Stratix IV Device Handbook Volume 1 101 Innovation Drive San Jose, CA 95134 www.altera.com SIV5V1-4.6 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as ISO trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and 9001:2008 services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service Registered described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 Contents Chapter Revision Dates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Section I. Device Core Chapter 1. Overview for the Stratix IV Device Family Feature Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 Stratix IV GX Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3 Stratix IV E Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 Stratix IV GT Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5 Architecture Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6 High-Speed Transceiver Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6 Highest Aggregate Data Bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6 Wide Range of Protocol Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6 Diagnostic Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7 Signal Integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7 FPGA Fabric and I/O Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 Device Core Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 Embedded Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 Digital Signal Processing (DSP) Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 Clock Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9 I/O Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9 High-Speed Differential I/O with DPA and Soft-CDR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9 External Memory Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9 System Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10 Integrated Software Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-19 Ordering Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-19 Chapter 2. Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Logic Array Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 LAB Interconnects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4 LAB Control Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4 Adaptive Logic Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5 ALM Operating Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 Normal Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 Extended LUT Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Arithmetic Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12 Shared Arithmetic Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14 LUT-Register Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15 Register Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17 ALM Interconnects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18 Clear and Preset Logic Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18 LAB Power Management Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19 Chapter 3. TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 TriMatrix Memory Block Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 Parity Bit Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 iv Contents Byte Enable Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 Packed Mode Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5 Address Clock Enable Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6 Mixed Width Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7 Asynchronous Clear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7 Error Correction Code (ECC) Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8 Memory Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9 Single-Port RAM Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10 Simple Dual-Port Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11 True Dual-Port Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14 Shift-Register Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16 ROM Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 FIFO Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 Clocking Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 Independent Clock Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Input/Output Clock Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Read/Write Clock Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Single Clock Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Selecting TriMatrix Memory Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18 Conflict Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19 Read-During-Write Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19 Same-Port Read-During-Write Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19 Mixed-Port Read-During-Write Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21 Power-Up Conditions and Memory Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23 Power Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-24 Chapter 4. DSP Blocks in Stratix IV Devices Stratix IV DSP Block Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2 Stratix IV Simplified DSP Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4 Stratix IV Operational Modes Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8 Stratix IV DSP Block Resource Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9 Input Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10 Multiplier and First-Stage Adder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12 Pipeline Register Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13 Second-Stage Adder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13 Rounding and Saturation Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-14 Second Adder and Output Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-14 Stratix IV Operational Mode Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15 Independent Multiplier Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15 9-, 12-, and 18-Bit Multiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15 36-Bit Multiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-19 Double Multiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-20 Two-Multiplier Adder Sum Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22 18 x 18 Complex Multiply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24 Four-Multiplier Adder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26 High-Precision Multiplier Adder Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27 Multiply Accumulate Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29 Shift Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30 Rounding and Saturation Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-32 DSP Block Control Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-34 Software Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-35 Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Contents v Chapter 5. Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 Global Clock Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3 Regional Clock Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4 Periphery Clock Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6 Clock Sources Per Quadrant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9 Clock Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9 Clock Network Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10 Dedicated Clock Input Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10 LABs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10 PLL Clock Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11 Clock Input Connections to the PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12 Clock Output Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13 Clock Control Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14 Clock Enable Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-17 Clock Source Control for PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18 Cascading PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19 PLLs in Stratix IV Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19 Stratix IV PLL Hardware Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23 PLL Clock I/O Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-24 PLL Control Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27 pfdena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27 areset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27 locked . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27 Clock Feedback Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-28 Source Synchronous Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-29 Source-Synchronous Mode for LVDS Compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-30 No-Compensation Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-30 Normal Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-31 Zero-Delay Buffer (ZDB) Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-31 External Feedback Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-32 Clock Multiplication and Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-33 Post-Scale Counter Cascading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-34 Programmable Duty Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-35 Programmable Phase Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-35 Programmable Bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-37 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-37 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-38 Spread-Spectrum Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-39 Clock Switchover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-39 Automatic Clock Switchover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-40 Manual Clock Switchover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43 Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43 PLL Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-44 PLL Reconfiguration Hardware Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-45 Post-Scale Counters (C0 to C9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-47 Scan Chain Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-48 Charge Pump and Loop Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-50 Bypassing a PLL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-51 Dynamic Phase-Shifting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-51 PLL Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-54 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 vi Contents Section II. I/O Interfaces Chapter 6. I/O Features in Stratix IV Devices I/O Standards Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 I/O Standards and Voltage Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3 I/O Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5 Modular I/O Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8 I/O Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-17 3.3-V I/O Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-19 External Memory Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-19 High-Speed Differential I/O with DPA Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-20 Programmable Current Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-20 Programmable Slew Rate Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21 Programmable I/O Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22 Programmable IOE Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22 Programmable Output Buffer Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22 Open-Drain Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22 Bus Hold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22 Programmable Pull-Up Resistor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23 Programmable Pre-Emphasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23 Programmable Differential Output Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23 MultiVolt I/O Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23 On-Chip Termination Support and I/O Termination Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24 On-Chip Series (RS) Termination Without Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-25 On-Chip Series Termination with Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26 Left-Shift Series Termination Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27 On-Chip Parallel Termination with Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-28 Expanded On-Chip Series Termination with Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-29 Dynamic On-Chip Termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-29 LVDS Input OCT (RD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-31 Summary of OCT Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-31 OCT Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-32 OCT Calibration Block Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-32 Sharing an OCT Calibration Block on Multiple I/O Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-34 OCT Calibration Block Modes of Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-35 Power-Up Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-35 User Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-36 OCT Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-37 Serial Data Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-37 Example of Using Multiple OCT Calibration Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38 RS Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38 Termination Schemes for I/O Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38 Single-Ended I/O Standards Termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38 Differential I/O Standards Termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-41 LVDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-43 Differential LVPECL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-44 RSDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-45 Mini-LVDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-46 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-46 I/O Bank Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-46 Non-Voltage-Referenced Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-47 Voltage-Referenced Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-47 Mixing Voltage-Referenced and Non-Voltage-Referenced Standards . . . . . . . . . . . . . . . . . . . . . 6-47 Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Contents vii Chapter 7. External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3 Using the RUP and RDN Pins in a DQS/DQ Group Used for Memory Interfaces . . . . . . . . . . . . . . 7-26 Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface . . . . . . . . . . . 7-26 Rules to Combine Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-27 Stratix IV External Memory Interface Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29 DQS Phase-Shift Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29 DLL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-31 Phase Offset Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-41 DQS Logic Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-43 DQS Delay Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-44 Update Enable Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-44 DQS Postamble Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-45 Leveling Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-46 Dynamic On-Chip Termination Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-48 I/O Element Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-49 Delay Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-52 I/O Configuration Block and DQS Configuration Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-54 Chapter 8. High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1 Locations of the I/O Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3 LVDS Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-4 LVDS SERDES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-8 ALTLVDS Port List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9 Differential Transmitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-11 Programmable VOD and Programmable Pre-Emphasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-14 Programmable VOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-15 Programmable Pre-Emphasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-16 Differential Receiver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-17 Differential I/O Termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-18 Receiver Hardware Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-19 DPA Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-19 Synchronizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20 Data Realignment Block (Bit Slip) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20 Deserializer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22 Receiver Data Path Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22 Non-DPA Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22 DPA Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-24 Soft-CDR Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25 LVDS Interface with the Use External PLL Option Enabled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26 Left and Right PLLs (PLL_Lx and PLL_Rx) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-29 Stratix IV Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-30 Source-Synchronous Timing Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-31 Differential Data Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-31 Differential I/O Bit Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-31 Transmitter Channel-to-Channel Skew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-33 Receiver Skew Margin for Non-DPA Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-33 Differential Pin Placement Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38 Guidelines for DPA-Enabled Differential Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38 DPA-Enabled Channels and Single-Ended I/Os . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38 DPA-Enabled Channel Driving Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38 Using Corner and Center Left and Right PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 viii Contents Using Both Center Left and Right PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-40 Guidelines for DPA-Disabled Differential Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42 DPA-Disabled Channels and Single-Ended I/Os . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42 DPA-Disabled Channel Driving Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42 Using Corner and Center Left and Right PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42 Using Both Center Left and Right PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-45 Section III. System Integration Chapter 9. Hot Socketing and Power-On Reset in Stratix IV Devices Stratix IV Hot-Socketing Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1 Stratix IV Devices can be Driven Before Power Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2 I/O Pins Remain Tri-Stated During Power Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2 Insertion or Removal of a Stratix IV Device from a Powered-Up System . . . . . . . . . . . . . . . . . . . . . . 9-2 Hot-Socketing Feature Implementation in Stratix IV Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3 Power-On Reset Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-4 Power-On Reset Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5 Chapter 10. Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1 Configuration Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2 Configuration Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2 Configuration Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-4 Power-On Reset Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5 VCCPGM Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5 VCCPD Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5 Fast Passive Parallel Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6 FPP Configuration Using a MAX II Device as an External Host . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6 FPP Configuration Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12 FPP Configuration Using a Microprocessor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16 Fast Active Serial Configuration (Serial Configuration Devices) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16 Estimating Active Serial Configuration Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22 Programming Serial Configuration Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-23 Guidelines for Connecting Serial Configuration Devices on an AS Interface . . . . . . . . . . . . . . . . . 10-25 Passive Serial Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-25 PS Configuration Using a MAX II Device as an External Host . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-26 PS Configuration Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-31 PS Configuration Using a Microprocessor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-32 PS Configuration Using a Download Cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-32 JTAG Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-35 Jam STAPL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-40 Device Configuration Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-40 Configuration Data Decompression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-48 Remote System Upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-50 Functional Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-51 Enabling Remote Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-53 Configuration Image Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-54 Remote System Upgrade Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-54 Remote Update Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-54 Dedicated Remote System Upgrade Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-57 Remote System Upgrade Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-58 Remote System Upgrade Control Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-58 Remote System Upgrade Status Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-59 Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Contents ix Remote System Upgrade State Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-60 User Watchdog Timer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-62 Quartus II Software Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-63 ALTREMOTE_UPDATE Megafunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-63 Design Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-64 Stratix IV Security Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 Security Against Copying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 Security Against Reverse Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 Security Against Tampering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 AES Decryption Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 Flexible Security Key Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65 Stratix IV Design Security Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-66 Security Modes Available . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67 Volatile Key . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67 Non-Volatile Key . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67 Non-Volatile Key with Tamper Protection Bit Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67 No Key Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-68 Supported Configuration Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-68 Chapter 11. SEU Mitigation in Stratix IV Devices Error Detection Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2 Configuration Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2 User Mode Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2 Automated Single-Event Upset Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5 Error Detection Pin Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5 CRC_ERROR Pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5 Error Detection Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6 Error Detection Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-7 Error Detection Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8 Software Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-10 Recovering From CRC Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-11 Chapter 12. JTAG Boundary-Scan Testing in Stratix IV Devices BST Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1 BST Operation Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-2 I/O Voltage Support in a JTAG Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4 BST Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4 BSDL Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4 Chapter 13. Power Management in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1 Stratix IV Power Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-2 Programmable Power Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-2 Stratix IV External Power Supply Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3 Temperature Sensing Diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4 External Pin Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4 Additional Information About this Handbook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 How to Contact Altera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 Typographic Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 x Stratix IV Device Handbook Volume 1 Contents September 2012 Altera Corporation Chapter Revision Dates The chapters in this document, Stratix IV Device Handbook, were revised on the following dates. Where chapters or groups of chapters are available separately, part numbers are listed. Chapter 1. Overview for the Stratix IV Device Family Revised: September 2012 Part Number: SIV51001-3.4 Chapter 2. Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Revised: February 2011 Part Number: SIV51002-3.1 Chapter 3. TriMatrix Embedded Memory Blocks in Stratix IV Devices Revised: December 2011 Part Number: SIV51003-3.3 Chapter 4. DSP Blocks in Stratix IV Devices Revised: February 2011 Part Number: SIV51004-3.1 Chapter 5. Clock Networks and PLLs in Stratix IV Devices Revised: September 2012 Part Number: SIV51005-3.4 Chapter 6. I/O Features in Stratix IV Devices Revised: September 2012 Part Number: SIV51006-3.4 Chapter 7. External Memory Interfaces in Stratix IV Devices Revised: February 2011 Part Number: SIV51007-3.2 Chapter 8. High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Revised: September 2012 Part Number: SIV51008-3.4 Chapter 9. Hot Socketing and Power-On Reset in Stratix IV Devices Revised: February 2011 Part Number: SIV51009-3.2 Chapter 10. Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Revised: September 2012 Part Number: SIV51010-3.5 Chapter 11. SEU Mitigation in Stratix IV Devices Revised: February 2011 Part Number: SIV51011-3.2 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 xii Chapter Revision Dates Chapter 12. JTAG Boundary-Scan Testing in Stratix IV Devices Revised: February 2011 Part Number: SIV51012-3.2 Chapter 13. Power Management in Stratix IV Devices Revised: February 2011 Part Number: SIV51013-3.2 Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Section I. Device Core This section provides a complete overview of all features relating to the Stratix(R) IV device family, which is the most architecturally advanced, high-performance, low-power FPGA in the market place. This section includes the following chapters: Chapter 1, Overview for the Stratix IV Device Family Chapter 2, Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Chapter 3, TriMatrix Embedded Memory Blocks in Stratix IV Devices Chapter 4, DSP Blocks in Stratix IV Devices Chapter 5, Clock Networks and PLLs in Stratix IV Devices Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 I-2 Stratix IV Device Handbook Volume 1 Section I: Device Core September 2012 Altera Corporation 1. Overview for the Stratix IV Device Family September 2012 SIV51001-3.4 SIV51001-3.4 Altera(R) Stratix(R) IV FPGAs deliver a breakthrough level of system bandwidth and power efficiency for high-end applications, allowing you to innovate without compromise. Stratix IV FPGAs are based on the Taiwan Semiconductor Manufacturing Company (TSMC) 40-nm process technology and surpass all other high-end FPGAs, with the highest logic density, most transceivers, and lowest power requirements. The Stratix IV device family contains three optimized variants to meet different application requirements: Stratix IV E (Enhanced) FPGAs--up to 813,050 logic elements (LEs), 33,294 kilobits (Kb) RAM, and 1,288 18 x 18 bit multipliers Stratix IV GX transceiver FPGAs--up to 531,200 LEs, 27,376 Kb RAM, 1,288 18 x 18-bit multipliers, and 48 full-duplex clock data recovery (CDR)-based transceivers at up to 8.5 Gbps Stratix IV GT--up to 531,200 LEs, 27,376 Kb RAM, 1,288 18 x 18-bit multipliers, and 48 full-duplex CDR-based transceivers at up to 11.3 Gbps The complete Altera high-end solution includes the lowest risk, lowest total cost path to volume using HardCopy(R) IV ASICs for all the family variants, a comprehensive portfolio of application solutions customized for end-markets, and the industry leading Quartus(R) II software to increase productivity and performance. f For information about upcoming Stratix IV device features, refer to the Upcoming Stratix IV Device Features document. f For information about changes to the currently published Stratix IV Device Handbook, refer to the Addendum to the Stratix IV Device Handbook chapter. This chapter contains the following sections: "Feature Summary" on page 1-2 "Architecture Features" on page 1-6 "Integrated Software Platform" on page 1-19 "Ordering Information" on page 1-19 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 September 2012 Feedback Subscribe 1-2 Chapter 1: Overview for the Stratix IV Device Family Feature Summary Feature Summary The following list summarizes the Stratix IV device family features: Up to 48 full-duplex CDR-based transceivers in Stratix IV GX and GT devices supporting data rates up to 8.5 Gbps and 11.3 Gbps, respectively Dedicated circuitry to support physical layer functionality for popular serial protocols, such as PCI Express (PCIe) (PIPE) Gen1 and Gen2, Gbps Ethernet (GbE), Serial RapidIO, SONET/SDH, XAUI/HiGig, (OIF) CEI-6G, SD/HD/3G-SDI, Fibre Channel, SFI-5, and Interlaken Complete PCIe protocol solution with embedded PCIe hard IP blocks that implement PHY-MAC layer, Data Link layer, and Transaction layer functionality f For more information, refer to the IP Compiler for PCI Express User Guide. Stratix IV Device Handbook Volume 1 Programmable transmitter pre-emphasis and receiver equalization circuitry to compensate for frequency-dependent losses in the physical medium Typical physical medium attachment (PMA) power consumption of 100 mW at 3.125 Gbps and 135 mW at 6.375 Gbps per channel 72,600 to 813,050 equivalent LEs per device 7,370 to 33,294 Kb of enhanced TriMatrix memory consisting of three RAM block sizes to implement true dual-port memory and FIFO buffers High-speed digital signal processing (DSP) blocks configurable as 9 x 9-bit, 12 x 12-bit, 18 x 18-bit, and 36 x 36-bit full-precision multipliers at up to 600 MHz Up to 16 global clocks (GCLK), 88 regional clocks (RCLK), and 132 periphery clocks (PCLK) per device Programmable power technology that minimizes power while maximizing device performance Up to 1,120 user I/O pins arranged in 24 modular I/O banks that support a wide range of single-ended and differential I/O standards Support for high-speed external memory interfaces including DDR, DDR2, DDR3 SDRAM, RLDRAM II, QDR II, and QDR II+ SRAM on up to 24 modular I/O banks High-speed LVDS I/O support with serializer/deserializer (SERDES), dynamic phase alignment (DPA), and soft-CDR circuitry at data rates up to 1.6 Gbps Support for source-synchronous bus standards, including SGMII, GbE, SPI-4 Phase 2 (POS-PHY Level 4), SFI-4.1, XSBI, UTOPIA IV, NPSI, and CSIX-L1 Pinouts for Stratix IV E devices designed to allow migration of designs from Stratix III to Stratix IV E with minimal PCB impact September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Feature Summary 1-3 Stratix IV GX Devices Stratix IV GX devices provide up to 48 full-duplex CDR-based transceiver channels per device: Thirty-two out of the 48 transceiver channels have dedicated physical coding sublayer (PCS) and physical medium attachment (PMA) circuitry and support data rates between 600 Mbps and 8.5 Gbps The remaining 16 transceiver channels have dedicated PMA-only circuitry and support data rates between 600 Mbps and 6.5 Gbps 1 The actual number of transceiver channels per device varies with device selection. For more information about the exact transceiver count in each device, refer to Table 1-1 on page 1-11. 1 For more information about transceiver architecture, refer to the Transceiver Architecture in Stratix IV Devices chapter. Figure 1-1 shows a high-level Stratix IV GX chip view. Figure 1-1. Stratix IV GX Chip View (1) PLL General Purpose I/O and Memory Interface PLL PLL General Purpose I/O and Memory Interface Transceiver Block General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL PLL PLL PCI Express Hard IP Block PLL PCI Express Hard IP Block General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL FPGA Fabric (Logic Elements, DSP, Embedded Memory, Clock Networks) General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PCI Express Hard IP Block PLL Transceiver Transceiver Transceiver Transceiver Block Block Block Block PLL PLL PCI Express Hard IP Block Transceiver Transceiver Transceiver Transceiver Block Block Block Block General Purpose I/O and Memory Interface General Purpose I/O and Memory Interface 600 Mbps-8.5 Gbps CDR-based Transceiver General Purpose I/O and 150 Mbps-1.6 Gbps LVDS interface with DPA and Soft-CDR Note to Figure 1-1: (1) Resource counts vary with device selection, package selection, or both. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 1-4 Chapter 1: Overview for the Stratix IV Device Family Feature Summary Stratix IV E Device Stratix IV E devices provide an excellent solution for applications that do not require high-speed CDR-based transceivers, but are logic, user I/O, or memory intensive. Figure 1-2 shows a high-level Stratix IV E chip view. Figure 1-2. Stratix IV E Chip View (1) General Purpose I/O and Memory PLL Interface PLL General Purpose I/O and Memory Interface PLL PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft-CDR General Purpose I/O and High-Speed LVDS I/O with DPA and Soft-CDR PLL PLL FPGA Fabric PLL (Logic Elements, DSP, Embedded Memory, Clock Networks) PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft-CDR General Purpose I/O and High-Speed LVDS I/O with DPA and Soft-CDR PLL PLL General Purpose I/O and Memory PLL Interface General Purpose I/O and High-Speed LVDS I/O with DPA and Soft-CDR PLL General Purpose I/O and Memory Interface General Purpose I/O and 150 Mbps-1.6 Gbps LVDS interface with DPA and Soft-CDR Note to Figure 1-2: (1) Resource counts vary with device selection, package selection, or both. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Feature Summary 1-5 Stratix IV GT Devices Stratix IV GT devices provide up to 48 CDR-based transceiver channels per device: Thirty-two out of the 48 transceiver channels have dedicated PCS and PMA circuitry and support data rates between 600 Mbps and 11.3 Gbps The remaining 16 transceiver channels have dedicated PMA-only circuitry and support data rates between 600 Mbps and 6.5 Gbps 1 The actual number of transceiver channels per device varies with device selection. For more information about the exact transceiver count in each device, refer to Table 1-7 on page 1-16. 1 For more information about Stratix IV GT devices and transceiver architecture, refer to the Transceiver Architecture in Stratix IV Devices chapter. Figure 1-3 shows a high-level Stratix IV GT chip view. (1) FPGA Fabric PLL (Logic Elements, DSP, Embedded Memory, Clock Networks) PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PCI Express Hard IP Block PLL Transceiver Transceiver Block Block General Purpose I/O and Memory Interface Transceiver Transceiver Block Block PLL PCI Express Hard IP Block PLL PLL PCI Express Hard IP Block Transceiver Transceiver Block Block Transceiver Transceiver Block Block General Purpose I/O and Memory Interface PCI Express Hard IP Block Figure 1-3. Stratix IV GT Chip View PLL General Purpose I/O and Memory Interface Transceiver Block General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL PLL PLL General Purpose I/O and Memory Interface 600 Mbps-11.3 Gbps CDR-based Transceiver General Purpose I/O and up to 1.6 Gbps LVDS interface with DPA and Soft-CDR Note to Figure 1-3: (1) Resource counts vary with device selection, package selection, or both. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 1-6 Chapter 1: Overview for the Stratix IV Device Family Architecture Features Architecture Features The Stratix IV device family features are divided into high-speed transceiver features and FPGA fabric and I/O features. 1 The high-speed transceiver features apply only to Stratix IV GX and Stratix IV GT devices. High-Speed Transceiver Features The following sections describe high-speed transceiver features for Stratix IV GX and GT devices. Highest Aggregate Data Bandwidth Up to 48 full-duplex transceiver channels supporting data rates up to 8.5 Gbps in Stratix IV GX devices and up to 11.3 Gbps in Stratix IV GT devices. Wide Range of Protocol Support Physical layer support for the following serial protocols: Stratix IV GX--PCIe Gen1 and Gen2, GbE, Serial RapidIO, SONET/SDH, XAUI/HiGig, (OIF) CEI-6G, SD/HD/3G-SDI, Fibre Channel, SFI-5, GPON, SAS/SATA, HyperTransport 1.0 and 3.0, and Interlaken Stratix IV GT--40G/100G Ethernet, SFI-S, Interlaken, SFI-5.1, Serial RapidIO, SONET/SDH, XAUI/HiGig, (OIF) CEI-6G, 3G-SDI, and Fibre Channel Extremely flexible and easy-to-configure transceiver data path to implement proprietary protocols PCIe Support Complete PCIe Gen1 and Gen2 protocol stack solution compliant to PCI Express base specification 2.0 that includes PHY-MAC, Data Link, and transaction layer circuitry embedded in PCI Express hard IP blocks f For more information, refer to the PCI Express Compiler User Guide. Stratix IV Device Handbook Volume 1 Root complex and end-point applications x1, x4, and x8 lane configurations PIPE 2.0-compliant interface Embedded circuitry to switch between Gen1 and Gen2 data rates Built-in circuitry for electrical idle generation and detection, receiver detect, power state transitions, lane reversal, and polarity inversion 8B/10B encoder and decoder, receiver synchronization state machine, and 300 parts per million (ppm) clock compensation circuitry Transaction layer support for up to two virtual channels (VCs) September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Architecture Features 1-7 XAUI/HiGig Support Compliant to IEEE802.3ae specification Embedded state machine circuitry to convert XGMII idle code groups (||I||) to and from idle ordered sets (||A||, ||K||, ||R||) at the transmitter and receiver, respectively 8B/10B encoder and decoder, receiver synchronization state machine, lane deskew, and 100 ppm clock compensation circuitry GbE Support Compliant to IEEE802.3-2005 specification Automatic idle ordered set (/I1/, /I2/) generation at the transmitter, depending on the current running disparity 8B/10B encoder and decoder, receiver synchronization state machine, and 100 ppm clock compensation circuitry Support for other protocol features such as MSB-to-LSB transmission in SONET/SDH configuration and spread-spectrum clocking in PCIe configurations Diagnostic Features Serial loopback from the transmitter serializer to the receiver CDR for transceiver PCS and PMA diagnostics Reverse serial loopback pre- and post-CDR to transmitter buffer for physical link diagnostics Loopback master and slave capability in PCI Express hard IP blocks f For more information, refer to the PCI Express Compiler User Guide. Signal Integrity Stratix IV devices simplify the challenge of signal integrity through a number of chip, package, and board-level enhancements to enable efficient high-speed data transfer into and out of the device. These enhancements include: September 2012 Programmable 3-tap transmitter pre-emphasis with up to 8,192 pre-emphasis levels to compensate for pre-cursor and post-cursor inter-symbol interference (ISI) Up to 900% boost capability on the first pre-emphasis post-tap User-controlled and adaptive 4-stage receiver equalization with up to 16 dB of high-frequency gain On-die power supply regulators for transmitter and receiver phase-locked loop (PLL) charge pump and voltage controlled oscillator (VCO) for superior noise immunity On-package and on-chip power supply decoupling to satisfy transient current requirements at higher frequencies, thereby reducing the need for on-board decoupling capacitors Calibration circuitry for transmitter and receiver on-chip termination (OCT) resistors Altera Corporation Stratix IV Device Handbook Volume 1 1-8 Chapter 1: Overview for the Stratix IV Device Family Architecture Features FPGA Fabric and I/O Features The following sections describe the Stratix IV FPGA fabric and I/O features. Device Core Features Up to 531,200 LEs in Stratix IV GX and GT devices and up to 813,050 LEs in Stratix IV E devices, efficiently packed in unique and innovative adaptive logic modules (ALMs) Ten ALMs per logic array block (LAB) deliver faster performance, improved logic utilization, and optimized routing Programmable power technology, including a variety of process, circuit, and architecture optimizations and innovations Programmable power technology available to select power-driven compilation options for reduced static power consumption Embedded Memory TriMatrix embedded memory architecture provides three different memory block sizes to efficiently address the needs of diversified FPGA designs: 640-bit MLAB 9-Kb M9K 144-Kb M144K Up to 33,294 Kb of embedded memory operating at up to 600 MHz Each memory block is independently configurable to be a single- or dual-port RAM, FIFO, ROM, or shift register Digital Signal Processing (DSP) Blocks Flexible DSP blocks configurable as 9 x 9-bit, 12 x 12-bit, 18 x 18-bit, and 36 x 36-bit full-precision multipliers at up to 600 MHz with rounding and saturation capabilities Faster operation due to fully pipelined architecture and built-in addition, subtraction, and accumulation units to combine multiplication results Optimally designed to support advanced features such as adaptive filtering, barrel shifters, and finite and infinite impulse response (FIR and IIR) filters Clock Networks Stratix IV Device Handbook Volume 1 Up to 16 global clocks and 88 regional clocks optimally routed to meet the maximum performance of 800 MHz Up to 112 and 132 periphery clocks in Stratix IV GX and Stratix IV E devices, respectively Up to 66 (16 GCLK + 22 RCLK + 28 PCLK) clock networks per device quadrant in Stratix IV GX and Stratix IV GT devices Up to 71 (16 GCLK + 22 RCLK + 33 PCLK) clock networks per device quadrant in Stratix IV E devices September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Architecture Features 1-9 PLLs Three to 12 PLLs per device supporting spread-spectrum input tracking, programmable bandwidth, clock switchover, dynamic reconfiguration, and delay compensation On-chip PLL power supply regulators to minimize noise coupling I/O Features Sixteen to 24 modular I/O banks per device with 24 to 48 I/Os per bank designed and packaged for optimal simultaneous switching noise (SSN) performance and migration capability Support for a wide range of industry I/O standards, including single-ended (LVTTL/CMOS/PCI/PCIX), differential (LVDS/mini-LVDS/RSDS), voltage-referenced single-ended and differential (SSTL/HSTL Class I/II) I/O standards On-chip series (RS) and on-chip parallel (RT) termination with auto-calibration for single-ended I/Os and on-chip differential (RD) termination for differential I/Os Programmable output drive strength, slew rate control, bus hold, and weak pull-up capability for single-ended I/Os User I/O:GND:VCC ratio of 8:1:1 to reduce loop inductance in the package--PCB interface Programmable transmitter differential output voltage (VOD) and pre-emphasis for high-speed LVDS I/O High-Speed Differential I/O with DPA and Soft-CDR Dedicated circuitry on the left and right sides of the device to support differential links at data rates from 150 Mbps to 1.6 Gbps Up to 98 differential SERDES in Stratix IV GX devices, up to 132 differential SERDES in Stratix IV E devices, and up to 47 differential SERDES in Stratix IV GT devices DPA circuitry at the receiver automatically compensates for channel-to-channel and channel-to-clock skew in source synchronous interfaces Soft-CDR circuitry at the receiver allows implementation of asynchronous serial interfaces with embedded clocks at up to 1.6 Gbps data rate (SGMII and GbE) External Memory Interfaces September 2012 Support for existing and emerging memory interface standards such as DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM, QDRII SRAM, QDRII+ SRAM, and RLDRAM II DDR3 up to 1,067 Mbps/533 MHz Programmable DQ group widths of 4 to 36 bits (includes parity bits) Dynamic OCT, trace mismatch compensation, read-write leveling, and half-rate register capabilities provide a robust external memory interface solution Altera Corporation Stratix IV Device Handbook Volume 1 1-10 Chapter 1: Overview for the Stratix IV Device Family Architecture Features System Integration All Stratix IV devices support hot socketing Four configuration modes: Passive Serial (PS) Fast Passive Parallel (FPP) Fast Active Serial (FAS) JTAG configuration Ability to perform remote system upgrades 256-bit advanced encryption standard (AES) encryption of configuration bits protects your design against copying, reverse engineering, and tampering Built-in soft error detection for configuration RAM cells f For more information about how to connect the PLL, external memory interfaces, I/O, high-speed differential I/O, power, and the JTAG pins to PCB, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines and the Stratix IV GT Device Family Pin Connection Guidelines. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Table 1-1. Stratix IV GX Device Features (Part 1 of 2) F1932 F1760 F1932 EP4SGX530 F1760 F1517 F1152 F780 F1932 EP4SGX360 F1760 F1517 F780 F1152 EP4SGX290 F1517 F780 F1152 EP4SGX230 F1517 F780 F1152 EP4SGX180 F1152 EP4SGX110 F780 F1152 Altera Corporation Package Option EP4SGX70 F780 Feature ALMs 29,040 42,240 70,300 91,200 116,480 141,440 212,480 LEs 72,600 105,600 175,750 228,000 291,200 353,600 531,200 0.6 Gbps8.5 Gbps Transceivers (PMA + PCS) -- 16 -- -- 16 -- -- 16 24 -- -- 16 24 -- -- 16 24 24 32 -- -- 16 24 24 32 24 32 8 -- 8 16 -- 8 16 -- -- 8 16 -- -- 16 16 -- -- -- -- 16 16 -- -- -- -- -- -- PMA-only CMU Channels (0.6 Gbps6.5 Gbps) -- 8 -- -- 8 -- -- 8 12 -- -- 8 12 -- -- 8 12 12 16 -- -- 8 12 12 16 12 16 PCI Express hard IP Blocks 1 2 1 28 56 28 (1) 0.6 Gbps6.5 Gbps Transceivers (PMA + PCS) Chapter 1: Overview for the Stratix IV Device Family Architecture Features September 2012 Table 1-1 lists the Stratix IV GX device features. (1) High-Speed LVDS SERDES (up to 1.6 Gbps) 2 28 1 56 2 1 2 2 28 44 88 28 44 88 -- 44 1 2 4 1 2 4 -- 2 4 88 88 98 2 -- 44 -- 2 4 4 88 88 98 88 98 (4) 1 1 4 4 4 1-11 Stratix IV Device Handbook Volume 1 SPI-4.2 Links F1932 F1760 F1932 EP4SGX530 F1760 F1517 F1152 F780 F1932 EP4SGX360 F1760 F1517 F780 F1152 EP4SGX290 F1517 F780 F1152 EP4SGX230 F1517 F1152 EP4SGX180 F780 F780 F1152 EP4SGX110 F1152 Package Option EP4SGX70 F780 Feature 1-12 Stratix IV Device Handbook Volume 1 Table 1-1. Stratix IV GX Device Features (Part 2 of 2) M9K Blocks (256 x 36 bits) 462 660 950 1,235 936 1,248 1,280 M144K Blocks (2048 x 72 bits) 16 16 20 22 36 48 64 7,370 9,564 13,627 17,133 17,248 22,564 27,376 384 512 920 1,288 832 Total Memory (MLAB+M9K +M144K) Kb Embedded Multipliers 18 x 18 (2) PLLs 4 3 (3) 372 488 372 372 48 8 372 56 4 Speed Grade (fastest to slowest) (5) -2, -3, -4 -2, -3, -4 -2 , -3, -4 -2 , -3, -4 -2, -3, -4 -2 , -3, -4 -2 , -3, -4 User I/Os 4 3 6 8 3 6 8 4 56 4 74 4 372 564 74 4 289 564 56 4 -2 , -3 , -4 -2, -3, -4 -2 , -3, -4 -2 -2 -2, -2, , , -3, -3, -3, -3, -4 -4 -4 -4 -2 , -3, -4 56 4 6 8 12 12 4 6 74 4 88 0 92 0 289 564 -2, -2, -2, -2, -3, -3, -3, -3, -4 -4 -4 -4 -2 , -3, -4 1,024 8 12 12 12 12 74 4 88 0 920 880 920 -2 -2, -2, -2, , -3, -3, -3, -3, -4 -4 -4 -4 -2, -3, -4 -2, -3, -4 -2, -3, -4 56 4 Notes to Table 1-1: (1) The total number of transceivers is divided equally between the left and right side of each device, except for the devices in the F780 package. These devices have eight transceiver channels located only on the right side of the device. September 2012 Altera Corporation (2) Four multiplier adder mode. (3) The user I/Os count from pin-out files includes all general purpose I/O, dedicated clock pins, and dual purpose configuration pins. Transceiver pins and dedicated configuration pins are not included in the pin count. (4) Total pairs of high-speed LVDS SERDES take the lowest channel count of RX/TX. (5) The difference between the Stratix IV GX devices in the -2 and -2x speed grades is the number of available transceiver channels. The -2 device allows you to use the transceiver CMU blocks as transceiver channels. The -2x device does NOT allow you to use the CMU blocks as transceiver channels. In addition to the reduction of available transceiver channels in the Stratix IV GX -2x device, the data rates in the -2x device are limited to 6.5 Gbps. Chapter 1: Overview for the Stratix IV Device Family Architecture Features 3 1,02 4 1,040 Table 1-2. Stratix IV GX Device Package Options Altera Corporation F780 (29 mm x 29 mm) Device (1), (2) F1152 (35 mm x 35 mm) (6) (6) F1152 (35 mm x 35 mm) (5), (7) F1517 (40 mm x 40 mm) (5), (7) F1760 (42.5 mm x 42.5 mm) F1932 (45 mm x 45 mm) (7) (7) EP4SGX70 DF29 -- -- HF35 -- -- -- -- EP4SGX110 DF29 -- FF35 HF35 -- -- -- -- EP4SGX180 DF29 -- FF35 -- HF35 KF40 -- -- EP4SGX230 DF29 -- EP4SGX290 -- FF35 -- HF35 KF40 -- -- FH29 (3) FF35 -- HF35 KF40 KF43 NF45 (3) FF35 -- -- -- EP4SGX360 -- FH29 EP4SGX530 -- -- HF35 HH35 KF40 (4) KH40 (4) KF43 NF45 KF43 NF45 Chapter 1: Overview for the Stratix IV Device Family Architecture Features September 2012 Table 1-2 lists the Stratix IV GX device package options. Notes to Table 1-2: (1) Device packages in the same column and marked under the same arrow sign have vertical migration capability. (2) Use the Pin Migration Viewer in the Pin Planner to verify the pin migration compatibility when migrating devices. For more information, refer to I/O Management in the Quartus II Handbook, Volume 2. (3) The 780-pin EP4SGX290 and EP4SGX360 devices are available only in 33 mm x 33 mm Hybrid flip chip package. (4) The 1152-pin and 1517-pin EP4SGX530 devices are available only in 42.5 mm x 42.5 mm Hybrid flip chip packages. (5) When migrating between hybrid and flip chip packages, there is an additional keep-out area. For more information, refer to the Package Information Datasheet for Altera Devices. (6) Devices listed in this column are available in -2x, -3, and -4 speed grades. These devices do not have on-package decoupling capacitors. (7) Devices listed in this column are available in -2, -3, and -4 speed grades. These devices have on-package decoupling capacitors. For more information about on-package decoupling capacitor value in each device, refer to Table 1-3. 1 On-package decoupling reduces the need for on-board or PCB decoupling capacitors by satisfying the transient current requirements at higher frequencies. The Power Delivery Network design tool for Stratix IV devices accounts for the on-package decoupling and reflects the reduced requirements for PCB decoupling capacitors. 1-13 Stratix IV Device Handbook Volume 1 1-14 Stratix IV Device Handbook Volume 1 Table 1-3 lists the Stratix IV GX device on-package decoupling information. Table 1-3. Stratix IV GX Device On-Package Decoupling Information (1) Ordering Information EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 HF35 HF35 HF35 KF40 HF35 KF40 VCC VCCL_GXB VCCA_L/R VCCT and VCCR (Shared) 10nF per bank (2) 100nF per transceiver block 100nF 1470nF + 147nF per side 21uF + 2470nF 10nF per bank (2) 100nF per transceiver block 100nF 1470nF + 147nF per side 21uF + 2470nF 10nF per bank (2) 100nF per transceiver block 100nF 1470nF + 147nF per side 21 uF + 2470 nF 10 nF per bank (2) 100 nF per transceiver block 100 nF 1470 nF + 147 nF per side 41 uF + 4470 nF 10 nF per bank (2) 100 nF per transceiver block 100nF 1470 nF + 147 nF per side 41 uF + 4470 nF 10 nF per bank (2) 100 nF per transceiver block 100 nF 1470 nF + 147 nF per side 41 uF + 4470 nF 10 nF per bank (2) 100 nF per transceiver block 100 nF 1470 nF + 147 nF per side 21uF + 2470nF VCCIO HF35 EP4SGX290 KF40 KF43 NF45 HF35 EP4SGX360 KF40 KF43 NF45 HH35 KH40 KF43 NF45 Notes to Table 1-3: September 2012 Altera Corporation (1) Table 1-3 refers to production devices on-package decoupling. For more information about decoupling design of engineering sample (ES) devices, contact Altera Technical Support. (2) For I/O banks 3(*), 4(*), 7(*), and 8(*) only. There is no OPD for I/O bank 1(*), 2(*), 5(*), and 6(*). Chapter 1: Overview for the Stratix IV Device Family Architecture Features EP4SGX530 Table 1-4. Stratix IV E Device Features Feature EP4SE230 Altera Corporation Package Pin Count 780 EP4SE360 780 EP4SE530 1152 1152 EP4SE820 1517 1760 1152 1517 ALMs 91,200 141,440 212,480 325,220 LEs 228,000 353,600 531,200 813,050 High-Speed LVDS SERDES (up to 1.6 Gbps) (1) 56 56 88 88 SPI-4.2 Links 3 3 4 4 M9K Blocks (256 x 36 bits) 1,235 1,248 1,280 1610 M144K Blocks (2048 x 72 bits) 22 48 64 60 Total Memory (MLAB+M9K+ M144K) Kb 17,133 22,564 27,376 33,294 Embedded Multipliers (18 x 18) (2) 1,288 1,040 1,024 960 PLLs User I/Os 112 6 88 112 132 4 6 6 4 4 8 8 12 12 8 488 488 744 744 976 976 744 -2, -3, -4 -2, -3, -4 -2, -3, -4 -2, -3, -4 -2, -3, -4 -2, -3, -4 (3) Speed Grade (fastest to slowest) 112 1760 12 (4) -3, -4 976 Chapter 1: Overview for the Stratix IV Device Family Architecture Features September 2012 Table 1-4 lists the Stratix IV E device features. 12 (4) -3, -4 1120 (4) -3, -4 Notes to Table 1-4: (1) The user I/O count from the pin-out files include all general purpose I/Os, dedicated clock pins, and dual purpose configuration pins. Transceiver pins and dedicated configuration pins are not included in the pin count. (2) Four multiplier adder mode. (4) This data is preliminary. 1-15 Stratix IV Device Handbook Volume 1 (3) Total pairs of high-speed LVDS SERDES take the lowest channel count of RX/TX. 1-16 Chapter 1: Overview for the Stratix IV Device Family Architecture Features Table 1-5 summarizes the Stratix IV E device package options. Table 1-5. Stratix IV E Device Package Options (1), F780 (29 mm x 29 mm) Device EP4SE230 (5), (6) (2) F1152 (35 mm x 35 mm) F29 EP4SE360 H29 (3) F1517 (40 mm x 40 mm) (5), (7) (7) F1760 (42.5 mm x 42.5 mm) -- -- -- F35 -- -- EP4SE530 -- H35 (4) H40 (4) F43 EP4SE820 -- H35 (4) H40 (4) F43 (7) Notes to Table 1-5: (1) Device packages in the same column and marked under the same arrow sign have vertical migration capability. (2) Use the Pin Migration Viewer in the Pin Planner to verify the pin migration compatibility when migrating devices. For more information, refer to I/O Management in the Quartus II Handbook, Volume 2. (3) The 780-pin EP4SE360 device is available only in the 33 mm x 33 mm Hybrid flip chip package. (4) The 1152-pin and 1517-pin for EP4SE530 and EP4SE820 devices are available only in the 42.5 mm x 42.5 mm Hybrid flip chip package. (5) When migrating between hybrid and flip chip packages, there is an additional keep-out area. For more information, refer to the Package Information Datasheet for Altera Devices. (6) Devices listed in this column do not have on-package decoupling capacitors. (7) Devices listed in this column have on-package decoupling capacitors. For more information about on-package decoupling capacitor value for each device, refer to Table 1-6. Table 1-6 lists the Stratix IV E on-package decoupling information. Table 1-6. Stratix IV E Device On-Package Decoupling Information (1) Ordering Information EP4SE360 F35 VCC VCCIO 41 uF + 4470 nF 10 nF per bank 41 uF + 4470 nF 10 nF per bank 41 uF + 4470 nF 10 nF per bank H35 EP4SE530 H40 F43 H35 EP4SE820 H40 F43 Note to Table 1-6: (1) Table 1-6 refers to production devices on-package decoupling. For more information about decoupling design of engineering sample (ES) devices, contact Altera Technical Support. Table 1-7 lists the Stratix IV GT device features. Table 1-7. Stratix IV GT Device Features (Part 1 of 2) Feature EP4S40G5 EP4S100G2 EP4S100G3 EP4S100G4 1517 1517 1517 1932 1932 ALMs 91,200 212,480 91,200 116,480 141,440 212,480 LEs 228,000 531,200 228,000 291,200 353,600 531,200 36 36 36 48 48 Package Pin Count Total Transceiver Channels Stratix IV Device Handbook Volume 1 EP4S40G2 EP4S100G5 1517 36 1932 48 September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Architecture Features 1-17 Table 1-7. Stratix IV GT Device Features (Part 2 of 2) Feature EP4S40G2 EP4S40G5 EP4S100G2 EP4S100G3 EP4S100G4 EP4S100G5 10G Transceiver Channels (600 Mbps - 11.3 Gbps with PMA + PCS) 12 12 24 24 24 24 32 8G Transceiver Channels (600 Mbps - 8.5 Gbps with PMA + PCS) (1) 12 12 0 8 8 0 0 PMA-only CMU Channels (600 Mbps- 6.5 Gbps) 12 12 12 16 16 12 16 PCIe hard IP Blocks 2 2 2 4 4 2 4 High-Speed LVDS SERDES (up to 1.6 Gbps) (2) 46 46 46 47 47 46 47 SP1-4.2 Links 2 2 2 2 2 2 2 M9K Blocks (256 x 36 bits) 1,235 1,280 1,235 936 1,248 1,280 M144K Blocks (2048 x 72 bits) 22 64 22 36 48 64 Total Memory (MLAB + M9K + M144K) Kb 17,133 27,376 17,133 17,248 22,564 27,376 Embedded Multipliers 18 x 18 (3) 1,288 1,024 1,288 832 1,024 1,024 8 8 8 12 12 8 12 654 654 654 781 781 654 781 -1, -2, -3 -1, -2, -3 -1, -2, -3 -1, -2, -3 -1, -2, -3 -1, -2, -3 -1, -2, -3 PLLs User I/Os (4), (5) Speed Grade (fastest to slowest) Notes to Table 1-7: (1) You can configure all 10G transceiver channels as 8G transceiver channels. For example, the EP4S40G2F40 device has twenty-four 8G transceiver channels and the EP4S100G5F45 device has thirty-two 8G transceiver channels. (2) Total pairs of high-speed LVDS SERDES take the lowest channel count of RX/TX. (3) Four multiplier adder mode. (4) The user I/O count from the pin-out files include all general purpose I/Os, dedicated clock pins, and dual purpose configuration pins. Transceiver pins and dedicated configuration pins are not included in the pin count. (5) This data is preliminary. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 1-18 Chapter 1: Overview for the Stratix IV Device Family Architecture Features Table 1-8 lists the resource counts for the Stratix IV GT devices. Table 1-8. Stratix IV GT Device Package Options (1), (2) 1517 Pin (40 mm x 40 mm) Device 1932 Pin (45 mm x 45 mm) (3) Stratix IV GT 40 G Devices EP4S40G2 F40 EP4S40G5 -- (4), (5) H40 -- Stratix IV GT 100 G Devices EP4S100G2 F40 -- EP4S100G3 -- F45 EP4S100G4 -- F45 EP4S100G5 H40 (4), (5) F45 Notes to Table 1-8: (1) This table represents pin compatability; however, it does not include hard IP block placement compatability. (2) Devices under the same arrow sign have vertical migration capability. (3) When migrating between hybrid and flip chip packages, there is an additional keep-out area. For more information, refer to the Altera Device Package Information Data Sheet. (4) EP4S40G5 and EP4S100G5 devices with 1517 pin-count are only available in 42.5-mm x 42.5-mm Hybrid flip chip packages. (5) If you are using the hard IP block, migration is not possible. Table 1-9 lists the Stratix IV GT on-package decoupling information. Table 1-9. Stratix IV GT Device On-Package Decoupling Information (1) Ordering Information EP4S40G2F40 EP4S100G2F40 VCC VCCIO VCCL_GXB VCCA_L/R VCCT_L/R VCCR_L/R 21 uF + 2470 nF 10 nF per bank (2) 100 nF per transceiver block 100 nF 100 nF 100 nF 41 uF + 4470 nF 10 nF per bank (2) 100 nF per transceiver block 100 nF 100 nF 100 nF EP4S100G3F45 EP4S100G4F45 EP4S40G5H40 EP4S100G5H40 EP4S100G5F45 Notes to Table 1-9: (1) Table 1-9 refers to production devices on-package decoupling. For more information about decoupling design of engineering sample (ES) devices, contact Altera Technical Support. (2) For I/O banks 3(*), 4(*), 7(*), and 8(*) only. There is no OPD for I/O bank 1(*), 2(*), 5(*), and 6(*). Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Integrated Software Platform 1-19 Integrated Software Platform The Quartus II software provides an integrated environment for HDL and schematic design entry, compilation and logic synthesis, full simulation and advanced timing analysis, SignalTap II Logic Analyzer, and device configuration of Stratix IV designs. The Quartus II software provides the MegaWizard Plug-In Manager user interface to generate different functional blocks, such as memory, PLL, and digital signal processing logic. For transceivers, the Quartus II software provides the ALTGX MegaWizard Plug-In Manager interface that guides you through configuration of the transceiver based on your application requirements. The Stratix IV GX and GT transceivers allow you to implement low-power and reliable high-speed serial interface applications with its fully reconfigurable hardware, optimal signal integrity, and integrated Quartus II software platform. f For more information about the Quartus II software features, refer to the Quartus II Handbook. Ordering Information This section describes the Stratix IV E, GT, and GX devices ordering information. Figure 1-4 shows the ordering codes for Stratix IV GX and E devices. Figure 1-4. Stratix IV GX and E Device Packaging Ordering Information EP4SGX 230 K F 40 C Family Signature 2 ES Optional Suffix Indicates specific device options ES: Engineering sample N: Lead-free devices EP4SGX: Stratix IV Transceiver EP4SE: Stratix IV Logic/Memory Device Density Speed Grade 70 110 180 230 290 360 530 820 D: 8 F: 16 H: 24 K: 36 N: 48 Package Type Operating Temperature C: Commercial Temperature (tJ=0 C to 85 C) I: Industrial Temperature (tJ=-40 C to 100 C) M: Military Temperature (tJ=-55 C to 125 C) Ball Array Dimension F: FineLine BGA (FBGA) H: Hybrid FineLine BGA September 2012 2, 2x, 3, or 4, with 2 being the fastest Transceiver Count Altera Corporation Corresponds to pin count 29 = 780 pins 35 = 1152 pins 40 = 1517 pins 43 = 1760 pins 45 = 1932 pins Stratix IV Device Handbook Volume 1 1-20 Chapter 1: Overview for the Stratix IV Device Family Ordering Information Figure 1-5 shows the ordering codes for Stratix IV GT devices. Figure 1-5. Stratix IV GT Device Packaging Ordering Information EP4SEP4S 40G 2 230 40 F C 2 ES Family Signature Optional Suffix Aggregate Bandwidth Indicates specific device options ES: Engineering sample N: Lead-free devices 40G 100G Device Density Speed Grade 2 = 230k LEs 3 = 290k LEs 4 = 360k LEs 5 = 530k LEs 1, 2, 3 with 1 being the fastest Operating Temperature Package Type C: Commercial temperature (t J = 0 C to 85 C) I : Industrial temperature (t J = 0C to 100C) F: FineLine BGA (FBGA) H: Hybrid FineLine BGA Ball Array Dimension Corresponds to pin count 40 = 1517 pins 45 = 1932 pins o Document Revision History Table 1-10 lists the revision history for this chapter. Table 1-10. Document Revision History (Part 1 of 2) Date Version September 2012 3.4 June 2011 3.3 February 2011 March 2010 Stratix IV Device Handbook Volume 1 3.2 3.1 Changes Updated Table 1-1 to close FB #30986. Updated Table 1-2 and Table 1-5 to close FB #31127. Added military temperature to Figure 1-4. Updated Table 1-7 and Table 1-8. Applied new template. Minor text edits. Updated Table 1-1, Table 1-2, and Table 1-7. Updated Figure 1-3. Updated the "Stratix IV GT Devices" section. Added two new references to the Introduction section. Minor text edits. September 2012 Altera Corporation Chapter 1: Overview for the Stratix IV Device Family Ordering Information 1-21 Table 1-10. Document Revision History (Part 2 of 2) Date Version November 2009 June 2009 3.0 2.4 April 2009 2.3 March 2009 2.2 March 2009 2.1 November 2008 2.0 Changes Updated the "Stratix IV Device Family Overview", "Feature Summary", "Stratix IV GT Devices", "High-Speed Transceiver Features", "FPGA Fabric and I/O Features", "Highest Aggregate Data Bandwidth", "System Integration", and "Integrated Software Platform" sections. Added Table 1-3, Table 1-6, and Table 1-9. Updated Table 1-1, Table 1-2, Table 1-4, Table 1-5, Table 1-7, and Table 1-8. Updated Figure 1-3, Figure 1-4, and Figure 1-5. Minor text edits. Updated Table 1-1. Minor text edits. Added Table 1-5, Table 1-6, and Figure 1-3. Updated Figure 1-5. Updated Table 1-1, Table 1-2, Table 1-3, and Table 1-4. Updated "Introduction", "Feature Summary", "Stratix IV GX Devices", "Stratix IV GT Devices", "Architecture Features", and "FPGA Fabric and I/O Features" Updated "Feature Summary", "Stratix IV GX Devices", "Stratix IV E Device", "Stratix IV GT Devices", "Signal Integrity" Removed Tables 1-5 and 1-6 Updated Figure 1-4 Updated "Introduction", "Feature Summary", "Stratix IV Device Diagnostic Features", "Signal Integrity", "Clock Networks","High-Speed Differential I/O with DPA and SoftCDR", "System Integration", and "Ordering Information" sections. Added "Stratix IV GT 100G Devices" and "Stratix IV GT 100G Transceiver Bandwidth" sections. Updated Table 1-1, Table 1-2, Table 1-3, and Table 1-4. Added Table 1-5 and Table 1-6. Updated Figure 1-3 and Figure 1-4. Added Figure 1-5. Removed "Referenced Documents" section. Updated "Feature Summary" on page 1-1. Updated "Stratix IV Device Diagnostic Features" on page 1-7. Updated "FPGA Fabric and I/O Features" on page 1-8. Updated Table 1-1. Updated Table 1-2. Updated "Table 1-5 shows the total number of transceivers available in the Stratix IV GT Device." on page 1-15. July 2008 1.1 Revised "Introduction". May 2008 1.0 Initial release. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 1-22 Stratix IV Device Handbook Volume 1 Chapter 1: Overview for the Stratix IV Device Family Ordering Information September 2012 Altera Corporation 2. Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices February 2011 SIV51002-3.1 SIV51002-3.1 This chapter describes the features of the logic array blocks (LABs) in the Stratix(R) IV core fabric. LABs are made up of adaptive logic modules (ALMs) that you can configure to implement logic functions, arithmetic functions, and register functions. LABs and ALMs are the basic building blocks of the Stratix IV device. Use these to configure logic, arithmetic, and register functions. The ALM provides advanced features with efficient logic usage and is completely backward-compatible. This chapter contains the following sections: "Logic Array Blocks" "Adaptive Logic Modules" on page 2-5 Logic Array Blocks Each LAB consists of ten ALMs, various carry chains, shared arithmetic chains, LAB control signals, local interconnect, and register chain connection lines. The local interconnect transfers signals between ALMs in the same LAB. The direct link interconnect allows the LAB to drive into the local interconnect of its left and right neighbors. Register chain connections transfer the output of the ALM register to the adjacent ALM register in the LAB. The Quartus(R) II Compiler places associated logic in the LAB or adjacent LABs, allowing the use of local, shared arithmetic chain, and register chain connections for performance and area efficiency. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 2-2 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Logic Array Blocks Figure 2-1 shows the Stratix IV LAB structure and interconnects. Figure 2-1. Stratix IV LAB Structure and Interconnects C4 C12 Row Interconnects of Variable Speed & Length R20 R4 ALMs Direct link interconnect from adjacent block Direct link interconnect from adjacent block Direct link interconnect to adjacent block Direct link interconnect to adjacent block Local Interconnect LAB MLAB Local Interconnect is Driven from Either Side by Columns & LABs, & from Above by Rows Column Interconnects of Variable Speed & Length The LAB of the Stratix IV device has a derivative called memory LAB (MLAB), which adds look-up table (LUT)-based SRAM capability to the LAB, as shown in Figure 2-2. The MLAB supports a maximum of 640 bits of simple dual-port static random access memory (SRAM). You can configure each ALM in an MLAB as either a 64 x 1 or a 32 x 2 block, resulting in a configuration of either a 64 x 10 or a 32 x 20 simple dual-port SRAM block. MLAB and LAB blocks always coexist as pairs in all Stratix IV families. MLAB is a superset of the LAB and includes all LAB features. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Logic Array Blocks 2-3 f For more information about the MLAB, refer to the chapter. Figure 2-2. Stratix IV LAB and MLAB Structure LUT-based-64 x 1 Simple dual-port SRAM (1) ALM LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) ALM LAB Control Block (1) ALM LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) LUT-based-64 x 1 Simple dual-port SRAM (1) MLAB ALM ALM LAB Control Block LUT-based-64 x 1 Simple dual-port SRAM ALM ALM ALM ALM ALM LAB Note to Figure 2-2: (1) You can use the MLAB ALM as a regular LAB ALM or configure it as a dual-port SRAM, as shown. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-4 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Logic Array Blocks LAB Interconnects The LAB local interconnect can drive ALMs in the same LAB. It is driven by column and row interconnects and ALM outputs in the same LAB. Neighboring LABs/MLABs, M9K RAM blocks, M144K blocks, or digital signal processing (DSP) blocks from the left or right can also drive the LAB's local interconnect through the direct link connection. The direct link connection feature minimizes the use of row and column interconnects, providing higher performance and flexibility. Each LAB can drive 30 ALMs through fast-local and direct-link interconnects. Figure 2-3 shows the direct-link connection. Figure 2-3. Direct-Link Connection Direct-link interconnect from the left LAB, TriMatrix memory block, DSP block, or IOE output Direct-link interconnect from the right LAB, TriMatrix memory block, DSP block, or IOE output ALMs ALMs Direct-link interconnect to right Direct-link interconnect to left Local Interconnect MLAB LAB LAB Control Signals Each LAB contains dedicated logic for driving control signals to its ALMs. Control signals include three clocks, three clock enables, two asynchronous clears, a synchronous clear, and synchronous load control signals. This gives a maximum of 10 control signals at a time. Although you generally use synchronous-load and clear signals when implementing counters, you can also use them with other functions. Each LAB has two unique clock sources and three clock enable signals, as shown in Figure 2-4. The LAB control block can generate up to three clocks using two clock sources and three clock enable signals. Each LAB's clock and clock enable signals are linked. For example, any ALM in a particular LAB using the labclk1 signal also uses the labclkena1 signal. If the LAB uses both the rising and falling edges of a clock, it also uses two LAB-wide clock signals. De-asserting the clock enable signal turns off the corresponding LAB-wide clock. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-5 The LAB row clocks [5..0] and LAB local interconnects generate the LAB-wide control signals. The MultiTrack interconnect's inherent low skew allows clock and control signal distribution in addition to data. Figure 2-4. LAB-Wide Control Signals There are two unique clock signals per LAB. 6 Dedicated Row LAB Clocks 6 6 Local Interconnect Local Interconnect Local Interconnect Local Interconnect Local Interconnect Local Interconnect labclk0 labclk2 labclk1 labclkena0 or asyncload or labpreset labclkena1 labclkena2 labclr1 syncload labclr0 synclr Adaptive Logic Modules The ALM is the basic building block of logic in the Stratix IV architecture. It provides advanced features with efficient logic usage. Each ALM contains a variety of LUT-based resources that can be divided between two combinational adaptive LUTs (ALUTs) and two registers. With up to eight inputs for the two combinational ALUTs, one ALM can implement various combinations of two functions. This adaptability allows an ALM to be completely backward-compatible with four-input LUT architectures. One ALM can also implement any function with up to six inputs and certain seven-input functions. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-6 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules In addition to the adaptive LUT-based resources, each ALM contains two programmable registers, two dedicated full adders, a carry chain, a shared arithmetic chain, and a register chain. Through these dedicated resources, an ALM can efficiently implement various arithmetic functions and shift registers. Each ALM drives all types of interconnects: local, row, column, carry chain, shared arithmetic chain, register chain, and direct link. Figure 2-5 shows a high-level block diagram of the Stratix IV ALM. Figure 2-5. High-Level Block Diagram of the Stratix IV ALM shared_arith_in carry_in Combinational/Memory ALUT0 reg_chain_in labclk To general or local routing dataf0 datae0 6-Input LUT adder0 D Q dataa To general or local routing reg0 datab datac datad datae1 adder1 D Q 6-Input LUT To general or local routing reg1 dataf1 To general or local routing Combinational/Memory ALUT1 reg_chain_out shared_arith_out Stratix IV Device Handbook Volume 1 carry_out February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-7 Figure 2-6 shows a detailed view of all the connections in an ALM. Figure 2-6. Stratix IV ALM Connection Details syncload aclr[1:0] shared_arith_in carry_in clk[2:0] sclr reg_chain_in dataf0 datae0 dataa datab GND 4-INPUT LUT datac0 + CLR D Q 3-INPUT LUT local interconnect row, column direct link routing row, column direct link routing 3-INPUT LUT 4-INPUT LUT datac1 + CLR D Q 3-INPUT LUT local interconnect row, column direct link routing row, column direct link routing 3-INPUT LUT VCC datae1 dataf1 shared_arith_out carry_out reg_chain_out One ALM contains two programmable registers. Each register has data, clock, clock enable, synchronous and asynchronous clear, and synchronous load and clear inputs. Global signals, general-purpose I/O pins, or any internal logic can drive the register's clock and clear-control signals. Either general-purpose I/O pins or internal logic can drive the clock enable. For combinational functions, the register is bypassed and the output of the LUT drives directly to the outputs of an ALM. Each ALM has two sets of outputs that drive the local, row, and column routing resources. The LUT, adder, or register outputs can drive these output drivers (refer to Figure 2-6). For each set of output drivers, two ALM outputs can drive column, row, or direct-link routing connections. One of these ALM outputs can also drive local interconnect resources. This allows the LUT or adder to drive one output while the register drives another output. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-8 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules This feature, called register packing, improves device utilization because the device can use the register and the combinational logic for unrelated functions. Another special packing mode allows the register output to feed back into the LUT of the same ALM so that the register is packed with its own fan-out LUT. This provides another mechanism for improved fitting. The ALM can also drive out registered and unregistered versions of the LUT or adder output. ALM Operating Modes The Stratix IV ALM operates in one of the following modes: Normal Extended LUT Arithmetic Shared Arithmetic LUT-Register Each mode uses ALM resources differently. In each mode, eleven available inputs to an ALM--the eight data inputs from the LAB local interconnect, carry-in from the previous ALM or LAB, the shared arithmetic chain connection from the previous ALM or LAB, and the register chain connection--are directed to different destinations to implement the desired logic function. LAB-wide signals provide clock, asynchronous clear, synchronous clear, synchronous load, and clock enable control for the register. These LAB-wide signals are available in all ALM modes. For more information about the LAB-wide control signals, refer to "LAB Control Signals" on page 2-4. The Quartus II software and supported third-party synthesis tools, in conjunction with parameterized functions such as the library of parameterized modules (LPM) functions, automatically choose the appropriate mode for common functions such as counters, adders, subtractors, and arithmetic functions. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-9 Normal Mode Normal mode is suitable for general logic applications and combinational functions. In this mode, up to eight data inputs from the LAB local interconnect are inputs to the combinational logic. Normal mode allows two functions to be implemented in one Stratix IV ALM, or a single function of up to six inputs. The ALM can support certain combinations of completely independent functions and various combinations of functions that have common inputs. Figure 2-7 shows the supported LUT combinations in normal mode. Figure 2-7. ALM in Normal Mode dataf0 datae0 datac dataa (1) 4-Input LUT combout0 datab datad datae1 dataf1 4-Input LUT combout1 dataf0 datae0 datac dataa datab 5-Input LUT combout0 datad datae1 dataf1 3-Input LUT dataf0 datae0 datac dataa datab 5-Input LUT 4-Input LUT datad datae1 dataf1 dataf0 datae0 datac dataa datab 5-Input LUT combout0 5-Input LUT combout1 dataf0 datae0 dataa datab datac datad 6-Input LUT combout0 dataf0 datae0 dataa datab datac datad 6-Input LUT combout0 6-Input LUT combout1 datad datae1 dataf1 combout1 combout0 combout1 datae1 dataf1 Note to Figure 2-7: (1) Combinations of functions with fewer inputs than those shown are also supported. For example, combinations of functions with the following number of inputs are supported: 4 and 3, 3 and 3, 3 and 2, and 5 and 2. Normal mode provides complete backward-compatibility with four-input LUT architectures. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-10 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules For the packing of 2 five-input functions into one ALM, the functions must have at least two common inputs. The common inputs are dataa and datab. The combination of a four-input function with a five-input function requires one common input (either dataa or datab). In the case of implementing 2 six-input functions in one ALM, four inputs must be shared and the combinational function must be the same. In a sparsely used device, functions that could be placed in one ALM may be implemented in separate ALMs by the Quartus II software to achieve the best possible performance. As a device begins to fill up, the Quartus II software automatically uses the full potential of the Stratix IV ALM. The Quartus II Compiler automatically searches for functions using common inputs or completely independent functions to be placed in one ALM to make efficient use of device resources. In addition, you can manually control resource usage by setting location assignments. You can implement any six-input function using inputs dataa, datab, datac, datad, and either datae0 and dataf0 or datae1 and dataf1. If you use datae0 and dataf0, the output is driven to register0, and/or register0 is bypassed and the data drives out to the interconnect using the top set of output drivers (refer to Figure 2-8). If you use datae1 and dataf1, the output either drives to register1 or bypasses register1 and drives to the interconnect using the bottom set of output drivers. The Quartus II Compiler automatically selects the inputs to the LUT. ALMs in normal mode support register packing. Figure 2-8. Input Function in Normal Mode dataf0 datae0 dataa datab datac datad (1) To general or local routing 6-Input LUT D Q To general or local routing reg0 datae1 dataf1 (2) D labclk Q To general or local routing reg1 These inputs are available for register packing. Notes to Figure 2-8: (1) If you use datae1 and dataf1 as inputs to a six-input function, datae0 and dataf0 are available for register packing. (2) The dataf1 input is available for register packing only if the six-input function is unregistered. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-11 Extended LUT Mode Use extended LUT mode to implement a specific set of seven-input functions. The set must be a 2-to-1 multiplexer fed by two arbitrary five-input functions sharing four inputs. Figure 2-9 shows the template of supported seven-input functions using extended LUT mode. In this mode, if the seven-input function is unregistered, the unused eighth input is available for register packing. Functions that fit into the template shown in Figure 2-9 occur naturally in designs. These functions often appear in designs as "if-else" statements in Verilog HDL or VHDL code. Figure 2-9. Template for Supported Seven-Input Functions in Extended LUT Mode datae0 datac dataa datab datad dataf0 5-Input LUT To general or local routing combout0 D 5-Input LUT Q To general or local routing reg0 datae1 dataf1 (1) This input is available for register packing. Note to Figure 2-9: (1) If the seven-input function is unregistered, the unused eighth input is available for register packing. The second register, reg1, is not available. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-12 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules Arithmetic Mode Arithmetic mode is ideal for implementing adders, counters, accumulators, wide parity functions, and comparators. The ALM in arithmetic mode uses two sets of 2 four-input LUTs along with two dedicated full adders. The dedicated adders allow the LUTs to be available to perform pre-adder logic; therefore, each adder can add the output of 2 four-input functions. The four LUTs share dataa and datab inputs. As shown in Figure 2-10, the carry-in signal feeds to adder0 and the carry-out from adder0 feeds to the carry-in of adder1. The carry-out from adder1 drives to adder0 of the next ALM in the LAB. ALMs in arithmetic mode can drive out registered and/or unregistered versions of the adder outputs. Figure 2-10. ALM in Arithmetic Mode carry_in datae0 adder0 4-Input LUT To general or local routing D dataf0 datac datab dataa datad datae1 Q To general or local routing reg0 4-Input LUT adder1 4-Input LUT To general or local routing D 4-Input LUT Q To general or local routing reg1 dataf1 carry_out While operating in arithmetic mode, the ALM can support simultaneous use of the adder's carry output along with combinational logic outputs. In this operation, adder output is ignored. Using the adder with combinational logic output provides resource savings of up to 50% for functions that can use this ability. Arithmetic mode also offers clock enable, counter enable, synchronous up/down control, add/subtract control, synchronous clear, and synchronous load. The LAB local interconnect data inputs generate the clock enable, counter enable, synchronous up/down, and add/subtract control signals. These control signals are good candidates for the inputs that are shared between the four LUTs in the ALM. The synchronous clear and synchronous load options are LAB-wide signals that affect all registers in the LAB. These signals can also be individually disabled or enabled per register. The Quartus II software automatically places any registers that are not used by the counter into other LABs. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-13 Carry Chain The carry chain provides a fast carry function between the dedicated adders in arithmetic or shared-arithmetic mode. The two-bit carry select feature in Stratix IV devices halves the propagation delay of carry chains within the ALM. Carry chains can begin in either the first ALM or the fifth ALM in the LAB. The final carry-out signal is routed to the ALM, where it is fed to local, row, or column interconnects. The Quartus II Compiler automatically creates carry-chain logic during design processing, or you can create it manually during design entry. Parameterized functions such as LPM functions automatically take advantage of carry chains for the appropriate functions. The Quartus II Compiler creates carry chains longer than 20 (10 ALMs in arithmetic or shared arithmetic mode) by linking LABs together automatically. For enhanced fitting, a long carry chain runs vertically, allowing fast horizontal connections to TriMatrix memory and DSP blocks. A carry chain can continue as far as a full column. To avoid routing congestion in one small area of the device when a high fan-in arithmetic function is implemented, the LAB can support carry chains that only use either the top half or bottom half of the LAB before connecting to the next LAB. This leaves the other half of the ALMs in the LAB available for implementing narrower fan-in functions in normal mode. Carry chains that use the top five ALMs in the first LAB carry into the top half of the ALMs in the next LAB within the column. Carry chains that use the bottom five ALMs in the first LAB carry into the bottom half of the ALMs in the next LAB within the column. In every alternate LAB column, the top half can be bypassed; in the other MLAB columns, the bottom half can be bypassed. For more information about carry-chain interconnects, refer to "ALM Interconnects" on page 2-18. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-14 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules Shared Arithmetic Mode In shared arithmetic mode, the ALM can implement a three-input add within the ALM. In this mode, the ALM is configured with 4 four-input LUTs. Each LUT either computes the sum of three inputs or the carry of three inputs. The output of the carry computation is fed to the next adder (either to adder1 in the same ALM or to adder0 of the next ALM in the LAB) using a dedicated connection called the shared arithmetic chain. This shared arithmetic chain can significantly improve the performance of an adder tree by reducing the number of summation stages required to implement an adder tree. Figure 2-11 shows the ALM using this feature. Figure 2-11. ALM in Shared Arithmetic Mode shared_arith_in carry_in labclk 4-Input LUT To general or local routing D datae0 datac datab dataa datad datae1 Q To general or local routing reg0 4-Input LUT 4-Input LUT To general or local routing D 4-Input LUT Q To general or local routing reg1 carry_out shared_arith_out You can find adder trees in many different applications. For example, the summation of the partial products in a logic-based multiplier can be implemented in a tree structure. Another example is a correlator function that can use a large adder tree to sum filtered data samples in a given time frame to recover or de-spread data that was transmitted using spread-spectrum technology. Shared Arithmetic Chain The shared arithmetic chain available in enhanced arithmetic mode allows the ALM to implement a three-input add. This significantly reduces the resources necessary to implement large adder trees or correlator functions. Shared arithmetic chains can begin in either the first or sixth ALM in the LAB. The Quartus II Compiler creates shared arithmetic chains longer than 20 (10 ALMs in arithmetic or shared arithmetic mode) by linking LABs together automatically. For enhanced fitting, a long shared arithmetic chain runs vertically, allowing fast horizontal connections to the TriMatrix memory and DSP blocks. A shared arithmetic chain can continue as far as a full column. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-15 Similar to the carry chains, the top and bottom halves of shared arithmetic chains in alternate LAB columns can be bypassed. This capability allows the shared arithmetic chain to cascade through half of the ALMs in an LAB while leaving the other half available for narrower fan-in functionality. Every other LAB column is top-half by-passable, while the other LAB columns are bottom-half by-passable. For more information about the shared arithmetic chain interconnect, refer to "ALM Interconnects" on page 2-18. LUT-Register Mode LUT-register mode allows third-register capability within an ALM. Two internal feedback loops allow combinational ALUT1 to implement the master latch and combinational ALUT0 to implement the slave latch needed for the third register. The LUT register shares its clock, clock enable, and asynchronous clear sources with the top dedicated register. Figure 2-12 shows the register constructed using two combinational blocks within the ALM. Figure 2-12. LUT Register from Two Combinational Blocks sumout clk aclr LUT regout 4-input LUT combout 5-input LUT combout sumout datain(datac) sclr February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-16 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules Figure 2-13 shows the ALM in LUT-register mode. Figure 2-13. ALM in LUT-Register Mode with Three-Register Capability clk [2:0] DC1 aclr [1:0] reg_chain_in datain lelocal 0 aclr aclr sclr regout datain latchout sdata leout 0 a regout leout 0 b E0 F1 lelocal 1 aclr datain E1 sdata F0 leout 1 a regout leout 1 b reg_chain_out Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-17 Register Chain In addition to general routing outputs, ALMs in the LAB have register-chain outputs. Register-chain routing allows registers in the same LAB to be cascaded together. The register-chain interconnect allows the LAB to use LUTs for a single combinational function and the registers to be used for an unrelated shift-register implementation. These resources speed up connections between ALMs while saving local interconnect resources (refer to Figure 2-14). The Quartus II Compiler automatically takes advantage of these resources to improve utilization and performance. Figure 2-14. Register Chain within the LAB (1) From previous ALM within the LAB reg_chain_in labclk To general or local routing adder0 D Q To general or local routing reg0 Combinational Logic adder1 D Q To general or local routing reg1 To general or local routing To general or local routing adder0 D Q To general or local routing reg0 Combinational Logic adder1 D Q To general or local routing reg1 To general or local routing reg_chain_out To next ALM within the LAB Note to Figure 2-14: (1) You can use the combinational or adder logic to implement an unrelated, un-registered function. For more information about the register chain interconnect, refer to "ALM Interconnects" on page 2-18. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 2-18 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules ALM Interconnects There are three dedicated paths between the ALMs--register cascade, carry chain, and shared arithmetic chain. Stratix IV devices include an enhanced interconnect structure in LABs for routing shared arithmetic chains and carry chains for efficient arithmetic functions. The register chain connection allows the register output of one ALM to connect directly to the register input of the next ALM in the LAB for fast shift registers. These ALM-to-ALM connections bypass the local interconnect. The Quartus II Compiler automatically takes advantage of these resources to improve utilization and performance. Figure 2-15 shows the shared arithmetic chain, carry chain, and register chain interconnects. Figure 2-15. Shared Arithmetic Chain, Carry Chain, and Register Chain Interconnects Local interconnect routing among ALMs in the LAB Carry chain & shared arithmetic chain routing to adjacent ALM ALM 1 Register chain routing to adjacent ALM's register input ALM 2 Local interconnect ALM 3 ALM 4 ALM 5 ALM 6 ALM 7 ALM 8 ALM 9 ALM 10 Clear and Preset Logic Control LAB-wide signals control the logic for the register's clear signal. The ALM directly supports an asynchronous clear function. You can achieve the register preset through the Quartus II software's NOT-gate push-back logic option. Each LAB supports up to two clears. Stratix IV devices provide a device-wide reset pin (DEV_CLRn) that resets all the registers in the device. An option set before compilation in the Quartus II software controls this pin. This device-wide reset overrides all other control signals. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules 2-19 LAB Power Management Techniques The following techniques are used to manage static and dynamic power consumption within the LAB: To save AC power, the Quartus II software forces all adder inputs low when ALM adders are not in use. Stratix IV LABs operate in high-performance mode or low-power mode. The Quartus II software automatically chooses the appropriate mode for the LAB, based on the design, to optimize speed versus leakage trade-offs. Clocks represent a significant portion of dynamic power consumption due to their high switching activity and long paths. The LAB clock that distributes a clock signal to registers within an LAB is a significant contributor to overall clock power consumption. Each LAB's clock and clock enable signal are linked. For example, a combinational ALUT or register in a particular LAB using the labclk1 signal also uses the labclkena1 signal. To disable LAB-wide clock power consumption without disabling the entire clock tree, use LAB-wide clock enable to gate the LAB-wide clock. The Quartus II software automatically promotes register-level clock enable signals to the LAB-level. All registers within the LAB that share a common clock and clock enable are controlled by a shared, gated clock. To take advantage of these clock enables, use a clock-enable construct in your HDL code for the registered logic. f For more information about implementing static and dynamic power consumption within the LAB, refer to the Power Optimization chapter in volume 2 of the Quartus II Handbook. Document Revision History Table 2-1 lists the revision history for this chapter. Table 2-1. Document Revision History Date Version February 2011 3.1 November 2009 3.0 June 2009 2.2 March 2009 2.1 November 2008 2.0 May 2008 1.0 February 2011 Altera Corporation Changes Updated Figure 2-6. Applied new template. Minor text edits. Updated graphics. Minor text edits. Removed the Conclusion section. Added introductory sentences to improve search ability. Minor text edits. Removed "Referenced Documents" section. Updated Figure 2-6. Made minor editorial changes. Initial release. Stratix IV Device Handbook Volume 1 2-20 Stratix IV Device Handbook Volume 1 Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices Adaptive Logic Modules February 2011 Altera Corporation 3. TriMatrix Embedded Memory Blocks in Stratix IV Devices December 2011 SIV51003-3.3 SIV51003-3.3 This chapter describes the TriMatrix embedded memory blocks in Stratix(R) IV devices. TriMatrix embedded memory blocks provide three different sizes of embedded SRAM to efficiently address the needs of Stratix IV FPGA designs. TriMatrix memory includes 640-bit memory logic array blocks (MLABs), 9-Kbit M9K blocks, and 144-Kbit M144K blocks. MLABs have been optimized to implement filter delay lines, small FIFO buffers, and shift registers. You can use the M9K blocks for general purpose memory applications and the M144K blocks for processor code storage, packet buffering, and video frame buffering. You can independently configure each embedded memory block to be a single- or dual-port RAM, FIFO buffer, ROM, or shift register using the Quartus(R) II MegaWizardTM Plug-In Manager. You can stitch together multiple blocks of the same type to produce larger memories with minimal timing penalty. TriMatrix memory provides up to 31,491 Kbits of embedded SRAM at up to 600 MHz operation. This chapter contains the following sections: "Overview" "Memory Modes" on page 3-9 "Clocking Modes" on page 3-17 "Design Considerations" on page 3-18 Overview Table 3-1 lists the features supported by the three sizes of TriMatrix memory. Table 3-1. Summary of TriMatrix Memory Features (Part 1 of 2) Feature Maximum performance Total RAM bits (including parity bits) MLABs M9K Blocks M144K Blocks 600 MHz 600 MHz 540 MHz 640 9216 147,456 (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 December 2011 Feedback Subscribe 3-2 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview Table 3-1. Summary of TriMatrix Memory Features (Part 2 of 2) Feature MLABs M9K Blocks 8K x 1 4K x 2 64 x 8 2K x 4 64 x 9 Configurations (depth x width) 1K x 8 64 x 10 1K x 9 32 x 16 512 x 16 32 x 18 512 x 18 32 x 20 256 x 32 256 x 36 M144K Blocks 16K x 8 16K x 9 8K x 16 8K x 18 4K x 32 4K x 36 2K x 64 2K x 72 Parity bits Supported Supported Supported Byte enable Supported Supported Supported -- Supported Supported Address clock enable Supported Supported Supported Single-port memory Supported Supported Supported Simple dual-port memory Supported Supported Supported True dual-port memory -- Supported Supported Embedded shift register Supported Supported Supported ROM Supported Supported Supported FIFO buffer Packed mode Supported Supported Supported Simple dual-port mixed width support -- Supported Supported True dual-port mixed width support -- Supported Supported Memory Initialization File (.mif) Supported Supported Supported Mixed clock mode Supported Supported Supported Power-up condition Outputs cleared if registered, otherwise reads memory contents Outputs cleared Outputs cleared Register clears Output registers Output registers Output registers Write/Read operation triggering Write: Falling clock edges Write and Read: Rising clock edges Write and Read: Rising clock edges Same-port read-during-write Outputs set to don't care Outputs set to old data or new data Outputs set to old data or new data Mixed-port read-during-write Outputs set to old data, new data, or don't care Outputs set to old data or don't care Outputs set to old data or don't care ECC Support Soft IP support using the Quartus II software Soft IP support using the Quartus II software Built-in support in x64-wide SDP mode or soft IP support using the Quartus II software Read: Rising clock edges (1) Note to Table 3-1: (1) The mixed-port read-during-write options of new data or old data are only supported for MLABs when you use both the read address registers and the output registers. Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview 3-3 Table 3-2 lists the capacity and distribution of the TriMatrix memory blocks in each Stratix IV family member. Table 3-2. TriMatrix Memory Capacity and Distribution in Stratix IV Devices MLABs M9K Blocks M144K Blocks Total Dedicated RAM Bits (Dedicated Memory Blocks Only) (Kb) Total RAM Bits (Including MLABs) (Kb) EP4SE230 4,560 1,235 22 14,283 17,133 EP4SE360 7,072 1,248 48 18,144 22,564 EP4SE530 10,624 1,280 64 20,736 27,376 EP4SE820 16,261 1,610 60 23,130 33,294 EP4SGX70 1,452 462 16 6,462 7,370 EP4SGX110 2,112 660 16 8,244 9,564 Device EP4SGX180 3,515 950 20 11,430 13,627 EP4SGX230 4,560 1,235 22 14,283 17,133 EP4SGX290 5,824 936 36 13,608 17,248 EP4SGX360 7,072 1,248 48 18,144 22,564 EP4SGX530 10,624 1,280 64 20,736 27,376 EP4S40G2 4,560 1,235 22 14,283 17,133 EP4S40G5 10,624 1280 64 20,736 27,376 EP4S100G2 4,560 1,235 22 14,283 17,133 EP4S100G3 5,824 936 36 13,608 17,248 EP4S100G4 7,072 1,248 48 18,144 22,564 EP4S100G5 10,624 1,280 64 20,736 27,376 TriMatrix Memory Block Types While the M9K and M144K memory blocks are dedicated resources, the MLABs are dual-purpose blocks. They can be configured as regular logic array blocks (LABs) or as MLABs. Ten adaptive logic modules (ALMs) make up one MLAB. You can configure each ALM in an MLAB as either a 64 x 1 or a 32 x 2 block, resulting in a 64 x 10 or 32 x 20 simple dual-port SRAM block in a single MLAB. Parity Bit Support All TriMatrix memory blocks have built-in parity-bit support. The ninth bit associated with each byte can store a parity bit or serve as an additional data bit. No parity function is actually performed on the ninth bit. Byte Enable Support All TriMatrix memory blocks support byte enables that mask the input data so that only specific bytes of data are written. The unwritten bytes retain the previously written values. The write enable (wren) signals, along with the byte enable (byteena) signals, control the RAM blocks' write operations. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-4 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview The default value for the byte enable signals is high (enabled), in which case writing is controlled only by the write enable signals. The byte enable registers have no clear port. When using parity bits on the M9K and M144K blocks, the byte enable controls all nine bits (eight bits of data plus one parity bit). When using parity bits on the MLAB, the byte-enable controls all 10 bits in the widest mode. The MSB for the byteena signal corresponds to the MSB of the data bus and the LSB of the byteena signal corresponds to the LSB of the data bus. For example, if you use a RAM block in x18 mode, with byteena = 01, data[8..0] is enabled, and data[17..9] id disabled. Similarly, if byteena = 11, both data[8..0] and data[17..9] are enabled. Byte enables are active high. 1 You cannot use the byte enable feature when using the error correction coding (ECC) feature on M144K blocks. 1 Byte enables are only supported for true dual-port memory configurations when both the PortA and PortB data widtByths of the individual M9K memory blocks are multiples of 8 or 9 bits. For example, if you implement a mixed data width memory configured with portA = 32 and portB = 8 as two separate 16 x 4 bit memories, you cannot use the byte enable feature. Figure 3-1 shows how the write enable (wren) and byte enable (byteena) signals control the operations of the RAM blocks. When a byte-enable bit is de-asserted during a write cycle, the corresponding data byte output can appear as either a "don't care" value or the current data at that location. The output value for the masked byte is controllable using the Quartus II software. When a byte-enable bit is asserted during a write cycle, the corresponding data byte output also depends on the setting chosen in the Quartus II software. Figure 3-1. Byte Enable Functional Waveform inclock wren address data byteena contents at a0 contents at a1 a0 an a1 a2 a0 a1 ABCD XXXX 10 XX a2 XXXX 01 11 FFFF XX ABFF FFFF FFCD FFFF contents at a2 ABCD don't care: q (asynch) doutn ABXX XXCD ABCD ABFF FFCD ABCD current data: q (asynch) doutn ABFF FFCD ABCD ABFF FFCD ABCD Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview 3-5 Packed Mode Support Stratix IV M9K and M144K blocks support packed mode. The packed mode feature packs two independent single-port RAMs into one memory block. The Quartus II software automatically implements packed mode where appropriate by placing the physical RAM block into true dual-port mode and using the MSB of the address to distinguish between the two logical RAMs. The size of each independent single-port RAM must not exceed half of the target block size. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-6 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview Address Clock Enable Support All Stratix IV memory blocks support address clock enable, which holds the previous address value for as long as the signal is enabled (addressstall = 1). When the memory blocks are configured in dual-port mode, each port has its own independent address clock enable. The default value for the address clock enable signals is low (disabled). Figure 3-2 shows an address clock enable block diagram. The address clock enable is referred to by the port name addressstall. Figure 3-2. Address Clock Enable address[0] 1 0 address[N] 1 0 address[0] register address[0] address[N] register address[N] addressstall clock Figure 3-3 shows the address clock enable waveform during the read cycle. Figure 3-3. Address Clock Enable During Read Cycle Waveform inclock rdaddress a0 a1 a2 a3 a4 a5 a6 rden addressstall latched address (inside memory) an q (synch) doutn-1 q (asynch) Stratix IV Device Handbook Volume 1 doutn a1 a0 doutn dout0 dout0 a4 dout4 dout1 dout1 a5 dout4 dout5 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview 3-7 Figure 3-4 shows the address clock enable waveform during the write cycle. Figure 3-4. Address Clock Enable During the Write Cycle Waveform inclock wraddress a0 a1 a2 a3 a4 a5 a6 00 01 02 03 04 05 06 data wren addressstall latched address (inside memory) contents at a0 contents at a1 an a1 a0 a4 a5 00 XX XX 01 02 contents at a2 XX contents at a3 XX contents at a4 contents at a5 03 04 XX XX 05 Mixed Width Support M9K and M144K memory blocks support mixed data widths inherently. MLABs can support mixed data widths through emulation using the Quartus II software. When using simple dual-port, true dual-port, or FIFO modes, mixed width support allows you to read and write different data widths to a memory block. For more information about the different widths supported per memory mode, refer to "Memory Modes" on page 3-9. 1 MLABs do not support mixed-width FIFO mode. Asynchronous Clear Stratix IV TriMatrix memory blocks support asynchronous clears on output latches and output registers. Therefore, if your RAM is not using output registers, you can still clear the RAM outputs using the output latch asynchronous clear. Figure 3-5 shows a waveform of the output latch asynchronous clear function. Figure 3-5. Output Latch Asynchronous Clear Waveform outclk aclr aclr at latch q You can selectively enable asynchronous clears per logical memory using the Quartus II RAM MegaWizard Plug-In Manager. f For more information, refer to the Internal Memory (RAM and ROM) User Guide. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-8 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Overview Error Correction Code (ECC) Support Stratix IV M144K blocks have built-in support for error correction code (ECC) when in x64-wide simple dual-port mode. ECC allows you to detect and correct data errors in the memory array. The M144K blocks have a single-error-correction double-error-detection (SECDED) implementation. SECDED can detect and fix a single bit error in a 64-bit word, or detect two bit errors in a 64-bit word. It cannot detect three or more errors. The M144K ECC status is communicated using a three-bit status flag eccstatus[2..0]. The status flag can be either registered or unregistered. When registered, it uses the same clock and asynchronous clear signals as the output registers. When unregistered, it cannot be asynchronously cleared. Table 3-3 lists the truth table for the ECC status flags. Table 3-3. Truth Table for ECC Status Flags Status eccstatus[2] eccstatus[1] eccstatus[0] No error 0 0 0 Single error and fixed 0 1 1 Double error and no fix 1 0 1 Illegal 0 0 1 Illegal 0 1 0 Illegal 1 0 0 Illegal 1 1 X 1 You cannot use the byte enable feature when ECC is engaged. 1 Read-during-write "old data mode" is not supported when ECC is engaged. Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes 3-9 Figure 3-6 shows a diagram of the ECC block of the M144K block. Figure 3-6. ECC Block Diagram of the M144K Block 8 64 64 SECDED Encoder Data Input 8 72 RAM Array 72 64 SECDED Encoder Comparator 8 64 8 8 64 Error Locator 64 Error Correction Block Flag Generator 3 Status Flags 64 Data Output Memory Modes Stratix IV TriMatrix memory blocks allow you to implement fully synchronous SRAM memory in multiple modes of operation. M9K and M144K blocks do not support asynchronous memory (unregistered inputs). MLABs support asynchronous (flow-through) read operations. Depending on which TriMatrix memory block you target, you can use the following: "Single-Port RAM Mode" on page 3-10 "Simple Dual-Port Mode" on page 3-11 "True Dual-Port Mode" on page 3-14 "Shift-Register Mode" on page 3-16 "ROM Mode" on page 3-17 "FIFO Mode" on page 3-17 c When using the memory blocks in ROM, single-port, simple dual-port, or true dual-port mode, you can corrupt the memory contents if you violate the setup or hold-time on any of the memory block input registers. This applies to both read and write operations. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-10 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes Single-Port RAM Mode All TriMatrix memory blocks support single-port mode. Single-port mode allows you to do either one-read or one-write operation at a time. Simultaneous reads and writes are not supported in single-port mode. Figure 3-7 shows the single-port RAM configuration. Figure 3-7. Single-Port RAM (1) data[ ] address[ ] wren byteena[] addressstall inclock clockena rden aclr q[] outclock Note to Figure 3-7: (1) You can implement two single-port memory blocks in a single M9K or M144K block. For more information, refer to "Packed Mode Support" on page 3-5. During a write operation, RAM output behavior is configurable. If you use the read-enable signal and perform a write operation with read enable de-activated, the RAM outputs retain the values they held during the most recent active read enable. If you activate read enable during a write operation, or if you are not using the read-enable signal at all, the RAM outputs either show the "new data" being written, the "old data" at that address, or a "don't care" value. To choose the desired behavior, set the read-during-write behavior to either new data, old data, or don't care in the RAM MegaWizard Plug-In Manager in the Quartus II software. For more information, refer to "Read-During-Write Behavior" on page 3-19. Table 3-4 lists the possible port width configurations for TriMatrix memory blocks in single-port mode. Table 3-4. Port Width Configurations for MLABs, M9K, and M144K Blocks (Single-Port Mode) MLABs M9K Blocks 8K x 1 64 x 8 64 x 9 Port Width Configurations 64 x 10 32 x 16 32 x 18 32 x 20 4K x 2 2K x 4 1K x 8 1K x 9 512 x 16 512 x 18 256 x 32 256 x 36 Stratix IV Device Handbook Volume 1 M144K Blocks 16K x 8 16K x 9 8K x 16 8K x 18 4K x 32 4K x 36 2K x 64 2K x 72 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes 3-11 Figure 3-8 shows timing waveforms for read and write operations in single-port mode with unregistered outputs. Registering the RAM's outputs simply delays the q output by one clock cycle. Figure 3-8. Timing Waveform for Read-Write Operations (Single-Port Mode) clk_a A0 address A1 rdena wrena bytenna data_a 01 10 00 A123 B456 C789 A0 (old data) DoldDold23 q_a (asyn) 11 DDDD B423 EEEE A1(old data) FFFF DDDD EEEE Simple Dual-Port Mode All TriMatrix memory blocks support simple dual-port mode. Simple dual-port mode allows you to perform one read and one write operation to different locations at the same time. Write operation happens on port A; read operation happens on port B. Figure 3-9 shows a simple dual-port configuration. Figure 3-9. Stratix IV Simple Dual-Port Memory data[ ] wraddress[ ] wren byteena[] wr_addressstall wrclock wrclocken aclr (1) rdaddress[ ] rden q[ ] rd_addressstall rdclock rdclocken ecc_status Note to Figure 3-9: (1) Simple dual-port RAM supports input/output clock mode in addition to read/write clock mode. Simple dual-port mode supports different read and write data widths (mixed-width support). Table 3-5 lists the mixed width configurations for M9K blocks in simple dual-port mode. MLABs do not have native support for mixed-width operation. The Quartus II software implements mixed-width memories in MLABs by using more than one MLAB. Table 3-5. M9K Block Mixed-Width Configurations (Simple Dual-Port Mode) (Part 1 of 2) Write Port Read Port 8K x 1 4K x 2 2K x 4 1K x 8 512 x 16 256 x 32 1K x 9 512 x 18 256 x 36 8K x 1 Y Y Y Y Y Y -- -- -- 4K x 2 Y Y Y Y Y Y -- -- -- 2K x 4 Y Y Y Y Y Y -- -- -- December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-12 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes Table 3-5. M9K Block Mixed-Width Configurations (Simple Dual-Port Mode) (Part 2 of 2) Write Port Read Port 8K x 1 4K x 2 2K x 4 1K x 8 512 x 16 256 x 32 1K x 9 512 x 18 256 x 36 1K x 8 Y Y Y Y Y Y -- -- -- 512 x 16 Y Y Y Y Y Y -- -- -- 256 x 32 Y Y Y Y Y Y -- -- -- 1K x 9 -- -- -- -- -- -- Y Y Y 512 x 18 -- -- -- -- -- -- Y Y Y 256 x 36 -- -- -- -- -- -- Y Y Y Table 3-6 lists the mixed-width configurations for M144K blocks in simple dual-port mode. Table 3-6. M144K Block Mixed-Width Configurations (Simple Dual-Port Mode) Write Port Read Port 16K x 8 8K x 16 4K x 32 2K x 64 16K x 9 8K x 18 4K x 36 2K x 72 16K x 8 Y Y Y Y -- -- -- -- 8K x 16 Y Y Y Y -- -- -- -- 4K x 32 Y Y Y Y -- -- -- -- 2K x 64 Y Y Y Y -- -- -- -- 16K x 9 -- -- -- -- Y Y Y Y 8K x 18 -- -- -- -- Y Y Y Y 4K x 36 -- -- -- -- Y Y Y Y 2K x 72 -- -- -- -- Y Y Y Y In simple dual-port mode, M9K and M144K blocks support separate write-enable and read-enable signals. You can save power by keeping the read-enable signal low (inactive) when not reading. Read-during-write operations to the same address can either output a "don't care" value or "old data" value. To choose the desired behavior, set the read-during-write behavior to either don't care or old data in the RAM MegaWizard Plug-In Manager in the Quartus II software. For more information, refer to "Read-During-Write Behavior" on page 3-19. MLABs only support a write-enable signal. For MLABs, you can set the same-port read-during-write behavior to don't care and the mixed-port read-during-write behavior to either don't care or old data. The available choices depend on the configuration of the MLAB. There is no "new data" option for MLABs. Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes 3-13 Figure 3-10 shows timing waveforms for read and write operations in simple dual-port mode with unregistered outputs. Registering the RAM outputs simply delays the q output by one clock cycle. Figure 3-10. Simple Dual-Port Timing Waveforms wrclock wren wraddress an-1 data din-1 a0 an a1 a2 a3 din a4 a5 din4 din5 a6 din6 rdclock rden rdaddress q (asynch) bn b1 b0 doutn-1 b2 b3 dout0 doutn Figure 3-11 shows timing waveforms for read and write operations in mixed-port mode with unregistered outputs. Figure 3-11. Mixed-Port Read-During-Write Timing Waveforms wrclock wren wraddress an-1 data din-1 a0 an a1 a2 din a3 a4 a5 din4 din5 a6 din6 rdclock rden rdaddress q (asynch) December 2011 bn doutn-1 Altera Corporation b0 doutn b1 b2 b3 dout0 Stratix IV Device Handbook Volume 1 3-14 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes True Dual-Port Mode Stratix IV M9K and M144K blocks support true dual-port mode. Sometimes called bi-directional dual-port, this mode allows you to perform any combination of two port operations: two reads, two writes, or one read and one write at two different clock frequencies. Figure 3-12 shows the true dual-port RAM configuration. Figure 3-12. Stratix IV True Dual-Port Memory (1) data_a[ ] address_a[ ] wren_a byteena_a[] addressstall_a clock_a rden_a aclr_a q_a[] data_b[ ] address_b[] wren_b byteena_b[] addressstall_b clock_b rden_b aclr_b q_b[] Note to Figure 3-12: (1) True dual-port memory supports input/output clock mode in addition to independent clock mode. The widest bit configuration of the M9K and M144K blocks in true dual-port mode is as follows: M9K: 512 x 16-bit (or 512 x18-bit with parity) M144K: 4K x 32-bit (or 4K x36-bit with parity) Wider configurations are unavailable because the number of output drivers is equivalent to the maximum bit width of the respective memory block. Because true dual-port RAM has outputs on two ports, its maximum width equals half of the total number of output drivers. Table 3-7 lists the possible M9K block mixed-port width configurations in true dual-port mode. Table 3-7. M9K Block Mixed-Width Configuration (True Dual-Port Mode) Write Port Read Port Stratix IV Device Handbook Volume 1 8K x 1 4K x 2 2K x 4 1K x 8 512 x 16 1K x 9 512 x 18 8K x 1 Y Y Y Y Y -- -- 4K x 2 Y Y Y Y Y -- -- 2K x 4 Y Y Y Y Y -- -- 1K x 8 Y Y Y Y Y -- -- 512 x 16 Y Y Y Y Y -- -- 1K x 9 -- -- -- -- -- Y Y 512 x 18 -- -- -- -- -- Y Y December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes 3-15 Table 3-8 lists the possible M144K block mixed-port width configurations in true dual-port mode. Table 3-8. M144K Block Mixed-Width Configurations (True Dual-Port Mode) Write Port Read Port 16K x 8 8K x 16 4K x 32 16K x 9 8K x 18 4K x 36 16K x 8 Y Y Y -- -- -- 8K x 16 Y Y Y -- -- -- 4K x 32 Y Y Y -- -- -- 16K x 9 -- -- -- Y Y Y 8K x 18 -- -- -- Y Y Y 4K x 36 -- -- -- Y Y Y In true dual-port mode, M9K and M144K blocks support separate write-enable and read-enable signals. You can save power by keeping the read-enable signal low (inactive) when not reading. Read-during-write operations to the same address can either output "new data" at that location or "old data". To choose the desired behavior, set the read-during-write behavior to either new data or old data in the RAM MegaWizard Plug-In Manager in the Quartus II software. For more information, refer to "Read-During-Write Behavior" on page 3-19. In true dual-port mode, you can access any memory location at any time from either port. When accessing the same memory location from both ports, you must avoid possible write conflicts. A write conflict happens when you attempt to write to the same address location from both ports at the same time. This results in unknown data being stored to that address location. No conflict resolution circuitry is built into the Stratix IV TriMatrix memory blocks. You must handle address conflicts external to the RAM block. Figure 3-13 shows true dual-port timing waveforms for the write operation at port A and the read operation at port B, with the read-during-write behavior set to new data. Registering the RAM's outputs simply delays the q outputs by one clock cycle. Figure 3-13. True Dual-Port Timing Waveform clk_a wren_a address_a an-1 an data_a din-1 din q_a (asynch) din-1 a0 din a1 dout0 a2 dout1 a3 dout2 a4 a5 a6 din4 din5 din6 dout3 din4 din5 clk_b wren_b address_b q_b (asynch) December 2011 bn doutn-1 Altera Corporation b0 b1 b2 b3 doutn dout0 dout1 dout2 Stratix IV Device Handbook Volume 1 3-16 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Memory Modes Shift-Register Mode All Stratix IV memory blocks support shift register mode. Embedded memory block configurations can implement shift registers for digital signal processing (DSP) applications, such as finite impulse response (FIR) filters, pseudo-random number generators, multi-channel filtering, and auto- and cross-correlation functions. These and other DSP applications require local data storage, traditionally implemented with standard flipflops that quickly exhaust many logic cells for large shift registers. A more efficient alternative is to use embedded memory as a shift-register block, which saves logic cell and routing resources. The size of a shift register (w x m x n) is determined by the input data width (w), the length of the taps (m), and the number of taps (n). You can cascade memory blocks to implement larger shift registers. Figure 3-14 shows the TriMatrix memory block in shift-register mode. Figure 3-14. Shift-Register Memory Configuration w x m x n Shift Register m-Bit Shift Register W W m-Bit Shift Register W W n Number of Taps m-Bit Shift Register W W m-Bit Shift Register W Stratix IV Device Handbook Volume 1 W December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Clocking Modes 3-17 ROM Mode All Stratix IV TriMatrix memory blocks support ROM mode. A .mif file initializes the ROM contents of these blocks. The address lines of the ROM are registered on M9K and M144K blocks, but can be unregistered on MLABs. The outputs can be registered or unregistered. Output registers can be asynchronously cleared. The ROM read operation is identical to the read operation in the single-port RAM configuration. FIFO Mode All TriMatrix memory blocks support FIFO mode. MLABs are ideal for designs with many small, shallow FIFO buffers. To implement FIFO buffers in your design, use the Quartus II software FIFO MegaWizard Plug-In Manager. Both single- and dual-clock (asynchronous) FIFO buffers are supported. f For more information about implementing FIFO buffers, refer to the SCFIFO and DCFIFO Megafunctions User Guide. 1 MLABs do not support mixed-width FIFO mode. Clocking Modes Stratix IV TriMatrix memory blocks support the following clocking modes: "Independent Clock Mode" on page 3-18 "Input/Output Clock Mode" on page 3-18 "Read/Write Clock Mode" on page 3-18 "Single Clock Mode" on page 3-18 c Violating the setup or hold time on the memory block address registers could corrupt memory contents. This applies to both read and write operations. Table 3-9 lists which clocking mode/memory mode combinations are supported. Table 3-9. TriMatrix Memory Clock Modes True Dual-Port Mode Simple Dual-Port Mode Single-Port Mode ROM Mode FIFO Mode Independent Y -- -- Y -- Input/output Y Y Y Y -- Read/write -- Y -- -- Y Single clock Y Y Y Y Y Clocking Mode December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-18 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations Independent Clock Mode Stratix IV TriMatrix memory blocks can implement independent clock mode for true dual-port memories. In this mode, a separate clock is available for each port (clock A and clock B). Clock A controls all registers on the port A side; clock B controls all registers on the port B side. Each port also supports independent clock enables for both port A and port B registers, respectively. Asynchronous clears are supported only for output latches and output registers on both ports. Input/Output Clock Mode Stratix IV TriMatrix memory blocks can implement input/output clock mode for true dual-port and simple dual-port memories. In this mode, an input clock controls all registers related to the data input to the memory block including data, address, byte enables, read enables, and write enables. An output clock controls the data output registers. Asynchronous clears are available on output latches and output registers only. Read/Write Clock Mode Stratix IV TriMatrix memory blocks can implement read/write clock mode for simple dual-port memories. In this mode, a write clock controls the data-input, write-address, and write-enable registers. Similarly, a read clock controls the data-output, read-address, and read-enable registers. The memory blocks support independent clock enables for both read and write clocks. Asynchronous clears are available on data output latches and registers only. When using read/write clock mode, if you perform a simultaneous read/write to the same address location, the output read data is unknown. If you require the output data to be a known value, use either single-clock mode or input/output clock mode and choose the appropriate read-during-write behavior in the MegaWizard Plug-In Manager. Single Clock Mode Stratix IV TriMatrix memory blocks can implement single-clock mode for true dual-port, simple dual-port, and single-port memories. In this mode, a single clock, together with a clock enable, is used to control all registers of the memory block. Asynchronous clears are available on output latches and output registers only. Design Considerations This section describes guidelines for designing with TriMatrix memory blocks. Selecting TriMatrix Memory Blocks The Quartus II software automatically partitions user-defined memory into embedded memory blocks by taking into account both speed and size constraints placed on your design. For example, the Quartus II software may spread memory out across multiple memory blocks when resources are available to increase the performance of the design. You can manually assign memory to a specific block size using the RAM MegaWizard Plug-In Manager. Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations 3-19 MLABs can implement single-port SRAM through emulation using the Quartus II software. Emulation results in minimal additional logic resources being used. Because of the dual-purpose architecture of the MLAB, it only has data input registers and output registers in the block. MLABs gain input address registers and additional data output registers from ALMs. f For more information about register packing, refer to the Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices chapter. Conflict Resolution When using memory blocks in true dual-port mode, it is possible to attempt two write operations to the same memory location (address). Because no conflict resolution circuitry is built into the memory blocks, this results in unknown data being written to that location. Therefore, you must implement conflict resolution logic external to the memory block to avoid address conflicts. Read-During-Write Behavior You can customize the read-during-write behavior of the Stratix IV TriMatrix memory blocks to suit your design needs. Two types of read-during-write operations are available: same port and mixed port. Figure 3-15 shows the difference between the two types. Figure 3-15. Stratix IV Read-During-Write Data Flow Port A data in Port B data in Mixed-port data flow Same-port data flow Port A data out Port B data out Same-Port Read-During-Write Mode This mode applies to either a single-port RAM or the same port of a true dual-port RAM. For MLABs, the output of the MLABs can only be set to don't care in same-port read-during-write mode. In this mode, the output of the MLABs is unknown during a write cycle. There is a window near the falling edge of the clock during which the output is unknown. Prior to that window, "old data" is read out; after that window, "new data" is seen at the output. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-20 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations Figure 3-16 shows sample functional waveforms of same-port read-during-write behavior in don't care mode for MLABs. Figure 3-16. MLABs Same-Port Read-During Write: Don't Care Mode clk_a address XX A0 data_in XX FFFF A1 A2 AAAA XXXX wrena q(unregistered) q(registered) XX FFFF A0(old data) AAAA A1(old data) XX FFFF A2(old data) AAAA For M9K and M144K memory blocks, three output choices are available in same-port read-during-write mode: "new data" (or flow-through) or "old data". In new data mode, the "new data" is available on the rising edge of the same clock cycle on which it was written. In old data mode, the RAM outputs reflect the "old data" at that address before the write operation proceeds. In don't care mode, the RAM outputs "unknown values" for a read-during-write operation. Figure 3-17 shows sample functional waveforms of same-port read-during-write behavior in new data mode for M9K and M144K blocks. Figure 3-17. M9K and M144K Blocks Same-Port Read-During-Write: New Data Mode clk_a 0A address 0B rdena wrena bytenna data_a q_a (asyn) Stratix IV Device Handbook Volume 1 01 10 00 A123 B456 C789 XX23 B423 11 B423 DDDD EEEE DDDD EEEE FFFF FFFF December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations 3-21 Figure 3-18 shows sample functional waveforms of same-port read-during-write behavior in old data mode for M9K and M144K blocks. Figure 3-18. M9K and M144K Blocks Same-Port Read-During-Write: Old Data Mode clk_a A0 address A1 rdena wrena 01 10 00 A123 B456 C789 bytenna data_a A0 (old data) DoldDold23 q_a (asyn) 11 DDDD B423 EEEE A1(old data) FFFF DDDD EEEE Mixed-Port Read-During-Write Mode This mode applies to a RAM in simple or true dual-port mode that has one port reading from and the other port writing to the same address location with the same clock. In this mode, you have two output choices if you use the output register: "old data," or "don't care". With MLABs, you also have the output register "new data." In old data mode, a read-during-write operation to different ports causes the RAM outputs to reflect the "old data" at that address location. In don't care mode, the same operation results in a "don't care" or "unknown" value on the RAM outputs. f Read-during-write behavior is controlled with the RAM MegaWizard Plug-In Manager. For more information, refer to the Internal Memory (RAM and ROM) User Guide. Figure 3-19 shows a sample functional waveform of mixed-port read-during-write behavior for old data mode in MLABs. Figure 3-19. MLABs Mixed-Port Read-During-Write: Old Data Mode clk_a wraddress A0 A1 rdaddress A0 A1 data_in AAAA BBBB CCCC DDDD EEEE FFFF 01 10 11 01 10 AAAA AABB A1(old data) DDDD wrena byteena_a q_b(registered) December 2011 Altera Corporation 11 A0 (old data) DDEE Stratix IV Device Handbook Volume 1 3-22 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations Figure 3-20 shows a sample functional waveform of mixed-port read-during-write behavior for don't care mode in MLABs. Figure 3-20. MLABs Mixed-Port Read-During-Write: Don't Care Mode clk_a wraddress A0 A1 rdaddress A0 A1 data_in AAAA BBBB CCCC DDDD EEEE FFFF 01 10 11 01 10 AAAA AABB CCBB DDDD DDEE wrena byteena_a 11 q_b(registered) FFEE Figure 3-21 shows a sample functional waveform of mixed-port read-during-write behavior for old data mode in M9K and M144K blocks. Figure 3-21. M9K and M144K Blocks Mixed-Port Read-During Write: Old Data Mode clk_a&b wrena address_a A1 A0 data_a AAAA BBBB CCCC bytenna 11 01 10 DDDD EEEE FFFF 11 rdenb address_b q_b_(asyn) Stratix IV Device Handbook Volume 1 A1 A0 A0 (old data) AAAA AABB A1(old data) DDDD EEEE December 2011 Altera Corporation Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations 3-23 Figure 3-22 shows a sample functional waveform of mixed-port read-during-write behavior for don't care mode in M9K and M144K blocks. Figure 3-22. M9K and M144K Blocks Mixed-Port Read-During Write: Don't Care Mode clk_a&b wrena A0 address_a A1 data_a AAAA BBBB CCCC bytenna 11 01 10 DDDD EEEE FFFF 11 rdenb address_b q_b_(asyn) A0 A1 XXXX (unknown data) Mixed-port read-during-write is not supported when two different clocks are used in a dual-port RAM. The output value is unknown during a dual-clock mixed-port read-during-write operation. Power-Up Conditions and Memory Initialization M9K memory cells are initialized to all zeros through a default .mif file in the Quartus II software. However, you may specify your own initialization of the memory cells through a defined .mif file. M144K memory cells are not initialized and; therefore, come up in an undefined state. This is to prevent the programming file from being too large. Again, you may specify your own initialization of the memory cells through a defined .mif file. MLABs power up to zero if output registers are used and power up reading the memory contents if output registers are not used. You must take this into consideration when designing logic that might evaluate the initial power-up values of the MLAB memory block. For Stratix IV devices, the Quartus II software initializes the RAM cells to zero unless there is a .mif file specified. As mentioned, all memory blocks support initialization using a .mif file. You can create .mif files in the Quartus II software and specify their use with the RAM MegaWizard Plug-In Manager when instantiating a memory in your design. Even if a memory is pre-initialized (for example, using a .mif file), it still powers up with its outputs cleared. f For more information about .mif files, refer to the Internal Memory (RAM and ROM) User Guide and the Quartus II Handbook. December 2011 Altera Corporation Stratix IV Device Handbook Volume 1 3-24 Chapter 3: TriMatrix Embedded Memory Blocks in Stratix IV Devices Design Considerations Power Management Stratix IV memory block clock-enables allow you to control clocking of each memory block to reduce AC power consumption. Use the read-enable signal to ensure that read operations only occur when you need them to. If your design does not need read-during-write, you can reduce your power consumption by de-asserting the read-enable signal during write operations, or any period when no memory operations occur. The Quartus II software automatically places any unused memory blocks in low-power mode to reduce static power. Document Revision History Table 3-10 lists the revision history for this chapter. Table 3-10. Document Revision History Date Version December 2011 February 2011 March 2010 November 2009 June 2009 April 2009 3.3 3.2 3.1 3.0 Changes Updated the "Byte Enable Support" and "Mixed-Port Read-During-Write Mode" sections. Updated Table 3-1. Updated the "Byte Enable Support" and "Power-Up Conditions and Memory Initialization" sections. Applied new template. Minor text edits. Updated the "Simple Dual-Port Mode", "Same-Port Read-During-Write Mode", and "Mixed-Port Read-During-Write Mode" sections. Updated Figure 3-14. Minor text edits. Updated Table 3-2. Updated the "Simple Dual-Port Mode" section. Minor text edits. Updated graphics. Updated Table 3-1 and Figure 3-2. Updated the "Introduction", "Byte Enable Support", "Mixed Width Support", "Asynchronous Clear", "Single-Port RAM", "Simple Dual-Port Mode", "True Dual-Port Mode", "FIFO Mode", and "Read/Write Clock Mode" sections. Added introductory sentences to improve search ability. Removed the Conclusion section. Minor text edits. Updated Table 3-2. Updated Table 3-2. Removed "Referenced Documents" section. 2.3 2.2 March 2009 2.1 November 2008 2.0 Updated "Power-Up Conditions and Memory Initialization" on page 3-20 May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 1 December 2011 Altera Corporation 4. DSP Blocks in Stratix IV Devices February 2011 SIV51004-3.1 SIV51004-3.1 This chapter describes how the Stratix(R) IV device digital signal processing (DSP) blocks are optimized to support DSP applications requiring high data throughput, such as finite impulse response (FIR) filters, infinite impulse response (IIR) filters, fast Fourier transform (FFT) functions, and encoders. You can configure the DSP blocks to implement one of several operational modes to suit your application. The built-in shift register chain, multipliers, and adders/subtractors minimize the amount of external logic to implement these functions, resulting in efficient resource usage and improved performance and data throughput for DSP applications. Many complex systems, such as WiMAX, 3GPP WCDMA, high-performance computing (HPC), voice over Internet protocol (VoIP), H.264 video compression, medical imaging, and HDTV use sophisticated digital signal processing techniques, which typically require a large number of mathematical computations. Stratix IV devices are ideally suited for these tasks because the DSP blocks consist of a combination of dedicated elements that perform multiplication, addition, subtraction, accumulation, summation, and dynamic shift operations. Along with the high-performance Stratix IV soft logic fabric and TriMatrix memory structures, you can configure DSP blocks to build sophisticated fixed-point and floating-point arithmetic functions. These can be manipulated easily to implement common, larger computationally intensive subsystems such as FIR filters, complex FIR filters, IIR filters, FFT functions, and discrete cosine transform (DCT) functions. This chapter contains the following sections: "Stratix IV DSP Block Overview" on page 4-2 "Stratix IV Simplified DSP Operation" on page 4-4 "Stratix IV Operational Modes Overview" on page 4-8 "Stratix IV DSP Block Resource Descriptions" on page 4-9 "Stratix IV Operational Mode Descriptions" on page 4-15 "Software Support" on page 4-35 (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 4-2 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Overview Stratix IV DSP Block Overview Each Stratix IV device has two to seven columns of DSP blocks that implement multiplication, multiply-add, multiply-accumulate (MAC), and dynamic shift functions efficiently. Architectural highlights of the Stratix IV DSP block include: High-performance, power optimized, fully registered, and pipelined multiplication operations Natively supported 9-, 12-, 18-, and 36-bit wordlengths Natively supported 18-bit complex multiplications Efficiently supported floating-point arithmetic formats (24-bit for single precision and 53-bit for double precision) Signed and unsigned input support Built-in addition, subtraction, and accumulation units to combine multiplication results efficiently Cascading 18-bit input bus to form the tap-delay line for filtering applications Cascading 44-bit output bus to propagate output results from one block to the next block without external logic support Rich and flexible arithmetic rounding and saturation units Efficient barrel shifter support Loopback capability to support adaptive filtering Table 4-1 lists the number of DSP blocks for the Stratix IV device family. Family Stratix IV E Stratix IV GX DSP Blocks Table 4-1. Number of DSP Blocks in Stratix IV Devices (Part 1 of 2) Device Independent Input and Output Multiplication Operators High-Precision Multiplier Adder Mode Four Multiplier Adder Mode 9x9 Multipliers 12 x 12 Multipliers 18 x 18 Multipliers 18 x 18 Complex 36 x 36 Multipliers 18 x 36 Multipliers 18 x 18 Multipliers EP4SE230 161 1,288 966 644 322 322 644 1288 EP4SE360 130 1,040 780 520 260 260 520 1040 EP4SE530 128 1,024 768 512 256 256 512 1024 EP4SE820 120 960 720 480 240 240 480 960 EP4SGX70 48 384 288 192 96 96 192 384 EP4SGX110 64 512 384 256 128 128 256 512 EP4SGX180 115 920 690 460 230 230 460 920 EP4SGX230 161 1,288 966 644 322 322 644 1288 EP4SGX290 104 832 624 416 208 208 416 832 EP4SGX360 (1) 130 1,040 780 520 260 260 520 1,040 EP4SGX360 (2) 128 1,024 768 512 256 256 512 1,024 128 1,024 768 512 256 256 512 1,024 EP4SGX530 Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Overview 4-3 Family DSP Blocks Table 4-1. Number of DSP Blocks in Stratix IV Devices (Part 2 of 2) Device Stratix IV GT Independent Input and Output Multiplication Operators High-Precision Multiplier Adder Mode Four Multiplier Adder Mode 9x9 Multipliers 12 x 12 Multipliers 18 x 18 Multipliers 18 x 18 Complex 36 x 36 Multipliers 18 x 36 Multipliers 18 x 18 Multipliers EP4S40G2 161 1,288 966 644 322 322 644 1,288 EP4S40G5 128 1,024 768 512 256 256 512 1,024 EP4S100G2 161 1,288 966 644 322 322 644 1,288 EP4S100G3 104 832 624 416 208 208 416 832 EP4S100G4 128 1,024 768 512 256 256 512 1,024 EP4S100G5 128 1,024 768 512 256 256 512 1,024 Notes to Table 4-1: (1) This is applicable for all packages in EP4SGX360 except F1932. (2) This is applicable for EP4SGX360F1932 only. Table 4-1 shows that the largest Stratix IV DSP-centric device provides up to 1288 18 x 18 multiplier functionality in the 36 x 36, complex 18 x 18, and summation modes. Each DSP block occupies four LABs in height and can be divided further into two half blocks that share some common clock signals, but are for all common purposes identical in functionality. Figure 4-1 shows the layout of each DSP block. Figure 4-1. Overview of DSP Block Signals 34 Control 144 Input Data Half-DSP Block 72 Output Data 72 Output Data 288 144 Half-DSP Block Full DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-4 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Simplified DSP Operation Stratix IV Simplified DSP Operation In Stratix IV devices, the fundamental building block is a pair of 18 x 18-bit multipliers followed by a first-stage 37-bit addition/subtraction unit, as shown in Equation 4-1 and Figure 4-2. 1 All signed numbers, input, and output data are represented in 2's-complement format only. Equation 4-1. Multiplier Equation P[36..0] = A0[17..0] x B0[17..0] A1[17..0] x B1[17..0] Figure 4-2. Basic Two-Multiplier Adder Building Block A0[17..0] B0[17..0] +/A1[17..0] D Q B1[17..0] D Q P[36..0] The structure shown in Figure 4-2 is useful for building more complex structures, such as complex multipliers and 36 x 36 multipliers, as described in later sections. Each Stratix IV DSP block contains four two-multiplier adder units (2 two-multiplier adder units per half block). Therefore, there are eight 18 x 18 multiplier functionalities per DSP block. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Simplified DSP Operation 4-5 Following the two-multiplier adder units are the pipeline registers, the second-stage adders, and an output register stage. You can configure the second-stage adders to provide the alternative functions per half block, as shown in Equation 4-2 and Equation 4-3. Equation 4-2. Four-Multiplier Adder Equation Z[37..0] = P0[36..0] + P1[36..0] Equation 4-3. Four-Multiplier Adder Equation (44-Bit Accumulation) Wn[43..0] = Wn-1[43..0] Zn[37..0] In these equations, n denotes sample time and P[36..0] denotes the result from the two-multiplier adder units. Equation 4-2 provides a sum of four 18 x 18-bit multiplication operations (four-multiplier adder). Equation 4-3 provides a four 18 x 18-bit multiplication operation but with a maximum 44-bit accumulation capability by feeding the output of the unit back to itself, as shown in Figure 4-3. Depending on the mode you select, you can bypass all register stages except accumulation and loopback mode. In these two modes, one set of registers must be enabled. If the register set is not enabled, an infinite loop occurs. Output Register Bank Adder/ Accumulator 144 Pipeline Register Bank Input Data Input Register Bank Figure 4-3. Four-Multiplier Adder and Accumulation Capability 44 Result[] Half-DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-6 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Simplified DSP Operation To support commonly found FIR-like structures efficiently, a major addition to the DSP block in Stratix IV devices is the ability to propagate the result of one half block to the next half block completely within the DSP block without additional soft logic overhead. This is achieved by the inclusion of a dedicated addition unit and routing that adds the 44-bit result of a previous half block with the 44-bit result of the current block. The 44-bit result is either fed to the next half block or out of the DSP block using the output register stage, as shown in Figure 4-4. Detailed examples are described in later sections. The combination of a fast, low-latency four-multiplier adder unit and the "chained cascade" capability of the output chaining adder provides the optimal FIR and vector multiplication capability. To support single-channel type FIR filters efficiently, you can configure one of the multiplier input's registers to form a tap delay line input, saving resources and providing higher system performance. Figure 4-4. Output Cascading Feature for FIR Structures From Previous Half DSP Block Half DSP Block Output Register Bank Round/Saturate Adder/ Accumulator 144 Pipeline Register Bank Input Data Input Register Bank 44 44 Result[] 44 To Next Half DSP Block Also shown in Figure 4-4 is the optional rounding and saturation unit (RSU). This unit provides a rich set of commonly found arithmetic rounding and saturation functions used in signal processing. In addition to the independent multipliers and sum modes, you can use DSP blocks to perform shift operations. DSP blocks can dynamically switch between logical shift left/right, arithmetic shift left/right, and rotation operation in one clock cycle. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Simplified DSP Operation 4-7 Figure 4-5 shows a top-level view of the Stratix IV DSP block. Figure 4-6 on page 4-9 shows a more detailed top-level view of the DSP block. Figure 4-5. Stratix IV Full DSP Block From Previous Half DSP Block Output Multiplexer Round/Saturate Output Register Bank Output Multiplexter Round/Saturate Output Register Bank Adder/Accumulator 144 Pipeline Register Bank Input Data Input Register Bank 44 Result[] Top Half DSP Block Adder/Accumulator 144 Pipeline Register Bank Input Data Input Register Bank 44 Result[] Bottom Half DSP Block To Next Half DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-8 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Modes Overview Stratix IV Operational Modes Overview You can use each Stratix IV DSP block in one of five basic operational modes. Table 4-2 lists the five basic operational modes and the number of multipliers that you can implement within a single DSP block, depending on the mode. Table 4-2. Stratix IV DSP Block Operation Modes Multiplier in Width # of Mults # per Block Signed or Unsigned RND, SAT In Shift Register Chainout Adder 1st Stage Add/Sub 2nd Stage Add/Acc 9 bits 1 8 Both No No No -- -- 12 bits 1 6 Both No No No -- -- 18 bits 1 4 Both Yes Yes No -- -- 36 bits 1 2 Both No No No -- -- Double 1 2 Both No No No -- -- Two-Multiplier Adder (1) 18 bits 2 4 Signed Yes No No Both -- Four-Multiplier Adder 18 bits 4 2 Both Yes Yes Yes Both Add Only Multiply Accumulate 18 bits 4 2 Both Yes Yes Yes Both Both 1 2 Both No No -- -- -- 2 2 Both No No No -- Add Only Mode Independent Multiplier Shift (2) High Precision Multiplier Adder 36 bits (3) 1836 (4) Notes to Table 4-2: (1) This mode also supports loopback mode. In loopback mode, the number of loopback multipliers per DSP block is two. You can use the remaining multipliers in regular two-multiplier adder mode. (2) Dynamic shift mode supports arithmetic shift left, arithmetic shift right, logical shift left, logical shift right, and rotation operation. (3) Dynamic shift mode operates on a 32-bit input vector but the multiplier width is configured as 36 bits. (4) Unsigned value is also supported but you must ensure that the result can be contained within 36 bits. The DSP block consists of two identical halves (the top half and bottom half). Each half has four 18 x 18 multipliers. The Quartus(R) II software includes megafunctions used to control the mode of operation of the multipliers. After making the appropriate parameter settings using the megafunction's MegaWizard Plug-In Manager, the Quartus II software automatically configures the DSP block. Stratix IV DSP blocks can operate in different modes simultaneously. Each half block is fully independent except for the sharing of the three clock, ena, and aclr signals. For example, you can break down a single DSP block to operate a 9 x 9 multiplier in one half block and an 18 x 18 two-multiplier adder in the other half block. This increases DSP block resource efficiency and allows you to implement more multipliers within a Stratix IV device. The Quartus II software automatically places multipliers that can share the same DSP block resources within the same block. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions 4-9 Stratix IV DSP Block Resource Descriptions The DSP block consists of the following elements: Input register bank Four two-multiplier adders Pipeline register bank Two second-stage adders Four rounding and saturation logic units Second adder register and output register bank Figure 4-6 shows a detailed overall architecture of the top half of the DSP block. Table 4-9 on page 4-34 shows a list of DSP block dynamic signals. Figure 4-6. Half DSP Block Architecture clock[3..0] ena[3..0] alcr[3..0] chainin[ ] (3) signa signb output_round output_saturate rotate shift_right zero_loopback accum_sload zero_chainout chainout_round chainout_saturate overflow (1) chainout_sat_overflow (2) scanina[ ] datab_3[ ] Multiplexer Shift/Rotate Output Register Bank Second Round/Saturate Chainout Adder Second Adder Register Bank dataa_3[ ] First Round/Saturate datab_2[ ] Second Stage Adder/Accumulator dataa_2[ ] Pipeline Register Bank datab_1[ ] Input Register Bank datab_0[ ] dataa_1[ ] First Stage Adder loopback First Stage Adder dataa_0[ ] result[ ] Half-DSP Block scanouta chainout Notes to Figure 4-6: (1) Block output for accumulator overflow and saturate overflow. (2) Block output for saturation overflow of chainout. (3) The chainin port must only be connected to chainout of the previous DSP blocks and must not be connected to general routings. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-10 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions Input Registers All of the DSP block registers are triggered by the positive edge of the clock signal and are cleared after power up. Each multiplier operand can feed an input register or go directly to the multiplier, bypassing the input registers. The following DSP block signals control the input registers within the DSP block: clock[3..0] ena[3..0] aclr[3..0] Every DSP block has nine 18-bit data input register banks per half DSP block. Every half DSP block has the option to use the eight data register banks as inputs to the four multipliers. The special ninth register bank is a delay register required by modes that use both the cascade and chainout features of the DSP block. Use the ninth register bank to balance the latency requirements when using the chained cascade feature. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions 4-11 A feature of the input register bank is to support a tap delay line. Therefore, the top leg of the multiplier input (A) can be driven from general routing or from the cascade chain, as shown in Figure 4-7. Table 4-9 on page 4-34 lists the DSP block dynamic signals. Figure 4-7. Input Register of a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb scanina[17..0] dataa_0[17..0] loopback datab_0[17..0] +/- dataa_1[17..0] datab_1[17..0] dataa_2[17..0] datab_2[17..0] +/- dataa_3[17..0] datab_3[17..0] Delay Register scanouta At compile time, you must select whether the A-input comes from general routing or from the cascade chain. In cascade mode, the dedicated shift outputs from one multiplier block and directly feeds the input registers of the adjacent multiplier below it (within the same half DSP block) or the first multiplier in the next half DSP block, to form an 8-tap shift register chain per DSP Block. The DSP block can increase the length of the shift register chain by cascading to the lower DSP blocks. The dedicated shift register chain spans a single column, but you can implement longer shift register chains requiring multiple columns using the regular FPGA routing resources. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-12 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions Shift registers are useful in DSP functions such as FIR filters. When implementing 18 x 18 or smaller width multipliers, you do not need external logic to create the shift register chain because the input shift registers are internal to the DSP block. This implementation significantly reduces the logical element (LE) resources required, avoids routing congestion, and results in predictable timing. The first multiplier in every half DSP block (top- and bottom-half) in Stratix IV devices has a multiplexer for the first multiplier B-input (lower-leg input) register to select between general routing and loopback, as shown in Figure 4-6 on page 4-9. In loopback mode, the most significant 18-bit registered outputs are connected as feedback to the multiplier input of the first top multiplier in each half DSP block. Loopback modes are used by recursive filters where the previous output is needed to compute the current output. Loopback mode is described in "Two-Multiplier Adder Sum Mode" on page 4-22. Table 4-3 lists input register modes for the DSP block. Table 4-3. Input Register Modes Register Input Mode (1) Parallel input Shift register input Loopback input (2) (3) 9x9 12 x 12 18 x 18 36 x 36 Double Y Y Y Y Y -- -- Y -- -- -- -- Y -- -- Notes to Table 4-3: (1) Multiplier operand input wordlengths are statically configured at compile time. (2) Available only on the A-operand. (3) Only one loopback input is allowed per half block. For more information, refer to Figure 4-15 on page 4-24. Multiplier and First-Stage Adder The multiplier stage natively supports 9 x 9, 12 x 12, 18 x 18, or 36 x 36 multipliers. Other wordlengths are padded up to the nearest appropriate native wordlength; for example, 16 x 16 would be padded up to use 18 x 18. For more information, refer to "Independent Multiplier Modes" on page 4-15. Depending on the data width of the multiplier, a single DSP block can perform many multiplications in parallel. Each multiplier operand can be a unique signed or unsigned number. Two dynamic signals, signa and signb, control the representation of each operand, respectively. A logic 1 value on the signa/signb signal indicates that data A/data B is a signed number; a logic 0 value indicates an unsigned number. Table 4-4 lists the sign of the multiplication result for the various operand sign representations. The result of the multiplication is signed if any one of the operands is a signed value. Table 4-4. Multiplier Sign Representation Stratix IV Device Handbook Volume 1 Data A (signa Value) Data B (signb Value) Result Unsigned (logic 0) Unsigned (logic 0) Unsigned Unsigned (logic 0) Signed (logic 1) Signed Signed (logic 1) Unsigned (logic 0) Signed Signed (logic 1) Signed (logic 1) Signed February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions 4-13 Each half block has its own signa and signb signal. Therefore, all of the data A inputs feeding the same half DSP block must have the same sign representation. Similarly, all of the data B inputs feeding the same half DSP block must have the same sign representation. The multiplier offers full precision regardless of the sign representation in all operational modes except for full precision 18 x 18 loopback and two-multiplier adder modes. For more information, refer to "Two-Multiplier Adder Sum Mode" on page 4-22. 1 By default, when the signa and signb signals are unused, the Quartus II software sets the multiplier to perform unsigned multiplication. Figure 4-6 on page 4-9 shows that the outputs of the multipliers are the only outputs that can feed into the first-stage adder. There are four first-stage adders in a DSP block (two adders per half DSP block). The first-stage adder block has the ability to perform addition and subtraction. The control signal for addition or subtraction is static and has to be configured after compile time. The first-stage adders are used by the sum modes to compute the sum of two multipliers, 18 x 18-complex multipliers, and to perform the first stage of a 36 x 36 multiply and shift operations. Depending on your specifications, the output of the first-stage adder has the option to feed into the pipeline registers, second-stage adder, rounding and saturation unit, or output registers. Pipeline Register Stage Figure 4-6 on page 4-9 shows that the output from the first-stage adder can either feed or bypass the pipeline registers. Pipeline registers increase the DSP block's maximum performance (at the expense of extra cycles of latency), especially when using the subsequent DSP block stages. Pipeline registers split up the long signal path between the input registers/multiplier/first-stage adder and the second-stage adder/ round-and-saturation/output registers, creating two shorter paths. Second-Stage Adder There are four individual 44-bit second-stage adders per DSP block (two adders per half DSP block). You can configure the second-stage adders as follows: 1 The final stage of a 36-bit multiplier A sum of four (18 x 18) An accumulator (44-bits maximum) A chained output summation (44-bits maximum) You can use the chained-output adder at the same time as a second-level adder in chained output summation mode. The output of the second-stage adder has the option to go into the rounding and saturation logic unit or the output register. 1 February 2011 You cannot use the second-stage adder independently from the multiplier and first-stage adder. Altera Corporation Stratix IV Device Handbook Volume 1 4-14 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV DSP Block Resource Descriptions Rounding and Saturation Stage The rounding and saturation logic units are located at the output of the 44-bit second-stage adder (the rounding logic unit followed by the saturation logic unit). There are two rounding and saturation logic units per half DSP block. The input to the rounding and saturation logic unit can come from one of the following stages: Output of the multiplier (independent multiply mode in 18 x 18) Output of the first-stage adder (two-multiplier adder) Output of the pipeline registers Output of the second-stage adder (four-multiplier adder and multiply-accumulate mode in 18 x 18) These stages are described in "Stratix IV Operational Mode Descriptions" on page 4-15. The rounding and saturation logic unit is controlled by the dynamic rounding and saturate signals, respectively. A logic 1 value on the rounding and/or saturate signals enables the rounding and/or saturate logic unit, respectively. 1 You can use the rounding and saturation logic units together or independently. Second Adder and Output Registers The second adder register and output register banks are two banks of 44-bit registers that you can combine to form larger 72-bit banks to support 36 x 36 output results. The outputs of the different stages in the Stratix IV devices are routed to the output registers through an output selection unit. Depending on the operational mode of the DSP block, the output selection unit selects whether the outputs of the DSP blocks comes from the outputs of the multiplier block, first-stage adder, pipeline registers, second-stage adder, or the rounding and saturation logic unit. The output selection unit is set automatically by the software, based on the DSP block operational mode you specified, and has the option to either drive or bypass the output registers. The exception is when you use the block in shift mode, in which case you dynamically control the output-select multiplexer directly. When the DSP block is configured in chained cascaded output mode, both of the second-stage adders are used. Use the first one for performing a four-multiplier adder; use the second for the chainout adder. The outputs of the four-multiplier adder are routed to the second-stage adder registers before they enter the chainout adder. The output of the chainout adder goes to the regular output register bank. Depending on the configuration, you can route the chainout results to the input of the next half block's chainout adder input or to the general fabric (functioning as regular output registers). For more information, refer to "Stratix IV Operational Mode Descriptions" on page 4-15. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-15 The second-stage and output registers are triggered by the positive edge of the clock signal and are cleared after power up. The following DSP block signals control the output registers within the DSP block: clock[3..0] ena[3..0] aclr[3..0] Stratix IV Operational Mode Descriptions This section contains an explanation of different operational modes in Stratix IV devices. Independent Multiplier Modes In independent input and output multiplier mode, the DSP block performs individual multiplication operations for general-purpose multipliers. 9-, 12-, and 18-Bit Multiplier You can configure each DSP block multiplier for 9-, 12-, or 18-bit multiplication. A single DSP block can support up to eight individual 9 x 9 multipliers, six individual 12 x 12 multipliers, or four individual 18 x 18 multipliers. For operand widths up to 9 bits, a 9 x 9 multiplier is implemented. For operand widths from 10 to 12 bits, a 12 x 12 multiplier is implemented, and for operand widths from 13 to 18 bits, an 18 x 18 multiplier is implemented. This is done by the Quartus II software by zero-padding the LSBs. Figure 4-8, Figure 4-9, and Figure 4-10 show the DSP block in the independent multiplier operation. Table 4-9 on page 4-34 lists the dynamic signals for the DSP block. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-16 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Figure 4-8. 18-Bit Independent Multiplier Mode Shown for a Half DSP Block signa clock[3..0] signb overflow (1) Pipeline Register Bank 18 dataa_1[17..0] Input Register Bank 18 datab_0[17..0] 18 datab_1[17..0] 36 result_0[ ] Output Register Bank 18 dataa_0[17..0] Round/Saturate output_round output_saturate Round/Saturate ena[3..0] aclr[3..0] 36 result_1[ ] Half-DSP Block Note to Figure 4-8: (1) Block output for accumulator overflow and saturate overflow. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-17 Figure 4-9. 12-Bit Independent Multiplier Mode Shown for a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb 12 dataa_0[11..0] 24 result_0[ ] 12 Output Register Bank 12 datab_1[11..0] Pipeline Register Bank 12 dataa_1[11..0] Input Register Bank datab_0[11..0] 24 result_1[ ] 12 dataa_2[11..0] 24 result_2[ ] 12 datab_2[11..0] Half-DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-18 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Figure 4-10. 9-Bit Independent Multiplier Mode Shown for a Half Block clock[3..0] ena[3..0] aclr[3..0] signa signb 9 dataa_0[8..0] 18 result_0[ ] 9 datab_0[8..0] 9 dataa_1[8..0] Output Register Bank 9 dataa_2[8..0] Pipeline Register Bank 9 datab_1[8..0] Input Register Bank 18 result_1[ ] 18 result_2[ ] 9 datab_2[8..0] 9 dataa_3[8..0] 18 result_3[ ] 9 datab_3[8..0] Half-DSP Block The multiplier operands can accept signed integers, unsigned integers, or a combination of both. You can change the signa and signb signals dynamically and can register the signals in the DSP block. Additionally, the multiplier inputs and results can be registered independently. You can use the pipeline registers within the DSP block to pipeline the multiplier result, increasing the performance of the DSP block. 1 Stratix IV Device Handbook Volume 1 The rounding and saturation logic unit is supported for 18-bit independent multiplier mode only. February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-19 36-Bit Multiplier You can efficiently construct a 36 x 36 multiplier using four 18 x 18 multipliers. This simplification fits conveniently into one half DSP block and is implemented in the DSP block automatically by selecting 36 x 36 mode. Stratix IV devices can have up to two 36-bit multipliers per DSP block (one 36-bit multiplier per half DSP block). The 36-bit multiplier is also under the independent multiplier mode but uses the entire half DSP block, including the dedicated hardware logic after the pipeline registers to implement the 36 x 36 bit multiplication operation, as shown in Figure 4-11. The 36-bit multiplier is useful for applications requiring more than 18-bit precision; for example, for the mantissa multiplication portion of single precision and extended single precision floating-point arithmetic applications. Figure 4-11. 36-Bit Independent Multiplier Mode Shown for a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb dataa_0[35..18] datab_0[35..18] datab_0[17..0] + Output Register Bank dataa_0[35..18] Input Register Bank datab_0[35..18] Pipeline Register Bank + dataa_0[17..0] 72 result[ ] + dataa_0[17..0] datab_0[17..0] Half-DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-20 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Double Multiplier You can configure the Stratix IV DSP block to efficiently support a signed or unsigned 54 x 54-bit multiplier that is required to compute the mantissa portion of an IEEE double-precision floating point multiplication. You can build a 54 x 54-bit multiplier using basic 18 x 18 multipliers, shifters, and adders. In order to efficiently use the Stratix IV DSP block's built-in shifters and adders, a special double mode (partial 54 x 54 multiplier) is available that is a slight modification to the basic 36 x 36 multiplier mode, as shown in Figure 4-12 and Figure 4-13. Figure 4-12. Double Mode Shown for a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb dataa_0[35..18] datab_0[35..18] datab_0[17..0] + Output Register Bank dataa_0[35..18] Input Register Bank datab_0[35..18] Pipeline Register Bank + dataa_0[17..0] 72 result[ ] + dataa_0[17..0] datab_0[17..0] Half-DSP Block Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-21 Figure 4-13. Unsigned 54 x 54 Multiplier for a Half-DSP Block clock[3..0] ena[3..0] aclr[3..0] "0" "0" dataa[53..36] signa signb Two Multiplier Adder Mode + 36 datab[53..36] dataa[35..18] Double Mode 55 datab[35..18] dataa[53..36] datab[17..0] dataa[35..18] Final Adder (implemented with ALUT logic) datab[53..36] dataa[53..36] Shifters and Adders datab[53..36] dataa[17..0] 108 result[ ] 36 x 36 Mode datab[35..18] dataa[35..18] Shifters and Adders datab[35..18] dataa[17..0] 72 datab[17..0] dataa[17..0] datab[17..0] Unsigned 54 X 54 Multiplier February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-22 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Two-Multiplier Adder Sum Mode In a two-multiplier adder configuration, the DSP block can implement four 18-bit two-multiplier adders (2 two-multiplier adders per half DSP block). You can configure the adders to take the sum or difference of two multiplier outputs. You must select summation or subtraction at compile time. The two-multiplier adder function is useful for applications such as FFTs, complex FIR, and IIR filters. Figure 4-14 on page 4-23 shows the DSP block configured in two-multiplier adder mode. Loopback mode is the other sub-feature of the two-multiplier adder mode. Figure 4-15 on page 4-24 shows the DSP block configured in the loopback mode. This mode takes the 36-bit summation result of the two multipliers and feeds back the most significant 18-bits to the input. The lower 18-bits are discarded. You have the option to disable or zero-out the loopback data by using the dynamic zero_loopback signal. A logic 1 value on the zero_loopback signal selects the zeroed data or disables the looped back data, while a logic 0 selects the looped back data. 1 You must select the option to use loopback mode or the general two-multiplier adder mode at compile time. For two-multiplier adder mode, if all the inputs are full 18-bit and unsigned, the result requires 37 bits. As the output data width in two-multiplier adder mode is limited to 36 bits, this 37-bit output requirement is not allowed. Any other combination that does not violate the 36-bit maximum result is permitted; for example, two 16 x 16 signed two-multiplier adders is valid. Two-multiplier adder mode supports the rounding and saturation logic unit. You can use the pipeline registers and output registers within the DSP block to pipeline the multiplier-adder result, increasing the performance of the DSP block. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-23 Figure 4-14. Two-Multiplier Adder Mode Shown for a Half DSP Block signa clock[3..0] ena[3..0] aclr[3..0] signb output_round output_saturate overflow (1) Output Register Bank Round/Saturate dataa_1[17..0] + Pipeline Register Bank datab_0[17..0] Input Register Bank dataa_0[17..0] result[ ] datab_1[17..0] Half-DSP Block Note to Figure 4-14: (1) Block output for accumulator overflow and saturate overflow. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-24 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Figure 4-15. Loopback Mode for a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb output_round output_saturate zero_loopback overflow (1) Output Register Bank dataa_1[17..0] + Round/Saturate datab_0[17..0] Pipeline Register Bank loopback Input Register Bank dataa_0[17..0] result[ ] datab_1[17..0] Half-DSP Block Note to Figure 4-15: (1) Block output for accumulator overflow and saturate overflow. 18 x 18 Complex Multiply You can configure the DSP block to implement complex multipliers using two-multiplier adder mode. A single half DSP block can implement one 18-bit complex multiplier. Equation 4-4 shows a complex multiplication. Equation 4-4. Complex Multiplication Equation (a + jb) x (c + jd) = ((a x c) - (b x d)) + j((a x d) + (b x c)) Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-25 To implement this complex multiplication within the DSP block, the real part ((a x c) - (b x d)) is implemented using two multipliers feeding one subtractor block while the imaginary part ((a x d) + (b x c)) is implemented using another two multipliers feeding an adder block. Figure 4-16 shows an 18-bit complex multiplication. This mode automatically assumes all inputs are using signed numbers. Figure 4-16. Complex Multiplier Using Two-Multiplier Adder Mode clock[3..0] ena[3..0] aclr[3..0] signa signb A C 36 AxC BxD Real Part Output Register Bank Pipeline Register Bank D Input Register Bank B 36 AxD BxC Imaginary Part Half-DSP Block February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-26 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Four-Multiplier Adder In the four-multiplier adder configuration shown in Figure 4-17, the DSP block can implement two four-multiplier adders (one four-multiplier adder per half DSP block). These modes are useful for implementing one-dimensional and two-dimensional filtering applications. The four-multiplier adder is performed in two addition stages. The outputs of two of the four multipliers are initially summed in the two first-stage adder blocks. The results of these two adder blocks are then summed in the second-stage adder block to produce the final four-multiplier adder result, as shown by Equation 4-2 on page 4-5 and Equation 4-3 on page 4-5. Figure 4-17. Four-Multiplier Adder Mode Shown for a Half DSP Block signa signb clock[3..0] ena[3..0] aclr[3..0] output_round output_saturate overflow (1) dataa_0[ ] datab_0[ ] + Output Register Bank + Round/Saturate dataa_2[ ] Input Register Bank datab_1[ ] Pipeline Register Bank dataa_1[ ] result[ ] datab_2[ ] + dataa_3[ ] datab_3[ ] Half-DSP Block Note to Figure 4-17: (1) Block output for accumulator overflow and saturate overflow. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-27 Four-multiplier adder mode supports the rounding and saturation logic unit. You can use the pipeline registers and output registers within the DSP block to pipeline the multiplier-adder result, increasing the performance of the DSP block. High-Precision Multiplier Adder Mode In a high-precision multiplier adder configuration, shown in Figure 4-18 on page 4-28, the DSP block can implement 2 two-multiplier adders, with multiplier precision of 18 x 36 (one two-multiplier adder per half DSP block). This mode is useful in filtering or FFT applications where a data path greater than 18 bits is required, yet 18 bits is sufficient for the coefficient precision. This can occur where the data has a high dynamic range. If the coefficients are fixed, as in FFT and most filter applications, the precision of 18 bits provide a dynamic range over 100 dB, if the largest coefficient is normalized to the maximum 18-bit representation. In these situations, the data path can be up to 36 bits, allowing sufficient capacity for bit growth or gain changes in the signal source without loss of precision. This mode is also extremely useful in single precision block floating point applications. The high-precision multiplier adder is performed in two stages. The 18 x 36 multiply is divided into two 18 x 18 multipliers. The multiplier with the LSB of the data source is performed unsigned, while the multiplier with the MSB of the data source can be signed or unsigned. The latter multiplier has its result left shifted by 18 bits prior to the first adder stage, creating an effective 18 x 36 multiplier. The results of these two adder blocks are then summed in the second stage adder block to produce the final result: Z[54..0] = P0[53..0] + P1[53..0] where: P0 = A[17..0] x B[35..0] P1 = C[17..0] x D[35..0] February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-28 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Figure 4-18. High-Precision Multiplier Adder Configuration signa signb clock[3..0] ena[3..0] aclr[3..0] overflow (1) dataA[0:17] dataB[0:17] Pipeline Register Bank dataC[0:17] Input Register Bank dataB[18:35] P0 <<18 + Output Register Bank + dataA[0:17] result[ ] dataD[0:17] + P1 dataC[0:17] <<18 dataD[18:35] Half-DSP Block Note to Figure 4-18: (1) Block output for accumulator overflow and saturate overflow. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-29 Multiply Accumulate Mode In multiply accumulate mode, the second-stage adder is configured as a 44-bit accumulator or subtractor. The output of the DSP block is looped back to the second-stage adder and added or subtracted with the two outputs of the first-stage adder block according to Equation 4-3 on page 4-5. Figure 4-19 shows the DSP block configured to operate in multiply accumulate mode. Figure 4-19. Multiply Accumulate Mode Shown for a Half DSP Block clock[3..0] ena[3..0] aclr[3..0] signa signb output_round output_saturate chainout_sat_overflow (1) accum_sload dataa_0[ ] datab_0[ ] + Output Register Bank Round/Saturate + Second Register Bank dataa_2[ ] Input Register Bank datab_1[ ] Pipeline Register Bank dataa_1[ ] 44 result[ ] datab_2[ ] + dataa_3[ ] datab_3[ ] Half-DSP Block Note to Figure 4-19: (1) Block output for saturation overflow of chainout. A single DSP block can implement up to two independent 44-bit accumulators. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-30 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Use the dynamic accum_sload control signal to clear the accumulation. A logic 1 value on the accum_sload signal synchronously loads the accumulator with the multiplier result only, while a logic 0 enables accumulation by adding or subtracting the output of the DSP block (accumulator feedback) to the output of the multiplier and first-stage adder. 1 You must configure the control signal for the accumulator and subtractor if static at compile time. This mode supports the rounding and saturation logic unit because it is configured as an 18-bit multiplier accumulator. You can use the pipeline registers and output registers within the DSP block to increase the performance of the DSP block. Shift Modes Stratix IV devices support the following shift modes for 32-bit input only: 1 Arithmetic shift left, ASL[N] Arithmetic shift right, ASR[32-N] Logical shift left, LSL[N] Logical shift right, LSR[32-N] 32-bit rotator or barrel shifter, ROT[N] You can switch between these modes using the dynamic rotate and shift control signals. You can use shift mode in a Stratix IV device by using a soft embedded processor such as Nios(R) II to perform the dynamic shift and rotate operation. Figure 4-20 on page 4-31 shows the shift mode configuration. Shift mode makes use of the available multipliers to logically or arithmetically shift left, right, or rotate the desired 32-bit data. You can configure the DSP block similar to the independent 36-bit multiplier mode to perform shift mode operations. Arithmetic shift right requires a signed input vector. During an arithmetic shift right, the sign is extended to fill the MSB of the 32-bit vector. The logical shift right uses an unsigned input vector. During a logical shift right, zeros are padded in the MSBs, shifting the 32-bit vector to the right. The barrel shifter uses unsigned input vector and implements a rotation function on a 32-bit word length. Two control signals, rotate and shift_right, together with the signa and signb signals, determine the shifting operation. Table 4-5 on page 4-31 lists examples of shift operations. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-31 Figure 4-20. Shift Operation Mode Shown for a Half DSP Block signa signb rotate shift_right clock[3..0] ena[3..0] aclr[3..0] dataa_0[35..18] datab_0[35..18] + dataa_0[35..18] + Shift/Rotate datab_0[35..18] Output Register Bank Input Register Bank Pipeline Register Bank dataa_0[17..0] 32 result[ ] datab_0[17..0] + dataa_0[17..0] datab_0[17..0] Half-DSP Block Table 4-5. Examples of Shift Operations Example Signa Signb Shift Rotate A-input B-input Result Logical Shift Left LSL[N] Unsigned Unsigned 0 0 0xAABBCCDD 0x0000100 0xBBCCDD00 Logical Shift Right LSR[32-N] Unsigned Unsigned 1 0 0xAABBCCDD 0x0000100 0x000000AA Arithmetic Shift Left ASL[N] Signed Unsigned 0 0 0xAABBCCDD 0x0000100 0xBBCCDD00 Arithmetic Shift Right ASR[32-N] Signed Unsigned 1 0 0xAABBCCDD 0x0000100 0xFFFFFFAA Unsigned Unsigned 0 1 0xAABBCCDD 0x0000100 0xBBCCDDAA Rotation ROT[N] February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-32 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Rounding and Saturation Mode Rounding and saturation functions are often required in DSP arithmetic. Use rounding to limit bit growth and its side effects; use saturation to reduce overflow and underflow side effects. Two rounding modes are supported in Stratix IV devices: 1 Round-to-nearest-integer mode Round-to-nearest-even mode You must select one of these two options at compile time. Round-to-nearest-integer provides the biased rounding support and is the simplest form of rounding commonly used in DSP arithmetic. The round-to-nearest-even method provides unbiased rounding support and is used where DC offsets are a concern. Table 4-6 lists how round-to-nearest-even works. Table 4-6. Example of Round-To-Nearest-Even Mode 6- to 4-bits Rounding Odd/Even (Integer) Fractional Add to Integer Result 010111 x > 0.5 (11) 1 0110 001101 x < 0.5 (01) 0 0011 001010 Even (0010) = 0.5 (10) 0 0010 001110 Odd (0011) = 0.5 (10) 1 0100 110111 x > 0.5 (11) 1 1110 101101 x < 0.5 (01) 0 1011 110110 Odd (1101) = 0.5 (10) 1 1110 110010 Even (1100) = 0.5 (10) 0 1100 Table 4-7 lists examples of the difference between the two modes. In this example, a 6-bit input is rounded to 4 bits. The main difference between the two rounding options is when the residue bits are exactly halfway between its nearest two integers and the LSB is zero (even). Table 4-7. Comparison of Round-to-Nearest-Integer and Round-to-Nearest-Even Stratix IV Device Handbook Volume 1 Round-To-Nearest-Integer Round-To-Nearest-Even 010111 0110 010111 0110 001101 0011 001101 0011 001010 0011 001010 0010 001110 0100 001110 0100 110111 1110 110111 1110 101101 1011 101101 1011 110110 1110 110110 1110 110010 1101 110010 1100 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions 4-33 Two saturation modes are supported in Stratix IV: 1 Asymmetric saturation mode Symmetric saturation mode You must select one of the two options at compile time. In 2's-complement format, the maximum negative number that can be represented is -2(n-1), while the maximum positive number is 2(n-1) - 1. Symmetrical saturation limits the maximum negative number to -2(n-1) + 1. For example, for 32 bits: Asymmetric 32-bit saturation: Max = 0x7FFFFFFF, Min = 0x80000000 Symmetric 32-bit saturation: Max = 0x7FFFFFFF, Min = 0x80000001 Table 4-8 lists how saturation works. In this example, a 44-bit input is saturated to 36-bits. Table 4-8. Examples of Saturation 44- to 36-Bits Saturation Symmetric SAT Result Asymmetric SAT Result 5926AC01342h 7FFFFFFFFh 7FFFFFFFFh ADA38D2210h 800000001h 800000000h Stratix IV devices have up to 16 configurable bit positions out of the 44-bit bus ([43:0]) for the rounding and saturate logic unit, providing higher flexibility. These 16-bit positions are located at bits [21:6] for rounding and [43:28] for saturation, as shown in Figure 4-21. 1 You must select the 16 configurable bit positions at compile time. Figure 4-21. Rounding and Saturation Locations 16 User defined SAT Positions (bit 43-28) 43 42 29 28 1 0 16 User defined RND Positions (bit 21-6) 43 42 1 February 2011 21 20 7 6 0 For symmetric saturation, the RND bit position is also used to determine where the LSP for the saturated data is located. Altera Corporation Stratix IV Device Handbook Volume 1 4-34 Chapter 4: DSP Blocks in Stratix IV Devices Stratix IV Operational Mode Descriptions Use the rounding and saturation function just described in regular supported multiplication operations, as specified in Table 4-2 on page 4-8. However, for accumulation-type operations, use the following convention: The functionality of the round logic unit is in the format of: Result = RND[S(A x B)], when used for an accumulation type of operation. Likewise, the functionality of the saturation logic unit is in the format of: Result = SAT[S(A x B)], when used for an accumulation type of operation. If you use both the rounding and saturation logic units for an accumulation type of operation, the format is: Result = SAT[RND[S(A x B)]] DSP Block Control Signals The Stratix IV DSP block is configured using a set of static and dynamic signals. You can configure the DSP block dynamic signals. You can set the signals to toggle or not toggle at run time. Table 4-9 lists the dynamic signals for the DSP block. Table 4-9. DSP Block Dynamic Signals (Part 1 of 2) Signal Name Function Count signa Signed/unsigned control for all multipliers and adders. signb signa for "multiplicand" input bus to dataa[17:0] to each multiplier signb for "multiplier" input bus datab[17:0] to each multiplier signa = 1, signb = 1 for signed-signed multiplication signa = 1, signb = 0 for signed-unsigned multiplication signa = 0, signb = 1 for unsigned-signed multiplication signa = 0, signb = 0 for unsigned-unsigned multiplication 2 Round control for the first stage round and saturation block. output_round output_round = 1 for rounding on multiply output output_round = 0 for normal multiply output 1 Round control for the second stage round and saturation block. chainout_round output_saturate chainout_saturate Stratix IV Device Handbook Volume 1 chainout_round = 1 for rounding multiply output chainout_round = 0 for normal multiply output 1 Saturation control for the first stage round and saturation block for Q-format multiply. If you enable both rounding and saturation, saturation is done on the rounded result. output_saturate = 1 for saturation support output_saturate = 0 for no saturation support Saturation control for the second stage round and saturation block for Q-format multiply. If you enable both rounding and saturation, saturation is done on the rounded result. chainout_saturate = 1 for saturation support chainout_saturate = 0 for no saturation support 1 1 February 2011 Altera Corporation Chapter 4: DSP Blocks in Stratix IV Devices Software Support 4-35 Table 4-9. DSP Block Dynamic Signals (Part 2 of 2) Signal Name Function Count Dynamically specifies whether the accumulator value is zero. accum_sload accum_sload = 0, accumulation input is from the output registers accum_sload = 1, accumulation input is set to zero 1 zero_chainout Dynamically specifies whether the chainout value is zero. 1 zero_loopback Dynamically specifies whether the loopback value is zero. 1 rotate rotate = 1, the rotation feature is enabled 1 shift_right shift_right = 1, the shift right feature is enabled 1 Total Signals per Half Block 11 clock0 clock1 clock2 DSP-block-wide clock signals. 4 Input and Pipeline Register enable signals. 4 DSP block-wide asynchronous clear signals (active low). 4 clock3 ena0 ena1 ena2 ena3 aclr0 aclr1 aclr2 aclr3 Total Count per Full Block 34 Software Support Altera provides two distinct methods for implementing various modes of the DSP block in a design--instantiation and inference. Both methods use the following Quartus II megafunctions: lpm_mult altmult_add altmult_accum altfp_mult To use the DSP block, instantiate the megafunctions in the Quartus II software. Alternatively, with inference, create an HDL design and synthesize it using a third-party synthesis tool (such as LeonardoSpectrum, Synplify, or Quartus II Native Synthesis) that infers the appropriate megafunction by recognizing multipliers, multiplier adders, multiplier accumulators, and shift functions. Using either method, the Quartus II software maps the functionality to the DSP blocks during compilation. f For instructions about using these megafunctions and the MegaWizard Plug-In Manager, refer to Quartus II software Help. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 4-36 Chapter 4: DSP Blocks in Stratix IV Devices Software Support f For more information, refer to the "Synthesis" section in volume 1 of the Quartus II Handbook. Document Revision History Table 4-10 lists the revision history for this chapter. Table 4-10. Document Revision History Date Version February 2011 November 2009 3.1 3.0 June 2009 2.3 April 2009 2.2 March 2009 2.1 November 2008 May 2008 Stratix IV Device Handbook Volume 1 2.0 1.0 Changes Applied new template. Minor text edits. Updated Table 4-1. Updated "Stratix IV Simplified DSP Operation" section. Updated graphics. Minor text edits. Added an introductory paragraph to increase search ability. Removed the Conclusion section. Updated Table 4-1. Updated Table 4-1. Removed "Referenced Documents" section. Updated Table 4-2. Updated Figure 4-16. Updated Figure 4-18. Initial release. February 2011 Altera Corporation 5. Clock Networks and PLLs in Stratix IV Devices September 2012 SIV51005-3.4 SIV51005-3.4 This chapter describes the hierarchical clock networks and phase-locked loops (PLLs) which have advanced features in Stratix(R) IV devices. It includes details about the ability to reconfigure the PLL counter clock frequency and phase shift in real time, allowing you to sweep PLL output frequencies and dynamically adjust the output clock phase shift. The Quartus(R) II software enables the PLLs and their features without external devices. The following sections describe the Stratix IV clock networks and PLLs in detail: "Clock Networks in Stratix IV Devices" on page 5-1 "PLLs in Stratix IV Devices" on page 5-19 Clock Networks in Stratix IV Devices The global clock networks (GCLKs), regional clock networks (RCLKs), and periphery clock networks (PCLKs) available in Stratix IV devices are organized into hierarchical clock structures that provide up to 236 unique clock domains (16 GCLKs + 88 RCLKs + 132 PCLKs) within the Stratix IV device and allow up to 71 unique GCLK, RCLK, and PCLK clock sources (16 GCLKs + 22 RCLKs + 33 PCLKs) per device quadrant. Table 5-1 lists the clock resources available in Stratix IV devices. Table 5-1. Clock Resources in Stratix IV Devices (Part 1 of 2) Clock Resource Number of Resources Available Source of Clock Resource Clock input pins 32 Single-ended (16 Differential) CLK[0..15]p and CLK[0..15]n pins GCLK networks 16 CLK[0..15]p and CLK[0..15]n pins, PLL clock outputs, and logic array RCLK networks 64/88 PCLK networks GCLKs/RCLKs per quadrant (1) 56/88/112/132 (33 per device quadrant) (2) 32/38 (3) CLK[0..15]p and CLK[0..15]n pins, PLL clock outputs, and logic array DPA clock outputs, PLD-transceiver interface clocks, horizontal I/O pins, and logic array 16 GCLKs + 16 RCLKs 16 GCLKs + 22 RCLKs (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 September 2012 Feedback Subscribe 5-2 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Table 5-1. Clock Resources in Stratix IV Devices (Part 2 of 2) Clock Resource GCLKs/RCLKs per device Number of Resources Available 80/104 (4) Source of Clock Resource 16 GCLKs + 64 RCLKs 16 GCLKs + 88 RCLKs Notes to Table 5-1: (1) There are 64 RCLKs in the EP4S40G2, EP4S100G2, EP4SE230, EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 devices. There are 88 RCLKs in the EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE360, EP4SE530, EP4SE820, EP4SGX290, EP4SGX360, and EP4SGX530 devices. (2) There are 56 PCLKs in the EP4SGX70, and EP4SGX110 devices. There are 88 PCLKs in the EP4S40G2, EP4S100G2, EP4SE230, EP4SE360, EP4SGX180, EP4SGX230, EP4SGX290, and EP4SGX360 devices. There are 112 PCLKs in the EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE530 and EP4SGX530 devices. There are 132 PCLKs in the EP4SE820 device. (3) There are 32 GCLKs/RCLKs per quadrant in the EP4S40G2, EP4S100G2, EP4SE230, EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 devices. There are 38 GCLKs/RCLKs per quadrant in the EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE360, EP4SE530, EP4SE820, EP4SGX290, EP4SGX360, and EP4SGX530 devices. (4) There are 80 GCLKs/RCLKs per entire device in the EP4S40G2, EP4S100G2, EP4SE230, EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 devices. There are 104 GCLKs/RCLKS per entire device in the EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE360, EP4SE530, EP4SE820, EP4SGX290, EP4SGX360, and EP4SGX530 devices. Stratix IV devices have up to 32 dedicated single-ended clock pins or 16 dedicated differential clock pins (CLK[0..15]p and CLK[0..15]n) that can drive either the GCLK or RCLK networks. These clock pins are arranged on the four sides of the Stratix IV device, as shown in Figure 5-1 through Figure 5-4 on page 5-5. f For more information about how to connect the clock input pins, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-3 Global Clock Networks Stratix IV devices provide up to 16 GCLKs that can drive throughout the device, serving as low-skew clock sources for functional blocks such as adaptive logic modules (ALMs), digital signal processing (DSP) blocks, TriMatrix memory blocks, and PLLs. Stratix IV device I/O elements (IOEs) and internal logic can also drive GCLKs to create internally generated global clocks and other high fan-out control signals; for example, synchronous or asynchronous clears and clock enables. Figure 5-1 shows the CLK pins and PLLs that can drive the GCLK networks in Stratix IV devices. Figure 5-1. GCLK Networks CLK[12..15] T1 T2 L1 R1 GCLK[12..15] CLK[0..3] L2 GCLK[0..3] L3 GCLK[8..11] R2 CLK[8..11] R3 GCLK[4..7] L4 R4 B1 B2 CLK[4..7] September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-4 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Regional Clock Networks RCLK networks only pertain to the quadrant they drive into. RCLK networks provide the lowest clock delay and skew for logic contained within a single device quadrant. The Stratix IV device IOEs and internal logic within a given quadrant can also drive RCLKs to create internally generated regional clocks and other high fan-out control signals; for example, synchronous or asynchronous clears and clock enables. Figure 5-2 through Figure 5-4 on page 5-5 show the CLK pins and PLLs that can drive the RCLK networks in Stratix IV devices. Figure 5-2. RCLK Networks (EP4SE230, EP4SGX70, and EP4SGX110 Devices) (1) CLK[12..15] T1 RCLK[54..63] RCLK[44..53] RCLK[38..43] RCLK[0..5] CLK[0..3] L2 Q1 Q2 Q4 Q3 RCLK[6..11] R2 CLK[8..11] RCLK[32..37] RCLK[12..21] RCLK[22..31] B1 CLK[4..7] Note to Figure 5-2: (1) A maximum of four signals from the core can drive into each group of RCLKs. For example, only four core signals can drive into RCLK[0..5] and another four core signals can drive into RCLK[54..63] at any one time. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-5 Figure 5-3. RCLK Networks (EP4S40G2, EP4S100G2, EP4SGX180, and EP4SGX230 Devices) (1) CLK[12..15] T1 T2 RCLK[54..63] RCLK[44..53] RCLK[0..5] RCLK[38..43] Q1 Q2 L2 CLK[0..3] L3 R2 CLK[8..11] R3 Q4 Q3 RCLK[6..11] RCLK[32..37] RCLK[12..21] RCLK[22..31] B1 B2 CLK[4..7] Note to Figure 5-3: (1) A maximum of four signals from the core can drive into each group of RCLKs. For example, only four core signals can drive into RCLK[0..5] and another four core signals can drive into RCLK[54..63] at any one time. Figure 5-4. RCLK Networks (EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE360, EP4SE530, EP4SE820, EP4SGX290, EP4SGX360, and EP4SGX530 Devices) (1), (2), (3) CLK[12..15] T1 T2 L1 R1 RCLK[82..87] RCLK[54..63] RCLK[44..53] RCLK[76..81] RCLK[0..5] CLK[0..3] L2 L3 RCLK[38..43] Q1 Q2 Q4 Q3 RCLK[6..11] R2 CLK[8..11] R3 RCLK[32..37] RCLK[64..69] RCLK[12..21] RCLK[22..31] RCLK[70..75] L4 R4 B1 B2 CLK[4..7] Notes to Figure 5-4: (1) The corner RCLK[64..87] can only be fed by their respective corner PLL outputs. For more information about connectivity, refer to Table 5-6 on page 5-13. (2) The EP4S40G5 and EP4SE360 devices have up to eight PLLs. For more information about PLL availability, refer to Table 5-7 on page 5-19. (3) A maximum of four signals from the core can drive into each group of RCLKs. For example, only four core signals can drive into RCLK[0..5] and another four core signals can drive into RCLK[54..63] at any one time. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-6 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Periphery Clock Networks PCLK networks shown in Figure 5-5 through Figure 5-8 on page 5-8 are collections of individual clock networks driven from the periphery of the Stratix IV device. Clock outputs from the dynamic phase aligner (DPA) block, programmable logic device (PLD)-transceiver interface clocks, I/O pins, and internal logic can drive the PCLK networks. PCLKs have higher skew when compared with GCLK and RCLK networks. You can use PCLKs for general purpose routing to drive signals into and out of the Stratix IV device. Figure 5-5. PCLK Networks (EP4SGX70 and EP4SGX110 Devices) CLK[12..15] T1 PCLK[42..56] PCLK[0..13] CLK[0..3] L2 Q1 Q2 Q4 Q3 PCLK[14..27] R2 CLK[8..11] PCLK[28..41] B1 CLK[4..7] Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-7 Figure 5-6. PCLK Networks (EP4S40G2, EP4S100G2, EP4SE230, EP4SE360, EP4SGX180, EP4SGX230, EP4SGX290, and EP4SGX360 Devices) (1) CLK[12..15] T1 T2 CLK[0..3] PCLK[0..10] PCLK[77..87] PCLK[11..21] PCLK[66..76] L2 Q1 Q2 R2 L3 Q4 Q3 R3 PCLK[22..32] PCLK[55..65] PCLK[33..43] PCLK[44..54] CLK[8..11] B1 B2 CLK[4..7] Note to Figure 5-6: (1) The EP4SE230 device has four PLLs. The EP4SGX290 and EP4SGX360 devices have up to 12 PLLs. For more information about PLL availability, refer to Table 5-7 on page 5-19. Figure 5-7. PCLK Networks (EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, EP4SE530, and EP4SGX530 Devices) (1) CLK[12..15] T1 T2 L1 R1 PCLK[98..111] PCLK[0..13] PCLK[14..27] CLK[0..3] PCLK[84..97] L2 Q1 Q2 R2 L3 Q4 Q3 R3 PCLK[28..41] PCLK[70..83] PCLK[42..55] PCLK[56..69] L4 CLK[8..11] R4 B1 B2 CLK[4..7] Note to Figure 5-7: (1) The EP4S40G5 device has eight PLLs. For more information about PLL availability, refer to Table 5-7 on page 5-19. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-8 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Figure 5-8. PCLK Networks (EP4SE820 Device) CLK[12..15] T1 T2 L1 R1 PCLK[0..15] PCLK[116..131] PCLK[16..32] CLK[0..3] PCLK[99..115] L2 Q1 Q2 R2 L3 Q4 Q3 R3 PCLK[33..49] PCLK[82..98] PCLK[50..65] PCLK[66..81] L4 CLK[8..11] R4 B1 B2 CLK[4..7] Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-9 Clock Sources Per Quadrant There are 26 section clock (SCLK) networks available in each spine clock that can drive six row clocks in each logic array block (LAB) row, nine column I/O clocks, and three core reference clocks. The SCLKs are the clock resources to the core functional blocks, PLLs, and I/O interfaces of the device. Figure 5-9 shows that the SCLKs can be driven by the GCLK, RCLK, PCLK, or the PLL feedback clock networks in each spine clock. 1 A spine clock is another layer of routing below the GCLKs, RCLKs, and PCLKs before each clock is connected to clock routing for each LAB row. The settings for spine clocks are transparent to all users. The Quartus II software automatically routes the spine clock based on the GCLK, RCLK, and PCLKs. Figure 5-9. Hierarchical Clock Networks per Spine Clock (1) 9 GCLK PLL feedback clock (4) 16 3 16 (2) PCLK Column I/O clock (5) SCLK 26 3 Core reference clock (6) 22 (3) RCLK 6 Row clock (7) Notes to Figure 5-9: (1) The GCLK, RCLK, PCLK, and PLL feedback clocks share the same routing to the SCLKs. The total number of clock resources must not exceed the SCLK limits in each region to ensure successful design fitting in the Quartus II software. (2) There are up to 16 PCLKs that can drive the SCLKs in each spine clock in the largest device. (3) There are up to 22 RCLKs that can drive the SCLKs in each spine clock in the largest device. (4) The PLL feedback clock is the clock from the PLL that drives into the SCLKs. (5) The column I/O clock is the clock that drives the column I/O core registers and I/O interfaces. (6) The core reference clock is the clock that feeds into the PLL as the PLL reference clock. (7) The row clock is the clock source to the LAB, memory blocks, and row I/O interfaces in the core row. Clock Regions Stratix IV devices provide up to 104 distinct clock domains (16 GCLKs + 88 RCLKs) in the entire device. You can use these clock resources to form the following types of clock regions: Entire device Regional Dual-regional To form the entire device clock region, a source (not necessarily a clock signal) drives a GCLK network that can be routed through the entire device. This clock region has the maximum delay when compared with other clock regions, but allows the signal to reach every destination within the device. This is a good option for routing global reset and clear signals or routing clocks throughout the device. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-10 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices To form a RCLK region, a source drives a single quadrant of the device. This clock region provides the lowest skew within a quadrant and is a good option if all the destinations are within a single device quadrant. To form a dual-regional clock region, a single source (a clock pin or PLL output) generates a dual-regional clock by driving two RCLK networks (one from each quadrant). This technique allows destinations across two device quadrants to use the same low-skew clock. The routing of this signal on an entire side has approximately the same delay as a RCLK region. Internal logic can also drive a dual-regional clock network. Corner PLL outputs only span one quadrant, they cannot generate a dual-regional clock network. Figure 5-10 shows the dual-regional clock region. Figure 5-10. Stratix IV Dual-Regional Clock Region Clock pins or PLL outputs can drive half of the device to create side-wide clocking regions for improved interface timing. Clock Network Sources In Stratix IV devices, clock input pins, PLL outputs, and internal logic can drive the GCLK and RCLK networks. For connectivity between dedicated pins CLK[0..15] and the GCLK and RCLK networks, refer to Table 5-2 and Table 5-3 on page 5-11. Dedicated Clock Input Pins Clock pins can be either differential clocks or single-ended clocks. Stratix IV devices support 16 differential clock inputs or 32 single-ended clock inputs. You can also use dedicated clock input pins CLK[15..0] for high fan-out control signals such as asynchronous clears, presets, and clock enables for protocol signals such as TRDY and IRDY for PCIe through the GCLK or RCLK networks. LABs You can drive each GCLK and RCLK network using LAB-routing to enable internal logic to drive a high fan-out, low-skew signal. 1 Stratix IV Device Handbook Volume 1 Stratix IV PLLs cannot be driven by internally generated GCLKs or RCLKs. The input clock to the PLL has to come from dedicated clock input pins or pin/PLL-fed GCLKs or RCLKs. September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-11 PLL Clock Outputs Stratix IV PLLs can drive both GCLK and RCLK networks, as described in Table 5-5 on page 5-13 and Table 5-6 on page 5-13. Table 5-2 lists the connection between the dedicated clock input pins and GCLKs. Table 5-2. Clock Input Pin Connectivity to the GCLK Networks CLK (p/n Pins) Clock Resources 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 GCLK0 Y Y Y Y -- -- -- -- -- -- -- -- -- -- -- -- GCLK1 Y Y Y Y -- -- -- -- -- -- -- -- -- -- -- -- GCLK2 Y Y Y Y -- -- -- -- -- -- -- -- -- -- -- -- GCLK3 Y Y Y Y -- -- -- -- -- -- -- -- -- -- -- -- GCLK4 -- -- -- -- Y Y Y Y -- -- -- -- -- -- -- -- GCLK5 -- -- -- -- Y Y Y Y -- -- -- -- -- -- -- -- GCLK6 -- -- -- -- Y Y Y Y -- -- -- -- -- -- -- -- GCLK7 -- -- -- -- Y Y Y Y -- -- -- -- -- -- -- -- GCLK8 -- -- -- -- -- -- -- -- Y Y Y Y -- -- -- -- GCLK9 -- -- -- -- -- -- -- -- Y Y Y Y -- -- -- -- GCLK10 -- -- -- -- -- -- -- -- Y Y Y Y -- -- -- -- GCLK11 -- -- -- -- -- -- -- -- Y Y Y Y -- -- -- -- GCLK12 -- -- -- -- -- -- -- -- -- -- -- -- Y Y Y Y GCLK13 -- -- -- -- -- -- -- -- -- -- -- -- Y Y Y Y GCLK14 -- -- -- -- -- -- -- -- -- -- -- -- Y Y Y Y GCLK15 -- -- -- -- -- -- -- -- -- -- -- -- Y Y Y Y Table 5-3 lists the connectivity between the dedicated clock input pins and RCLKs in Stratix IV devices. A given clock input pin can drive two adjacent RCLK networks to create a dual-regional clock network. Table 5-3. Clock Input Pin Connectivity to the RCLK Networks (Part 1 of 2) CLK (p/n Pins) Clock Resource 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RCLK [0, 4, 6, 10] Y -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- RCLK [1, 5, 7, 11] -- Y -- -- -- -- -- -- -- -- -- -- -- -- -- -- RCLK [2, 8] -- -- Y -- -- -- -- -- -- -- -- -- -- -- -- -- RCLK [3, 9] -- -- -- Y -- -- -- -- -- -- -- -- -- -- -- -- RCLK [13, 17, 21, 23, 27, 31] -- -- -- -- Y -- -- -- -- -- -- -- -- -- -- -- RCLK [12, 16, 20, 22, 26, 30] -- -- -- -- -- Y -- -- -- -- -- -- -- -- -- -- RCLK [15, 19, 25, 29] -- -- -- -- -- -- Y -- -- -- -- -- -- -- -- -- RCLK [14, 18, 24, 28] -- -- -- -- -- -- -- Y -- -- -- -- -- -- -- -- RCLK [35, 41] -- -- -- -- -- -- -- -- Y -- -- -- -- -- -- -- September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-12 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Table 5-3. Clock Input Pin Connectivity to the RCLK Networks (Part 2 of 2) CLK (p/n Pins) Clock Resource 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RCLK [34, 40] -- -- -- -- -- -- -- -- -- Y -- -- -- -- -- -- RCLK [33, 37, 39, 43] -- -- -- -- -- -- -- -- -- -- Y -- -- -- -- -- RCLK [32, 36, 38, 42] -- -- -- -- -- -- -- -- -- -- -- Y -- -- -- -- RCLK [47, 51, 57, 61] -- -- -- -- -- -- -- -- -- -- -- -- Y -- -- -- RCLK [46, 50, 56, 60] -- -- -- -- -- -- -- -- -- -- -- -- -- Y -- -- RCLK [45, 49, 53, 55, 59, 63] -- -- -- -- -- -- -- -- -- -- -- -- -- -- Y -- RCLK [44, 48, 52, 54, 58, 62] -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Y Clock Input Connections to the PLLs Table 5-4 lists the dedicated clock input pin connectivity to Stratix IV PLLs. Table 5-4. Device PLLs and PLL Clock Pin Drivers Dedicated Clock Input Pin CLK (p/n Pins) (1), (2) PLL Number L1 (3) L2 L3 L4 (3) B1 B2 R1 (3) R2 R3 R4 (3) T1 T2 CLK0 Y Y Y Y -- -- -- -- -- -- -- -- CLK1 Y Y Y Y -- -- -- -- -- -- -- -- CLK2 Y Y Y Y -- -- -- -- -- -- -- -- CLK3 Y Y Y Y -- -- -- -- -- -- -- -- CLK4 -- -- -- -- Y Y -- -- -- -- -- -- CLK5 -- -- -- -- Y Y -- -- -- -- -- -- CLK6 -- -- -- -- Y Y -- -- -- -- -- -- CLK7 -- -- -- -- Y Y -- -- -- -- -- -- CLK8 -- -- -- -- -- -- Y Y Y Y -- -- CLK9 -- -- -- -- -- -- Y Y Y Y -- -- CLK10 -- -- -- -- -- -- Y Y Y Y -- -- CLK11 -- -- -- -- -- -- Y Y Y Y -- -- CLK12 -- -- -- -- -- -- -- -- -- -- Y Y CLK13 -- -- -- -- -- -- -- -- -- -- Y Y CLK14 -- -- -- -- -- -- -- -- -- -- Y Y CLK15 -- -- -- -- -- -- -- -- -- -- Y Y Notes to Table 5-4: (1) For single-ended clock inputs, only the CLK<#>p pin has a dedicated connection to the PLL. If you use the CLK<#>n pin, a global clock is used. (2) For the availability of the clock input pins in each device density, refer to the "Stratix IV Device Pin-Out Files" section of the Pin-Out Files for Altera Devices site. (3) These are non-compensated clock input paths. For the compensated input for these PLLs, use the corresponding PLL_[L, R][1,4]_CLK input pin. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 1 5-13 Dedicated clock pins can drive PLLs over dedicated routing; they do not require the global or regional network. Compensated inputs, which are a subset of dedicated clock pins, drive PLLs that can only compensate the input delay when a dedicated clock pin is in the same I/O bank as the PLL used. Clock Output Connections PLLs in Stratix IV devices can drive up to 20 RCLK networks and four GCLK networks. For Stratix IV PLL connectivity to GCLK networks, refer to Table 5-5. The Quartus II software automatically assigns PLL clock outputs to RCLK and GCLK networks. Table 5-5 lists how the PLL clock outputs connect to the GCLK networks. Table 5-5. Stratix IV PLL Connectivity to the GCLK Networks (1) PLL Number Clock Network L1 L2 L3 L4 B1 B2 R1 R2 R3 R4 T1 T2 GCLK0 Y Y Y Y -- -- -- -- -- -- -- -- GCLK1 Y Y Y Y -- -- -- -- -- -- -- -- GCLK2 Y Y Y Y -- -- -- -- -- -- -- -- GCLK3 Y Y Y Y -- -- -- -- -- -- -- -- GCLK4 -- -- -- -- Y Y -- -- -- -- -- -- GCLK5 -- -- -- -- Y Y -- -- -- -- -- -- GCLK6 -- -- -- -- Y Y -- -- -- -- -- -- GCLK7 -- -- -- -- Y Y -- -- -- -- -- -- GCLK8 -- -- -- -- -- -- Y Y Y Y -- -- GCLK9 -- -- -- -- -- -- Y Y Y Y -- -- GCLK10 -- -- -- -- -- -- Y Y Y Y -- -- GCLK11 -- -- -- -- -- -- Y Y Y Y -- -- GCLK12 -- -- -- -- -- -- -- -- -- -- Y Y GCLK13 -- -- -- -- -- -- -- -- -- -- Y Y GCLK14 -- -- -- -- -- -- -- -- -- -- Y Y GCLK15 -- -- -- -- -- -- -- -- -- -- Y Y Note to Table 5-5: (1) Only PLL counter outputs C0 - C3 can drive the GCLK networks. Table 5-6 lists how the PLL clock outputs connect to the RCLK networks. Table 5-6. Stratix IV RCLK Outputs From the PLL Clock Outputs (1) (Part 1 of 2) PLL Number Clock Resource September 2012 L1 L2 L3 L4 B1 B2 R1 R2 R3 R4 T1 T2 RCLK[0..11] -- Y Y -- -- -- -- -- -- -- -- -- RCLK[12..31] -- -- -- -- Y Y -- -- -- -- -- -- RCLK[32..43] -- -- -- -- -- -- -- Y Y -- -- -- RCLK[44..63] -- -- -- -- -- -- -- -- -- -- Y Y Altera Corporation Stratix IV Device Handbook Volume 1 5-14 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices Table 5-6. Stratix IV RCLK Outputs From the PLL Clock Outputs (1) (Part 2 of 2) PLL Number Clock Resource L1 L2 L3 L4 B1 B2 R1 R2 R3 R4 T1 T2 RCLK[64..69] -- -- -- Y -- -- -- -- -- -- -- -- RCLK[70..75] -- -- -- -- -- -- -- -- -- Y -- -- RCLK[76..81] -- -- -- -- -- -- Y -- -- -- -- -- RCLK[82..87] Y -- -- -- -- -- -- -- -- -- -- -- Note to Table 5-6: (1) All PLL counter outputs can drive the RCLK networks. Clock Control Block Every GCLK and RCLK network has its own clock control block. The control block provides the following features: Clock source selection (dynamic selection for GCLKs) Global clock multiplexing Clock power down (static or dynamic clock enable or disable) Figure 5-11 and Figure 5-12 show the GCLK and RCLK select blocks, respectively. You can select the clock source for the GCLK select block either statically or dynamically. You can statically select the clock source using a setting in the Quartus II software or you can dynamically select the clock source using internal logic to drive the multiplexer-select inputs. When selecting the clock source dynamically, you can select either PLL outputs (such as C0 or C1) or a combination of clock pins or PLL outputs. Figure 5-11. Stratix IV GCLK Control Block CLKp Pins PLL Counter Outputs CLKSELECT[1..0] (1) 2 2 CLKn Pin Internal Logic 2 Static Clock Select (2) This multiplexer supports user-controllable dynamic switching Enable/ Disable Internal Logic GCLK Notes to Figure 5-11: (1) When the device is operating in user mode, you can dynamically control the clock select signals through internal logic. (2) When the device is operation in user mode, you can only set the clock select signals through a configuration file (SRAM object file [.sof] or programmer object file [.pof]) and cannot be dynamically controlled. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-15 The mapping between the input clock pins, PLL counter outputs, and clock control block inputs is as follows: inclk[0] and inclk[1]--can be fed by any of the four dedicated clock pins on the same side of the Stratix IV device inclk[2]--can be fed by PLL counters C0 and C2 from the two center PLLs on the same side of the Stratix IV device inclk[3]--can be fed by PLL counters C1 and C3 from the two center PLLs on the same side of the Stratix IV device The corner PLLs (L1, L4, R1, and R4) and the corresponding clock input pins (PLL_L1_CLK and so forth) do not support dynamic selection for the GCLK network. The clock source selection for the GCLK and RCLK networks from the corner PLLs (L1, L4, R1, and R4) and the corresponding clock input pins (PLL_L1_CLK and so forth) are controlled statically using configuration bit settings in the configuration file (.sof or .pof) generated by the Quartus II software. Figure 5-12. RCLK Control Block CLKp Pin PLL Counter Outputs CLKn Pin (2) 2 Internal Logic Static Clock Select (1) Enable/ Disable Internal Logic RCLK Notes to Figure 5-12: (1) When the device is operation in user mode, you can only set the clock select signals through a configuration file (.sof or .pof) and cannot be dynamically controlled. (2) The CLKn pin is not a dedicated clock input when used as a single-ended PLL clock input. You can only control the clock source selection for the RCLK select block statically using configuration bit settings in the configuration file (.sof or .pof) generated by the Quartus II software. You can power down the Stratix IV clock networks using both static and dynamic approaches. When a clock network is powered down, all the logic fed by the clock network is in off-state, thereby reducing the overall power consumption of the device. The unused GCLK and RCLK networks are automatically powered down through configuration bit settings in the configuration file (.sof or .pof) generated by the Quartus II software. The dynamic clock enable or disable feature allows the internal logic to control power-up or power-down synchronously on the GCLK and RCLK networks, including dual-regional clock regions. This function is independent of the PLL and is applied directly on the clock network, as shown in Figure 5-11 and Figure 5-12. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-16 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices You can set the input clock sources and the clkena signals for the GCLK and RCLK network multiplexers through the Quartus II software using the ALTCLKCTRL megafunction. You can also enable or disable the dedicated external clock output pins using the ALTCLKCTRL megafunction. Figure 5-13 shows the external PLL output clock control block. 1 When using the ALTCLKCTRL megafunction to implement dynamic clock source selection, the inputs from the clock pins feed the inclk[0..1] ports of the multiplexer, while the PLL outputs feed the inclk[2..3] ports. You can choose from among these inputs using the CLKSELECT[1..0] signal. f For more information, refer to the Clock Control Block (ALTCLKCTRL) Megafunction User Guide. Figure 5-13. Stratix IV External PLL Output Clock Control Block PLL Counter Outputs 7 or 10 Static Clock Select (1) Enable/ Disable Internal Logic IOE (2) Internal Logic Static Clock Select (1) PLL_<#>_CLKOUT pin Notes to Figure 5-13: (1) When the device is operation in user mode, you can only set the clock select signals through a configuration file (.sof or .pof) and cannot be dynamically controlled. (2) The clock control block feeds to a multiplexer within the PLL_<#>_CLKOUT pin's IOE. The PLL_<#>_CLKOUT pin is a dual-purpose pin. Therefore, this multiplexer selects either an internal signal or the output of the clock control block. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices 5-17 Clock Enable Signals Figure 5-14 shows how the clock enable and disable circuit of the clock control block is implemented in Stratix IV devices. Figure 5-14. clkena Implementation (1) (1) clkena output of clock select mux Q D R1 (2) Q D R2 GCLK/ RCLK/ PLL_<#>_CLKOUT (1) Notes to Figure 5-14: (1) The R1 and R2 bypass paths are not available for the PLL external clock outputs. (2) The select line is statically controlled by a bit setting in the configuration file (.sof or .pof). In Stratix IV devices, the clkena signals are supported at the clock network level instead of at the PLL output counter level. This allows you to gate off the clock even when you are not using a PLL. You can also use the clkena signals to control the dedicated external clocks from the PLLs. Figure 5-15 shows a waveform example for a clock output enable. clkena is synchronous to the falling edge of the clock output. Stratix IV devices also have an additional metastability register that aids in asynchronous enable and disable of the GCLK and RCLK networks. You can optionally bypass this register in the Quartus II software. Figure 5-15. clkena Signals (1) output of clock select mux clkena output of AND gate with R2 bypassed output of AND gate with R2 not bypassed Note to Figure 5-15: (1) You can use the clkena signals to enable or disable the GCLK and RCLK networks or the PLL_<#>_CLKOUT pins. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-18 Chapter 5: Clock Networks and PLLs in Stratix IV Devices Clock Networks in Stratix IV Devices The PLL can remain locked independent of the clkena signals because the loop-related counters are not affected. This feature is useful for applications that require a low-power or sleep mode. The clkena signal can also disable clock outputs if the system is not tolerant of frequency over-shoot during resynchronization. Clock Source Control for PLLs The clock input to Stratix IV PLLs comes from clock input multiplexers. The clock multiplexer inputs come from dedicated clock input pins, PLLs through the GCLK and RCLK networks, or from dedicated connections between adjacent top/bottom and left/right PLLs. The clock input sources to top/bottom and left/right PLLs (L2, L3, T1, T2, B1, B2, R2, and R3) are shown in Figure 5-16; the corresponding clock input sources to left and right PLLs (L1, L4, R1, and R4) are shown in Figure 5-17. The multiplexer select lines are only set in the configuration file (.sof or .pof). After programmed, this block cannot be changed without loading a new configuration file (.sof or .pof). The Quartus II software automatically sets the multiplexer select signals depending on the clock sources selected in the design. Figure 5-16. Clock Input Multiplexer Logic for L2, L3, T1, T2, B1, B2, R2, and R3 PLLs (1) clk[n+3..n] (2) GCLK / RCLK input (3) 4 inclk0 To the clock switchover block Adjacent PLL output (1) inclk1 4 Notes to Figure 5-16: (1) When the device is operating in user mode, input clock multiplexing is controlled through a configuration file (.sof or .pof) only and cannot be dynamically controlled. (2) n=0 for L2 and L3 PLLs; n=4 for B1 and B2 PLLs; n=8 for R2 and R3 PLLs, and n=12 for T1 and T2 PLLs. (3) You can drive the GCLK or RCLK input using an output from another PLL, a pin-driven GCLK or RCLK, or through a clock control block provided the clock control block is fed by an output from another PLL or a pin-driven dedicated GCLK or RCLK. An internally generated global signal or general purpose I/O pin cannot drive the PLL. Figure 5-17. Clock Input Multiplexer Logic for L1, L4, R1, and R4 PLLs PLL_<L1/L4/R1/R4>_CLK (1) inclk0 GCLK/RCLK (2) 4 CLK[0..3] or CLK[8..11] (3) inclk1 4 Notes to Figure 5-17: (1) Dedicated clock input pins to the PLLs are L1, L4, R1, and R4, respectively. For example, PLL_L1_CLK is the dedicated clock input for PLL_L1. (2) You can drive the GCLK or RCLK input using an output from another PLL, a pin-driven GCLK or RCLK, or through a clock control block provided the clock control block is fed by an output from another PLL or a pin-driven dedicated GCLK or RCLK. An internally generated global signal or general purpose I/O pin cannot drive the PLL. (3) The center clock pins can feed the corner PLLs on the same side directly through a dedicated path. However, these paths may not be fully compensated. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-19 Cascading PLLs You can cascade the left/right and top/bottom PLLs through the GCLK and RCLK networks. In addition, where two left/right or top/bottom PLLs exist next to each other, there is a direct connection between them that does not require the GCLK or RCLK network. Using this path reduces clock jitter when cascading PLLs. 1 Stratix IV GX devices allow cascading the left and right PLLs to transceiver PLLs (CMU PLLs and receiver CDRs). f For more information, refer to the "FPGA Fabric PLLs -Transceiver PLLs Cascading" section in the Transceiver Clocking in Stratix IV Devices chapter. When cascading PLLs in Stratix IV devices, the source (upstream) PLL must have a low-bandwidth setting while the destination (downstream) PLL must have a high-bandwidth setting. Ensure that there is no overlap of the bandwidth ranges of the two PLLs. f For more information about PLL cascading in external memory interfaces designs, refer to the External Memory PHY Interface (ALTMEMPHY) (nonAFI) Megafunction User Guide. PLLs in Stratix IV Devices Stratix IV devices offer up to 12 PLLs that provide robust clock management and synthesis for device clock management, external system clock management, and high-speed I/O interfaces. The nomenclature for the PLLs follows their geographical location in the device floor plan. The PLLs that reside on the top and bottom sides of the device are named PLL_T1, PLL_T2, PLL_B1 and PLL_B2; the PLLs that reside on the left and right sides of the device are named PLL_L1, PLL_L2, PLL_L3, PLL_L4, PLL_R1, PLL_R2, PLL_R3, and PLL_R4. Table 5-7 lists the number of PLLs available in the Stratix IV device family. Table 5-7. PLL Availability for Stratix IV Devices (Part 1 of 2) Device Package L1 L2 L3 L4 T1 T2 B1 B2 R1 R2 R3 R4 EP4S40G2 F1517 -- Y Y -- Y Y Y Y -- Y Y -- EP4S40G5 H1517 -- Y Y -- Y Y Y Y -- Y Y -- EP4S100G2 F1517 -- Y Y -- Y Y Y Y -- Y Y -- EP4S100G3 F1932 Y Y Y Y Y Y Y Y Y Y Y Y EP4S100G4 F1932 Y Y Y Y Y Y Y Y Y Y Y Y H1517 -- Y Y -- Y Y Y Y -- Y Y -- F1932 Y Y Y Y Y Y Y Y Y Y Y Y F780 -- Y -- -- Y -- Y -- -- Y -- -- H780 -- Y -- -- Y -- Y -- -- Y -- -- F1152 -- Y Y -- Y Y Y Y -- Y Y -- EP4S100G5 EP4SE230 EP4SE360 September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-20 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Table 5-7. PLL Availability for Stratix IV Devices (Part 2 of 2) Device EP4SE530 EP4SE820 EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 EP4SGX290 EP4SGX360 EP4SGX530 Package L1 L2 L3 L4 T1 T2 B1 B2 R1 R2 R3 R4 H1152 -- Y Y -- Y Y Y Y -- Y Y -- H1517 Y Y Y Y Y Y Y Y Y Y Y Y F1760 Y Y Y Y Y Y Y Y Y Y Y Y H1152 -- Y Y -- Y Y Y Y -- Y Y -- H1517 Y Y Y Y Y Y Y Y Y Y Y Y F1760 Y Y Y Y Y Y Y Y Y Y Y Y F780 -- Y -- -- Y -- Y -- -- -- -- -- F1152 -- Y -- -- Y -- Y -- -- Y -- -- F780 -- Y -- -- Y -- Y -- -- -- -- -- F1152 -- Y -- -- Y -- Y -- -- Y -- -- F780 -- Y -- -- Y -- Y -- -- -- -- -- F1152 -- Y -- -- Y Y Y Y -- Y -- -- F1517 -- Y Y -- Y Y Y Y -- Y Y -- F780 -- Y -- -- Y -- Y -- -- -- -- -- F1152 -- Y -- -- Y Y Y Y -- Y -- -- F1517 -- Y Y -- Y Y Y Y -- Y Y -- H780 -- -- -- -- Y Y Y Y -- -- -- -- F1152 -- Y -- -- Y Y Y Y -- Y -- -- F1517 -- Y Y -- Y Y Y Y -- Y Y -- F1760 Y Y Y Y Y Y Y Y Y Y Y Y F1932 Y Y Y Y Y Y Y Y Y Y Y Y H780 -- -- -- -- Y Y Y Y -- -- -- -- F1152 -- Y -- -- Y Y Y Y -- Y -- -- F1517 -- Y Y -- Y Y Y Y -- Y Y -- F1760 Y Y Y Y Y Y Y Y Y Y Y Y F1932 Y Y Y Y Y Y Y Y Y Y Y Y H1152 -- Y -- -- Y Y Y Y -- Y -- -- H1517 -- Y Y -- Y Y Y Y -- Y Y -- F1760 Y Y Y Y Y Y Y Y Y Y Y Y F1932 Y Y Y Y Y Y Y Y Y Y Y Y All Stratix IV PLLs have the same core analog structure with only minor differences in the features that are supported. Table 5-8 lists the features of top/bottom and left/right PLLs in Stratix IV devices. Table 5-8. PLL Features in Stratix IV Devices (Part 1 of 2) Feature (1) Stratix IV Top/Bottom PLLs Stratix IV Left/Right PLLs C (output) counters 10 7 M, N, C counter sizes 1 to 512 1 to 512 6 single-ended or 4 single-ended and 1 differential pair 2 single-ended or 1 differential pair Dedicated clock outputs Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-21 Table 5-8. PLL Features in Stratix IV Devices (Part 2 of 2) Feature Clock input pins (2) External feedback input pin (1) Stratix IV Top/Bottom PLLs Stratix IV Left/Right PLLs 4 single-ended or 4 differential pin pairs 4 single-ended or 4 differential pin pairs Single-ended or differential Single-ended only Spread-spectrum input clock tracking Yes (3) Yes Through GCLK and RCLK and a dedicated path between adjacent PLLs PLL cascading Compensation modes Through GCLK and RCLK and dedicated path between adjacent PLLs (4) All except LVDS clock network compensation All except external feedback mode when using differential I/Os No Yes PLL drives LVDSCLK and LOADEN VCO output drives the DPA clock Phase shift resolution (3) No Down to 96.125 ps Yes (5) Down to 96.125 ps Programmable duty cycle Yes Yes Output counter cascading Yes Yes Input clock switchover Yes Yes (5) Notes to Table 5-8: (1) While there is pin compatibility, there is no hard IP block placement compatibility. (2) General purpose I/O pins cannot drive the PLL clock input pins. (3) Provided input clock jitter is within input jitter tolerance specifications. (4) The dedicated path between adjacent PLLs is not available on L1, L4, R1, and R4 PLLs. (5) The smallest phase shift is determined by the voltage-controlled oscillator (VCO) period divided by eight. For degree increments, the Stratix IV device can shift all output frequencies in increments of at least 45. Smaller degree increments are possible depending on the frequency and divide parameters. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-22 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Figure 5-18 shows the location of PLLs in Stratix IV devices. Figure 5-18. PLL Locations in Stratix IV Devices Top/Bottom PLLs Top/Bottom PLLs CLK[12..15] T1 T2 PLL_L1_CLK Left/Right PLLs CLK[0..3] Left/Right PLLs PLL_L4_CLK L1 Q1 Q2 L2 L3 Q4 Q3 L4 R1 PLL_R1_CLK R2 R3 CLK[8..11] R4 PLL-R4_CLK Left/Right PLLs Left/Right PLLs B1 B2 CLK[4..7] Top/Bottom PLLs Stratix IV Device Handbook Volume 1 Top/Bottom PLLs September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-23 Stratix IV PLL Hardware Overview Stratix IV devices contain up to 12 PLLs with advanced clock management features. The goal of a PLL is to synchronize the phase and frequency of an internal or external clock to an input reference clock. There are a number of components that comprise a PLL to achieve this phase alignment. Stratix IV PLLs align the rising edge of the input reference clock to a feedback clock using the phase-frequency detector (PFD). The falling edges are determined by the duty-cycle specifications. The PFD produces an up or down signal that determines whether the VCO must operate at a higher or lower frequency. The output of the PFD feeds the charge pump and loop filter, which produces a control voltage for setting the VCO frequency. If the PFD produces an up signal, the VCO frequency increases. A down signal decreases the VCO frequency. The PFD outputs these up and down signals to a charge pump. If the charge pump receives an up signal, current is driven into the loop filter. Conversely, if the charge pump receives a down signal, current is drawn from the loop filter. The loop filter converts these up and down signals to a voltage that is used to bias the VCO. The loop filter also removes glitches from the charge pump and prevents voltage over-shoot, which filters the jitter on the VCO. The voltage from the loop filter determines how fast the VCO operates. A divide counter (m) is inserted in the feedback loop to increase the VCO frequency above the input reference frequency. VCO frequency (fVCO) is equal to (m) times the input reference clock (fREF). The input reference clock (fREF) to the PFD is equal to the input clock (fIN) divided by the pre-scale counter (N). Therefore, the feedback clock (fFB) applied to one input of the PFD is locked to the fREF that is applied to the other input of the PFD. The VCO output from the left and right PLLs can feed seven post-scale counters (C[0..6]), while the corresponding VCO output from the top and bottom PLLs can feed ten post-scale counters (C[0..9]). These post-scale counters allow a number of harmonically related frequencies to be produced by the PLL. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-24 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Figure 5-19 shows a simplified block diagram of the major components of the Stratix IV PLL. Figure 5-19. Stratix IV PLL Block Diagram To DPA block on Left/Right PLLs Lock Circuit pfdena Casade output to adjacent PLL locked /2, /4 /C0 4 /n inclk0 inclk1 GCLK/RCLK Clock Switchover Block PFD CP LF VCO 8 /2 (2) 8 /C1 8 /C2 clkswitch clkbad0 clkbad1 activeclock /C3 Cascade input from adjacent PLL /Cn (1) /m no compensation mode ZDB, External feedback modes LVDS Compensation mode Source Synchronous, normal modes PLL Output Mux GCLKs Dedicated clock inputs RCLKs External clock outputs DIFFIOCLK from Left/Right PLLs LOAD_EN from Left/Right PLLs FBOUT (3) External memory interface DLL FBIN DIFFIOCLK network GCLK/RCLK network Notes to Figure 5-19: (1) The number of post-scale counters is seven for left and right PLLs and ten for top and bottom PLLs. (2) This is the VCO post-scale counter K. (3) The FBOUT port is fed by the M counter in Stratix IV PLLs. 1 You can drive the GCLK or RCLK inputs using an output from another PLL, a pin-driven GCLK or RCLK, or through a clock control block provided the clock control block is fed by an output from another PLL or a pin-driven dedicated GCLK or RCLK. An internally generated global signal or general purpose I/O pin cannot drive the PLL. PLL Clock I/O Pins Each top and bottom PLL supports six clock I/O pins, organized as three pairs of pins: Stratix IV Device Handbook Volume 1 1st pair--two single-ended I/O or one differential I/O 2nd pair--two single-ended I/O or one differential external feedback input (FBp/FBn) 3rd pair--two single-ended I/O or one differential input September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-25 Figure 5-20 shows the clock I/O pins associated with the top and bottom PLLs. Figure 5-20. External Clock Outputs for Top and Bottom PLLs Internal Logic C0 C1 C2 C3 Top/Bottom PLLs C4 C5 C6 C7 C8 C9 m(fbout) clkena0 (3) clkena2 (3) clkena4 (3) clkena1 (3) clkena3 (3) clkena5 (3) PLL_<#>_CLKOUT0p (1), (2) PLL_<#>_FBp/CLKOUT1 (1), (2) PLL_<#>_CLKOUT0n (1), (2) PLL_<#>_CLKOUT3 (1), (2) PLL_<#>_FBn/CLKOUT2 (1), (2) PLL_<#>_CLKOUT4 (1), (2) Notes to Figure 5-20: (1) You can feed these clock output pins using any one of the C[9..0], m counters. (2) The CLKOUT0p and CLKOUT0n pins can be either single-ended or differential clock outputs. The CLKOUT1 and CLKOUT2 pins are dual-purpose I/O pins that you can use as two single-ended outputs or one differential external feedback input pin. The CLKOUT3 and CLKOUT4 pins are two single-ended output pins. (3) These external clock enable signals are available only when using the ALTCLKCTRL megafunction. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-26 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Any of the output counters (C[9..0] on the top and bottom PLLs and C[6..0] on the left and right PLLs) or the M counter can feed the dedicated external clock outputs, as shown in Figure 5-20 and Figure 5-21. Therefore, one counter or frequency can drive all output pins available from a given PLL. Each left and right PLL supports two clock I/O pins, configured as either two single-ended I/Os or one differential I/O pair. When using both pins as single-ended I/Os, one of them can be the clock output while the other pin is the external feedback input (FB) pin. Therefore, for single-ended I/O standards, the left and right PLLs only support external feedback mode. Figure 5-21. External Clock Outputs for Left and Right PLLs Internal Logic C0 C1 C2 LEFT/RIGHT PLLs C3 C4 C5 C6 m(fbout) clkena0 (3) clkena1 (3) PLL_<L2, L3, R2, R3>_CLKOUT0n/FB_CLKOUT0p (1), (2) PLL_<L2, L3, R2, R3>_FB_CLKOUT0p/CLKOUT0n (1), (2) Notes to Figure 5-21: (1) You can feed these clock output pins using any one of the C[6..0], m counters. (2) The CLKOUT0p and CLKOUT0n pins are dual-purpose I/O pins that you can use as two single-ended outputs or one single-ended output and one external feedback input pin. (3) These external clock enable signals are available only when using the ALTCLKCTRL megafunction. Each pin of a single-ended output pair can either be in-phase or 180 out-of-phase. The Quartus II software places the NOT gate in the design into the IOE to implement the 180 phase with respect to the other pin in the pair. The clock output pin pairs support the same I/O standards as standard output pins (in the top and bottom banks) as well as LVDS, LVPECL, differential High-Speed Transceiver Logic (HSTL), and differential SSTL. f To determine which I/O standards are supported by the PLL clock input and output pins, refer to the I/O Features in Stratix IV Devices chapter. Stratix IV PLLs can also drive out to any regular I/O pin through the GCLK or RCLK network. You can also use the external clock output pins as user I/O pins if you do not need external PLL clocking. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-27 PLL Control Signals You can use the pfdena, areset, and locked signals to observe and control PLL operation and resynchronization. pfdena Use the pfdena signal to maintain the most recent locked frequency so your system has time to store its current settings before shutting down. The pfdena signal controls the PFD output with a programmable gate. If you disable PFD, the VCO operates at its most recent set value of control voltage and frequency, with some long-term drift to a lower frequency. The PLL continues running even if it goes out-of-lock or the input clock is disabled. You can use either your own control signal or the control signals available from the clock switchover circuit (activeclock, clkbad[0], or clkbad[1]) to control pfdena. areset The areset signal is the reset or resynchronization input for each PLL. The device input pins or internal logic can drive these input signals. When areset is driven high, the PLL counters reset, clearing the PLL output and placing the PLL out-of-lock. The VCO is then set back to its nominal setting. When areset is driven low again, the PLL resynchronizes to its input as it re-locks. You must assert the areset signal every time the PLL loses lock to guarantee the correct phase relationship between the PLL input and output clocks. You can set up the PLL to automatically reset (self reset) after a loss-of-lock condition using the Quartus II MegaWizard Plug-In Manager. You must include the areset signal in designs if either of the following conditions is true: 1 PLL reconfiguration or clock switchover is enabled in the design Phase relationships between the PLL input and output clocks must be maintained after a loss-of-lock condition If the input clock to the PLL is not toggling or is unstable after power up, assert the areset signal after the input clock is stable and within specifications. locked The locked signal output of the PLL indicates that the PLL has locked onto the reference clock and the PLL clock outputs are operating at the desired phase and frequency set in the Quartus II MegaWizard Plug-In Manager. The lock detection circuit provides a signal to the core logic that gives an indication when the feedback clock has locked onto the reference clock both in phase and frequency. 1 September 2012 Altera recommends using the areset and locked signals in your designs to control and observe the status of your PLL. Altera Corporation Stratix IV Device Handbook Volume 1 5-28 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Clock Feedback Modes Stratix IV PLLs support up to six different clock feedback modes. Each mode allows clock multiplication and division, phase shifting, and programmable duty cycle. Table 5-9 lists the clock feedback modes supported by the Stratix IV device PLLs. Table 5-9. Clock Feedback Mode Availability Availability Clock Feedback Mode Top and Bottom PLLs Left and Right PLLs Source-synchronous Yes Yes No-compensation Yes Yes Normal Yes Yes Yes Yes Zero-delay buffer (ZDB) (1) Yes LVDS compensation No External feedback Yes (2) Yes Notes to Table 5-9: (1) The high-bandwidth PLL setting is not supported in external feedback mode. (2) External feedback mode is supported for single-ended inputs and outputs only on the left and right PLLs. 1 Stratix IV Device Handbook Volume 1 The input and output delays are fully compensated by a PLL only when using the dedicated clock input pins associated with a given PLL as the clock source. For example, when using PLL_T1 in normal mode, the clock delays from the input pin to the PLL clock output-to-destination register are fully compensated, provided the clock input pin is one of the following two pins: CLK14 and CLK15. Compensated pins are only in the same I/O bank as the PLL. When an RCLK or GCLK network drives the PLL, the input and output delays may not be fully compensated in the Quartus II software. Another example is when you configure PLL_T2 in zero-delay buffer mode and the PLL input is driven by a dedicated clock input pin, a fully compensated clock path results in zero-delay between the clock input and one of the output clocks from the PLL. If the PLL input is instead fed by a non-dedicated input (using the GCLK network), the output clock may not be perfectly aligned with the input clock. September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-29 Source Synchronous Mode If data and clock arrive at the same time on the input pins, the same phase relationship is maintained at the clock and data ports of any IOE input register. Figure 5-22 shows an example waveform of the clock and data in this mode. Altera recommends source synchronous mode for source-synchronous data transfers. Data and clock signals at the IOE experience similar buffer delays as long as you use the same I/O standard. Figure 5-22. Phase Relationship Between Clock and Data in Source-Synchronous Mode Data pin PLL reference clock at input pin Data at register Clock at register Source-synchronous mode compensates for the delay of the clock network used plus any difference in the delay between these two paths: Data pin to the IOE register input Clock input pin to the PLL PFD input The Stratix IV PLL can compensate multiple pad-to-input-register paths, such as a data bus when it is set to use source-synchronous compensation mode. You can use the "PLL Compensation" assignment in the Quartus II software Assignment Editor to select which input pins are used as the PLL compensation targets. You can include your entire data bus, provided the input registers are clocked by the same output of a source-synchronous-compensated PLL. In order for the clock delay to be properly compensated, all of the input pins must be on the same side of the device. The PLL compensates for the input pin with the longest pad-to-register delay among all input pins in the compensated bus. If you do not make the "PLL Compensation" assignment, the Quartus II software automatically selects all of the pins driven by the compensated output of the PLL as the compensation target. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-30 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Source-Synchronous Mode for LVDS Compensation The goal of source-synchronous mode is to maintain the same data and clock timing relationship seen at the pins of the internal serializer/deserializer (SERDES) capture register, except that the clock is inverted (180 phase shift). Thus, source-synchronous mode ideally compensates for the delay of the LVDS clock network plus any difference in delay between these two paths: Data pin-to-SERDES capture register Clock input pin-to-SERDES capture register. In addition, the output counter must provide the 180 phase shift Figure 5-23 shows an example waveform of the clock and data in LVDS mode. Figure 5-23. Phase Relationship Between the Clock and Data in LVDS Mode Data pin PLL reference clock at input pin Data at register Clock at register No-Compensation Mode In no-compensation mode, the PLL does not compensate for any clock networks. This mode provides better jitter performance because the clock feedback into the PFD passes through less circuitry. Both the PLL internal- and external-clock outputs are phase-shifted with respect to the PLL clock input. Figure 5-24 shows an example waveform of the PLL clocks' phase relationship in no-compensation mode. Figure 5-24. Phase Relationship Between the PLL Clocks in No Compensation Mode Phase Aligned PLL Reference Clock at the Input Pin PLL Clock at the Register Clock Port (1) External PLL Clock Outputs (1) Note to Figure 5-24: (1) The PLL clock outputs lag the PLL input clocks depending on routine delays. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-31 Normal Mode An internal clock in normal mode is phase-aligned to the input clock pin. The external clock-output pin has a phase delay relative to the clock input pin if connected in this mode. The Quartus II software timing analyzer reports any phase difference between the two. In normal mode, the delay introduced by the GCLK or RCLK network is fully compensated. Figure 5-25 shows an example waveform of the PLL clocks' phase relationship in normal mode. Figure 5-25. Phase Relationship Between the PLL Clocks in Normal Mode Phase Aligned PLL Reference Clock at the Input Pin PLL Clock at the Register Clock Port Dedicated PLL Clock Outputs (1) Note to Figure 5-25: (1) The external clock output can lead or lag the PLL internal clock signals. Zero-Delay Buffer (ZDB) Mode In ZDB mode, the external clock output pin is phase-aligned with the clock input pin for zero-delay through the device. When using this mode, you must use the same I/O standard on the input clocks and output clocks to guarantee clock alignment at the input and output pins. ZDB mode is supported on all Stratix IV PLLs. When using Stratix IV PLLs in ZDB mode, along with single-ended I/O standards, to ensure phase alignment between the CLK pin and the external clock output (CLKOUT) pin, you must instantiate a bi-directional I/O pin in the design to serve as the feedback path connecting the FBOUT and FBIN ports of the PLL. The PLL uses this bi-directional I/O pin to mimic, and compensate for, the output delay from the clock output port of the PLL to the external clock output pin. Figure 5-26 shows ZDB mode in Stratix IV PLLs. When using ZDB mode, you cannot use differential I/O standards on the PLL clock input or output pins. 1 September 2012 The bi-directional I/O pin that you instantiate in your design must always be assigned a single-ended I/O standard. Altera Corporation Stratix IV Device Handbook Volume 1 5-32 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 1 When using ZDB mode, to avoid signal reflection, do not place board traces on the bi-directional I/O pin. Figure 5-26. ZDB Mode in Stratix IV PLLs inclk /n PFD CP/LF VCO /C0 PLL_<#>_CLKOUT# /C1 PLL_<#>_CLKOUT# /m fbout bidirectional I/O pin (1) fbin Note to Figure 5-26: (1) The bidirectional I/O pin must be assigned to the PLL_<#>_FB_CLKOUT0p pin for left and right PLLs and to the PLL_<#>_FBp_/CLKOUT1 pin for top and bottom PLLs. Figure 5-27 shows an example waveform of the PLL clocks' phase relationship in ZDB mode. Figure 5-27. Phase Relationship Between the PLL Clocks in ZDB Mode Phase Aligned PLL Reference Clock at the Input Pin PLL Clock at the Register Clock Port (1) Dedicated PLL Clock Outputs Note to Figure 5-27: (1) The internal PLL clock output can lead or lag the external PLL clock outputs. External Feedback Mode In external feedback mode, the external feedback input pin (fbin) is phase-aligned with the clock input pin, as shown in Figure 5-28. Aligning these clocks allows you to remove clock delay and skew between devices. This mode is supported on all Stratix IV PLLs. In external feedback mode, the output of the M counter (FBOUT) feeds back to the PLL fbin input (using a trace on the board) becoming part of the feedback loop. Also, use one of the dual-purpose external clock outputs as the fbin input pin in this mode. When using external feedback mode, you must use the same I/O standard on the input clock, feedback input, and output clocks. Left and right PLLs support this mode when using single-ended I/O standards only. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-33 Figure 5-28 shows an example waveform of the phase relationship between the PLL clocks in external feedback mode. Figure 5-28. Phase Relationship Between the PLL Clocks in External Feedback Mode Phase Aligned PLL Reference Clock at the Input Pin PLL Clock at the Register Clock Port (1) Dedicated PLL Clock Outputs (1) fbin Clock Input Pin Note to Figure 5-28: (1) The PLL clock outputs can lead or lag the fbin clock input. Figure 5-29 shows external feedback mode implementation in Stratix IV devices. Figure 5-29. External Feedback Mode in Stratix IV Devices inclk /n PFD CP/LF VCO PLL_<#>_CLKOUT# /C0 PLL_<#>_CLKOUT# /C1 /m fbout fbin external board trace Clock Multiplication and Division Each Stratix IV PLL provides clock synthesis for PLL output ports using M/(N* post-scale counter) scaling factors. The input clock is divided by a pre-scale factor, n, and is then multiplied by the m feedback factor. The control loop drives the VCO to match fin (M/N). Each output port has a unique post-scale counter that divides down the high-frequency VCO. For multiple PLL outputs with different frequencies, the VCO is set to the least common multiple of the output frequencies that meets its frequency specifications. For example, if the output frequencies required from one PLL are 33 and 66 MHz, the Quartus II software sets the VCO to 660 MHz (the least common multiple of 33 and 66 MHz within the VCO range). Then the post-scale counters scale down the VCO frequency for each output port. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-34 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Each PLL has one pre-scale counter, n, and one multiply counter, m, with a range of 1 to 512 for both m and n. The n counter does not use duty-cycle control because the only purpose of this counter is to calculate frequency division. There are seven generic post-scale counters per left or right PLL and ten post-scale counters per top or bottom PLL that can feed the GCLKs, RCLKs, or external clock outputs. These post-scale counters range from 1 to 512 with a 50% duty cycle setting. The high- and low-count values for each counter range from 1 to 256. The sum of the high- and low-count values chosen for a design selects the divide value for a given counter. The Quartus II software automatically chooses the appropriate scaling factors according to the input frequency, multiplication, and division values entered into the ALTPLL megafunction. Post-Scale Counter Cascading Stratix IV PLLs support post-scale counter cascading to create counters larger than 512. This is automatically implemented in the Quartus II software by feeding the output of one C counter into the input of the next C counter, as shown in Figure 5-30. Figure 5-30. Counter Cascading VCO Output C0 VCO Output C1 VCO Output C2 VCO Output C3 VCO Output C4 from preceding post-scale counter VCO Output Cn (1) Note to Figure 5-30: (1) N = 6 or N = 9 When cascading post-scale counters to implement a larger division of the high-frequency VCO clock, the cascaded counters behave as one counter with the product of the individual counter settings. For example, if C0 = 40 and C1 = 20, the cascaded value is C0 x C1 = 800. 1 Stratix IV Device Handbook Volume 1 Post-scale counter cascading is set in the configuration file. You cannot set this using PLL reconfiguration. September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-35 Programmable Duty Cycle The programmable duty cycle allows PLLs to generate clock outputs with a variable duty cycle. This feature is supported on the PLL post-scale counters. The duty-cycle setting is achieved by a low and high time-count setting for the post-scale counters. To determine duty cycle choices, the Quartus II software uses the frequency input and the required multiply or divide rate. The post-scale counter value determines the precision of the duty cycle. Precision is defined as 50% divided by the post-scale counter value. For example, if the C0 counter is 10, steps of 5% are possible for duty-cycle choices from 5% to 90%. If the PLL is in external feedback mode, set the duty cycle for the counter driving the fbin pin to 50%. Combining the programmable duty cycle with programmable phase shift allows the generation of precise non-overlapping clocks. Programmable Phase Shift Use phase shift to implement a robust solution for clock delays in Stratix IV devices. Implement phase shift by using a combination of the VCO phase output and the counter starting time. A combination of VCO phase output and counter starting time is the most accurate method of inserting delays because it is only based on counter settings, which are independent of process, voltage, and temperature (PVT). You can phase-shift the output clocks from the Stratix IV PLLs in either of these two resolutions: Fine resolution using VCO phase taps Coarse resolution using counter starting time Implement fine-resolution phase shifts by allowing any of the output counters (C[n..0]) or the m counter to use any of the eight phases of the VCO as the reference clock. This allows you to adjust the delay time with a fine resolution. Equation 5-1 shows the minimum delay time that you can insert using this method. Equation 5-1. Fine-Resolution Phase Shift fine = 1 T = 8 VCO N 1 = 8fVCO 8MfREF where fREF is the input reference clock frequency. For example, if fREF is 100 MHz, N is 1, and M is 8, then fVCO is 800 MHz and fine equals 156.25 ps. This phase shift is defined by the PLL operating frequency, which is governed by the reference clock frequency and the counter settings. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-36 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Equation 5-2 shows the coarse-resolution phase shifts are implemented by delaying the start of the counters for a predetermined number of counter clocks. Equation 5-2. Coarse-Resolution Phase Shift coarse = C - 1 (C - 1)N = fVco MfREF where C is the count value set for the counter delay time (this is the initial setting in the "PLL usage" section of the compilation report in the Quartus II software). If the initial value is 1, C - 1 = 0 phase shift. Figure 5-31 shows an example of phase-shift insertion with fine resolution using the VCO phase-taps method. The eight phases from the VCO are shown and labeled for reference. For this example, CLK0 is based on the 0phase from the VCO and has the C value for the counter set to one. The CLK1 signal is divided by four, two VCO clocks for high time and two VCO clocks for low time. CLK1 is based on the 135 phase tap from the VCO and also has the C value for the counter set to one. In this case, the two clocks are offset by 3 FINE. CLK2 is based on the 0phase from the VCO but has the C value for the counter set to three. This arrangement creates a delay of 2 COARSE (two complete VCO periods). Figure 5-31. Delay Insertion Using VCO Phase Output and Counter Delay Time 1/8 tVCO tVCO 0 45 90 135 180 225 270 315 CLK0 td0-1 CLK1 td0-2 CLK2 You can use coarse- and fine-phase shifts to implement clock delays in Stratix IV devices. Stratix IV devices support dynamic phase-shifting of VCO phase taps only. You can reconfigure the phase shift any number of times. Each phase shift takes about one SCANCLK cycle, allowing you to implement large phase shifts quickly. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-37 Programmable Bandwidth Stratix IV PLLs provide advanced control of the PLL bandwidth using the PLL loop's programmable characteristics, including loop filter and charge pump. Background PLL bandwidth is the measure of the PLL's ability to track the input clock and its associated jitter. The closed-loop gain 3 dB frequency in the PLL determines PLL bandwidth. Bandwidth is approximately the unity gain point for open loop PLL response. As Figure 5-32 shows, these points correspond to approximately the same frequency. Stratix IV PLLs provide three bandwidth settings--low, medium (default), and high. Figure 5-32. Open- and Closed-Loop Response Bode Plots Open-Loop Reponse Bode Plot Increasing the PLL's bandwidth in effect pushes the open loop response out. 0 dB Gain Frequency Closed-Loop Reponse Bode Plot Gain Frequency September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-38 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices A high-bandwidth PLL provides a fast lock time and tracks jitter on the reference clock source, passing it through to the PLL output. A low-bandwidth PLL filters out reference clock jitter but increases lock time. Stratix IV PLLs allow you to control the bandwidth over a finite range to customize the PLL characteristics for a particular application. The programmable bandwidth feature in Stratix IV PLLs benefits applications requiring clock switchover. A high-bandwidth PLL can benefit a system that must accept a spread-spectrum clock signal. Stratix IV PLLs can track a spread-spectrum clock by using a high-bandwidth setting. Using a low-bandwidth setting in this case could cause the PLL to filter out the jitter on the input clock. A low-bandwidth PLL can benefit a system using clock switchover. When clock switchover occurs, the PLL input temporarily stops. A low-bandwidth PLL reacts more slowly to changes on its input clock and takes longer to drift to a lower frequency (caused by input stopping) than a high-bandwidth PLL. Implementation Traditionally, external components such as the VCO or loop filter control a PLL's bandwidth. Most loop filters consist of passive components such as resistors and capacitors that take up unnecessary board space and increase cost. With Stratix IV PLLs, all the components are contained within the device to increase performance and decrease cost. When you specify the bandwidth setting (low, medium, or high) in the ALTPLL MegaWizardPlug-in Manager, the Quartus II software automatically sets the corresponding charge pump and loop filter (Icp, R, C) values to achieve the desired bandwidth range. Figure 5-33 shows the loop filter and components that you can set using the Quartus II software. The components are the loop filter resistor, R, the high frequency capacitor, Ch, and the charge pump current, IUP or IDN. Figure 5-33. Loop Filter Programmable Components IUP PFD R Ch IDN C Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-39 Spread-Spectrum Tracking Stratix IV devices can accept a spread-spectrum input with typical modulation frequencies. However, the device cannot automatically detect that the input is a spread-spectrum signal. Instead, the input signal looks like deterministic jitter at the input of the PLL. Stratix IV PLLs can track a spread-spectrum input clock as long as it is within input-jitter tolerance specifications. Stratix IV devices cannot internally generate spread-spectrum clocks. Clock Switchover The clock switchover feature allows the PLL to switch between two reference input clocks. Use this feature for clock redundancy or for a dual-clock domain application such as in a system that turns on the redundant clock if the previous clock stops running. The design can perform clock switchover automatically when the clock is no longer toggling or based on a user control signal, clkswitch. The following clock switchover modes are supported in Stratix IV PLLs: September 2012 Automatic switchover--The clock sense circuit monitors the current reference clock and if it stops toggling, automatically switches to the other inclk0 or inclk1 clock. Manual clock switchover--Clock switchover is controlled using the clkswitch signal. When the clkswitch signal goes from logic low to logic high, and stays high for at least three clock cycles, the reference clock to the PLL is switched from inclk0 to inclk1, or vice-versa. Automatic switchover with manual override--This mode combines automatic switchover and manual clock switchover. When the clkswitch signal goes high, it overrides the automatic clock switchover function. As long as the clkswitch signal is high, further switchover action is blocked. Altera Corporation Stratix IV Device Handbook Volume 1 5-40 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Stratix IV PLLs support a fully configurable clock switchover capability. Figure 5-34 shows a block diagram of the automatic switchover circuit built into the PLL. When the current reference clock is not present, the clock sense block automatically switches to the backup clock for PLL reference. The clock switchover circuit also sends out three status signals--clkbad[0], clkbad[1], and activeclock--from the PLL to implement a custom switchover circuit in the logic array. You can select a clock source as the backup clock by connecting it to the inclk1 port of the PLL in your design. Figure 5-34. Automatic Clock Switchover Circuit Block Diagram clkbad[0] clkbad[1] activeclock Switchover State Machine Clock Sense clksw Clock Switch Control Logic clkswitch inclk0 n Counter inclk1 muxout PFD refclk fbclk Automatic Clock Switchover Use the switchover circuitry to automatically switch between inclk0 and inclk1 when the current reference clock to the PLL stops toggling. For example, in applications that require a redundant clock with the same frequency as the reference clock, the switchover state machine generates a signal (clksw) that controls the multiplexer select input, as shown in Figure 5-34. In this case, inclk1 becomes the reference clock for the PLL. When using automatic switchover mode, you can switch back and forth between inclk0 and inclk1 any number of times when one of the two clocks fails and the other clock is available. When using automatic clock switchover mode, the following requirements must be satisfied: Both clock inputs must be running The period of the two clock inputs can differ by no more than 100% (2x) If the current clock input stops toggling while the other clock is also not toggling, switchover is not initiated and the clkbad[0..1] signals are not valid. Also, if both clock inputs are not the same frequency, but their period difference is within 100%, the clock sense block detects when a clock stops toggling, but the PLL may lose lock after the switchover is completed and needs time to re-lock. 1 Stratix IV Device Handbook Volume 1 Altera recommends resetting the PLL using the areset signal to maintain the phase relationships between the PLL input and output clocks when using clock switchover. September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-41 In automatic switchover mode, the clkbad[0] and clkbad[1] signals indicate the status of the two clock inputs. When they are asserted, the clock sense block has detected that the corresponding clock input has stopped toggling. These two signals are not valid if the frequency difference between inclk0 and inclk1 is greater than 20%. The activeclock signal indicates which of the two clock inputs (inclk0 or inclk1) is being selected as the reference clock to the PLL. When the frequency difference between the two clock inputs is more than 20%, the activeclock signal is the only valid status signal. Figure 5-35 shows an example waveform of the switchover feature when using automatic switchover mode. In this example, the inclk0 signal is stuck low. After the inclk0 signal is stuck at low for approximately two clock cycles, the clock sense circuitry drives the clkbad[0] signal high. Also, because the reference clock signal is not toggling, the switchover state machine controls the multiplexer through the clkswitch signal to switch to the backup clock, inclk1. Figure 5-35. Automatic Switchover After Loss of Clock Detection inclk0 inclk1 (1) muxout clkbad0 clkbad1 activeclock Note to Figure 5-35: (1) Switchover is enabled on the falling edge of inclk0 or inclk1, depending on which clock is available. In this figure, switchover is enabled on the falling edge of inclk1. Manual Override In automatic switchover with manual override mode, you can use the clkswitch input for user- or system-controlled switch conditions. You can use this mode for same-frequency switchover, or to switch between inputs of different frequencies. For example, if inclk0 is 66 MHz and inclk1 is 200 MHz, you must control switchover using clkswitch because the automatic clock-sense circuitry cannot monitor clock input (inclk0 and inclk1) frequencies with a frequency difference of more than 100% (2x). This feature is useful when the clock sources originate from multiple cards on the backplane, requiring a system-controlled switchover between the frequencies of operation. You must choose the backup clock frequency and set the m, n, c, and k counters accordingly so the VCO operates within the recommended operating frequency range of 600 to 1,600 MHz. The ALTPLL MegaWizard Plug-in Manager notifies you if a given combination of inclk0 and inclk1 frequencies cannot meet this requirement. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-42 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Figure 5-36 shows a clock switchover waveform controlled by clkswitch. In this case, both clock sources are functional and inclk0 is selected as the reference clock; clkswitch goes high, which starts the switchover sequence. On the falling edge of inclk0, the counter's reference clock, muxout, is gated off to prevent clock glitching. On the falling edge of inclk1, the reference clock multiplexer switches from inclk0 to inclk1 as the PLL reference and the activeclock signal changes to indicate which clock is currently feeding the PLL. Figure 5-36. Clock Switchover Using the clkswitch (Manual) Control (1) inclk0 inclk1 muxout clkswitch activeclock clkbad0 clkbad1 Note to Figure 5-36: (1) To initiate a manual clock switchover event, both inclk0 and inclk1 must be running when the clkswitch signal goes high. In automatic override with manual switchover mode, the activeclock signal mirrors the clkswitch signal. As both clocks are still functional during the manual switch, neither clkbad signal goes high. Because the switchover circuit is positive-edge sensitive, the falling edge of the clkswitch signal does not cause the circuit to switch back from inclk1 to inclk0. When the clkswitch signal goes high again, the process repeats. clkswitch and automatic switch only work if the clock being switched to is available. If the clock is not available, the state machine waits until the clock is available. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-43 Manual Clock Switchover In manual clock switchover mode, the clkswitch signal controls whether inclk0 or inclk1 is selected as the input clock to the PLL. By default, inclk0 is selected. A low-to-high transition on clkswitch and clkswitch being held high for at least three inclk cycles initiates a clock switchover event. You must bring clkswitch back low again in order to perform another switchover event in the future. If you do not require another switchover event in the future, you can leave clkswitch in a logic high state after the initial switch. Pulsing clkswitch high for at least three inclk cycles performs another switchover event. If inclk0 and inclk1 are different frequencies and are always running, the clkswitch minimum high time must be greater than or equal to three of the slower frequency inclk0 or inclk1 cycles. Figure 5-37 shows a block diagram of the manual switchover circuit. Figure 5-37. Manual Clock Switchover Circuitry in Stratix IV PLLs clkswitch Clock Switch Control Logic inclk0 n Counter PFD inclk1 muxout refclk fbclk f For more information about PLL software support in the Quartus II software, refer to the Phase-Locked Loop (ALTPLL) Megafunction User Guide. Guidelines When implementing clock switchover in Stratix IV PLLs, use the following guidelines: Automatic clock switchover requires that the inclk0 and inclk1 frequencies be within 100% (2x) of each other. Failing to meet this requirement causes the clkbad[0] and clkbad[1] signals to not function properly. When using manual clock switchover, the difference between inclk0 and inclk1 can be more than 100% (2x). However, differences in frequency, phase, or both, of the two clock sources will likely cause the PLL to lose lock. Resetting the PLL ensures that the correct phase relationships are maintained between the input and output clocks. 1 September 2012 Both inclk0 and inclk1 must be running when the clkswitch signal goes high to initiate the manual clock switchover event. Failing to meet this requirement causes the clock switchover to not function properly. Applications that require a clock switchover feature and a small frequency drift must use a low-bandwidth PLL. The low-bandwidth PLL reacts more slowly than a high-bandwidth PLL to reference input clock changes. When switchover happens, a low-bandwidth PLL propagates the stopping of the clock to the output more slowly than a high-bandwidth PLL. However, be aware that the low-bandwidth PLL also increases lock time. Altera Corporation Stratix IV Device Handbook Volume 1 5-44 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices After a switchover occurs, there may be a finite resynchronization period for the PLL to lock onto a new clock. The exact amount of time it takes for the PLL to re-lock depends on the PLL configuration. The phase relationship between the input clock to the PLL and the output clock from the PLL is important in your design. Assert areset for at least 10 ns after performing a clock switchover. Wait for the locked signal to go high and be stable before re-enabling the output clocks from the PLL. Figure 5-38 shows how the VCO frequency gradually decreases when the current clock is lost and then increases as the VCO locks on to the backup clock. Figure 5-38. VCO Switchover Operating Frequency Primary Clock Stops Running Switchover Occurs VCO Tracks Secondary Clock Fvco Disable the system during clock switchover if it is not tolerant of frequency variations during the PLL resynchronization period. You can use the clkbad[0] and clkbad[1] status signals to turn off the PFD (PFDENA = 0) so the VCO maintains its most recent frequency. You can also use the state machine to switch over to the secondary clock. When the PFD is re-enabled, output clock-enable signals (clkena) can disable clock outputs during the switchover and resynchronization period. When the lock indication is stable, the system can re-enable the output clocks. PLL Reconfiguration PLLs use several divide counters and different VCO phase taps to perform frequency synthesis and phase shifts. In Stratix IV PLLs, you can reconfigure both the counter settings and phase-shift the PLL output clock in real time. You can also change the charge pump and loop-filter components, which dynamically affects PLL bandwidth. You can use these PLL components to update the output-clock frequency and PLL bandwidth and to phase-shift in real time, without reconfiguring the entire Stratix IV device. The ability to reconfigure the PLL in real time is useful in applications that operate at multiple frequencies. It is also useful in prototyping environments, allowing you to sweep PLL output frequencies and adjust the output-clock phase dynamically. For instance, a system generating test patterns is required to generate and transmit patterns at 75 or 150 MHz, depending on the requirements of the device under test. Reconfiguring the PLL components in real time allows you to switch between two such output frequencies within a few microseconds. You can also use this feature to adjust clock-to-out (tCO) delays in real time by changing the PLL output clock phase shift. This approach eliminates the need to regenerate a configuration file with the new PLL settings. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-45 PLL Reconfiguration Hardware Implementation The following PLL components are reconfigurable in real time: Pre-scale counter (n) Feedback counter (m) Post-scale output counters (C0 - C9) Post VCO Divider (K) Dynamically adjust the charge-pump current (Icp) and loop-filter components (R, C) to facilitate reconfiguration of the PLL bandwidth Figure 5-39 shows how you can dynamically adjust the PLL counter settings by shifting their new settings into a serial shift-register chain or scan chain. Serial data is input to the scan chain using the scandata port. Shift registers are clocked by scanclk. The maximum scanclk frequency is 100 MHz. Serial data is shifted through the scan chain as long as the scanclkena signal stays asserted. After the last bit of data is clocked, asserting the configupdate signal for at least one scanclk clock cycle causes the PLL configuration bits to be synchronously updated with the data in the scan registers. Figure 5-39. PLL Reconfiguration Scan Chain (1) from m counter from n counter LF/K/CP (3) PFD VCO scandata scanclkena configupdate /Ci (2) inclk scandataout /Ci-1 /C2 /C1 /C0 /m /n scandone scanclk Notes to Figure 5-39: (1) Stratix IV left and right PLLs support C0 - C6 counters. (2) i = 6 or i = 9. (3) This figure shows the corresponding scan register for the K counter in between the scan registers for the charge pump and loop filter. The K counter is physically located after the VCO. 1 September 2012 The counter settings are updated synchronously to the clock frequency of the individual counters. Therefore, all counters are not updated simultaneously. Altera Corporation Stratix IV Device Handbook Volume 1 5-46 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Table 5-10 lists how these signals can be driven by the PLD logic array or I/O pins. Table 5-10. Real-Time PLL Reconfiguration Ports PLL Port Name Description Source Destination scandata Serial input data stream to scan chain. Logic array or I/O pin PLL reconfiguration circuit scanclk Serial clock input signal. This clock can be free running. GCLK, RCLK or I/O pins PLL reconfiguration circuit scanclkena Enables scanclk and allows the scandata to be loaded in the scan chain. Active high. Logic array or I/O pin PLL reconfiguration circuit configupdate Writes the data in the scan chain to the PLL. Active high. Logic array or I/O pin PLL reconfiguration circuit scandone Indicates when the PLL has finished reprogramming. A rising edge indicates the PLL has begun reprogramming. A falling edge indicates the PLL has finished reprogramming. PLL reconfiguration circuit Logic array or I/O pins scandataout Used to output the contents of the scan chain. PLL reconfiguration circuit Logic array or I/O pins To reconfigure the PLL counters, follow these steps: 1. The scanclkena signal is asserted at least one scanclk cycle prior to shifting in the first bit of scandata (D0). 2. Serial data (scandata) is shifted into the scan chain on the second rising edge of scanclk. 3. After all 234 bits (top and bottom PLLs) or 180 bits (left and right PLLs) have been scanned into the scan chain, the scanclkena signal is de-asserted to prevent inadvertent shifting of bits in the scan chain. 4. The configupdate signal is asserted for one scanclk cycle to update the PLL counters with the contents of the scan chain. 5. The scandone signal goes high, indicating the PLL is being reconfigured. A falling edge indicates the PLL counters have been updated with new settings. 6. Reset the PLL using the areset signal if you make any changes to the M, N, or post-scale output C counters or to the Icp, R, or C settings. 7. You can repeat steps 1-5 to reconfigure the PLL any number of times. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-47 Figure 5-40 shows a functional simulation of the PLL reconfiguration feature. Figure 5-40. PLL Reconfiguration Waveform (LSB) D0 SCANDATA (MSB) Dn SCANCLK SCANCLKENA D0_old SCANDATAOUT Dn_old Dn CONFIGUPDATE SCANDONE ARESET 1 When you reconfigure the counter clock frequency, you cannot reconfigure the corresponding counter phase shift settings using the same interface. Instead, reconfigure the phase shifts in real time using the dynamic phase shift reconfiguration interface. If you reconfigure the counter frequency, but wish to keep the same non-zero phase shift setting (for example, 90) on the clock output, you must reconfigure the phase shift immediately after reconfiguring the counter clock frequency. Post-Scale Counters (C0 to C9) You can reconfigure the multiply or divide values and duty cycle of post-scale counters in real time. Each counter has an 8-bit high-time setting and an 8-bit low-time setting. The duty cycle is the ratio of output high- or low-time to the total cycle time, which is the sum of the two. Additionally, these counters have two control bits, rbypass, for bypassing the counter, and rselodd, to select the output clock duty cycle. When the rbypass bit is set to 1, it bypasses the counter, resulting in a divide by 1. When the rbypass bit is set to 0, the high- and low-time counters are added to compute the effective division of the VCO output frequency. For example, if the post-scale divide factor is 10, the high- and low-count values can be set to 5 and 5, respectively, to achieve a 50% - 50% duty cycle. The PLL implements this duty cycle by transitioning the output clock from high to low on the rising edge of the VCO output clock. However, a 4 and 6 setting for the high- and low-count values, respectively, produces an output clock with a 40% - 60% duty cycle. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-48 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices The rselodd bit indicates an odd divide factor for the VCO output frequency along with a 50% duty cycle. For example, if the post-scale divide factor is 3, the high- and low-time count values could be set to 2 and 1, respectively, to achieve this division. This implies a 67% - 33% duty cycle. If you need a 50% - 50% duty cycle, you can set the rselodd control bit to 1 to achieve this duty cycle despite an odd division factor. The PLL implements this duty cycle by transitioning the output clock from high to low on a falling edge of the VCO output clock. When you set rselodd = 1, you subtract 0.5 cycles from the high time and you add 0.5 cycles to the low time. For example: High-time count = 2 cycles Low-time count = 1 cycle rselodd = 1 effectively equals: High-time count = 1.5 cycles Low-time count = 1.5 cycles Duty cycle = (1.5/3) % high-time count and (1.5/3) % low-time count Scan Chain Description The length of the scan chain varies for different Stratix IV PLLs. The top and bottom PLLs have ten post-scale counters and a 234-bit scan chain, while the left and right PLLs have seven post-scale counters and a 180-bit scan chain. Table 5-11 lists the number of bits for each component of a Stratix IV PLL. Table 5-11. Top and Bottom PLL Reprogramming Bits (Part 1 of 2) Number of Bits Block Name Total Counter C9 (1) 16 2 18 C8 16 2 18 C7 16 2 18 16 2 18 C5 16 2 18 C4 16 2 18 C3 16 2 18 C2 16 2 18 C1 16 2 18 C0 16 2 18 M 16 2 18 N 16 2 18 Charge Pump Current 0 3 3 VCO Post-Scale divider (K) 1 0 1 C6 Stratix IV Device Handbook Volume 1 (2) Other (3) September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-49 Table 5-11. Top and Bottom PLL Reprogramming Bits (Part 2 of 2) Number of Bits Block Name Total Counter Loop Filter Capacitor (4) Other (1) 0 2 2 Loop Filter Resistor 0 5 5 Unused CP/LF 0 7 7 Total number of bits -- -- 234 Notes to Table 5-11: (1) Includes two control bits, rbypass, for bypassing the counter, and rselodd, to select the output clock duty cycle. (2) The LSB for the C9 low-count value is the first bit shifted into the scan chain for the top and bottom PLLs. (3) The LSB for the C6 low-count value is the first bit shifted into the scan chain for the left and right PLLs. (4) The MSB for the loop filter is the last bit shifted into the scan chain. Table 5-11 lists the scan chain order of PLL components for the top and bottom PLLs, which have 10 post-scale counters. The order of bits is the same for the left and right PLLs, but the reconfiguration bits start with the C6 post-scale counter. Figure 5-41 shows the scan-chain order of PLL components for the top and bottom PLLs. (1) Figure 5-41. Scan-Chain Order of PLL Components for Top and Bottom PLLs DATAIN LF K CP LSB MSB C6 C4 C5 C7 C8 N M C0 C3 C2 C1 DATAOUT C9 Note to Figure 5-41: (1) Left and right PLLs have the same scan-chain order. The post-scale counters end at C6. Figure 5-42 shows the scan-chain bit-order sequence for post-scale counters in all Stratix IV PLLs. Figure 5-42. Scan-Chain Bit-Order Sequence for Post-Scale Counters in Stratix IV PLLs DATAOUT September 2012 HB HB HB HB HB HB HB HB 0 1 2 3 4 5 6 7 LB LB LB LB LB LB LB LB 0 1 2 3 4 5 6 7 Altera Corporation rbypass DATAIN rselodd Stratix IV Device Handbook Volume 1 5-50 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Charge Pump and Loop Filter You can reconfigure the charge-pump and loop-filter settings to update the PLL bandwidth in real time. Table 5-12 lists the possible settings for charge pump current (Icp) values for Stratix IV PLLs. Table 5-12. Charge Pump Current Bit Settings CP[2] CP[1] CP[0] Decimal Value for Setting 0 0 0 0 0 0 1 1 0 1 1 3 1 1 1 7 Table 5-13 lists the possible settings for loop-filter resistor (R) values for Stratix IV PLLs. Table 5-13. Loop-Filter Resistor Bit Settings LFR[4] LFR[3] LFR[2] LFR[1] LFR[0] Decimal Value for Setting 0 0 0 0 0 0 0 0 0 1 1 3 0 0 1 0 0 4 0 1 0 0 0 8 1 0 0 0 0 16 1 0 0 1 1 19 1 0 1 0 0 20 1 1 0 0 0 24 1 1 0 1 1 27 1 1 1 0 0 28 1 1 1 1 0 30 Table 5-14 lists the possible settings for loop-filter capacitor (C) values for Stratix IV PLLs. Table 5-14. Loop-Filter Capacitor Bit Settings Stratix IV Device Handbook Volume 1 LFC[1] LFC[0] Decimal Value for Setting 0 0 0 0 1 1 1 1 3 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-51 Bypassing a PLL Bypassing a PLL counter results in a multiply (m counter) or a divide (n and C0 to C9 counters) factor of one. Table 5-15 lists the settings for bypassing the counters in Stratix IV PLLs. Table 5-15. PLL Counter Settings PLL Scan Chain Bits [0..8] Settings LSB MSB X X X X X X X X 1 (1) X X X X X X X X 0 (1) Description PLL counter bypassed PLL counter not bypassed because bit 8 (MSB) is set to 0 Note to Table 5-15: (1) Counter-bypass bit. 1 To bypass any of the PLL counters, set the bypass bit to 1. The values on the other bits are ignored. To bypass the VCO post-scale counter (K), set the corresponding bit to 0. Dynamic Phase-Shifting The dynamic phase-shifting feature allows the output phases of individual PLL outputs to be dynamically adjusted relative to each other and to the reference clock, without having to send serial data through the scan chain of the corresponding PLL. This feature simplifies the interface and allows you to quickly adjust the clock-to-out (tCO) delays by changing the output clock phase-shift in real time. This adjustment is achieved by incrementing or decrementing the VCO phase-tap selection to a given C counter or to the M counter. The phase is shifted by 1/8 of the VCO frequency at a time. The output clocks are active during this phase-reconfiguration process. Table 5-16 lists the control signals that are used for dynamic phase-shifting. Table 5-16. Dynamic Phase-Shifting Control Signals (Part 1 of 2) Signal Name Description Source PHASECOUNTERSELECT [3..0] Counter select. Four bits decoded to select either the M or one of the C counters for phase adjustment. One address maps to select all C counters. This signal is registered in the PLL on the rising edge of SCANCLK. Logic array or I/O pins PLL reconfiguration circuit PHASEUPDOWN Selects dynamic phase shift direction; 1 = UP; 0 = DOWN. Signal is registered in the PLL on the rising edge of SCANCLK. Logic array or I/O pin PLL reconfiguration circuit PHASESTEP Logic high enables dynamic phase shifting. Logic array or I/O pin PLL reconfiguration circuit September 2012 Altera Corporation Destination Stratix IV Device Handbook Volume 1 5-52 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices Table 5-16. Dynamic Phase-Shifting Control Signals (Part 2 of 2) Signal Name Description Source Destination SCANCLK Free running clock from the core used in combination with PHASESTEP to enable and disable dynamic phase shifting. Shared with SCANCLK for dynamic reconfiguration. GCLK, RCLK or I/O pin PLL reconfiguration circuit PHASEDONE When asserted, this indicates to core-logic that the phase adjustment is complete and the PLL is ready to act on a possible second adjustment pulse. Asserts based on internal PLL timing. De-asserts on the rising edge of SCANCLK. PLL reconfiguration circuit Logic array or I/O pins Table 5-17 lists the PLL counter selection based on the corresponding PHASECOUNTERSELECT setting. Table 5-17. Phase Counter Select Mapping PHASECOUNTERSELECT[3] [2] [1] [0] Selects 0 0 0 0 All Output Counters 0 0 0 1 M Counter 0 0 1 0 C0 Counter 0 0 1 1 C1 Counter 0 1 0 0 C2 Counter 0 1 0 1 C3 Counter 0 1 1 0 C4 Counter 0 1 1 1 C5 Counter 1 0 0 0 C6 Counter 1 0 0 1 C7 Counter 1 0 1 0 C8 Counter 1 0 1 1 C9 Counter To perform one dynamic phase-shift, follow these steps: 1. Set PHASEUPDOWN and PHASECOUNTERSELECT as required. 2. Assert PHASESTEP for at least two SCANCLK cycles. Each PHASESTEP pulse enables one phase shift. 3. Deassert PHASESTEP after PHASEDONE goes low. 4. Wait for PHASEDONE to go high. 5. Repeat steps 1-4 as many times as required to perform multiple phase-shifts. The PHASEUPDOWN and PHASECOUNTERSELECT signals are synchronous to SCANCLK and must meet tsu/th requirements with respect to SCANCLK edges. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 1 5-53 You can repeat dynamic phase-shifting indefinitely. For example, in a design where the VCO frequency is set to 1000 MHz and the output clock frequency is 100 MHz, performing 40 dynamic phase shifts (each one yields 125 ps phase shift) results in shifting the output clock by 180, which is a phase shift of 5 ns. The PHASESTEP signal is latched on the negative edge of SCANCLK (a,c) and must remain asserted for at least two SCANCLK cycles. De-assert PHASESTEP after PHASEDONE goes low. On the second SCANCLK rising edge (b,d) after PHASESTEP is latched, the values of PHASEUPDOWN and PHASECOUNTERSELECT are latched and the PLL starts dynamic phase-shifting for the specified counters and in the indicated direction. PHASEDONE is de-asserted synchronous to SCANCLK at the second rising edge (b,d) and remains low until the PLL finishes dynamic phase-shifting. Depending on the VCO and SCANCLK frequencies, PHASEDONE low time may be greater than or less than one SCANCLK cycle. You can perform another dynamic phase-shift after the PHASEDONE signal goes from low to high. Each PHASESTEP pulse enables one phase shift. PHASESTEP pulses must be at least one SCANCLK cycle apart. Figure 5-43. Dynamic Phase Shifting Waveform SCANCLK PHASESTEP PHASEUPDOWN PHASECOUNTERSELECT PHASEDONE a b c d PHASEDONE goes low synchronous with SCANCLK t CONFIGPHASE Depending on the VCO and SCANCLK frequencies, PHASEDONE low time may be greater than or less than one SCANCLK cycle. After PHASEDONE goes from low to high, you can perform another dynamic phase shift. PHASESTEP pulses must be at least one SCANCLK cycle apart. f For information about the ALTPLL_RECONFIG MegaWizard Plug-In Manager, refer to the Phase-Locked Loops Reconfiguration (ALTPLL_RECONFIG) Megafunction User Guide. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 5-54 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices PLL Specifications f For information about PLL timing specifications, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Document Revision History Table 5-18 lists the revision history for this chapter. Table 5-18. Document Revision History (Part 1 of 2) Date Version September 2012 3.4 December 2011 3.3 February 2011 March 2010 November 2009 June 2009 Stratix IV Device Handbook Volume 1 3.2 3.1 3.0 2.3 Changes Updated the "Periphery Clock Networks" section. Updated the "Dynamic Phase-Shifting" section. Updated Figure 5-43. Updated the "Clock Input Connections to the PLLs," "PLL Clock I/O Pins," "Clock Feedback Modes," and "Clock Switchover" sections. Updated Table 5-4 and Table 5-8. Updated Figure 5-26, Figure 5-40, and Figure 5-43. Applied new template. Minor text edits. Updated Table 5-3. Updated notes to Figure 5-2, Figure 5-3, Figure 5-4, and Figure 5-9. Added a note to Table 5-5 and Table 5-6. Added two notes to Table 5-4. Updated Figure 5-43. Updated the "Dynamic Phase-Shifting" section. Minor text edits. Updated Table 5-1 and Table 5-7. Updated "Clock Networks in Stratix IV Devices", "Periphery Clock Networks", and "Cascading PLLs" sections. Added Figure 5-5, Figure 5-6, Figure 5-7, Figure 5-8, and Figure 5-9. Added "Clock Sources Per Region" section. Updated Figure 5-40. Removed EP4SE110, EP4SE290, and EP4SE680 devices. Added EP4S40G2, EP4S100G2, EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5, and EP4SE820 devices. Updated Table 5-7. Updated the "PLL Reconfiguration Hardware Implementation" and "Zero-Delay Buffer Mode" sections. Added introductory sentences to improve search ability. Removed the Conclusion section. Minor text edits. September 2012 Altera Corporation Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices 5-55 Table 5-18. Document Revision History (Part 2 of 2) Date Version April 2009 2.2 March 2009 2.1 November 2008 May 2008 September 2012 2.0 1.0 Altera Corporation Changes Updated Table 5-1 and Table 5-7. Updated Figure 5-3 and Figure 5-4. Updated the "Periphery Clock Networks" section. Updated Table 5-7. Updated Figure 5-34. Updated "Guidelines" section. Removed "Referenced Documents" section. Updated Table 5-7. Updated Note 1 of Figure 5-10. Updated Figure 5-15. Updated Figure 5-20. Added Figure 5-21. Made minor editorial changes. Initial release. Stratix IV Device Handbook Volume 1 5-56 Stratix IV Device Handbook Volume 1 Chapter 5: Clock Networks and PLLs in Stratix IV Devices PLLs in Stratix IV Devices September 2012 Altera Corporation Section II. I/O Interfaces This section provides information on Stratix(R) IV device I/O features, external memory interfaces, and high-speed differential interfaces with DPA. This section includes the following chapters: Chapter 6, I/O Features in Stratix IV Devices Chapter 7, External Memory Interfaces in Stratix IV Devices Chapter 8, High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 II-2 Stratix IV Device Handbook Volume 1 Section II: I/O Interfaces September 2012 Altera Corporation 6. I/O Features in Stratix IV Devices September 2012 SIV51006-3.4 SIV51006-3.4 This chapter describes how Stratix IV devices provide I/O capabilities that allow you to work in compliance with current and emerging I/O standards and requirements. With these device features, you can reduce board design interface costs and increase development flexibility. Altera Stratix IV FPGAs deliver a breakthrough level of system bandwidth and power efficiency for high-end applications, allowing you to innovate without compromise. Stratix IV I/Os are specifically designed for ease-of-use and rapid system integration while simultaneously providing the high bandwidth required to maximize internal logic capabilities and produce system-level performance. Stratix IV device I/O capability far exceeds the I/O bandwidth available from previous generation FPGAs. Independent modular I/O banks with a common bank structure for vertical migration lend efficiency and flexibility to the high-speed I/O. Package and die enhancements with dynamic termination and output control provide best-in-class signal integrity. Numerous I/O features assist high-speed data transfer into and out of the device, including: Up to 32 full-duplex clock data recovery (CDR)-based transceivers supporting data rates between 600 Mbps and 8.5 Gbps Dedicated circuitry to support physical layer functionality for popular serial protocols, such as PCI Express(R) (PIPE) (PCIe) Gen1 and Gen2, Gigabit Ethernet (GbE), Serial RapidIO(R), SONET/SDH, XAUI/HiGig, (OIF) CEI-6G, SD/HD/3G-SDI, Fibre Channel, SFI-5, and Interlaken Complete PCIe protocol solution with embedded PCIe hard IP blocks that implement PHY-MAC layer, data link layer, and transaction layer functionality Single-ended, non-voltage-referenced, and voltage-referenced I/O standards Low-voltage differential signaling (LVDS), reduced swing differential signaling (RSDS), mini-LVDS, high-speed transceiver logic (HSTL), and SSTL Single data rate (SDR) and half data rate (HDR--half frequency and twice data width of SDR) input and output options Up to 132 full duplex 1.6 Gbps true LVDS channels (132 Tx + 132 Rx) on the row I/O banks Hard dynamic phase alignment (DPA) block with serializer/deserializer (SERDES) Deskew, read and write leveling, and clock-domain crossing functionality Programmable output current strength Programmable slew rate (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 September 2012 Feedback Subscribe 6-2 Chapter 6: I/O Features in Stratix IV Devices I/O Standards Support Programmable delay Programmable bus-hold circuit Programmable pull-up resistor Open-drain output Serial, parallel, and dynamic on-chip termination (OCT) Differential OCT Programmable pre-emphasis Programmable equalization Programmable differential output voltage (VOD) This chapter contains the following sections: "I/O Standards Support" "I/O Banks" on page 6-5 "I/O Structure" on page 6-17 "On-Chip Termination Support and I/O Termination Schemes" on page 6-24 "OCT Calibration" on page 6-32 "Termination Schemes for I/O Standards" on page 6-38 "Design Considerations" on page 6-46 I/O Standards Support Stratix IV devices support a wide range of industry I/O standards. Table 6-1 lists the I/O standards Stratix IV devices support, as well as the typical applications. These devices support VCCIO voltage levels of 3.0, 2.5, 1.8, 1.5, and 1.2 V. Table 6-1. I/O Standards and Applications for Stratix IV Devices (Part 1 of 2) I/O Standard 3.3-V LVTTL/LVCMOS Stratix IV Device Handbook Volume 1 (1), (2) Application General purpose 2.5-V LVCMOS General purpose 1.8-V LVCMOS General purpose 1.5-V LVCMOS General purpose 1.2-V LVCMOS General purpose 3.0-V PCI/PCI-X PC and embedded system SSTL-2 Class I and II DDR SDRAM SSTL-18 Class I and II DDR2 SDRAM SSTL-15 Class I and II DDR3 SDRAM HSTL-18 Class I and II QDRII/RLDRAM II HSTL-15 Class I and II QDRII/QDRII+/RLDRAM II HSTL-12 Class I and II General purpose Differential SSTL-2 Class I and II DDR SDRAM Differential SSTL-18 Class I and II DDR2 SDRAM September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Standards Support 6-3 Table 6-1. I/O Standards and Applications for Stratix IV Devices (Part 2 of 2) I/O Standard Application Differential SSTL-15 Class I and II DDR3 SDRAM Differential HSTL-18 Class I and II Clock interfaces Differential HSTL-15 Class I and II Clock interfaces Differential HSTL-12 Class I and II Clock interfaces LVDS High-speed communications RSDS Flat panel display mini-LVDS Flat panel display LVPECL Video graphics and clock distribution Notes to Table 6-1: (1) The 3.3-V LVTTL/LVCMOS standard is supported using VCCIO at 3.0 V. (2) For more information about the 3.3-V LVTTL/LVCMOS standard supported in Stratix IV devices, refer to "3.3-V I/O Interface" on page 6-19. f For more information about transceiver supported I/O standards, refer to the Transceiver Architecture in Stratix IV Devices chapter. I/O Standards and Voltage Levels Stratix IV devices support a wide range of industry I/O standards, including single-ended, voltage-referenced single-ended, and differential I/O standards. Table 6-2 lists the supported I/O standards and typical values for input and output VCCIO, VCCPD, VREF, and board VTT. Table 6-2. I/O Standards and Voltage Levels for Stratix IV Devices (1) (Part 1 of 3) VCCIO (V) I/O Standard Standard Support Input Operation Column Row I/O Banks I/O Banks 3.3-V LVTTL (3) 3.3-V LVCMOS 2.5-V LVCMOS Output Operation Column I/O Banks Row I/O Banks VTT (V) VCCPD (V) VREF (V) (Board (Pre-Driver (Input Ref Termination Voltage) Voltage) Voltage) JESD8-B 3.0/2.5 3.0/2.5 3.0 3.0 3.0 -- -- JESD8-B 3.0/2.5 3.0/2.5 3.0 3.0 3.0 -- -- JESD8-5 3.0/2.5 3.0/2.5 2.5 2.5 2.5 -- -- 1.8-V LVCMOS JESD8-7 1.8/1.5 1.8/1.5 1.8 1.8 2.5 -- -- 1.5-V LVCMOS JESD8-11 1.8/1.5 1.8/1.5 1.5 1.5 2.5 -- -- 1.2-V LVCMOS JESD8-12 1.2 1.2 1.2 1.2 2.5 -- -- 3.0-V PCI PCI Rev 2.1 3.0 3.0 3.0 3.0 3.0 -- -- 3.0-V PCI-X PCI-X Rev 1.0 3.0 3.0 3.0 3.0 3.0 -- -- JESD8-9B (2) (2) 2.5 2.5 2.5 1.25 1.25 SSTL-2 Class II JESD8-9B (2) (2) 2.5 2.5 2.5 1.25 1.25 SSTL-18 Class I JESD8-15 (2) (2) 1.8 1.8 2.5 0.90 0.90 SSTL-18 Class II JESD8-15 (2) (2) 1.8 1.8 2.5 0.90 0.90 SSTL-2 Class I September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-4 Chapter 6: I/O Features in Stratix IV Devices I/O Standards Support Table 6-2. I/O Standards and Voltage Levels for Stratix IV Devices (1) (Part 2 of 3) VCCIO (V) I/O Standard Standard Support Input Operation Column Row I/O Banks I/O Banks Output Operation Column I/O Banks Row I/O Banks VTT (V) VCCPD (V) VREF (V) (Board (Pre-Driver (Input Ref Termination Voltage) Voltage) Voltage) -- (2) (2) 1.5 1.5 2.5 0.75 0.75 -- (2) (2) 1.5 -- 2.5 0.75 0.75 JESD8-6 (2) (2) 1.8 1.8 2.5 0.90 0.90 JESD8-6 (2) (2) 1.8 1.8 2.5 0.90 0.90 JESD8-6 (2) (2) 1.5 1.5 2.5 0.75 0.75 HSTL-15 Class II JESD8-6 (2) (2) 1.5 -- 2.5 0.75 0.75 HSTL-12 Class I JESD8-16A (2) (2) 1.2 1.2 2.5 0.6 0.6 HSTL-12 Class II JESD8-16A (2) (2) 1.2 -- 2.5 0.6 0.6 Differential SSTL-2 Class I JESD8-9B (2) (2) 2.5 2.5 2.5 -- 1.25 Differential SSTL-2 Class II JESD8-9B (2) (2) 2.5 2.5 2.5 -- 1.25 Differential SSTL-18 Class I JESD8-15 (2) (2) 1.8 1.8 2.5 -- 0.90 Differential SSTL-18 Class II JESD8-15 (2) (2) 1.8 1.8 2.5 -- 0.90 Differential SSTL-15 Class I -- (2) (2) 1.5 1.5 2.5 -- 0.75 Differential SSTL-15 Class II -- (2) (2) 1.5 -- 2.5 -- 0.75 Differential HSTL-18 Class I JESD8-6 (2) (2) 1.8 1.8 2.5 -- 0.90 Differential HSTL-18 Class II JESD8-6 (2) (2) 1.8 1.8 2.5 -- 0.90 Differential HSTL-15 Class I JESD8-6 (2) (2) 1.5 1.5 2.5 -- 0.75 Differential HSTL-15 Class II JESD8-6 (2) (2) 1.5 -- 2.5 -- 0.75 Differential HSTL-12 Class I JESD8-16A (2) (2) 1.2 1.2 2.5 -- 0.60 Differential HSTL-12 Class II JESD8-16A (2) (2) 1.2 -- 2.5 -- 0.60 SSTL-15 Class I SSTL-15 Class II HSTL-18 Class I HSTL-18 Class II HSTL-15 Class I LVDS (4), (5), (8) ANSI/TIA/ EIA-644 (2) (2) 2.5 2.5 2.5 -- -- RSDS (6), (7), (8) -- (2) (2) 2.5 2.5 2.5 -- -- -- (2) (2) 2.5 2.5 2.5 -- -- mini-LVDS (6), (7), (8) Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-5 Table 6-2. I/O Standards and Voltage Levels for Stratix IV Devices (1) (Part 3 of 3) VCCIO (V) I/O Standard Standard Support Input Operation Column Row I/O Banks I/O Banks LVPECL -- (4) 2.5 Output Operation Column I/O Banks Row I/O Banks -- -- VTT (V) VCCPD (V) VREF (V) (Board (Pre-Driver (Input Ref Termination Voltage) Voltage) Voltage) 2.5 -- -- Notes to Table 6-2: (1) VCCPD is either 2.5 or 3.0 V. For VCCIO = 3.0 V, VCCPD = 3.0 V. For VCCIO = 2.5 V or less, VCCPD = 2.5 V. (2) Single-ended HSTL/SSTL, differential SSTL/HSTL, and LVDS input buffers are powered by VCCPD. Row I/O banks support both true differential input buffers and true differential output buffers. Column I/O banks support true differential input buffers, but not true differential output buffers. I/O pins are organized in pairs to support differential standards. Column I/O differential HSTL and SSTL inputs use LVDS differential input buffers without on-chip RD support. (3) For more information about the 3.3-V LVTTL/LVCMOS standard supported in Stratix IV devices, refer to "3.3-V I/O Interface" on page 6-19. (4) Column I/O banks support LVPECL I/O standards for input clock operation. Clock inputs on column I/Os are powered by VCCCLKIN when configured as differential clock inputs. They are powered by VCCIO when configured as single-ended clock inputs. Differential clock inputs in row I/Os are powered by VCCPD. (5) Column and row I/O banks support LVDS outputs using two single-ended output buffers, an external one-resistor (LVDS_E_1R), and a three-resistor (LVDS_E_3R) network. (6) Row I/O banks support RSDS and mini-LVDS I/O standards using a true LVDS output buffer without a resistor network. (7) Column and row I/O banks support RSDS and mini-LVDS I/O standards using two single-ended output buffers with one-resistor (RSDS_E_1R and mini-LVDS_E_1R) and three-resistor (RSDS_E_3R and mini-LVDS_E_3R) networks. (8) The emulated differential output standard that supports the tri-state feature includes: LVDS_E_1R, LVDS_E_3R, RSDS_E_1R, RSDS_E_3R, Mini_LVDS_E_1R, and Mini_LVDS_E_3R. For more information, refer to the I/O Buffer (ALTIOBUF) Megafunction User Guide. f For more information about the electrical characteristics of each I/O standard, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. I/O Banks Stratix IV devices contain up to 24 I/O banks, as shown in Figure 6-1 and Figure 6-2. The row I/O banks contain true differential input and output buffers and dedicated circuitry to support differential standards at speeds up to 1.6 Gbps. Each I/O bank in Stratix IV devices can support high-performance external memory interfaces with dedicated circuitry. The I/O pins are organized in pairs to support differential standards. Each I/O pin pair can support both differential input and output buffers. The only exceptions are the clk[1,3,8,10], PLL_L[1,4]_clk, and PLL_R[1,4]_clk pins, which support differential input operations only. f For information about the number of channels available for the LVDS I/O standard, refer to the High-Speed Differential I/O Interface and DPA in Stratix IV Devices chapter. For more information about transceiver-bank-related features, refer to the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-6 Chapter 6: I/O Features in Stratix IV Devices I/O Banks Bank 1A Bank 8A Bank 8B (1), (2), (3), (4), (5), (6), (7), (8) Bank 8C Bank 7B Bank 7C I/O banks 8A, 8B, and 8C support all single-ended and differential input and output operations except LVPECL, which is supported on clk input pins only. Bank 7A Bank 6C Bank 5C Bank 2C LVPECL I/O standard for input operation on dedicated clock input pins. Bank 2B SSTL-15 Class II, HSTL-15 Class II, HSTL-12 Class II, differential SSTL-15 Class II, differential HSTL-15 Class II, differential HSTL-12 Class II standards are only supported for input operations. I/O banks 4A, 4B, and 4C support all single-ended and differential input and output operations except LVPECL, which is supported on clk input pins only. Bank 2A I/O banks 3A, 3B, and 3C support all single-ended and differential input and output operations except LVPECL, which is supported on clk input pins only. Bank 3A Bank 3B Bank 3C Bank 4C Bank 4B Bank 5B Bank 1C Row I/O banks support LVTTL, LVCMOS, 2.5-V, 1.8-V, 1.5-V, 1.2-V, SSTL-2 Class I & II, SSTL-18 Class I & II, SSTL-15 Class I, HSTL-18 Class I & II, HSTL-15 Class I, HSTL-12 Class I, LVDS, RSDS, mini-LVDS, differential SSTL-2 Class I & II, differential SSTL-18 Class I & II, differential SSTL-15 Class I, differential HSTL-18 Class I & II, differential HSTL-15 Class I, and differential HSTL-12 Class I standards for input and output operations. Bank 5A Bank 1B Bank 6B I/O banks 7A, 7B, and 7C support all single-ended and differential input and output operations except LVPECL, which is supported on clk input pins only. Bank 6A Figure 6-1. Stratix IV E Devices I/0 Banks Bank 4A Notes to Figure 6-1: (1) Differential HSTL and SSTL outputs are not true differential outputs. They use two single-ended outputs with the second output programmed as inverted. (2) Column I/O differential HSTL and SSTL inputs use LVDS differential input buffers without differential OCT support. (3) Column I/O supports LVDS outputs using single-ended buffers and external resistor networks. (4) Column I/O supports PCI/PCI-X with on-chip clamp diode. Row I/O supports PCI/PCI-X with external clamp diode. (5) Clock inputs on column I/Os are powered by VCCCLKIN when configured as differential clock inputs. They are powered by VCCIO when configured as single-ended clock inputs. All outputs use the corresponding bank VCCIO. (6) Row I/O supports the true LVDS output buffer. (7) Column and row I/O banks support LVPECL standards for input clock operation. (8) Figure 6-1 is a top view of the silicon die that corresponds to a reverse view for flip chip packages. It is a graphical representation only. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-7 Bank 3A Bank 3B Bank 3C Bank 4C Bank 4B Transceiver Bank GXBR2 Bank 6C Bank5C I/O banks 4A, 4B & 4C support all single-ended and differential input and output operation. I/O banks 3A, 3B & 3C support all single-ended and differential input and output operation. Bank 5A Bank 2B Bank 5B Bank 2C Bank 1C Row I/O banks support LVTTL, LVCMOS, 2.5-V, 1.8V, 1.5-V, 1.2-V, SSTL-2 Class I & II, SSTL-18 Class I & II, SSTL-15 Class I, HSTL-18 Class I & II, HSTL-15 Class I, HSTL-12 Class I, LVDS, RSDS, mini-LVDS, differential SSTL-2 Class I & II, differential SSTL-18 Class I & II, differential SSTL-15 Class I, differential HSTL-18 Class I & II, differential HSTL-15 Class I and differential HSTL-12 Class I standards for input and output operation. SSTL-15 class II, HSTL-15 Class II, HSTL-12 Class II, differential SSTL-15 Class II, differential HSTL-15 Class II, differential HSTL-12 Class II standards are only supported for input operations Transceiver Bank GXBR1 Bank 6B Bank 1B I/O banks 7A, 7B & 7C support all single-ended and differential input and output operation. Transceiver Bank GXBR3 Bank 7A Transceiver Bank GXBR0 Bank 1A Bank 7B Bank 7C I/O banks 8A, 8B & 8C support all single-ended and differential input and output operation. Bank 2A Transceiver Bank GXBL3 Transceiver Bank GXBL2 Transceiver Bank GXBL1 Transceiver Bank GXBL0 Bank 8C Bank 8B Bank 8A (1), (2), (3), (4), (5), (6), (7), (8), (9) Bank 6A Figure 6-2. Stratix IV GX Devices I/O Banks Bank 4A Notes to Figure 6-2: (1) Differential HSTL and SSTL outputs are not true differential outputs. They use two single-ended outputs with the second output programmed as inverted. (2) Column I/O differential HSTL and SSTL inputs use LVDS differential input buffers without differential OCT support. (3) Column I/O supports LVDS outputs using single-ended buffers and external resistor networks. (4) Column I/O supports PCI/PCI-X with an on-chip clamp diode. Row I/O supports PCI/PCI-X with an external clamp diode. (5) Clock inputs on column I/Os are powered by VCCCLKIN when configured as differential clock inputs. They are powered by VCCIO when configured as single-ended clock inputs. All outputs use the corresponding bank VCCIO. (6) Row I/O supports the true LVDS output buffer. (7) Column and row I/O banks support LVPECL standards for input clock operation. (8) Figure 6-2 is a top view of the silicon die that corresponds to a reverse view for flip chip packages. It is a graphical representation only. (9) Stratix IV devices do not support the PCI clamp diode when VCCIO is 1.2 V, 1.5 V, or 1.8 V. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-8 Chapter 6: I/O Features in Stratix IV Devices I/O Banks Modular I/O Banks The I/O pins in Stratix IV devices are arranged in groups called modular I/O banks. Depending on device densities, the number of Stratix IV device I/O banks range from 16 to 24. The number of I/O pins on each bank is 24, 32, 36, 40, or 48. Figure 6-4 through Figure 6-16 show the number of I/O pins available in each I/O bank. In Stratix IV devices, the maximum number of I/O banks per side is either four or six, depending on the device density. When migrating between devices with a different number of I/O banks per side, it is the middle or "B" bank that is removed or inserted. For example, when moving from a 24-bank device to a 16-bank device, the banks that are dropped are "B" banks, namely: 1B, 2B, 3B, 4B, 5B, 6B, 7B, and 8B. Similarly, when moving from a 16-bank device to a 24-bank device, the banks that are added are the same "B" banks. After migration from a smaller device to a larger device, the bank size increases or remains the same, but never decreases. For example, the number of I/O pins to a bank may increase from 24 to 26, 32, 36, 40, 42, or 48, but will never decrease. This is shown in Figure 6-3. Figure 6-3. Bank Migration Path with Increasing Device Size 24 Stratix IV Device Handbook Volume 1 26 32 36 40 42 48 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-9 Figure 6-4 through Figure 6-16 show the number of I/O pins and packaging information for different sets of available devices. They show the top view of the silicon die that corresponds to a reverse view for flip chip packages. They are graphical representations only. 1 For Figure 6-4 through Figure 6-16, the pin count includes all general purpose I/Os, dedicated clock pins, and dual purpose configuration pins. Transceiver pins and dedicated configuration pins are not included in the pin count. 40 Bank 2C Bank 5C 26 Bank 2A Bank 5A 32 40 Bank 4A Bank Name Number of I/Os 40 32 Bank 7A 26 Bank 4C 26 24 Bank 6C EP4SE230 EP4SE360 24 Bank 1C Bank 7C 32 Bank 3C 26 24 Bank 6A 24 Bank 1A Bank 3A 32 Bank 8C Bank Name 40 Number of I/Os Bank 8A Figure 6-4. Number of I/Os in Each Bank in EP4SE230 and EP4SE360 Devices in the 780-Pin FineLine BGA Package 24 32 32 24 40 Bank 8B Bank 8C Bank 7C Bank 7B Altera Corporation Bank 7A Bank 4A 40 40 September 2012 Bank 4B Bank 2A 24 48 Bank 4C Bank 2C 32 42 EP4SE360 EP4SE530 EP4SE820 Bank 3C Bank 1C 32 42 Bank 3B Bank 1A Bank 3A 48 24 Bank Name 40 Number of I/Os Bank 8A Figure 6-5. Number of I/Os in Each Bank in EP4SE360, EP4SE530, and EP4SE820 Devices in the 1152-Pin FineLine BGA Package Bank 6A 48 Bank 6C 42 Bank 5C 42 Bank 5A 48 Bank Name Number of I/Os Stratix IV Device Handbook Volume 1 6-10 Chapter 6: I/O Features in Stratix IV Devices I/O Banks 48 32 32 48 48 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 48 Number of I/Os Bank 8A Figure 6-6. Number of I/Os in Each Bank in EP4SE530 and EP4SE820 Devices in the 1517-Pin FineLine BGA Package 50 Bank 1A Bank 6A 50 24 Bank 1B Bank 6B 24 42 Bank 1C Bank 6C 42 42 Bank 2C Bank 5C 42 24 Bank 2B Bank 5B 24 50 Bank 2A Bank 5A 50 48 Bank 4A 48 Bank 4B 32 Bank 4C 32 Bank 3C 48 Bank 3B 48 Bank 3A EP4SE530 EP4SE820 Bank Name Number of I/Os 48 48 48 48 48 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 48 Number of I/Os Bank 8A Figure 6-7. Number of I/Os in Each Bank in EP4SE530 and EP4SE820 Devices in the 1760-Pin Fineline BGA Package 50 Bank 1A Bank 6A 50 36 Bank 1B Bank 6B 36 50 Bank 1C Bank 6C 50 50 Bank 2C Bank 5C 50 36 Bank 2B Bank 5B 36 50 Bank 2A Bank 5A 50 Stratix IV Device Handbook Volume 1 48 Bank 4A 48 Bank 4B 48 Bank 4C 48 Bank 3C 48 Bank 3B 48 Bank 3A EP4SE530 EP4SE820 Bank Name Number of I/Os September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-11 40 Bank GXBR1 Bank 2A 24 24 40 Bank 7C Bank 7A Number of Transceiver Channels 4 Bank GXBR0 32 Bank 4A Bank 2C 40 26 EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 Bank 4C Bank 1C 24 26 Bank 3C Bank 1A Bank 3A 32 24 Bank Name Bank 8A 40 Number of I/Os Bank 8C Figure 6-8. Number of I/Os in Each Bank in EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 Devices in the 780-Pin FineLine BGA Package 4 Bank Name Number of I/Os Altera Corporation 32 40 Bank 7C Bank 7A Bank 4C Bank 4A 32 40 Bank GXBL0 Bank 3C 4 Number of Transceiver Channels Bank GXBR1 4 Bank GXBR0 4 EP4SGX290 EP4SGX360 32 Bank GXBL1 Bank 3A 4 Number of Transceiver Channels September 2012 32 Bank 1C 40 1 Bank 8C Bank Name 40 Number of I/Os Bank 8A Figure 6-9. Number of I/Os in Each Bank in EP4SGX290 and EP4SGX360 Devices in the 780-Pin FineLine BGA Package Bank Name Number of I/Os Stratix IV Device Handbook Volume 1 6-12 Chapter 6: I/O Features in Stratix IV Devices I/O Banks Bank GXBL0 24 24 40 Bank 7C Bank 7A 40 *Number of Transceiver Channels Bank 4A 4* Bank 6A 32 Bank 6C 26 Bank GXBR1 4* Bank GXBR0 4* Bank Name Number of I/Os 40 Bank GXBL1 Bank 4C 4* EP4SGX70 EP4SGX110 24 Bank 1C Bank 3C 26 24 Bank 1A Bank 3A 32 Bank 8C Bank Name 40 Number of I/Os Bank 8A Figure 6-10. Number of I/Os in Each Bank in EP4SGX70 and EP4SGX110 Devices in the 1152-Pin FineLine BGA Package 32 32 24 40 Bank 8C Bank 7C Bank 7B Bank 7A 24 Bank 8B Bank 4B Bank 4A 40 40 24 Bank GXBL0 Bank 4C 4 (2) 32 4 (2) Bank GXBL1 Bank 3C Bank 1C 32 42 EP4SGX180 EP4SGX230 EP4SGX290 EP4SGX360 EP4SGX530 Bank 3B Bank 1A Bank 3A 48 24 Bank Name 40 Number of I/Os Bank 8A Figure 6-11. Number of I/Os in Each Bank in EP4SGX180, EP4SGX230, EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1152-Pin FineLine BGA Package (1), (2) Bank 6A 48 Bank 6C 42 Bank GXBR1 4 (2) Bank GXBR0 4 (2) Bank Name Number of I/Os Notes to Figure 6-11: (1) Except for the EP4SGX530 device, all listed devices have two variants in the F1152 package option--one with no PMA-only transceiver channels and the other with two PMA-only transceiver channels for each transceiver bank. The EP4SGX530 device is only offered with two PMA-only transceiver channels for each transceiver bank in the F1152 package option. (2) There are two additional PMA-only transceiver channels in each transceiver bank for devices with the PMA-only transceiver package option. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-13 24 32 32 24 40 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 40 Number of I/Os Bank 8A Figure 6-12. Number of I/Os in Each Bank in EP4SGX180, EP4SGX230, EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1517-Pin FineLine BGA Package (1) 48 Bank 1A Bank 6A 48 42 Bank 1C Bank 6C 42 42 Bank 2C Bank 5C 42 48 Bank 2A Bank 5A 48 EP4SGX180 EP4SGX230 EP4SGX290 EP4SGX360 EP4SGX530 Bank GXBR1 4 (1) 4 (1) Bank GXBL0 Bank GXBR0 4 (1) 40 24 32 32 24 40 Bank 4A Bank GXBL1 Bank 4B 4 (1) Bank 4C 4 (1) Bank 3C Bank GXBR2 Bank 3B Bank GXBL2 Bank 3A 4 (1) Bank Name Number of I/Os Note to Figure 6-12: (1) There are two additional PMA-only transceiver channels in each transceiver bank. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-14 Chapter 6: I/O Features in Stratix IV Devices I/O Banks 48 32 32 48 48 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 48 Number of I/Os Bank 8A Figure 6-13. Number of I/Os in Each Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1932-Pin FineLine BGA Package (1) 50 Bank 1A Bank 6A 50 42 Bank 1C Bank 6C 42 42 Bank 2C Bank 5C 42 20 Bank 2B Bank 5B 20 50 Bank 2A Bank 5A 50 4 (1) Bank GXBL3 Bank GXBR3 4 (1) 4 (1) Bank GXBL2 Bank GXBR2 4 (1) 4 (1) Bank GXBL1 4 (1) 4 (1) Bank GXBL0 Bank GXBR1 Bank GXBR0 48 Bank 4A 48 Bank 4B 32 Bank 4C 32 Bank 3C 48 Bank 3B 48 Bank 3A EP4SGX530 EP4SGX290 EP4SGX360 4 (1) Bank Name Number of I/Os Note to Figure 6-13: (1) There are two additional PMA-only transceiver channels in each transceiver bank. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Banks 6-15 48 32 32 48 48 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 48 Number of I/Os Bank 8A Figure 6-14. Number of I/Os in Each Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1760-Pin FineLine BGA Package (1) 50 Bank 1A Bank 6A 50 42 Bank 1C Bank 6C 42 42 Bank 2C Bank 5C 42 Bank 5A 50 50 EP4SGX290 EP4SGX360 EP4SGX530 Bank 2A Bank GXBR1 4 (1) 4 (1) Bank GXBL0 Bank GXBR0 4 (1) 48 48 32 32 48 48 Bank 4A Bank GXBL1 Bank 4B 4 (1) Bank 4C 4 (1) Bank 3C Bank GXBR2 Bank 3B Bank GXBL2 Bank 3A 4 (1) Bank Name Number of I/Os Note to Figure 6-14: (1) There are two additional PMA-only transceiver channels in each transceiver bank. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-16 Chapter 6: I/O Features in Stratix IV Devices I/O Banks 1 The information in Figure 6-15 and Figure 6-16 applies to Stratix IV GX and GT devices. 48 32 32 48 48 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 48 Number of I/Os Bank 8A Figure 6-15. Number of I/Os in Each Bank in EP4S100G3, EP4S100G4, and EP4S100G5 Devices in the 1932-Pin FineLine BGA Package (1) 40 Bank 1A Bank 6A 38 21 Bank 1C Bank 6C 22 21 Bank 2C Bank 5C 19 13 Bank 2B Bank 5B 12 41 Bank 2A Bank 5A 42 4 (1) Bank GXBL2 Bank GXBR2 4 (1) 4 (1) Bank GXBL1 Bank GXBR1 4 (1) 4 (1) Bank GXBL0 Bank GXBR0 4 (1) Bank 4A 48 Bank 4B 48 Bank 4C 32 Bank 3C 32 Bank 3B 48 48 Bank 3A EP4S100G3 EP4S100G4 EP4S100G5 Bank Name Number of I/Os Note to Figure 6-15: (1) There are two additional PMA-only transceiver channels in each transceiver bank. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Structure 6-17 24 32 32 24 40 Bank 8B Bank 8C Bank 7C Bank 7B Bank 7A Bank Name 40 Number of I/Os Bank 8A Figure 6-16. Number of I/Os in Each Bank in EP4S40G2, EP4S40G5, EP4S100G2, and EP4S100G5 Devices in the 1517-Pin FineLine BGA Package (1) 43 Bank 1A Bank 6A 44 22 Bank 1C Bank 6C 23 23 Bank 2C Bank 5C 23 46 Bank 2A Bank 5A 46 EP4S40G2 EP4S40G5 EP4S100G2 EP4S100G5 Bank GXBR1 4 (1) 4 (1) Bank GXBL0 Bank GXBR0 4 (1) 40 24 32 32 24 40 Bank 4A Bank GXBL1 Bank 4B 4 (1) Bank 4C 4 (1) Bank 3C Bank GXBR2 Bank 3B Bank GXBL2 Bank 3A 4 (1) Bank Name Number of I/Os Note to Figure 6-16: (1) There are two additional PMA-only transceiver channels in each transceiver bank. I/O Structure The I/O element (IOE) in Stratix IV devices contain a bidirectional I/O buffer and I/O registers to support a complete embedded bidirectional single data rate or DDR transfer. The IOEs are located in I/O blocks around the periphery of the Stratix IV device. There are up to four IOEs per row I/O block and four IOEs per column I/O block. The row IOEs drive row, column, or direct link interconnects. The column IOEs drive column interconnects. The Stratix IV bidirectional IOE also supports the following features: September 2012 Programmable input delay Programmable output-current strength Programmable slew rate Programmable output delay Programmable bus-hold Programmable pull-up resistor Open-drain output On-chip series termination with calibration Altera Corporation Stratix IV Device Handbook Volume 1 6-18 Chapter 6: I/O Features in Stratix IV Devices I/O Structure On-chip series termination without calibration On-chip parallel termination with calibration On-chip differential termination PCI clamping diode I/O registers are composed of the input path for handling data from the pin to the core, the output path for handling data from the core to the pin, and the output-enable (OE) path for handling the OE signal to the output buffer. These registers allow faster source-synchronous register-to-register transfers and resynchronization. The input path consists of the DDR input registers, alignment and synchronization registers, and HDR. You can bypass each block of the input path. The output and OE paths are divided into output or OE registers, alignment registers, and HDR blocks. You can bypass each block of the output and OE paths. Figure 6-17 shows the Stratix IV IOE structure. Figure 6-17. IOE Structure in Stratix IV Devices (1), (2), (3), (4) Firm Core DQS Logic Block OE Register D OE from Core 2 Half Data Rate Block D6_OCT D5_OCT PRN Q Dynamic OCT Control (2) Alignment Registers OE Register D VCCIO D5, D6 Delay PRN Q VCCIO PCI Clamp Programmable Pull-Up Resistor Programmable Current Strength and Slew Rate Control Output Register Write Data from Core Half Data Rate Block 4 Alignment Registers PRN D Q From OCT Calibration Block Output Buffer D5, D6 Delay Output Register D Open Drain PRN Q D2 Delay Input Buffer D3_0 Delay clkout To Core D3_1 Delay To Core Read Data to Core 4 Half Data Rate Block Alignment and Synchronization Registers D1 Delay Bus-Hold Circuit Input Register PRN D Q Input Register Input Register PRN D DQS CQn On-Chip Termination PRN Q D Q D4 Delay clkin Notes to Figure 6-17: (1) The following features are not supported by true differential standards: open drain or tri-state output,; programmable current strength and slew rate control; PCI Clamp; programmable pull-up resistor; bus-hold circuit. (2) The D3_0 and D3_1 delays have the same available settings in the Quartus(R) II software (3) One dynamic OCT control is available per DQ/DQS group. (4) Column I/O supports PCI/PCI-X with an on-chip clamp diode. Row I/O supports PCI/PCI-X with an external clamp diode. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Structure 6-19 f For more information about I/O registers and how they are used for memory applications, refer to the External Memory Interfaces in Stratix IV Devices chapter. 3.3-V I/O Interface Stratix IV I/O buffers support 3.3-V I/O standards. You can use them as transmitters or receivers in your system. The output high voltage (VOH), output low voltage (VOL), input high voltage (VIH), and input low voltage (VIL) levels meet the 3.3-V I/O standards specifications defined by EIA/JEDEC Standard JESD8-B with margin when the Stratix IV VCCIO voltage is powered by 3.0 V. To ensure device reliability and proper operation, when interfacing with a 3.3-V I/O system using Stratix IV devices, ensure that you do not violate the absolute maximum ratings of the devices. Altera recommends performing IBIS simulation to determine that the overshoot and undershoot voltages are within the guidelines. When using the Stratix IV device as a transmitter, you can use slow slew rate and series termination to limit overshoot and undershoot at the I/O pins, but they are not required. Transmission line effects that cause large voltage deviations at the receiver are associated with an impedance mismatch between the driver and the transmission lines. By matching the impedance of the driver to the characteristic impedance of the transmission line, you can significantly reduce overshoot voltage. You can use a series termination resistor placed physically close to the driver to match the total driver impedance to the transmission line impedance. Stratix IV devices support series OCT for all LVTTL and LVCMOS I/O standards in all I/O banks. When using the Stratix IV device as a receiver, you can use a clamping diode (on-chip or off-chip) to limit overshoot, though this is not required. Stratix IV devices provide an optional on-chip PCI-clamping diode for column I/O pins. You can use this diode to protect the I/O pins against overshoot voltage. The 3.3-V I/O standard is supported using bank supply voltage (VCCIO) at 3.0 V. In this method, the clamping diode (on-chip or off-chip), when enabled, can sufficiently clamp overshoot voltage to within the DC and AC input voltage specifications. The clamped voltage can be expressed as the sum of the supply voltage (VCCIO) and the diode forward voltage. f For more information about the absolute maximum rating and maximum allowed overshoot during transitions, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. External Memory Interfaces In addition to the I/O registers in each IOE, Stratix IV devices also have dedicated registers and phase-shift circuitry on all I/O banks for interfacing with external memory interfaces. f For more information about external memory interfaces, refer to the External Memory Interfaces in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-20 Chapter 6: I/O Features in Stratix IV Devices I/O Structure High-Speed Differential I/O with DPA Support Stratix IV devices have the following dedicated circuitry for high-speed differential I/O support: Differential I/O buffer Transmitter serializer Receiver deserializer Data realignment Dynamic phase aligner (DPA) Synchronizer (FIFO buffer) Phase-locked loops (PLLs) f For more information about DPA support, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Programmable Current Strength The output buffer for each Stratix IV device I/O pin has a programmable current strength control for certain I/O standards. Use programmable current strength to mitigate the effects of high signal attenuation due to a long transmission line or a legacy backplane. The LVTTL, LVCMOS, SSTL, and HSTL standards have several levels of current strength that you can control. Table 6-3 lists the programmable current strength for Stratix IV devices. Table 6-3. Programmable Current Strength (Part 1 of 2) (1), I/O Standard 3.3-V LVTTL Stratix IV Device Handbook Volume 1 (2) IOH / IOL Current Strength Setting (mA) for Column I/O Pins IOH / IOL Current Strength Setting (mA) for Row I/O Pins 16, 12, 8, 4 12, 8, 4 3.3-V LVCMOS 16, 12, 8, 4 8, 4 2.5-V LVCMOS 16, 12, 8, 4 12, 8, 4 1.8-V LVCMOS 12, 10, 8, 6, 4, 2 8, 6, 4, 2 1.5-V LVCMOS 12, 10, 8, 6, 4, 2 8, 6, 4, 2 1.2-V LVCMOS 8, 6, 4, 2 4, 2 SSTL-2 Class I 12, 10, 8 12, 8 SSTL-2 Class II 16 16 SSTL-18 Class I 12, 10, 8, 6, 4 12, 10, 8, 6, 4 SSTL-18 Class II 16, 8 16, 8 SSTL-15 Class I 12, 10, 8, 6, 4 8, 6, 4 SSTL-15 Class II 16, 8 -- HSTL-18 Class I 12, 10, 8, 6, 4 12, 10, 8, 6, 4 HSTL-18 Class II 16 16 HSTL-15 Class I 12, 10, 8, 6, 4 8, 6, 4 HSTL-15 Class II 16 -- September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Structure 6-21 Table 6-3. Programmable Current Strength (Part 2 of 2) (1), (2) IOH / IOL Current Strength Setting (mA) for Column I/O Pins IOH / IOL Current Strength Setting (mA) for Row I/O Pins HSTL-12 Class I 12, 10, 8, 6, 4 8, 6, 4 HSTL-12 Class II 16 -- I/O Standard Notes to Table 6-3: (1) The default setting in the Quartus II software is 50-OCT RS without calibration for all non-voltage reference and HSTL and SSTL Class I I/O standards. The default setting is 25-OCT RS without calibration for HSTL and SSTL Class II I/O standards. (2) The 3.3-V LVTTL and 3.3-V LVCMOS are supported using VCCIO and VCCPD at 3.0 V. 1 Altera recommends performing IBIS or SPICE simulations to determine the best current strength setting for your specific application. Programmable Slew Rate Control The output buffer for each Stratix IV device regular- and dual-function I/O pin has a programmable output slew-rate control that you can configure for low-noise or high-speed performance. A faster slew rate provides high-speed transitions for high-performance systems. A slower slew rate can help reduce system noise, but adds a nominal delay to the rising and falling edges. Each I/O pin has an individual slew-rate control, allowing you to specify the slew rate on a pin-by-pin basis. 1 You cannot use the programmable slew rate feature when using OCT. The Quartus II software allows four settings for programmable slew rate control--0, 1, 2, and 3--where 0 is slow slew rate and 3 is fast slew rate. Figure 6-4 lists the default slew rate settings from the Quartus II software. Table 6-4. Default Slew Rate Settings I/O Standard Slew Rate Option Default Slew Rate 1.2-V, 1.5-V, 1.8-V, 2.5-V LVCMOS, and 3.3-V LVTTL/LVCMOS 0, 1, 2, 3 3 SSTL-2, SSTL-18, SSTL-15, HSTL-18, HSTL-15, and HSTL-12 0, 1, 2, 3 3 3.0-V PCI/PCI-X 0, 1, 2, 3 3 LVDS_E_1R, mini-LVDS_E_1R, and RSDS_E_1R 0, 1, 2, 3 3 LVDS_E_3R, mini-LVDS_E_3R, and RSDS_E_3R 0, 1, 2, 3 3 You can use faster slew rates to improve the available timing margin in memory-interface applications or when the output pin has high-capacitive loading. 1 September 2012 Altera recommends performing IBIS or SPICE simulations to determine the best slew rate setting for your specific application. Altera Corporation Stratix IV Device Handbook Volume 1 6-22 Chapter 6: I/O Features in Stratix IV Devices I/O Structure Programmable I/O Delay The following sections describe programmable IOE delay and programmable output buffer delay. Programmable IOE Delay The Stratix IV device IOE includes programmable delays, shown in Figure 6-17 on page 6-18, that you can activate to ensure zero hold times, minimize setup times, or increase clock-to-output times. Each pin can have a different input delay from pin-to-input register or a delay from output register-to-output pin values to ensure that the bus has the same delay going into or out of the device. This feature helps read and time margins because it minimizes the uncertainties between signals in the bus. f For more information about programmable IOE delay specifications, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Programmable Output Buffer Delay Stratix IV devices support delay chains built inside the single-ended output buffer, as shown in Figure 6-17 on page 6-18. The delay chains can independently control the rising and falling edge delays of the output buffer, providing the ability to adjust the output-buffer duty cycle, compensate channel-to-channel skew, reduce simultaneous switching output (SSO) noise by deliberately introducing channel-to-channel skew, and improve high-speed memory-interface timing margins. Stratix IV devices support four levels of output buffer delay settings. The default setting is No Delay. f For more information about programmable output buffer delay specifications, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Open-Drain Output Stratix IV devices provide an optional open-drain output (equivalent to an open collector output) for each I/O pin. When configured as open drain, the logic value of the output is either high-Z or 0. Typically, an external pull-up resistor is required to provide logic high. Bus Hold Each Stratix IV device I/O pin provides an optional bus-hold feature. Bus-hold circuitry can weakly hold the signal on an I/O pin at its last-driven state. Because the bus-hold feature holds the last-driven state of the pin until the next input signal is present, you do not need an external pull-up or pull-down resistor to hold a signal level when the bus is tri-stated. Bus-hold circuitry also pulls non-driven pins away from the input threshold voltage where noise can cause unintended high-frequency switching. You can select this feature individually for each I/O pin. The bus-hold output drives no higher than VCCIO to prevent over-driving signals. If you enable the bus-hold feature, you cannot use the programmable pull-up option. Disable the bus-hold feature if the I/O pin is configured for differential signals. Bus-hold circuitry uses a resistor with a nominal resistance (RBH) of approximately 7 k to weakly pull the signal level to the last-driven state. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices I/O Structure 6-23 f For more information about the specific sustaining current driven through this resistor and the overdrive current used to identify the next-driven input level, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Bus-hold circuitry is active only after configuration. When going into user mode, the bus-hold circuit captures the value on the pin present at the end of configuration. Programmable Pull-Up Resistor Each Stratix IV device I/O pin provides an optional programmable pull-up resistor during user mode. If you enable this feature for an I/O pin, the pull-up resistor (typically 25 K ) weakly holds the I/O to the VCCIO level. Programmable pull-up resistors are only supported on user I/O pins and are not supported on dedicated configuration pins, JTAG pins, or dedicated clock pins. If you enable the programmable pull-up option, you cannot use the bus-hold feature. 1 When the optional DEV_OE signal drives low, all the I/O pins remain tri-stated even with the programmable pull-up option enabled. Programmable Pre-Emphasis Stratix IV LVDS transmitters support programmable pre-emphasis to compensate for the frequency dependent attenuation of the transmission line. The Quartus II software allows four settings for programmable pre-emphasis. f For more information about programmable pre-emphasis, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Programmable Differential Output Voltage Stratix IV LVDS transmitters support programmable VOD. The programmable VOD settings allow you to adjust output eye height to optimize trace length and power consumption. A higher VOD swing improves voltage margins at the receiver end; a smaller VOD swing reduces power consumption. The Quartus II software allows four settings for programmable VOD. f For more information about programmable VOD, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. MultiVolt I/O Interface The Stratix IV architecture supports the MultiVolt I/O interface feature that allows the Stratix IV devices in all packages to interface with systems of different supply voltages. You can connect the VCCIO pins to a 1.2-, 1.5-, 1.8-, 2.5-, or 3.0-V power supply, depending on the output requirements. The output levels are compatible with systems of the same voltage as the power supply. (For example, when VCCIO pins are connected to a 1.5-V power supply, the output levels are compatible with 1.5-V systems.) September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-24 Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes f For more information about pin connection guidelines, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. The Stratix IV VCCPD power pins must be connected to a 2.5- or 3.0-V power supply. Using these power pins to supply the pre-driver power to the output buffers increases the performance of the output pins. Table 6-5 lists Stratix IV MultiVolt I/O support. Table 6-5. Stratix IV MultiVolt I/O Support (1) Input Signal (V) VCCIO (V) Output Signal (V) (3) 1.2 1.5 1.8 2.5 3.0 3.3 1.2 1.5 1.8 2.5 3.0 3.3 1.2 Y -- -- -- -- 1.5 -- Y Y -- -- -- Y -- -- -- -- -- -- -- Y -- -- -- -- 1.8 -- Y Y -- -- -- -- -- Y -- -- -- 2.5 -- -- -- Y Y (2) Y (2) -- -- -- Y -- -- 3.0 -- -- -- Y Y Y -- -- -- -- Y -- Notes to Table 6-5: (1) The pin current may be slightly higher than the default value. You must verify that the driving device's VOL maximum and VOH minimum voltages do not violate the applicable Stratix IV VIL maximum and VIH minimum voltage specifications. (2) Altera recommends that you use an external clamping diode on the I/O pins when the input signal is 3.0 V or 3.3 V. You have the option to use an internal clamping diode for column I/O pins. (3) Each I/O bank of a Stratix IV device has its own VCCIO pins and supports only one VCCIO, either 1.2, 1.5, 1.8, or 3.0 V. The LVDS I/O standard is not supported when VCCIO is 3.0 V. The LVDS input operations are supported when VCCIO is 1.2 V, 1.5 V, 1.8 V, or 2.5 V. The LVDS output operations are only supported when VCCIO is 2.5 V. On-Chip Termination Support and I/O Termination Schemes Stratix IV devices feature dynamic series and parallel OCT to provide I/O impedance matching and termination capabilities. OCT maintains signal quality, saves board space, and reduces external component costs. Stratix IV devices support: On-chip series termination (RS) with calibration On-chip series termination (RS) without calibration On-chip Parallel termination (RT) with calibration Dynamic series termination for single-ended I/O standards Dynamic Parallel termination for single-ended I/O standards On-chip differential termination (RD) for differential LVDS I/O standards Stratix IV devices support OCT in all I/O banks by selecting one of the OCT I/O standards. These devices also support OCT RS and RT in the same I/O bank for different I/O standards if they use the same VCCIO supply voltage. You can independently configure each I/O in an I/O bank to support OCT RS, programmable current strength, or OCT RT. 1 Stratix IV Device Handbook Volume 1 You cannot configure both OCT RS and programmable current strength for the same I/O buffer. September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes 6-25 A pair of RUP and RDN pins are available in a given I/O bank and are shared for series- and parallel-calibrated termination. The RUP and RDN pins share the same VCCIO and GND, respectively, with the I/O bank where they are located. The RUP and RDN pins are dual-purpose I/Os and function as regular I/Os if you do not use the calibration circuit. For calibration, the connections are as follows: The RUP pin is connected to VCCIO through an external 25- 1% or 50- 1% resistor for an on-chip series termination value of 25-or 50-, respectively. The RDN pin is connected to GND through an external 25- 1% or 50- 1% resistor for an on-chip series termination value of 25-or 50-, respectively. For on-chip parallel termination, the connections are as follows: The RUP pin is connected to VCCIO through an external 50- 1% resistor. The RDN pin is connected to GND through an external 50- 1% resistor. On-Chip Series (RS) Termination Without Calibration Stratix IV devices support driver-impedance matching to provide the I/O driver with controlled output impedance that closely matches the impedance of the transmission line. As a result, you can significantly reduce reflections. Stratix IV devices support on-chip series termination for single-ended I/O standards (Figure 6-18). The RS shown in Figure 6-18 is the intrinsic impedance of the output transistors. Typical RS values are 25 and 50 . When you select matching impedance, current strength is no longer selectable. Figure 6-18. On-Chip Series Termination Without Calibration Stratix IV Driver Series Termination Receiving Device VCCIO RS ZO = 50 RS GND To use on-chip termination for the SSTL Class I standard, you must select the 50- on-chip series termination setting, thus eliminating the external 25- RS (to match the 50- transmission line). For the SSTL Class II standard, you must select the 25- on-chip series termination setting (to match the 50- transmission line and the near-end external 50- pull-up to VTT). September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-26 Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes On-Chip Series Termination with Calibration Stratix IV devices support on-chip series termination with calibration in all banks. The on-chip series termination calibration circuit compares the total impedance of the I/O buffer to the external 25- 1% or 50- 1% resistors connected to the RUP and RDN pins and dynamically enables or disables the transistors until they match. The RS shown in Figure 6-19 is the intrinsic impedance of the transistors. Calibration occurs at the end of device configuration. When the calibration circuit finds the correct impedance, it powers down and stops changing the characteristics of the drivers. Figure 6-19. On-Chip Series Termination with Calibration Stratix IV Driver Series Termination Receiving Device VCCIO RS ZO = 50 RS GND Table 6-6 lists the I/O standards that support on-chip series termination with and without calibration. Table 6-6. Selectable I/O Standards for On-Chip Series Termination with and Without Calibration (Part 1 of 2) On-Chip Series Termination Setting I/O Standard 3.3-V LVTTL/LVCMOS 2.5-V LVCMOS 1.8-V LVCMOS Stratix IV Device Handbook Volume 1 Row I/O () Column I/O () 50 50 25 25 50 50 25 25 50 50 25 25 50 1.5-V LVCMOS 50 1.2-V LVCMOS 50 SSTL-2 Class I 50 50 SSTL-2 Class II 25 25 SSTL-18 Class I 50 50 SSTL-18 Class II 25 25 25 50 25 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes 6-27 Table 6-6. Selectable I/O Standards for On-Chip Series Termination with and Without Calibration (Part 2 of 2) On-Chip Series Termination Setting I/O Standard Row I/O () Column I/O () SSTL-15 Class I 50 50 SSTL-15 Class II -- 25 HSTL-18 Class I 50 50 HSTL-18 Class II 25 25 HSTL-15 Class I 50 50 HSTL-15 Class II -- 25 HSTL-12 Class I 50 50 HSTL-12 Class II -- 25 Left-Shift Series Termination Control Stratix IV devices support left-shift series termination control. You can use left-shift series termination control to get the calibrated OCT RS with half of the impedance value of the external reference resistors connected to the RUP and RDN pins. This feature is useful in applications that require both 25- and 50- calibrated OCT RS at the same VCCIO. For example, if your application requires 25- and 50- calibrated OCT RS for SSTL-2 Class I and Class II I/O standards, you only need one OCT calibration block with 50- external reference resistors. You can enable the left-shift series termination control feature in the ALTIOBUF megafunction in the Quartus II software. The Quartus II software only allows left-shift series termination control for 25- calibrated OCT RS with 50- external reference resistors connected to the RUP and RDN pins. You can only use left-shift series termination control for the I/O standards that support 25- calibrated OCT RS . 1 This feature is automatically enabled if you are using a bidirectional I/O with 25- calibrated OCT RS and 50- parallel OCT. f For more information about how to enable the left-shift series termination feature in the ALTIOBUF megafunction, refer to the I/O Buffer (ALTIOBUF) Megafunction User Guide. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-28 Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes On-Chip Parallel Termination with Calibration Stratix IV devices support on-chip parallel termination with calibration in all banks. On-chip parallel termination with calibration is only supported for input configuration of input and bidirectional pins. Output pin configurations do not support on-chip parallel termination with calibration. Figure 6-20 shows on-chip parallel termination with calibration. When you use parallel OCT, the VCCIO of the bank must match the I/O standard of the pin where the parallel OCT is enabled. Figure 6-20. On-Chip Parallel Termination with Calibration Stratix IV OCT VCCIO 100 ZO = 50 V REF 100 GND Transmitter Receiver The on-chip parallel termination calibration circuit compares the total impedance of the I/O buffer to the external 50- 1% resistors connected to the RUP and RDN pins and dynamically enables or disables the transistors until they match. Calibration occurs at the end of device configuration. When the calibration circuit finds the correct impedance, it powers down and stops changing the characteristics of the drivers. Table 6-7 lists the I/O standards that support on-chip parallel termination with calibration. Table 6-7. Selectable I/O Standards with On-Chip Parallel Termination with Calibration I/O Standard Stratix IV Device Handbook Volume 1 On-Chip Parallel Termination Setting (Column I/O) () On-Chip Parallel Termination Setting (Row I/O) () SSTL-2 Class I, II 50 50 SSTL-18 Class I, II 50 50 SSTL-15 Class I, II 50 50 HSTL-18 Class I, II 50 50 HSTL-15 Class I, II 50 50 HSTL-12 Class I, II 50 50 Differential SSTL-2 Class I, II 50 50 Differential SSTL-18 Class I, II 50 50 Differential SSTL-15 Class I, II 50 50 Differential HSTL-18 Class I, II 50 50 Differential HSTL-15 Class I, II 50 50 Differential HSTL-12 Class I, II 50 50 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes 6-29 Expanded On-Chip Series Termination with Calibration OCT calibration circuits always adjust OCT RS to match the external resistors connected to the RUP and RDN pin; however, it is possible to achieve OCT RS values other than the 25- and 50- resistors. Theoretically, if you need a different OCT RS value, you can change the resistance connected to the RUP and RDN pins accordingly. Practically, the OCT RS range that Stratix IV devices support is limited because of output buffer size and granularity limitations. The Quartus II software only allows discrete OCT RS calibration settings of 25, 40, 50, and 60 . You can select the closest discrete value of OCT RS with calibration settings in the Quartus II software to your system to achieve the closest timing. For example, if you are using 20- OCT RS with calibration in your system, you can select the 25- OCT RS with calibration setting in the Quartus II software to achieve the closest timing. Table 6-8 lists expanded OCT RS with calibration supported in Stratix IV devices. Use expanded on-chip series termination with calibration of SSTL and HSTL for impedance matching to improve signal integrity but do not use it to meet the JEDEC standard. Table 6-8. Selectable I/O Standards with Expanded On-Chip Series Termination with Calibration Range Expanded OCT RS Range I/O Standard Row I/O () Column I/O () 3.3-V LVTTL/LVCMOS 20-60 20-60 2.5-V LVTTL/LVCMOS 20-60 20-60 1.8-V LVTTL/LVCMOS 20-60 20-60 1.5-V LVTTL/LVCMOS 40-60 20-60 1.2-V LVTTL/LVCMOS 40-60 20-60 SSTL-2 20-60 20-60 SSTL-18 20-60 20-60 SSTL-15 40-60 20-60 HSTL-18 20-60 20-60 HSTL-15 40-60 20-60 HSTL-12 40-60 20-60 Dynamic On-Chip Termination Stratix IV devices support on and off dynamic termination, both series and parallel, for a bidirectional I/O in all I/O banks. Figure 6-21 shows the termination schemes supported in Stratix IV devices. Dynamic parallel termination is enabled only when the bidirectional I/O acts as a receiver and is disabled when it acts as a driver. Similarly, dynamic series termination is enabled only when the bidirectional I/O acts as a driver and is disabled when it acts as a receiver. This feature is useful for terminating any high-performance bidirectional path because signal integrity is optimized depending on the direction of the data. Using dynamic OCT helps save power because device termination is internal instead of external. Termination only switches on during input operation, thus drawing less static power. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-30 Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes 1 When using calibrated input parallel and calibrated output series termination on bidirectional pins, they must use the same termination value because each I/O pin can only reference one OCT calibration block. The only exception is when using 50 parallel OCT and 25 series OCT using the left shift series termination control. For example, you cannot use calibrated 50 parallel OCT on the input buffer of a bidirectional pin and calibrated 40 series OCT on the output buffer because these would require two separate calibration blocks with different RUP and RDN resistor values. Figure 6-21. Dynamic Parallel OCT in Stratix IV Devices VCCIO VCCIO Transmitter Receiver 100 100 50 ZO = 50 100 100 50 GND GND Stratix IV OCT Stratix IV OCT VCCIO VCCIO 100 100 50 ZO = 50 100 100 50 GND GND Transmitter Receiver Stratix IV OCT Stratix IV OCT f For more information about tolerance specifications for OCT with calibration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices On-Chip Termination Support and I/O Termination Schemes 6-31 LVDS Input OCT (RD) Stratix IV devices support OCT for differential LVDS input buffers with a nominal resistance value of 100 , as shown in Figure 6-22. Differential OCT RD can be enabled in row I/O banks when both the VCCIO and VCCPD is set to 2.5 V. Column I/O banks do not support OCT RD. Dedicated clock input pairs CLK[1,3,8,10][p,n], PLL_L[1,4]_CLK[p,n], and PLL_R[1,4]_CLK[p,n] on the row I/O banks of Stratix IV devices do not support RD termination. Figure 6-22. Differential Input OCT Transmitter Receiver ZO = 50 100 ZO = 50 f For more information about differential on-chip termination, refer to the High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices chapter. Summary of OCT Assignments Table 6-9 lists the OCT assignments for the Quartus II software version 9.1 and later. Table 6-9. Summary of OCT Assignments in the Quartus II Software Assignment Name Value Applies To Parallel 50 with calibration Input buffers for single-ended and differential HSTL/SSTL standards Differential Input buffers for LVDS receivers on row I/O banks (1) Input Termination Series 25 without calibration Series 50 without calibration Output Termination Series 25 with calibration Series 40 with calibration Output buffers for single-ended LVTTL/LVCMOS and HSTL/SSTL standards as well as differential HSTL/SSTL standards Series 50 with calibration Series 60 with calibration Note to Table 6-9: (1) You can enable differential OCT RD in row I/O banks when both VCCIO and VCCPD are set to 2.5 V. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-32 Chapter 6: I/O Features in Stratix IV Devices OCT Calibration OCT Calibration Stratix IV devices support calibrated on-chip series termination (RS) and calibrated on-chip parallel termination (RT) on all I/O pins. You can calibrate the device's I/O bank with any of the OCT calibration blocks available in the device provided the VCCIO of the I/O bank with the pins using calibrated OCT matches the VCCIO of the I/O bank with the calibration block and its associated RUP and RDN pins. OCT Calibration Block Location Table 6-10 and Table 6-11 list the location of OCT calibration blocks in Stratix IV devices. For both tables, the following legend applies: 1 "Y" indicates I/O banks with OCT calibration block "N" indicates I/O banks without OCT calibration block "--" indicates I/O banks that are not available in the device Table 6-10 and Table 6-11 do not show transceiver banks and transceiver calibration blocks. Table 6-10 lists the OCT calibration blocks in Banks 1A through 4C. Table 6-10. OCT Calibration Block Counts and Placement in Stratix IV Devices (1A through 4C) (Part 1 of 2) Device EP4SE230 EP4SE360 EP4SE530 EP4SE820 EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 Stratix IV Device Handbook Volume 1 Bank Number of OCT Blocks 1A 1B 1C 2A 2B 2C 3A 3B 3C 4A 4B 4C 780 8 Y -- N Y -- N Y -- N Y -- N 780 8 Y -- N Y -- N Y -- N Y -- N 1152 8 Y -- N Y -- N Y N N Y N N 1152 8 Y -- N Y -- N Y N N Y N N 1517 10 Y N N Y N N Y N Y Y N N 1760 10 Y N N Y N N Y N Y Y N N 1152 8 Y -- N Y -- N Y N N Y N N 1517 10 Y N N Y N N Y N Y Y N N 1760 10 Y N N Y N N Y N Y Y N N 780 8 Y -- N Y -- N Y -- N Y -- N 780 8 Y -- N Y -- N Y -- N Y -- N 1152 8 Y -- N -- -- -- Y -- N Y -- N 780 8 Y -- N Y -- N Y -- N Y -- N 1152 8 Y -- N -- -- -- Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N 780 8 Y -- N Y -- N Y -- N Y -- N 1152 8 Y -- N -- -- -- Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N Pin September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices OCT Calibration 6-33 Table 6-10. OCT Calibration Block Counts and Placement in Stratix IV Devices (1A through 4C) (Part 2 of 2) Bank Number of OCT Blocks 1A 1B 1C 2A 2B 2C 3A 3B 3C 4A 4B 4C 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 Y -- N -- -- -- Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N 1760 8 Y -- N Y -- N Y N N Y N N 1932 10 Y N N Y -- N Y N Y Y N N 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 Y -- N -- -- -- Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N 1760 8 Y -- N Y -- N Y N N Y N N 1932 10 Y N N Y -- N Y N Y Y N N 1152 8 Y -- N -- -- -- Y N Y Y N N 1517 10 Y -- N Y -- N Y N Y Y N N 1760 10 Y -- N Y -- N Y N Y Y N N 1932 10 Y -- N Y N N Y N Y Y N N EP4S40G2 1517 8 Y -- N Y -- N Y N N Y N N EP4S40G5 1517 10 Y -- N Y -- N Y N Y Y N N EP4S100G2 1517 8 Y -- N Y -- N Y N N Y N N EP4S100G3 1932 10 Y -- N Y N N Y N Y Y N N EP4S100G4 1932 10 Y -- N Y N N Y N Y Y N N 1517 10 Y -- N Y -- N Y N Y Y N N 1932 10 Y -- N Y N N Y N Y Y N N Device EP4SGX290 EP4SGX360 EP4SGX530 EP4S100G5 Pin Table 6-11 lists the OCT calibration blocks in Banks 5A through 8C. Table 6-11. OCT Calibration Block Counts and Placement in Stratix IV Devices (5A through 8C) (Part 1 of 2) Device EP4SE230 EP4SE360 EP4SE530 EP4SE820 EP4SGX70 September 2012 Bank Number of OCT Blocks 5A 5B 5C 6A 6B 6C 7A 7B 7C 8A 8B 8C 780 8 Y -- N Y -- N Y -- N Y -- N 780 8 Y -- N Y -- N Y -- N Y -- N 1152 8 Y -- N Y -- N Y N N Y N N 1152 8 Y -- N Y -- N Y N N Y N N 1517 10 Y N N Y N N Y N N Y N Y 1760 10 Y N N Y N N Y N N Y N Y 1152 8 Y -- N Y -- N Y N N Y N N 1517 10 Y N N Y N N Y N N Y N Y 1760 10 Y N N Y N N Y N N Y N Y 780 8 -- -- -- -- -- -- Y -- N Y -- N Pin Altera Corporation Stratix IV Device Handbook Volume 1 6-34 Chapter 6: I/O Features in Stratix IV Devices OCT Calibration Table 6-11. OCT Calibration Block Counts and Placement in Stratix IV Devices (5A through 8C) (Part 2 of 2) Device EP4SGX110 Pin 780 Bank Number of OCT Blocks 5A 5B 5C 6A 6B 6C 7A 7B 7C 8A 8B 8C 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 -- -- -- Y -- N Y -- N Y -- N 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 -- -- -- Y -- N Y N N Y Y N 1517 8 Y -- N Y -- N Y N N Y N N 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 -- -- -- Y -- N Y N N Y Y N 1517 8 Y -- N Y -- N Y N N Y N N 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 -- -- -- Y -- N Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N 1760 8 Y -- N Y -- N Y N N Y N N 1932 10 Y -- N Y N N Y N N Y N Y 780 8 -- -- -- -- -- -- Y -- N Y -- N 1152 8 -- -- -- Y -- N Y N N Y N N 1517 8 Y -- N Y -- N Y N N Y N N 1760 8 Y -- N Y -- N Y N N Y N N 1932 10 Y -- N Y N N Y N N Y N Y 1152 8 -- -- -- Y -- N Y N N Y N Y 1517 10 Y -- N Y -- N Y N N Y N Y 1760 10 Y -- N Y -- N Y N N Y N Y 1932 10 Y N N Y -- N Y N N Y N Y EP4S40G2 1517 8 Y -- N Y -- N Y N N Y N N EP4S40G5 1517 10 Y -- N Y -- N Y N N Y N Y EP4S100G2 1517 8 Y -- N Y -- N Y N N Y N N EP4S100G3 1932 10 Y N N Y -- N Y N N Y N Y EP4S100G4 1932 10 Y N N Y -- N Y N N Y N Y 1517 10 Y -- N Y -- N Y N N Y N Y 1932 10 Y N N Y -- N Y N N Y N Y EP4SGX180 EP4SGX230 EP4SGX290 EP4SGX360 EP4SGX530 EP4S100G5 Sharing an OCT Calibration Block on Multiple I/O Banks An OCT calibration block has the same VCCIO as the I/O bank that contains the block. OCT RS calibration is supported on all I/O banks with different VCCIO voltage standards, up to the number of available OCT calibration blocks. You can configure the I/O banks to receive calibration codes from any OCT calibration block with the same VCCIO. All I/O banks with the same VCCIO can share one OCT calibration block, even if that particular I/O bank has an OCT calibration block. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices OCT Calibration 6-35 For example, Figure 6-23 shows a group of I/O banks that has the same VCCIO voltage. If a group of I/O banks has the same VCCIO voltage, you can use one OCT calibration block to calibrate the group of I/O banks placed around the periphery. Because 3B, 4C, 6C, and 7B have the same VCCIO as bank 7A, you can calibrate all four I/O banks (3B, 4C, 6C, and 7B) with the OCT calibration block (CB7) located in bank 7A. You can enable this by serially shifting out OCT RS calibration codes from the OCT calibration block located in bank 7A to the I/O banks located around the periphery. 1 I/O banks that do not contain calibration blocks share calibration blocks with I/O banks that do contain calibration blocks. Figure 6-23 is a top view of the silicon die that corresponds to a reverse view for flip chip packages. It is a graphical representation only. This figure does not show transceiver banks and transceiver calibration blocks. Bank 7A Bank 7B Bank 7C Bank 8C Bank 8B Bank 8A CB 7 Figure 6-23. Example of Calibrating Multiple I/O Banks with One Shared OCT Calibration Block Bank 1A Bank 6A Bank 1B Bank 6B Bank 1C Bank 6C I/O bank with the same VCCIO Bank 2C Bank 5C I/O bank with different VCCIO Bank 2B Bank 5B Bank 2A Bank 5A Bank 4A Bank 4B Bank 4C Bank 3C Bank 3B Bank 3A Stratix IV OCT Calibration Block Modes of Operation Stratix IV devices support OCT RS and OCT RT on all I/O banks. The calibration can occur in either power-up or user mode. Power-Up Mode In power-up mode, OCT calibration is automatically performed at power up. Calibration codes are shifted to selected I/O buffers before transitioning to user mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-36 Chapter 6: I/O Features in Stratix IV Devices OCT Calibration User Mode In user mode, the OCTUSRCLK, ENAOCT, nCLRUSR, and ENASER[9..0] signals are used to calibrate and serially transfer calibration codes from each OCT calibration block to any I/O. Table 6-12 lists the user-controlled calibration block signal names and their descriptions. Table 6-12. OCT Calibration Block Ports for User Control Signal Name Description OCTUSRCLK Clock for OCT block. ENAOCT Enable OCT Termination (Generated by user IP). When ENOCT = 0, each signal enables the OCT serializer for the corresponding OCT calibration block. ENASER[9..0] When ENAOCT = 1, each signal enables OCT calibration for the corresponding OCT calibration block. S2PENA_<bank#> Serial-to-parallel load enable per I/O bank. nCLRUSR Clear user. Figure 6-24 shows the flow of the user signal. When ENAOCT is 1, all OCT calibration blocks are in calibration mode; when ENAOCT is 0, all OCT calibration blocks are in serial data transfer mode. The OCTUSRCLK clock frequency must be 20 MHz or less. 1 You must generate all user signals on the rising edge of OCTUSRCLK. Figure 6-24 does not show transceiver banks and transceiver calibration blocks. CB9 Bank 1A CB7 CB8 CB0 CB6 ENAOCT, nCLRUSR, Bank 1B Bank 1C S2PENA_1C Stratix IV Core Bank 2C Bank 6C S2PENA_6C Bank 5C OCTUSRCLK, ENASER[N] Bank 5B CB1 CB5 CB3 Bank 4A Bank 4B Bank 4C Bank 3C Bank 3B Bank 3A Bank 5A CB4 CB2 Stratix IV Device Handbook Volume 1 Bank 6A Bank 6B S2PENA_4C Bank 2B Bank 2A Bank 7A Bank 7B Bank 7C Bank 8C Bank 8B Bank 8A Figure 6-24. Signals Used for User Mode Calibration September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices OCT Calibration 6-37 OCT Calibration Figure 6-25 shows user mode signal-timing waveforms. To calibrate OCT block[N] (where N is a calibration block number), you must assert ENAOCT one cycle before asserting ENASER[N]. Also, nCLRUSR must be set to low for one OCTUSRCLK cycle before the ENASER[N] signal is asserted. Assert the ENASER[N] signals for 1000 OCTUSRCLK cycles to perform OCTRS and OCTRT calibration. You can de-assert ENAOCT one clock cycle after the last ENASER is de-asserted. Serial Data Transfer After you complete calibration, you must serially shift out the 28-bit OCT calibration codes (14-bit OCT RS and 14-bit OCT RT) from each OCT calibration block to the corresponding I/O buffers. Only one OCT calibration block can send out the codes at any time by asserting only one ENASER[N] signal at a time. After you de-assert ENAOCT, wait at least one OCTUSRCLK cycle to enable any ENASER[N] signal to begin serial transfer. To shift the 28-bit code from the OCT calibration block[N], you must assert ENASER[N] for exactly 28 OCTUSRCLK cycles. Between two consecutive asserted ENASER signals, there must be at least one OCTUSRCLK cycle gap. (Figure 6-25). Figure 6-25. OCT User Mode Signal--Timing Waveform for One OCT Block OCTUSRCLK ENAOCT Calibration Phase nCLRUSR ENASER0 1000 OCTUSRCLK Cycles 28 OCTUSRCLK Cycles ts2p (1) S2PENA_1A Note to Figure 6-25: (1) ts2p 25 ns. After calibrated codes are shifted in serially to each I/O bank, the calibrated codes must be converted from serial to parallel format before being used in the I/O buffers. Figure 6-25 shows the S2PENA signals that can be asserted at any time to update the calibration codes in each I/O bank. All I/O banks that received the codes from the same OCT calibration block can have S2PENA asserted at the same time, or at a different time, even while another OCT calibration block is calibrating and serially shifting codes. The S2PENA signal is asserted one OCTUSRCLK cycle after ENASER is de-asserted for at least 25 ns. You cannot use I/Os for transmitting or receiving data when their S2PENA is asserted for parallel codes transfer. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-38 Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards Example of Using Multiple OCT Calibration Blocks Figure 6-26 shows a signal timing waveform for two OCT calibration blocks doing RS and RT calibration. Calibration blocks can start calibrating at different times by asserting the ENASER signals at different times. ENAOCT must remain asserted while any calibration is ongoing. You must set nCLRUSR low for one OCTUSRCLK cycle before each ENASER[N] signal is asserted. In Figure 6-26, when you set nCLRUSR to 0 for the second time to initialize OCT calibration block 0, this does not affect OCT calibration block 1, whose calibration is already in progress. Figure 6-26. OCT User-Mode Signal Timing Waveform for Two OCT Blocks OCTUSRCLK Calibration Phase ENAOCT nCLRUSR 1000 OCTUSRCLK 28 OCTUSRCLK CY CLE S CY CLE S ENASER0 ENASER1 1000 OCTUSRCLK 28 OCTUSRCLK CY CLE S CY CLE S ts2p (1) S2PENA_1A (2) ts2p (1) S2PENA_2A (3) Notes to Figure 6-26: (1) ts2p 25 ns. (2) S2PENA_1A is asserted in Bank 1A for calibration block 0. (3) S2PENA_2A is asserted in Bank 2A for calibration block 1. RS Calibration If only RS calibration is used for an OCT calibration block, its corresponding ENASER signal only requires to be asserted for 240 OCTUSRCLK cycles. 1 You must assert the ENASER signal for 28 OCTUSRCLK cycles for serial transfer. Termination Schemes for I/O Standards The following sections describe the different termination schemes for the I/O standards used in Stratix IV devices. Single-Ended I/O Standards Termination Voltage-referenced I/O standards require both an input reference voltage, VREF, and a termination voltage, VTT. The reference voltage of the receiving device tracks the termination voltage of the transmitting device. Figure 6-27 and Figure 6-28 show the details of SSTL and HSTL I/O termination on Stratix IV devices. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards 1 6-39 In Stratix IV devices, you cannot use series and parallel OCT simultaneously. For more information, refer to "Dynamic On-Chip Termination" on page 6-29. Figure 6-27. SSTL I/O Standard Termination Termination SSTL Class I SSTL Class II External On-Board Termination 25 50 25 50 50 50 VREF Receiver Transmitter OCT Transmit VTT 50 Receiver Transmitter Stratix IV Series OCT 25 VTT VTT 50 50 50 50 8 VREF VREF Transmitter Receiver VCCIO 25 OCT Receive Receiver Transmitter Stratix IV Parallel OCT VTT 100 25 50 VREF Transmitter Receiver VCCIO Series OCT 50 100 Series OCT 25 100 100 Receiver VCCIO 100 50 100 100 50 8 Transmitter VCCIO VCCIO Stratix IV Parallel OCT VCCIO 50 VREF 100 OCT in BiDirectional Pins 50 VREF Stratix IV Series OCT 50 VTT VTT VTT 100 50 100 100 100 Series OCT 50 Stratix IV September 2012 Altera Corporation Stratix IV Series OCT 25 Stratix IV Stratix IV Stratix IV Device Handbook Volume 1 6-40 Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards Figure 6-28. HSTL I/O Standard Termination Termination HSTL Class II HSTL Class I VTT VTT VTT 50 50 50 External On-Board Termination 50 50 VREF VREF Transmitter Receiver VTT Stratix IV Series OCT 50 Receiver VTT Stratix IV Series OCT 25 50 50 VREF Receiver Transmitter VCCIO 100 50 VREF OCT Receive VTT 100 Stratix IV Stratix IV Device Handbook Volume 1 100 Series OCT 25 Stratix IV Parallel OCT 100 Transmitter Receiver VCCIO VCCIO 100 50 100 VCCIO 50 100 VCCIO 100 Receiver Stratix IV Parallel OCT Receiver VCCIO Transmitter 50 VREF Transmitter Series OCT 50 VTT 50 50 50 VREF OCT Transmit OCT in BiDirectional Pins Transmitter 100 50 8 100 Series OCT 50 Stratix IV 100 Stratix IV 100 Series OCT 25 Stratix IV September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards 6-41 Differential I/O Standards Termination Stratix IV devices support differential SSTL-18 and SSTL-2, differential HSTL-18, HSTL-15, HSTL-12, LVDS, LVPECL, RSDS, and mini-LVDS. Figure 6-29 through Figure 6-35 show the details of various differential I/O terminations on these devices. 1 Differential HSTL and SSTL outputs are not true differential outputs. They use two single-ended outputs with the second output programmed as inverted. Figure 6-29. Differential SSTL I/O Standard Termination Termination Differential SSTL Class II Differential SSTL Class I VTT VTT 50 External On-Board Termination 25 25 VTT VTT 25 50 Receiver Differential SSTL Class I Z0= 50 VTT VCCIO 50 100 Z0= 50 100 VTT VCCIO GND 100 50 Z0= 50 100 GND Altera Corporation Receiver Transmitter Series OCT 25 VCCIO Z0= 50 September 2012 50 Differential SSTL Class II Series OCT 50 Transmitter 50 50 50 25 50 Transmitter OCT 50 50 50 VTT VTT Receiver 100 100 VCCIO GND 100 100 GND Transmitter Receiver Stratix IV Device Handbook Volume 1 6-42 Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards Figure 6-30. Differential HSTL I/O Standard Termination Termination Differential HSTL Class II Differential HSTL Class I VTT VTT 50 External On-Board Termination 50 50 50 50 50 Receiver 50 Receiver Transmitter Differential HSTL Class II Differential HSTL Class I Series OCT 50 Series OCT 25 VCCIO Z0= 50 OCT Z0= 50 VTT VCCIO 50 100 Z0= 50 100 VCCIO GND 100 VTT 50 Z0= 50 100 Receiver 100 100 VCCIO GND 100 100 GND GND Stratix IV Device Handbook Volume 1 50 50 50 Transmitter Transmitter VTT VTT VTT VTT Transmitter Receiver September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards 6-43 LVDS The LVDS I/O standard is a differential high-speed, low-voltage swing, low-power, general-purpose I/O interface standard. In Stratix IV devices, the LVDS I/O standard requires a 2.5-V VCCIO level. The LVDS input buffer requires 2.5-V VCCPD. Use this standard in applications requiring high-bandwidth data transfer, such as backplane drivers and clock distribution. LVDS requires a 100- termination resistor between the two signals at the input buffer. Stratix IV devices provide an optional 100- differential termination resistor in the device using on-chip differential termination. Figure 6-31 shows LVDS termination. The on-chip differential resistor is only available in the row I/O banks. Figure 6-31. LVDS I/O Standard Termination (1) Termination LVDS Differential Outputs Differential Inputs External On-Board Termination 50 100 50 Differential Inputs Differential Outputs 50 OCT Receive (True LVDS Output) (2) 100 50 Stratix IV OCT OCT Receive (Single-Ended LVDS Output with One-Resistor Network, LVDS_E_1R) (3) Differential Inputs Single-Ended Outputs 1 inch 50 100 Rp 50 External Resistor Stratix IV OCT Single-Ended Outputs OCT Receive (Single-Ended LVDS Output with Three-Resistor Network, LVDS_E_3R) (3) Differential Inputs 1 inch 50 Rs 100 Rp Rs 50 External Resistor Stratix IV OCT Notes to Figure 6-31: (1) For LVDS output with a three-resistor network, the RS and RP values are 120 and 170 , respectively. For LVDS output with a one-resistor network, the RP value is 120 . (2) Side I/O banks support true LVDS output buffers. (3) Column and side I/O banks support LVDS_E_1R and LVDS_E_3R I/O standards using two single-ended output buffers. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-44 Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards Differential LVPECL In Stratix IV devices, the LVPECL I/O standard is supported on input clock pins on column and row I/O banks. LVPECL output operation is not supported in Stratix IV devices. LVDS input buffers are used to support LVPECL input operation. AC coupling is required when the LVPECL common-mode voltage of the output buffer is higher than the LVPECL input common-mode voltage. Figure 6-32 shows the AC-coupled termination scheme. The 50- resistors used at the receiver end are external to the device. Figure 6-32. LVPECL AC-Coupled Termination (1) Altera FPGA LVPECL Output Buffer 0.1 F 0.1 F Stratix IV LVPECL Input Buffer ZO = 50 50 VICM 50 ZO = 50 Note to Figure 6-32: (1) The LVPECL AC-coupled termination is applicable only when you use an Altera FPGA LVPECL transmitter. DC-coupled LVPECL is supported if the LVPECL output common mode voltage is within the Stratix IV LVPECL input buffer specification (Figure 6-33). Figure 6-33. LVPECL DC-Coupled Termination (1) Altera FPGA LVPECL Output Buffer Stratix IV LVPECL Input Buffer ZO = 50 ZO = 50 100 Note to Figure 6-33: (1) The LVPECL DC-coupled termination is applicable only when you use an Altera FPGA LVPECL transmitter. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Termination Schemes for I/O Standards 6-45 RSDS Stratix IV devices support the RSDS output standard with data rates up to 230 Mbps using LVDS output buffer types. For transmitters, use two single-ended output buffers with the external one- or three-resistor networks in the column I/O bank, as shown in Figure 6-34. The one-resistor topology is for data rates up to 200 Mbps. The three-resistor topology is for data rates above 200 Mbps. The row I/O banks support RSDS output using true LVDS output buffers without an external resistor network. Figure 6-34. RSDS I/O Standard Termination (1) One-Resistor Network (RSDS_E_1R) Termination Three-Resistor Network (RSDS_E_3R) 1 inch External On-Board Termination RP 1 inch 50 50 RS 100 RP 50 50 100 RS Receiver Transmitter Stratix IV OCT 1 inch RP OCT Transmitter 50 50 Transmitter Receiver 1 inch RS RP 100 RS Receiver Transmitter Stratix IV OCT 50 50 100 Receiver Note to Figure 6-34: (1) The RS and RP values are pending characterization. A resistor network is required to attenuate the LVDS output-voltage swing to meet RSDS specifications. You can modify the three-resistor network values to reduce power or improve noise margin. The resistor values chosen must satisfy Equation 6-1. Equation 6-1. R R s ------p2 -------------------- = 50 R R s + ------p2 1 Altera recommends performing additional simulations using IBIS models to validate that custom resistor values meet the RSDS requirements. f For more information about the RSDS I/O standard, refer to the RSDS Specification from the National Semiconductor website at www.national.com. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-46 Chapter 6: I/O Features in Stratix IV Devices Design Considerations Mini-LVDS Stratix IV devices support the mini-LVDS output standard with data rates up to 340 Mbps using LVDS output buffer types. For transmitters, use two single-ended output buffers with external one- or three-resistor networks, as shown in Figure 6-35. The one-resistor topology is for data rates up to 200 Mbps. The three-resistor topology is for data rates above 200 Mbps. The row I/O banks support mini-LVDS output using true LVDS output buffers without an external resistor network. Figure 6-35. Mini-LVDS I/O Standard Termination (1) One-Resistor Network (mini-LVDS_E_1R) Termination Three-Resistor Network (mini-LVDS_E_3R) 1 inch External On-Board Termination R P 50 50 1 inch RS 100 Receiver R Receiver Stratix IV OCT 1 inch RS 50 R P 100 RS OCT Transmitter Receiver 100 Transmitter Stratix IV OCT 50 P 50 RS Transmitter 1 inch 50 R P Transmitter 50 50 100 Receiver Note to Figure 6-35: (1) The RS and RP values are pending characterization. A resistor network is required to attenuate the LVDS output voltage swing to meet the mini-LVDS specifications. You can modify the three-resistor network values to reduce power or improve noise margin. The resistor values chosen must satisfy Equation 6-1 on page 6-45. 1 Altera recommends that you perform additional simulations using IBIS models to validate that custom resistor values meet the RSDS requirements. f For more information about the mini-LVDS I/O standard, see the mini-LVDS Specification from the Texas Instruments website at www.ti.com. Design Considerations Although Stratix IV devices feature various I/O capabilities for high-performance and high-speed system designs, there are several other design considerations that require your attention to ensure the success of your designs. I/O Bank Restrictions Each I/O bank can simultaneously support multiple I/O standards. The following sections provide guidelines for mixing non-voltage-referenced and voltage-referenced I/O standards in Stratix IV devices. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Design Considerations 6-47 Non-Voltage-Referenced Standards Each I/O bank of a Stratix IV device has its own VCCIO pins and supports only one VCCIO, either 1.2, 1.5, 1.8, 2.5, or 3.0 V. An I/O bank can simultaneously support any number of input signals with different I/O standard assignments if it meets the VCCIO and VCCPD requirement, as shown in Table 6-2 on page 6-3. For output signals, a single I/O bank supports non-voltage-referenced output signals that are driving at the same voltage as VCCIO. Because an I/O bank can only have one VCCIO value, it can only drive out that one value for non-voltage-referenced signals. For example, an I/O bank with a 2.5-V VCCIO setting can support 2.5-V standard inputs and outputs as well as 3.0-V LVCMOS inputs (but not output or bidirectional pins). Voltage-Referenced Standards To accommodate voltage-referenced I/O standards, each Stratix IV device's I/O bank supports multiple VREF pins feeding a common VREF bus. The number of available VREF pins increases as device density increases. If these pins are not used as VREF pins, they cannot be used as generic I/O pins and must be tied to VCCIO or GND. Each bank can only have a single VCCIO voltage level and a single VREF voltage level at a given time. An I/O bank featuring single-ended or differential standards can support voltage-referenced standards if all voltage-referenced standards use the same VREF setting. For performance reasons, voltage-referenced input standards use their own VCCPD level as the power source. This feature allows you to place voltage-referenced input signals in an I/O bank with a VCCIO of 2.5 V or below. For example, you can place HSTL-15 input pins in an I/O bank with 2.5-V VCCIO. However, the voltage-referenced input with parallel OCT enabled requires the VCCIO of the I/O bank to match the voltage of the input standard. Voltage-referenced bidirectional and output signals must be the same as the I/O bank's VCCIO voltage. For example, you can only place SSTL-2 output pins in an I/O bank with a 2.5-V VCCIO. Mixing Voltage-Referenced and Non-Voltage-Referenced Standards An I/O bank can support both voltage-referenced and non-voltage-referenced pins by applying each of the rule sets individually. For example, an I/O bank can support SSTL-18 inputs and 1.8-V inputs and outputs with a 1.8-V VCCIO and a 0.9-V VREF. Similarly, an I/O bank can support 1.5-V standards, 1.8-V inputs (but not outputs), and HSTL and HSTL-15 I/O standards with a 1.5-V VCCIO and 0.75-V VREF. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-48 Chapter 6: I/O Features in Stratix IV Devices Design Considerations Document Revision History Table 6-13 lists the revision history for this chapter. Table 6-13. Document Revision History (Part 1 of 2) Date Version September 2012 3.4 December 2011 3.3 February 2011 March 2010 November 2009 June 2009 April 2009 March 2009 Stratix IV Device Handbook Volume 1 3.2 3.1 3.0 2.3 2.2 Changes Updated the "Programmable Slew Rate Control" section to close FB #68385. Updated Figure 6-17 to close FB #57979. Updated Figure 6-2 and Figure 6-17. Updated the "Modular I/O Banks", "On-Chip Termination Support and I/O Termination Schemes", "Dynamic On-Chip Termination", and "Programmable Pull-Up Resistor" sections. Updated Figure 6-17, Figure 6-32, and Figure 6-33. Applied new template. Minor text edits. Updated Table 6-2 and Table 6-5. Updated Figure 6-18, Figure 6-19, Figure 6-27, Figure 6-28, and Figure 6-31. Added the "Summary of OCT Assignments" section. Added a note to the "Sharing an OCT Calibration Block on Multiple I/O Banks" section. Updated the "OCT Calibration" section. Minor text edits. Updated Table 6-2, Table 6-4, Table 6-6, Table 6-9, and Table 6-10. Updated Figure 6-1, Figure 6-2, Figure 6-4, Figure 6-5, Figure 6-6, Figure 6-8, Figure 6-9, Figure 6-10, Figure 6-11, Figure 6-12, Figure 6-13, and Figure 6-31. Added Table 6-8. Added Figure 6-7, Figure 6-14, Figure 6-15, and Figure 6-16. Added "Left-Shift Series Termination Control" and "Expanded On-Chip Series Termination with Calibration" sections. Updated "MultiVolt I/O Interface", "RSDS", "Mini-LVDS", and "Non-Voltage-Referenced Standards" sections. Deleted Figure 6-5: Number of I/Os in Each Bank in EP4SE290 and EP4SE360 in the 1517-Pin FineLine BGA Package. Minor text edits. Added introductory sentences to improve search ability. Removed the Conclusion section. Updated Figure 6-2. Updated Table 6-8 and Table 6-9. Deleted Figure 6-14. Updated Table 6-1, Table 6-2,Table 6-3, Table 6-4, Table 6-6, Table 6-8, and Table 6-9. Updated Figure 6-2, Figure 6-7, Figure 6-8, Figure 6-9, Figure 6-10, Figure 6-11, and Figure 6-12. Added Figure 6-14. Removed Equation 6-2 and "Referenced Documents" section. 2.1 September 2012 Altera Corporation Chapter 6: I/O Features in Stratix IV Devices Design Considerations 6-49 Table 6-13. Document Revision History (Part 2 of 2) Date Version Changes Updated "Modular I/O Banks" on page 6-7. November 2008 2.0 Updated Figure 6-3 and Figure 6-21. Made minor editorial changes. May 2008 1.0 Initial release. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 6-50 Stratix IV Device Handbook Volume 1 Chapter 6: I/O Features in Stratix IV Devices Design Considerations September 2012 Altera Corporation 7. External Memory Interfaces in Stratix IV Devices February 2011 SIV51007-3.2 SIV51007-3.2 This chapter describes external memory interfaces available with the Stratix IV device family and that family's silicon capability to support external memory interfaces. To support the level of system bandwidth achievable with Altera Stratix IV FPGAs, the devices provide an efficient architecture to quickly and easily fit wide external memory interfaces within their small modular I/O bank structure. The I/Os are designed to provide high-performance support for existing and emerging external double data rate (DDR) memory standards, such as DDR3, DDR2, DDR SDRAM, QDR II+, QDR II SRAM, and RLDRAM II. Stratix IV I/O elements provide easy-to-use built-in functionality required for a rapid and robust implementation with features such as dynamic calibrated on-chip termination (OCT), trace mismatch compensation, read- and write-leveling circuit for DDR3 SDRAM interfaces, half data rate (HDR) blocks, and 4- to 36-bit programmable DQ group widths. The high-performance memory interface solution is backed-up by a self-calibrating megafunction (ALTMEMPHY), optimized to take advantage of the Stratix IV I/O structure and the TimeQuest Timing Analyzer, which completes the picture by providing the total solution for the highest reliable frequency of operation across process, voltage, and temperature (PVT) variations. This chapter contains the following sections: "Memory Interfaces Pin Support" on page 7-3 "Stratix IV External Memory Interface Features" on page 7-29 f For more information about external memory system performance specifications, board design guidelines, timing analysis, simulation, and debugging information, refer to the External Memory Interface Handbook. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 7-2 Chapter 7: External Memory Interfaces in Stratix IV Devices Figure 7-1 shows an overview of the memory interface data path that uses all the Stratix IV I/O element (IOE) features. (1), (2) Figure 7-1. External Memory Interface Data Path Overview Memory Stratix IV FPGA Postamble Enable Postamble Clock 4n DPRAM (2) DLL DQS Logic Block Postamble Control Circuit DQS Enable Circuit 2n 2n Alignment & Synchronization Registers Half Data Rate Input Registers DQS (Read) (3) DDR Input Registers n DQ (Read) (3) Resynchronization Clock n 2n 4n Half-Rate Resynchronization Clock Clock Management & Reset DQ Write Clock Half-Rate Clock 2n Alignment Registers Half Data Rate Output Registers 2 4 Half Data Rate Output Registers 2 Alignment Registers DQ (Write) (3) DDR Output and Output Enable Registers DQS (Write) (3) DDR Output and Output Enable Registers Alignment Clock DQS Write Clock Notes to Figure 7-1: (1) You can bypass each register block. (2) The blocks used for each memory interface may differ slightly. The shaded blocks are part of the Stratix IV IOE. (3) These signals may be bidirectional or unidirectional, depending on the memory standard. When bidirectional, the signal is active during both read and write operations. Memory interfaces use Stratix IV device features such as delay-locked loops (DLLs), dynamic OCT control, read- and write-leveling circuitry, and I/O features such as OCT, programmable input delay chains, programmable output delay, slew rate adjustment, and programmable drive strength. f For more information about I/O features, refer to the I/O Features in Stratix IV Devices chapter. The ALTMEMPHY megafunction instantiates a phase-locked loop (PLL) and PLL reconfiguration logic to adjust the phase shift based on VT variation. vs f For more information about the Stratix IV PLL, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. For more information about the ALTMEMPHY megafunction, refer to the External Memory PHY Interface (ALTMEMPHY) (nonAFI) Megafunction User Guide. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-3 Memory Interfaces Pin Support A typical memory interface requires data (D, Q, or DQ), data strobe (DQS/CQ and DQSn/CQn), address, command, and clock pins. Some memory interfaces use data mask (DM, BWSn, or NWSn) pins to enable write masking and QVLD pins to indicate that the read data is ready to be captured. This section describes how Stratix IV devices support all these different pins. 1 If you have more than one clock pair, you must place them in the same DQ group. For example, if you have two clock pairs, you must place both of them in the same x4 DQS group. f For more information about pin connections, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. f For more information about pin planning and pin connections between a Stratix IV device and an external memory device, refer to the External Memory Interface Handbook. DDR3, DDR2, DDR SDRAM, and RLDRAM II devices use the CK and CK# signals to capture the address and command signals. Generate these signals to mimic the write-data strobe using Stratix IV DDR I/O registers (DDIOs) to ensure that the timing relationships between the CK/CK# and DQS signals (tDQSS, tDSS, and tDSH in DDR3, DDR2, and DDR SDRAM devices or tCKDK in RLDRAM II devices) are met. QDR II+ and QDR II SRAM devices use the same clock (K/K#) to capture write data, address, and command signals. Memory clock pins in Stratix IV devices are generated using a DDIO register going to differential output pins (refer to Figure 7-2), marked in the pin table with DIFFOUT, DIFFIO_TX, or DIFFIO_RX prefixes. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-4 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support f For more information about which pins to use for memory clock pins, refer to the External Memory Interface Handbook. Figure 7-2. Memory Clock Generation FPGA LEs I/O Elements VCC D Q 1 D Q mem_clk (2) 0 mem_clk_n (2) System Clock (3) Notes to Figure 7-2: (1) For pin location requirements,refer to the External Memory Interface Handbook. (2) The mem_clk[0] and mem_clk_n[0] pins for DDR3, DDR2, and DDR SDRAM interfaces use the I/O input buffer for feedback required by the ALTMEMPHY megafunction for tracking; therefore, use bidirectional I/O buffers for these pins. For memory interfaces using a differential DQS input, the input feedback buffer is configured as differential input. For memory interfaces using a single-ended DQS input, the input buffer is configured as a single-ended input. Using a single-ended input feedback buffer requires that I/O standard's VREF voltage is provided to that I/O bank's VREF pins. (3) To minimize jitter, regional clock networks are required for memory output clock generation. Stratix IV devices offer differential input buffers for differential read-data strobe and clock operations. In addition, Stratix IV devices also provide an independent DQS logic block for each CQn pin for complementary read-data strobe and clock operations. In the Stratix IV pin tables, the differential DQS pin pairs are denoted as DQS and DQSn pins, while the complementary CQ signals are denoted as CQ and CQn pins. DQSn and CQn pins are marked separately in the pin table. Each CQn pin connects to a DQS logic block and the shifted CQn signals go to the negative-edge input registers in the DQ IOE registers. 1 Use differential DQS signaling for DDR2 SDRAM interfaces running at or above 333 MHz. DQ pins can be bidirectional signals, as in DDR3, DDR2, and DDR SDRAM, and RLDRAM II common I/O (CIO) interfaces, or unidirectional signals, as in QDR II+, QDR II SRAM, and RLDRAM II separate I/O (SIO) devices. Connect the unidirectional read-data signals to Stratix IV DQ pins and the unidirectional write-data signals to a different DQS/DQ group than the read DQS/DQ group. Furthermore, the write clocks must be assigned to the DQS/DQSn pins associated to this write DQS/DQ group. Do not use the CQ/CQn pin-pair for write clocks. 1 Stratix IV Device Handbook Volume 1 Using a DQS/DQ group for the write-data signals minimizes output skew, allows access to the write-leveling circuitry (for DDR3 SDRAM interfaces), and allows vertical migration. These pins also have access to deskewing circuitry (using programmable delay chains) that can compensate for delay mismatch between signals on the bus. February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-5 The DQS and DQ pin locations are fixed in the pin table. Memory interface circuitry is available in every Stratix IV I/O bank that does not support transceivers. All the memory interface pins support the I/O standards required to support DDR3, DDR2, DDR SDRAM, QDR II+, QDR II SRAM, and RLDRAM II devices. The Stratix IV device family supports DQS and DQ signals with DQ bus modes of x4, x8/x9, x16/x18, or x32/x36, although not all devices support DQS bus mode x32/x36. When any of these pins are not used for memory interfacing, you can use them as user I/Os. In addition, you can use any DQSn or CQn pins not used for clocking as DQ (data) pins. Table 7-1 lists pin support per DQS/DQ bus mode, including the DQS/CQ and DQSn/CQn pin pair. Table 7-1. Stratix IV DQS/DQ Bus Mode Pins Mode x4 x8/x9 (3) DQSn Support CQn Support Yes No Parity or DM (Optional) No Typical Number of Data Pins per Group Maximum Number of Data Pins per Group (2) No 4 5 QVLD (Optional) (6) (1) Yes Yes Yes Yes 8 or 9 11 x16/x18 (4) Yes Yes Yes Yes 16 or 18 23 x32/x36 (5) Yes Yes Yes Yes 32 or 36 47 x32/x36 (7) Yes 32 or 36 39 Yes Yes No (8) Notes to Table 7-1: (1) The QVLD pin is not used in the ALTMEMPHY megafunction. (2) This represents the maximum number of DQ pins (including parity, data mask, and QVLD pins) connected to the DQS bus network with single-ended DQS signaling. When you use differential or complementary DQS signaling, the maximum number of data per group decreases by one. This number may vary per DQS/DQ group in a particular device. Check the pin table for the exact number per group. For DDR3, DDR2, and DDR interfaces, the number of pins is further reduced for an interface larger than x8 due to the need of one DQS pin for each x8/x9 group that is used to form the x16/x18 and x32/x36 groups. (3) Two x4 DQS/DQ groups are stitched to make a x8/x9 group so there are a total of 12 pins in this group. (4) Four x4 DQS/DQ groups are stitched to make a x16/x18 group. (5) Eight x4 DQS/DQ groups are stitched to make a x32/x36 group. (6) The DM pin can be supported if differential DQS is not used and the group does not have additional signals. (7) These x32/x36 DQS/DQ groups are available in EP4SGX290, EP4SGX360, and EP4SGX530 devices in 1152- and 1517-pin FineLine BGA packages. There are 40 pins in each of these DQS/DQ groups. (8) There are 40 pins in each of these DQS/DQ groups. The BWSn pins cannot be placed within the same DQS/DQ group as the write data pins because of insufficient pins available. Table 7-2 lists the number of DQS/DQ groups available per side in each Stratix IV device. For a more detailed listing of the number of DQS/DQ groups available per bank in each Stratix IV device, see Figure 7-3 through Figure 7-19. These figures represent the die-top view of the Stratix IV device. Table 7-2. Number of DQS/DQ Groups in Stratix IV Devices per Side (Part 1 of 3) Device Package EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 780-pin FineLine BGA EP4SGX290 EP4SGX360 780-pin FineLine BGA February 2011 Altera Corporation Side x4 (2) (1) x8/x9 x16/x18 x32/x36 Left 14 6 2 0 Top/Bottom 17 8 2 0 Right 0 0 0 0 Left/Right 0 0 0 0 Top/Bottom 18 8 2 0 (3) Refer to: Figure 7-3 Figure 7-5 Stratix IV Device Handbook Volume 1 7-6 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Table 7-2. Number of DQS/DQ Groups in Stratix IV Devices per Side (Part 2 of 3) Device Package EP4SE230 EP4SE360 780-pin FineLine BGA EP4SGX110 1152-pin FineLine BGA (with 16 transceivers) EP4SGX70 EP4SGX110 1152-pin FineLine BGA (with 24 transceivers) EP4SGX180 EP4SGX230 1152-pin FineLine BGA EP4SGX290 EP4SGX360 EP4SGX530 1152-pin FineLine BGA EP4SE360 EP4SE530 EP4SE820 Side x4 (2) (1) x8/x9 x16/x18 x32/x36 Left/Right 14 6 2 0 Top/Bottom 17 8 2 0 Right/Left 7 3 1 0 Top/Bottom 17 8 2 0 Right/Left 14 6 2 0 Top/Bottom 17 8 2 0 (3) Refer to: Figure 7-4 Figure 7-6 Figure 7-7 Right/Left 13 6 2 0 Top/Bottom 26 12 4 0 Right/Left 13 6 2 0 Top/Bottom 26 12 4 1152-pin FineLine BGA All sides 26 12 4 0 Figure 7-10 EP4SGX180 EP4SGX230 1517-pin FineLine BGA All sides 26 12 4 0 Figure 7-11 EP4SGX290 EP4SGX360 EP4SGX530 1517-pin FineLine BGA Right/Left 26 12 4 0 Top/Bottom 26 12 4 EP4SE530 EP4SE820 1517-pin FineLine BGA Right/Left 34 16 6 0 Top/Bottom 38 18 8 4 EP4S40G2 EP4S40G5 EP4S100G2 EP4S100G5 Left 12 3 1 0 1517-pin FineLine BGA Top/Bottom 26 12 4 0 Right 11 4 1 0 EP4SGX290 EP4SGX360 EP4SGX530 1760-pin FineLine BGA Right/Left 26 12 4 0 Top/Bottom 38 18 8 4 EP4SE530 1760-pin FineLine BGA Right/Left 34 16 6 0 Top/Bottom 38 18 8 4 EP4SE820 1760-pin FineLine BGA Right/Left 40 18 6 0 Top/Bottom 44 22 10 4 EP4SGX290 EP4SGX360 EP4SGX530 1932-pin FineLine BGA Right/Left 29 13 4 0 Top/Bottom 38 18 8 4 Stratix IV Device Handbook Volume 1 2 2 (4) (4) Figure 7-8 Figure 7-9 Figure 7-12 Figure 7-13 Figure 7-14 Figure 7-15 Figure 7-16 Figure 7-17 Figure 7-18 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-7 Table 7-2. Number of DQS/DQ Groups in Stratix IV Devices per Side (Part 3 of 3) Device Package EP4S100G3 EP4S100G4 EP4S100G5 1932-pin FineLine BGA Side x4 (2) (1) x8/x9 x16/x18 x32/x36 Left 8 2 0 0 Top/Bottom 38 18 8 4 Right 7 1 0 0 (3) Refer to: Figure 7-19 Notes to Table 7-2: (1) These numbers are preliminary until the devices are available. (2) Some of the x4 groups may use RUP and RDN pins. You cannot use these groups if you use the Stratix IV calibrated OCT feature. (3) To interface with a x36 QDR II+/QDR II SRAM device in a Stratix IV FPGA that does not support the x32/x36 DQS/DQ group, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (4) These x32/x36 DQS/DQ groups have 40 pins instead of 48 pins per group. BWSn pins cannot be placed within the same DQS/DQ group as the write data pins because of insufficient pins available. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-8 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-3. Number of DQS/DQ Groups per Bank in EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 Devices in the 780-Pin FineLine BGA Package (1), (2), (3), (4). (5) DLL0 I/O Bank 8A I/O Bank 8C I/O Bank 7C I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A 32 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 1C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 Devices in the 780-Pin FineLine BGA I/O Bank 2C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 2A 32 User I/Os x4=4 x8/x9=2 x16/x18=1 DLL1 I/O Bank 3A I/O Bank 3C I/O Bank 4C I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-3: (1) These numbers are preliminary until the devices are available. (2) EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. (5) All I/O pin counts include dedicated clock inputs that you can use for data inputs. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-9 Figure 7-4. Number of DQS/DQ Groups per Bank in EP4SE230 and EP4SE360 Devices in the 780-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A I/O Bank 8C I/O Bank 7C I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 32 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 1C I/O Bank 6C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 26 User I/Os x4=3 x8/x9=1 x16/x18=0 EP4SE230 and EP4SE360 Devices in the 780-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 26 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 2A I/O Bank 5A 32 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=4 x8/x9=2 x16/x18=1 DLL1 I/O Bank 3A I/O Bank 3C I/O Bank 4C I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-4: (1) These numbers are preliminary until the devices are available. (2) EP4SE230 and EP4SE360 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. (5) All I/O pin counts include dedicated clock inputs that you can use for data inputs. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-10 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-5. Number of DQS/DQ Groups per Bank in EP4SGX290 and EP4SGX360 Devices in the 780-Pin FineLine BGA Package (1), (2) DLL0 I/O Bank 8A I/O Bank 8C I/O Bank 7C I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 EP4SGX290 and EP4SGX360 Devices in the 780-Pin FineLine BGA DLL1 I/O Bank 3A I/O Bank 3C I/O Bank 4C I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-5: (1) These numbers are preliminary until the devices are available. (2) EP4SGX290 and EP4SGX360 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-11 Figure 7-6. Number of DQS/DQ Groups per Bank in EP4SGX110 Devices with 16 Transceivers in the 1152-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A I/O Bank 8C I/O Bank 7C I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 32 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=4 x8/x9=2 x16/x18=1 EP4SGX110 Devices in the 1152-Pin FineLine BGA (with 16 Transceivers) I/O Bank 1C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 DLL1 I/O Bank 6C 26 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 3A I/O Bank 3C I/O Bank 4C I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-6: (1) These numbers are preliminary until the devices are available. (2) EP4SGX110 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. (5) All I/O pin counts include dedicated clock inputs that you can use for data inputs. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-12 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-7. Number of DQS/DQ Groups per Bank in EP4SGX70 and EP4SGX110 Devices with 24 Transceivers in the 1152-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A (3) I/O Bank 8C I/O Bank 7C I/O Bank 7A (3) 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=2 x8/x9=1 x16/x18=0 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 1A (3) DLL3 I/O Bank 6A (3) 32 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 1C (4) 26 User I/Os (5) x4=3 x8/x9=1 x16/x18=0 I/O Bank 6C 26 User I/Os (5) x4=3 x8/x9=1 x16/x18=0 EP4SGX70 and EP4SGX110 Devices in the 1152-Pin FineLine BGA (with 24 Transceivers) I/O Bank 6A (3) 32 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 1C (4) 26 User I/Os (5) x4=3 x8/x9=1 x16/x18=0 I/O Bank 6C 26 User I/Os (5) x4=3 x8/x9=1 x16/x18=0 I/O Bank 3A (3) DLL1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 3C 24 User I/Os x4=2 x8/x9=1 x16/x18=0 I/O Bank 4C I/O Bank 4A (3) 24 User I/Os x4=3 x8/x9=1 x16/x18=0 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-7: (1) These numbers are preliminary until the devices are available. (2) EP4SGX70 and EP4SGX110 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-13 Figure 7-8. Number of DQS/DQ Groups per Bank in EP4SGX180 and EP4SGX230 Devices in the 1152-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A I/O Bank 8B I/O Bank 8C I/O Bank 7C I/O Bank 7B I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 EP4SGX180 and EP4SGX230 Devices in the 1152-Pin FineLine BGA I/O Bank 1C I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-8: (1) These numbers are preliminary until the devices are available. (2) EP4SGX180 and EP4SGX230 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-14 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-9. Number of DQS/DQ Groups per Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1152-Pin FineLine BGA Package (1), (3), (4), (5) DLL0 I/O Bank 8A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) I/O Bank 8B I/O Bank 8C I/O Bank 7C I/O Bank 7B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) DLL3 I/O Bank 1A I/O Bank 6A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1152-Pin FineLine BGA I/O Bank 1C I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL1 I/O Bank 3A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) DLL2 Notes to Figure 7-9: (1) These numbers are preliminary until the devices are available. (2) These x32/x36 DQS/DQ groups have 40 pins instead of 48 pins per group. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group; however, there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-15 Figure 7-10. Number of DQS/DQ Groups per Bank in EP4SE360, EP4SE530, and EP4SE820 Devices in the 1152-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A I/O Bank 8B I/O Bank 8C I/O Bank 7C I/O Bank 7B I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 I/O Bank 1C I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 EP4SE360, EP4SE530 and EP4SE820 Devices in the 1152-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 2A I/O Bank 5A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 DLL1 I/O Bank 3A I/O Bank 3B 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 3C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 4C I/O Bank 4B I/O Bank 4A 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-10: (1) These numbers are preliminary until the devices are available. (2) EP4SE360, EP4SE530, and EP4SE820 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. (5) All I/O pin counts include dedicated clock inputs that you can use for data inputs. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-16 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-11. Number of DQS/DQ Groups per Bank in EP4SGX180 and EP4SGX230 Devices in the 1517-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A I/O Bank 8B I/O Bank 8C I/O Bank 7C I/O Bank 7B I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 I/O Bank 1C I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 EP4SGX180 and EP4SGX230 Devices in the 1517-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 2A I/O Bank 5A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 DLL1 I/O Bank 3A I/O Bank 3B 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 3C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 4C I/O Bank 4B I/O Bank 4A 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-11: (1) These numbers are preliminary until the devices are available. (2) EP4SGX180 and EP4SGX230 devices do not support x32/x36 mode. To interface with a x36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-17 Figure 7-12. Number of DQS/DQ Groups per Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1517-Pin FineLine BGA Package (1), (3), (4), (5) DLL0 I/O Bank 8A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) I/O Bank 8B I/O Bank 8C I/O Bank 7C I/O Bank 7B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16//x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) DLL3 I/O Bank 1A I/O Bank 6A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 I/O Bank 1C I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1517-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 42 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 2A I/O Bank 5A 48 User I/Os x4=7 x8/x9=3 x16/x18=1 48 User I/Os x4=7 x8/x9=3 x6/x18=1 DLL1 I/O Bank 3A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) I/O Bank 3B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 3C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 4C I/O Bank 4B 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=1 (2) DLL2 Notes to Figure 7-12: (1) These numbers are preliminary until the devices are available. (2) These x32/x36 DQS/DQ groups have 40 pins instead of 48 pins per group. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-18 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-13. Number of DQS/DQ Groups per Bank in EP4SE530 and EP4SE820 Devices in the 1517-pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 1A DLL3 I/O Bank 6A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 1B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 6B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 1C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 EP4SE530 and EP4SE820 Devices in the 1517-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2B I/O Bank 5B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 2A I/O Bank 5A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-13: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-19 Figure 7-14. Number of DQS/DQ Groups per Bank in EP4S40G2, EP4S40G5, EP4S100G2, and EP4S100G5 Devices in the 1517-Pin FineLine BGA Package (1), (2), (3), (4), (5) DLL0 I/O Bank 8A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 I/O Bank 8B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 I/O Bank 7B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 I/O Bank 7A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL3 I/O Bank 1A I/O Bank 6A 43 User I/Os x4=5 x8/x9=1 x16/x18=0 44 User I/Os x4=5 x8/x9=1 x16/x18=0 I/O Bank 1C 20 User I/Os x4=0 x8/x9=0 x16/x18=0 I/O Bank 6C 21 User I/Os x4=0 x8/x9=0 x16/x18=0 EP4S40G2, EP4S40G5, EP4S100G2, and EP4S100G5 Devices in the 1517-Pin FineLine BGA I/O Bank 2C 21 User I/Os x4=1 x8/x9=0 x16/x18=0 I/O Bank 5C 21 User I/Os x4=0 x8/x9=0 x16/x18=0 I/O Bank 2A I/O Bank 5A 46 User I/Os x4=6 x8/x9=2 x16/x18=1 46 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 40 User I/Os x4=6 x8/x9=3 x16/x18=1 24 User I/Os x4=4 x8/x9=2 x16/x18=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 40 User I/Os x4=6 x8/x9=3 x16/x18=1 DLL2 Notes to Figure 7-14: (1) These numbers are preliminary until the devices are available. (2) EP4S40G2, EP4S40G5, EP4S100G2, and EP4S100G5 devices do not support 32/36 mode. To interface with a 36 QDR II+/QDR II SRAM device, refer to "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" on page 7-26. (3) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (4) All I/O pin counts include dedicated clock inputs that you can use for data inputs. (5) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a 4 DQS/DQ group with any of its pin members used for configuration purposes. Make sure that the DQS/DQ groups that you have chosen are not used for configuration as you may lose up to four 4 DQS/DQ groups, depending on your configuration scheme. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-20 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-15. Number of DQS/DQ Groups per Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1760-Pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16//x18=0 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL3 I/O Bank 1A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6A 50 User I/Os x4=7 x8/x9=3 x6/x18=1 x32/x36=0 I/O Bank 1C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1760-Pin FineLine BGA I/O Bank 2C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 5A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 DLL1 I/O Bank 3A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 3B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 3C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 4C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 4B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-15: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-21 Figure 7-16. Number of DQS/DQ Groups per Bank in EP4SE530 Devices in the 1760-Pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL3 I/O Bank 1A I/O Bank 6A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 1B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 6B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 1C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 EP4SE530 Devices in the 1760-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2B I/O Bank 5B 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 24 User I/Os x4=4 x8/x9=2 x16/x18=1 x32/x36=0 I/O Bank 2A I/O Bank 5A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-16: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-22 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-17. Number of DQS/DQ Groups per Bank in EP4SE820 Devices in the 1760-pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 48 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 7C 48 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL3 I/O Bank 1A I/O Bank 6A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6B 36 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 1B 36 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 1C 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6C 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 EP4SE820 Devices in the 1760-Pin FineLine BGA I/O Bank 2C I/O Bank 5C 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2B I/O Bank 5B 36 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 36 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2A I/O Bank 5A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 48 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-17: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-23 Figure 7-18. Number of DQS/DQ Groups per Bank in EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1932-Pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 1A DLL3 I/O Bank 6A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 6C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 1C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 I/O Bank 2C I/O Bank 5C 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 42 User I/Os x4=6 x8/x9=3 x16/x18=1 x32/x36=0 EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1932-Pin FineLine BGA I/O Bank 2B I/O Bank 5B 20 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 20 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 2A I/O Bank 5A 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 50 User I/Os x4=7 x8/x9=3 x16/x18=1 x32/x36=0 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-18: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-24 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Figure 7-19. Number of DQS/DQ Groups per Bank in EP4S100G3, EP4S100G4, and EP4S100G5 Devices in the 1932-Pin FineLine BGA Package (1), (2), (3), (4) DLL0 I/O Bank 8A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 8C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7C 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 7B 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 7A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 I/O Bank 1A DLL3 I/O Bank 6A 38 User I/Os x4=3 x8/x9=0 x16/x18=0 x32/x36=0 40 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 I/O Bank 6C 20 User I/Os x4=0 x8/x9=0 x16/x18=0 x32/x36=0 I/O Bank 1C 19 User I/Os x4=0 x8/x9=0 x16/x18=0 x32/x36=0 I/O Bank 2C I/O Bank 5C 19 User I/Os x4=0 x8/x9=0 x16/x18=0 x32/x36=0 17 User I/Os x4=0 x8/x9=0 x16/x18=0 x32/x36=0 EP4S100G3, EP4S100G4, and EP4S100G5 Devices in the 1932-Pin FineLine BGA I/O Bank 2B I/O Bank 5B 13 User I/Os x4=1 x8/x9=0 x16/x18=0 x32/x36=0 12 User I/Os x4=0 x8/x9=0 x16/x18=0 x32/x36=0 I/O Bank 2A I/O Bank 5A 39 User I/Os x4=4 x8/x9=1 x16/x18=0 x32/x36=0 40 User I/Os x4=4 x8/x9=1 x16/x18=0 x32/x36=0 DLL1 I/O Bank 3A I/O Bank 3B I/O Bank 3C I/O Bank 4C I/O Bank 4B I/O Bank 4A 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 32 User I/Os x4=3 x8/x9=1 x16/x18=0 x32/x36=0 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 48 User I/Os x4=8 x8/x9=4 x16/x18=2 x32/x36=1 DLL2 Notes to Figure 7-19: (1) These numbers are preliminary until the devices are available. (2) You can also use DQS/DQSn pins in some of the x4 groups as RUP and RDN pins, but you cannot use a x4 group for memory interfaces if two pins of the x4 group are used as RUP and RDN pins for OCT calibration. If two pins of a x4 group are used as RUP and RDN pins for OCT calibration, you can use the x16/x18 or x32/x36 groups that include that x4 group, however there are restrictions on using x8/x9 groups that include that x4 group. (3) All I/O pin counts include dedicated clock inputs and dedicated corner PLL clock inputs that you can use for data inputs. (4) You can also use some of the DQS/DQ pins in I/O Bank 1C as configuration pins. You cannot use a x4 DQS/DQ group with any of its pin members used for configuration purposes. Ensure that the DQS/DQ groups that you have chosen are not also used for configuration because you may lose up to four x4 DQS/DQ groups, depending on your configuration scheme. The DQS and DQSn pins are listed in the Stratix IV pin tables as DQSXY and DQSnXY, respectively, where X indicates the DQS/DQ grouping number and Y indicates whether the group is located on the top (T), bottom (B), left (L), or right (R) side of the device. The DQS/DQ pin numbering is based on x4 mode. The corresponding DQ pins are marked as DQXY, where X indicates which DQS group the pins belong to and Y indicates whether the group is located on the top (T), bottom (B), left (L), or right (R) side of the device. For example, DQS1L indicates a DQS pin located on the left side of the device. The DQ pins belonging to that group are shown as DQ1L in the pin table. For more information, refer to Figure 7-20. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 1 7-25 The parity, DM, BWSn, NWSn, ECC, and QVLD pins are shown as DQ pins in the pin table. The numbering scheme starts from the top-left corner of the device going counter-clockwise in a die-top view. Figure 7-20 shows how the DQS/DQ groups are numbered in a die-top view of the device. The top and bottom sides of the device can contain up to 38 x4 DQS/DQ groups. The left and right sides of the device can contain up to 34 x4 DQS/DQ groups. Figure 7-20. DQS Pins in Stratix IV I/O Banks DQS20T DQS38T DQS19T DQS1T DLL0 DLL3 PLL_T1 PLL_T2 PLL_R1 PLL_L1 8A 8B 8C 7C 7B 7A DQS1L DQS34R 1A 6A 1B 6B 1C 6C DQS17L DQS18R PLL_R2 PLL_L2 Stratix IV Device PLL_R3 PLL_L3 DQS18L DQS17R 2C 5C 2B 5B 2A 5A DQS34L DQS1R 3A 3B 3C 4C 4B 4A PLL_R4 PLL_L4 PLL_B1 PLL_B2 DLL2 DLL1 DQS1B February 2011 Altera Corporation DQS19B DQS20B DQS38B Stratix IV Device Handbook Volume 1 7-26 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Using the RUP and RDN Pins in a DQS/DQ Group Used for Memory Interfaces You can use the DQS/DQSn pins in some of the x4 groups as RUP and RDN pins (listed in the pin table). You cannot use a x4 DQS/DQ group for memory interfaces if any of its pin members are used as RUP and RDN pins for OCT calibration. You may be able to use the x8/x9 group that includes this x4 DQS/DQ group, if either of the following applies: You are not using DM pins with your differential DQS pins You are not using complementary or differential DQS pins You can use the x8/x9 group because a DQS/DQ x8/x9 group actually comprises 12 pins, as the groups are formed by stitching two DQS/DQ groups in x4 mode with six pins each (refer to Table 7-1 on page 7-5). A typical x8 memory interface consists of one DQS, one DM, and eight DQ pins that add up to 10 pins. If you choose your pin assignment carefully, you can use the two extra pins for RUP and RDN. In a DDR3 SDRAM interface, you must use differential DQS, which means that you only have one extra pin. In this case, pick different pin locations for the RUP and RDN pins (for example, in the bank that contains the address and command pins). You cannot use the RUP and RDN pins shared with DQS/DQ group pins when using x9 QDR II+/QDR II SRAM devices, as the RUP and RDN pins are dual purpose with the CQn pins. In this case, pick different pin locations for RUP and RDN pins to avoid conflict with memory interface pin placement. In this case, you have the choice of placing the RUP and RDN pins in the data-write group or in the same bank as the address and command pins. There is no restriction on using x16/x18 or x32/x36 DQS/DQ groups that include the x4 groups whose pins are being used as RUP and RDN pins, because there are enough extra pins that can be used as DQS pins. 1 For x8, x16/x18, or x32/x36 DQS/DQ groups whose members are used for RUP and RDN, you must assign DQS and DQ pins manually. The Quartus(R) II software might not be able to place DQS and DQ pins without manual pin assignments, resulting in a "no-fit". Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface This implementation combines x16/x18 DQS/DQ groups to interface with a x36 QDR II+/QDR II SRAM device. The x36 read data bus uses two x16/x18 groups while the x36 write data uses another two x16/x18 or four x8/x9 groups. The CQ/CQn signal traces are split on the board trace to connect to two pairs of CQ/CQn pins in the FPGA. This is the only connection on the board that you need to change for this implementation. Other QDR II+/QDR II SRAM interface rules for Stratix IV devices also apply for this implementation. 1 The ALTMEMPHY megafunction and UniPHY-based external memory interface IPs do not use the QVLD signal, so you can leave the QVLD signal unconnected as in any QDR II+/QDR II SRAM interface. f For more information about the ALTMEMPHY megafunction or UniPHY-based IPs, refer to the External Memory Interface Handbook. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support 7-27 Rules to Combine Groups In 780-, 1152-, and some 1517-pin package devices, there is at most one x16/x18 group per I/O sub-bank. You can combine two x16/x18 groups from a single side of the device for a x36 interface. For devices that do not have four x16/x18 groups in a single side of the device to form two x36 groups for read and write data, you can form one x36 group on one side of the device and another x36 group on the other side of the device. For vertical migration with the x36 emulation implementation, check if migration is possible by enabling device migration in the Quartus II project. The Quartus II software supports the use of four x8/x9 DQ groups for write data pins and migration of these groups across device density. Table 7-3 lists the possible combinations to use two x16/x18 DQS/DQ groups to form a x32/x36 group on Stratix IV devices lacking a native x32/x36 DQS/DQ group. Table 7-3. Possible Group Combinations in Stratix IV Devices (Part 1 of 2) Package 780-Pin FineLine BGA 1152-Pin FineLine BGA February 2011 Altera Corporation Device Density EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 EP4SGX290 EP4SGX360 EP4SE230 EP4SE360 EP4SGX70 EP4SGX110 EP4SGX180 EP4SGX230 EP4SGX290 (2) EP4SGX360 (2) EP4SGX530 (2) EP4SE360 EP4SE530 EP4SE820 I/O Sub-Bank Combinations 3A and 4A, 7A and 8A (bottom and top I/O banks) 1A and 2A, 5A and 6A (left and right I/O banks) 3A and 4A, 7A and 8A (bottom and top I/O banks) 3A and 4A, 7A and 8A (bottom and top I/O banks) (1) (1) (1) 1A and 1C, 6A and 6C (left and right I/O banks) 3A and 3B, 4A and 4B (bottom I/O banks) 7A and 7B, 8A and 8B (top I/O banks) 1A and 1C, 2A and 2C (left I/O banks) 3A and 3B, 4A and 4B (bottom I/O banks) 5A and 5C, 6A and 6C (right I/O banks) 7A and 7B, 8A and 8B (top I/O banks) Stratix IV Device Handbook Volume 1 7-28 Chapter 7: External Memory Interfaces in Stratix IV Devices Memory Interfaces Pin Support Table 7-3. Possible Group Combinations in Stratix IV Devices (Part 2 of 2) Package 1517-Pin FineLine BGA 1760-Pin FineLine BGA 1932-Pin FineLine BGA Device Density EP4SGX180 EP4SGX230 EP4SGX290 (2) EP4SGX360 (2) EP4SGX530 (2) EP4SE530 (2) EP4SE820 (2) EP4S40G2 EP4S40G5 EP4S100G2 EP4S100G5 EP4SGX290 EP4SGX360 EP4SGX530 EP4SE530 (2) EP4SE820 (2) EP4SGX290 (2) EP4SGX360 (2) EP4SGX530 (2) I/O Sub-Bank Combinations 1A and 1C, 2A and 2C (left I/O banks) 3A and 3B, 4A and 4B (bottom I/O banks) 5A and 5C, 6A and 6C (right I/O banks) 7A and 7B, 8A and 8B (top I/O banks) 1A and 1B, 2A and 2B or 1B and 1C, 2B and 2C (left I/O banks) (3) 5A and 5B, 6A and 6B or 5B and 5C, 6B and 6C (right I/O banks) (3) 3A and 3B, 4A and 4B (bottom I/O banks) 7A and 7B, 8A and 8B (top I/O banks) 1A and 1C, 2A and 2C (left I/O banks) 3A and 3B, 4A and 4B (bottom I/O banks) 5A and 5C, 6A and 6C (right I/O banks) 7A and 7B, 8A and 8B (top I/O banks) 1A and 1B, 2A and 2B or 1B and 1C, 2B and 2C (left I/O banks) (3) 5A and 5B, 6A and 6B or 5B and 5C, 6B and 6C (right I/O banks) (3) 1A and 1C, 2A and 2C (left I/O banks) 5A and 5C, 6A and 6C (right I/O banks) Notes to Table 7-3: (1) Each side of the device in these packages has four remaining x8/x9 groups. You can combine them for the write side (only) if you want to keep the x36 QDR II+/QDR II SRAM interface on one side of the device. You must change the Memory Interface Data Group default assignment from the default 18 to 9 in this case. (2) This device supports x36 DQS/DQ groups on the top and bottom I/O banks natively. (3) Although it is possible to combine the x16/x18 DQS/DQ groups from I/O banks 1A and 1C, 2A and 2C, 5A and 5C, and 6A and 6C, Altera does not recommend this due to the size of the package. Similarly, crossing a bank number (for example, combining groups from I/O banks 6C and 5C) is not supported in this package. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-29 Stratix IV External Memory Interface Features Stratix IV devices are rich with features that allow robust high-performance external memory interfacing. The ALTMEMPHY megafunction allows you to use these external memory interface features and helps set up the physical interface (PHY) best suited for your system. This section describes each Stratix IV device feature that is used in external memory interfaces from the DQS phase-shift circuitry, DQS logic block, leveling multiplexers, and dynamic OCT control block. 1 The ALTMEMPHY megafunction and the Altera memory controller MegaCore(R) functions can run at half the frequency of the I/O interface of the memory devices to allow better timing management in high-speed memory interfaces. Stratix IV devices have built-in registers in the IOE to convert data from full-rate (the I/O frequency) to half-rate (the controller frequency) and vice versa. You can bypass these registers if your memory controller is not running at half the rate of the I/O frequency. When using the Altera memory controller MegaCore functions, the ALTMEMPHY megafunction is instantiated for you. f For more information about the ALTMEMPHY megafunction, refer to the External Memory PHY Interface (ALTMEMPHY) (nonAFI) Megafunction User Guide. DQS Phase-Shift Circuitry Stratix IV phase-shift circuitry provides phase shift to the DQS/CQ and CQn pins on read transactions when the DQS/CQ and CQn pins are acting as input clocks or strobes to the FPGA. The DQS phase-shift circuitry consists of DLLs that are shared between multiple DQS pins and the phase-offset module to further fine-tune the DQS phase shift for different sides of the device. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-30 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Figure 7-21 shows how the DQS phase-shift circuitry is connected to the DQS/CQ and CQn pins in the device where memory interfaces are supported on all sides of the Stratix IV device. Figure 7-21. DQS/CQ and CQn Pins and DQS Phase-Shift Circuitry DQS/CQ Pin CQn Pin DLL Reference Clock (1), (2) DQS/CQ Pin CQn Pin DLL Reference Clock DQS Logic Blocks t DQS Phase-Shift Circuitry t to IOE to IOE t t to IOE to IOE DQS Phase-Shift Circuitry DQS Logic Blocks DQS/CQ Pin CQn Pin DQS/CQ Pin CQn Pin t to IOE t to IOE t to IOE t to IOE DQS Phase-Shift Circuitry to IOE to IOE to IOE to IOE t t t t to IOE t CQn Pin to IOE t DQS/CQ Pin to IOE t CQn Pin to IOE t DQS/CQ Pin DQS Phase-Shift Circuitry DLL Reference Clock DLL Reference Clock CQn Pin DQS/CQ Pin CQn Pin DQS/CQ Pin Notes to Figure 7-21: (1) For possible reference input clock pins for each DLL, refer to "DLL" on page 7-31. (2) You can configure each DQS/CQ and CQn pin with a phase shift based on one of two possible DLL output settings. DQS phase-shift circuitry is connected to the DQS logic blocks that control each DQS/CQ or CQn pin. The DQS logic blocks allow the DQS delay settings to be updated concurrently at every DQS/CQ or CQn pin. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-31 DLL DQS phase-shift circuitry uses a DLL to dynamically control the clock delay needed by the DQS/CQ and CQn pin. The DLL, in turn, uses a frequency reference to dynamically generate control signals for the delay chains in each of the DQS/CQ and CQn pins, allowing it to compensate for PVT variations. The DQS delay settings are Gray-coded to reduce jitter when the DLL updates the settings. The phase-shift circuitry needs 1,280 clock cycles to lock and calculate the correct input clock period when the DLL is in low jitter mode. Otherwise, only 256 clock cycles are needed. Do not send data during these clock cycles because there is no guarantee that it will be captured properly. As the settings from the DLL may not be stable until this lock period has elapsed, be aware that anything using these settings (including the leveling delay system) may be unstable during this period. 1 You can still use the DQS phase-shift circuitry for any memory interfaces that are less than 100 MHz. However, the DQS signal may not shift over 2.5 ns. Even if the DQS signal is not shifted exactly to the middle of the DQ valid window, the I/O element should still be able to capture the data in low-frequency applications in which a large amount of timing margin is available. There are a maximum of four DLLs in a Stratix IV device, located in each corner of the device. These four DLLs support a maximum of four unique frequencies, with each DLL running at one frequency. Each DLL can have two outputs with different phase offsets, which allows one Stratix IV device to have eight different DLL phase shift settings. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-32 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Figure 7-22 shows the DLL and I/O bank locations in Stratix IV devices from a die-top view if all sides of the device support external memory interfaces. Figure 7-22. Stratix IV DLL and I/O Bank Locations (Die-Top View) PLL_L1 8A 8B 8C PLL_T1 PLL_T2 7C 7B PLL_R1 7A 6 6 DLL0 DLL3 6 6 1A 6A 1B 6B 1C 6C PLL_R2 PLL_L2 Stratix IV FPGA PLL_L3 PLL_R3 2C 5C 2B 5B 5A 2A 6 6 DLL1 6 DLL2 6 PLL_L4 3A 3B 3C PLL_B1 PLL_B2 4C 4B 4A PLL_R4 The DLL can access the two adjacent sides from its location within the device. For example, DLL0 on the top left of the device can access the top side (I/O banks 7A, 7B, 7C, 8A, 8B, and 8C) and the left side of the device (I/O banks 1A, 1B, 1C, 2A, 2B, and 2C). This means that each I/O bank is accessible by two DLLs, giving more flexibility to create multiple frequencies and multiple-type interfaces. You can have two different interfaces with the same frequency on the two sides adjacent to a DLL, where the DLL controls the DQS delay settings for both interfaces. Each bank can use settings from either or both DLLs the bank is adjacent to. For example, DQS1L can get its phase-shift settings from DLL0, while DQS2L can get its phase-shift settings from DLL1. Table 7-4 lists the DLL location and supported I/O banks for Stratix IV devices. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 1 7-33 You can only have one memory interface in each I/O sub-bank (such as I/O sub-banks 1A, 1B, and 1C) when you use leveling delay chains. This is because there is only one leveling delay chain per I/O sub-bank. Table 7-4. DLL Location and Supported I/O Banks DLL Location Accessible I/O Banks (1) DLL0 Top-left corner 1A, 1B, 1C, 2A, 2B, 2C, 7A, 7B, 7C, 8A, 8B, 8C DLL1 Bottom-left corner 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B, 3C, 4A, 4B, 4C DLL2 Bottom-right corner 3A, 3B, 3C, 4A, 4B, 4C, 5A, 5B, 5C, 6A, 6B, 6C DLL3 Top-right corner 5A, 5B, 5C, 6A, 6B, 6C, 7A, 7B, 7C, 8A, 8B, 8C Note to Table 7-4: (1) The DLL can access these I/O banks if they are available for memory interfacing. The reference clock for each DLL may come from PLL output clocks or any of the two dedicated clock input pins located in either side of the DLL. Table 7-5 through Table 7-17 lists the available DLL reference clock input resources for the Stratix IV device family. 1 When you have a dedicated PLL that only generates the DLL input reference clock, set the PLL mode to No Compensation to achieve better performance or the Quartus II software changes it automatically. Because the PLL does not use any other outputs, it does not need to compensate for any clock paths. Table 7-5. DLL Reference Clock Input for EP4SGX70, EP4SGX110, EP4SGX180, and EP4SGX230 Devices in the 780-Pin FineLine BGA Package DLL DLL0 DLL1 CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P PLL_T1 PLL_L2 -- CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P PLL_B1 -- -- CLK7P CLK3P -- PLL_B1 -- -- -- PLL_T1 -- -- CLK4P DLL2 CLK5P CLK6P CLK7P CLK12P DLL3 CLK13P CLK14P CLK15P February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-34 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Table 7-6. DLL Reference Clock Input for EP4SE230 and EP4SE360 Devices in the 780-Pin FineLine BGA Package DLL DLL0 DLL1 DLL2 DLL3 CLKIN (Top/Bottom) CLKIN (Left/Right) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P CLK7P CLK3P CLK4P CLK8P CLK5P CLK9P CLK6P CLK10P CLK7P CLK11P CLK12P CLK8P CLK13P CLK9P CLK14P CLK10P CLK15P CLK11P PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 -- PLL_B1 PLL_L2 -- PLL_B1 PLL_R2 -- PLL_T1 PLL_R2 -- Table 7-7. DLL Reference Clock Input for EP4SGX290 and EP4SGX360 Devices in the 780-Pin FineLine BGA Package DLL CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) -- PLL_T1 -- -- -- PLL_B1 -- -- -- PLL_B2 -- -- -- PLL_T2 -- -- CLK12P DLL0 CLK13P CLK14P CLK15P CLK4P DLL1 CLK5P CLK6P CLK7P CLK4P DLL2 CLK5P CLK6P CLK7P CLK12P DLL3 CLK13P CLK14P CLK15P Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-35 Table 7-8. DLL Reference Clock Input for EP4SGX70 and EP4SGX110 Devices in the 1152-Pin FineLine BGA Package (with 24 Transceivers) CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) DLL0 CLK12P CLK13P CLK14P CLK15P CLK0P CLK1P CLK2P CLK3P PLL_T1 PLL_L2 -- DLL1 CLK4P CLK5P CLK6P CLK7P CLK0P CLK1P CLK2P CLK3P PLL_B1 PLL_L2 -- DLL2 CLK4P CLK5P CLK6P CLK7P CLK8P CLK9P CLK10P CLK11P PLL_B1 PLL_R2 -- DLL3 CLK12P CLK13P CLK14P CLK15P CLK8P CLK9P CLK10P CLK11P PLL_T1 PLL_R2 -- DLL Table 7-9. DLL Reference Clock Input for EP4SGX110 Devices in the 1152-Pin FineLine BGA Package (with 16 Transceivers) DLL CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 -- PLL_B1 -- -- PLL_B1 -- -- PLL_T1 PLL_R2 -- CLK12P DLL0 CLK13P CLK0P CLK14P CLK1P CLK15P CLK4P DLL1 CLK5P CLK0P CLK6P CLK1P CLK7P CLK4P DLL2 CLK5P CLK10P CLK6P CLK11P CLK7P CLK12P DLL3 CLK13P CLK10P CLK14P CLK11P CLK15P February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-36 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Table 7-10. DLL Reference Clock Input for EP4SGX180, EP4SGX230, EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1152-Pin FineLine BGA Package DLL CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 -- PLL_B1 -- -- PLL_B2 -- -- PLL_T2 PLL_R2 -- CLK12P DLL0 CLK13P CLK0P CLK14P CLK1P CLK15P CLK4P DLL1 CLK5P CLK0P CLK6P CLK1P CLK7P CLK4P DLL2 CLK5P CLK10P CLK6P CLK11P CLK7P CLK12P DLL3 CLK13P CLK10P CLK14P CLK11P CLK15P Table 7-11. DLL Reference Clock Input for EP4SE360, EP4SE530, and EP4SE820 Devices in the 1152-Pin FineLine BGA Packages DLL DLL0 DLL1 DLL2 DLL3 CLKIN (Top/Bottom) CLKIN (Left/Right) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P CLK7P CLK3P CLK4P CLK8P CLK5P CLK9P CLK6P CLK10P CLK7P CLK11P CLK12P CLK8P CLK13P CLK9P CLK14P CLK10P CLK15P CLK11P Stratix IV Device Handbook Volume 1 PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 -- PLL_B1 PLL_L3 -- PLL_B2 PLL_R3 -- PLL_T2 PLL_R2 -- February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-37 Table 7-12. DLL Reference Clock Input for EP4SE530 and EP4SE820 Devices in the 1517- and 1760-Pin FineLine BGA Packages DLL DLL0 DLL1 DLL2 DLL3 CLKIN (Top/Bottom) CLKIN (Left/Right) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P CLK7P CLK3P CLK4P CLK8P CLK5P CLK9P CLK6P CLK10P CLK7P CLK11P CLK12P CLK8P CLK13P CLK9P CLK14P CLK10P CLK15P CLK11P PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 PLL_L1 PLL_B1 PLL_L3 PLL_L4 PLL_B2 PLL_R3 PLL_R4 PLL_T2 PLL_R2 PLL_R1 Table 7-13. DLL Reference Clock Input for EP4SGX180, EP4SGX230, EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1517-Pin FineLine BGA Package DLL DLL0 DLL1 DLL2 DLL3 February 2011 CLKIN (Top/Bottom) CLKIN (Left/Right) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P CLK7P CLK3P CLK4P CLK8P CLK5P CLK9P CLK6P CLK10P CLK7P CLK11P CLK12P CLK8P CLK13P CLK9P CLK14P CLK10P CLK15P CLK11P Altera Corporation PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 -- PLL_B1 PLL_L3 -- PLL_B2 PLL_R3 -- PLL_T2 PLL_R2 -- Stratix IV Device Handbook Volume 1 7-38 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Table 7-14. DLL Reference Clock Input for EP4S40G2, EP4S40G5, EP4S100G2, and EP4S100G5 Devices in the 1517-Pin FineLine BGA Package CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) DLL0 CLK12P CLK13P CLK14P CLK15P CLK1P CLK3P PLL_T1 PLL_L2 -- DLL1 CLK4P CLK5P CLK6P CLK7P CLK1P CLK3P PLL_B1 PLL_L3 -- DLL2 CLK4P CLK5P CLK6P CLK7P CLK8P CLK10P PLL_B2 PLL_R3 -- DLL3 CLK12P CLK13P CLK14P CLK15P CLK8P CLK10P PLL_T2 PLL_R2 -- DLL Table 7-15. DLL Reference Clock Input for EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1760-Pin FineLine BGA Package CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) DLL0 CLK12P CLK13P CLK14P CLK15P CLK0P CLK1P CLK2P CLK3P PLL_T1 PLL_L2 -- DLL1 CLK4P CLK5P CLK6P CLK7P CLK0P CLK1P CLK2P CLK3P PLL_B1 PLL_L3 -- DLL2 CLK4P CLK5P CLK6P CLK7P CLK8P CLK9P CLK10P CLK11P PLL_B2 PLL_R3 -- DLL3 CLK12P CLK13P CLK14P CLK15P CLK8P CLK9P CLK10P CLK11P PLL_T2 PLL_R2 -- DLL Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-39 Table 7-16. DLL Reference Clock Input for EP4SGX290, EP4SGX360, and EP4SGX530 Devices in the 1932-Pin FineLine BGA Package DLL DLL0 DLL1 DLL2 DLL3 CLKIN (Top/Bottom) CLKIN (Left/Right) CLK12P CLK0P CLK13P CLK1P CLK14P CLK2P CLK15P CLK3P CLK4P CLK0P CLK5P CLK1P CLK6P CLK2P CLK7P CLK3P CLK4P CLK8P CLK5P CLK9P CLK6P CLK10P CLK7P CLK11P CLK12P CLK8P CLK13P CLK9P CLK14P CLK10P CLK15P CLK11P PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) PLL_T1 PLL_L2 PLL_L1 PLL_B1 PLL_L3 PLL_L4 PLL_B2 PLL_R3 PLL_R4 PLL_T2 PLL_R2 PLL_R1 Table 7-17. DLL Reference Clock Input for EP4S100G3, EP4S100G4, and EP4S100G5 Devices in the 1932-Pin FineLine BGA Package CLKIN (Top/Bottom) CLKIN (Left/Right) PLL (Top/Bottom) PLL (Left/Right) PLL (Corner) DLL0 CLK12P CLK13P CLK14P CLK15P -- PLL_T1 PLL_L2 PLL_L1 DLL1 CLK4P CLK5P CLK6P CLK7P -- PLL_B1 PLL_L3 PLL_L4 DLL2 CLK4P CLK5P CLK6P CLK7P CLK9P CLK11P PLL_B2 PLL_R3 PLL_R4 DLL3 CLK12P CLK13P CLK14P CLK15P CLK9P CLK11P PLL_T2 PLL_R2 PLL_R1 DLL February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-40 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Figure 7-23 shows a simple block diagram of the DLL. The input reference clock goes into the DLL to a chain of up to 16 delay elements. The phase comparator compares the signal coming out of the end of the delay chain block to the input reference clock. The phase comparator then issues the upndn signal to the Gray-code counter. This signal increments or decrements a six-bit delay setting (DQS delay settings) that increases or decreases the delay through the delay element chain to bring the input reference clock and the signals coming out of the delay element chain in phase. Figure 7-23. Simplified Diagram of the DQS Phase-Shift Circuitry (1) addnsub Phase offset settings from the logic array ( offset [5:0] ) 6 offsetdelayctrlin [5:0] DLL aload Input Reference Clock (2) offsetdelayctrlout [5:0] Phase Comparator upndninclkena 6 Phase Offset Control B offsetdelayctrlout [5:0] offsetdelayctrlin [5:0] 6 delayctrlout [5:0] 6 6 Phase offset settings to DQS pins on top or bottom edge (3) ( offsetctrlout [5:0] ) addnsub Phase offset settings from the logic array ( offset [5:0] ) Up/Down Counter Delay Chains 6 (dll_offset_ctrl_a) upndnin clk Phase Offset Control A 6 (dll_offset_ctrl_b) Phase offset settings to DQS pin on left or right edge (3) ( offsetctrlout [5:0] ) DQS Delay Settings (4) dqsupdate Notes to Figure 7-23: (1) All features of the DQS phase-shift circuitry are accessible from the ALTMEMPHY megafunction in the Quartus II software. (2) The input reference clock for the DQS phase-shift circuitry can come from a PLL output clock or an input clock pin. For more information, refer to Table 7-5 on page 7-33 through Table 7-17 on page 7-39. (3) Phase offset settings can only go to the DQS logic blocks. (4) DQS delay settings can go to the logic array, DQS logic block, and leveling circuitry. 1 In the Quartus II assignment, phase offset control block `A' is designated as DLLOFFSETCTRL_<coordinate x>_<coordinate y>_N1 and phase offset control block `B' is designated as DLLOFFSETCTRL_<coordinate x>_<coordinate y>_N2. You can reset the DLL from either the logic array or a user I/O pin. Each time the DLL is reset, you must wait for 1,280 clock cycles for the DLL to lock before you can capture the data properly. Depending on the DLL frequency mode, the DLL can shift the incoming DQS signals by 0, 22.5, 30, 36, 45, 60, 67.5, 72, 90, 108, 120, 135, 144, 180, or 240. The shifted DQS signal is then used as the clock for the DQ IOE input registers. All DQS/CQ and CQn pins, referenced to the same DLL, can have their input signal phase shifted by a different degree amount but all must be referenced at one particular frequency. For example, you can have a 90 phase shift on DQS1T and a 60 phase shift on DQS2T, referenced from a 200-MHz clock. Not all phase-shift combinations are supported. The phase shifts on the DQS pins referenced by the same DLL must all be a multiple of 22.5 (up to 90), 30 (up to 120), 36 (up to 144), 45 (up to 180), or 60 (up to 240). Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-41 There are eight different frequency modes for the Stratix IV DLL, as listed in Table 7-18. Each frequency mode provides different phase shift selections. In frequency mode 0, 1, 2, and 3, the 6-bit DQS delay settings vary with PVT to implement the phase-shift delay. In frequency modes 4, 5, 6, and 7, only 5 bits of the DQS delay settings vary with PVT to implement the phase-shift delay; the most significant bit of the DQS delay setting is set to 0. Table 7-18. Stratix IV DLL Frequency Modes Frequency Mode Available Phase Shift Number of Delay Chains 0 22.5, 45, 67.5, 90 16 1 30, 60, 90, 120 12 2 36, 72, 108, 144 10 3 45, 90, 135, 180 8 4 30, 60, 90, 120 12 5 36, 72, 108, 144 10 6 45, 90, 135, 180 8 7 60, 120, 180, 240 6 f For the frequency range of each mode, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. For 0 shift, the DQS/CQ signal bypasses both the DLL and DQS logic blocks. The Quartus II software automatically sets the DQ input delay chains so that the skew between the DQ and DQS/CQ pin at the DQ IOE registers is negligible when 0 shift is implemented. You can feed the DQS delay settings to the DQS logic block and logic array. The shifted DQS/CQ signal goes to the DQS bus to clock the IOE input registers of the DQ pins. The signal can also go into the logic array for resynchronization if you are not using IOE resynchronization registers. The shifted CQn signal can only go to the negative-edge input register in the DQ IOE and is only used for QDR II+ and QDR II SRAM interfaces. Phase Offset Control Each DLL has two phase-offset modules and can provide two separate DQS delay settings with independent offsets, one for the top and bottom I/O bank and one for the left and right I/O bank, so you can fine-tune the DQS phase-shift settings between two different sides of the device. Even though you have independent phase offset control, the frequency of the interface using the same DLL must be the same. Use the phase offset control module for making small shifts to the input signal and use the DQS phase-shift circuitry for larger signal shifts. For example, if the DLL only offers a multiple of 30 phase shift, but your interface needs a 67.5 phase shift on the DQS signal, you can use two delay chains in the DQS logic blocks to give you 60 phase shift and use the phase offset control feature to implement the extra 7.5 phase shift. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-42 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features You can use either a static phase offset or a dynamic phase offset to implement the additional phase shift. The available additional phase shift is implemented in 2's: complement in Gray-code between settings -64 to +63 for frequency mode 0, 1, 2, and 3, and between settings -32 to +31 for frequency modes 4, 5, 6, and 7. An additional bit indicates whether the setting has a positive or negative value. The settings are linear, each phase offset setting adds a delay amount specified in the DC and Switching Characteristics for Stratix IV Devices chapter. The DQS phase shift is the sum of the DLL delay settings and the user-selected phase offset settings whose top setting is 64 for frequency modes 0, 1, 2, and 3; and 32 for frequency modes 4, 5, 6, and 7, so the actual physical offset setting range is 64 or 32 subtracted by the DQS delay settings from the DLL. 1 When using this feature, you need to monitor the DQS delay settings to know how many offsets you can add and subtract in the system. Note that the DQS delay settings output by the DLL are also Gray coded. For example, if the DLL determines that DQS delay settings of 28 is needed to achieve a 30 phase shift in DLL frequency mode 1, you can subtract up to 28 phase offset settings and you can add up to 35 phase offset settings to achieve the optimal delay that you need. However, if the same DQS delay settings of 28 is needed to achieve 30 phase shift in DLL frequency mode 4, you can still subtract up to 28 phase offset settings, but you can only add up to 3 phase offset settings before the DQS delay settings reach their maximum settings because DLL frequency mode 4 only uses 5-bit DLL delay settings. f For more information about the value for each step, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. When using static phase offset, you can specify the phase offset amount in the ALTMEMPHY megafunction as a positive number for addition or a negative number for subtraction. You can also have a dynamic phase offset that is always added to, subtracted from, or both added to and subtracted from the DLL phase shift. When you always add or subtract, you can dynamically input the phase offset amount into the dll_offset[5..0] port. When you want to both add and subtract dynamically, you control the addnsub signal in addition to the dll_offset[5..0] signals. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation 7-43 Stratix IV Device Handbook Volume 1 DQS Logic Block Each DQS/CQ and CQn pin is connected to a separate DQS logic block, which consists of the DQS delay chains, update enable circuitry, and DQS postamble circuitry, as shown in Figure 7-24. Figure 7-24. Stratix IV DQS Logic Block DQS Delay Chain DQS Enable dqsenable (2) 1xx 000 dqsbusout 001 010 011 Bypass dqsin DQS bus 6 6 DQS Enable Control 0 1 0 1 6 D Q dqsupdateen Input Reference Clock (1) Update Enable Circuitry phasectrlin 6 <dqs_ctrl_latches_enable> 6 delayctrlin Resynchronization Clock clk 4 phaseinvertctrl 0111 0110 0101 0100 0011 0010 0001 0000 Postamble Enable 0 1 <level_dqs_enable> postamble control clock 0 0 dqsenableout 0 1 1 1 dqsenablein enaphasetransferreg <delay_dqs_enable_by_half_cycle> February 2011 Altera Corporation Notes to Figure 7-24: (1) The input reference clock for the DQS phase-shift circuitry can come from a PLL output clock or an input clock pin. For more information, refer to Table 7-5 on page 7-33 through Table 7-17 on page 7-39. (2) The dqsenable signal can also come from the Stratix IV FPGA fabric. Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 6 offsetctrlin [5:0] 6 Phase offset 1 D Q settings from the 0 DQS phase-shift circuitry <dqs_offsetctrl_enable> 6 DQS delay settings from the delayctrlin [5:0] DQS phase-shift circuitry dqsbusout phasectrlin[2:0] dqsin DQS/CQ or CQn Pin 6 PRE Q D 7-44 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features DQS Delay Chain DQS delay chains consist of a set of variable delay elements to allow the input DQS/CQ and CQn signals to be shifted by the amount specified by the DQS phase-shift circuitry or the logic array. There are four delay elements in the DQS delay chain; the first delay chain closest to the DQS/CQ pin can be shifted either by the DQS delay settings or by the sum of the DQS delay setting and the phase-offset setting. The number of delay chains required is transparent because the ALTMEMPHY megafunction automatically sets it when you choose the operating frequency. The DQS delay settings can come from the DQS phase-shift circuitry on either end of the I/O banks or from the logic array. The delay elements in the DQS logic block have the same characteristics as the delay elements in the DLL. When the DLL is not used to control the DQS delay chains, you can input your own Gray-coded 6-bit or 5-bit settings using the dqs_delayctrlin[5..0] signals available in the ALTMEMPHY megafunction. These settings control 1, 2, 3, or all 4 delay elements in the DQS delay chains. The ALTMEMPHY megafunction can also dynamically choose the number of DQS delay chains needed for the system. The amount of delay is equal to the sum of the delay element's intrinsic delay and the product of the number of delay steps and the value of the delay steps. You can also bypass the DQS delay chain to achieve a 0 phase shift. Update Enable Circuitry Both the DQS delay settings and the phase-offset settings pass through a register before going into the DQS delay chains. The registers are controlled by the update enable circuitry to allow enough time for any changes in the DQS delay setting bits to arrive at all the delay elements. This allows them to be adjusted at the same time. The update enable circuitry enables the registers to allow enough time for the DQS delay settings to travel from the DQS phase-shift circuitry or core logic to all the DQS logic blocks before the next change. It uses the input reference clock or a user clock from the core to generate the update enable output. The ALTMEMPHY megafunction uses this circuit by default. Figure 7-25 shows an example waveform of the update enable circuitry output. Figure 7-25. DQS Update Enable Waveform DLL Counter Update (Every 8 cycles) DLL Counter Update (Every 8 cycles) System Clock DQS Delay Settings (Updated every 8 cycles) 6 bit Update Enable Circuitry Output Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-45 DQS Postamble Circuitry For external memory interfaces that use a bidirectional read strobe such as in DDR3, DDR2, and DDR SDRAM, the DQS signal is low before going to or coming from a high-impedance state. The state in which DQS is low, just after a high-impedance state, is called the preamble; the state in which DQS is low, just before it returns to a high-impedance state, is called the postamble. There are preamble and postamble specifications for both read and write operations in DDR3, DDR2, and DDR SDRAM. The DQS postamble circuitry ensures that data is not lost if there is noise on the DQS line during the end of a read operation that occurs while DQS is in a postamble state. Stratix IV devices have dedicated postamble registers that you can control to ground the shifted DQS signal used to clock the DQ input registers at the end of a read operation. This ensures that any glitches on the DQS input signals during the end of a read operation that occurs while DQS is in a postamble state do not affect the DQ IOE registers. In addition to the dedicated postamble register, Stratix IV devices also have an HDR block inside the postamble enable circuitry. Use these registers if the controller is running at half the frequency of the I/Os. Using the HDR block as the first stage capture register in the postamble enable circuitry block is optional. The HDR block is clocked by the half-rate resynchronization clock, which is the output of the I/O clock divider circuit (shown in Figure 7-31 on page 7-49). There is an AND gate after the postamble register outputs that is used to avoid postamble glitches from a previous read burst on a non-consecutive read burst. This scheme allows a half-a-clock cycle latency for dqsenable assertion and zero latency for dqsenable de-assertion, as shown in Figure 7-26. Figure 7-26. Avoiding Glitch on a Non-Consecutive Read Burst Waveform Postamble glitch Postamble Preamble DQS Postamble Enable dqsenable Delayed by 1/2T logic February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-46 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Leveling Circuitry DDR3 SDRAM unbuffered modules use a fly-by clock distribution topology for better signal integrity. This means that the CK/CK# signals arrive at each DDR3 SDRAM device in the module at different times. The difference in arrival time between the first DDR3 SDRAM device and the last device on the module can be as long as 1.6 ns. Figure 7-27 shows the clock topology in DDR3 SDRAM unbuffered modules. Figure 7-27. DDR3 SDRAM Unbuffered Module Clock Topology DQS/DQ DQS/DQ DQS/DQ DQS/DQ CK/CK# DQS/DQ DQS/DQ DQS/DQ DQS/DQ Stratix IV Device Because the data and read strobe signals are still point-to-point, take special care to ensure that the timing relationship between the CK/CK# and DQS signals (tDQSS, tDSS, and tDSH) during a write is met at every device on the modules. Furthermore, read data coming back into the FPGA from the memory is also staggered in a similar way. Stratix IV FPGAs have leveling circuitry to address these two situations. There is one leveling circuitry per I/O sub-bank (for example, I/O sub-bank 1A, 1B, and 1C each has one leveling circuitry). These delay chains are PVT-compensated by the same DQS delay settings as the DLL and DQS delay chains. For frequencies equal to and above 400 MHz, the DLL uses eight delay chains, such that each delay chain generates a 45 delay. The generated clock phases are distributed to every DQS logic block that is available in the I/O sub-bank. The delay chain taps then feeds a multiplexer controlled by the ALTMEMPHY megafunction to select which clock phases are to be used for that x4 or x 8 DQS group. Each group can use a different tap output from the read-leveling and write-leveling delay chains to compensate for the different CK/CK# delay going into each device on the module. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-47 Figure 7-28 and Figure 7-29 show the Stratix IV write- and read-leveling circuitry. Figure 7-28. Stratix IV Write-Leveling Delay Chains and Multiplexers Write clk (-900) (1) Write-Leveled DQS Clock Write-Leveled DQ Clock Note to Figure 7-28: (1) There is one leveling delay chain per I/O sub-bank (for example, I/O sub-banks 1A, 1B, and 1C). You can only have one memory interface in each I/O sub-bank when you use the leveling delay chain. Figure 7-29. Stratix IV Read-Leveling Delay Chains and Multiplexers (1) I/O Clock Divider (2) use_masterin slaveout masterin DQS delayctrlin 1 0 Half-Rate Resynchronization Clock DFF 1 0 clkout Half-Rate Source Synchronous Clock phaseselect phasectrlin 6 4 phaseinvertctrl Resynchronization Clock (resync_clk_2x) 0111 0110 0101 0100 0011 0010 0001 0000 0 1 Read-Leveled Resynchronization Clock Notes to Figure 7-29: (1) There is one leveling delay chain per I/O sub-bank (for example, I/O sub-banks 1A, 1B, and 1C). You can only have one memory interface in each I/O sub-bank when you use the leveling delay chain. (2) Each divider feeds up to six pins (from a 4 DQS group) in the device. To feed wider DQS groups, you must chain multiple clock dividers together by feeding the slaveout output of one divider to the masterin input of the neighboring pins' divider. The -90 write clock of the ALTMEMPHY megafunction feeds the write-leveling circuitry to produce the clock to generate the DQS and DQ signals. During initialization, the ALTMEMPHY megafunction picks the correct write-leveled clock for the DQS and DQ clocks for each DQS/DQ group after sweeping all the available clocks in the write calibration process. The DQ clock output is -90 phase-shifted compared to the DQS clock output. Similarly, the resynchronization clock feeds the read-leveling circuitry to produce the optimal resynchronization and postamble clock for each DQS/DQ group in the calibration process. The resynchronization and postamble clocks can use different clock outputs from the leveling circuitry. The output from the read-leveling circuitry can also generate the half-rate resynchronization clock that goes to the FPGA fabric. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-48 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 1 The ALTMEMPHY megafunction dynamically calibrates the alignment for read- and write-leveling during the initialization process. f For more information about the ALTMEMPHY megafunction, refer to the External Memory PHY Interface (ALTMEMPHY) (nonAFI) Megafunction User Guide. Dynamic On-Chip Termination Control Figure 7-30 shows the dynamic OCT control block. The block includes all the registers needed to dynamically turn on OCT RT during a read and turn OCT RT off during a write. f For more information about dynamic on-chip termination control, refer to the I/O Features in Stratix IV Devices chapter. Figure 7-30. Stratix IV Dynamic OCT Control Block OCT Control OCT Enable 2 DFF OCT HalfRate Clock HDR Block DFF Resynchronization Registers Write Clock (1) OCT Control Path Note to Figure 7-30: (1) The write clock comes from either the PLL or the write-leveling delay chain. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-49 I/O Element Registers The IOE registers are expanded to allow source-synchronous systems to have faster register-to-register transfers and resynchronization. Both top and bottom and left and right IOEs have the same capability. Left and right IOEs have extra features to support LVDS data transfer. Figure 7-31 shows the registers available in the Stratix IV input path. The input path consists of the DDR input registers, resynchronization registers, and HDR block. You can bypass each block of the input path. Figure 7-31. Stratix IV IOE Input Registers DQ (1) Double Data Rate Input Registers D Q DFF Input Reg AI D DQS/CQ (3), (9) Differential Input Buffer DQSn (9) CQn (4) Q neg_reg_out DFF Input Reg BI 0 1 D Q Half Data Rate Registers DFF Input Reg C I directin Alignment & Synchronization Registers D Q D 0 1 Q datain [0] D Q dataout D DFF DFF 1 Q To Core dataout [0] (7) DFF Q DFF DFF D enaphasetransferreg enainputcycledelay <bypass_output_register>(10) D Q 1 D DFF Q 0 1 DFF dataout D DFF DFF 1 Q DFF D Q DFF I/O Clock Divider (6) dataoutbypass (8) Q DFF Q (2) Resynchronization Clock (resync_clk_2x) (5) D 0 D 0 Q Q DFF datain [1] D To Core dataout[2] (7) 0 D 0 1 D To Core dataout [1] (7) To Core dataout [3] (7) Q DFF Half-Rate Resynchronization Clock (resync_clk_1x) to core (7) Notes to Figure 7-31: (1) You can bypass each register block in this path. (2) This is the 0-phase resynchronization clock (from the read-leveling delay chain). (3) The input clock can be from the DQS logic block (whether the postamble circuitry is bypassed or not) or from a global clock line. (4) This input clock comes from the CQn logic block. (5) This resynchronization clock comes from a PLL through the clock network (resync_ck_2x). (6) The I/O clock divider resides adjacent to the DQS logic block. In addition to the PLL and read-leveled resync clock, the I/O clock divider can also be fed by the DQS bus or CQn bus. (7) The half-rate data and clock signals feed into a dual-port RAM in the FPGA core. (8) You can dynamically change the dataoutbypass signal after configuration to select either the directin input or the output from the half data rate register to feed dataout. (9) The DQS and DQSn signals must be inverted for DDR, DDR2, and DDR3 interfaces. When using Altera's memory interface IPs, the DQS and DQSn signals are automatically inverted. (10) The bypass_output_register option allows you to select either the output from the second mux or the output of the fourth alignment/ synchronization register to feed dataout. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-50 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features There are three registers in the DDR input registers block. Two registers capture data on the positive and negative edges of the clock, while the third register aligns the captured data. You can choose to use the same clock for the positive edge and negative edge registers, or two complementary clocks (DQS/CQ for the positive-edge register and DQSn/CQn for the negative-edge register). The third register that aligns the captured data uses the same clock as the positive edge registers. The resynchronization registers consist of up to three levels of registers to resynchronize the data to the system clock domain. These registers are clocked by the resynchronization clock that is either generated by the PLL or the read-leveling delay chain. The outputs of the resynchronization registers can go straight to the core or to the HDR blocks, which are clocked by the divided-down resynchronization clock. For more information about the read-leveling delay chain, refer to "Leveling Circuitry" on page 7-46. Figure 7-32 shows the registers available in the Stratix IV output and output-enable paths. The path is divided into the HDR block, resynchronization registers, and output and output-enable registers. The device can bypass each block of the output and output-enable path. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation 7-51 Stratix IV Device Handbook Volume 1 Figure 7-32. Stratix IV IOE Output and Output-Enable Path Registers (1) Half Data Rate to Single Data Rate Output-Enable Registers From Core (2) Alignment Registers (4) D Q Double Data Rate Output-Enable Registers DFF DFF From Core (2) 0 1 D Q D DFF D Q D D D Q Q Q DFF Q OE Reg A OE DFF OR2 1 DFF DFF 0 DFF D Half Data Rate to Single Data Rate Output Registers Q Alignment Registers (4) OE Reg B OE From Core (wdata2) (2) D Q Double Data Rate Output Registers DFF DFF 0 D Q D Q D 1 From Core (wdata0) (2) D DFF D Q D Q Q Q TRI DFF Output Reg Ao DFF DFF D Output Reg Bo 0 1 D Q DFF D Q Q DFF Q DFF From Core (wdata1) (2) D Q D D Q Q DFF DFF DFF Half-Rate Clock (3) February 2011 Altera Corporation Alignment Clock (3) Write Clock (5) Notes to Figure 7-32: (1) You can bypass each register block of the output and output-enable paths. (2) Data coming from the FPGA core are at half the frequency of the memory interface clock frequency in half-rate mode. (3) The half-rate clock comes from the PLL, while the alignment clock comes from the write-leveling delay chains. (4) These registers are only used in DDR3 SDRAM interfaces for write-leveling purposes. (5) The write clock can come from either the PLL or from the write-leveling delay chain. The DQ write clock and DQS write clock have a 90 offset between them. Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features D DQ or DQS DFF DFF From Core (wdata3) (2) 1 0 7-52 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features The output path is designed to route combinatorial or registered SDR outputs and full-rate or half-rate DDR outputs from the FPGA core. Half-rate data is converted to full-rate using the HDR block, clocked by the half-rate clock from the PLL. The resynchronization registers are also clocked by the same 0 system clock, except in the DDR3 SDRAM interface. In DDR3 SDRAM interfaces, the leveling registers are clocked by the write-leveling clock. For more information about the write-leveling delay chain, refer to "Leveling Circuitry" on page 7-46. The output-enable path has a structure similar to the output path. You can have a combinatorial or registered output in SDR applications and you can use half-rate or full-rate operation in DDR applications. Also, the ouput-enable path's resynchronization registers have a structure similar to the output path registers, ensuring that the output-enable path goes through the same delay and latency as the output path. Delay Chain Stratix IV devices have run-time adjustable delay chains in the I/O blocks and the DQS logic blocks. You can control the delay chain setting through the I/O or the DQS configuration block output. Figure 7-33 shows the delay chain ports. Figure 7-33. Delay Chain delayctrlin [3..0] <use finedelayctrlin> finedelayctrlin datain t 0 dataout t 1 Every I/O block contains the following: Stratix IV Device Handbook Volume 1 Two delay chains in a series between the output registers and the output buffer One delay chain between the input buffer and the input register Two delay chains between the output enable and the output buffer Two delay chains between the OCT RT enable control register and the output buffer February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-53 Figure 7-34 shows the delay chains in an I/O block. Figure 7-34. Delay Chains in an I/O Block rtena oe octdelaysetting1 (only) D5 OCT Delay Chain D5 OutputEnable Delay Chain octdelaysetting2 (only) D6 OCT Delay Chain D6 OutputEnable Delay Chain (outputdelaysetting1 + outputfinedelaysetting1) (outputdelaysetting2 + outputfinedelaysetting2) D5 Delay Delay Chain D6 Delay Delay Chain 0 1 (outputdelaysetting2 + outputfinedelaysetting2) or (outputonlydelaysetting2 + outputonlyfinedelaysetting2) D1 Delay Delay Chain (padtoinputregisterdelaysetting + padtoinputregisterfinedelaysetting) Each DQS logic block contains a delay chain after the dqsbusout output and another delay chain before the dqsenable input. Figure 7-35 shows the delay chains in the DQS input path. Figure 7-35. Delay Chains in the DQS Input Path (dqsbusoutdelaysetting + dqsbusoutfinedelaysetting) DQS DQS Delay Chain DQS Enable D4 Delay Chain dqsin dqsbusout dqsenable (dqsenabledelaysetting + dqsenablefinedelaysetting) T11 Delay Chain DQS Enable Control February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 7-54 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features I/O Configuration Block and DQS Configuration Block The I/O configuration block and the DQS configuration block are shift registers that you can use to dynamically change the settings of various device configuration bits. The shift registers power-up low. Every I/O pin contains one I/O configuration register, while every DQS pin contains one DQS configuration block in addition to the I/O configuration register. Figure 7-36 shows the I/O configuration block and the DQS configuration block circuitry. Figure 7-36. I/O Configuration Block and DQS Configuration Block bit 0 bit 1 MSB bit 2 datain update ena clk Table 7-19 lists the I/O configuration block bit sequence. Table 7-19. I/O Configuration Block Bit Sequence Bit Bit Name 0..3 outputdelaysetting1[0..3] 4..6 outputdelaysetting2[0..2] 7..10 padtoinputregisterdelaysetting[0..3] Table 7-20 lists the DQS configuration block bit sequence. Table 7-20. DQS Configuration Block Bit Sequence (Part 1 of 2) Stratix IV Device Handbook Volume 1 Bit Bit Name 0..3 dqsbusoutdelaysetting[0..3] 4..6 dqsinputphasesetting[0..2] 7..10 dqsenablectrlphasesetting[0..3] 11..14 dqsoutputphasesetting[0..3] 15..18 dqoutputphasesetting[0..3] 19..22 resyncinputphasesetting[0..3] 23 dividerphasesetting 24 enaoctcycledelaysetting 25 enainputcycledelaysetting 26 enaoutputcycledelaysetting 27..29 dqsenabledelaysetting[0..2] 30..33 octdelaysetting1[0..3] February 2011 Altera Corporation Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features 7-55 Table 7-20. DQS Configuration Block Bit Sequence (Part 2 of 2) Bit Bit Name 34..36 octdelaysetting2[0..2] 37 enadataoutbypass 38 enadqsenablephasetransferreg 39 enaoctphasetransferreg 40 enaoutputphasetransferreg 41 enainputphasetransferreg 42 resyncinputphaseinvert 43 dqsenablectrlphaseinvert 44 dqoutputphaseinvert 45 dqsoutputphaseinvert Document Revision History Table 7-21 lists the revision history for this chapter. Table 7-21. Document Revision History (Part 1 of 2) Date Version February 2011 March 2010 February 2011 3.2 3.1 Altera Corporation Changes Updated Table 7-5, Table 7-6, Table 7-11, Table 7-19, and Table 7-20. Added Table 7-12. Updated Figure 7-36. Removed Table 7-1 and Table 7-6. Applied new template. Minor text edits. Updated Figure 7-8, Figure 7-11, Figure 7-23, Figure 7-24, Figure 7-29, Figure 7-31, and Figure 7-36. Added Figure 7-9 and Figure 7-12. Added Table 7-7. Updated Table 7-1, Table 7-2, Table 7-3, Table 7-4, Table 7-6, Table 7-8 and Table 7-19. Added note to the "Memory Interfaces Pin Support" section. Changed "DLL1 through DLL4" to "DLL0 through DLL3" throughout. Added frequency mode 7 throughout. Minor text edits. Stratix IV Device Handbook Volume 1 7-56 Chapter 7: External Memory Interfaces in Stratix IV Devices Stratix IV External Memory Interface Features Table 7-21. Document Revision History (Part 2 of 2) Date Version November 2009 June 2009 April 2009 March 2009 November 2008 May 2008 Stratix IV Device Handbook Volume 1 3.0 2.3 2.2 2.1 2.0 1.0 Changes Updated the "Memory Interfaces Pin Support" and "Combining x16/x18 DQS/DQ Groups for a x36 QDR II+/QDR II SRAM Interface" sections. Updated Table 7-1, Table 7-2, Table 7-7, and Table 7-12. Updated Figure 7-3, Figure 7-4, Figure 7-5, Figure 7-6, Figure 7-7, Figure 7-8, Figure 7-9, Figure 7-10, Figure 7-11, Figure 7-13, Figure 7-14, Figure 7-15, and Figure 7-16. Added Figure 7-12 and Figure 7-17. Added Table 7-14, Table 7-17, Table 7-19, and Table 7-20. Added "Delay Chain" and "I/O Configuration Block and DQS Configuration Block" sections. Removed Figure 7-8 and Figure 7-12. Removed Table 7-1, Table 7-2, and Table 7-24. Minor text edits. Updated "Overview" and "Leveling Circuitry". Updated Figure 7-26 and Figure 7-27. Updated Table 7-3. Added introductory sentences to improve search ability. Removed the Conclusion section. Updated Table 7-5, Table 7-6, Table 7-15, and Table 7-17 Removed Figure 7-12, Figure 7-13, and Figure 7-20 Updated Table 7-1, Table 7-5, Table 7-8, Table 7-12, Table 7-13, Table 7-14, Table 7-15, and Table 7-17. Replaced Table 7-6. Added Table 7-11 and Table 7-16. Updated Figure 7-3, Figure 7-6, Figure 7-8, Figure 7-9, and Figure 7-11. Added Figure 7-7, Figure 7-11, Figure 7-12, Figure 7-13, and Figure 7-20. Updated "Combining x16/x18 DQS/DQ Groups for x36 QDR II+/QDR II SRAM Interface". Updated "Rules to Combine Groups". Removed "Referenced Documents" section. Updated Table 7-1, Table 7-2, Table 7-3, Table 7-4, Table 7-5, and Table 7-6. Added Table 7-7. Updated Figure 7-1 and Figure 7-19. Updated "Combining x16/x18 DQS/DQ groups for x36 QDR II+/QDR II SRAM Interface" on page 7-26. Updated "Rules to Combine Groups" on page 7-27. Updated "DQS Phase-Shift Circuitry" on page 7-29. Updated Table 7-9, Table 7-10, Table 7-11, Table 7-13, Table 7-13, Table 7-14, Table 7-15, Table 7-15, Table 7-16, and Table 7-18. Updated Figure 7-30 and Figure 7-31. Made minor editorial changes. Initial release. February 2011 Altera Corporation 8. High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices September 2012 SIV51008-3.4 SIV51008-3.4 This chapter describes the significant advantages of the high-speed differential I/O interfaces and the dynamic phase aligner (DPA) over single-ended I/Os and their contribution to the overall system bandwidth achievable with Stratix(R) IV FPGAs. All references to Stratix IV devices in this chapter apply to Stratix IV E, GT, and GX devices. The Stratix IV device family consists of the Stratix IV E (Enhanced) devices without high-speed clock data recovery (CDR) based transceivers, Stratix IV GT devices with up to 48 CDR-based transceivers running up to 11.3 Gbps, and Stratix IV GX devices with up to 48 CDR-based transceivers running up to 8.5 Gbps. The following sections describe the Stratix IV high-speed differential I/O interfaces and DPA: "Locations of the I/O Banks" on page 8-3 "LVDS Channels" on page 8-4 "LVDS SERDES" on page 8-8 "ALTLVDS Port List" on page 8-9 "Differential Transmitter" on page 8-11 "Differential Receiver" on page 8-17 "LVDS Interface with the Use External PLL Option Enabled" on page 8-26 "Left and Right PLLs (PLL_Lx and PLL_Rx)" on page 8-29 "Stratix IV Clocking" on page 8-30 "Source-Synchronous Timing Budget" on page 8-31 "Differential Pin Placement Guidelines" on page 8-38 Overview All Stratix IV E, GX, and GT devices have built-in serializer/deserializer (SERDES) circuitry that supports high-speed LVDS interfaces at data rates of up to 1.6 Gbps. SERDES circuitry is configurable to support source-synchronous communication protocols such as Utopia, Rapid I/O, XSBI, small form factor interface (SFI), serial peripheral interface (SPI), and asynchronous protocols such as SGMII and Gigabit Ethernet. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 September 2012 Feedback Subscribe 8-2 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Overview The Stratix IV device family has the following dedicated circuitry for high-speed differential I/O support: Differential I/O buffer Transmitter serializer Receiver deserializer Data realignment DPA Synchronizer (FIFO buffer) Phase-locked loops (PLLs) (located on left and right sides of the device) For high-speed differential interfaces, the Stratix IV device family supports the following differential I/O standards: LVDS Mini-LVDS Reduced swing differential signaling (RSDS) In the Stratix IV device family, I/Os are divided into row and column I/Os. Figure 8-1 shows I/O bank support for the Stratix IV device family. The row I/Os provide dedicated SERDES circuitry. Figure 8-1. I/O Bank Support in the Stratix IV Device Family (1), (2), (3), (4) LVDS I/Os Row I/Os with Dedicated SERDES Circuitry (3), (4) LVDS Interface with 'Use External PLL' Option Enabled Column I/Os (1), (2) LVDS Interface with 'Use External PLL' Option Disabled Notes to Figure 8-1: (1) Column input buffers are true LVDS buffers, but do not support 100-differential on-chip termination. (2) Column output buffers are single ended and need external termination schemes to support LVDS, mini-LVDS, and RSDS standards. For more information, refer to the I/O Features in Stratix IV Devices chapter. (3) Row input buffers are true LVDS buffers and support 100-differential on-chip termination. (4) Row output buffers are true LVDS buffers. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Locations of the I/O Banks 8-3 The ALTLVDS transmitter and receiver requires various clock and load enable signals from a left or right PLL. The Quartus(R) II software provides the following two choices when configuring the LVDS SERDES circuitry when using the PLL: 1 LVDS interface with the Use External PLL option enabled--You control the PLL settings, such as dynamically reconfiguring the PLL to support different data rates, dynamic phase shift, and so on. You must enable the Use External PLL option in the ALTLVDS_TX and ALTLVDS_RX megafunctions, using the ALTLVDS MegaWizard Plug-in Manager software. You also must instantiate an ALTPLL megafunction to generate the various clocks and load enable signals. For more information, refer to "LVDS Interface with the Use External PLL Option Enabled" on page 8-26. LVDS interface with the Use External PLL option disabled--The Quartus II software configures the PLL settings automatically. The software is also responsible for generating the various clock and load enable signals based on the input reference clock and data rate selected. Both choices target the same physical PLL; the only difference is the additional flexibility provided when an LVDS interface has the Use External PLL option enabled. Locations of the I/O Banks Stratix IV I/Os are divided into 16 to 24 I/O banks. The dedicated circuitry that supports high-speed differential I/Os is located in banks in the right and left side of the device. Figure 8-2 shows a high-level chip overview of the Stratix IV E device. Figure 8-2. High-Speed Differential I/Os with DPA Locations in Stratix IV E Devices General Purpose I/O and Memory Interface PLL PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL FPGA Fabric (Logic Elements, DSP, Embedded Memory, Clock Networks) PLL General Purpose I/O and Memory Interface September 2012 Altera Corporation PLL PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and Memory Interface PLL PLL PLL General Purpose I/O and Memory Interface Stratix IV Device Handbook Volume 1 8-4 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Channels Figure 8-3 shows a high-level chip overview of the Stratix IV GT and GX devices. Figure 8-3. High-Speed Differential I/Os with DPA Locations in Stratix IV GT and GX Devices PLL PLL General Purpose I/O and Memory Interface PLL PCI Express Hard IP Block PLL PLL PLL PCI Express Hard IP Block PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL FPGA Fabric (Logic Elements, DSP, Embedded Memory, Clock Networks) General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PLL General Purpose I/O and High-Speed LVDS I/O with DPA and Soft CDR PCI Express Hard IP Block PLL General Purpose I/O and Memory Interface Transceiver Transceiver Transceiver Transceiver Block Block Block Block PLL PLL PCI Express Hard IP Block Transceiver Transceiver Transceiver Transceiver Block Block Block Block General Purpose I/O and Memory Interface General Purpose I/O and Memory Interface LVDS Channels The Stratix IV device family supports LVDS on both row and column I/O banks. Row I/Os support true LVDS input with 100- differential input termination (OCT RD), and true LVDS output buffers. Column I/Os supports true LVDS input buffers without OCT RD. Alternately, you can configure the row and column LVDS pins as emulated LVDS output buffers that use two single-ended output buffers with an external resistor network to support LVDS, mini-LVDS, and RSDS standards. Stratix IV devices offer single-ended I/O refclk support for the LVDS. Dedicated SERDES and DPA circuitries are implemented on the row I/O banks to further enhance LVDS interface performance in the device. For column I/O banks, SERDES is implemented in the core logic because there is no dedicated SERDES circuitry on column I/O banks. 1 Stratix IV Device Handbook Volume 1 Emulated differential output buffers support tri-state capability starting with the Quartus II software version 9.1. September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Channels 8-5 Table 8-1 and Table 8-2 list the maximum number of row and column LVDS I/Os supported in Stratix IV E devices. You can design the LVDS I/Os as true LVDS buffers or emulated LVDS buffers, as long as the combination of the two do not exceed the maximum count. For example, there are a total of 112 LVDS pairs on row I/Os in the 780-pin EP4SE230 device (refer to Table 8-1). You can design up to a maximum of 56 true LVDS input buffers and 56 true LVDS output buffers, or up to a maximum of 112 emulated LVDS output buffers. For the 780-pin EP4SE230 device (refer to Table 8-2), there are a total of 128 LVDS pairs on column I/Os. You can design up to a maximum of 64 true LVDS input buffers and 64 emulated LVDS output buffers, or up to a maximum of 128 emulated LVDS output buffers. Table 8-1. LVDS Channels Supported in Stratix IV E Device Row I/O Banks (1), (2), (3) Device 780-Pin FineLine BGA 1152-Pin FineLine BGA 1517-Pin FineLine BGA 1760- Pin FineLine BGA EP4SE230 56 Rx or eTx + 56 Tx or eTx -- -- -- EP4SE360 56 Rx or eTx + 56 Tx or eTx (4) 88 Rx or eTx + 88 Tx or eTx -- -- EP4SE530 -- 88 Rx or eTx + 88 Tx or eTx (5) 112 Rx or eTx + 112 Tx or eTx (6) 112 Rx or eTx + 112 Tx or eTx EP4SE820 -- 88 Rx or eTx + 88 Tx or eTx 112 Rx or eTx + 112 Tx or eTx 132 Rx or eTx + 132 Tx or eTx Notes to Table 8-1: (1) Receiver (Rx) = true LVDS input buffers with OCT RD, Transmitter (Tx) = true LVDS output buffers, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channels are equally divided between the left and right sides of the device. (3) The LVDS channel count does not include dedicated clock input pins. (4) EP4SE360 devices are offered in the H780 package instead of the F780 package. (5) EP4SE530 devices are offered in the H1152 package instead of the F1152 package. (6) EP4SE530 devices are offered in the H1517 package instead of the F1517 package. Table 8-2. LVDS Channels Supported in Stratix IV E Device Column I/O Banks (1), (2), (3) Device 780-Pin FineLine BGA 1152-Pin FineLine BGA 1517-Pin FineLine BGA 1760-Pin FineLine BGA EP4SE230 64 Rx or eTx + 64 eTx -- -- -- 96 Rx or eTx + 96 eTx -- -- 96 Rx or eTx + 96 eTx 128 Rx or eTx + 128 eTx (5) (6) 64 Rx or eTx + 64 eTx EP4SE360 (4) EP4SE530 -- EP4SE820 -- 96 Rx or eTx + 96 eTx 128 Rx or eTx + 128 eTx 128 Rx or eTx + 128 eTx 144 Rx or eTx + 144 eTx Notes to Table 8-2: (1) Rx = true LVDS input buffers without OCT RD, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channels are equally divided between the top and bottom sides of the device. (3) The LVDS channel count does not include dedicated clock input pins. (4) EP4SE360 devices are offered in the H780 package instead of the F780 package. (5) EP4SE530 devices are offered in the H1152 package instead of the F1152 package. (6) EP4SE530 devices are offered in the H1517 package instead of the F1517 package. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-6 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Channels Table 8-3 and Table 8-4 list the maximum number of row and column LVDS I/Os supported in Stratix IV GT devices. Table 8-3. LVDS Channels Supported in Stratix IV GT Device Row I/O Banks (1), (2) Device 1517-pin FineLine BGA 1932-pin FineLine BGA EP4S40G2 46 Rx or eTx + 73 Tx or eTx -- EP4S40G5 46 Rx or eTx + 73 Tx or eTx -- EP4S100G2 46 Rx or eTx + 73 Tx or eTx -- EP4S100G3 -- 47 Rx or eTx + 56 Tx or eTx EP4S100G4 -- 47 Rx or eTx + 56 Tx or eTx EP4S100G5 46 Rx or eTx + 73 Tx or eTx 47 Rx or eTx + 56 Tx or eTx Notes to Table 8-3: (1) Rx = true LVDS input buffers with OCT RD, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channel count does not include dedicated clock input pins. Table 8-4. LVDS Channels Supported in Stratix IV GT Device Column I/O Banks (1), (2) Device 1517-pin FineLine BGA 1932-pin FineLine BGA EP4S40G2 96 Rx or eTx + 96 eTx -- EP4S40G5 96 Rx or eTx + 96 eTx -- EP4S100G2 96 Rx or eTx + 96 eTx -- EP4S100G3 -- 128 Rx or eTx + 128 eTx EP4S100G4 -- 128 Rx or eTx + 128 eTx EP4S100G5 96 Rx or eTx + 96 eTx 128 Rx or eTx + 128 eTx Notes to Table 8-4: (1) Rx = true LVDS input buffers without OCT RD, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channel count does not include dedicated clock input pins. Table 8-5 and Table 8-6 list the maximum number of row and column LVDS I/Os supported in Stratix IV GX devices. Table 8-5. LVDS Channels Supported in Stratix IV GX Device Row I/O Banks 780-Pin FineLine BGA 1152-Pin FineLine BGA EP4SGX70 28 Rx or eTx + 28 Tx or eTx -- EP4SGX110 28 Rx or eTx + 28 Tx or eTx EP4SGX180 EP4SGX230 Device EP4SGX290 1152-Pin FineLine BGA (1), (2), (3) (Part 1 of 2) 1517-Pin FineLine BGA 1760-Pin FineLine BGA 1932-Pin FineLine BGA 56 Rx or eTx + 56 Tx or eTx -- -- -- 28 Rx or eTx + 28 Tx or eTx 56 Rx or eTx + 56 Tx or eTx -- -- -- 28 Rx or eTx + 28 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 88 Rx or eTx + 88 Tx or eTx -- -- 28 Rx or eTx + 28 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 88 Rx or eTx + 88 Tx or eTx -- -- 44 Rx or eTx + 44 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 88 Rx or eTx + 88 Tx or eTx 88 Rx or eTx + 88 Tx or eTx 98 Rx or eTx + 98 Tx or eTx -- Stratix IV Device Handbook Volume 1 (5) (4) September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Channels 8-7 Table 8-5. LVDS Channels Supported in Stratix IV GX Device Row I/O Banks Device 780-Pin FineLine BGA EP4SGX360 -- EP4SGX530 (5) -- 1152-Pin FineLine BGA 1152-Pin FineLine BGA (1), (2), (3) (Part 2 of 2) (4) 1517-Pin FineLine BGA 1760-Pin FineLine BGA 1932-Pin FineLine BGA 44 Rx or eTx + 44 Tx or eTx 44 Rx or eTx + 44 Tx or eTx 88 Rx or eTx + 88 Tx or eTx 88 Rx or eTx + 88 Tx or eTx 98 Rx or eTx + 98 Tx or eTx -- 44 Rx or eTx + 44 Tx or eTx (6) 88 Rx or eTx + 88 Tx or eTx (7) 88 Rx or eTx + 88 Tx or eTx 98 Rx or eTx + 98 Tx or eTx Notes to Table 8-5: (1) Rx = true LVDS input buffers with OCT RD, Tx = true LVDS output buffers, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channels are equally divided between the left and right sides of the device, except for the devices in the 780-pin Fineline BGA. These devices have the LVDS Rx and Tx located on the left side of the device. (3) The LVDS channel count does not include dedicated clock input pins. (4) This package supports PMA-only transceiver channels. (5) EP4SGX290 and EP4SGX360 devices are offered in the H780 package instead of the F780 package. (6) EP4SGX530 devices are offered in the H1152 package instead of the F1152 package. (7) EP4SGX530 devices are offered in the H1517 package instead of the F1517 package. Table 8-6. LVDS Channels Supported in Stratix IV GX Device Column I/O Banks (1), 1152-Pin FineLine BGA (2), (3) 1517-Pin FineLine BGA 1760-Pin FineLine BGA 1932-Pin FineLine BGA 64 Rx or eTx + 64 eTx -- -- -- 64 Rx or eTx + 64 eTx 64 Rx or eTx + 64 eTx -- -- -- 64 Rx or eTx + 64 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx -- -- EP4SGX230 64 Rx or eTx + 64 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx -- -- EP4SGX290 72 Rx or eTx + 72 eTx (5) 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 128 Rx or eTx + 128 eTx 128 Rx or eTx + 128 eTx (8) EP4SGX360 72 Rx or eTx + 72 eTx (5) 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 96 Rx or eTx + 96 eTx 128 Rx or eTx + 128 eTx 128 Rx or eTx + 128 eTx (8) EP4SGX530 -- -- 96 Rx or eTx + 96 eTx (6) 96 Rx or eTx + 96 eTx (7) 128 Rx or eTx + 128 eTx 128 Rx or eTx + 128 eTx 780-Pin FineLine BGA 1152-Pin FineLine BGA EP4SGX70 64 Rx or eTx + 64 eTx -- EP4SGX110 64 Rx or eTx + 64 eTx EP4SGX180 Device (4) Notes to Table 8-6: (1) Rx = true LVDS input buffers without OCT RD, eTx = emulated LVDS output buffers (either LVDS_E_1R or LVDS_E_3R). (2) The LVDS Rx and Tx channels are equally divided between the left and right sides of the device. (3) The LVDS channel count does not include dedicated clock input pins. (4) This package supports PMA-only transceiver channels. (5) EP4SGX290 and EP4SGX360 devices are offered in the H780 package instead of the F780 package. (6) EP4SGX530 devices are offered in the H1152 package instead of the F1152 package. (7) EP4SGX530 devices are offered in the H1517 package instead of the F1517 package. (8) The Quartus II software version 9.0 does not support EP4SGX290 and EP4SGX360 devices in the 1932-Pin FineLine BGA package. These devices will be supported in a future release of the Quartus II software. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-8 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS SERDES LVDS SERDES Figure 8-4 shows a transmitter and receiver block diagram for the LVDS SERDES circuitry in the left and right banks. This diagram shows the interface signals of the transmitter and receiver data path. For more information, refer to "Differential Transmitter" on page 8-11 and "Differential Receiver" on page 8-17. Figure 8-4. LVDS SERDES (1), (2), (3) Serializer tx_in 2 IOE Supports SDR, DDR, or Non-Registered Datapath IOE tx_out + - 10 DIN DOUT LVDS Transmitter tx_coreclock 3 (LVDS_LOAD_EN, diffioclk, tx_coreclock) IOE Supports SDR, DDR, or Non-Registered Datapath 10 2 LVDS Receiver + - IOE rx_out rx_in Synchronizer FPGA Fabric Deserializer Bit Slip DOUT DIN DOUT DIN DPA Circuitry Retimed Data DOUT DIN DIN diffioclk 2 (LOAD_EN, diffioclk) Clock MUX DPA_diffioclk LVDS_diffioclk DPA Clock 3 (DPA_LOAD_EN, DPA_diffioclk, rx_divfwdclk) rx_divfwdclk rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclock LVDS Clock Domain DPA Clock Domain 8 Serial LVDS Clock Phases Left/Right PLL rx_inclock/tx_inclock Notes to Figure 8-4: (1) This diagram shows a shared PLL between the transmitter and receiver. If the transmitter and receiver are not sharing the same PLL, the two left and right PLLs are required. (2) In SDR and DDR modes, the data width is 1 and 2 bits, respectively. (3) The tx_in and rx_out ports have a maximum data width of 10 bits. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices ALTLVDS Port List 8-9 ALTLVDS Port List Table 8-7 lists the interface signals for an LVDS transmitter and receiver. Table 8-7. Port List of the LVDS Interface (ALTLVDS) Port Name Input / Output (1), (2) (Part 1 of 3) Description PLL Signals pll_areset Input Asynchronous reset to the LVDS transmitter and receiver PLL. The minimum pulse width requirement for this signal is 10 ns. Input The data bus width per channel is the same as the serialization factor (SF). Input data must be synchronous to the tx_coreclock signal. LVDS Transmitter Interface Signals tx_in[ ] Reference clock input for the transmitter PLL. Input tx_inclock The ALTLVDS MegaWizard Plug-In Manager software automatically selects the appropriate PLL multiplication factor based on the data rate and reference clock frequency selection. For more information about the allowed frequency range for this reference clock, refer to the "High-Speed I/O Specification" section in the DC and Switching Characteristics for Stratix IV Devices chapter. Input This port is instantiated only when you select the Use External PLL option in the MegaWizard Plug-In Manager software. This input port must be driven by the PLL instantiated though the ALTPLL MegaWizard Plug-In Manager software. tx_out Output LVDS transmitter serial data output port. tx_out is clocked by a serial clock generated by the left and right PLL. tx_outclock Output The frequency of this clock is programmable to be the same as the data rate, half the data rate, or one-fourth the data rate. The phase offset of this clock, with respect to the serial data, is programmable in increments of 45. tx_enable (3) FPGA fabric-transmitter interface clock. The parallel transmitter data generated in the FPGA fabric must be clocked with this clock. tx_coreclock (3) Output tx_locked September 2012 Output Altera Corporation This port is not available when you select the Use External PLL option in the MegaWizard Plug-In Manager software. The FPGA fabric-transmitter interface clock must be driven by the PLL instantiated through the ALTPLL MegaWizard Plug-In Manager software. When high, this signal indicates that the transmitter PLL is locked to the input reference clock. Stratix IV Device Handbook Volume 1 8-10 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices ALTLVDS Port List Table 8-7. Port List of the LVDS Interface (ALTLVDS) (1), (2) Port Name Input / Output (Part 2 of 3) Description LVDS Receiver Interface Signals rx_in Input LVDS receiver serial data input port. Reference clock input for the receiver PLL. rx_inclock Input The ALTLVDS MegaWizard Plug-In Manager software automatically selects the appropriate PLL multiplication factor based on the data rate and reference clock frequency selection. For more information about the allowed frequency range for this reference clock, refer to the "High-Speed I/O Specification" section in the DC and Switching Characteristics for Stratix IV Devices chapter. Input Edge-sensitive bit-slip control signal. Each rising edge on this signal causes the data re-alignment circuitry to shift the word boundary by one bit. The minimum pulse width requirement is one parallel clock cycle. There is no maximum pulse width requirement. Input When low, the DPA tracks any dynamic phase variations between the clock and data. When high, the DPA holds the last locked phase and does not track any dynamic phase variations between the clock and data. This port is not available in non-DPA mode. Input This port is instantiated only when you select the Use External PLL option in the MegaWizard Plug-In Manager software. This input port must be driven by the PLL instantiated though the ALTPLL MegaWizard Plug-In Manager software. Output Receiver parallel data output. The data bus width per channel is the same as the deserialization factor (DF). The output data is synchronous to the rx_outclock signal in non-DPA and DPA modes. It is synchronous to the rx_divfwdclk signal in soft-CDR mode. rx_outclock Output Parallel output clock from the receiver PLL. The parallel data output from the receiver is synchronous to this clock in non-DPA and DPA modes. This port is not available when you select the Use External PLL option in the MegaWizard Plug-In Manager software. The FPGA fabric-receiver interface clock must be driven by the PLL instantiated through the ALTPLL MegaWizard Plug-In Manager software. rx_locked Output When high, this signal indicates that the receiver PLL is locked to rx_inclock. rx dpa locked Output This signal only indicates an initial DPA lock condition to the optimum phase after power up or reset. This signal is not de-asserted if the DPA selects a new phase out of the eight clock phases to sample the received data. You must not use the rx_dpa_locked signal to determine a DPA loss-of-lock condition. rx_cda_max Output Data re-alignment (bit slip) roll-over signal. When high for one parallel clock cycle, this signal indicates that the user-programmed number of bits for the word boundary to roll-over have been slipped. rx_divfwdclk Output Parallel DPA clock to the FPGA fabric logic array. The parallel receiver output data to the FPGA fabric logic array is synchronous to this clock in soft-CDR mode. This signal is not available in non-DPA and DPA modes. dpa_pll_recal Input Enable PLL calibration dynamically without resetting the DPA circuitry or the PLL. rx_channel_data_align rx_dpll_hold rx_enable (3) rx_out[ ] Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter Table 8-7. Port List of the LVDS Interface (ALTLVDS) Input / Output Port Name (1), (2) 8-11 (Part 3 of 3) Description Output Busy signal that is asserted high when the PLL calibration occurs. Input Asynchronous reset to the DPA circuitry and FIFO. The minimum pulse width requirement for this reset is one parallel clock cycle. This signal resets DPA and FIFO blocks. rx_fifo_reset Input Asynchronous reset to the FIFO between the DPA and the data realignment circuits. The synchronizer block must be reset after a DPA loses lock condition and the data checker shows corrupted received data. The minimum pulse width requirement for this reset is one parallel clock cycle. This signal resets the FIFO block. rx_cda_reset Input Asynchronous reset to the data realignment circuitry. The minimum pulse width requirement for this reset is one parallel clock cycle. This signal resets the data realignment block. dpa_pll_cal_busy Reset Signals rx_reset Notes to Table 8-7: (1) Unless stated, signals are valid in all three modes (non-DPA, DPA, and soft-CDR) for a single channel. (2) All reset and control signals are active high. (3) For more information, refer to "LVDS Interface with the Use External PLL Option Enabled" on page 8-26. f For more information about the LVDS transmitter and receiver settings using ALTLVDS_TX and ALTLVDS_RX megafunction, refer to the ALTLVDS Megafunction User Guide. Differential Transmitter The Stratix IV transmitter has a dedicated circuitry to provide support for LVDS signaling. The dedicated circuitry consists of a differential buffer, a serializer, and left and right PLLs that can be shared between the transmitter and receiver. The differential buffer can drive out LVDS, mini-LVDS, and RSDS signaling levels. The serializer takes up to 10 bits wide parallel data from the FPGA fabric, clocks it into the load registers, and serializes it using shift registers clocked by the left and right PLL before sending the data to the differential buffer. The MSB of the parallel data is transmitted first. 1 September 2012 When using emulated LVDS I/O standards at the differential transmitter, the SERDES circuitry must be implemented in logic cells but not hard SERDES. Altera Corporation Stratix IV Device Handbook Volume 1 8-12 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter The load enable (LVDS_LOAD_EN) signal and the diffioclk signal (the clock running at serial data rate) generated from PLL_Lx (left PLL) or PLL_Rx (right PLL) clocks the load and shift registers. You can statically set the serialization factor to x3, x4, x6, x7, x8, or x10 using the Quartus II software. The load enable signal is derived from the serialization factor setting. Figure 8-5 shows a block diagram of the Stratix IV transmitter. Figure 8-5. Stratix IV Transmitter (1), (2) Serializer tx_in 10 DIN 2 IOE IOE supports SDR, DDR, or Non-Registered Datapath + - DOUT tx_out FPGA Fabric LVDS Transmitter tx_coreclock 3 (LVDS_LOAD_EN, diffioclk, tx_coreclock) Left/Right PLL tx_inclock LVDS Clock Domain Notes to Figure 8-5: (1) In SDR and DDR modes, the data width is 1 and 2 bits, respectively. (2) The tx_in port has a maximum data width of 10 bits. You can configure any Stratix IV transmitter data channel to generate a source-synchronous transmitter clock output. This flexibility allows the placement of the output clock near the data outputs to simplify board layout and reduce clock-to-data skew. Different applications often require specific clock-to-data alignments or specific data-rate-to-clock-rate factors. The transmitter can output a clock signal at the same rate as the data with a maximum frequency of 800 MHz. The output clock can also be divided by a factor of 1, 2, 4, 6, 8, or 10, depending on the serialization factor. You can set the phase of the clock in relation to the data at 0 or 180 (edge or center aligned). The left and right PLLs (PLL_Lx and PLL_Rx) provide additional support for other phase shifts in 45 increments. These settings are made statically in the Quartus II MegaWizard Plug-In Manager software. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter 8-13 Figure 8-6 shows the Stratix IV transmitter in clock output mode. In clock output mode, you can use an LVDS channel as a clock output channel. Figure 8-6. Stratix IV Transmitter in Clock Output Mode Transmitter Circuit Parallel Series Txclkout+ Txclkout- FPGA Fabric Left/Right PLL diffioclk LVDS_LOAD_EN You can bypass the Stratix IV serializer to support DDR (x2) and SDR (x1) operations to achieve a serialization factor of 2 and 1, respectively. The I/O element (IOE) contains two data output registers that can each operate in either DDR or SDR mode. Figure 8-7 shows the serializer bypass path. Figure 8-7. Serializer Bypass in Stratix IV Devices (1), Serializer tx_in 2 DIN 2 (2), (3) IOE supports SDR, DDR, or Non-Registered Datapath IOE + - DOUT tx_out FPGA Fabric LVDS Transmitter tx_coreclock (LVDS_LOAD_EN, diffioclk, tx_coreclock) 3 Left/Right PLL Notes to Figure 8-7: (1) All disabled blocks and signals are grayed out. (2) In DDR mode, tx_inclock clocks the IOE register. In SDR mode, data is directly passed through the IOE. (3) In SDR and DDR modes, the data width to the IOE is 1 and 2 bits, respectively. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-14 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter Programmable VOD and Programmable Pre-Emphasis Stratix IV LVDS transmitters support programmable pre-emphasis and programmable VOD. Pre-emphasis increases the amplitude of the high-frequency component of the output signal, and thus helps to compensate for the frequency-dependent attenuation along the transmission line. Figure 8-8 shows the differential LVDS output. Figure 8-8. Differential VOD Single-Ended Waveform Positive Channel (p) VOD Negative Channel (n) VCM Ground VOD (diff peak - peak) = 2 x VOD(single-ended) Differential Waveform VOD p - n = 0V VOD Figure 8-9 shows the LVDS output with pre-emphasis. Figure 8-9. Programmable Pre-Emphasis (1) OUT VP VOD OUT VP Note to Figure 8-9: (1) VP-- voltage boost from pre-emphasis. VOD-- Differential output voltage (peak-peak). Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter 8-15 Pre-emphasis is an important feature for high-speed transmission. Without pre-emphasis, the output current is limited by the VOD setting and the output impedance of the driver. At high frequency, the slew rate may not be fast enough to reach full VOD before the next edge, producing pattern-dependent jitter. With pre-emphasis, the output current is boosted momentarily during switching to increase the output slew rate. The overshoot introduced by the extra current happens only during switching and does not ring, unlike the overshoot caused by signal reflection. The amount of pre-emphasis needed depends on the attenuation of the high-frequency component along the transmission line. The Quartus II software allows four settings for programmable pre-emphasis--zero (0), low (1), medium (2), and high (3). The default setting is low. The VOD is also programmable with four settings: low (0), medium low (1), medium high (2), and high (3). The default setting is medium low. Programmable VOD You can statically assign the VOD settings from the Assignment Editor. Table 8-8 lists the assignment name for programmable VOD and its possible values in the Quartus II software Assignment Editor. Table 8-8. Quartus II Software Assignment Editor To tx_out Assignment name Programmable Differential Output Voltage (VOD) Allowed values 0, 1, 2, 3 Figure 8-10 shows the assignment of programmable VOD for a transmit data output from the Quartus II software Assignment Editor. Figure 8-10. Quartus II Software Assignment Editor--Programmable VOD September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-16 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Transmitter Programmable Pre-Emphasis Four different settings are allowed for pre-emphasis from the Assignment Editor for each LVDS output channel. Table 8-9 lists the assignment name and its possible values for programmable pre-emphasis in the Quartus II software Assignment Editor. Table 8-9. Quartus II Software Assignment Editor To tx_out Assignment name Programmable Pre-emphasis Allowed values 0, 1, 2, 3 Figure 8-11 shows the assignment of programmable pre-emphasis for a transmit data output port from the Quartus II software Assignment Editor. Figure 8-11. Quartus II Software Assignment Editor - Programmable Pre-Emphasis Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver 8-17 Differential Receiver The Stratix IV device family has a dedicated circuitry to receive high-speed differential signals in row I/Os. Figure 8-12 shows the hardware blocks of the Stratix IV receiver. The receiver has a differential buffer and left and right PLLs that can be shared between the transmitter and receiver, a DPA block, a synchronizer, a data realignment block, and a deserializer. The differential buffer can receive LVDS, mini-LVDS, and RSDS signal levels, which are statically set in the Quartus II software Assignment Editor. The left and right PLL receives the external clock input and generates different phases of the same clock. The DPA block chooses one of the clocks from the left and right PLL and aligns the incoming data on each channel. The synchronizer circuit is a 1 bit wide by 6 bit deep FIFO buffer that compensates for any phase difference between the DPA clock and the data realignment block. If necessary, the user-controlled data realignment circuitry inserts a single bit of latency in the serial bit stream to align to the word boundary. The deserializer includes shift registers and parallel load registers, and sends a maximum of 10 bits to the internal logic. The Stratix IV device family supports three different receiver modes: "Non-DPA Mode" on page 8-22 "DPA Mode" on page 8-24 "Soft-CDR Mode" on page 8-25 The physical medium connecting the transmitter and receiver LVDS channels may introduce skew between the serial data and the source-synchronous clock. The instantaneous skew between each LVDS channel and the clock also varies with the jitter on the data and clock signals as seen by the receiver. The three different modes-- non-DPA, DPA, and soft-CDR--provide different options to overcome skew between the source synchronous clock (non-DPA, DPA) /reference clock (soft-CDR) and the serial data. 1 September 2012 Only non-DPA mode requires manual skew adjustment. Altera Corporation Stratix IV Device Handbook Volume 1 8-18 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver Non-DPA mode allows you to statically select the optimal phase between the source synchronous clock and the received serial data to compensate skew. In DPA mode, the DPA circuitry automatically chooses the best phase to compensate for the skew between the source synchronous clock and the received serial data. Soft-CDR mode provides opportunities for synchronous and asynchronous applications for chip-to-chip and short reach board-to-board applications for SGMII protocols. Figure 8-12. Receiver Block Diagram (1), (2) LVDS Receiver IOE Supports SDR, DDR, or Non-Registered Datapath 2 rx_out + IOE 10 rx_in Synchronizer Deserializer Bit Slip DOUT DIN DOUT DIN DPA Circuitry Retimed Data DOUT DIN DIN FPGA Fabric 2 DPA Clock LVDS_diffiioclk Clock Mux rx_divfwdclk DPA_diffioclk diffioclk (LOAD_EN, diffioclk) 3 (DPA_LOAD_EN, DPA_diffioclk, rx_divfwdclk) rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclk) 8 Serial LVDS Clock Phases LVDS Clock Domain DPA Clock Domain Left/Right PLL rx_inclock Notes to Figure 8-12: (1) In SDR and DDR modes, the data width from the IOE is 1 and 2 bits, respectively. (2) The rx_out port has a maximum data width of 10 bits. Differential I/O Termination The Stratix IV device family provides a 100- on-chip differential termination option on each differential receiver channel for LVDS standards. On-chip termination saves board space by eliminating the need to add external resistors on the board. You can enable on-chip termination in the Quartus II software Assignment Editor. On-chip differential termination is supported on all row I/O pins and dedicated clock input pins (CLK[0,2,9,11]). It is not supported for column I/O pins, dedicated clock input pins (CLK[1,3,8,10]), or the corner PLL clock inputs. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver 8-19 Figure 8-13 shows device on-chip termination. Figure 8-13. On-Chip Differential I/O Termination Stratix IV Differential Receiver with On-Chip 100 Termination LVDS Transmitter Z0 = 50 RD Z0 = 50 Receiver Hardware Blocks The differential receiver has the following hardware blocks: "DPA Block" on page 8-19 "Synchronizer" on page 8-20 "Data Realignment Block (Bit Slip)" on page 8-20 "Deserializer" on page 8-22 DPA Block The DPA block takes in high-speed serial data from the differential input buffer and selects one of the eight phases generated by the left and right PLL to sample the data. The DPA chooses a phase closest to the phase of the serial data. The maximum phase offset between the received data and the selected phase is 1/8 UI, which is the maximum quantization error of the DPA. The eight phases of the clock are equally divided, offering a 45 resolution. Figure 8-14 shows the possible phase relationships between the DPA clocks and the incoming serial data. Figure 8-14. DPA Clock Phase to Serial Data Timing Relationship rx_in D0 D1 D2 D3 D4 (1) Dn 0 45 90 135 180 225 270 315 Tvco 0.125Tvco Note to Figure 8-14: (1) TVCO is defined as the PLL serial clock period. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-20 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver The DPA block continuously monitors the phase of the incoming serial data and selects a new clock phase if needed. You can prevent the DPA from selecting a new clock phase by asserting the optional RX_DPLL_HOLD port, which is available for each channel. DPA circuitry does not require a fixed training pattern to lock to the optimum phase out of the eight phases. After reset or power up, DPA circuitry requires transitions on the received data to lock to the optimum phase. An optional output port, RX_DPA_LOCKED, is available to indicate an initial DPA lock condition to the optimum phase after power up or reset. This signal is not de-asserted if the DPA selects a new phase out of the eight clock phases to sample the received data. Do not use the rx_dpa_locked signal to determine a DPA loss-of-lock condition. Use data checkers such as a cyclic redundancy check (CRC) or diagonal interleaved parity (DIP-4) to validate the data. An independent reset port, RX_RESET, is available to reset the DPA circuitry. DPA circuitry must be retrained after reset. 1 The DPA block is bypassed in non-DPA mode. Synchronizer The synchronizer is a 1 bit wide and 6 bit deep FIFO buffer that compensates for the phase difference between DPA_diffioclk, which is the optimal clock selected by the DPA block, and LVDS_diffioclk, which is produced by the left and right PLL. The synchronizer can only compensate for phase differences, not frequency differences between the data and the receiver's input reference clock. An optional port, RX_FIFO_RESET, is available to the internal logic to reset the synchronizer. The synchronizer is automatically reset when the DPA first locks to the incoming data. Altera recommends using RX_FIFO_RESET to reset the synchronizer when the DPA signals a loss-of-lock condition and the data checker indicates corrupted received data. 1 The synchronizer circuit is bypassed in non-DPA and soft-CDR mode. Data Realignment Block (Bit Slip) Skew in the transmitted data along with skew added by the link causes channel-to-channel skew on the received serial data streams. If the DPA is enabled, the received data is captured with different clock phases on each channel. This may cause the received data to be misaligned from channel to channel. To compensate for this channel-to-channel skew and establish the correct received word boundary at each channel, each receiver channel has a dedicated data realignment circuit that realigns the data by inserting bit latencies into the serial stream. An optional RX_CHANNEL_DATA_ALIGN port controls the bit insertion of each receiver independently controlled from the internal logic. The data slips one bit on the rising edge of RX_CHANNEL_DATA_ALIGN. The requirements for the RX_CHANNEL_DATA_ALIGN signal include: Stratix IV Device Handbook Volume 1 The minimum pulse width is one period of the parallel clock in the logic array. The minimum low time between pulses is one period of the parallel clock. This is an edge-triggered signal. September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver 8-21 Valid data is available two parallel clock cycles after the rising edge of RX_CHANNEL_DATA_ALIGN. Figure 8-15 shows receiver output (RX_OUT) after one bit slip pulse with the deserialization factor set to 4. Figure 8-15. Data Realignment Timing rx_inclock rx_in 3 2 1 0 3 2 1 0 3 2 1 0 rx_outclock rx_channel_data_align rx_out 3210 321x xx21 0321 The data realignment circuit can have up to 11 bit-times of insertion before a rollover occurs. The programmable bit rollover point can be from 1 to 11 bit-times, independent of the deserialization factor. The programmable bit rollover point must be set equal to or greater than the deserialization factor, allowing enough depth in the word alignment circuit to slip through a full word. You can set the value of the bit rollover point using the MegaWizard Plug-In Manager software. An optional status port, RX_CDA_MAX, is available to the FPGA fabric from each channel to indicate when the preset rollover point is reached. Figure 8-16 shows a preset value of four bit-times before rollover occurs. The rx_cda_max signal pulses for one rx_outclock cycle to indicate that rollover has occurred. Figure 8-16. Receiver Data Re-alignment Rollover rx_inclock rx_channel_data_align rx_outclock rx_cda_max September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-22 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver Deserializer You can statically set the deserialization factor to 3, 4, 6, 7, 8, or 10 by using the Quartus II software. You can bypass the Stratix IV deserializer in the Quartus II MegaWizard Plug-In Manager software to support DDR (x2) or SDR (x1) operations, as shown Figure 8-17. The DPA and data realignment circuit cannot be used when the deserializer is bypassed. The IOE contains two data input registers that can operate in DDR or SDR mode. Figure 8-17. Deserializer Bypass in Stratix IV Devices (1), (2), (3) LVDS Receiver IOE Supports SDR, DDR, or Non-Registered Datapath 2 rx_out + IOE 2 rx_in Synchronizer Deserializer Bit Slip DOUT DIN DOUT DIN DPA Circuitry Retimed Data DOUT DIN DIN FPGA Fabric 2 DPA Clock rx_divfwdclk DPA_diffioclk Clock Mux LVDS_diffiioclk diffioclk (LOAD_EN, diffioclk) 3 (DPA_LOAD_EN, DPA_diffioclk, rx_divfwdclk) rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclk) 8 Serial LVDS Clock Phases Left/Right PLL Notes to Figure 8-17: (1) All disabled blocks and signals are grayed out. (2) In DDR mode, rx_inclock clocks the IOE register. In SDR mode, data is directly passed through the IOE. (3) In SDR and DDR modes, the data width from the IOE is 1 and 2 bits, respectively. Receiver Data Path Modes The Stratix IV device family supports three receiver datapath modes--non-DPA mode, DPA mode, and soft-CDR mode. Non-DPA Mode Figure 8-18 shows the non-DPA datapath block diagram. In non-DPA mode, the DPA and synchronizer blocks are disabled. Input serial data is registered at the rising or falling edge of the serial LVDS_diffioclk clock produced by the left and right PLL. You can select the rising/falling edge option using the ALTLDVS MegaWizard Plug-In Manager software. Both data realignment and deserializer blocks are clocked by the LVDS_diffioclk clock, which is generated by the left and right PLL. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver 1 8-23 When using non-DPA receivers, you must drive the PLL from a dedicated and compensated clock input pin. Compensated clock inputs are dedicated clock pins in the same I/O bank as the PLL. f For more information about dedicated and compensated clock inputs, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. Figure 8-18. Receiver Data Path in Non-DPA Mode (1), (2) LVDS Receiver IOE Supports SDR, DDR, or Non-Registered Datapath 2 rx_out + IOE 10 rx_in Synchronizer Deserializer Bit Slip DOUT DIN DOUT DIN DPA P Circuitr y Retimed Data DOUT DIN N DIN FPGA Fabric 2 DPA P Clock L LVDS_diffiioclk Clock Mux rx_divfwdclk DPA_diffioclk P diffioclk (LOAD_EN, diffioclk) 3 (DPA_LO P AD_EN, DPA_diffioclk, P rx_divfwdclk) rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclk) 8 Serial LVDS L Clock Phases Left/Right PLL rx_inclock LVDS Clock Domain Notes to Figure 8-18: (1) In SDR and DDR modes, the data width from the IOE is 1 and 2 bits, respectively. (2) The rx_out port has a maximum data width of 10 bits. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-24 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver DPA Mode Figure 8-19 shows the DPA mode datapath, where all the hardware blocks mentioned in "Receiver Hardware Blocks" on page 8-19 are active. The DPA block chooses the best possible clock (DPA_diffioclk) from the eight fast clocks sent by the left and right PLL. This serial DPA_diffioclk clock is used for writing the serial data into the synchronizer. A serial LVDS_diffioclk clock is used for reading the serial data from the synchronizer. The same LVDS_diffioclk clock is used in data realignment and deserializer blocks. Figure 8-19. Receiver Datapath in DPA Mode (1), (2), (3) LVDS Receiver IOE Supports SDR, DDR, or Non-Registered Datapath 2 rx_out + IOE 10 rx_in Synchronizer Deserializer Bit Slip DOUT DIN DOUT DIN DPA Circuitry Retimed Data DOUT DIN DIN FPGA Fabric 2 DPA Clock LVDS_diffiioclk Clock Mux rx_divfwdclk DPA_diffioclk diffioclk (LOAD_EN, diffioclk) 3 (DPA_LOAD_EN, DPA_diffioclk, rx_divfwdclk) rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclk) 8 Serial LVDS Clock Phases LVDS Clock Domain DPA Clock Domain Left/Right PLL rx_inclock Notes to Figure 8-19: (1) All disabled blocks and signals are grayed out. (2) In SDR and DDR modes, the data width from the IOE is 1 and 2 bits, respectively. (3) The rx_out port has a maximum data width of 10 bits. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Receiver 8-25 Soft-CDR Mode The Stratix IV LVDS channel offers soft-CDR mode to support the Gigabit Ethernet and SGMII protocols. A receiver PLL uses the local clock source for reference. Figure 8-20 shows the soft-CDR mode datapath. Figure 8-20. Receiver Datapath in Soft-CDR Mode (1), (2), (3) LVDS Receiver IOE Supports SDR, DDR, or Non-Registered Datapath 2 rx_out + IOE 10 rx_in Synchronizer Deserializer Bit Slip DOUT DIN DOUT DIN DPA Circuitry Retimed Data DOUT DIN DIN FPGA Fabric 2 DPA Clock LVDS_diffiioclk Clock Mux rx_divfwdclk DPA_diffioclk diffioclk (LOAD_EN, diffioclk) 3 (DPA_LOAD_EN, DPA_diffioclk, rx_divfwdclk) rx_outclock 3 (LVDS_LOAD_EN, LVDS_diffioclk, rx_outclk) 8 Serial LVDS Clock Phases LVDS Clock Domain DPA Clock Domain Left/Right PLL rx_inclock Notes to Figure 8-20: (1) All disabled blocks and signals are grayed out. (2) In SDR and DDR modes, the data width from the IOE is 1 and 2 bits, respectively. (3) The rx_out port has a maximum data width of 10 bits. In soft-CDR mode, the synchronizer block is inactive. The DPA circuitry selects an optimal DPA clock phase to sample the data. Use the selected DPA clock for bit-slip operation and deserialization. The DPA block also forwards the selected DPA clock, divided by the deserialization factor called rx_divfwdclk, to the FPGA fabric, along with the deserialized data. This clock signal is put on the periphery clock (PCLK) network. When using soft-CDR mode, the rx_reset port must not be asserted after the DPA training is asserted because the DPA will continuously choose new phase taps from the PLL to track parts per million (PPM) differences between the reference clock and incoming data. f For more information about periphery clock networks, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-26 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Interface with the Use External PLL Option Enabled You can use every LVDS channel in soft-CDR mode and can drive the FPGA fabric using the periphery clock network in the Stratix IV device family. The rx_dpa_locked signal is not valid in soft-CDR mode because the DPA continuously changes its phase to track PPM differences between the upstream transmitter and the local receiver input reference clocks. The parallel clock rx_outclock, generated by the left and right PLL, is also forwarded to the FPGA fabric. LVDS Interface with the Use External PLL Option Enabled The ALTLVDS MegaWizard Plug-In Manager software provides an option for implementing the LVDS interface with the Use External PLL option. With this option enabled you can control the PLL settings, such as dynamically reconfiguring the PLL to support different data rates, dynamic phase shift, and other settings. You also must instantiate an ALTPLL megafunction to generate the various clock and load enable signals. When you enable the Use External PLL option with the ALTLVDS transmitter and receiver, the following signals are required from the ALTPLL megafunction: 1 Serial clock input to the SERDES of the ALTLVDS transmitter and receiver Load enable to the SERDES of the ALTLVDS transmitter and receiver Parallel clock used to clock the transmitter FPGA fabric logic and parallel clock used for the receiver rx_syncclock port and receiver FPGA fabric logic Asynchronous PLL reset port of the ALTLVDS receiver As an example, Table 8-10 describes the serial clock output, load enable output, and parallel clock output generated on ports c0, c1, and c2, respectively, along with the locked signal of the ALTPLL instance. You can choose any of the PLL output clock ports to generate the interface clocks. f With soft SERDES, a different clocking requirement is needed. For more information, refer to the LVDS SERDES Transmitter/Receiver (ALTLVDS_RX/TX) Megafunction User Guide. 1 The high-speed clock generated from the PLL is intended to clock the LVDS SERDES circuitry only. Do not use the high-speed clock to drive other logic because the allowed frequency to drive the core logic is restricted by the PLL FOUT specification. For more information about the FOUT specification, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Table 8-10 lists the signal interface between the output ports of the ALTPLL megafunction and the input ports of the ALTLVDS transmitter and receiver. Table 8-10. Signal Interface Between ALTPLL and ALTLVDS_TX and ALTLVDS_RX Megafunctions (Part 1 of 2) From the ALTPLL Megafunction Serial clock output (c0) Load enable output (c1) Stratix IV Device Handbook Volume 1 To the ALTLVDS Transmitter (1) To the ALTLVDS Receiver tx_inclock (serial clock input to the transmitter) rx_inclock (serial clock input) tx_enable (load enable to the transmitter) rx_enable (load enable for the deserializer) September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Interface with the Use External PLL Option Enabled 8-27 Table 8-10. Signal Interface Between ALTPLL and ALTLVDS_TX and ALTLVDS_RX Megafunctions (Part 2 of 2) From the ALTPLL Megafunction Parallel clock output (c2) To the ALTLVDS Transmitter To the ALTLVDS Receiver Parallel clock used inside the transmitter core logic in the FPGA fabric rx_syncclock (parallel clock input) and parallel clock used inside the receiver core logic in the FPGA fabric ~(locked) pll_areset (asynchronous PLL reset port) (2) -- Notes to Table 8-10: (1) The serial clock output (c0) can only drive tx_inclock on the ALTLVDS transmitter and rx_inclock on the ALTLVDS receiver. This clock cannot drive the core logic. (2) The pll_areset signal is automatically enabled for the LVDS receiver in external PLL mode. This signal does not exist for LVDS transmitter instantiation when the external PLL option is enabled. 1 The rx_syncclock port is automatically enabled in an LVDS receiver in external PLL mode. The Quartus II compiler errors out if this port is not connected, as shown in Figure 8-21. When generating the ALTPLL megafunction, the Left/Right PLL option is configured to set up the PLL in LVDS mode. Figure 8-21 shows the connection between the ALTPLL and ALTLVDS_TX and ALTLVDS_RX megafunctions. Figure 8-21. LVDS Interface with the ALTPLL Megafunction (1) FPGA Fabric LVDS Transmitter (ALTLVDS) tx_inclock Transmitter Core Logic tx_in tx_enable tx_coreclk c0 c1 c2 rx_coreclk Receiver Core Logic LVDS Receiver (ALTLVDS) rx_inclock rx_out ALTPLL inclk0 pll_areset locked rx_enable rx_syncclock pll_areset Note to Figure 8-21: (1) Instantiation of pll_areset is optional for the ALTPLL instantiation. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-28 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices LVDS Interface with the Use External PLL Option Enabled Example 8-1 shows how to generate three output clocks using an ALTPLL megafunction. Example 8-1. Generating Three Output Clocks Using an ALTPLL Megafunction LVDS data rate = 1 Gbps; serialization factor = 10; input reference clock = 100 MHz The following settings are used when generating the three output clocks using an ALTPLL megafunction. The serial clock must be 1000 MHz and the parallel clock must be 100 MHz (serial clock divided by the serialization factor): c0 Frequency = 1000 MHz (multiplication factor = 10 and division factor = 1) Phase shift = -180 with respect to the voltage-controlled oscillator (VCO) clock Duty cycle = 50% c1 Frequency = (1000/10) = 100 MHz (multiplication factor = 1 and division factor = 1) Phase shift = (10 - 2) x 360/10 = 288 [(deserialization factor - 2)/deserialization factor] x 360 Duty cycle = (100/10) = 10% (100 divided by the serialization factor) c2 Frequency = (1000/10) = 100 MHz (multiplication factor = 1 and division factor = 1) Phase shift = (-180/10) = -18 (c0 phase shift divided by the serialization factor) Duty cycle = 50% Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Left and Right PLLs (PLL_Lx and PLL_Rx) 8-29 The Equation 8-1 calculations for phase shift assume that the input clock and serial data are edge aligned. Introducing a phase shift of -180 to sampling clock (c0) ensures that the input data is center-aligned with respect to the c0, as shown in Figure 8-22. Figure 8-22. Phase Relationship for External PLL Interface Signals inclk0 VCO clk (internal PLL clk) c0 (-180 phase shift) c1 (288 phase shift) c2 (-18 phase shift) Serial data D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 Left and Right PLLs (PLL_Lx and PLL_Rx) The Stratix IV device family contains up to eight left and right PLLs with up to four PLLs located on the left side and four on the right side of the device. The left PLLs can support high-speed differential I/O banks on the left side; the right PLLs can support high-speed differential I/O banks on the right side of the device. The high-speed differential I/O receiver and transmitter channels use these left and right PLLs to generate the parallel clocks (rx_outclock and tx_outclock) and high-speed clocks (diffioclk). Figure 8-2 on page 8-3 and Figure 8-3 on page 8-4 show the locations of the left and right PLLs for Stratix IV E, GT, and GX devices. The PLL VCO operates at the clock frequency of the data rate. Clock switchover and dynamic reconfiguration are allowed using the left and right PLL in high-speed differential I/O support mode. f For more information, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-30 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Stratix IV Clocking Stratix IV Clocking The left and right PLLs feed into the differential transmitter and receive channels through the LVDS and DPA clock network. The center left and right PLLs can clock the transmitter and receive channels above and below them. The corner left and right PLLs can drive I/Os in the banks adjacent to them. Figure 8-23 shows center PLL clocking in the Stratix IV device family. For more information about PLL clocking restrictions, refer to "Differential Pin Placement Guidelines" on page 8-38. Figure 8-23. LVDS/DPA Clocks in the Stratix IV Device Family with Center PLLs 4 LVDS Clock DPA Clock Quadrant Quadrant DPA Clock LVDS 4 Clock 4 4 2 Center PLL_L2 Center PLL_R2 Center PLL_L3 Center PLL_R3 2 2 2 4 4 4 LVDS Clock DPA Clock Quadrant Quadrant DPA Clock LVDS 4 Clock Figure 8-24 shows center and corner PLL clocking in the Stratix IV device family. For more information about PLL clocking restrictions, refer to "Differential Pin Placement Guidelines" on page 8-38. Figure 8-24. LVDS/DPA Clocks in the Stratix IV Device Family with Center and Corner PLLs Corner PLL_R1 Corner PLL_L1 2 2 4 LVDS Clock DPA Clock Quadrant Quadrant DPA Clock LVDS 4 Clock 4 4 2 2 Center PLL_L2 Center PLL_R2 Center PLL_L3 Center PLL_R3 2 2 4 4 4 LVDS Clock DPA Clock Quadrant Quadrant DPA Clock LVDS 4 Clock 2 2 Corner PLL_L4 Stratix IV Device Handbook Volume 1 Corner PLL_R4 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget 8-31 Source-Synchronous Timing Budget This section describes the timing budget, waveforms, and specifications for source-synchronous signaling in the Stratix IV device family. LVDS I/O standards enable high-speed data transmission. This high data transmission rate results in better overall system performance. To take advantage of fast system performance, it is important to understand how to analyze timing for these high-speed signals. Timing analysis for the differential block is different from traditional synchronous timing analysis techniques. Instead of focusing on clock-to-output and setup times, source synchronous timing analysis is based on the skew between the data and the clock signals. High-speed differential data transmission requires the use of timing parameters provided by IC vendors and is strongly influenced by board skew, cable skew, and clock jitter. This section defines the source-synchronous differential data orientation timing parameters, the timing budget definitions for the Stratix IV device family, and how to use these timing parameters to determine a design's maximum performance. Differential Data Orientation There is a set relationship between an external clock and the incoming data. For operations at 1 Gbps and a serialization factor of 10, the external clock is multiplied by 10. You can set phase-alignment in the PLL to coincide with the sampling window of each data bit. The data is sampled on the falling edge of the multiplied clock. Figure 8-25 shows the data bit orientation of the x10 mode. Figure 8-25. Bit Orientation in the Quartus II Software inclock/outclock 10 LVDS Bits MSB data in 9 8 7 6 5 4 3 LSB 2 1 0 Differential I/O Bit Position Data synchronization is necessary for successful data transmission at high frequencies. Figure 8-26 shows the data bit orientation for a channel operation. This figure is based on the following: September 2012 Serialization factor equals the clock multiplication factor Edge alignment is selected for phase alignment Implemented in hard SERDES Altera Corporation Stratix IV Device Handbook Volume 1 8-32 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget For other serialization factors, use the Quartus II software tools to find the bit position within the word. Table 8-11 lists the bit positions after deserialization. Figure 8-26. Bit-Order and Word Boundary for One Differential Channel (1) Transmitter Channel Operation (x8 Mode) tx_outclock tx_out X Current Cycle Next Cycle Previous Cycle X X X X X X X 7 6 5 4 3 2 1 0 X X X X X X X X MSB LSB Receiver Channel Operation (x8 Mode) rx_inclock rx_in 7 6 5 4 3 2 1 0 X X X X X X X X X X X X X X X X X X X X X X X X rx_outclock rx_out [7..0] XXXXXXXX XXXXXXXX XXXX7654 3210XXXX Note to Figure 8-26: (1) These are only functional waveforms and are not intended to convey timing information. Table 8-11 lists the conventions for differential bit naming for 18 differential channels. The MSB and LSB positions increase with the number of channels used in a system. Table 8-11. Differential Bit Naming Internal 8-Bit Parallel Data Receiver Channel Data Number Stratix IV Device Handbook Volume 1 MSB Position LSB Position 1 7 0 2 15 8 3 23 16 4 31 24 5 39 32 6 47 40 7 55 48 8 63 56 9 71 64 10 79 72 11 87 80 12 95 88 13 103 96 14 111 104 15 119 112 16 127 120 17 135 128 18 143 136 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget 8-33 Transmitter Channel-to-Channel Skew Transmitter channel-to-channel skew (TCCS) is an important parameter based on the Stratix IV transmitter in a source synchronous differential interface. This parameter is used in receiver skew margin calculation. For more information, refer to "Receiver Skew Margin for Non-DPA Mode" on page 8-33. TCCS is the difference between the fastest and slowest data output transitions, including the TCO variation and clock skew. For LVDS transmitters, the TimeQuest Timing Analyzer provides a TCCS report, which shows TCCS values for serial output ports. f You can get the TCCS value from the TCCS report (report_TCCS) in the Quartus II compilation report under the TimeQuest Timing Analyzer, or from the DC and Switching Characteristics for Stratix IV Devices chapter. Receiver Skew Margin for Non-DPA Mode Changes in system environment, such as temperature, media (cable, connector, or PCB), and loading effect the receiver's setup and hold times; internal skew affects the sampling ability of the receiver. Different modes of LVDS receivers use different specifications that can help in deciding the ability to sample the received serial data correctly. In DPA mode, you must use DPA jitter tolerance instead of receiver input skew margin (RSKM). In non-DPA mode, use TCCS, RSKM, and sampling window (SW) specifications for high-speed source-synchronous differential signals in the receiver data path. The relationship between RSKM, TCCS, and SW is expressed by the RSKM equation shown in Equation 8-1. Equation 8-1. RSKM TUI - SW - TCCS RSKM = ---------------------------------------------2 Conventions used for the equation: September 2012 Time unit interval (TUI)--Time period of the serial data. RSKM--The timing margin between the receiver's clock input and the data input sampling window. SW--The period of time that the input data must be stable to ensure that data is successfully sampled by the LVDS receiver. The SW is a device property and varies with device speed grade. TCCS--The timing difference between the fastest and the slowest output edges, including tCO variation and clock skew, across channels driven by the same PLL. The clock is included in the TCCS measurement. Altera Corporation Stratix IV Device Handbook Volume 1 8-34 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget Figure 8-27 shows the relationship between the RSKM, TCCS, and the receiver's SW. You must calculate the RSKM value to decide whether or not data can be sampled properly by the LVDS receiver with the given data rate and device. A positive RSKM value indicates that the LVDS receiver can sample the data properly, whereas a negative RSKM indicates that it cannot. Figure 8-27. Differential High-Speed Timing Diagram and Timing Budget for Non-DPA Mode Timing Diagram External Input Clock Time Unit Interval (TUI) Internal Clock TCCS TCCS Receiver Input Data RSKM SW RSKM Internal Clock Falling Edge Timing Budget TUI External Clock Clock Placement Internal Clock Synchronization Transmitter Output Data RSKM RSKM TCCS TCCS 2 Receiver Input Data SW Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget 8-35 For LVDS receivers, the Quartus II software provides an RSKM report showing the SW, TUI, and RSKM values for non-DPA mode. You can generate the RSKM report by executing the report_RSKM command in the TimeQuest Timing Analyzer. You can find the RSKM report in the Quartus II compilation report under the TimeQuest Timing Analyzer section. 1 In order to obtain the RSKM value, you must assign an appropriate input delay to the LVDS receiver through the TimeQuest Timing Analyzer constraints menu. For assigning input delay, follow these steps: 1. The Quartus II TimeQuest Timing Analyzer GUI has many options for setting the constraints and analyzing the design. Figure 8-28 shows various commands on the Constraints menu. For setting input delay, you must select the Set Input Delay option. Figure 8-28. Selection of Constraint Menu in TimeQuest Timing Analyzer September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-36 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget 2. Figure 8-29 shows the setting parameters for the Set Input Delay option. The clock name must reference the source synchronous clock that feeds the LVDS receiver. Select the desired clock using the pull-down menu. Figure 8-29. Input Time Delay Assignment Through TimeQuest Timing Analyzer 3. Figure 8-30 shows the Targets option. You can view a list of all available ports using the List option in the Name Finder window. Figure 8-30. Name Finder Window in Set Input Delay Option Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Source-Synchronous Timing Budget 8-37 4. Select the LVDS receiver serial input ports (from the list) according to the input delay you set. Click OK. 5. In the Set Input Delay window, set the appropriate values in the Input Delay Options section and Delay value. 6. Click Run to incorporate these values in the TimeQuest Timing Analyzer. 7. Assign the appropriate delay for all the LVDS receiver input ports following these steps. If you have already assigned Input Delay and you need to add more delay to that input port, use the Add Delay option in the Set Input Delay window. 1 If no input delay is set in the TimeQuest Timing Analyzer, the receiver channel-to-channel skew (RCCS) defaults to zero. You can also directly set the input delay in a Synopsys Design Constraint file (.sdc) using the set_input_delay command. f For more information about .sdc commands and the TimeQuest Timing Analyzer, refer to the Quartus II TimeQuest Timing Analyzer chapter in volume 3 of the Quartus II Development Software Handbook. Example 8-2 shows the RSKM calculation. Example 8-2. RSKM Data Rate: 1 Gbps, Board channel-to-channel skew = 200 ps For Stratix IV devices: TCCS = 100 ps (pending characterization) SW = 300 ps (pending characterization) TUI = 1000 ps Total RCCS = TCCS + Board channel-to-channel skew= 100 ps + 200 ps = 300 ps RSKM= TUI - SW - RCCS = 1000 ps - 300 ps - 300 ps = 400 ps > 0 Because the RSKM > 0 ps, receiver non-DPA mode must work correctly. 1 September 2012 You can also calculate RSKM using the steps described in "Guidelines for DPAEnabled Differential Channels" on page 8-38. Altera Corporation Stratix IV Device Handbook Volume 1 8-38 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines Differential Pin Placement Guidelines To ensure proper high-speed operation, differential pin placement guidelines have been established. The Quartus II compiler automatically checks that these guidelines are followed and issues an error message if they are not met. This section is divided into pin placement guidelines with and without DPA usage because DPA usage adds some constraints on the placement of high-speed differential channels. 1 DPA-enabled differential channels refer to DPA mode or soft-CDR mode; DPA disabled channels refer to non-DPA mode. Guidelines for DPA-Enabled Differential Channels The Stratix IV device family has differential receivers and transmitters in I/O banks on the left and right sides of the device. Each receiver has a dedicated DPA circuit to align the phase of the clock to the data phase of its associated channel. When you use DPA-enabled channels in differential banks, you must adhere to the guidelines listed in the following sections. DPA-Enabled Channels and Single-Ended I/Os When you enable a DPA channel in a bank, both single-ended I/Os and differential I/O standards are allowed in the bank. Single-ended I/Os are allowed in the same I/O bank, as long as the single-ended I/O standard uses the same VCCIO as the DPA-enabled differential I/O bank. Single-ended inputs can be in the same logic array block (LAB) row as a differential channel using the SERDES circuitry. DDIO can be placed within the same LAB row as a SERDES differential channel but half rate DDIO (single data rate) output pins cannot be placed within the same LAB row as a receiver SERDES differential channel. The input register must be implemented within the FPGA fabric logic. DPA-Enabled Channel Driving Distance If the number of DPA channels driven by each left and right PLL exceeds 25 LAB rows, Altera recommends implementing data realignment (bit slip) circuitry for all the DPA channels. Using Corner and Center Left and Right PLLs If a differential bank is being driven by two left and right PLLs, where the corner left and right PLL is driving one group and the center left and right PLL is driving another group, there must be at least one row of separation between the two groups of DPA-enabled channels (refer to Figure 8-31). The two groups can operate at independent frequencies. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines 8-39 You do not need a separation if a single left and right PLL is driving the DPA-enabled channels as well as DPA-disabled channels. Figure 8-31. Corner and Center Left and Right PLLs Driving DPA-Enabled Differential I/Os in the Same Bank Corner Left /Right PLL Reference CLK DPA -enabled Diff I/O DPA - enabled Diff I/O DPA - enabled Diff I/O Channels driven by Corner Left/Right PLL DPA - enabled Diff I/O DPA - enabled Diff I/O Diff I/O One Unused Channel for Buffer DPA-enabled Diff I/O DPA-enabled Diff I/O DPA -enabled Diff I/O Channels driven by Center Left/Right PLL DPA- enabled Diff I/O Reference CLK Center Left /Right PLL September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-40 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines Using Both Center Left and Right PLLs You can use both center left and right PLLs to drive DPA-enabled channels simultaneously, as long as they drive these channels in their adjacent banks only, as shown in Figure 8-32. If one of the center left and right PLLs drives the top and bottom banks, you cannot use the other center left and right PLL to drive differential channels, as shown in Figure 8-32. If the top PLL_L2 and PLL_R2 drives DPA-enabled channels in the lower differential bank, the PLL_L3 and PLL_R3 cannot drive DPA-enabled channels in the upper differential banks and vice versa. In other words, the center left and right PLLs cannot drive cross-banks simultaneously, as shown in Figure 8-33. Figure 8-32. Center Left and Right PLLs Driving DPA-Enabled Differential I/Os Stratix IV Device Handbook Volume 1 DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O Reference CLK DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O Reference CLK Center Left/Right PLL (PLL_L2/PLL_R2) Center Left/Right PLL (PLL_L2/PLL_R2) Center Left/Right PLL (PLL_L3/PLL_R3) Center Left/Right PLL (PLL_L3/PLL_R3) Reference CLK DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O Reference CLK DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O Unused PLL September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines 8-41 Figure 8-33. Invalid Placement of DPA-Enabled Differential I/Os Driven by Both Center Left and Right PLLs DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O Reference CLK Center Left /Right PLL Center Left /Right PLL Reference CLK DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O DPA-enabled Diff I/O September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-42 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines Guidelines for DPA-Disabled Differential Channels When you use DPA-disabled channels in the left and right banks of a Stratix IV device, you must adhere to the guidelines in the following sections. 1 When using non-DPA receivers, you must drive the PLL from a dedicated and compensated clock input pin. Compensated clock inputs are dedicated clock pins in the same I/O bank as the PLL. f For more information about dedicated and compensated clock inputs, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. DPA-Disabled Channels and Single-Ended I/Os The placement rules for DPA-disabled channels and single-ended I/Os are the same as those for DPA-enabled channels and single-ended I/Os. For more information, refer to "DPA-Enabled Channels and Single-Ended I/Os" on page 8-38. DPA-Disabled Channel Driving Distance Each left and right PLL can drive all the DPA-disabled channels in the entire bank. Using Corner and Center Left and Right PLLs You can use a corner left and right PLL to drive all transmitter channels and a center left and right PLL to drive all DPA-disabled receiver channels within the same differential bank. In other words, a transmitter channel and a receiver channel in the same LAB row can be driven by two different PLLs, as shown in Figure 8-34. A corner left and right PLL and a center left and right PLL can drive duplex channels in the same differential bank, as long as the channels driven by each PLL are not interleaved. Separation is not necessary between the group of channels driven by the corner and center left and right PLLs, as shown in Figure 8-34 and Figure 8-35. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines 8-43 Figure 8-34. Corner and Center Left and Right PLLs Driving DPA-Disabled Differential I/Os in the Same Bank Corner Left/Right Corner Left/ Right PLL PLL Reference CLK Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX Diff RX Diff TX DPA-disabled Diff I/O Diff RX Diff TX DPA -disabled Diff I /O Reference CLK Center Left/Right PLL September 2012 Altera Corporation Reference CLK DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O Channels driven by Corner Left/Right PLL No separation buffer needed Channels driven by Center Left/Right PLL Reference CLK Center Left/Right PLL Stratix IV Device Handbook Volume 1 8-44 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines Figure 8-35. Invalid Placement of DPA-Disabled Differential I/Os Due to Interleaving of Channels Driven by the Corner and Center Left and Right PLLs Corner Left/Right PLL Reference CLK DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O Reference CLK Center Left/Right PLL Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines 8-45 Using Both Center Left and Right PLLs You can use both center left and right PLLs simultaneously to drive DPA-disabled channels on upper and lower differential banks. Unlike DPA-enabled channels, the center left and right PLLs can drive cross-banks. For example, the upper-center left and right PLL can drive the lower differential bank at the same time the lower center left and right PLL is driving the upper differential bank, and vice versa, as shown in Figure 8-36. Figure 8-36. Both Center Left and Right PLLs Driving Cross-Bank DPA-Disabled Channels Simultaneously DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O Reference CLK Center Left/Right PLL Center Left/Right PLL Reference CLK DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O DPA-disabled Diff I/O September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 8-46 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines Document Revision History Table 8-12 lists the revision history for this chapter. Table 8-12. Document Revision History (Part 1 of 2) Date Version September 2012 December 2011 February 2011 March 2010 November 2009 June 2009 April 2009 March 2009 Stratix IV Device Handbook Volume 1 3.4 3.3 Changes Updated Figure 8-22 to close FB case #.28708. Updated the "Soft-CDR Mode" section to close FB #41886. Updated the "Overview" and "ALTLVDS Port List" sections. Updated Table 8-10. Updated Table 8-10. Updated the "Differential Transmitter", "Non-DPA Mode", "LVDS Interface with the Use External PLL Option Enabled", "Deserializer", and "Guidelines for DPA-Disabled Differential Channels" sections. Applied new template. Minor text edits. Removed note 7 from Table 8-1 and Table 8-2. Updated Figure 8-5. Updated the "LVDS Channels" section. Updated Table 8-7. Added a note to the "LVDS Interface with the Use External PLL Option Enabled" and "ALTLVDS Port List" sections. Minor text edits. Changed "dedicated LVDS" to "true LVDS". Removed EP4SE110, EP4SE290, and EP4SE680 devices. Added EP4SE820 and Stratix IV GT devices. Updated "LVDS Channels", "Differential Transmitter", "Soft-CDR Mode", and "DPAEnabled Channels and Single-Ended I/Os" sections. Updated Table 8-1, Table 8-2, Table 8-5, and Table 8-6. Added Table 8-3 and Table 8-4. Updated Example 8-1. Updated Figure 8-22. Minor text edits. Added an introductory paragraph to increase search ability. Minor text edits. Updated "Introduction". Updated Figure 8-3. Removed Table 8-5 and Table 8-6. Updated "Introduction", "Stratix IV LVDS Channels", "Stratix IV Differential Transmitter", "Differential I/O Termination", and "Dynamic Phase Alignment (DPA) Block" sections. Updated Table 8-1, Table 8-2, Table 8-3, Table 8-4, and Table 8-7. Added Table 8-5 and Table 8-6. Updated Figure 8-2. Removed "Referenced Documents" section. 3.2 3.1 3.0 2.3 2.2 2.1 September 2012 Altera Corporation Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines 8-47 Table 8-12. Document Revision History (Part 2 of 2) Date Version November 2008 May 2008 September 2012 2.0 1.0 Altera Corporation Changes Updated Figure 8-2, Figure 8-3, Figure 8-21, Figure 8-34. Removed Figure 8-31. Updated Table 8-1, Table 8-10. Updated "Differential Pin Placement Guidelines" section. Initial release. Stratix IV Device Handbook Volume 1 8-48 Stratix IV Device Handbook Volume 1 Chapter 8: High-Speed Differential I/O Interfaces and DPA in Stratix IV Devices Differential Pin Placement Guidelines September 2012 Altera Corporation Section III. System Integration This section includes the following chapters: Chapter 9, Hot Socketing and Power-On Reset in Stratix IV Devices Chapter 10, Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Chapter 11, SEU Mitigation in Stratix IV Devices Chapter 12, JTAG Boundary-Scan Testing in Stratix IV Devices Chapter 13, Power Management in Stratix IV Devices Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 III-2 Stratix IV Device Handbook Volume 1 Section III: System Integration September 2012 Altera Corporation 9. Hot Socketing and Power-On Reset in Stratix IV Devices February 2011 SIV51009-3.2 SIV51009-3.2 This chapter describes hot-socketing specifications, power-on reset (POR) requirements, and their implementation in Stratix(R) IV devices. Stratix IV devices offer hot socketing, also known as hot plug-in or hot swap, and power sequencing support without the use of external devices. You can insert or remove a Stratix IV device or a board in a system during system operation without causing undesirable effects to the running system bus or board that is inserted into the system. The hot-socketing feature also removes some of the difficulty when you use Stratix IV devices on PCBs that contain a mixture of 3.0-, 2.5-, 1.8-, 1.5-, and 1.2-V devices. The Stratix IV hot-socketing feature provides: Board or device insertion and removal without external components or board manipulation Support for any power-up sequence with the exception that VCC must power up fully before VCCAUX for all Stratix IV production devices I/O buffers non-intrusive to system buses during hot insertion This section also describes POR circuitry in Stratix IV devices. POR circuitry keeps the devices in the reset state until the power supply outputs are within operating range (provided VCC powers up fully before VCCAUX). This chapter contains the following sections: "Stratix IV Hot-Socketing Specifications" "Hot-Socketing Feature Implementation in Stratix IV Devices" on page 9-2 "Power-On Reset Circuitry" on page 9-3 "Power-On Reset Specifications" on page 9-4 Stratix IV Hot-Socketing Specifications Stratix IV devices are hot-socketing compliant without the need for external components or special design requirements. Hot-socketing support in Stratix IV devices has the following advantages: "Stratix IV Devices can be Driven Before Power Up" on page 9-2 "I/O Pins Remain Tri-Stated During Power Up" on page 9-2 "Insertion or Removal of a Stratix IV Device from a Powered-Up System" on page 9-2 (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 9-2 Chapter 9: Hot Socketing and Power-On Reset in Stratix IV Devices Stratix IV Hot-Socketing Specifications Stratix IV Devices can be Driven Before Power Up You can drive signals into I/O pins, dedicated input pins, and dedicated clock pins of Stratix IV devices before or during power up or power down without damaging the device. I/O Pins Remain Tri-Stated During Power Up A device that does not support hot socketing can interrupt system operation or cause contention by driving out before or during power up. In a hot-socketing situation, the Stratix IV device's output buffers are turned off during system power up or power down. Also, the Stratix IV device does not drive out until the device is configured and working within the recommended operating conditions. Insertion or Removal of a Stratix IV Device from a Powered-Up System Devices that do not support hot socketing can short power supplies when powered up through the device signal pins. This irregular power up can damage both the driving and driven devices and can disrupt card power up. You can insert a Stratix IV device into or remove it from a powered-up system board without damaging the system board or interfering with its operation. You can power up or power down the VCCIO, VCC, VCCPGM, and VCCPD supplies in any sequence (with any time between them) which are monitored by the hot socket circuit. In addition, all other power supplies for the device can be powered up or down in any sequence. Individual power supply ramp-up and ramp-down rates range from 50 s to 100 ms. During hot socketing, the I/O pin capacitance is less than 15 pF and the clock pin capacitance is less than 20 pF. 1 To successfully power-up and exit POR on production devices, fully power VCC before VCCAUX begins to ramp. A possible concern regarding hot socketing is the potential for "latch-up." Stratix IV devices are immune to latch-up when hot socketing. Latch-up can occur when electrical subsystems are hot socketed into an active system. During hot socketing, the signal pins can be connected and driven by the active system before the power supply can provide current to the device's power and ground planes. This condition can lead to latch-up and cause a low-impedance path from power to ground within the device. As a result, the device draws a large amount of current, possibly causing electrical damage. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 9: Hot Socketing and Power-On Reset in Stratix IV Devices Hot-Socketing Feature Implementation in Stratix IV Devices 9-3 Hot-Socketing Feature Implementation in Stratix IV Devices The hot-socketing feature turns off the output buffer during power up and power down of the VCC, VCCAUX, VCCIO, VCCPGM, or VCCPD power supplies. The hot-socketing circuitry generates an internal HOTSCKT signal when the VCC, VCCAUX, VCCIO, VCCPGM, or VCCPD power supplies are below the threshold voltage. Hot-socketing circuitry is designed to prevent excess I/O leakage during power up. When the voltage ramps up very slowly, it is still relatively low, even after the POR signal is released and the configuration is completed. The CONF_DONE, nCEO, and nSTATUS pins fail to respond, as the output buffer cannot flip from the state set by the hot-socketing circuit at this low voltage. Therefore, the hot-socketing circuitry has been removed from these configuration pins to make sure that they are able to operate during configuration. Thus, it is expected behavior for these pins to drive out during power-up and power-down sequences. Figure 9-1 shows the Stratix IV device's I/O pin circuitry. Figure 9-1. Hot-Socketing Circuitry for Stratix IV Devices Power On Reset Monitor VCCIO Weak Pull-Up Resistor PAD R Output Enable Voltage Tolerance Control Hot Socket Output Pre-Driver Input Buffer to Logic Array The POR circuit monitors the voltage level of the power supplies (VCC, VCCAUX, VCCPT, VCCPGM, and VCCPD) and keeps the I/O pins tri-stated until the device is in user mode. The weak pull-up resistor (R) in the Stratix IV input/output element (IOE) keeps the I/O pins from floating. The 3.0-V tolerance control circuit permits the I/O pins to be driven by 3.0 V before the VCC, VCCAUX, VCCPT, VCCPGM, or VCCPD supplies are powered. It also prevents the I/O pins from driving out when the device is not in user mode. To successfully power-up and exit POR on production devices, fully power VCC before VCCAUX begins to ramp. 1 February 2011 Altera uses GND as a reference for hot-socketing operations and I/O buffer designs. To ensure proper operation, you must connect the GND between boards before connecting the power supplies. This prevents the GND on your board from being pulled up inadvertently by a path to power through other components on your board. A pulled up GND could otherwise cause an out-of-specification I/O voltage or current condition with the Altera device. Altera Corporation Stratix IV Device Handbook Volume 1 9-4 Chapter 9: Hot Socketing and Power-On Reset in Stratix IV Devices Power-On Reset Circuitry Power-On Reset Circuitry When power is applied to a Stratix IV device, a POR event occurs if the power supply reaches the recommended operating range within the maximum power supply ramp time (tRAMP). If tRAMP is not met, the device I/O pins and programming registers remain tri-stated, during which device configuration could fail. The maximum tRAMP for Stratix IV devices is 100 ms; the minimum tRAMP is 50 s. When the PORSEL pin is high, the maximum TRAMP for Stratix IV devices is 4 ms. Stratix IV devices provide a dedicated input pin (PORSEL) to select a POR delay time during power up. When the PORSEL pin is connected to GND, the POR delay time is 100 to 300 ms. When the PORSEL pin is set to high, the POR delay time is 4 to 12 ms. The POR block consists of a regulator POR, satellite POR, and main POR to check the power supply levels for proper device configuration. The satellite POR monitors the following: 1 VCCPD and VCCPGM power supplies that are used in the I/O buffers and for device programming VCCAUX power supply which is the auxiliary supply for the programmable power technology VCC and VCCPT power supplies that are used in the device core Altera requires powering up VCC before VCCAUX. The main POR waits for satellite POR and the regulator POR to release the POR signal. Until the release of the POR signal, the device configuration cannot start. The internal configuration memory supply that is used during device configuration is checked by the regulator POR block and is gated in the main POR block for the final POR trip. Figure 9-2 shows a simplified diagram of the POR block. 1 All configuration-related dedicated and dual function I/O pins must be powered by VCCPGM. Figure 9-2. Simplified POR Diagram for Stratix IV Devices Regulator POR Main POR V CCPGM V CCPD POR Pulse Setting V CC Satellite POR POR V CCPT V CCAUX Stratix IV Device Handbook Volume 1 PORSEL February 2011 Altera Corporation Chapter 9: Hot Socketing and Power-On Reset in Stratix IV Devices Power-On Reset Specifications 9-5 Power-On Reset Specifications Table 9-1 lists the power supplies that the POR circuit monitors. 1 Altera requires powering up VCC before VCCAUX. Table 9-1. Power Supplies Monitored by the POR Circuitry Power Supply Description Setting (V) VCC Core and periphery power supply 0.9 VCCPT Programmable power technology power supply 1.5 VCCPD I/O pre-driver power supply VCCPGM Configuration pins power supply VCCAUX Auxiliary supply for the programmable power technology 2.5, 3.0 1.8, 2.5, 3.0 2.5 Table 9-2 lists the power supplies that the POR circuit does not monitor. Table 9-2. Power Supplies Not Monitored by the POR Circuitry (Note 1) Power Supply Description Setting (V) 1.2, 1.5, 1.8, 2.5, 3.0 VCCIO I/O power supply VCCA_PLL PLL analog global power supply 2.5 VCCD_PLL PLL digital power supply 0.9 VCC_CLKIN PLL differential clock input power supply (top and bottom I/O banks only) 2.5 VCCBAT Battery back-up power supply for design security volatile key storage 1.2-3.3 Note to Table 9-2: (1) The transceiver supplies are not monitored by POR. 1 VCCIO, VCCA_PLL, VCCD_PLL, VCC_CLKIN, and VCCBAT are not monitored by POR and have no affect on the device configuration. The POR specification is designed to ensure that all the circuits in the Stratix IV device are at certain known states during power up. The POR signal pulse width is programmable using the PORSEL input pin. When the PORSEL pin is connected to GND, the POR delay time is 100 to 300 ms. When the PORSEL pin is set to high, the POR delay time is 4 to 12 ms. f For more information about the POR specification, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 9-6 Chapter 9: Hot Socketing and Power-On Reset in Stratix IV Devices Power-On Reset Specifications Document Revision History Table 9-3 lists the revision history for this chapter. Table 9-3. Document Revision History Date Version February 2011 March 2010 November 2009 June 2009 March 2009 November 2008 Changes Updated Table 9-2. Updated the "Power-On Reset Circuitry", "Power-On Reset Specifications", and "Insertion or Removal of a Stratix IV Device from a Powered-Up System" sections. Applied new template. Minor text edits. Updated the introduction and the "Stratix IV Hot-Socketing Specifications", "Insertion or Removal of a Stratix IV Device from a Powered-Up System", "Hot-Socketing Feature Implementation in Stratix IV Devices", "Power-On Reset Circuitry", and "Power-On Reset Specifications" sections. Updated Table 9-1 and Table 9-2. Updated Figure 9-2. Minor text edits. Updated graphics. Minor text edits. Updated Table 9-2. Added introductory sentences to improve search ability. Removed the Conclusion section. Minor text edits. Changed all "Stratix IV E" to "Stratix IV". Updated "Stratix IV Hot-Socketing Specifications" and "Hot-Socketing Feature Implementation in Stratix IV Devices" sections. Updated Figure 9-2. Removed "Referenced Documents" section. Updated "Hot-Socketing Feature Implementation in Stratix IV Devices" on page 9-2. Updated "Power-On Reset Circuitry" on page 9-4. Updated Table 9-1. Made minor editorial changes. 3.2 3.1 3.0 2.2 2.1 2.0 July 2008 1.1 Revised "Introduction". May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation 10. Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices September 2012 SIV51010-3.5 SIV51010-3.5 This chapter describes the configuration, design security, and remote system upgrades in Stratix(R) IV devices. To save configuration memory space and time, Stratix IV devices provide configuration data decompression. They also provide a built-in design security feature that protects your designs against IP theft and tampering of your configuration files. Stratix IV devices also offer remote system upgrade capability so that you can upgrade your system in real-time through any network. This helps to deliver feature enhancements and bug fixes and provides error detection, recovery, and status information to ensure reliable reconfiguration. Overview This chapter describes supported configuration schemes for Stratix IV devices, instructions about how to execute the required configuration schemes, and the necessary pin settings. Stratix IV devices use SRAM cells to store configuration data. As SRAM is volatile, you must download configuration data to the Stratix IV device each time the device powers up. You can configure Stratix IV devices using one of four configuration schemes: Fast passive parallel (FPP) Fast active serial (AS) Passive serial (PS) Joint Test Action Group (JTAG) All configuration schemes use either an external controller (for example, a MAX(R) II device or microprocessor), a configuration device, or a download cable. For more information, refer to "Configuration Features" on page 10-4. This chapter includes the following sections: "Configuration Schemes" on page 10-2 "Configuration Features" on page 10-4 "Fast Passive Parallel Configuration" on page 10-6 "Fast Active Serial Configuration (Serial Configuration Devices)" on page 10-16 "Passive Serial Configuration" on page 10-24 "JTAG Configuration" on page 10-34 "Device Configuration Pins" on page 10-39 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 September 2012 Feedback Subscribe 10-2 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Schemes "Configuration Data Decompression" on page 10-47 "Remote System Upgrades" on page 10-49 "Remote System Upgrade Mode" on page 10-53 "Dedicated Remote System Upgrade Circuitry" on page 10-56 "Quartus II Software Support" on page 10-62 "Design Security" on page 10-63 Configuration Devices Altera(R) serial configuration devices support a single-device and multi-device configuration solution for Stratix IV devices and are used in the fast AS configuration scheme. Serial configuration devices offer a low-cost, low pin-count configuration solution. f For information about serial configuration devices, refer to the Serial Configuration Devices (EPCS1, EPCS4, EPCS16, EPCS64, and EPCS128) Data Sheet in volume 2 of the Configuration Handbook. 1 All minimum timing information in this chapter covers the entire Stratix IV family. Some devices may work at less than the minimum timing stated in this handbook due to process variation. Configuration Schemes Select the configuration scheme by driving the Stratix IV device MSEL pins either high or low, as shown in Table 10-1. The MSEL input buffers are powered by the VCC power supply. Altera recommends hard wiring the MSEL[] pins to VCCPGM and GND. The MSEL[2..0] pins have 5-k internal pull-down resistors that are always active. During power-on reset (POR) and during reconfiguration, the MSEL pins must be at VIL and VIH levels of VCCPGM voltage to be considered logic low and logic high. 1 To avoid problems with detecting an incorrect configuration scheme, hardwire the MSEL[] pins to VCCPGM and GND without pull-up or pull-down resistors. Do not drive the MSEL[] pins by a microprocessor or another device. Table 10-1. Configuration Schemes for Stratix IV Devices (Part 1 of 2) Configuration Scheme MSEL2 MSEL1 MSEL0 Fast passive parallel 0 0 0 Passive serial 0 1 0 Fast AS (40 MHz) (1) 0 1 1 Remote system upgrade fast AS (40 MHz) (1) 0 1 1 FPP with design security feature and/or decompression enabled (2) 0 0 1 Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Schemes 10-3 Table 10-1. Configuration Schemes for Stratix IV Devices (Part 2 of 2) Configuration Scheme JTAG-based configuration (4) MSEL2 MSEL1 MSEL0 (3) (3) (3) Notes to Table 10-1: (1) Stratix IV devices only support fast AS configuration. You must use either EPCS64 or EPCS128 devices to configure a Stratix IV device in fast AS mode. (2) These modes are only supported when using a MAX II device or a microprocessor with flash memory for configuration. In these modes, the host system must output a DCLK that is x4 the data rate. (3) Do not leave the MSEL pins floating, connect them to VCCPGM or GND. These pins support the non-JTAG configuration scheme used in production. If you only use the JTAG configuration, connect the MSEL pins to GND. (4) The JTAG-based configuration takes precedence over other configuration schemes, which means the MSEL pin settings are ignored. The JTAG-based configuration does not support the design security or decompression features. Table 10-2 lists the uncompressed raw binary file (.rbf) configuration file sizes for Stratix IV devices. Table 10-2. Uncompressed Raw Binary File (.rbf) Sizes for Stratix IV Devices Device Data Size (Bits) EP4SE230 94,557,472 EP4SE360 128,395,584 EP4SE530 171,722,064 EP4SE820 241,684,472 EP4SGX70 47,833,352 EP4SGX110 47,833,352 EP4SGX180 94,557,472 EP4SGX230 94,557,472 EP4SGX290 EP4SGX360 128,395,584 171,722,064 (1) 128,395,584 171,722,064 (1) EP4SGX530 171,722,064 EP4S40G2 94,557,472 EP4S40G5 171,722,064 EP4S100G2 94,557,472 EP4S100G3 171,722,064 EP4S100G4 171,722,064 EP4S100G5 171,722,064 Note to Table 10-2: (1) This only applies to the F45 package. Use the data in Table 10-2 to estimate the file size before design compilation. Different configuration file formats; for example, a hexidecimal (.hex) or tabular text file (.ttf) format, have different file sizes. Refer to the Quartus(R) II software for the different types of configuration file and file sizes. However, for any specific version of the Quartus II software, any design targeted for the same device has the same uncompressed configuration file size. If you are using compression, the file size can vary after each compilation because the compression ratio depends on the design. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-4 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Features f For more information about setting device configuration options or creating configuration files, refer to the Device Configuration Options and Configuration File Formats chapters in volume 2 of the Configuration Handbook. Configuration Features Stratix IV devices offer design security, decompression, and remote system upgrade features. Design security using configuration bitstream encryption is available in Stratix IV devices, which protects your designs. Stratix IV devices can receive a compressed configuration bitstream and decompress this data in real-time, reducing storage requirements and configuration time. You can make real-time system upgrades from remote locations of your Stratix IV designs with the remote system upgrade feature. Table 10-3 lists which configuration features you can use in each configuration scheme. Table 10-3. Configuration Features for Stratix IV Devices Configuration Scheme Configuration Method Decompression Design Security Remote System Upgrade FPP MAX II device or a microprocessor with flash memory Y (1) Y (1) Y Fast AS Serial configuration device Y Y Y MAX II device or a microprocessor with flash memory Y Y -- Download cable Y Y -- MAX II device or a microprocessor with flash memory -- -- -- Download cable -- -- -- PS JTAG Note to Table 10-3: (1) In these modes, the host system must send a DCLK that is x4 the data rate. You can also refer to the following: For more information about the configuration data decompression feature, refer to "Configuration Data Decompression" on page 10-47. For more information about the remote system upgrade feature, refer to "Remote System Upgrades" on page 10-49. For more information about the design security feature, refer to "Design Security" on page 10-63. If your system already contains a common flash interface (CFI) flash memory, you can use it for Stratix IV device configuration storage as well. The MAX II parallel flash loader (PFL) feature in MAX II devices provides an efficient method to program CFI flash memory devices through the JTAG interface and provides the logic to control configuration from the flash memory device to the Stratix IV device. Both PS and FPP configuration modes are supported using this PFL feature. f For more information about PFL, refer to Parallel Flash Loader Megafunction User Guide. For more information about programming Altera serial configuration devices, refer to "Programming Serial Configuration Devices" on page 10-22. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Features 10-5 Power-On Reset Circuit The POR circuit keeps the entire system in reset until the power supply voltage levels have stabilized on power-up. After power-up, the device does not release nSTATUS until VCC, VCCAUX, VCCPT, VCCPGM, and VCCPD are above the device's POR trip point. On power down, brown-out occurs if the VCC, VCCAUX, VCCPT, VCCPGM, or VCCPD drops below the threshold voltage. In Stratix IV devices, a pin-selectable option (PORSEL) is provided that allows you to select between the standard POR time or fast POR time. When PORSEL is driven low, the standard POR time is 100 ms < TPOR < 300 ms, which has a lower power-ramp rate. When PORSEL is driven high, the fast POR time is 4 ms < TPOR < 12 ms. VCCPGM Pins Stratix IV devices have a power supply, VCCPGM, for all the dedicated configuration pins and dual function pins. The supported configuration voltage is 1.8, 2.5, and 3.0 V. Stratix IV devices do not support 1.5 V configuration. Use the VCCPGM pin to power all dedicated configuration inputs, dedicated configuration outputs, dedicated configuration bidirectional pins, and some of the dual functional pins that you use for configuration. With VCCPGM, the configuration input buffers do not have to share power lines with the regular I/O buffer in Stratix IV devices. The operating voltage for the configuration input pin is independent of the I/O banks power supply VCCIO during configuration. Therefore, Stratix IV devices do not need configuration voltage constraints on VCCIO . VCCPD Pins Stratix IV devices have a dedicated programming power supply, VCCPD, which must be connected to 3.0 V/2.5 V to power the I/O pre-drivers and JTAG I/O pins (TCK, TMS, TDI, TDO, and TRST). 1 VCCPGM and VCCPD must ramp up from 0 V to the desired voltage level within 100 ms when PORSEL is low or 4 ms when PORSEL is high. If these supplies are not ramped up within this specified time, your Stratix IV device will not configure successfully. If your system cannot ramp up the power supplies within 100 ms or 4 ms, you must hold nCONFIG low until all the power supplies are stable. 1 VCCPD must be greater than or equal to VCCIO of the same bank. If VCCIO of the bank is set to 3.0 V, VCCPD must be powered up to 3.0 V. If the VCCIO of the bank is powered to 2.5 V or lower, VCCPD must be powered up to 2.5 V. For more information about configuration pins power supply, refer to "Device Configuration Pins" on page 10-39. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-6 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration Fast Passive Parallel Configuration Fast passive parallel configuration in Stratix IV devices is designed to meet the continuously increasing demand for faster configuration times. Stratix IV devices are designed with the capability of receiving byte-wide configuration data per clock cycle. You can perform FPP configuration of Stratix IV devices using an intelligent host, such as a MAX II device or a microprocessor. FPP Configuration Using a MAX II Device as an External Host FPP configuration using an external host provides the fastest method to configure Stratix IV devices. In this configuration scheme, you can use a MAX II device as an intelligent host that controls the transfer of configuration data from a storage device, such as flash memory, to the target Stratix IV device. You can store configuration data in .rbf, .hex, or .ttf format. When using the MAX II device as an intelligent host, a design that controls the configuration process, such as fetching the data from flash memory and sending it to the device, must be stored in the MAX II device. 1 If you are using the Stratix IV decompression and/or design security features, the external host must be able to send a DCLK frequency that is x4 the data rate. The x4 DCLK signal does not require an additional pin and is sent on the DCLK pin. The maximum DCLK frequency is 125 MHz, which results in a maximum data rate of 250 Mbps. If you are not using the Stratix IV decompression or design security features, the data rate is x8 the DCLK frequency. Figure 10-1 shows the configuration interface connections between the Stratix IV device and a MAX II device for single device configuration. Figure 10-1. Single Device FPP Configuration Using an External Host Memory ADDR DATA[7..0] VCCPGM (1) VCCPGM (1) VCCPGM/VCCIO (2) 10 k 10 k 10 k Stratix IV Device MSEL[2..0] CONF_DONE GND nSTATUS External Host (MAX II Device or Microprocessor) nCE GND nCEO N.C. DATA[7..0] nCONFIG DCLK Note to Figure 10-1: (1) Connect the resistor to a supply that provides an acceptable input signal for the Stratix IV device. VCCPGM must be high enough to meet the VIH specification of the I/O on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration 10-7 After power-up, the Stratix IV device goes through a POR. The POR delay depends on the PORSEL pin setting. When PORSEL is driven low, the standard POR time is 100 ms < TPOR < 300 ms. When PORSEL is driven high, the fast POR time is 4 ms < TPOR < 12 ms. During POR, the device resets, holds nSTATUS low, and tri-states all user I/O pins. After the device successfully exits POR, all user I/O pins continue to be tri-stated. If nIO_pullup is driven low during power up and configuration, the user I/O pins and dual-purpose I/O pins have weak pull-up resistors, which are on (after POR) before and during configuration. If nIO_pullup is driven high, the weak pull-up resistors are disabled. The configuration cycle consists of three stages: reset, configuration, and initialization. While nCONFIG or nSTATUS are low, the device is in the reset stage. To initiate configuration, the MAX II device must drive the nCONFIG pin from low to high. 1 To begin the configuration process, you must fully power VCCPT, VCC, VCCPD, and VCCPGM of the banks where the configuration pins reside to the appropriate voltage levels. When nCONFIG goes high, the device comes out of reset and releases the open-drain nSTATUS pin, which is then pulled high by an external 10-k pull-up resistor. After nSTATUS is released, the device is ready to receive configuration data and the configuration stage begins. When nSTATUS is pulled high, the MAX II device places the configuration data one byte at a time on the DATA[7..0] pins. 1 A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host (a Max II CPLD or a microcontroller) is not driving the line. For example, during external host reprogramming or power-up where the I/O driving nCONFIG may be tri-stated. If a pull-up resistor is added to the nCONFIG line, the FPGA stays in user mode if the external host is being reprogrammed. If a pull-down resistor is added to the nCONFIG line, the FPGA goes into reset mode if the external host is being reprogrammed. Whenever the nCONFIG line is released high, ensure that the first DCLK and DATA are not driven unintentionally. 1 Stratix IV devices receive configuration data on the DATA[7..0] pins and the clock is received on the DCLK pin. Data is latched into the device on the rising edge of DCLK. If you are using the Stratix IV decompression and/or design security features, configuration data is latched on the rising edge of every first DCLK cycle out of the four DCLK cycles. Altera recommends that you to keep the data on DATA[7. . 0] stable for the next 3 clock cycles when the data is being processed. You can only stop DCLK after three clock cycles after the last data is latched. Data is continuously clocked into the target device until CONF_DONE goes high. The CONF_DONE pin goes high one byte early in FPP modes. The last byte is required for FPP mode. After the device has received the next-to-last byte of the configuration data successfully, it releases the open-drain CONF_DONE pin, which is pulled high by an external 10-kpull-up resistor. A low-to-high transition on CONF_DONE indicates configuration is complete and initialization of the device can begin. The CONF_DONE pin must have an external 10-k pull-up resistor for the device to initialize. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-8 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration In Stratix IV devices, the initialization clock source is either the internal oscillator or the optional CLKUSR pin. By default, the internal oscillator is the clock source for initialization. If you use the internal oscillator, the Stratix IV device provides itself with enough clock cycles for proper initialization. Therefore, if the internal oscillator is the initialization clock source, sending the entire configuration file to the device is sufficient to configure and initialize the device. Driving DCLK to the device after configuration is complete does not affect device operation. You can also synchronize initialization of multiple devices or delay initialization with the CLKUSR option. You can turn on the Enable user-supplied start-up clock (CLKUSR) option in the Quartus II software from the General tab of the Device and Pin Options dialog box. Supplying a clock on CLKUSR does not affect the configuration process. The CONF_DONE pin goes high one byte early in FPP modes. The last byte is required for FPP mode. After the CONF_DONE pin transitions high, CLKUSR is enabled after the time specified at tCD2CU. After this time period elapses, Stratix IV devices require 8,532 clock cycles to initialize properly and enter user mode. Stratix IV devices support a CLKUSR fMAX of 125 MHz. An optional INIT_DONE pin is available, which signals the end of initialization and the start of user-mode with a low-to-high transition. This Enable INIT_DONE Output option is available in the Quartus II software from the General tab of the Device and Pin Options dialog box. If you use the INIT_DONE pin, it is high because of an external 10-k pull-up resistor when nCONFIG is low and during the beginning of configuration. After the option bit to enable INIT_DONE is programmed into the device (during the first frame of configuration data), the INIT_DONE pin goes low. When initialization is complete, the INIT_DONE pin is released and pulled high. The MAX II device must be able to detect this low-to-high transition, which signals the device has entered user mode. When initialization is complete, the device enters user mode. In user-mode, the user I/O pins no longer have weak pull-up resistors and function as assigned in your design. 1 Two DCLK falling edges are required after CONF_DONE goes high to begin the initialization of the device for both uncompressed and compressed bitstream in FPP. To ensure DCLK and DATA[7..0] are not left floating at the end of configuration, the MAX II device must drive them either high or low, whichever is convenient on your board. The DATA[7..0] pins are available as user I/O pins after configuration. When you select the FPP scheme as a default in the Quartus II software, these I/O pins are tri-stated in user mode. To change this default option in the Quartus II software, select the Dual-Purpose Pins tab of the Device and Pin Options dialog box. The configuration clock (DCLK) speed must be below the specified frequency to ensure correct configuration. No maximum DCLK period exists, which means you can pause configuration by halting DCLK for an indefinite amount of time. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration 1 10-9 If you need to stop DCLK, it can only be stopped: three clock cycles after the last data byte was latched into the Stratix IV device when you use the decompression and/or design security features. two clock cycles after the last data byte was latched into the Stratix IV device when you do not use the Stratix IV decompression and/or design security features. By stopping DCLK, the configuration circuit allows enough clock cycles to process the last byte of latched configuration data. When the clock restarts, the MAX II device must provide data on the DATA[7..0] pins prior to sending the first DCLK rising edge. If an error occurs during configuration, the device drives its nSTATUS pin low, resetting itself internally. The low signal on the nSTATUS pin also alerts the MAX II device that there is an error. If the Auto-restart configuration after error option (available in the Quartus II software from the General tab of the Device and Pin Options dialog box) is turned on, the device releases nSTATUS after a reset time-out period (a maximum of 500 s). After nSTATUS is released and pulled high by a pull-up resistor, the MAX II device can try to reconfigure the target device without needing to pulse nCONFIG low. If this option is turned off, the MAX II device must generate a low-to-high transition (with a low pulse of at least 2 s) on nCONFIG to restart the configuration process. 1 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. The MAX II device can also monitor the CONF_DONE and INIT_DONE pins to ensure successful configuration. The MAX II device must monitor the CONF_DONE pin to detect errors and determine when programming completes. If all the configuration data is sent, but the CONF_DONE or INIT_DONE signals have not gone high, the MAX II device reconfigures the target device. 1 If you use the optional CLKUSR pin and nCONFIG is pulled low to restart the configuration during device initialization, ensure that CLKUSR continues toggling during the time nSTATUS is low (a maximum of 500 s). When the device is in user mode, initiating reconfiguration is done by transitioning the nCONFIG pin low-to-high. The nCONFIG pin must be low for at least 2 s. When nCONFIG is pulled low, the device also pulls nSTATUS and CONF_DONE low and all I/O pins are tri-stated. After nCONFIG returns to a logic high level and nSTATUS is released by the device, reconfiguration begins. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-10 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration Figure 10-2 shows how to configure multiple Stratix IV devices using a MAX II device. This circuit is similar to the FPP configuration circuit for a single device, except the devices are cascaded for multi-device configuration. Figure 10-2. Multi-Device FPP Configuration Using an External Host Memory ADDR DATA[7..0] VCCPGM (1) VCCPGM (1) VCCPGM/VCCIO (2) 10 k 10 k 10 k Stratix IV Device 2 Stratix IV Device 1 MSEL[2..0] MSEL[2..0] CONF_DONE CONF_DONE GND nSTATUS External Host (MAX II Device or Microprocessor) nCE nCEO GND nSTATUS nCE nCEO N.C. GND DATA[7..0] DATA[7..0] nCONFIG nCONFIG DCLK DCLK Note to Figure 10-2: (1) Connect the pull-up resistor to a supply that provides an acceptable input signal for all Stratix IV devices in the chain. VCCPGM must be high enough to meet the VIH specification of the I/O standard on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. In a multi-device FPP configuration, the first device's nCE pin is connected to GND while its nCEO pin is connected to nCE of the next device in the chain. The last device's nCE input comes from the previous device, while its nCEO pin is left floating. After the first device completes configuration in a multi-device configuration chain, its nCEO pin drives low to activate the second device's nCE pin, which prompts the second device to begin configuration. The second device in the chain begins configuration within one clock cycle; therefore, the transfer of data destinations is transparent to the MAX II device. All other configuration pins (nCONFIG, nSTATUS, DCLK, DATA[7..0], and CONF_DONE) are connected to every device in the chain. The configuration signals may require buffering to ensure signal integrity and prevent clock skew problems. Ensure that the DCLK and DATA lines are buffered for every fourth device. Because all device CONF_DONE pins are tied together, all devices initialize and enter user mode at the same time. All nSTATUS and CONF_DONE pins are tied together; if any device detects an error, configuration stops for the entire chain and you must reconfigure the entire chain. For example, if the first device flags an error on nSTATUS, it resets the chain by pulling its nSTATUS pin low. This behavior is similar to a single device detecting an error. If the Auto-restart configuration after error option is turned on, the devices release their nSTATUS pins after a reset time-out period (a maximum of 500 s). After all nSTATUS pins are released and pulled high, the MAX II device tries to reconfigure the chain without pulsing nCONFIG low. If this option is turned off, the MAX II device must generate a low-to-high transition (with a low pulse of at least 2 s) on nCONFIG to restart the configuration process. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration 1 10-11 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. In a multi-device FPP configuration chain, all Stratix IV devices in the chain must either enable or disable the decompression and/or design security features. You cannot selectively enable the decompression and/or design security features for each device in the chain because of the DATA and DCLK relationship. If the chain contains devices that do not support design security, use a serial configuration scheme. If a system has multiple devices that contain the same configuration data, tie all device nCE inputs to GND and leave the nCEO pins floating. All other configuration pins (nCONFIG, nSTATUS, DCLK, DATA[7..0], and CONF_DONE) are connected to every device in the chain. Configuration signals may require buffering to ensure signal integrity and prevent clock skew problems. Ensure that the DCLK and DATA lines are buffered for every fourth device. Devices must be the same density and package. All devices start and complete configuration at the same time. Figure 10-3 shows a multi-device FPP configuration when both Stratix IV devices are receiving the same configuration data. Figure 10-3. Multiple-Device FPP Configuration Using an External Host When Both Devices Receive the Same Data Memory ADDR DATA[7..0] VCCPGM (1) VCCPGM (1) VCCPGM/VCCIO (2) 10 k 10 k Stratix IV Device 10 k Stratix IV Device MSEL[2..0] MSEL[2..0] GND CONF_DONE nSTATUS nCE External Host (MAX II Device or Microprocessor) GND nCEO GND CONF_DONE nSTATUS nCE N.C. (3) nCEO N.C. GND DATA[7..0] DATA[7..0] nCONFIG nCONFIG DCLK DCLK Notes to Figure 10-3: (1) Connect the resistor to a supply that provides an acceptable input signal for all Stratix IV devices in the chain. VCCPGM must be high enough to meet the VIH specification of the I/O on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. (3) The nCEO pins of both Stratix IV devices are left unconnected when configuring the same configuration data into multiple devices. You can use a single configuration chain to configure Stratix IV devices with other Altera devices that support FPP configuration, such as other types of Stratix devices. To ensure that all devices in the chain complete configuration at the same time, or that an error flagged by one device initiates reconfiguration in all devices, tie all of the device CONF_DONE and nSTATUS pins together. f For more information about configuring multiple Altera devices in the same configuration chain, refer to the Configuring Mixed Altera FPGA Chains in volume 2 of the Configuration Handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-12 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration FPP Configuration Timing Figure 10-4 shows the timing waveform for an FPP configuration when using a MAX II device as an external host. This waveform shows the timing when you have not enabled the decompression and design security features. Figure 10-4. FPP Configuration Timing Waveform (Note 1), (2) tCF2ST1 tCFG tCF2CK nCONFIG nSTATUS (3) tSTATUS tCF2ST0 t CLK CONF_DONE (4) tCF2CD tST2CK tCH tCL (5) DCLK tDH DATA[7..0] (6) Byte 0 Byte 1 Byte 2 Byte 3 Byte n-2 Byte n-1 User Mode Byte n tDSU User I/O High-Z User Mode INIT_DONE tCD2UM Notes to Figure 10-4: (1) Use this timing waveform when you have not enabled the decompression and design security features. (2) The beginning of this waveform shows the device in user mode. In user mode, nCONFIG, nSTATUS, and CONF_DONE are at logic high levels. When nCONFIG is pulled low, a reconfiguration cycle begins. (3) After power-up, the Stratix IV device holds nSTATUS low for the time of the POR delay. (4) After power-up, before and during configuration, CONF_DONE is low. (5) Do not leave DCLK floating after configuration. You can drive it high or low, whichever is more convenient. (6) DATA[7..0] are available as user I/O pins after configuration except for some exceptions on Stratix IV GT devices. The state of these pins depends on the dual-purpose pin settings. Table 10-4 lists the timing parameters for Stratix IV devices for an FPP configuration when you have not enabled the decompression and design security features. Table 10-4. FPP Timing Parameters for Stratix IV Devices (Part 1 of 2) (Note 1), (2) Minimum Symbol Parameter Stratix IV (7) Stratix IV (8) Maximum Stratix IV (9) Stratix IV (7) Stratix IV (8) Stratix IV (9) Units tCF2CD nCONFIG low to CONF_DONE low -- 800 ns tCF2ST0 nCONFIG low to nSTATUS low -- 800 ns tCFG nCONFIG low pulse width 2 -- s tSTATUS nSTATUS low pulse width 10 500 (3) s tCF2ST1 nCONFIG high to nSTATUS high -- 500 (4) s tCF2CK nCONFIG high to first rising edge on DCLK 500 -- s Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration 10-13 Table 10-4. FPP Timing Parameters for Stratix IV Devices (Part 2 of 2) (Note 1), (2) Minimum Symbol Parameter Stratix IV (7) Stratix IV (8) Maximum Stratix IV (9) Stratix IV (7) Stratix IV (8) Stratix IV (9) Units tST2CK nSTATUS high to first rising edge of DCLK 2 -- s tDSU Data setup time before rising edge on DCLK 4 -- ns tDH Data hold time after rising edge on DCLK 1 -- ns TR Input rise time -- 40 ns t Input fall time -- 40 ns tCD2UM CONF_DONE high to user mode (5) 55 150 s tCD2CU CONF_DONE high to CLKUSR enabled -- -- tCD2UMC CONF_DONE high to user mode with CLKUSR option on -- -- tCH DCLK high time (6) 3.6 4.5 5.6 -- ns tCL DCLK low time (6) 3.6 4.5 5.6 -- ns tCLK DCLK period (6) 8 10 12.5 -- ns fMAX DCLK frequency 4 x maximum DCLK period tCD2CU + (8532 x CLKUSR period) -- 125 100 80 MHz Notes to Table 10-4: (1) This information is preliminary. (2) Use these timing parameters when you have not enabled the decompression and design security features. (3) You can obtain this value if you do not delay the configuration by extending the nCONFIG or nSTATUS low pulse width. (4) This value is applicable if you do not delay the configuration by externally holding nSTATUS low. (5) The minimum and maximum numbers apply only if you chose the internal oscillator as the clock source for starting the device. (6) Adding up tCH and tCL equals to tCLK. When EP4SE230 tCH is 3.6 ns (min), tCL must be 4.4 ns and vice versa. (7) Applicable to EP4SE230, EP4SE360, EP4SGX70, EP4SGX110, EP4SGX180, EP4SGX230, EP4SGX290 (except F45 package), EP4SGX360 (except F45 package), EP4S40G2, EP4S100G2 devices. (8) Applicable to EP4SE530, EP4SGX290 (only for F45 package), EP4SGX360 (only for F45 package), EP4SGX530, EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5 devices. (9) Applicable to EP4SE820 only. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-14 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration Figure 10-5 shows the timing waveform for an FPP configuration when using a MAX II device as an external host. This waveform shows the timing when you have enabled the decompression and/or design security features. Figure 10-5. FPP Configuration Timing Waveform with Decompression or Design Security Feature Enabled (Note 1), (2) tCF2ST1 tCFG tCF2CK nCONFIG nSTATUS (3) CONF_DONE (4) tSTATUS tCF2ST0 tCF2CD DCLK tCL tST2CK tCH 1 2 3 4 1 2 3 (7) 4 1 3 (5) 4 tCLK DATA[7..0] Byte 0 Byte 1 tDH tDH tDSU Byte 2 Byte (n-1) Byte n User Mode User Mode High-Z User I/O (6) INIT_DONE tCD2UM Notes to Figure 10-5: (1) Use this timing waveform when you have enabled the decompression and/or design security features. (2) The beginning of this waveform shows the device in user-mode. In user-mode, nCONFIG, nSTATUS, and CONF_DONE are at logic high levels. When nCONFIG is pulled low, a reconfiguration cycle begins. (3) After power-up, the Stratix IV device holds nSTATUS low for the time of the POR delay. (4) After power-up, before and during configuration, CONF_DONE is low. (5) Do not leave DCLK floating after configuration. You can drive it high or low, whichever is more convenient. (6) DATA[7..0] are available as user I/O pins after configuration except for some exceptions on Stratix IV GT devices. The state of these pins depends on the dual-purpose pin settings. (7) If needed, you can pause DCLK by holding it low. When DCLK restarts, the external host must provide data on the DATA[7..0] pins prior to sending the first DCLK rising edge. Table 10-5 lists the timing parameters for Stratix IV devices for an FPP configuration when you enable the decompression and/or the design security features. Table 10-5. FPP Timing Parameters for Stratix IV Devices with the Decompression and/or Design Security Features Enabled (Note 1), (2) (Part 1 of 2) Minimum Symbol Parameter Stratix IV (7) Stratix IV (8) Maximum Stratix IV (9) Stratix IV (7) Stratix IV (8) Stratix IV (9) Units tCF2CD nCONFIG low to CONF_DONE low -- 800 ns tCF2ST0 nCONFIG low to nSTATUS low -- 800 ns tCFG nCONFIG low pulse width 2 -- s tSTATUS nSTATUS low pulse width 10 500 (3) s tCF2ST1 nCONFIG high to nSTATUS high -- 500 (4) s tCF2CK nCONFIG high to first rising edge on DCLK 500 -- s Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Passive Parallel Configuration 10-15 Table 10-5. FPP Timing Parameters for Stratix IV Devices with the Decompression and/or Design Security Features Enabled (Note 1), (2) (Part 2 of 2) Minimum Symbol Parameter Stratix IV (7) Stratix IV (8) Maximum Stratix IV (9) Stratix IV (7) Stratix IV (8) Stratix IV (9) Units tST2CK nSTATUS high to first rising edge of DCLK 2 -- s tDSU Data setup time before rising edge on DCLK 4 -- ns tDH Data hold time after rising edge on DCLK 3/(DCLK frequency) + 1 -- s tDATA Data rate -- 250 Mbps tR Input rise time -- 40 ns t Input fall time -- 40 ns tCD2UM CONF_DONE high to user mode (5) 55 150 s tCD2CU CONF_DONE high to CLKUSR enabled -- -- tCD2UMC CONF_DONE high to user mode with CLKUSR option on (5) -- -- tCH DCLK high time (6) 3.6 4.5 5.6 -- ns tCL DCLK low time (6) 3.6 4.5 5.6 -- ns tCLK DCLK period (6) 8 10 12.5 -- ns fMAX DCLK frequency 4 x maximum DCLK period tCD2CU + (8532 x CLKUSR period) -- 125 100 80 MHz Notes to Table 10-5: (1) This information is preliminary. (2) Use these timing parameters when you use the decompression and/or design security features. (3) You can obtain this value if you do not delay the configuration by extending the nCONFIG or nSTATUS low pulse width. (4) This value is applicable if you do not delay the configuration by externally holding nSTATUS low. (5) The minimum and maximum numbers apply only if you chose the internal oscillator as the clock source for starting the device. (6) Adding up tCH and tCL equals to tCLK. When EP4SE230 tCH is 3.6 ns (min), tCL must be 4.4 ns and vice versa. (7) Applicable for EP4SE230, EP4SE360, EP4SGX70, EP4SGX110, EP4SGX180, EP4SGX230, EP4SGX290 (except F45 package), EP4SGX360 (except F45 package), EP4S40G2, EP4S100G2 devices. (8) Applicable for EP4SE530, EP4SGX290 (only for F45 package), EP4SGX360 (only for F45 package), EP4SGX530, EP4S40G5, EP4S100G3, EP4S100G4, EP4S100G5 devices. (9) Applicable to EP4SE820 only. f For more information about device configuration options and how to create configuration files, refer to the Device Configuration Options and Configuration File Formats chapters in volume 2 of the Configuration Handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-16 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) FPP Configuration Using a Microprocessor In this configuration scheme, a microprocessor can control the transfer of configuration data from a storage device, such as flash memory, to the target Stratix IV device. All information in "FPP Configuration Using a MAX II Device as an External Host" on page 10-6 is also applicable when using a microprocessor as an external host. Refer to this section for all configuration and timing information. Fast Active Serial Configuration (Serial Configuration Devices) In the fast AS configuration scheme, Stratix IV devices are configured using a serial configuration device. These configuration devices are low-cost devices with non-volatile memory that feature a simple four-pin interface and a small form factor. The largest serial configuration device currently supports 128 MBits of configuration bitstream. Use the Stratix IV decompression features or select an FPP or PS configuration scheme for EP4SE360, EP4SGX290, EP4S40G5, EP4S100G3 and larger devices. f For more information about serial configuration devices, refer to the Serial Configuration Devices (EPCS1, EPCS4, EPCS16, EPCS64, and EPCS128) Data Sheet chapter in volume 2 of the Configuration Handbook. Serial configuration devices provide a serial interface to access configuration data. During device configuration, Stratix IV devices read configuration data using the serial interface, decompress data if necessary, and configure their SRAM cells. This scheme is referred to as the AS configuration scheme because the Stratix IV device controls the configuration interface. This scheme contrasts with the PS configuration scheme where the configuration device controls the interface. 1 Stratix IV Device Handbook Volume 1 The Stratix IV decompression and design security features are fully available when configuring your Stratix IV device using fast AS mode. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) 10-17 Serial configuration devices have a four-pin interface--serial clock input (DCLK), serial data output (DATA), AS data input (ASDI), and an active-low chip select (nCS). This four-pin interface connects to Stratix IV device pins, as shown in Figure 10-6. Figure 10-6. Single Device Fast AS Configuration VCCPGM (1) VCCPGM (1) VCCPGM (1) 10 k 10 k 10 k Serial Configuration Device Stratix IV Device nSTATUS CONF_DONE nCONFIG nCE nCEO GND VCCPGM DATA DATA0 DCLK DCLK MSEL2 nCS nCSO MSEL1 ASDO MSEL0 ASDI (2) N.C. GND Notes to Figure 10-6: (1) Connect the pull-up resistors to VCCPGM at a 3.0-V supply. (2) Stratix IV devices use the ASDO-to-ASDI path to control the configuration device. You can power the EPCS serial configuration device with 3.0 V when you configure the Stratix IV FPGA using Active Serial (AS) configuration mode. This is feasible because the power supply to the EPCS device ranges between 2.7 V and 3.6 V. You do not need a dedicated 3.3 V power supply to power the EPCS device. The EPCS device and the VCCPGM pins on the Stratix IV device may share the same 3.0 V power supply. After power-up, the Stratix IV devices go through a POR. The POR delay depends on the PORSEL pin setting. When PORSEL is driven low, the standard POR time is 100 ms < TPOR < 300 ms. When PORSEL is driven high, the fast POR time is 4 ms < TPOR < 12 ms. During POR, the device resets, holds nSTATUS and CONF_DONE low, and tri-states all user I/O pins. After the device successfully exits POR, all the user I/O pins continue to be tri-stated. If nIO_pullup is driven low during power-up and configuration, the user I/O pins and dual-purpose I/O pins will have weak pull-up resistors, which are on (after POR) before and during configuration. If nIO_pullup is driven high, the weak pull-up resistors are disabled. The configuration cycle consists of three stages--reset, configuration, and initialization. While nCONFIG or nSTATUS are low, the device is in reset. After POR, the Stratix IV device releases nSTATUS, which is pulled high by an external 10-k pull-up resistor and enters configuration mode. 1 To begin configuration, power the VCC, VCCIO, VCCPGM, and VCCPD voltages (for the banks where the configuration pins reside) to the appropriate voltage levels. The serial clock (DCLK) generated by the Stratix IV device controls the entire configuration cycle and provides timing for the serial interface. Stratix IV devices use an internal oscillator to generate DCLK. Using the MSEL[] pins, you can select to use a 40 MHz oscillator. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-18 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) In fast AS configuration schemes, Stratix IV devices drive out control signals on the falling edge of DCLK. The serial configuration device responds to the instructions by driving out configuration data on the falling edge of DCLK. Then the data is latched into the Stratix IV device on the following falling edge of DCLK. In configuration mode, Stratix IV devices enable the serial configuration device by driving the nCSO output pin low, which connects to the chip select (nCS) pin of the configuration device. The Stratix IV device uses the serial clock (DCLK) and serial data output (ASDO) pins to send operation commands and/or read address signals to the serial configuration device. The configuration device provides data on its serial data output (DATA) pin, which connects to the DATA0 input of the Stratix IV devices. After all the configuration bits are received by the Stratix IV device, it releases the open-drain CONF_DONE pin, which is pulled high by an external 10-k resistor. Initialization begins only after the CONF_DONE signal reaches a logic high level. All AS configuration pins (DATA0, DCLK, nCSO, and ASDO) have weak internal pull-up resistors that are always active. After configuration, these pins are set as input tri-stated and are driven high by the weak internal pull-up resistors. The CONF_DONE pin must have an external 10-k pull-up resistor in order for the device to initialize. In Stratix IV devices, the initialization clock source is either the internal oscillator or the optional CLKUSR pin. By default, the internal oscillator is the clock source for initialization. If you use the internal oscillator, the Stratix IV device provides itself with enough clock cycles for proper initialization. You also have the flexibility to synchronize initialization of multiple devices or to delay initialization with the CLKUSR option. You can turn on the Enable user-supplied start-up clock (CLKUSR) option in the Quartus II software from the General tab of the Device and Pin Options dialog box. When you select the Enable user supplied start-up clock option, the CLKUSR pin is the initialization clock source. Supplying a clock on CLKUSR does not affect the configuration process. After all configuration data is accepted and CONF_DONE goes high, CLKUSR is enabled after four clock cycles of DCLK. After this time period elapses, Stratix IV devices require 8,532 clock cycles to initialize properly and enter user mode. Stratix IV devices support a CLKUSR fMAX of 125 MHz. An optional INIT_DONE pin is available, which signals the end of initialization and the start of user-mode with a low-to-high transition. The Enable INIT_DONE Output option is available in the Quartus II software from the General tab of the Device and Pin Options dialog box. If you use the INIT_DONE pin, it is high due to an external 10-k pull-up resistor when nCONFIG is low and during the beginning of configuration. After the option bit to enable INIT_DONE is programmed into the device (during the first frame of configuration data), the INIT_DONE pin goes low. When initialization is complete, the INIT_DONE pin is released and pulled high. This low-to-high transition signals that the device has entered user mode. When initialization is complete, the device enters user mode. In user mode, the user I/O pins no longer have weak pull-up resistors and function as assigned in your design. If an error occurs during configuration, Stratix IV devices assert the nSTATUS signal low, indicating a data frame error, and the CONF_DONE signal stays low. If the Auto-restart configuration after error option (available in the Quartus II software from the General tab of the Device and Pin Options dialog box) is turned on, the Stratix IV device resets the configuration device by pulsing nCSO, releases nSTATUS after a reset time-out period (a maximum of 500 s), and retries configuration. If this option is turned off, the system must monitor nSTATUS for errors and then pulse nCONFIG low for at least 2 s to restart configuration. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) 1 10-19 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. When the Stratix IV device is in user mode, you can initiate reconfiguration by pulling the nCONFIG pin low. The nCONFIG pin must be low for at least 2 s. When nCONFIG is pulled low, the device also pulls nSTATUS and CONF_DONE low and all I/O pins are tri-stated. After nCONFIG returns to a logic high level and nSTATUS is released by the Stratix IV device, reconfiguration begins. 1 If you wish to gain control of the EPCS pins, hold the nCONFIG pin low and pull the nCE pin high. This causes the device to reset and tri-state the AS configuration pins. The timing parameters for AS mode are not listed here because the tCF2CD, tCF2ST0, tCFG, tSTATUS, tCF2ST1, and tCD2UM timing parameters are identical to the timing parameters for PS mode listed in Table 10-7 on page 10-30. You can configure multiple Stratix IV devices using a single serial configuration device. You can cascade multiple Stratix IV devices using the chip-enable (nCE) and chip-enable-out (nCEO) pins. The first device in the chain must have its nCE pin connected to GND. You must connect its nCEO pin to the nCE pin of the next device in the chain. When the first device captures all of its configuration data from the bitstream, it drives the nCEO pin low, enabling the next device in the chain. You must leave the nCEO pin of the last device unconnected. The nCONFIG, nSTATUS, CONF_DONE, DCLK, and DATA0 pins of each device in the chain are connected (refer to Figure 10-7). The first Stratix IV device in the chain is the configuration master and controls configuration of the entire chain. You must connect its MSEL pins to select the AS configuration scheme. The remaining Stratix IV devices are configuration slaves. You must connect their MSEL pins to select the PS configuration scheme. Any other Altera device that supports PS configuration can also be part of the chain as a configuration slave. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-20 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) Figure 10-7 shows the pin connections for the multi-device fast AS configuration. Figure 10-7. Multi-Device Fast AS Configuration VCCPGM (1) VCCPGM (1) VCCPGM (1) 10 k 10 k 10 k Serial Configuration Device Stratix IV Device Master Stratix IV Device Slave nSTATUS CONF_DONE nCONFIG nCE nCEO GND DATA nSTATUS CONF_DONE nCONFIG nCE nCEO VCCPGM DATA0 VCCPGM DATA0 DCLK DCLK MSEL2 nCS nCSO MSEL1 ASDI ASDO MSEL0 DCLK GND N.C. MSEL2 MSEL1 MSEL0 GND Buffers (2) Notes to Figure 10-7: (1) Connect the pull-up resistors to VCCPGM at a 3.0-V supply. (2) Connect the repeater buffers between the Stratix IV master and slave device(s) for DATA[0] and DCLK. This is to prevent potential signal integrity and clock skew problems. As shown in Figure 10-7, the nSTATUS and CONF_DONE pins on all target devices are connected together with external pull-up resistors. These pins are open-drain bidirectional pins on the devices. When the first device asserts nCEO (after receiving all of its configuration data), it releases its CONF_DONE pin. But the subsequent devices in the chain keep this shared CONF_DONE line low until they have received their configuration data. When all target devices in the chain have received their configuration data and have released CONF_DONE, the pull-up resistor drives a high level on this line and all devices simultaneously enter initialization mode. If an error occurs at any point during configuration, the nSTATUS line is driven low by the failing device. If you enable the Auto-restart configuration after error option, reconfiguration of the entire chain begins after a reset time-out period (a maximum of 500 s). If you did not enable the Auto-restart configuration after error option, the external system must monitor nSTATUS for errors and then pulse nCONFIG low to restart configuration. The external system can pulse nCONFIG if it is under system control rather than tied to VCCGPM. 1 Stratix IV Device Handbook Volume 1 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) 1 10-21 While you can cascade Stratix IV devices, you cannot cascade or chain together serial configuration devices. If the configuration bitstream size exceeds the capacity of a serial configuration device, you must select a larger configuration device and/or enable the compression feature. When configuring multiple devices, the size of the bitstream is the sum of the individual device configuration bitstreams. A system may have multiple devices that contain the same configuration data. In active serial chains, you can implement this by storing one copy of the .sof in the serial configuration device. The same copy of the .sof configures the master Stratix IV device and all remaining slave devices concurrently. All Stratix IV devices must be the same density and package. To configure four identical Stratix IV devices with the same .sof, set up the chain as shown in Figure 10-8. The first device is the master device and its MSEL pins must be set to select AS configuration. The other three slave devices are set up for concurrent configuration and their MSEL pins must be set to select PS configuration. The nCE input pins from the master and slave are connected to GND, and the DATA and DCLK pins connect in parallel to all four devices. During the configuration cycle, the master device reads its configuration data from the serial configuration device and transmits the configuration data to all three slave devices, configuring all of them simultaneously. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-22 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) Figure 10-8 shows the multi-device fast AS configuration when the devices receive the same data using a single .sof. Figure 10-8. Multi-Device Fast AS Configuration When the Devices Receive the Same Data Using a Single .sof Stratix IV Device Slave nSTATUS CONF_DONE nCONFIG nCE nCEO N.C. VCCPGM (1) VCCPGM (1) VCCPGM (1) DATA0 10 k 10 k MSEL2 DCLK 10 k VCCPGM MSEL1 MSEL0 GND Stratix IV Device Master Serial Configuration Device nSTATUS CONF_DONE nCONFIG nCE nCEO DATA0 DCLK DCLK nCS nCSO ASDI ASDO N.C. GND VCCPGM GND DATA Stratix IV Device Slave MSEL2 nSTATUS CONF_DONE nCONFIG nCE DATA0 MSEL2 DCLK MSEL1 nCEO N.C. VCCPGM MSEL1 MSEL0 MSEL0 GND GND Stratix IV Device Slave Buffers (2) nSTATUS CONF_DONE nCONFIG nCE nCEO DATA0 DCLK MSEL2 N.C. VCCPGM MSEL1 MSEL0 GND Notes to Figure 10-8: (1) Connect the pull-up resistors to VCCPGM at a 3.0-V supply. (2) Connect the repeater buffers between the Stratix IV master and slave device(s) for DATA[0] and DCLK. This is to prevent potential signal integrity and clock skew problems. Estimating Active Serial Configuration Time Active serial configuration time is dominated by the time it takes to transfer data from the serial configuration device to the Stratix IV device. This serial interface is clocked by the Stratix IV DCLK output (generated from an internal oscillator) and must be set to 40 MHz (25 ns).Therefore, the minimum configuration time estimate for an EP4SE230 device (94, 600, 000 bits of uncompressed data) is: RBF Size x (minimum DCLK period / 1 bit per DCLK cycle) = estimated minimum configuration time 94, 600, 000 bits x (25 ns / 1 bit) = 2365 ms Enabling compression reduces the amount of configuration data that is transmitted to the Stratix IV device, which also reduces configuration time. On average, compression reduces configuration time, depending on the design. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) 10-23 Programming Serial Configuration Devices Serial configuration devices are non-volatile, flash-memory-based devices. You can program these devices in-system using the USB-BlasterTM, EthernetBlasterTM, or ByteBlasterTM II download cable. Alternatively, you can program them using the Altera programming unit (APU), supported third-party programmers, or a microprocessor with the SRunner software driver. You can perform in-system programming of serial configuration devices using the conventional AS programming interface or the JTAG interface solution. Because serial configuration devices do not support the JTAG interface, the conventional method to program them is using the AS programming interface. The configuration data used to program serial configuration devices is downloaded using programming hardware. During in-system programming, the download cable disables device access to the AS interface by driving the nCE pin high. Stratix IV devices are also held in reset by a low level on nCONFIG. After programming is complete, the download cable releases nCE and nCONFIG, allowing the pull-down and pull-up resistors to drive GND and VCCPGM, respectively. Figure 10-9 shows the download cable connections for the serial configuration device. Altera has developed Serial FlashLoader (SFL), an in-system programming solution for serial configuration devices using the JTAG interface. This solution requires the Stratix IV device to be a bridge between the JTAG interface and the serial configuration device. f For more information about SFL, refer to AN 370: Using the Serial FlashLoader with Quartus II Software. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-24 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Fast Active Serial Configuration (Serial Configuration Devices) f For more information about the USB-Blaster download cable, refer to the USB-Blaster Download Cable User Guide. For more information about the ByteBlaster II cable, refer to the ByteBlaster II Download Cable User Guide. For more information about the EthernetBlaster download cable, refer to the EthernetBlaster Communications Cable User Guide. Figure 10-9. In-System Programming of Serial Configuration Devices VCCPGM (1) VCCPGM (1) VCCPGM (1) 10 k 10 k 10 k Stratix IV Device CONF_DONE nSTATUS Serial Configuration Device nCEO N.C. nCONFIG nCE 10 k VCCPGM DATA DATA0 DCLK DCLK nCS nCSO MSEL1 ASDI ASDO MSEL0 MSEL2 GND Pin 1 VCCPGM (2) USB Blaster or ByteBlaser II (AS Mode) 10-Pin Male Header Notes to Figure 10-9: (1) Connect these pull-up resistors to VCCPGM at a 3.0-V supply. (2) Power up the USB-ByteBlaster, ByteBlaster II, or EthernetBlaster cable's VCC(TRGT) with VCCPGM. You can program serial configuration devices with the Quartus II software using the Altera programming hardware and the appropriate configuration device programming adapter. In production environments, you can program serial configuration devices using multiple methods. You can use Altera programming hardware or other third-party programming hardware to program blank serial configuration devices before they are mounted on PCBs. Alternatively, you can use an on-board microprocessor to program the serial configuration device in-system using C-based software drivers provided by Altera. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 10-25 You can program a serial configuration device in-system by an external microprocessor using SRunner. SRunner is a software driver developed for embedded serial configuration device programming, which can be easily customized to fit in different embedded systems. SRunner is able to read raw programming data (.rpd) and write to serial configuration devices. The serial configuration device programming time using SRunner is comparable to the programming time with the Quartus II software. f For more information about SRunner, refer to AN 418: SRunner: An Embedded Solution for Serial Configuration Device Programming and the source code on the Altera website at www.altera.com. f For more information about programming serial configuration devices, refer to the Serial Configuration Devices (EPCS1, EPCS4, EPCS16, EPCS64, and EPCS128) Data Sheet chapter in volume 2 of the Configuration Handbook. Guidelines for Connecting Serial Configuration Devices on an AS Interface For single- and multi-device AS configurations, the board trace length and loading between the supported serial configuration device and the Stratix IV device family must follow the recommendations listed in Table 10-6. Table 10-6. Maximum Trace Length and Loading for the AS Configuration Maximum Board Trace Length from the Stratix IV Device to the Serial Configuration Device (Inches) Maximum Board Load (pF) DCLK 10 15 DATA[0] 10 30 nCSO 10 30 ASDO 10 30 Stratix IV Device AS Pins Passive Serial Configuration You can program a PS configuration for Stratix IV devices using an intelligent host, such as a MAX II device or microprocessor with flash memory, or a download cable. In the PS scheme, an external host (a MAX II device, embedded processor, or host PC) controls configuration. Configuration data is clocked into the target Stratix IV device using the DATA0 pin at each rising edge of DCLK. 1 September 2012 The Stratix IV decompression and design security features are fully available when configuring your Stratix IV device using PS mode. Altera Corporation Stratix IV Device Handbook Volume 1 10-26 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration PS Configuration Using a MAX II Device as an External Host In this configuration scheme, you can use a MAX II device as an intelligent host that controls the transfer of configuration data from a storage device, such as flash memory, to the target Stratix IV device. You can store configuration data in .rbf, .hex, or .ttf format. Figure 10-10 shows the configuration interface connections between a Stratix IV device and a MAX II device for single device configuration. Figure 10-10. Single Device PS Configuration Using an External Host Memory ADDR VCCPGM (1) VCCPGM (1) VCCPGM/VCCIO (2) DATA0 10 k 10 k 10 k Stratix IV Device CONF_DONE nSTATUS External Host (MAX II Device or Microprocessor) nCE nCEO GND DATA0 MSEL2 nCONFIG MSEL1 DCLK N.C. VCCPGM MSEL0 GND Note to Figure 10-10: (1) Connect the resistor to a supply that provides an acceptable input signal for the Stratix IV device. VCCPGM must be high enough to meet the VIH specification of the I/O on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. After power-up, Stratix IV devices go through a POR. The POR delay depends on the PORSEL pin setting. When PORSEL is driven low, the standard POR time is 100 ms < TPOR < 300 ms. When PORSEL is driven high, the fast POR time is 4 ms < TPOR < 12 ms. During POR, the device resets, holds nSTATUS low, and tri-states all user I/O pins. After the device successfully exits POR, all user I/O pins continue to be tri-stated. If nIO_pullup is driven low during power-up and configuration, the user I/O pins and dual-purpose I/O pins will have weak pull-up resistors that are on (after POR) before and during configuration. If nIO_pullup is driven high, the weak pull-up resistors are disabled. The configuration cycle consists of three stages--reset, configuration, and initialization. While nCONFIG or nSTATUS are low, the device is in reset. To initiate configuration, the MAX II device must generate a low-to-high transition on the nCONFIG pin. 1 Stratix IV Device Handbook Volume 1 VCC, VCCIO, VCCPGM, and VCCPD of the banks where the configuration pins reside must be fully powered to the appropriate voltage levels to begin the configuration process. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 10-27 When nCONFIG goes high, the device comes out of reset and releases the open-drain nSTATUS pin, which is then pulled high by an external 10-k pull-up resistor. After nSTATUS is released, the device is ready to receive configuration data and the configuration stage begins. When nSTATUS is pulled high, the MAX II device places the configuration data one bit at a time on the DATA0 pin. If you are using configuration data in .rbf, .hex, or .ttf format, you must send the LSB of each data byte first. For example, if the .rbf contains the byte sequence 02 1B EE 01 FA, the serial bitstream you must transmit to the device is 0100-0000 1101-1000 0111-0111 1000-0000 0101-1111. 1 A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host (a Max II CPLD or a microcontroller) is not driving the line. For example, during external host reprogramming or power-up where the I/O driving nCONFIG may be tri-stated. If a pull-up resistor is added to the nCONFIG line, the FPGA stays in user mode if the external host is being reprogrammed. If a pull-down resistor is added to the nCONFIG line, the FPGA goes into reset mode if the external host is being reprogrammed. Whenever the nCONFIG line is released high, ensure that the first DCLK and DATA are not driven unintentionally. The Stratix IV device receives configuration data on the DATA0 pin and the clock is received on the DCLK pin. Data is latched into the device on the rising edge of DCLK. Data is continuously clocked into the target device until CONF_DONE goes high. After the device has received all configuration data successfully, it releases the open-drain CONF_DONE pin, which is pulled high by an external 10-kpull-up resistor. A low-to-high transition on CONF_DONE indicates configuration is complete and initialization of the device can begin. The CONF_DONE pin must have an external 10-k pull-up resistor for the device to initialize. In Stratix IV devices, the initialization clock source is either the internal oscillator or the optional CLKUSR pin. By default, the internal oscillator is the clock source for initialization. If you use the internal oscillator, the Stratix IV device provides itself with enough clock cycles for proper initialization. Therefore, if the internal oscillator is the initialization clock source, sending the entire configuration file to the device is sufficient to configure and initialize the device. Driving DCLK to the device after configuration is complete does not affect device operation. You also have the flexibility to synchronize initialization of multiple devices or to delay initialization with the CLKUSR option. You can turn on the Enable user-supplied start-up clock (CLKUSR) option in the Quartus II software from the General tab of the Device and Pin Options dialog box. If you supply a clock on CLKUSR, it will not affect the configuration process. After all configuration data has been accepted and CONF_DONE goes high, CLKUSR is enabled after the time specified at tCD2CU. After this time period elapses, Stratix IV devices require 8,532 clock cycles to initialize properly and enter user mode. Stratix IV devices support a CLKUSR fMAX of 125 MHz. An optional INIT_DONE pin is available that signals the end of initialization and the start of user-mode with a low-to-high transition. The Enable INIT_DONE Output option is available in the Quartus II software from the General tab of the Device and Pin Options dialog box. If you use the INIT_DONE pin, it is high due to an external 10-k pull-up resistor when nCONFIG is low and during the beginning of configuration. After the option bit to enable INIT_DONE is programmed into the device (during the first frame of configuration data), the INIT_DONE pin goes low. When September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-28 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration initialization is complete, the INIT_DONE pin is released and pulled high. The MAX II device must be able to detect this low-to-high transition that signals the device has entered user mode. When initialization is complete, the device enters user mode. In user-mode, the user I/O pins no longer have weak pull-up resistors and function as assigned in your design. 1 Two DCLK falling edges are required after CONF_DONE goes high to begin the initialization of the device for both uncompressed and compressed bitstream in PS. To ensure DCLK and DATA0 are not left floating at the end of configuration, the MAX II device must drive them either high or low, whichever is convenient on your board. The DATA[0] pin is available as a user I/O pin after configuration. When you chose the PS scheme as a default in the Quartus II software, this I/O pin is tri-stated in user mode and must be driven by the MAX II device. To change this default option in the Quartus II software, select the Dual-Purpose Pins tab of the Device and Pin Options dialog box. The configuration clock (DCLK) speed must be below the specified frequency to ensure correct configuration. No maximum DCLK period exists, which means you can pause the configuration by halting DCLK for an indefinite amount of time. If an error occurs during configuration, the device drives its nSTATUS pin low, resetting itself internally. The low signal on the nSTATUS pin also alerts the MAX II device that there is an error. If the Auto-restart configuration after error option (available in the Quartus II software from the General tab of the Device and Pin Options dialog box) is turned on, the Stratix IV device releases nSTATUS after a reset time-out period (a maximum of 500 s). After nSTATUS is released and pulled high by a pull-up resistor, the MAX II device can try to reconfigure the target device without needing to pulse nCONFIG low. If this option is turned off, the MAX II device must generate a low-to-high transition (with a low pulse of at least 2 s) on nCONFIG to restart the configuration process. 1 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. The MAX II device can also monitor the CONF_DONE and INIT_DONE pins to ensure successful configuration. The CONF_DONE pin must be monitored by the MAX II device to detect errors and determine when programming completes. If all configuration data is sent, but CONF_DONE or INIT_DONE have not gone high, the MAX II device must reconfigure the target device. 1 If you use the optional CLKUSR pin and nCONFIG is pulled low to restart configuration during device initialization, you must ensure that CLKUSR continues toggling during the time nSTATUS is low (a maximum of 500 s). When the device is in user-mode, you can initiate a reconfiguration by transitioning the nCONFIG pin low-to-high. The nCONFIG pin must be low for at least 2 s. When nCONFIG is pulled low, the device also pulls nSTATUS and CONF_DONE low and all I/O pins are tri-stated. After nCONFIG returns to a logic high level and nSTATUS is released by the device, reconfiguration begins. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 10-29 Figure 10-11 shows how to configure multiple devices using a MAX II device. This circuit is similar to the PS configuration circuit for a single device, except the Stratix IV devices are cascaded for multi-device configuration. Figure 10-11. Multi-Device PS Configuration Using an External Host Memory ADDR VCCPGM (1) VCCPGM (1) VCCPGM/VCCIO (2) DATA0 10 k 10 k 10 k Stratix IV Device 1 CONF_DONE CONF_DONE nSTATUS nSTATUS nCE nCE External Host (MAX II Device or Microprocessor) Stratix IV Device 2 nCEO GND DATA0 MSEL2 VCCPGM nCEO MSEL2 DATA0 nCONFIG MSEL1 nCONFIG MSEL1 DCLK MSEL0 DCLK MSEL0 N.C. VCCPGM GND GND Note to Figure 10-11: (1) Connect the resistor to a supply that provides an acceptable input signal for all Stratix IV devices in the chain. VCCPGM must be high enough to meet the VIH specification of the I/O on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. In multi-device PS configuration, the first device's nCE pin is connected to GND, while its nCEO pin is connected to nCE of the next device in the chain. The last device's nCE input comes from the previous device, while its nCEO pin is left floating. After the first device completes configuration in a multi-device configuration chain, its nCEO pin drives low to activate the second device's nCE pin, which prompts the second device to begin configuration. The second device in the chain begins configuration within one clock cycle. Therefore, the transfer of data destinations is transparent to the MAX II device. All other configuration pins (nCONFIG, nSTATUS, DCLK, DATA0, and CONF_DONE) are connected to every device in the chain. Configuration signals can require buffering to ensure signal integrity and prevent clock skew problems. Ensure that the DCLK and DATA lines are buffered for every fourth device. Because all device CONF_DONE pins are tied together, all devices initialize and enter user mode at the same time. Because all nSTATUS and CONF_DONE pins are tied together, if any device detects an error, configuration stops for the entire chain and you must reconfigure the entire chain. For example, if the first device flags an error on nSTATUS, it resets the chain by pulling its nSTATUS pin low. This behavior is similar to a single device detecting an error. If the Auto-restart configuration after error option is turned on, the devices release their nSTATUS pins after a reset time-out period (a maximum of 500 s). After all nSTATUS pins are released and pulled high, the MAX II device can try to reconfigure the chain without needing to pulse nCONFIG low. If this option is turned off, the MAX II device must generate a low-to-high transition (with a low pulse of at least 2 s) on nCONFIG to restart the configuration process. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-30 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 1 If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the nSTATUS pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. In your system, you can have multiple devices that contain the same configuration data. To support this configuration scheme, all device nCE inputs are tied to GND, while the nCEO pins are left floating. All other configuration pins (nCONFIG, nSTATUS, DCLK, DATA0, and CONF_DONE) are connected to every device in the chain. Configuration signals can require buffering to ensure signal integrity and prevent clock skew problems. Ensure that the DCLK and DATA lines are buffered for every fourth device. Devices must be the same density and package. All devices start and complete configuration at the same time. Figure 10-12 shows multi-device PS configuration when both Stratix IV devices are receiving the same configuration data. Figure 10-12. Multiple-Device PS Configuration When Both Devices Receive the Same Data Memory ADDR VCCPGM (1) VCCPGM (1) V CCPGM/VCCIO (2) DATA0 10 k 10 k 10 k Stratix IV Device Stratix IV Device CONF_DONE CONF_DONE nSTATUS External Host (MAX II Device or Microprocessor) nCE nCEO GND DATA0 MSEL2 N.C. (3) nCEO nSTATUS nCE VCCPGM GND DATA0 MSEL2 nCONFIG MSEL1 nCONFIG MSEL1 DCLK MSEL0 DCLK MSEL0 N.C. (3) VCCPGM GND GND Notes to Figure 10-12: (1) Connect the resistor to a supply that provides an acceptable input signal for all Stratix IV devices in the chain. VCCPGM must be high enough to meet the VIH specification of the I/O on the device and the external host. Altera recommends powering up all configuration system I/Os with VCCPGM. (2) A pull-up or pull-down resistor helps keep the nCONFIG line in a known state when the external host is not driving the line. (3) The nCEO pins of both devices are left unconnected when configuring the same configuration data into multiple devices. You can use a single configuration chain to configure Stratix IV devices with other Altera devices. To ensure that all devices in the chain complete configuration at the same time, or that an error flagged by one device initiates reconfiguration in all devices, all of the device CONF_DONE and nSTATUS pins must be tied together. f For more information about configuring multiple Altera devices in the same configuration chain, refer to the Configuring Mixed Altera FPGA Chains chapter in volume 2 of the Configuration Handbook. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 10-31 PS Configuration Timing Figure 10-13 shows the timing waveform for PS configuration when using a MAX II device as an external host. Figure 10-13. PS Configuration Timing Waveform (Note 1) tCF2ST1 tCFG tCF2CK nCONFIG nSTATUS (2) tSTATUS tCF2ST0 t CLK CONF_DONE (3) tCF2CD tST2CK tCH tCL (4) DCLK tDH Bit 0 Bit 1 Bit 2 Bit 3 DATA (5) Bit n tDSU High-Z User I/O User Mode INIT_DONE tCD2UM Notes to Figure 10-13: (1) The beginning of this waveform shows the device in user mode. In user mode, nCONFIG, nSTATUS, and CONF_DONE are at logic high levels. When nCONFIG is pulled low, a reconfiguration cycle begins. (2) After power-up, the Stratix IV device holds nSTATUS low for the time of the POR delay. (3) After power-up, before and during configuration, CONF_DONE is low. (4) Do not leave DCLK floating after configuration. You can drive it high or low, whichever is more convenient. (5) DATA[0] is available as a user I/O pin after configuration. The state of this pin depends on the dual-purpose pin settings. Table 10-7 lists the timing parameters for Stratix IV devices for PS configuration. Table 10-7. PS Timing Parameters for Stratix IV Devices (Part 1 of 2) (Note 1) Symbol Parameter Minimum Maximum Units ns tCF2CD nCONFIG low to CONF_DONE low -- 800 tCF2ST0 nCONFIG low to nSTATUS low -- 800 ns tCFG nCONFIG low pulse width 2 -- s tSTATUS nSTATUS low pulse width 10 500 (2) s tCF2ST1 nCONFIG high to nSTATUS high -- 500 (3) s tCF2CK nCONFIG high to first rising edge on DCLK 500 -- s tST2CK nSTATUS high to first rising edge of DCLK 2 -- s tDSU Data setup time before rising edge on DCLK 4 -- ns tDH Data hold time after rising edge on DCLK 0 -- ns tCH DCLK high time (5) 3.2 -- ns tCL DCLK low time (5) 3.2 -- ns tCLK DCLK period (5) 8 -- ns fMAX DCLK frequency -- 125 MHz tR Input rise time -- 40 ns September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-32 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration Table 10-7. PS Timing Parameters for Stratix IV Devices (Part 2 of 2) (Note 1) Symbol Parameter Minimum Maximum Units tF Input fall time -- 40 ns tCD2UM CONF_DONE high to user mode (4) 55 150 s tCD2CU CONF_DONE high to CLKUSR enabled 4 x maximum DCLK period -- -- tCD2UMC CONF_DONE high to user mode with CLKUSR option on tCD2CU + (8532 CLKUSR period) -- -- Notes to Table 10-7: (1) This information is preliminary. (2) This value is applicable if you do not delay the configuration by extending the nCONFIG or nSTATUS low pulse width. (3) This value is applicable if you do not delay the configuration by externally holding nSTATUS low. (4) The minimum and maximum numbers apply only if you choose the internal oscillator as the clock source for starting the device. (5) Adding up tCH and tCL equals to tCLK. When tCH is 3.2 ns (min), tCL must be 4.8 ns and vice versa. f Device configuration options and how to create configuration files are described in the Device Configuration Options and Configuration File Formats chapters in volume 2 of the Configuration Handbook. PS Configuration Using a Microprocessor In this PS configuration scheme, a microprocessor controls the transfer of configuration data from a storage device, such as flash memory, to the target Stratix IV device. For more information about configuration and timing information, refer to "PS Configuration Using a MAX II Device as an External Host" on page 10-25. This section is also applicable when using a microprocessor as an external host. PS Configuration Using a Download Cable 1 In this section, the generic term "download cable" includes the Altera USB-Blaster universal serial bus (USB) port download cable, MasterBlaster serial/USB communications cable, ByteBlaster II parallel port download cable, ByteBlasterMV parallel port download cable, and EthernetBlaster download cable. In a PS configuration with a download cable, an intelligent host (such as a PC) transfers data from a storage device to the device using the USB Blaster, MasterBlaster, ByteBlaster II, EthernetBlaster, or ByteBlasterMV cable. After power-up, Stratix IV devices go through a POR. The POR delay depends on the PORSEL pin setting. When PORSEL is driven low, the standard POR time is 100 ms < TPOR < 300 ms. When PORSEL is driven high, the fast POR time is 4 ms < TPOR < 12 ms. During POR, the device resets, holds nSTATUS low, and tri-states all user I/O pins. After the device successfully exits POR, all user I/O pins continue to be tri-stated. If nIO_pullup is driven low during power-up and configuration, the user I/O pins and dual-purpose I/O pins will have weak pull-up resistors, which are on (after POR) before and during configuration. If nIO_pullup is driven high, the weak pull-up resistors are disabled. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration 10-33 The configuration cycle consists of three stages--reset, configuration, and initialization. While nCONFIG or nSTATUS are low, the device is in reset. To initiate configuration in this scheme, the download cable generates a low-to-high transition on the nCONFIG pin. 1 To begin configuration, power the VCC, VCCIO, VCCPGM, and VCCPD voltages (for the banks where the configuration pins reside) to the appropriate voltage levels. When nCONFIG goes high, the device comes out of reset and releases the open-drain nSTATUS pin, which is then pulled high by an external 10-k pull-up resistor. After nSTATUS is released, the device is ready to receive configuration data and the configuration stage begins. The programming hardware or download cable then places the configuration data one bit at a time on the device's DATA0 pin. The configuration data is clocked into the target device until CONF_DONE goes high. The CONF_DONE pin must have an external 10-k pull-up resistor for the device to initialize. When using a download cable, setting the Auto-restart configuration after error option does not affect the configuration cycle because you must manually restart configuration in the Quartus II software when an error occurs. Additionally, the Enable user-supplied start-up clock (CLKUSR) option has no affect on the device initialization because this option is disabled in the .sof when programming the device using the Quartus II programmer and download cable. Therefore, if you turn on the CLKUSR option, you do not need to provide a clock on CLKUSR when you are configuring the device with the Quartus II programmer and a download cable. Figure 10-14 shows PS configuration for Stratix IV devices using a USB Blaster, EthernetBlaster, MasterBlaster, ByteBlaster II, or ByteBlasterMV cable. Figure 10-14. PS Configuration Using a USB Blaster, EthernetBlaster, MasterBlaster, ByteBlaster II, or ByteBlasterMV Cable VCCPGM (1) VCCPGM (1) 10 k (2) VCCPGM (1) VCCPGM (1) VCCPGM (1) 10 k 10 k Stratix IV Device VCCPGM 10 k (2) MSEL2 10 k CONF_DONE nSTATUS MSEL1 MSEL0 GND nCE GND DCLK DATA0 nCONFIG nCEO Download Cable 10-Pin Male Header (PS Mode) N.C. Pin 1 VCCPGM (1) GND VIO (3) Shield GND Notes to Figure 10-14: (1) Connect the pull-up resistor to the same supply voltage (VCCPGM) as the USB Blaster, MasterBlaster (VIO pin), ByteBlaster II, ByteBlasterMV, or EthernetBlaster cable. (2) You only need the pull-up resistors on DATA0 and DCLK if the download cable is the only configuration scheme used on your board. This ensures that DATA0 and DCLK are not left floating after configuration. For example, if you are also using a configuration device, you do not need the pull-up resistors on DATA0 and DCLK. (3) Pin 6 of the header is a VIO reference voltage for the MasterBlaster output driver. VIO must match the device's VCCPGM. For more information about this value, refer to the MasterBlaster Serial/USB Communications Cable User Guide. In the USB-Blaster, ByteBlaster II, and ByteBlasterMV cable, this pin is a no connect. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-34 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Passive Serial Configuration You can use a download cable to configure multiple Stratix IV devices by connecting each device's nCEO pin to the subsequent device's nCE pin. The first device's nCE pin is connected to GND, while its nCEO pin is connected to the nCE of the next device in the chain. The last device's nCE input comes from the previous device, while its nCEO pin is left floating. All other configuration pins (nCONFIG, nSTATUS, DCLK, DATA0, and CONF_DONE) are connected to every device in the chain. Because all CONF_DONE pins are tied together, all devices in the chain initialize and enter user mode at the same time. In addition, because the nSTATUS pins are tied together, the entire chain halts configuration if any device detects an error. The Auto-restart configuration after error option does not affect the configuration cycle because you must manually restart the configuration in the Quartus II software when an error occurs. Figure 10-15 shows how to configure multiple Stratix IV devices with a download cable. Figure 10-15. Multi-Device PS Configuration Using a USB Blaster, EthernetBlaster, MasterBlaster, ByteBlaster II, or ByteBlasterMV Cable VCCPGM (1) 10 k VCCPGM (1) VCCPGM (1) 10 k VCCPGM (1) GND GND VCCPGM (1) (2) Pin 1 VCCPGM (1) GND VIO (3) nCEO nCE 10 k VCCPGM (1) 10 k CONF_DONE nSTATUS DCLK MSEL2 MSEL1 MSEL0 (2) 10 k Stratix IV Device 1 Download Cable 10-Pin Male Header (PS Mode) VCCPGM (1) DATA0 nCONFIG GND Stratix IV Device 2 MSEL2 MSEL1 MSEL0 CONF_DONE nSTATUS DCLK GND nCEO N.C. nCE DATA0 nCONFIG Notes to Figure 10-15: (1) Connect the pull-up resistor to the same supply voltage (VCCPGM) as the USB Blaster, MasterBlaster (VIO pin), ByteBlaster II, ByteBlasterMV, or EthernetBlaster cable. (2) You only need the pull-up resistors on DATA0 and DCLK if the download cable is the only configuration scheme used on your board. This is to ensure that DATA0 and DCLK are not left floating after configuration. For example, if you are also using a configuration device, you do not need the pull-up resistors on DATA0 and DCLK. (3) Pin 6 of the header is a VIO reference voltage for the MasterBlaster output driver. VIO must match the device's VCCPGM. For more information about this value, refer to the MasterBlaster Serial/USB Communications Cable User Guide. In the USB-Blaster, ByteBlaster II, and ByteBlasterMV cables, this pin is a no connect. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices JTAG Configuration 10-35 f For more information about how to use the USB Blaster, MasterBlaster, ByteBlaster II, or ByteBlasterMV cables, refer to the following user guides: USB-Blaster Download Cable User Guide MasterBlaster Serial/USB Communications Cable User Guide ByteBlaster II Download Cable User Guide ByteBlasterMV Download Cable User Guide EthernetBlaster Communications Cable User Guide JTAG Configuration JTAG has developed a specification for boundary-scan testing. This boundary-scan test (BST) architecture offers the capability to efficiently test components on PCBs with tight lead spacing. The BST architecture can test pin connections without using physical test probes and capture functional data while a device is operating normally. You can also use JTAG circuitry to shift configuration data into the device. The Quartus II software automatically generates .sofs that you can use for JTAG configuration with a download cable in the Quartus II software programmer. f For more information about JTAG boundary-scan testing and commands available using Stratix IV devices, refer to the following documents: JTAG Boundary Scan Testing in Stratix IV Devices chapter Programming Support for Jam STAPL Language Stratix IV devices are designed such that JTAG instructions have precedence over any device configuration modes. Therefore, JTAG configuration can take place without waiting for other configuration modes to complete. For example, if you attempt JTAG configuration of Stratix IV devices during PS configuration, PS configuration is terminated and JTAG configuration begins. 1 You cannot use the Stratix IV decompression or design security features if you are configuring your Stratix IV device when using JTAG-based configuration. 1 A device operating in JTAG mode uses four required pins, TDI, TDO, TMS, and TCK, and one optional pin, TRST. The TCK pin has an internal weak pull-down resistor, while the TDI, TMS, and TRST pins have weak internal pull-up resistors (typically 25 k). The JTAG output pin TDO and all JTAG input pins are powered by 2.5-V/3.0-V VCCPD. All the JTAG pins only support the LVTTL I/O standard. All user I/O pins are tri-stated during JTAG configuration. f All the JTAG pins are powered by the VCCPD power supply of I/O bank 1A. For more information about how to connect a JTAG chain with multiple voltages across the devices in the chain, refer to the JTAG Boundary Scan Testing in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-36 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices JTAG Configuration During JTAG configuration, you can download data to the device on the PCB through the USB Blaster, MasterBlaster, ByteBlaster II, EthernetBlaster, or ByteBlasterMV download cable. Configuring devices through a cable is similar to programming devices in-system, except you must connect the TRST pin to VCCPD. This ensures that the TAP controller is not reset. Figure 10-16 shows JTAG configuration of a single Stratix IV device when using a download cable. Figure 10-16. JTAG Configuration of a Single Device Using a Download Cable VCCPD (1) (5) VCCPGM VCCPD (1) VCCPGM 10 k Stratix IV Device 10 k nCE (4) GND N.C. (2) (2) (2) nCE0 nSTATUS CONF_DONE nCONFIG MSEL[2..0] DCLK (5) TCK TDO TMS TDI Download Cable 10-Pin Male Header (JTAG Mode) (Top View) VCCPD (1) TRST Pin 1 VCCPD (1) GND VIO (3) 1 k GND GND Notes to Figure 10-16: (1) Connect the pull-up resistor to the same supply voltage as the USB Blaster, MasterBlaster (VIO pin), ByteBlaster II, ByteBlasterMV, or EthernetBlaster cable. The voltage supply can be connected to the VCCPD of the device. (2) Connect the nCONFIG and MSEL[2..0] pins to support a non-JTAG configuration scheme. If you only use the JTAG configuration, connect nCONFIG to VCCPGM and MSEL[2..0] to GND. Pull DCLK either high or low, whichever is convenient on your board. (3) Pin 6 of the header is a VIO reference voltage for the MasterBlaster output driver. VIO must match the device's VCCPD. For more information about this value, refer to the MasterBlaster Serial/USB Communications Cable User Guide. In the USB-Blaster, ByteBlaster II, and ByteBlasterMV cable, this pin is a no connect. (4) You must connect nCE to GND or driven low for successful JTAG configuration. (5) The pull-up resistor value can vary from 1 k to 10 k . Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices JTAG Configuration 10-37 To configure a single device in a JTAG chain, the programming software places all other devices in bypass mode. In bypass mode, devices pass programming data from the TDI pin to the TDO pin through a single bypass register without being affected internally. This scheme enables the programming software to program or verify the target device. Configuration data driven into the device appears on the TDO pin one clock cycle later. The Quartus II software verifies successful JTAG configuration upon completion. At the end of configuration, the software checks the state of CONF_DONE through the JTAG port. When the Quartus II software generates a JAM file (.jam) for a multi-device chain, it contains instructions so that all the devices in the chain are initialized at the same time. If CONF_DONE is not high, the Quartus II software indicates that configuration has failed. If CONF_DONE is high, the software indicates that configuration was successful. After the configuration bitstream is transmitted serially using the JTAG TDI port, the TCK port is clocked an additional 1,094 cycles to perform device initialization. Stratix IV devices have dedicated JTAG pins that always function as JTAG pins. Not only can you perform JTAG testing on Stratix IV devices before and after, but also during configuration. While other device families do not support JTAG testing during configuration, Stratix IV devices support the bypass, ID code, and sample instructions during configuration without interrupting configuration. All other JTAG instructions may only be issued by first interrupting configuration and reprogramming the I/O pins using the CONFIG_IO instruction. The CONFIG_IO instruction allows I/O buffers to be configured using the JTAG port and when issued, interrupts configuration. This instruction allows you to perform board-level testing prior to configuring the Stratix IV device or waiting for a configuration device to complete configuration. After configuration has been interrupted and JTAG testing is complete, you must reconfigure the part using JTAG (PULSE_CONFIG instruction) or by pulsing nCONFIG low. The chip-wide reset (DEV_CLRn) and chip-wide output enable (DEV_OE) pins on Stratix IV devices do not affect JTAG boundary-scan or programming operations. Toggling these pins does not affect JTAG operations (other than the usual boundary-scan operation). When designing a board for JTAG configuration for Stratix IV devices, consider the dedicated configuration pins. Table 10-8 lists how these pins are connected during JTAG configuration. Table 10-8. Dedicated Configuration Pin Connections During JTAG Configuration (Part 1 of 2) Signal September 2012 Description nCE On all Stratix IV devices in the chain, nCE must be driven low by connecting it to GND, pulling it low using a resistor, or driving it by some control circuitry. For devices that are also in multi-device FPP, AS, or PS configuration chains, the nCE pins must be connected to GND during JTAG configuration or JTAG must be configured in the same order as the configuration chain. nCEO On all Stratix IV devices in the chain, you can leave nCEO floating or connected to the nCE of the next device. MSEL Do not leave these pins floating. These pins support whichever non-JTAG configuration is used in production. If you only use JTAG configuration, tie these pins to GND. Altera Corporation Stratix IV Device Handbook Volume 1 10-38 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices JTAG Configuration Table 10-8. Dedicated Configuration Pin Connections During JTAG Configuration (Part 2 of 2) Signal Description nCONFIG Driven high by connecting to VCCPGM, pulling up using a resistor, or driven high by some control circuitry. nSTATUS Pull to VCCPGM using a 10-k resistor. When configuring multiple devices in the same JTAG chain, each nSTATUS pin must be pulled up to VCCPGM individually. CONF_DONE Pull to VCCPGM using a 10-k resistor. When configuring multiple devices in the same JTAG chain, each CONF_DONE pin must be pulled up to VCCPGM individually. CONF_DONE going high at the end of JTAG configuration indicates successful configuration. DCLK Do not leave DCLK floating. Drive low or high, whichever is more convenient on your board. When programming a JTAG device chain, one JTAG-compatible header is connected to several devices. The number of devices in the JTAG chain is limited only by the drive capability of the download cable. When four or more devices are connected in a JTAG chain, Altera recommends buffering the TCK, TDI, and TMS pins with an on-board buffer. JTAG-chain device programming is ideal when the system contains multiple devices, or when testing your system using JTAG BST circuitry. Figure 10-17 shows a multi-device JTAG configuration when using a download cable. Figure 10-17. JTAG Configuration of Multiple Devices Using a Download Cable Stratix IV Device Download Cable 10-Pin Male Header (JTAG Mode) VCCPGM (2) Pin 1 VCCPGM (5) VCCPD (1) (1) VCCPD (2) (5) VCCPD (1) VIO (3) VCCPGM (2) DCLK MSEL[2..0] nCE (4) 10 k nSTATUS nCONFIG TRST TDI TMS TDO TCK 10 k (2) nSTATUS nCONFIG (2) DCLK (2) MSEL[2..0] CONF_DONE CONF_DONE (2) DCLK (2) MSEL[2..0] VCCPD (1) VCCPGM VCCPGM 10 k 10 k nSTATUS nCONFIG CONF_DONE (2) VCCPGM 10 k 10 k VCCPD (1) Stratix II or Stratix II GX Stratix IV Device Device Stratix IV Device nCE (4) TRST TDI TMS VCCPD (1) TDO TCK nCE (4) TRST TDI TMS TDO TCK 1 k Notes to Figure 10-17: (1) Connect the pull-up resistor to the same supply voltage as the USB Blaster, MasterBlaster (VIO pin), ByteBlaster II, ByteBlasterMV, or EthernetBlaster cable. Connect the voltage supply to VCCPD of the device. (2) Connect the nCONFIG and MSEL[2..0] pins to support a non-JTAG configuration scheme. If you only use a JTAG configuration, connect nCONFIG to VCCPGM and MSEL[2..0] to GND. Pull DCLK either high or low, whichever is convenient on your board. (3) Pin 6 of the header is a VIO reference voltage for the MasterBlaster output driver. VIO must match the device's VCCPD. For more information about this value, refer to the MasterBlaster Serial/USB Communications Cable User Guide. In the USB-Blaster, ByteBlaster II, and ByteBlasterMV cables, this pin is a no connect. (4) You must connect nCE to GND or drive it low for successful JTAG configuration. (5) The pull-up resistor value can vary from 1 k to 10 k . Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices JTAG Configuration 10-39 You must connect the nCE pin to GND or drive it low during JTAG configuration. In multi-device FPP, AS, and PS configuration chains, the first device's nCE pin is connected to GND, while its nCEO pin is connected to nCE of the next device in the chain. The last device's nCE input comes from the previous device, while its nCEO pin is left floating. In addition, the CONF_DONE and nSTATUS signals are all shared in multi-device FPP, AS, or PS configuration chains so the devices can enter user mode at the same time after configuration is complete. When the CONF_DONE and nSTATUS signals are shared among all the devices, you must configure every device when JTAG configuration is performed. If you only use JTAG configuration, Altera recommends connecting the circuitry as shown in Figure 10-17, where each of the CONF_DONE and nSTATUS signals are isolated, so that each device can enter user mode individually. After the first device completes configuration in a multi-device configuration chain, its nCEO pin drives low to activate the second device's nCE pin, which prompts the second device to begin configuration. Therefore, if these devices are also in a JTAG chain, ensure the nCE pins are connected to GND during JTAG configuration or that the devices are JTAG configured in the same order as the configuration chain. As long as the devices are JTAG configured in the same order as the multi-device configuration chain, the nCEO of the previous device drives the nCE of the next device low when it has successfully been JTAG configured. You can place other Altera devices that have JTAG support in the same JTAG chain for device programming and configuration. 1 JTAG configuration support is enhanced and allows more than 17 Stratix IV devices to be cascaded in a JTAG chain. f For more information about configuring multiple Altera devices in the same configuration chain, refer to the Configuring Mixed Altera FPGA Chains chapter in volume 2 of the Configuration Handbook. You can configure Stratix IV devices using multiple configuration schemes on the same board. Combining JTAG configuration with AS configuration on your board is useful in the prototyping environment because it allows multiple methods to configure your FPGA. f For more information about combining JTAG configuration with other configuration schemes, refer to the Combining Different Configuration Schemes chapter in volume 2 of the Configuration Handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-40 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins Figure 10-18 shows JTAG configuration of a Stratix IV device using a microprocessor. Figure 10-18. JTAG Configuration of a Single Device Using a Microprocessor VCCPGM (1) VCCPGM (1) Memory ADDR Stratix IV Device 10 k 10 k DATA nSTATUS VCCPD TRST TDI (4) TCK (4) TMS (4) TDO (4) Microprocessor CONF_DONE DCLK nCONFIG MSEL[2..0] nCEO (2) (2) (2) N.C. (3) nCE GND Notes to Figure 10-18: (1) Connect the pull-up resistor to a supply that provides an acceptable input signal for all Stratix IV devices in the chain. VCCPGM must be high enough to meet the VIH specification of the I/O on the device. (2) Connect the nCONFIG and MSEL[2..0] pins to support a non-JTAG configuration scheme. If you use only a JTAG configuration, connect nCONFIG to VCCGPM and MSEL[2..0] to GND. Pull DCLK either high or low, whichever is convenient on your board. (3) Connect nCE to GND or drive it low for successful JTAG configuration. (4) The microprocessor must use the same I/O standard as VCCPD to drive the JTAG pins. Jam STAPL Jam STAPL, JEDEC standard JESD-71, is a standard file format for in-system programmability (ISP) purposes. Jam STAPL supports programming or configuration of programmable devices and testing of electronic systems, using the IEEE 1149.1 JTAG interface. Jam STAPL is a freely licensed open standard. The Jam Player provides an interface for manipulating the IEEE Std. 1149.1 JTAG TAP state machine. f For more information about JTAG and Jam STAPL in embedded environments, refer to Using Jam STAPL for ISP via an Embedded Processor. To download the Jam Player, visit the Altera website at www.altera.com. Device Configuration Pins The following tables list the connections and functionality of all the configuration-related pins on Stratix IV devices. Table 10-9 lists the Stratix IV configuration pins and their power supply. Table 10-9. Stratix IV Configuration Pin Summary (Part 1 of 2) (Note 1) Description Input/Output Dedicated Powered By Configuration Mode TDI Input Yes VCCPD JTAG TMS Input Yes VCCPD JTAG TCK Input Yes VCCPD JTAG TRST Input Yes VCCPD JTAG TDO Output Yes VCCPD JTAG CRC_ERROR Output -- Pull-up Optional, all modes Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins 10-41 Table 10-9. Stratix IV Configuration Pin Summary (Part 2 of 2) (Note 1) Description Input/Output Dedicated Powered By Configuration Mode DATA0 Input -- VCCPGM/VCCIO (3) All modes except JTAG DATA[7..1] Input -- VCCPGM/VCCIO (3) FPP INIT_DONE Output -- Pull-up Optional, all modes CLKUSR Input -- VCCPGM/VCCIO (3) Optional nSTATUS Bidirectional Yes VCCPGM/Pull-up All modes Input Yes VCCPGM All modes Bidirectional Yes VCCPGM/Pull-up All modes nCONFIG Input Yes VCCPGM All modes PORSEL Input Yes VCC (2) All modes Yes VCCPGM AS nCE CONF_DONE ASDO (4) Output nCSO (4) Output Yes VCCPGM AS Input Yes VCCPGM PS, FPP Output Yes VCCPGM AS Input Yes VCC (2) All modes Output Yes VCCPGM All modes Input Yes VCC (2) All modes DCLK (4) nIO_PULLUP nCEO MSEL[2..0] Notes to Table 10-9: (1) The total number of pins is 29. The total number of dedicated pins is 18. (2) Although MSEL[2..0], PORSEL, and nIO_PULLUP are powered up by VCC, Altera recommends connecting these pins to VCCPGM or GND directly without using a pull-up or pull-down resistor. (3) These pins are powered up by VCCPGM during configuration. These pins are powered up by VCCIO if they are used as regular I/O in user mode. (4) To tri-state this pin, in the Quartus II software, on the Assignments menu, select Device. On the Device page, select Device and Pin Options... On the Device and Pin Options page, select Configuration and select the Enable input tri-state on active configuration pins in user mode option. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-42 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins Table 10-10 lists the dedicated configuration pins. You must connect these pins properly on your board for successful configuration. Some of these pins may not be required for your configuration schemes. Table 10-10. Dedicated Configuration Pins on the Stratix IV Device (Part 1 of 4) Pin Name User Mode Configuration Scheme Pin Type Description Dedicated power pin. Use this pin to power all dedicated configuration inputs, dedicated configuration outputs, dedicated configuration bidirectional pins, and some of the dual functional pins that are used for configuration. VCCPGM N/A All Power You must connect this pin to 1.8, 2.5, or 3.0 V. VCCPGM must ramp-up from 0 V to VCCPGM within 100 ms when PORSEL is low or 4 ms when PORSEL is high. If VCCPGM is not ramped up within this specified time, your Stratix IV device will not configure successfully. If your system does not allow a VCCPGM ramp-up within 100 ms or 4 ms, you must hold nCONFIG low until all power supplies are stable. Dedicated power pin. Use this pin to power the I/O pre-drivers, JTAG input and output pins, and design security circuitry. VCCPD N/A All Power You must connect this pin to 2.5 V or 3.0 V, depending on the I/O standards selected. For the 3.0-V I/O standard, VCCPD = 3.0 V. For the 2.5 V or below I/O standards, VCCPD = 2.5 V. VCCPD must ramp-up from 0 V to 2.5 V / 3.0 V within 100 ms when PORSEL is low or 4 ms when PORSEL is high. If VCCPD is not ramped up within this specified time, your Stratix IV device will not configure successfully. If your system does not allow a VCCPD to ramp-up time within 100 ms or 4 ms, you must hold nCONFIG low until all power supplies are stable. PORSEL N/A All Input Dedicated input that selects between a standard POR time or a fast POR time. A logic low selects a standard POR time setting of 100 ms < TPOR < 300 ms and a logic high selects a fast POR time setting of 4 ms < TPOR < 12 ms. The PORSEL input buffer is powered by VCC and has an internal 5-kpull-down resistor that is always active. Tie the PORSEL pin directly to VCCPGM or GND. nIO_PULLUP N/A All Input Dedicated input that chooses whether the internal pull-up resistors on the user I/O pins and dual-purpose I/O pins (nCSO, nASDO, DATA[7..0], CLKUSR, and INIT_DONE) are on or off before and during configuration. A logic high turns off the weak internal pull-up resistors; a logic low turns them on. The nIO-PULLUP input buffer is powered by VCC and has an internal 5-k pull-down resistor that is always active. The nIO-PULLUP can be tied directly to VCCPGM, using a 1-k pull-up resistor or tied directly to GND, depending on your device requirements. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins 10-43 Table 10-10. Dedicated Configuration Pins on the Stratix IV Device (Part 2 of 4) Pin Name User Mode Configuration Scheme Pin Type Description Three-bit configuration input that sets the Stratix IV device configuration scheme. For the appropriate connections, refer to Table 10-1 on page 10-2. MSEL[2..0] N/A All Input You must hardwire these pins to VCCPGM or GND. The MSEL[2..0] pins have internal 5-k pull-down resistors that are always active. nCONFIG N/A All Input Configuration control input. Pulling this pin low during user-mode causes the device to lose its configuration data, enter a reset state, and tri-state all I/O pins. Returning this pin to a logic high level initiates a reconfiguration. Configuration is possible only if this pin is high, except in JTAG programming mode, when nCONFIG is ignored. The device drives nSTATUS low immediately after power-up and releases it after the POR time. During user mode and regular configuration, this pin is pulled high by an external 10-k resistor. This pin, when driven low by the Stratix IV device, indicates that the device has encountered an error during configuration. nSTATUS N/A All Bidirectional open-drain Status output--If an error occurs during configuration, nSTATUS is pulled low by the target device. Status input--If an external source drives the nSTATUS pin low during configuration or initialization, the target device enters an error state. Driving nSTATUS low after configuration and initialization does not affect the configured device. If you use a configuration device, driving nSTATUS low causes the configuration device to attempt to configure the device, but because the device ignores transitions on nSTATUS in user mode, the device does not reconfigure. To initiate a reconfiguration, nCONFIG must be pulled low. If you have enabled the Auto-restart configuration after error option, the nSTATUS pin transitions from high to low and back again to high when a configuration error is detected. This appears as a low pulse at the pin with a minimum pulse width of 10 s to a maximum pulse width of 500 s, as defined in the tSTATUS specification. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-44 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins Table 10-10. Dedicated Configuration Pins on the Stratix IV Device (Part 3 of 4) Pin Name User Mode Configuration Scheme Pin Type Description If VCCPGM is not fully powered up, the following could occur: nSTATUS (continued) -- -- -- VCCPGM is powered high enough for the nSTATUS buffer to function properly and nSTATUS is driven low. When VCCPGM is ramped up, POR trips and nSTATUS is released after POR expires. VCCPGM is not powered high enough for the nSTATUS buffer to function properly. In this situation, nSTATUS might appear logic high, triggering a configuration attempt that would fail because POR did not yet trip. When VCCPD is powered up, nSTATUS is pulled low because POR did not yet trip. When POR trips after VCCPGM is powered up, nSTATUS is released and pulled high. At that point, reconfiguration is triggered and the device is configured. Status output. The target device drives the CONF_DONE pin low before and during configuration. After all the configuration data is received without error and the initialization cycle starts, the target device releases CONF_DONE. CONF_DONE N/A All Bidirectional open-drain Status input. After all the data is received and CONF_DONE goes high, the target device initializes and enters user mode. The CONF_DONE pin must have an external 10-k pull-up resistor for the device to initialize. Driving CONF_DONE low after configuration and initialization does not affect the configured device. nCE N/A All Input Active-low chip enable. The nCE pin activates the device with a low signal to allow configuration. The nCE pin must be held low during configuration, initialization, and user mode. In single device configuration, it must be tied low. In multi-device configuration, nCE of the first device is tied low, while its nCEO pin is connected to nCE of the next device in the chain. The nCE pin must also be held low for successful JTAG programming of the device. nCEO N/A All Output Output that drives low when device configuration is complete. In single device configuration, this pin is left floating. In multi-device configuration, this pin feeds the next device's nCE pin. The nCEO of the last device in the chain is left floating. The nCEO pin is powered by VCCPGM. ASDO N/A AS Output Control signal from the Stratix IV device to the serial configuration device in AS mode used to read out configuration data. In AS mode, ASDO has an internal pull-up resistor that is always active. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins 10-45 Table 10-10. Dedicated Configuration Pins on the Stratix IV Device (Part 4 of 4) Pin Name nCSO User Mode Configuration Scheme N/A AS Pin Type Description Output Output control signal from the Stratix IV device to the serial configuration device in AS mode that enables the configuration device. In AS mode, nCSO has an internal pull-up resistor that is always active. In PS and FPP configurations, DCLK is the clock input used to clock data from an external source into the target device. Data is latched into the device on the rising edge of DCLK. DCLK N/A Synchronous configuration schemes (PS, FPP, AS) In AS mode, DCLK is an output from the Stratix IV device that provides timing for the configuration interface. In AS mode, DCLK has an internal pull-up resistor (typically 25 k) that is always active. Input (PS, FPP) Output (AS) In AS configuration schemes, this pin is driven into an inactive state after configuration completes. You can use this pin as a user I/O during user mode. In PS or FPP schemes that use a control host, you must drive DCLK either high or low, whichever is more convenient. In passive schemes, you cannot use DCLK as a user I/O during user mode. Toggling this pin after configuration does not affect the configured device. DATA0 N/A in AS mode. I/O in PS or FPP mode. Data input. In serial configuration modes, bit-wide configuration data is presented to the target device on the DATA0 pin. PS, FPP, AS Input In AS mode, DATA0 has an internal pull-up resistor that is always active. After PS or FPP configuration, DATA0 is available as a user I/O pin. The state of this pin depends on the Dual-Purpose Pin settings. Data inputs. Byte-wide configuration data is presented to the target device on DATA[7..0]. DATA[7..1] September 2012 I/O Altera Corporation Parallel configuration schemes (FPP) Inputs In serial configuration schemes, they function as user I/O pins during configuration, which means they are tri-stated. After FPP configuration, DATA[7..1] are available as user I/O pins. The state of these pins depends on the Dual-Purpose Pin settings. Stratix IV Device Handbook Volume 1 10-46 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins Table 10-11 lists the optional configuration pins. If these optional configuration pins are not enabled in the Quartus II software, they are available as general-purpose user I/O pins. Therefore, during configuration, these pins function as user I/O pins and are tri-stated with weak pull-up resistors. Table 10-11. Optional Configuration Pins Pin Name CLKUSR INIT_DONE DEV_OE DEV_CLRn User Mode N/A if option is on. I/O if option is off. N/A if option is on. I/O if option is off. N/A if option is on. I/O if option is off. N/A if option is on. I/O if option is off. Stratix IV Device Handbook Volume 1 Pin Type Input Description Optional user-supplied clock input synchronizes the initialization of one or more devices. Enable this pin by turning on the Enable user-supplied start-up clock (CLKUSR) option in the Quartus II software. Output open-drain Use as a status pin to indicate when the device has initialized and is in user mode. When nCONFIG is low and during the beginning of configuration, the INIT_DONE pin is tri-stated and pulled high due to an external 10-k pull-up resistor. After the option bit to enable INIT_DONE is programmed into the device (during the first frame of configuration data), the INIT_DONE pin goes low. When initialization is complete, the INIT_DONE pin is released and pulled high and the device enters user mode. Thus, the monitoring circuitry must be able to detect a low-to-high transition. Enable this pin by turning on the Enable INIT_DONE output option in the Quartus II software. Input Optional pin that allows you to override all tri-states on the device. When this pin is driven low, all I/O pins are tri-stated. When this pin is driven high, all I/O pins behave as programmed. Enable this pin by turning on the Enable device-wide output enable (DEV_OE) option in the Quartus II software. Input Optional pin that allows you to override all clears on all device registers. When this pin is driven low, all registers are cleared. When this pin is driven high, all registers behave as programmed. Enable this pin by turning on the Enable device-wide reset (DEV_CLRn) option in the Quartus II software. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Device Configuration Pins 10-47 Table 10-12 lists the dedicated JTAG pins. JTAG pins must be kept stable before and during configuration to prevent accidental loading of JTAG instructions. The TDI, TMS, and TRST pins have weak internal pull-up resistors, while TCK has a weak internal pull-down resistor (typically 25 k ). If you plan to use the SignalTap(R) embedded logic array analyzer, you must connect the JTAG pins of the Stratix IV device to a JTAG header on your board. Table 10-12. Dedicated JTAG Pins Pin Name User Mode TDI TDO N/A N/A Pin Type Test data input Test data output Description Serial input pin for instructions as well as test and programming data. Data is shifted on the rising edge of TCK. The TDI pin is powered by the 2.5-V/3.0-V VCCPD supply. If the JTAG interface is not required on your board, you can disable the JTAG circuitry by connecting this pin to logic high using a 1-k resistor. Serial data output pin for instructions as well as test and programming data. Data is shifted out on the falling edge of TCK. The pin is tri-stated if data is not being shifted out of the device. The TDO pin is powered by VCCPD. For recommendations about connecting a JTAG chain with multiple voltages across the devices in the chain, refer to the JTAG Boundary Scan Testing in Stratix IV Devices chapter. If the JTAG interface is not required on your board, you can disable the JTAG circuitry by leaving this pin unconnected. TMS N/A Input pin that provides the control signal to determine the transitions of the TAP controller state machine. TMS is evaluated on the rising edge of TCK. Therefore, you must set up TMS before the rising edge of TCK. Transitions within the state machine occur on the falling Test mode edge of TCK after the signal is applied to TMS. The TMS pin is powered by 2.5-V/3.0-V select VCCPD. If the JTAG interface is not required on your board, you can disable the JTAG circuitry by connecting this pin to logic high using a 1-k resistor. TCK N/A Test clock input Clock input to the BST circuitry. Some operations occur at the rising edge, while others occur at the falling edge. The TCK pin is powered by the 2.5-V/3.0-V VCCPD supply. It is expected that the clock input waveform have a nominal 50% duty cycle. If the JTAG interface is not required on your board, you can disable the JTAG circuitry by connecting TCK to GND. Active-low input to asynchronously reset the boundary-scan circuit. The TRST pin is optional according to IEEE Std. 1149.1. The TRST pin is powered by the 2.5-V/3.0-V VCCPD supply. TRST N/A Test reset input Hold TMS at 1 or keep TCK static while TRST is changed from 0 to 1. (optional) If the JTAG interface is not required on your board, you can disable the JTAG circuitry by connecting the TRST pin to GND. f For more information about the pin connection recommendations, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-48 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Data Decompression Configuration Data Decompression Stratix IV devices support configuration data decompression, which saves configuration memory space and time. This feature allows you to store compressed configuration data in configuration devices or other memory and transmit this compressed bitstream to Stratix IV devices. During configuration, the Stratix IV device decompresses the bitstream in real time and programs its SRAM cells. 1 Preliminary data indicates that compression typically reduces the configuration bitstream size by 30% to 55% based on the designs used. Stratix IV devices support decompression in the FPP (when using a MAX II device or microprocessor + flash), fast AS, and PS configuration schemes. The Stratix IV decompression feature is not available in the JTAG configuration scheme. In PS mode, use the Stratix IV decompression feature because sending compressed configuration data reduces configuration time. When you enable compression, the Quartus II software generates configuration files with compressed configuration data. This compressed file reduces the storage requirements in the configuration device or flash memory, and decreases the time needed to transmit the bitstream to the Stratix IV device. The time required by a Stratix IV device to decompress a configuration file is less than the time needed to transmit the configuration data to the device. There are two ways to enable compression for Stratix IV bitstreams--before design compilation (in the Compiler Settings menu) and after design compilation (in the Convert Programming Files window). To enable compression in the project's Compiler Settings menu, follow these steps: 1. On the Assignments menu, click Device to bring up the Settings dialog box. 2. After selecting your Stratix IV device, open the Device and Pin Options window. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Configuration Data Decompression 10-49 3. In the Configuration settings tab, turn on Generate compressed bitstreams (as shown in Figure 10-19). Figure 10-19. Enabling Compression for Stratix IV Bitstreams in Compiler Settings You can also enable compression when creating programming files from the Convert Programming Files window. To do this, follow these steps: 1. On the File menu, click Convert Programming Files. 2. Select the programming file type (.pof, .sram, .hex, .rbf, or .ttf). 3. For .pof output files, select a configuration device. 4. In the Input files to convert box, select SOF Data. 5. Select Add File and add a Stratix IV device .sof file. 6. Select the name of the file you added to the SOF Data area and click Properties. 7. Check the Compression check box. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-50 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrades When multiple Stratix IV devices are cascaded, you can selectively enable the compression feature for each device in the chain if you are using a serial configuration scheme. Figure 10-20 shows a chain of two Stratix IV devices. The first Stratix IV device has compression enabled; therefore, receives a compressed bitstream from the configuration device. The second Stratix IV device has the compression feature disabled and receives uncompressed data. In a multi-device FPP configuration chain (with a MAX II device or microprocessor + flash), all Stratix IV devices in the chain must either enable or disable the decompression feature. You cannot selectively enable the compression feature for each device in the chain because of the DATA and DCLK relationship. Figure 10-20. Compressed and Uncompressed Configuration Data in the Same Configuration File Serial Configuration Data Serial Configuration Device Uncompressed Configuration Data Compressed Configuration Data Decompression Controller Stratix IV Device Stratix IV Device nCE nCEO nCE nCEO N.C. GND You can generate programming files for this setup by clicking Convert Programming Files on the File menu in the Quartus II software. Remote System Upgrades This section describes the functionality and implementation of the dedicated remote system upgrade circuitry. It also defines several concepts related to remote system upgrade, including factory configuration, application configuration, remote update mode, and user watchdog timer. Additionally, this section provides design guidelines for implementing remote system upgrades with the supported configuration schemes. System designers sometimes face challenges such as shortened design cycles, evolving standards, and system deployments in remote locations. Stratix IV devices help overcome these challenges with their inherent reprogrammability and dedicated circuitry to perform remote system upgrades. Remote system upgrades help deliver feature enhancements and bug fixes without costly recalls, reduce time-to-market, extend product life, and avoid system downtime. Stratix IV devices feature dedicated remote system upgrade circuitry. Soft logic (either the Nios(R) II embedded processor or user logic) implemented in a Stratix IV device can download a new configuration image from a remote location, store it in configuration memory, and direct the dedicated remote system upgrade circuitry to initiate a reconfiguration cycle. The dedicated circuitry performs error detection during and after the configuration process, recovers from any error condition by reverting back to a safe configuration image, and provides error status information. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrades 10-51 Remote system upgrade is supported in fast AS Stratix IV configuration schemes. You can also implement remote system upgrade in conjunction with advanced Stratix IV features such as real-time decompression of configuration data and design security using the advanced encryption standard (AES) for secure and efficient field upgrades. The largest serial configuration device currently supports 128 Mbits of configuration bitstream. 1 Stratix IV devices only support remote system upgrade in the single device fast AS configuration scheme. Because the largest serial configuration device currently supports 128 Mbits of configuration bitstream, the remote system upgrade feature is not supported in EP4SGX290, EP4SE360, and larger devices. 1 The remote system upgrade feature is not supported in a multi-device chain. Functional Description The dedicated remote system upgrade circuitry in Stratix IV devices manages remote configuration and provides error detection, recovery, and status information. User logic or a Nios II processor implemented in the Stratix IV device logic array provides access to the remote configuration data source and an interface to the system's configuration memory. Stratix IV devices have remote system upgrade processes that involve the following steps: 1. A Nios II processor (or user logic) implemented in the Stratix IV device logic array receives new configuration data from a remote location. The connection to the remote source uses a communication protocol such as the transmission control protocol/Internet protocol (TCP/IP), peripheral component interconnect (PCI), user datagram protocol (UDP), universal asynchronous receiver/transmitter (UART), or a proprietary interface. 2. The Nios II processor (or user logic) stores this new configuration data in non-volatile configuration memory. 3. The Nios II processor (or user logic) initiates a reconfiguration cycle with the new or updated configuration data. 4. The dedicated remote system upgrade circuitry detects and recovers from any error(s) that might occur during or after the reconfiguration cycle and provides error status information to the user design. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-52 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrades Figure 10-21 shows the steps required for performing remote configuration updates. (The numbers in Figure 10-21 coincide with the steps just mentioned.) Figure 10-21. Functional Diagram of Stratix IV Remote System Upgrade 1 2 Development Location Data Stratix IV Device Control Module Data Configuration Memory Data Stratix IV Configuration 3 Figure 10-22 shows a block diagram for implementing a remote system upgrade with the Stratix IV fast AS configuration scheme. Figure 10-22. Remote System Upgrade Block Diagram for Stratix IV Fast AS Configuration Scheme Stratix IV Device Nios II Processor or User Logic Serial Configuration Device You must set the mode select pins (MSEL[2..0]) to fast AS mode to use remote system upgrade in your system. Table 10-13 lists the MSEL pin settings for Stratix IV devices in standard configuration mode and remote system upgrade mode. The following sections describe remote update of the remote system upgrade mode. For more information about standard configuration schemes supported in Stratix IV devices, refer to "Configuration Schemes" on page 10-2. Table 10-13. Remote System Upgrade Modes in Stratix IV Devices Configuration Scheme Fast AS (40 MHz) MSEL[2..0] Remote System Upgrade Mode 011 Standard 011 Remote update (1) Note to Table 10-13: (1) All EPCS densities are able to support DCLK up to 40 MHz, but batches of EPCS1 and EPCS4 manufactured on 0.18-m process geometry can only support DCLK up to 20 MHz. For more information, refer to the Serial Configuration Devices (EPCS1, EPCS4, EPCS16, EPCS64, and EPCS128) Data Sheet chapter in volume 2 of the Configuration Handbook. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrades 1 10-53 When using fast AS mode, you must select remote update mode in the Quartus II software and insert the ALTREMOTE_UPDATE megafunction to access the circuitry. For more information, refer to "ALTREMOTE_UPDATE Megafunction" on page 10-62. Enabling Remote Update You can enable remote update for Stratix IV devices in the Quartus II software before design compilation (in the Compiler Settings menu). In remote update mode, the auto-restart configuration after error option is always enabled. To enable remote update in the project's compiler settings, in the Quartus II software, follow these steps: 1. On the Assignment menu, click Device. The Settings dialog box appears. 2. Click Device and Pin Options. The Device and Pin Options dialog box appears. 3. Click the Configuration tab. 4. From the Configuration scheme list, select Active Serial (you can also use Configuration Device) (Figure 10-23). 5. From the Configuration Mode list, select Remote (Figure 10-23). 6. Click OK. 7. In the Settings dialog box, click OK. Figure 10-23. Enabling Remote Update for Stratix IV Devices in the Compiler Settings Menu September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-54 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrade Mode Configuration Image Types When performing a remote system upgrade, Stratix IV device configuration bitstreams are classified as factory configuration images or application configuration images. An image, also referred to as a configuration, is a design loaded into the Stratix IV device that performs certain user-defined functions. Each Stratix IV device in your system requires one factory image or the addition of one or more application images. The factory image is a user-defined fall-back, or safe configuration, and is responsible for administering remote updates in conjunction with the dedicated circuitry. Application images implement user-defined functionality in the target Stratix IV device. You may include the default application image functionality in the factory image. A remote system upgrade involves storing a new application configuration image or updating an existing one using the remote communication interface. After an application configuration image is stored or updated remotely, the user design in the Stratix IV device initiates a reconfiguration cycle with the new image. Any errors during or after this cycle are detected by the dedicated remote system upgrade circuitry and cause the device to automatically revert to the factory image. The factory image then performs error processing and recovery. The factory configuration is written to the serial configuration device only once by the system manufacturer and must not be remotely updated. On the other hand, application configurations may be remotely updated in the system. Both images can initiate system reconfiguration. Remote System Upgrade Mode Remote system upgrade has one mode of operation--remote update mode. Remote update mode allows you to determine the functionality of your system after power-up and offers several features. Remote Update Mode In remote update mode, Stratix IV devices load the factory configuration image after power up. The user-defined factory configuration determines which application configuration is to be loaded and triggers a reconfiguration cycle. The factory configuration may also contain application logic. When used with serial configuration devices, remote update mode allows an application configuration to start at any flash sector boundary. For example, this translates to a maximum of 128 sectors in the EPCS64 device and 32 sectors in the EPCS16 device, where the minimum size of each page is 512 KBits. Altera recommends not using the same page in the serial configuration devices for two images. Additionally, remote update mode features a user watchdog timer that determines the validity of an application configuration. When a Stratix IV device is first powered up in remote update mode, it loads the factory configuration located at page zero (page registers PGM[23:0] = 24'b0). Always store the factory configuration image for your system at page address zero. This corresponds to the start address location 0x000000 in the serial configuration device. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrade Mode 10-55 The factory image is user-designed and contains soft logic to: Process any errors based on status information from the dedicated remote system upgrade circuitry Communicate with the remote host and receive new application configurations and store this new configuration data in the local non-volatile memory device Determine which application configuration is to be loaded into the Stratix IV device Enable or disable the user watchdog timer and load its time-out value (optional) Instruct the dedicated remote system upgrade circuitry to initiate a reconfiguration cycle Figure 10-24 shows the transitions between the factory and application configurations in remote update mode. Figure 10-24. Transitions Between Configurations in Remote Update Mode Configuration Error Set Control Register and Reconfigure Power Up Configuration Error Factory Configuration (page 0) Application 1 Configuration Reload a Different Application Reload a Different Application Set Control Register and Reconfigure Application n Configuration Configuration Error After power up or a configuration error, the factory configuration logic is loaded automatically. The factory configuration also must specify whether to enable the user watchdog timer for the application configuration and if enabled, to include the timer setting information. The user watchdog timer ensures that the application configuration is valid and functional. The timer must be continually reset within a specific amount of time during user mode operation of an application configuration. Only valid application configurations contain the logic to reset the timer in user mode. This timer reset logic must be part of a user-designed hardware and/or software health monitoring signal that indicates error-free system operation. If the timer is not reset in a specific amount of time; for example, the user application configuration detects a functional problem or if the system hangs, the dedicated circuitry updates the remote system upgrade status register, triggering the loading of the factory configuration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-56 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Remote System Upgrade Mode 1 The user watchdog timer is automatically disabled for factory configurations. For more information about the user watchdog timer, refer to "User Watchdog Timer" on page 10-61. If there is an error while loading the application configuration, the cause of the reconfiguration is written by the dedicated circuitry to the remote system upgrade status register. Actions that cause the remote system upgrade status register to be written are: nSTATUS driven low externally Internal CRC error User watchdog timer time-out A configuration reset (logic array nCONFIG signal or external nCONFIG pin assertion to low) Stratix IV devices automatically load the factory configuration located at page address zero. This user-designed factory configuration can read the remote system upgrade status register to determine the reason for the reconfiguration. The factory configuration then takes appropriate error recovery steps and writes to the remote system upgrade control register to determine the next application configuration to be loaded. When Stratix IV devices successfully load the application configuration, they enter into user mode. In user mode, the soft logic (Nios II processor or state machine and the remote communication interface) assists the Stratix IV device in determining when a remote system update is arriving. When a remote system update arrives, the soft logic receives the incoming data, writes it to the configuration memory device, and triggers the device to load the factory configuration. The factory configuration reads the remote system upgrade status register and control register, determines the valid application configuration to load, writes the remote system upgrade control register accordingly, and initiates system reconfiguration. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry 10-57 Dedicated Remote System Upgrade Circuitry This section describes the implementation of the Stratix IV remote system upgrade dedicated circuitry. The remote system upgrade circuitry is implemented in hard logic. This dedicated circuitry interfaces to the user-defined factory and application configurations implemented in the Stratix IV device logic array to provide the complete remote configuration solution. The remote system upgrade circuitry contains the remote system upgrade registers, a watchdog timer, and a state machine that controls those components. Figure 10-25 shows the data path for the remote system upgrade block. Figure 10-25. Remote System Upgrade Circuit Data Path (Note 1) Internal Oscillator Status Register (SR) [4..0] Control Register [37..0] Logic Array Update Register [37..0] update Shift Register dout Bit [4..0] din dout capture RSU State Machine din Bit [37..0] capture time-out User Watchdog Timer clkout capture update Logic Array clkin RU_DOUT RU_SHIFTnLD RU_CAPTnUPDT RU_CLK RU_DIN RU_nCONFIG RU_nRSTIMER Logic Array Note to Figure 10-25: (1) The RU_DOUT, RU_SHIFTnLD, RU_CAPTnUPDT, RU_CLK, RU_DIN, RU_nCONFIG, and RU_nRSTIMER signals are internally controlled by the ALTREMOTE_UPDATE megafunction. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-58 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry Remote System Upgrade Registers The remote system upgrade block contains a series of registers that store the page addresses, watchdog timer settings, and status information. Table 10-14 lists these registers. Table 10-14. Remote System Upgrade Registers Register Description Shift register This register is accessible by the logic array and allows the update, status, and control registers to be written and sampled by user logic. Control register This register contains the current page address, user watchdog timer settings, and one bit specifying whether the current configuration is a factory configuration or an application configuration. During a read operation in an application configuration, this register is read into the shift register. When a reconfiguration cycle is initiated, the contents of the update register are written into the control register. Update register This register contains data similar to that in the control register. However, it can only be updated by the factory configuration by shifting data into the shift register and issuing an update operation. When a reconfiguration cycle is triggered by the factory configuration, the control register is updated with the contents of the update register. During a capture in a factory configuration, this register is read into the shift register. Status register This register is written to by the remote system upgrade circuitry on every reconfiguration to record the cause of the reconfiguration. This information is used by the factory configuration to determine the appropriate action following a reconfiguration. During a capture cycle, this register is read into the shift register. The remote system upgrade control and status registers are clocked by the 10-MHz internal oscillator (the same oscillator that controls the user watchdog timer). However, the remote system upgrade shift and update registers are clocked by the user clock input (RU_CLK). Remote System Upgrade Control Register The remote system upgrade control register stores the application configuration page address and user watchdog timer settings. The control register functionality depends on the remote system upgrade mode selection. In remote update mode, the control register page address bits are set to all zeros (24'b0 = 0x000000) at power up to load the factory configuration. A factory configuration in remote update mode has write access to this register. Figure 10-26 and Table 10-15 specify the control register bit positions. In the figure, the numbers show the bit position of a setting within a register. For example, bit number 25 is the enable bit for the watchdog timer. Figure 10-26. Remote System Upgrade Control Register 37 36 35 34 33 32 31 30 29 28 27 26 Wd_timer[11..0] Stratix IV Device Handbook Volume 1 25 Wd_en 24 23 22 .. 3 PGM[23..0] 2 1 0 AnF September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry 10-59 The application-not-factory (AnF) bit indicates whether the current configuration loaded in the Stratix IV device is the factory configuration or an application configuration. This bit is set low by the remote system upgrade circuitry when an error condition causes a fall-back to the factory configuration. When the AnF bit is high, the control register access is limited to read operations. When the AnF bit is low, the register allows write operations and disables the watchdog timer. In remote update mode, the factory configuration design sets this bit high (1'b1) when updating the contents of the update register with the application page address and watchdog timer settings. Table 10-15 lists the remote system upgrade control register contents. Table 10-15. Remote System Upgrade Control Register Contents Remote System Upgrade Mode Value (2) AnF (1) Remote update 1'b0 PGM[23..0] Remote update 24'b0x000000 AS configuration start address (StAdd[23..0]) Wd_en Remote update 1'b0 User watchdog timer enable bit Control Register Bit Definition Application not factory User watchdog time-out value Remote update Wd_timer[11..0] 12'b000000000000 (most significant 12 bits of 29-bit count value: {Wd_timer[11..0], 17'b0}) Notes to Table 10-15: (1) In remote update mode, the remote configuration block does not update the AnF bit automatically (you can update it manually). (2) This is the default value of the control register bit. Remote System Upgrade Status Register The remote system upgrade status register specifies the reconfiguration trigger condition. The various trigger and error conditions include: September 2012 Cyclic redundancy check (CRC) error during application configuration nSTATUS assertion by an external device due to an error Stratix IV device logic array triggered a reconfiguration cycle, possibly after downloading a new application configuration image External configuration reset (nCONFIG) assertion User watchdog timer time-out Altera Corporation Stratix IV Device Handbook Volume 1 10-60 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry Figure 10-27 and Table 10-16 specify the contents of the status register. The numbers in the figure show the bit positions within a 5-bit register. Figure 10-27. Remote System Upgrade Status Register 4 Wd 3 2 1 nCONFIG Core_nCONFIG nSTATUS 0 CRC Table 10-16. Remote System Upgrade Status Register Contents Status Register Bit Definition POR Reset Value CRC (from the configuration) CRC error caused reconfiguration 1 bit '0' nSTATUS nSTATUS caused reconfiguration 1 bit '0' Device logic array caused reconfiguration 1 bit '0' nCONFIG caused reconfiguration 1 bit '0' Watchdog timer caused reconfiguration 1 bit '0' CORE_nCONFIG (1) nCONFIG Wd Note to Table 10-16: (1) Logic array reconfiguration forces the system to load the application configuration data into the Stratix IV device. This occurs after the factory configuration specifies the appropriate application configuration page address by updating the update register. Remote System Upgrade State Machine The remote system upgrade control and update registers have identical bit definitions, but serve different roles (refer to Table 10-14 on page 10-57). While both registers can only be updated when the device is loaded with a factory configuration image, the update register writes are controlled by the user logic; the control register writes are controlled by the remote system upgrade state machine. In factory configurations, the user logic sends the AnF bit (set high), the page address, and the watchdog timer settings for the next application configuration bit to the update register. When the logic array configuration reset (RU_nCONFIG) goes low, the remote system upgrade state machine updates the control register with the contents of the update register and initiates system reconfiguration from the new application page. 1 To ensure successful reconfiguration between the pages, assert the RU_nCONFIG signal for a minimum of 250 ns. This is equivalent to strobing the reconfiguration input of the ALTREMOTE_UPDATE megafunction high for a minimum of 250 ns. In the event of an error or reconfiguration trigger condition, the remote system upgrade state machine directs the system to load a factory or application configuration (page zero or page one, based on the mode and error condition) by setting the control register accordingly. Table 10-17 lists the contents of the control register after such an event occurs for all possible error or trigger conditions. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry 10-61 The remote system upgrade status register is updated by the dedicated error monitoring circuitry after an error condition but before the factory configuration is loaded. Table 10-17. Control Register Contents after an Error or Reconfiguration Trigger Condition Reconfiguration Error/Trigger Control Register Setting Remote Update nCONFIG reset All bits are 0 nSTATUS error All bits are 0 CORE triggered reconfiguration Update register CRC error All bits are 0 Wd time out All bits are 0 Capture operations during factory configuration access the contents of the update register. This feature is used by the user logic to verify that the page address and watchdog timer settings were written correctly. Read operations in application configurations access the contents of the control register. This information is used by the user logic in the application configuration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-62 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Dedicated Remote System Upgrade Circuitry User Watchdog Timer The user watchdog timer prevents a faulty application configuration from stalling the device indefinitely. The system uses the timer to detect functional errors after an application configuration is successfully loaded into the Stratix IV device. The user watchdog timer is a counter that counts down from the initial value loaded into the remote system upgrade control register by the factory configuration. The counter is 29 bits wide and has a maximum count value of 229. When specifying the user watchdog timer value, specify only the most significant 12 bits. The granularity of the timer setting is 217 cycles. The cycle time is based on the frequency of the 10-MHz internal oscillator. Table 10-18 lists the operating range of the 10-MHz internal oscillator. Table 10-18. 10-MHz Internal Oscillator Specifications (Note 1) Minimum Typical Maximum Units 4.3 5.3 10 MHz Note to Table 10-18: (1) These values are preliminary. The user watchdog timer begins counting after the application configuration enters device user mode. This timer must be periodically reloaded or reset by the application configuration before the timer expires by asserting RU_nRSTIMER. If the application configuration does not reload the user watchdog timer before the count expires, a time-out signal is generated by the remote system upgrade dedicated circuitry. The time-out signal tells the remote system upgrade circuitry to set the user watchdog timer status bit (Wd) in the remote system upgrade status register and reconfigures the device by loading the factory configuration. 1 To allow remote system upgrade dedicated circuitry to reset the watchdog timer, you must assert the RU_nRSTIMER signal active for a minimum of 250 ns. This is equivalent to strobing the reset_timer input of the ALTREMOTE_UPDATE megafunction high for a minimum of 250 ns. The user watchdog timer is not enabled during the configuration cycle of the device. Errors during configuration are detected by the CRC engine. Also, the timer is disabled for factory configurations. Functional errors should not exist in the factory configuration because it is stored and validated during production and is never updated remotely. 1 Stratix IV Device Handbook Volume 1 The user watchdog timer is disabled in factory configurations and during the configuration cycle of the application configuration. It is enabled after the application configuration enters user mode. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Quartus II Software Support 10-63 Quartus II Software Support The Quartus II software provides the flexibility to include the remote system upgrade interface between the Stratix IV device logic array and the dedicated circuitry, generate configuration files for production, and allows remote programming of the system configuration memory. The ALTREMOTE_UPDATE megafunction is the implementation option in the Quartus II software that you use for the interface between the remote system upgrade circuitry and the device logic array interface. Using the megafunction block instead of creating your own logic saves design time and offers more efficient logic synthesis and device implementation. ALTREMOTE_UPDATE Megafunction The ALTREMOTE_UPDATE megafunction provides a memory-like interface to the remote system upgrade circuitry and handles the shift register read and write protocol in the Stratix IV device logic. This implementation is suitable for designs that implement the factory configuration functions using a Nios II processor or user logic in the device. Figure 10-28 shows the interface signals between the ALTREMOTE_UPDATE megafunction and Nios II processor or user logic. Figure 10-28. Interface Signals between the ALTREMOTE_UPDATE Megafunction and the Nios II Processor ALTREMOTE_UPDATE read_param write_param param[2..0] data_in[23..0] Nios II Processor or User Logic reconfig reset_timer clock reset busy data_out[23..0] f For more information about the ALTREMOTE_UPDATE megafunction and the description of ports shown in Figure 10-28, refer to the Remote Update Circuitry (ALTREMOTE_UPDATE) Megafunction User Guide. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-64 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security Design Security This section provides an overview of the design security feature and its implementation on Stratix IV devices using the advanced encryption standard (AES). It also covers the new security modes available in Stratix IV devices. As Stratix IV devices continue play a role in larger and more critical designs in competitive commercial and military environments, it is increasingly important to protect the designs from copying, reverse engineering, and tampering. Stratix IV devices address these concerns with both volatile and non-volatile security feature support. Stratix IV devices have the ability to decrypt configuration bitstreams using the AES algorithm, an industry-standard encryption algorithm that is FIPS-197 certified. Stratix IV devices have a design security feature that utilizes a 256-bit security key. Stratix IV devices store configuration data in SRAM configuration cells during device operation. Because SRAM is volatile, the SRAM cells must be loaded with configuration data each time the device powers up. It is possible to intercept configuration data when it is being transmitted from the memory source (flash memory or a configuration device) to the device. The intercepted configuration data could then be used to configure another device. When using the Stratix IV design security feature, the security key is stored in the Stratix IV device. Depending on the security mode, you can configure the Stratix IV device using a configuration file that is encrypted with the same key, or for board testing, configured with a normal configuration file. The design security feature is available when configuring Stratix IV devices using FPP configuration mode with an external host (such as a MAX II device or microprocessor), or when using fast AS or PS configuration schemes. The design security feature is also available in remote update with fast AS configuration mode. The design security feature is not available when you are configuring your Stratix IV device using JTAG-based configuration. For more information, refer to "Supported Configuration Schemes" on page 10-67. 1 Stratix IV Device Handbook Volume 1 When using a serial configuration scheme such as PS or fast AS, configuration time is the same whether or not you enable the design security feature. If the FPP scheme is used with the design security or decompression feature, a x4 DCLK is required. This results in a slower configuration time when compared with the configuration time of a Stratix IV device that has neither the design security nor the decompression feature enabled. September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security 10-65 Stratix IV Security Protection Stratix IV device designs are protected from copying, reverse engineering, and tampering using configuration bitstream encryption. Security Against Copying The security key is securely stored in the Stratix IV device and cannot be read out through any interfaces. In addition, as configuration file read-back is not supported in Stratix IV devices, the design information cannot be copied. Security Against Reverse Engineering Reverse engineering from an encrypted configuration file is very difficult and time consuming because the Stratix IV configuration file formats are proprietary and the file contains millions of bits which require specific decryption. Reverse engineering the Stratix IV device is just as difficult because the device is manufactured on the most advanced 40-nm process technology. Security Against Tampering The non-volatile keys are one-time programmable. After the Tamper Protection bit is set in the key programming file generated by the Quartus II software, the Stratix IV device can only be configured with configuration files encrypted with the same key. AES Decryption Block The main purpose of the AES decryption block is to decrypt the configuration bitstream prior to entering data decompression or configuration. Prior to receiving encrypted data, you must enter and store the 256-bit security key in the device. You can choose between a non-volatile security key and a volatile security key with battery backup. The security key is scrambled prior to storing it in the key storage to make it more difficult for anyone to retrieve the stored key using de-capsulation of the device. Flexible Security Key Storage Stratix IV devices support two types of security key programming--volatile and non-volatile keys. Table 10-19 lists the differences between volatile keys and non-volatile keys. Table 10-19. Security Key Options (Part 1 of 2) Options September 2012 Volatile Key Non-Volatile Key Key programmability Reprogrammable and erasable One-time programmable External battery Required Not required Key programming method (1) On-board On and off board Altera Corporation Stratix IV Device Handbook Volume 1 10-66 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security Table 10-19. Security Key Options (Part 2 of 2) Options Design protection Volatile Key Secure against copying and reverse engineering Non-Volatile Key Secure against copying and reverse engineering. Tamper resistant if tamper protection bit is set. Note to Table 10-19: (1) Key programming is carried out using the JTAG interface. You can program the non-volatile key to the Stratix IV device without an external battery. Also, there are no additional requirements to any of the Stratix IV power supply inputs. VCCBAT is a dedicated power supply for volatile key storage and not shared with other on-chip power supplies, such as VCCIO or VCC. VCCBAT continuously supplies power to the volatile register regardless of the on-chip supply condition. 1 After power-up, you must wait 300 ms (PORSEL = 0) or 12 ms (PORSEL = 1) before beginning key programming to ensure that VCCBAT is at full rail. 1 For more information about how to calculate the key retention time of the battery used for volatile key storage, refer to the Stratix III, Stratix IV, Stratix V, HardCopy III and HardCopy IV PowerPlay Early Power Estimator. f For more information about battery specifications, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. f For more information about the VCCBAT pin connection recommendations, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. Stratix IV Design Security Solution Stratix IV devices are SRAM-based devices. To provide design security, Stratix IV devices require a 256-bit security key for configuration bitstream encryption. You can carry out secure configuration in the following steps, as shown in Figure 10-29: 1. Program the security key into the Stratix IV device. 2. Program the user-defined 256-bit AES keys to the Stratix IV device through the JTAG interface. 3. Encrypt the configuration file and store it in the external memory. 4. Encrypt the configuration file with the same 256-bit keys used to program the Stratix IV device. Encryption of the configuration file is done using the Quartus II software. The encrypted configuration file is then loaded into the external memory, such as a configuration or flash device. 5. Configure the Stratix IV device. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security 10-67 At system power-up, the external memory device sends the encrypted configuration file to the Stratix IV device. Figure 10-29. Design Security (Note 1) Stratix IV Device User-Defined Step 1 Key Storage AES Key AES Decryption Step 3 Encrypted Step 2 Memory or Configuration Configuration File Device Note to Figure 10-29: (1) Step 1, Step 2, and Step 3 correspond to the procedure described in "Design Security" on page 10-63. Security Modes Available The following security modes are available on the Stratix IV device. Volatile Key Secure operation with volatile key programmed and required external battery: this mode accepts both encrypted and unencrypted configuration bitstreams. Use the unencrypted configuration bitstream support for board-level testing only. Non-Volatile Key Secure operation with one time programmable (OTP) security key programmed: this mode accepts both encrypted and unencrypted configuration bitstreams. Use the unencrypted configuration bitstream support for board level testing only. Non-Volatile Key with Tamper Protection Bit Set Secure operation in tamper resistant mode with OTP security key programmed: only encrypted configuration bitstreams are allowed to configure the device. Tamper protection disables JTAG configuration with unencrypted configuration bitstream. September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 10-68 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security 1 Enabling the tamper protection bit disables test mode in Stratix IV devices. This process is irreversible and prevents Altera from conducting carry-out failure analysis if test mode is disabled. Contact Altera Technical Support to enable the tamper protection bit. No Key Operation Only unencrypted configuration bitstreams are allowed to configure the device. Table 10-20 lists the different security modes and configuration bitstream supported for each mode. Table 10-20. Security Modes Supported Mode (1) Volatile key Non-volatile key Non-volatile key with tamper protection bit set Function Configuration File Secure Encrypted Board-level testing Unencrypted Secure Encrypted Board-level testing Unencrypted Secure (tamper resistant) (2) Encrypted Notes to Table 10-20: (1) In No key operation, only the unencrypted configuration file is supported. (2) The tamper protection bit setting does not prevent the device from being reconfigured. Supported Configuration Schemes The Stratix IV device supports only selected configuration schemes, depending on the security mode you select when you encrypt the Stratix IV device. Figure 10-30 shows the restrictions of each security mode when encrypting Stratix IV devices. Figure 10-30. Security Modes in Stratix IV Devices--Sequence and Restrictions No Key Volatile Key Unencrypted or Encrypted Configuration File Unencrypted Configuration File Non-Volatile Key Unencrypted or Encrypted Configuration File Non-Volatile Key with Tamper-Protection Bit Set Encrypted Configuration File Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security 10-69 Table 10-21 lists the configuration modes allowed in each of the security modes. Table 10-21. Allowed Configuration Modes for Various Security Modes Configuration File Security Mode No key Unencrypted Secure with volatile key Board-level testing with volatile key Encrypted Unencrypted Secure with non-volatile key Board-level testing with non-volatile key Encrypted Unencrypted Secure in tamper resistant mode using non-volatile key with tamper protection set Encrypted (Note 1) Allowed Configuration Modes All configuration modes that do not engage the design security feature. Passive serial with AES (and/or with decompression) Fast passive parallel with AES (and/or with decompression) Remote update fast AS with AES (and/or with decompression) Fast AS (and/or with decompression) All configuration modes that do not engage the design security feature. Passive serial with AES (and/or with decompression) Fast passive parallel with AES (and/or with decompression) Remote update fast AS with AES (and/or with decompression) Fast AS (and/or with decompression) All configuration modes that do not engage the design security feature. Passive serial with AES (and/or with decompression) Fast passive parallel with AES (and/or with decompression) Remote update fast AS with AES (and/or with decompression) Fast AS (and/or with decompression) Note to Table 10-21: (1) There is no impact to the configuration time required when compared with unencrypted configuration modes except FPP with AES (and/or decompression), which requires a DCLK that is x4 the data rate. You can use the design security feature with other configuration features, such as compression and remote system upgrade features. When you use compression with the design security feature, the configuration file is first compressed and then encrypted using the Quartus II software. During configuration, the Stratix IV device first decrypts and then decompresses the configuration file. Document Revision History Table 10-22 lists the revision history for this chapter. Table 10-22. Document Revision History (Part 1 of 2) Date Version September 2012 December 2011 September 2012 3.5 3.4 Altera Corporation Changes Updated the "FPP Configuration Using a MAX II Device as an External Host" section to close FB #36583 and #63157. Updated the "Estimating Active Serial Configuration Time" section to close FB #64163. Updated the "PS Configuration Using a MAX II Device as an External Host" section to close FB #63157. Updated Figure 10-1, Figure 10-2, Figure 10-3, Figure 10-10,Figure 10-11 and Figure 10-12 to close FB #63155. Updated Table 10-2 and Table 10-7. Stratix IV Device Handbook Volume 1 10-70 Chapter 10: Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices Design Security Table 10-22. Document Revision History (Part 2 of 2) Date Version April 2011 February 2011 March 2010 November 2009 June 2009 3.2 3.1 2.3 March 2009 2.1 Stratix IV Device Handbook Volume 1 Updated the "FPP Configuration Using a MAX II Device as an External Host", "Fast Active Serial Configuration (Serial Configuration Devices)", and "PS Configuration Using a MAX II Device as an External Host". Updated Table 10-10. Updated the "Fast Active Serial Configuration (Serial Configuration Devices)", "FPP Configuration Using a MAX II Device as an External Host" "Configuration Data Decompression", and "User Watchdog Timer" sections. Updated Table 10-2, Table 10-4, Table 10-5, Table 10-7, and Table 10-9. Applied new template. Minor text edits. Added the "Guidelines for Connecting Serial Configuration Devices on an AS Interface" section. Updated the "Power-On Reset Circuit" and "Fast Active Serial Configuration (Serial Configuration Devices)" sections. Updated Table 10-2, Table 10-4, Table 10-5, Table 10-10, and Table 10-13. Updated Figure 10-16 and Figure 10-17 with Note 5. Updated Figure 10-4, Figure 10-5, and Figure 10-13. Updated the reference in the "Configuration Schemes" section. Updated Table 10-1 and Table 10-2. Updated the "FPP Configuration Using a MAX II Device as an External Host","Fast Active Serial Configuration (Serial Configuration Devices)", "Device Configuration Pins", "Remote System Upgrades", "Remote System Upgrade Mode", "Estimating Active Serial Configuration Time", "Remote System Upgrade State Machine", and "User Watchdog Timer" sections. Removed Table 10-4, Table 10-7, Table 10-8, and Table 10-25. Minor text edits. Updated the "VCCPD Pins", "FPP Configuration Using a MAX II Device as an External Host", "Estimating Active Serial Configuration Time", "Fast Active Serial Configuration (Serial Configuration Devices)", "Remote System Upgrades", "PS Configuration Using a MAX II Device as an External Host", and "PS Configuration Using a Download Cable" sections. Updated Table 10-3, Table 10-13 and Table 10-2. Added introductory sentences to improve search ability. Removed the Conclusion section. Minor text edits. Updated Table 10-2. Updated Table 10-1, Table 10-2, and Table 10-9. Removed "Referenced Documents" section. Updated "Fast Active Serial Configuration (Serial Configuration Devices)" and "JTAG Configuration" sections. Updated Figure 10-4, Figure 10-5, Figure 10-6, and Figure 10-13. Updated Table 10-2 and Table 10-13. 3.0 2.2 May 2008 3.3 April 2009 November 2008 Changes 2.0 1.0 Initial release. September 2012 Altera Corporation 11. SEU Mitigation in Stratix IV Devices February 2011 SIV51011-3.2 SIV51011-3.2 This chapter describes how to use the error detection cyclical redundancy check (CRC) feature when a Stratix(R) IV device is in user mode and recovers from CRC errors. The purpose of the error detection CRC feature in the Stratix IV device is to detect a flip in any of the configuration random access memory (CRAM) bits in Stratix IV devices due to a soft error. With the error detection circuitry, you can continuously verify the integrity of the configuration CRAM bits. In critical applications such as avionics, telecommunications, system control, and military applications, it is important to be able to do the following: 1 Confirm that the configuration data stored in a Stratix IV device is correct Alert the system to the occurrence of a configuration error The error detection feature is enhanced in the Stratix IV device family. Similar to Stratix III devices, the error detection and recovery time for single-event upset (SEU) in Stratix IV devices is reduced when compared with Stratix II devices. f For more information about test methodology for enhanced error detection in Stratix IV devices, refer to AN 539: Test Methodology of Error Detection and Recovery using CRC in Altera FPGA Devices. Dedicated circuitry is built into Stratix IV devices and consists of a CRC error detection feature that optionally checks for SEUs continuously and automatically. 1 For Stratix IV devices, the error detection CRC feature is provided in the Quartus(R) II software version 8.0 and onwards. Using error detection CRC for the Stratix IV device family has no impact on fitting or performance of your device. This chapter contains the following sections: "Error Detection Fundamentals" on page 11-2 "Configuration Error Detection" on page 11-2 "User Mode Error Detection" on page 11-2 "Error Detection Pin Description" on page 11-5 "Error Detection Block" on page 11-6 "Error Detection Timing" on page 11-8 "Recovering From CRC Errors" on page 11-11 (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 11-2 Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Fundamentals Error Detection Fundamentals Error detection determines whether the data received is corrupted during transmission. To accomplish this, the transmitter uses a function to calculate a checksum value for the data and appends the checksum to the original data frame. The receiver uses the same calculation methodology to generate a checksum for the received data frame and compares the received checksum to the transmitted checksum. If the two checksum values are equal, the received data frame is correct and no data corruption occurred during transmission or storage. The error detection CRC feature uses the same concept. When Stratix IV devices are configured successfully and are in user mode, the error detection CRC feature ensures the integrity of the configuration data. 1 There are two CRC error checks. One CRC error check always runs during configuration and a second optional CRC error check runs in the background in user mode. Both CRC error checks use the same CRC polynomial but different error detection implementations. For more information, refer to the "Configuration Error Detection" and "User Mode Error Detection" sections. Configuration Error Detection In configuration mode, a frame-based CRC is stored within the configuration data and contains the CRC value for each data frame. During configuration, the Stratix IV device calculates the CRC value based on the frame of data that is received and compares it against the frame CRC value in the data stream. Configuration continues until either the device detects an error or configuration is completed. In Stratix IV devices, the CRC value is calculated during the configuration stage. A parallel CRC engine generates 16 CRC check bits per frame and then stores them in CRAM. The CRAM chain used for storing the CRC check bits is 16 bits wide and its length is equal to the number of frames in the device. User Mode Error Detection Stratix IV devices have built-in error detection circuitry to detect data corruption by soft errors in the CRAM cells. This feature allows all CRAM contents to be read and verified to match a configuration-computed CRC value. Soft errors are changes in a CRAM bit state due to an ionizing particle. The error detection capability continuously computes the CRC of the configured CRAM bits and compares it with the pre-calculated CRC. If the CRCs match, there is no error in the current configuration CRAM bits. The process of error detection continues until the device is reset (by setting nCONFIG low). If you enable the CRC error detection option in the Quartus II software, after the device transitions into user mode, the error detection process is enabled. The internal 100 MHz configuration oscillator is divided down by a factor of two to 256 (at powers of two) to be used as the clock source during the error detection process. You must set the clock divide factor in the Quartus II software. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 11: SEU Mitigation in Stratix IV Devices User Mode Error Detection 11-3 A single 16-bit CRC calculation is done on a per-frame basis. After it has finished the CRC calculation for a frame, the resulting 16-bit signature is hex 0000 if there are no CRAM bit errors detected in a frame by the error detection circuitry and the output signal CRC_ERROR is 0. If a CRAM bit error is detected by the circuitry within a frame in the device, the resulting signature is non-zero. This causes the CRC engine to start searching for the error bit location. Error detection in Stratix IV devices calculates CRC check bits for each frame and pulls the CRC_ERROR pin high when it detects bit errors in the chip. Within a frame, it can detect all single-bit, double-bit, and three-bit errors. The probability of more than three CRAM bits being flipped by an SEU event is very low. In general, for all error patterns the probability of detection is 99.998%. The CRC engine reports the bit location and determines the type of error for all single-bit errors and over 99.641% of double-adjacent errors. The probability of other error patterns is very low and report of the location of bit flips is not guaranteed by the CRC engine. You can also read-out the error bit location through the JTAG and the core interface. Shift these bits out through either the SHIFT_EDERROR_REG JTAG instruction or the core interface before the CRC detects the next error in another frame. If the next frame also has an error, you must shift these bits out within the amount of time of one frame CRC verification. You can choose to extend this time interval by slowing down the error detection clock frequency, but this slows down the error recovery time for the SEU event. For the minimum update interval for Stratix IV devices, refer to Table 11-6 on page 11-9. If these bits are not shifted out before the next error location is found, the previous error location and error message is overwritten by the new information. The CRC circuit continues to run, and if an error is detected, you must decide whether to complete a reconfiguration or to ignore the CRC error. The error detection logic continues to calculate the CRC_ERROR and 16-bit signatures for the next frame of data regardless if any error has occurred in the current frame or not. You need to monitor these signals and take the appropriate actions if a soft error occurs. The error detection circuitry in Stratix IV devices uses a 16-bit CRC-ANSI standard (16-bit polynomial) as the CRC generator. The computed 16-bit CRC signature for each frame is stored in the registers within the core. The total storage register size is 16 (the number of bits per frame) x the number of frames. The Stratix IV device error detection feature does not check memory blocks and I/O buffers. Thus, the CRC_ERROR signal might stay solid high or low depending on the error status of the previously checked CRAM frame. The I/O buffers are not verified during error detection because these bits use flipflops as storage elements that are more resistant to soft errors when compared with CRAM cells. The support parity bits of MLAB, M9K, and M144K are used to check the contents of the memory blocks for any errors. The M144K TriMatrix memory block has a built-in error correction code block that checks and corrects the errors in the block. f For more information, refer to the TriMatrix Embedded Memory Blocks in Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 11-4 Chapter 11: SEU Mitigation in Stratix IV Devices User Mode Error Detection A JTAG instruction, EDERROR_INJECT, is provided to test the capability of the error detection block. This instruction is able to change the content of the 21-bit JTAG fault injection register that is used for error injection in Stratix IV devices, enabling the testing of the error detection block. 1 You can only execute the EDERROR_INJECT JTAG instruction when the device is in user mode. Table 11-1 lists the description of the EDERROR_INJECT JTAG instruction. Table 11-1. EDERROR_INJECT JTAG Instruction JTAG Instruction Instruction Code Description EDERROR_INJECT 00 0001 0101 This instruction controls the 21-bit JTAG fault injection register, which is used for error injection. You can create a JamTM file (.jam) to automate the testing and verification process. This allows you to verify the CRC functionality in-system, on-the-fly, without having to reconfigure the device. You can then switch to the CRC circuit to check for real errors induced by an SEU. You can introduce a single-error or double-errors adjacent to each other to the configuration memory. This provides an extra way to facilitate design verification and system fault tolerance characterization. Use the JTAG fault injection register with the EDERROR_INJECT instruction to flip the readback bits. The Stratix IV device is then forced into error test mode. The content of the JTAG fault injection register is not loaded into the fault injection register during the processing of the last and first frame. It is only loaded at the end of this period. 1 You can only introduce error injection in the first data frame, but you can monitor the error information at any time. For more information about the JTAG fault injection register and fault injection register, refer to "Error Detection Registers" on page 11-7. Table 11-2 lists how the fault injection register is implemented and describes error injection. Table 11-2. Fault Injection Register Bit Bit[20..19] Bit[18..8] Bit[7..0] Description Error Type Byte Location of the Injected Error Error Byte Value Depicts the location of the injected error in the first data frame. Depicts the location of the bit error and corresponds to the error injection type selection. Error Type (1) Error injection type Bit[20] Bit[19] 0 1 Single-byte error injection 1 0 Double-adjacent byte error injection 0 0 No error injection Content Note to Table 11-2: (1) Bit[20] and Bit[19] cannot both be set to 1 as this is not a valid selection. The error detection circuitry decodes this as no error injection. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Pin Description 1 11-5 After the test completes, Altera recommends reconfiguring the device. Automated Single-Event Upset Detection Stratix IV devices offer on-chip circuitry for automated checking of SEU detection. Some applications that require the device to operate error-free in high-neutron flux environments require periodic checks to ensure continued data integrity. The error detection CRC feature ensures data reliability and is one of the best options for mitigating SEU. You can implement the error detection CRC feature with existing circuitry in Stratix IV devices, eliminating the need for external logic. The CRC_ERROR pin reports a soft error when the configuration CRAM data is corrupted. You must decide whether to reconfigure the device or to ignore the error. Error Detection Pin Description Depending on the type of error detection feature you choose, you must use different error detection pins to monitor the data during user mode. CRC_ERROR Pin Table 11-3 describes the CRC_ERROR pin. Table 11-3. CRC_ERROR Pin Description Pin Name CRC_ERROR Pin Type I/O and open-drain 1 Description Active-high signal indicates that the error detection circuit has detected errors in the configuration CRAM bits. This pin is optional and is used when the error detection CRC circuit is enabled. When the error detection CRC circuit is disabled, it is a user I/O pin. To use the CRC_ERROR pin, you can either tie this pin to VCCPGM through a 10k resistor or, depending on the input voltage specification of the system receiving the signal, you can tie this pin to a different pull-up voltage. The WYSIWYG function performs optimization on the Verilog Quartus Mapping (VQM) netlist within the Quartus II software. f For more information about the stratixiv_crcblock WYSIWYG function, refer to the AN 539: Test Methodology of Error Detection and Recovery using CRC in Altera FPGA Devices. f For more information about the CRC_ERROR pin for Stratix IV devices, refer to Device Pin-Outs on the Altera website. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 11-6 Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Block Error Detection Block You can enable the Stratix IV device error detection block in the Quartus II software (refer to "Software Support" on page 11-10). This block contains the logic necessary to calculate the 16-bit CRC signature for the configuration CRAM bits in the device. The CRC circuit continues running even if an error occurs. When a soft error occurs, the device sets the CRC_ERROR pin high. Two types of CRC detection checks the configuration bits: 1 Stratix IV Device Handbook Volume 1 CRAM error checking ability (16-bit CRC), which occurs during user mode to be used by the CRC_ERROR pin. For each frame of data, the pre-calculated 16-bit CRC enters the CRC circuit at the end of the frame data and determines whether there is an error or not. If an error occurs, the search engine starts to find the location of the error. The error messages are shifted out through the JTAG instruction or core interface logics while the error detection block continues running. The JTAG interface reads out the 16-bit CRC result for the first frame and also shifts the 16-bit CRC bits to the 16-bit CRC storage registers for test purposes. Single error, double errors, or double-errors adjacent to each other are deliberately introduced to configuration memory for testing and design verification. 16-bit CRC that is embedded in every configuration data frame. During configuration, after a frame of data is loaded into the Stratix IV device, the pre-computed CRC is shifted into the CRC circuitry. At the same time, the CRC value for the data frame shifted-in is calculated. If the pre-computed CRC and calculated CRC values do not match, nSTATUS is set low. Every data frame has a 16-bit CRC; therefore, there are many 16-bit CRC values for the whole configuration bitstream. Every device has different lengths of configuration data frame. The "Error Detection Block" section describes the 16-bit CRC only when the device is in user mode. February 2011 Altera Corporation Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Block 11-7 Error Detection Registers There is one set of 16-bit registers in the error detection circuitry that stores the computed CRC signature. A non-zero value on the syndrome register causes the CRC_ERROR pin to be set high. Figure 11-1 shows the error detection circuitry, syndrome registers, and error injection block. Figure 11-1. Error Detection Block Diagram 16-Bit CRC Calculation and Error Readback bit stream with expected CRC included Syndrome Search Engine Error Detection State Machine 8 Register Control Signals 30 16 Error Message CRC_ERROR Register 46 Error Injection Block Fault Injection Register JTAG Update User Update Register Register JTAG Shift User Shift Register Register JTAG Fault Injection Register JTAG TDO General Routing Table 11-4 lists the registers shown in Figure 11-1. Table 11-4. Error Detection Registers (Part 1 of 2) Register Description Syndrome Register This register contains the CRC signature of the current frame through the error detection verification cycle. The CRC_ERROR signal is derived from the contents of this register. Error Message Register This 46-bit register contains information on the error type, location of the error, and the actual syndrome. The types of errors and location reported are single- and double-adjacent bit errors. The location bits for other types of errors are not identified by the error message register. The content of the register can be shifted out through the SHIFT_EDERROR_REG JTAG instruction or to the core through the core interface. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 11-8 Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Timing Table 11-4. Error Detection Registers (Part 2 of 2) Register Description JTAG Update Register This register is automatically updated with the contents of the error message register one cycle after the 46-bit register content is validated. It includes a clock enable that must be asserted prior to being sampled into the JTAG shift register. This requirement ensures that the JTAG update register is not being written into by the contents of the error message register at the same time that the JTAG shift register is reading its contents. User Update Register This register is automatically updated with the contents of the Error Message Register, one cycle after the 46-bit register content is validated. It includes a clock enable that must be asserted prior to being sampled into the User Shift Register. This requirement ensures that the User Update Register is not being written into by the contents of the Error Message Register at exactly the same time that the User Shift Register is reading its contents. JTAG Shift Register This register is accessible by the JTAG interface and allows the contents of the JTAG Update Register to be sampled and read by the JTAG instruction SHIFT_EDERROR_REG. User Shift Register This register is accessible by the core logic and allows the contents of the User Update Register to be sampled and read by user logic. JTAG Fault Injection Register This 21-bit register is fully controlled by the JTAG instruction EDERROR_INJECT. This register holds the information of the error injection that you want in the bitstream. Fault Injection Register The content of the JTAG Fault Injection Register is loaded into this 21-bit register when it is being updated. Error Detection Timing When you enable the CRC feature through the Quartus II software, the device automatically activates the CRC process after entering user mode, after configuration, and after initialization is complete. If an error is detected within a frame, CRC_ERROR is driven high at the end of the error location search, after the error message register is updated. At the end of this cycle, the CRC_ERROR pin is pulled low for a minimum of 32 clock cycles. If the next frame contains an error, CRC_ERROR is driven high again after the error message register is overwritten by the new value. You can start to unload the error message on each rising edge of the CRC_ERROR pin. Error detection runs until the device is reset. The error detection circuitry runs off an internal configuration oscillator with a divisor that sets the maximum frequency. Table 11-5 lists the minimum and maximum error detection frequencies based on the best performance of the internal configuration oscillator. Table 11-5. Minimum and Maximum Error Detection Frequencies Device Type Error Detection Frequency Maximum Error Detection Frequency Minimum Error Detection Frequency Valid Divisors (n) Stratix IV 100 MHz / 2n 50 MHz 390 kHz 1, 2, 3, 4, 5, 6, 7, 8 Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Timing 11-9 You can set a lower clock frequency by specifying a division factor in the Quartus II software (refer to "Software Support" on page 11-10). The divisor is a power of two, in which n is between 1 and 8. The divisor ranges from 2 through 256. Refer to Equation 11-1. Equation 11-1. 100 MHz error detection frequency = ----------------------n 2 1 The error detection frequency reflects the frequency of the error detection process for a frame because the CRC calculation in the Stratix IV device is done on a per-frame basis. You must monitor the error message to avoid missing information in the error message register. The error message register is updated whenever an error occurs. The minimum interval time between each update for the error message register depends on the device and the error detection clock frequency. Table 11-6 lists the estimated minimum interval time between each update for the error message register for Stratix IV devices. Table 11-6. Minimum Update Interval for Error Message Register (1) Device Timing Interval (s) EP4SGX70 13.8 EP4SGX110 13.8 EP4SGX180 19.8 EP4SGX230 19.8 EP4SGX290 21.8 EP4SGX360 21.8 EP4SGX530 26.8 EP4SE230 19.8 EP4SE360 21.8 EP4SE530 26.8 EP4SE820 33.8 EP4S40G2 19.8 EP4S40G5 26.8 EP4S100G2 19.8 EP4S100G3 26.8 EP4S100G4 26.8 EP4S100G5 26.8 Note to Table 11-6: (1) These timing numbers are preliminary. CRC calculation time for the error detection circuitry to check from the first until the last frame depends on the device and the error detection clock frequency. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 11-10 Chapter 11: SEU Mitigation in Stratix IV Devices Error Detection Timing Table 11-7 lists the estimated time for each CRC calculation with minimum and maximum clock frequencies for Stratix IV devices. The minimum CRC calculation time is calculated by using the maximum error detection frequency with a divisor factor of one, and the maximum CRC calculation time is calculated by using the minimum error detection frequency with a divisor factor of eight. Table 11-7. CRC Calculation Time (1) Device Minimum Time (ms) Maximum Time (s) EP4SGX70 111 30.90 EP4SGX110 111 30.90 EP4SGX180 225 62.44 EP4SGX230 225 62.44 EP4SGX290 296 82.05 EP4SGX360 296 82.05 EP4SGX530 398 110.38 EP4SE230 225 62.44 EP4SE360 296 82.05 EP4SE530 398 110.38 EP4SE820 577 160.00 EP4S40G2 225 62.44 EP4S40G5 398 110.38 EP4S100G2 225 62.44 EP4S100G3 398 110.38 EP4S100G4 398 110.38 EP4S100G5 398 110.38 Note to Table 11-7: (1) These timing numbers are preliminary. Software Support The Quartus II software version 8.0 and onwards supports the error detection CRC feature for Stratix IV devices. Enabling this feature generates the CRC_ERROR output to the optional dual purpose CRC_ERROR pin. The error detection CRC feature is controlled by the Device and Pin Options dialog box in the Quartus II software. To enable the error detection feature using CRC, follow these steps: 1. Open the Quartus II software and load a project using a Stratix IV device. 2. On the Assignments menu, click Settings. The Settings dialog box is shown. 3. In the Category list, select Device. The Device page is shown. 4. Click Device and Pin Options. The Device and Pin Options dialog box is shown (refer to Figure 11-2). Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 11: SEU Mitigation in Stratix IV Devices Recovering From CRC Errors 11-11 5. In the Device and Pin Options dialog box, click the Error Detection CRC tab. 6. Turn on Enable error detection CRC (Figure 11-2). Figure 11-2. Enabling the Error Detection CRC Feature in the Quartus II Software 7. In the Divide error check frequency by pull-down list, enter a valid divisor as listed in Table 11-5 on page 11-8. 1 The divide value divides the frequency of the configuration oscillator output clock that clocks the CRC circuitry. 8. Click OK. Recovering From CRC Errors The system that the Stratix IV device resides in must control device reconfiguration. After detecting an error on the CRC_ERROR pin, strobing the nCONFIG signal low directs the system to perform the reconfiguration at a time when it is safe for the system to reconfigure the device. When the data bit is rewritten with the correct value by reconfiguring the device, the device functions correctly. While soft errors are uncommon in Altera devices, certain high-reliability applications require a design to account for these errors. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 11-12 Chapter 11: SEU Mitigation in Stratix IV Devices Recovering From CRC Errors Document Revision History Table 11-8 lists the revision history for this chapter. Table 11-8. Document Revision History Date Version February 2011 March 2010 November 2009 3.2 3.1 3.0 June 2009 2.3 April 2009 2.2 March 2009 2.1 Changes Applied new template. Minor Text edits. Updated Table 11-3 and Table 11-6. Minor text edits. Updated Table 11-3, Table 11-5, Table 11-6, and Table 11-7. Updated the "CRC_ERROR Pin" section. Minor text edits. Added an introductory paragraph to increase search ability. Removed the Conclusion section. Minor text edits. Updated Table 11-6 and Table 11-7. Updated "Error Detection Timing" section. Updated Table 11-6. Added Table 11-7. Removed "Critical Error Detection", "Critical Error Pin", and "Referenced Documents" sections. November 2008 2.0 Minor text edits. May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation 12. JTAG Boundary-Scan Testing in Stratix IV Devices February 2011 SIV51012-3.2 SIV51012-3.2 The IEEE Std. 1149.1 boundary-scan test (BST) circuitry available in Stratix(R) IV devices provides a cost-effective and efficient way to test systems that contain devices with tight lead spacing. Circuit boards with Altera and other IEEE Std. 1149.1-compliant devices can use EXTEST, SAMPLE/PRELOAD, and BYPASS modes to create serial patterns that internally test the pin connections between devices and check device operation. This chapter describes how to use the IEEE Std. 1149.1 BST circuitry in Stratix IV devices. The features are similar to Stratix III devices, unless stated otherwise in this chapter. This chapter contains the following sections: "BST Architecture" "BST Operation Control" on page 12-2 "I/O Voltage Support in a JTAG Chain" on page 12-4 "BST Circuitry" on page 12-4 "BSDL Support" on page 12-4 BST Architecture A device operating in IEEE Std. 1149.1 BST mode uses four required pins, TDI, TDO, TMS, TCK, and one optional pin, TRST. The TCK pin has an internal weak pull-down resistor, while the TDI, TMS, and TRST pins have internal weak pull-up resistors. The TDO output pin and all the JTAG input pins are powered by the 2.5-V/3.0-V VCCPD supply of I/O bank 1A. All user I/O pins are tri-stated during JTAG configuration. f For more information about the description and functionality of all JTAG pins, registers used by the IEEE Std. 1149.1 BST circuitry, and the test access port (TAP) controller, refer to the IEEE 1149.1 (JTAG) Boundary-Scan Testing in Stratix III Devices chapter in volume 1 of the Stratix III Device Handbook. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 12-2 Chapter 12: JTAG Boundary-Scan Testing in Stratix IV Devices BST Operation Control BST Operation Control Table 12-1 lists the boundary-scan register length for Stratix IV devices. Table 12-1. Boundary-Scan Register Length in Stratix IV Devices Device Boundary-Scan Register Length EP4SGX70 1506 EP4SGX110 1506 EP4SGX180 2274 EP4SGX230 2274 EP4SGX290 (1) 2682 EP4SGX360 (1) 2682 EP4SGX530 2970 EP4SE230 2274 EP4SE360 2682 EP4SE530 2970 EP4SE820 3402 EP4S40G2 2274 EP4S40G5 2970 EP4S100G2 2274 EP4S100G3 2970 EP4S100G4 2970 EP4S100G5 2970 Note to Table 12-1: (1) For the F1932 package of EP4SGX290 and EP4SGX360 devices, the boundary-scan register length is 2970. Table 12-2 lists the IDCODE information for Stratix IV devices. Table 12-2. IDCODE Information for Stratix IV Devices (Part 1 of 2) IDCODE (32 Bits) Device (1) Version (4 Bits) Part Number (16 Bits) Manufacturer Identity (11 Bits) LSB (1 Bit) (2) EP4SGX70 0000 0010 0100 0010 0000 000 0110 1110 1 EP4SGX110 0000 0010 0100 0000 0000 000 0110 1110 1 EP4SGX180 0000 0010 0100 0010 0001 000 0110 1110 1 EP4SGX230 0000 0010 0100 0000 1001 000 0110 1110 1 EP4SGX290 (3) 0000 0010 0100 0010 0010 000 0110 1110 1 EP4SGX290 (4) 0000 0010 0100 0100 0011 000 0110 1110 1 EP4SGX360 (3) 0000 0010 0100 0000 0010 000 0110 1110 1 EP4SGX360 (4) 0000 0010 0100 1000 0011 000 0110 1110 1 EP4SGX530 0000 0010 0100 0000 0011 000 0110 1110 1 EP4SE230 0000 0010 0100 0001 0001 000 0110 1110 1 EP4SE360 0000 0010 0100 0001 0010 000 0110 1110 1 Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 12: JTAG Boundary-Scan Testing in Stratix IV Devices BST Operation Control 12-3 Table 12-2. IDCODE Information for Stratix IV Devices (Part 2 of 2) IDCODE (32 Bits) (1) Device Version (4 Bits) Part Number (16 Bits) Manufacturer Identity (11 Bits) LSB (1 Bit) (2) EP4SE530 0000 0010 0100 0001 0011 000 0110 1110 1 EP4SE820 0000 0010 0100 0000 0100 000 0110 1110 1 EP4S40G2 (5) 0000 0010 0100 0100 0001 000 0110 1110 1 EP4S40G5 (6) 0000 0010 0100 0010 0011 000 0110 1110 1 0000 0010 0100 0100 0001 000 0110 1110 1 0000 0010 0100 1010 0011 000 0110 1110 1 EP4S100G2 (5) EP4S100G3 EP4S100G4 EP4S100G5 (6) 0000 0010 0100 0110 0011 000 0110 1110 1 0000 0010 0100 0010 0011 000 0110 1110 1 Notes to Table 12-2: (1) The MSB is on the left. (2) The LSB of the IDCODE is always 1. (3) The IDCODE is applicable for all packages except F1932. (4) The IDCODE is applicable for package F1932 only. (5) For the ES1 device, the IDCODE is the same as the IDCODE of EP4SGX230. (6) For the ES1 device, the IDCODE is the same as the IDCODE of EP4SGX530. 1 If the device is in reset state, when the nCONFIG or nSTATUS signal is low, the device IDCODE might not be read correctly. To read the device IDCODE correctly, you must issue the IDCODE JTAG instruction only when the nSTATUS signal is high. f For more information about the following topics, refer to the IEEE 1149.1 (JTAG) Boundary-Scan Testing in Stratix III Devices chapter in volume 1 of the Stratix III Device Handbook: February 2011 JTAG instruction codes with descriptions TAP controller state-machine Timing requirements for IEEE Std. 1149.1 signals Instruction mode Mandatory JTAG instructions (SAMPLE/PRELOAD, EXTEST, and BYPASS) Optional JTAG instructions (IDCODE, USERCODE, CLAMP, and HIGHZ) Altera Corporation Stratix IV Device Handbook Volume 1 12-4 Chapter 12: JTAG Boundary-Scan Testing in Stratix IV Devices I/O Voltage Support in a JTAG Chain I/O Voltage Support in a JTAG Chain The JTAG chain supports several devices. However, you must use caution if the chain contains devices that have different VCCIO levels. f For more information, refer to the IEEE 1149.1 (JTAG) Boundary-Scan Testing in Stratix III Devices chapter in volume 1 of the Stratix III Device Handbook. BST Circuitry The IEEE Std. 1149.1 BST circuitry is enabled after device power-up. You can perform BST on Stratix IV devices before, during, and after configuration. Stratix IV devices support BYPASS, IDCODE, and SAMPLE JTAG instructions during configuration without interrupting configuration. To send all other JTAG instructions, you must interrupt configuration using the CONFIG_IO JTAG instruction. f For more information, refer to AN 39: IEEE Std. 1149.1 (JTAG) Boundary-Scan Testing in Altera Devices. f For more information about using the CONFIG_IO JTAG instruction for dynamic I/O buffer configuration, considerations when performing BST for configured devices, and JTAG pin connections to mask-out the BST circuitry, refer to the IEEE 1149.1 (JTAG) Boundary-Scan Testing in Stratix III Devices chapter in volume 1 of the Stratix III Device Handbook. f For more information about using the IEEE Std.1149.1 circuitry for device configuration, refer to the Configuration, Design Security, Remote System Upgrades in Stratix IV Devices chapter. f If you must perform BST for configured devices, you must use the Quartus II software version 8.1 and onwards to generate the design-specific boundary-scan description language (BSDL) files. For the procedure to generate post-configured BSDL files using the Quartus II software, refer to the BSDL Files Generation in Quartus II on the Altera website. BSDL Support BSDL, a subset of VHDL, provides a syntax that allows you to describe the features of an IEEE Std. 1149.1 BST-capable device that can be tested. f For more information about BSDL files for IEEE Std. 1149.1-compliant Stratix IV devices, refer to the Stratix IV BSDL Files on the Altera website. f BSDL files for IEEE std. 1149.1-compliant Stratix IV devices can also be generated using the Quartus II software version 8.1 and onwards. For more information about the procedure to generate BSDL files using the Quartus II software, refer to the BSDL Files Generation in Quartus II on the Altera website. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 12: JTAG Boundary-Scan Testing in Stratix IV Devices BSDL Support 12-5 Document Revision History Table 12-3 lists the revision history for this chapter. Table 12-3. Document Revision History Date Version February 2011 March 2010 3.2 3.1 November 2009 3.0 Changes Applied new template. Minor text edits. Updated the hand note in the "BST Operation Control" section. Changed "IDCODE JTAG Instruction" to read "IDCODE" as needed. Minor text edits Updated Table 12-1 and Table 12-2. Minor text edits. Added an introductory paragraph to increase search ability. Removed the Conclusion section. Minor text edits. Updated Table 12-1. Updated Table 12-1 and Table 12-2. Removed "Referenced Documents" section. June 2009 2.3 April 2009 2.2 March 2009 2.1 November 2008 2.0 Minor text edits. April 2010 1.0 Initial release. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 12-6 Stratix IV Device Handbook Volume 1 Chapter 12: JTAG Boundary-Scan Testing in Stratix IV Devices BSDL Support February 2011 Altera Corporation 13. Power Management in Stratix IV Devices February 2011 SIV51013-3.2 SIV51013-3.2 This chapter describes power management in Stratix(R) IV devices. Stratix IV devices offer programmable power technology options for low-power operation. You can use these options, along with speed grade choices, in different permutations to give the best power and performance combination. For thermal management, use the Stratix IV internal temperature sensing device (TSD) with built-in analog-to-digital converter (ADC) circuitry or external TSD with an external temperature sensor to easily incorporate this feature in your designs. Being able to monitor the junction temperature of the device at any time also offers the ability to control air flow to the device and save power for the whole system. Overview Stratix IV FPGAs deliver a breakthrough level of system bandwidth and power efficiency for high-end applications, allowing you to innovate without compromise. Stratix IV devices use advanced power management techniques to enable both density and performance increases while simultaneously reducing power dissipation. The total power of an FPGA includes static and dynamic power. Static power is the power consumed by the FPGA when it is configured but no clocks are operating. Dynamic power is the switching power when the device is configured and running. You configure dynamic power with the equation shown in Equation 13-1. Equation 13-1. Dynamic Power Equation (1) 2 1 P = --- CV frequency 2 Note to Equation 13-1: (1) P = power; C = load capacitance; and V = supply voltage level. Equation 13-1 shows that frequency is design dependant. However, you can vary the voltage to lower dynamic power consumption by the square value of the voltage difference. Stratix IV devices minimize static and dynamic power with advanced process optimizations and programmable power technology. These technologies enable Stratix IV designs to optimally meet design-specific performance requirements with the lowest possible power. The Quartus(R) II software optimizes all designs with Stratix IV power technology to ensure performance is met at the lowest power consumption. This automatic process allows you to concentrate on the functionality of the design instead of the power consumption of the design. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 1 February 2011 Feedback Subscribe 13-2 Chapter 13: Power Management in Stratix IV Devices Stratix IV Power Technology Power consumption also affects thermal management. Stratix IV devices offer a TSD feature that self-monitors the device junction temperature and can be used with external circuitry for other activities, such as controlling air flow to the Stratix IV FPGA. This chapter contains the following sections: "Stratix IV Power Technology" "Stratix IV External Power Supply Requirements" "Temperature Sensing Diode" Stratix IV Power Technology The following sections describe Stratix IV programmable power technology. Programmable Power Technology Stratix IV devices offer the ability to configure portions of the core, called tiles, for high-speed or low-power mode of operation performed by the Quartus II software without user intervention. Setting a tile to high-speed or low-power mode is accomplished with on-chip circuitry and does not require extra power supplies brought into the Stratix IV device. In a design compilation, the Quartus II software determines whether a tile must be in high-speed or low-power mode based on the timing constraints of the design. f For more information about how the Quartus II software uses programmable power technology when compiling a design, refer to AN 514: Power Optimization in Stratix IV FPGAs. A Stratix IV tile can consist of the following: Memory logic array block (MLAB)/logic array block (LAB) pairs with routing to the pair MLAB/LAB pairs with routing to the pair and to adjacent digital signal processing (DSP)/memory block routing TriMatrix memory blocks DSP blocks All blocks and routing associated with the tile share the same setting of either high-speed or low-power mode. By default, tiles that include DSP blocks or memory blocks are set to high-speed mode for optimum performance. Unused DSP blocks and memory blocks are set to low-power mode to minimize static power. Clock networks do not support programmable power technology. With programmable power technology, faster speed grade FPGAs may require less power because there are fewer high-speed MLAB and LAB pairs, when compared with slower speed grade FPGAs. The slower speed grade device may have to use more high-speed MLAB and LAB pairs to meet performance requirements, while the faster speed grade device can meet performance requirements with MLAB and LAB pairs in low-power mode. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 13: Power Management in Stratix IV Devices Stratix IV External Power Supply Requirements 13-3 The Quartus II software sets unused device resources in the design to low-power mode to reduce static and dynamic power. It also sets the following resources to low-power mode when they are not used in the design: LABs and MLABs TriMatrix memory blocks DSP blocks If a phase-locked loop (PLL) is instantiated in the design, asserting the areset pin high keeps the PLL in low-power mode. Table 13-1 lists the available Stratix IV programmable power capabilities. Speed grade considerations can add to the permutations to give you flexibility in designing your system. Table 13-1. Programmable Power Capabilities in Stratix IV Devices Feature Programmable Power Technology LAB Yes Routing Yes Memory Blocks Fixed setting (1) DSP Blocks Fixed setting (1) Global Clock Networks No Note to Table 13-1: (1) Tiles with DSP blocks and memory blocks that are used in the design are always set to high-speed mode. By default, unused DSP blocks and memory blocks are set to low-power mode. Stratix IV External Power Supply Requirements This section describes the different external power supplies required to power Stratix IV devices. You can supply some of the power supply pins with the same external power supply, provided they have the same voltage level. f For power supply pin connection guidelines and power regulator sharing, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. f For each Altera recommended power supply's operating conditions, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 13-4 Chapter 13: Power Management in Stratix IV Devices Temperature Sensing Diode Temperature Sensing Diode The Stratix IV TSD uses the characteristics of a PN junction diode to determine die temperature. Knowing the junction temperature is crucial for thermal management. Historically, junction temperature is calculated using ambient or case temperature, junction-to-ambient (ja) or junction to-case (jc) thermal resistance, and device power consumption. Stratix IV devices can either monitor its die temperature with the internal TSD with built-in ADC circuitry or the external TSD with an external temperature sensor. This allows you to control the air flow to the device. You can use the Stratix IV internal TSD in two different modes of operation-- power-up mode and user mode. For power-up mode, the internal TSD reads the die's temperature during configuration if the ALTTEMP_SENSE megafunction is enabled in your design. The ALTTEMP_SENSE megafunction allows temperature sensing during device user mode by asserting the clken signal to the internal TSD circuitry. To reduce device static power, disable the internal TSD with built-in ADC circuitry when not in use. f For more information about using the ALTTEMP_SENSE megafunction, refer to the Thermal Sensor (ALTTEMP_SENSE) Megafunction User Guide. The external temperature sensor steers bias current through the Stratix IV external TSD, which measures forward voltage and converts this reading to temperature in the form of an 8-bit signed number (7 bits plus sign). The 8-bit output represents the junction temperature of the Stratix IV device and can be used for intelligent power management. External Pin Connections The Stratix IV external TSD requires two pins for voltage reference. Figure 13-1 shows how to connect the external TSD with an external temperature sensor device. As an example, external temperature sensing devices, such as MAX1619, MAX1617A, MAX6627, and ADT 7411, can be connected to the two external TSD pins for temperature reading. Figure 13-1. TSD External Pin Connections in Stratix IV Devices External TSD TEMPDIODEP External Temperature Sensor Stratix IV Device TEMPDIODEN f For more information about the external TSD specification, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Chapter 13: Power Management in Stratix IV Devices Temperature Sensing Diode 13-5 The TSD is a very sensitive circuit that can be influenced by noise coupled from other traces on the board and possibly within the device package itself, depending on your device usage. The interfacing device registers' temperature is based on millivolts (mV) of difference, as seen at the external TSD pins. Switching the I/O near the TSD pins can affect the temperature reading. Altera recommends taking temperature readings during periods of inactivity in the device or use the internal TSD with built-in ADC circuitry. The following are board connection guidelines for the TSD external pin connections: The maximum trace lengths for the TEMPDIODEP/TEMPDIODEN traces must be less than eight inches. Route both traces in parallel and place them close to each other with grounded guard tracks on each side. Altera recommends 10-mils width and space for both traces. Route traces through a minimum number of vias and crossunders to minimize the thermocouple effects. Ensure that the number of vias are the same on both traces. Ensure both traces are approximately the same length. Avoid coupling with toggling signals (for example, clocks and I/O) by having the GND plane between the diode traces and the high frequency signals. For high-frequency noise filtering, place an external capacitor (close to the external chip) between the TEMPDIODEP/TEMPDIODEN trace. For Maxim devices, use an external capacitor between 2200 pF to 3300 pF. Place a 0.1 uF bypass capacitor close to the external device. You can use internal TSD with built-in ADC circuitry and external TSD at the same time. If you only use internal ADC circuitry, the external TSD pins (TEMPDIODEP/TEMPDIODEN) can connect these pins to GND because the external TSD pins are not used. f For more information about the TEMPDIODEP/TEMPDIODEN pin connection when you are not using an external TSD, refer to the Stratix IV GX and Stratix IV E Pin Connection Guidelines. f For device specification and connection guidelines, refer to the external temperature sensor device data sheet from the device manufacturer. February 2011 Altera Corporation Stratix IV Device Handbook Volume 1 13-6 Chapter 13: Power Management in Stratix IV Devices Temperature Sensing Diode Document Revision History Table 13-2 lists the revision history for this chapter. Table 13-2. Document Revision History Date Version February 2011 March 2010 November 2009 June 2009 March 2009 3.2 3.1 3.0 2.2 2.1 Changes Applied new template. Minor text edits. Updated the "External Pin Connections" section. Minor text edits. Updated the "Temperature Sensing Diode" and "External Pin Connections" sections. Updated Equation 13-1. Removed Table 13-2: Stratix IV External Power Supply Pins. Minor text edits. Updated the "External Pin Connections" section. Added an introductory paragraph to increase search ability. Removed the Conclusion section. Updated "Temperature Sensing Diode" and "External Pin Connections" sections. Updated Figure 13-1. Removed "Referenced Documents" section. November 2008 2.0 Minor text edits. May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 1 February 2011 Altera Corporation Additional Information About this Handbook This chapter provides additional information about the document and Altera. How to Contact Altera To locate the most up-to-date information about Altera products, refer to the following table. Contact (1) Technical support Technical training Product literature Contact Method Address Website www.altera.com/support Website www.altera.com/training Email custrain@altera.com Website www.altera.com/literature Nontechnical support (general) Email nacomp@altera.com (software licensing) Email authorization@altera.com Note to Table: (1) You can also contact your local Altera sales office or sales representative. Typographic Conventions The following table shows the typographic conventions this document uses. Visual Cue Meaning Bold Type with Initial Capital Letters Indicate command names, dialog box titles, dialog box options, and other GUI labels. For example, Save As dialog box. For GUI elements, capitalization matches the GUI. bold type Indicates directory names, project names, disk drive names, file names, file name extensions, software utility names, and GUI labels. For example, \qdesigns directory, D: drive, and chiptrip.gdf file. Italic Type with Initial Capital Letters Indicate document titles. For example, Stratix IV Design Guidelines. Indicates variables. For example, n + 1. italic type Variable names are enclosed in angle brackets (< >). For example, <file name> and <project name>.pof file. Initial Capital Letters Indicate keyboard keys and menu names. For example, the Delete key and the Options menu. "Subheading Title" Quotation marks indicate references to sections in a document and titles of Quartus II Help topics. For example, "Typographic Conventions." September 2012 Altera Corporation Stratix IV Device Handbook Volume 1 Info-2 Additional Information Typographic Conventions Visual Cue Meaning Indicates signal, port, register, bit, block, and primitive names. For example, data1, tdi, and input. The suffix n denotes an active-low signal. For example, resetn. Courier type Indicates command line commands and anything that must be typed exactly as it appears. For example, c:\qdesigns\tutorial\chiptrip.gdf. Also indicates sections of an actual file, such as a Report File, references to parts of files (for example, the AHDL keyword SUBDESIGN), and logic function names (for example, TRI). r An angled arrow instructs you to press the Enter key. 1., 2., 3., and a., b., c., and so on Numbered steps indicate a list of items when the sequence of the items is important, such as the steps listed in a procedure. Bullets indicate a list of items when the sequence of the items is not important. 1 The hand points to information that requires special attention. h The question mark directs you to a software help system with related information. f The feet direct you to another document or website with related information. m The multimedia icon directs you to a related multimedia presentation. c A caution calls attention to a condition or possible situation that can damage or destroy the product or your work. w A warning calls attention to a condition or possible situation that can cause you injury. The envelope links to the Email Subscription Management Center page of the Altera website, where you can sign up to receive update notifications for Altera documents. The feedback icon allows you to submit feedback to Altera about the document. Methods for collecting feedback vary as appropriate for each document. The social media icons allow you to inform others about Altera documents. Methods for submitting information vary as appropriate for each medium. Stratix IV Device Handbook Volume 1 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 101 Innovation Drive San Jose, CA 95134 www.altera.com SIV5V2-4.4 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as ISO trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and 9001:2008 services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service Registered described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers Contents Chapter Revision Dates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Section I. Transceiver Architecture Chapter 1. Transceiver Architecture in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 Transceiver Channel Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 Stratix IV GX Device Offerings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 Stratix IV GT Device Offerings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 Transceiver Block Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16 Transceiver Channel Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-17 Transmitter Channel Datapath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-19 Receiver Channel Datapath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-40 CMU Channel Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-100 Configuring CMU Channels for Clock Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-101 Configuring CMU Channels as Transceiver Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-106 Other CMU Channel Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-109 Dynamic Reconfiguration of the CMU Channel Analog Controls . . . . . . . . . . . . . . . . . . . . . . . . 1-110 Functional Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-110 Basic Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-111 Deterministic Latency Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-122 PCIe Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-127 XAUI Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-153 GIGE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-164 SONET/SDH Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-172 SDI Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-178 (OIF) CEI PHY Interface Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-181 Serial RapidIO Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-182 Basic (PMA Direct) Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-187 Loopback Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-190 Serial Loopback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-190 Parallel Loopback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-191 Reverse Serial Loopback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-193 Reverse Serial Pre-CDR Loopback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-194 PCIe Reverse Parallel Loopback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-194 Auxiliary Transmit (ATX) PLL Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-195 6G ATX PLL Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-195 10G ATX PLL Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-196 Input Reference Clocks for the ATX PLL Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-198 Architecture of the ATX PLL Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-199 ATX Clock Divider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-200 The Differences Between 10G ATX PLL, 6G ATX PLL, and CMU PLL . . . . . . . . . . . . . . . . . . . . . . 1-200 Calibration Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-201 Calibration Block Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-201 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-205 Input Signals to the Calibration Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-205 Built-In Self Test Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-207 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers iv Contents BIST Mode Pattern Generators and Verifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-207 PRBS in Single-Width Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-208 PRBS in Double-Width Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-209 Transceiver Port Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-210 Reference Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-225 Chapter 2. Transceiver Clocking in Stratix IV Devices Glossary of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 Input Reference Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 Input Reference Clock Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 refclk0 and refclk1 Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 Inter-Transceiver Block (ITB) Clock Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 Dedicated CLK Input Pins on the FPGA Global Clock Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 Clock Output from Left and Right PLLs in the FPGA Fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 FPGA Fabric PLLs-Transceiver PLLs Cascading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 Example 1: Channel Configuration with a 4 Gbps Data Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 Dedicated Left and Right PLL Cascade Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 780-Pin Package . . . . . . . . . . . . . . . . . . . . . 2-11 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1152-Pin Package . . . . . . . . . . . . . . . . . . . . 2-11 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1517-Pin Package . . . . . . . . . . . . . . . . . . . . 2-12 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1932-Pin Package . . . . . . . . . . . . . . . . . . . . 2-13 FPGA Fabric PLLs-Transceiver PLLs Cascading Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14 Example 2: Design Target--EP4SGX530NF45 Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14 Left and Right, Left, or Right PLL in VCO Bypass Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17 Transceiver Channel Datapath Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-20 Transmitter Channel Datapath Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-20 Transmitter Channel-to-Channel Skew Optimization for Modes Other than Basic (PMA Direct) Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-21 Transmitter Channel Datapath Clocking Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-22 Transmitter Channel Clocking Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-23 Non-Bonded Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-24 Bonded Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27 Non-Bonded Basic (PMA Direct) Mode Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . 2-34 Bonded Basic (PMA Direct) xN Mode Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36 Receiver Channel Datapath Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-39 Non-Bonded Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-39 Bonded Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-43 Basic (PMA Direct) Mode Channel Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-49 FPGA Fabric-Transceiver Interface Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-51 FPGA Fabric-Transmitter Interface Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-52 Quartus II-Selected Transmitter Phase Compensation FIFO Write Clock . . . . . . . . . . . . . . . . . . 2-52 User-Selected Transmitter Phase Compensation FIFO Write Clock . . . . . . . . . . . . . . . . . . . . . . . 2-58 FPGA Fabric-Receiver Interface Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-61 Quartus II Software-Selected Receiver Phase Compensation FIFO Read Clock . . . . . . . . . . . . . 2-62 User-Selected Receiver Phase Compensation FIFO Read Clock . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-68 Using the CMU/ATX PLL for Clocking User Logic in the FPGA Fabric . . . . . . . . . . . . . . . . . . . . . . . . 2-71 Configuration Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-72 Configuration Example 1: Configuring 24 Channels in Basic (PMA Direct) xN Mode in the EP4S100G5F45 Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-73 Configuration Example 2: Configuring Sixteen Identical Channels Across Four Transceiver Blocks . 2-75 Configuration Example 3: Configuring Sixteen Channels Across Four Transceiver Blocks . . . . . . 2-76 Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode 2-78 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Contents v Chapter 3. Configuring Multiple Protocols and Data Rates in Stratix IV Devices Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 Glossary of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2 Creating Transceiver Channel Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 General Requirements to Combine Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 Transmitter Buffer Voltage (VCCH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 Transceiver Analog Power (VCCA_L/R) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 Control Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4 gxb_powerdown Port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4 reconfig_fromgxb and reconfig_togxb Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4 Calibration Clock and Power Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5 Sharing CMU PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5 Multiple Channels Sharing a CMU PLL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5 Example 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6 Example 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9 Sharing ATX PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10 Combining Receiver Only Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10 Combining Transmitter Channel and Receiver Channel Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11 Multiple Transmitter Channel and Receiver Channel Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11 Example 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11 Combining Transceiver Instances in Multiple Transceiver Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13 Example 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14 Combining Transceiver Instances Using PLL Cascade Clocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16 Combining Channels Configured in Protocol Functional Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 Combining Channels in Bonded Functional Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 Bonded x4 Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 Bonded x8 Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20 Combining Channels Configured in Deterministic Latency Mode . . . . . . . . . . . . . . . . . . . . . . . . . . 3-24 Combining Channels Using the PCIe hard IP Block with Other Channels . . . . . . . . . . . . . . . . . . . . 3-24 Combining Transceiver Channels in Basic (PMA Direct) Configurations . . . . . . . . . . . . . . . . . . . . . . . 3-25 Combining Multiple Channels Configured in Basic (PMA Direct) x1 Configurations . . . . . . . . . . 3-26 Multiple Basic (PMA Direct) x1 Configuration Instances with One Channel per Instance . . . . 3-26 One Instance in a Basic (PMA Direct) x1 Configuration with Multiple Transceiver Channels . 3-26 Combining Multiple Instances of Transmitter Only and Receiver Only Configurations in Basic (PMA Direct) x1 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-29 Combining Channels Configured in Basic (PMA Direct) x1 with Non-Basic (PMA Direct) Modes 3-29 Basic (PMA Direct) xN Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-33 Channel Placement in a Basic (PMA Direct) xN Mode Instance . . . . . . . . . . . . . . . . . . . . . . . . . . 3-33 Examples of Combining Multiple Instances of Basic (PMA Direct) xN Modes . . . . . . . . . . . . . . 3-35 Combination Requirements When You Enable Channel Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . 3-42 Combination Requirements When You Enable the Use Alternate CMU PLL Option . . . . . . . . . . . 3-42 Example 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-43 Combination Requirements When You Use Multiple TX PLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-44 Example 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-45 Combining Transceiver Channels When You Enable the Adaptive Equalization (AEQ) Feature . . . 3-47 Example 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-48 Combination Requirements for Stratix IV Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-49 Placement Rules for Transceiver Channels at 9.9 Gbps to 11.3 Gbps . . . . . . . . . . . . . . . . . . . . . . . . . 3-49 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-49 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers vi Contents Chapter 4. Reset Control and Power Down in Stratix IV Devices User Reset and Power-Down Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2 Blocks Affected by the Reset and Power-Down Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3 Transceiver Reset Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4 All Supported Functional Modes Except PCIe Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6 Bonded Channel Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6 Non-Bonded Channel Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15 PCIe Functional Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22 PCIe Reset Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22 PCIe Initialization/Compliance Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-23 PCIe Normal Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-23 PMA Direct Drive Mode Reset Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24 Basic (PMA Direct) Drive xN Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-25 Transmitter Only Channel with No PLL_L/R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-25 Transmitter Only Channel with a PLL_L/R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26 Basic (PMA Direct) Drive x1 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31 Receiver and Transmitter Channel Set-Up--Receiver CDR in Automatic Lock Mode . . . . . . . 4-32 Receiver and Transmitter Channel Set-up--Receiver CDR in Manual Lock Mode . . . . . . . . . . 4-34 Dynamic Reconfiguration Reset Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36 Reset Sequence when Using Dynamic Reconfiguration with the `data rate division in TX' Option . . . 4-36 Reset Sequence when Using Dynamic Reconfiguration with the `Channel and TX PLL select/reconfig' Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37 Power Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-38 Simulation Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39 Reference Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39 Chapter 5. Dynamic Reconfiguration in Stratix IV Devices Glossary of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 Dynamic Reconfiguration Controller Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3 Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration . . . . . . . . 5-4 ALTGX MegaWizard Plug-In Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4 The reconfig_clk Clock Requirements for the ALTGX Instance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4 ALTGX_RECONFIG MegaWizard Plug-In Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5 The reconfig_clk Clock Requirements for the ALTGX_RECONFIG Instance . . . . . . . . . . . . . . . . 5-5 Interfacing ALTGX and ALTGX_RECONFIG Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5 Logical Channel Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5 Total Number of Channels Option in the ALTGX_RECONFIG Instance . . . . . . . . . . . . . . . . . . . 5-10 Connecting the ALTGX and ALTGX_RECONFIG Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11 Dynamic Reconfiguration Modes Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12 PMA Controls Reconfiguration Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12 Dynamically Reconfiguring PMA Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13 Transceiver Channel Reconfiguration Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19 Memory Initialization File (.mif) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-20 Channel and CMU PLL Reconfiguration Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-24 Channel Reconfiguration with Transmitter PLL Select Mode Details . . . . . . . . . . . . . . . . . . . . . . 5-48 CMU PLL Reconfiguration Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-54 Central Control Unit Reconfiguration Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-57 Special Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-57 Data Rate Division in Transmitter Mode Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-63 Offset Cancellation Feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-67 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-67 ALTGX_RECONFIG Instance Signals Transition during Offset Cancellation . . . . . . . . . . . . . . . 5-68 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Contents vii EyeQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-69 Enabling the EyeQ Control Logic and the EyeQ Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-70 Connections Between the ALTGX and ALTGX_RECONFIG Instances . . . . . . . . . . . . . . . . . . . . 5-70 Controlling the EyeQ Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-71 Adaptive Equalization (AEQ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-75 Adaptive Equalization Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-75 Enabling the AEQ Control Logic and AEQ Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-75 Connections Between the ALTGX and ALTGX_RECONFIG Instances . . . . . . . . . . . . . . . . . . . . 5-76 One Time Mode for a Single Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-78 Dynamic Reconfiguration Controller Port List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-78 Error Indication During Dynamic Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-90 Dynamic Reconfiguration Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-91 PMA Controls Reconfiguration Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-91 Offset Cancellation Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-93 Dynamic Reconfiguration Duration for Channel and Transmitter PLL Select/Reconfig Modes . . . 5-94 Dynamic Reconfiguration (ALTGX_RECONFIG Instance) Resource Utilization . . . . . . . . . . . . . . . . . 5-94 Functional Simulation of the Dynamic Reconfiguration Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-95 Dynamic Reconfiguration Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-96 Example 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-96 Example 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-100 Additional Information How to Contact Altera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 Typographic Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers viii Stratix IV Device Handbook Volume 2: Transceivers Contents September 2012 Altera Corporation Chapter Revision Dates The chapters in this document, Stratix IV Device Handbook Volume 2, were revised on the following dates. Where chapters or groups of chapters are available separately, part numbers are listed. Chapter 1. Transceiver Architecture in Stratix IV Devices Revised: September 2012 Part Number: SIV52001-4.5 Chapter 2. Transceiver Clocking in Stratix IV Devices Revised: September 2012 Part Number: SIV52002-3.4 Chapter 3. Configuring Multiple Protocols and Data Rates in Stratix IV Devices Revised: September 2012 Part Number: SIV52003-4.2 Chapter 4. Reset Control and Power Down in Stratix IV Devices Revised: September 2012 Part Number: SIV52004-4.3 Chapter 5. Dynamic Reconfiguration in Stratix IV Devices Revised: September 2012 Part Number: SIV52005-3.4 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers x Stratix IV Device Handbook Volume 2: Transceivers Chapter Revision Dates September 2012 Altera Corporation Section I. Transceiver Architecture This section provides a description of transceiver architecture and transceiver clocking for the Stratix(R) IV device family. It also describes configuring for multiple protocols and data rates, reset control and power down, and dynamic reconfiguration for Stratix IV devices. This section includes the following chapters: Chapter 1, Transceiver Architecture in Stratix IV Devices Chapter 2, Transceiver Clocking in Stratix IV Devices Chapter 3, Configuring Multiple Protocols and Data Rates in Stratix IV Devices Chapter 4, Reset Control and Power Down in Stratix IV Devices Chapter 5, Dynamic Reconfiguration in Stratix IV Devices Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers I-2 Section I: Transceiver Architecture Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation 1. Transceiver Architecture in Stratix IV Devices September 2012 SIV52001-4.5 SIV52001-4.5 This chapter provides details about Stratix(R) IV GX and GT transceiver architecture, transceiver channels, available modes, and a description of transmitter and receiver channel datapaths. f For information about upcoming Stratix IV device features, refer to the Upcoming Stratix IV Device Features document. f For information about changes to the currently published Stratix IV Device Handbook, refer to the Addendum to the Stratix IV Device Handbook chapter. Overview Stratix IV FPGAs deliver a breakthrough level of system bandwidth and power efficiency for high-end applications, allowing you to innovate without compromise with two offerings: Stratix IV GX and Stratix IV GT. Stratix IV GX devices are part of the Altera(R) 40 nm Stratix IV device family. Stratix IV GX devices provide up to 32 full-duplex CDR-based transceivers with physical coding sublayer (PCS) and physical medium attachment (PMA), at serial data rates between 600 Mbps and 8.5 Gbps. Also, Stratix IV GX provides up to 16 additional full-duplex CDR-based transceivers with PMA supporting serial data rates between 600 Mbps and 6.5 Gbps. Table 1-1 lists the Stratix IV GX serial protocols the transceiver channels support. Table 1-1. Serial Protocols Supported by the Stratix IV GX Transceiver Channels Protocol Description PCI Express(R) (PCIe) Gen 1 at 2.5 Gbps and Gen 2 at 5.0 Gbps XAUI 3.125 Gbps to 3.75 Gbps for HiGig support GIGE Serial 1.25 Gbps RapidIO(R) 1.25 Gbps, 2.5 Gbps, and 3.125 Gbps SONET/SDH OC-12 at 622 Mbps, OC-48 at 2.488 Gbps, and OC-96 at 4.976 Gbps (OIF) CEI PHY Interface 4.976 Gbps to 6.375 Gbps for Interlaken support Serial Digital Interface (SDI) HD-SDI at 1.485 Gbps and 1.4835 Gbps 3G-SDI at 2.97 Gbps and 2.967 Gbps To implement proprietary protocols, the transceiver channels in the Stratix IV GX device supports the highly flexible Basic single-width (600 Mbps to 3.75 Gbps) and Basic double-width (1 Gbps to 8.5 Gbps) functional modes. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 2: Transceivers September 2012 Feedback Subscribe 1-2 Chapter 1: Transceiver Architecture in Stratix IV Devices Overview Stratix IV GT devices are also part of Altera's 40 nm Stratix IV device family and contain serial transceivers that support data rates between 600 Mbps and 11.3 Gbps. Stratix IV GT devices are targeted towards implementing 40 Gbps/100 Gbps transceiver links. Example applications include 40G/100G Ethernet and SFI-S. Stratix IV GT devices can be broadly classified into the following: Stratix IV GT devices targeted to achieve 100 Gbps ingress/egress data rates--48 full duplex clock and clock data recovery (CDR)-based transceivers, 32 of which support data rates up to 11.3 Gbps Stratix IV GT devices targeted to achieve 40 Gbps ingress/egress data rates--36 full duplex CDR-based transceivers, 12 of which support data rates up to 11.3 Gbps Though optimized for 40 Gbps/100 Gbps systems, Stratix IV GT transceivers also provide PMA and PCS support for the protocols shown in Table 1-2. Table 1-2. Serial Protocols Supported by the Stratix IV GT Transceiver Channels Protocol Description PCIe Gen 1 at 2.5 Gbps and Gen 2 at 5.0 Gbps XAUI 3.125 Gbps up to HiGig at 3.75 Gbps GIGE 1.25 Gbps Serial RapidIO 2.5 Gbps and 3.125 Gbps SONET/SDH OC-48 and OC-96 (OIF) CEI PHY Interface 4.976 Gbps to 6.375 Gbps Serial Digital Interface (SDI) 3G-SDI at 2.97Gbps and 2.967 Gbps To implement proprietary protocols, the transceiver channels in the Stratix IV GT device support the highly flexible Basic single-width (600 Mbps to 3.75 Gbps) and Basic double-width (600 Mbps to 11.3 Gbps) functional modes. f Stratix IV GX and GT devices have PCIe hard IP, PCS, and PMA blocks. For more information, refer to the PCI Express Compiler User Guide. f For more information about Stratix IV GX and GT protocols, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Overview 1-3 Figure 1-1 shows an example of the Stratix IV GX and GT transceiver architecture. Links to the corresponding transceiver architecture descriptions are listed below. This is an elementary diagram and does not represent an actual transceiver block. Figure 1-1. Example of a Transceiver Block Transceiver Block 1 Transmitter Channel Datapath 3 Transceiver Channel 3 BIST 10 2 Loopback 7 Receiver Channel Datapath 5 Transceiver Channel 2 2 Calibration Block 9 Calibration Block 9 Loopback 7 Transceiver Block GXBR1 Transceiver Block GXBL1 Channel 3 Channel 2 Channel 3 Channel 2 Channel 1 Channel 0 Channel 1 Channel 0 Local Clock Divider Block 4 (each Transceiver Channel) ATX PLL Block 8 ATX PLL Block 8 Transceiver Block GXBL0 Transceiver Block GXBR0 Channel 3 Channel 2 Channel 3 Channel 2 Channel 1 Channel 0 Channel 1 Channel 0 Calibration Block BIST 10 9 Calibration Block Central Control Unit (CCU) CMU1 Channel 6 1 CMU0 Channel 6 Transceiver Channel 1 2 BIST 10 Loopback 7 9 Transceiver Channel 0 2 BIST 10 Loopback 7 Descriptions for the example transceiver architecture are as follows: 1. "Transceiver Block Architecture" on page 1-16 2. "Transceiver Channel Architecture" on page 1-17 3. "Transmitter Channel Datapath" on page 1-19 4. "Transmitter Local Clock Divider Block" on page 1-39 5. "Receiver Channel Datapath" on page 1-40 6. "CMU Channel Architecture" on page 1-100 7. "Loopback Modes" on page 1-190 8. "Auxiliary Transmit (ATX) PLL Block" on page 1-195 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-4 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 9. "Calibration Blocks" on page 1-201 10. "Built-In Self Test Modes" on page 1-207 Transceiver Channel Locations Stratix IV GX and GT transceivers are structured into full-duplex (Transmitter and Receiver) four-channel groups called transceiver blocks. The total number of transceiver channels and the location of transceiver blocks varies from device to device. Stratix IV GX Device Offerings Table 1-3 lists the total number of transceiver channels and transceiver block locations in each Stratix IV GX device member structured into full-duplex four- and six-channel groups called transceiver blocks. Table 1-3. Number of Transceiver Channels and Transceiver Block Locations in Stratix IV GX Devices (Part 1 of 2) Device Member Total Number of Transceiver Channels EP4SGX70DF29 EP4SGX110DF29 EP4SGX180DF29 Transceiver Channel Location Eight transceiver channels located in two transceiver blocks: 8 Right side--GXBR0 and GXBR1 Refer to Figure 1-2 on page 1-5. EP4SGX230DF29 EP4SGX290FH29 EP4SGX360FH29 Eight transceiver channels located in two transceiver blocks: EP4SGX110FF35 EP4SGX180FF35 16 EP4SGX230FF35 Right side--GXBR0 and GXBR1 Left side--GXBL0 and GXBL1 Refer to Figure 1-2 on page 1-5. EP4SGX290FF35 EP4SGX360FF35 Eight regular transceiver channels supporting data rates between 600 Mbps and 8.5 Gbps and four clock multiplier unit (CMU) channels supporting data rates between 600 Mbps and 6.5 Gbps located in two transceiver blocks: EP4SGX180HF35 EP4SGX230HF35 EP4SGX290HF35 EP4SGX360HF35 24 Right side--GXBR0 and GXBR1 Left side--GXBL0 and GXBL1 Refer to Figure 1-3 on page 1-6. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-5 Table 1-3. Number of Transceiver Channels and Transceiver Block Locations in Stratix IV GX Devices (Part 2 of 2) Device Member Total Number of Transceiver Channels Transceiver Channel Location EP4SGX180KF40 Twelve regular transceiver channels supporting data rates between 600 Mbps and 8.5 Gbps and six CMU channels supporting data rates between 600 Mbps and 6.5 Gbps located in three transceiver blocks: EP4SGX230KF40 EP4SGX290KF40 36 EP4SGX290KF43 EP4SGX360KF40 EP4SGX360KF43 Right side--GXBR0, GXBR1, and GXBR2 Left side--GXBL0, GXBL1, and GXBL2 Refer to Figure 1-3 on page 1-6. EP4SGX530KF40 Sixteen regular transceiver channels supporting data rates between 600 Mbps and 8.5 Gbps and eight CMU channels supporting data rates between 600 Mbps and 6.5 Gbps located in four transceiver blocks: EP4SGX290NF45 EP4SGX360NF45 48 EP4SGX530NF45 Right side--GXBR0, GXBR1, GXBR2, and GXBR3 Left side--GXBL0, GXBL1, GXBL2, and GXBL3 Refer to Figure 1-4 on page 1-7. Figure 1-2 shows transceiver channel, PLL, and PCIe hard IP block locations in each Stratix IV GX device that has 8 or 16 transceiver channels. Figure 1-2. Transceiver Channel, PLL, and PCIe Hard IP Block Locations with 8 and 16 Stratix IV GX Transceiver Channels EP4SGX290FH29, EP4SGX360FH29, EP4SGX110FF35, EP4SGX180FF35, EP4SGX230FF35, EP4SGX290FF35, EP4SGX360FF35 (16 transceivers) EP4SGX70DF29 EP4SGX110DF29 EP4SGX180DF29 EP4SGX230DF29 (8 transceivers) Transceiver Block GXBL1 Channel 2 Channel 1 Channel 1 Channel 0 Channel 0 Transceiver Block GXBL0 Altera Corporation PCIe hard IP Channel 3 Channel 2 PCIe hard IP Channel 3 ATX PLL R0 (6G) September 2012 Transceiver Block GXBR1 ATX PLL R0 (6G) Transceiver Block GXBR0 Channel 3 Channel 3 Channel 2 Channel 2 Channel 1 Channel 1 Channel 0 Channel 0 Stratix IV Device Handbook Volume 2: Transceivers 1-6 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations Figure 1-3 shows transceiver channel, PLL, and PCIe hard IP block locations in each Stratix IV GX device that has 24 or 36 transceiver channels (except for the EP4SGX530 device). Figure 1-3. Transceiver Channel, PLL, and PCIe Hard IP Block Locations with 24 or 36 Stratix IV GX Transceiver Channels (Except the EP4SGX530 Device) EP4SGX180KF40, EP4SGX230KF40 (1), EP4SGX290KF40, EP4SGX290KF43, EP4SGX360KF43, EP4SGX360KF40, (36 transceivers) Transceiver Block GXBL2 Transceiver Block GXBR2 Channel 3 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 Channel 1 Channel 0 Channel 0 ATX PLL L1 (6G) (1) ATX PLL R1 (6G) (1) EP4SGX180HF35, EP4SGX230HF35, EP4SGX290HF35, EP4SGX360HF35, (24 transceivers) Transceiver Block GXBL1 Transceiver Block GXBR1 Channel 3 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 0 ATX PLL L0 (6G) Transceiver Block GXBL0 Channel 3 PCIe hard IP Channel 1 Channel 0 PCIe hard IP Channel 1 ATX PLL R01 (6G) Transceiver Block GXBR0 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 Channel 1 Channel 0 Channel 0 Note to Figure 1-3: (1) The 6G ATX PLL R1 and L1 blocks are not available in the EP4SGX230KF40 devices. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-7 Figure 1-4 shows transceiver channel, PLL, and PCIe hard IP block locations in each EP4SGX530 Stratix IV GX device that has 24 and 36 transceiver channels and the remaining Stratix IV GX devices that have 48 transceiver channels. Figure 1-4. Transceiver Channel, PLL, and PCIe Hard IP Block Locations with 48 Stratix IV GX Transceiver Channels and EP4SGX530 with 24, 36, and 48 Stratix IV GX Transceiver Channels EP4SGX290NF45, EP4SGX360NF45, EP4SGX530NF45 (48 transceivers) Transceiver Block GXBL3 Transceiver Block GXBR3 Channel 3 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 Channel 1 Channel 0 PCIe hard IP Transceiver Block GXBL2 PCIe hard IP EP4SGX530KF43, EP4SGX530KF40 (1), (2) (36 transceivers) Channel 0 Channel 3 Transceiver Block GXBR2 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 Channel 1 Channel 0 Channel 0 ATX PLL L1 (6G) (1) ATX PLL R1 (6G) (1) EP4SGX530HH35 (24 transceivers) Transceiver Block GXBL1 Transceiver Block GXBR1 Channel 3 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 0 ATX PLL L0 (6G) Transceiver Block GXBL0 Channel 3 PCIe hard IP Channel 1 Channel 0 PCIe hard IP Channel 1 ATX PLL R01 (6G) Transceiver Block GXBR0 Channel 3 Channel 2 Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 Channel 1 Channel 0 Channel 0 Notes to Figure 1-4: (1) The 6G ATX PLL R1 and L1 blocks are not available in the EP4SGX530KF40 devices. (2) Only the EP4SGX530 device in the six transceiver block package has four PCI hard IP blocks. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-8 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations Stratix IV GT Device Offerings Table 1-4 lists the Stratix IV GT device offerings along with the number of transceiver channels available in each device. Table 1-4. Stratix IV GT Device Offerings and Transceiver Channels Available in Each Device (4) EP4S100G5H40 EP4S100G3F45, EP4S100G4F45 EP4S100G5F45 36 36 36 48 48 12 12 24 24 24 32 8G Transceiver Channels (2) 12 12 0 0 8 0 CMU Channels (PMA-only) (3) 12 12 12 12 16 16 Transceiver Channels EP4S40G2F40 (4) EP4S40G5H40 Total Transceiver Channels 36 10G Transceiver Channels (1) EP4S100G2F40 Notes to Table 1-4: (1) 10G transceiver channels support data rates between 600 Mbps and 11.3 Gbps. (2) 8G transceiver channels support data rates between 600 Mbps and 8.5 Gbps. All 10G transceiver channels can also be configured as 8G transceiver channels. For example, the EP4S40G2F40 device has twenty-four 8G transceiver channels and the EP4S100G5F45 device has thirty-two 8G transceiver channels. (3) CMU channels that support data rates between 600 Mbps and 6.5 Gbps are PMA-only channels that do not have PCS circuitry. For more information, refer to "CMU Channel Architecture" on page 1-100. (4) F40 devices use 1517-pin flip chip packages. H40 devices use 1517-pin hybrid flip chip packages. F45 devices use 1932-pin flip chip packages. Table 1-5 lists the transceiver blocks in each Stratix IV GT device that support transceiver channels up to 11.3 Gbps. Table 1-5. Transceiver Blocks in Stratix IV GT Devices Supporting Transceiver Channels up to 11.3 Gbps (Part 1 of 2) Device Member Total number of Transceiver Channels Transceiver Channel Location 12 regular transceiver channels (six 10G and six 8G), located in three transceiver blocks: EP4S40G2F40 36 Left Side--4 in GXBL2, 2 in GXBL1 Right Side--4 in GXBR2, 2 in GXBR1 Refer to Figure 1-5 on page 1-10. 12 regular transceiver channels (six 10G and six 8G), located in three transceiver blocks: EP4S40G5H40 36 Left Side--4 in GXBL2, 4 in GXBL1, 4 in GXBL0 Right Side--4 in GXBR2, 4 in GXBR1, 4 in GXBR0 Refer to Figure 1-6 on page 1-11. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-9 Table 1-5. Transceiver Blocks in Stratix IV GT Devices Supporting Transceiver Channels up to 11.3 Gbps (Part 2 of 2) Device Member Total number of Transceiver Channels Transceiver Channel Location 12 regular transceiver channels, capable of 10G each, located in three transceiver blocks: EP4S100G2F40 36 Left Side--4 in GXBL2, 4 in GXBL1, 4 in GXBL0 Right Side--4 in GXBR2, 4 in GXBR1, 4 in GXBR0 Refer to Figure 1-7 on page 1-12. 12 regular transceiver channels, capable of 10G each, located in three transceiver blocks: EP4S100G5H40 36 Left Side--4 in GXBL2, 4 in GXBL1, 4 in GXBL0 Right Side--4 in GXBR2, 4 in GXBR1, 4 in GXBR0 Refer to Figure 1-8 on page 1-13. 16 regular transceiver channels (twelve 10G and four 8G) located in four transceiver blocks: EP4S100G3F45 EP4S100G4F45 48 Left Side--4 in GXBL3, 4 in GXBL2, 4 in GXBL1, 4 in GXBL0 (8G) Right Side--4 in GXBR3, 4 in GXBR2, 4 in GXBR1, 4 in GXBR0 (8G) Refer to Figure 1-9 on page 1-14. 16 regular transceiver channels, all capable of 10G each, located in four transceiver blocks: EP4S100G5F45 48 Left Side--4 in GXBL3, 4 in GXBL2, 4 in GXBL1, 4 in GXBL0 Right Side--4 in GXBR3, 4 in GXBR2, 4 in GXBR1, 4 in GXBR0 Refer to Figure 1-10 on page 1-15. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-10 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations Figure 1-5 shows the transceiver channel, PLL, and PCIe hard IP block locations for the EP4S40G2F40 Stratix IV GT devices. Figure 1-5. Transceiver Channel, PLL, and PCIe Hard IP Block Locations in EP4S40G2F40 Stratix IV GT Devices (1), (2) EP4S40G2F40 Transceiver Block GXBR2 Transceiver Block GXBL2 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (10G) ATX PLL (10G) Transceiver Block GXBL1 Transceiver Block GXBR1 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 8G Channel 1 8G Channel 1 8G Channel 0 8G Channel 0 ATX PLL (6G) Transceiver Block GXBL0 8G Channel 3 PCIe hard IP 10G Channel 3 PCIe hard IP 10G Channel 3 ATX PLL (6G) Transceiver Block GXBR0 8G Channel 3 8G Channel 2 8G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 8G Channel 1 8G Channel 1 8G Channel 0 8G Channel 0 Notes to Figure 1-5: (1) EP4S40G2F40ES1 devices do not have 10G auxiliary transmit (ATX) PLL blocks. Use the CMU PLL to generate transceiver clocks for channels configured at 11.3 Gbps. (2) If you are using the PCIe hard IP block, the EP4S40G2F40 device is not able to migrate to the EP4S40G5H40 device. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-11 Figure 1-6 shows the transceiver channel, PLL, and PCIe hard IP block locations for the EP4S40G5H40 Stratix IV GT devices. Figure 1-6. Transceiver Channel, PLL, and PCIe Hard IP Block Locations in EP4S40G5H40 Stratix IV GT Devices (1) EP4S40G5H40 Transceiver Block GXBR2 Transceiver Block GXBL2 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (10G) Transceiver Block GXBL1 PCIe hard IP 10G Channel 3 PCIe hard IP 10G Channel 3 ATX PLL (10G) Transceiver Block GXBR1 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 8G Channel 1 8G Channel 1 8G Channel 0 8G Channel 0 ATX PLL (6G) ATX PLL (6G) Transceiver Block GXBL0 Transceiver Block GXBR0 8G Channel 3 8G Channel 3 8G Channel 2 8G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 8G Channel 1 8G Channel 1 8G Channel 0 8G Channel 0 Note to Figure 1-6: (1) If you are using the PCIe hard IP block, the EP4S40G2F40 device is not able to migrate to the EP4S40G5H40 device. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-12 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations Figure 1-7 shows the transceiver channel, PLL, and PCIe hard IP block locations in EP4S100G2F40 Stratix IV GT devices. Figure 1-7. Transceiver Channel, PLL, and PCIe Hard IP Block Locations in EP4S100G2F40 Stratix IV GT Devices EP4S100G2F40 Transceiver Block QR2 Transceiver Block QL2 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (10G) ATX PLL (10G) Transceiver Block QR1 Transceiver Block QL1 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (6G) Transceiver Block QL0 PCIe hard IP 10G Channel 3 PCIe hard IP 10G Channel 3 ATX PLL (6G) Transceiver Block QR0 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 Notes to Figure 1-7: (1) EP4S100G2F40ES1 devices do not have 10G ATX PLL blocks. Use the CMU PLL to generate transceiver clocks for channels configured at 11.3 Gbps. (2) If you are using the PCIe hard IP block, the EP4S100G2F40 device is not able to migrate to the EP4S100G5H40 device. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-13 Figure 1-8 shows the transceiver channel, PLL, and PCIe hard IP block locations for the EP4S100G5H40 Stratix IV GT devices. Figure 1-8. Transceiver Channel, PLL, and Hard IP Block Locations in EP4S100G5H40 Stratix IV GT Devices EP4S100G5H40 Transceiver Block QR2 Transceiver Block QL2 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (10G) Transceiver Block QL1 PCIe hard IP 10G Channel 3 PCIe hard IP 10G Channel 3 ATX PLL (10G) Transceiver Block QR1 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 ATX PLL (6G) Transceiver Block QL0 ATX PLL (6G) Transceiver Block QR0 10G Channel 3 10G Channel 3 10G Channel 2 10G Channel 2 CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 10G Channel 1 10G Channel 1 10G Channel 0 10G Channel 0 Note to Figure 1-8: (1) If you are using the PCIe hard IP block, the EP4S100G2F40 device is not able to migrate to the EP4S100G5H40 device. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-14 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations Figure 1-9 shows the transceiver channel, PLL, and PCIe hard IP block locations for the EP4S100G3F45 and EP4S100G4F45 Stratix IV GT devices. Figure 1-9. Transceiver Channel, PLL, and PCIe Hard IP Block Locations in EP4S100G3F45 and EP4S100G4F45 Stratix IV GT Devices (1) EP4S100G3F45, EP4S100G4F45 Transceiver Block QR3 Transceiver Block QL3 Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (10G) Transceiver Block QL2 PCIe hard IP Channel 3 (10G) PCIe hard IP Channel 3 (10G) ATX PLL (10G) Transceiver Block QR2 Channel 3 (10G) Channel 3 (10G) Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (6G) ATX PLL (6G) Transceiver Block QR1 Transceiver Block QL1 Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (6G) Transceiver Block QL0 PCIe hard IP Channel 3 (10G) PCIe hard IP Channel 3 (10G) ATX PLL (6G) Transceiver Block QR0 Channel 3 (8G) Channel 3 (8G) Channel 2 (8G) Channel 2 (8G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (8G) Channel 1 (8G) Channel 0 (8G) Channel 0 (8G) Note to Figure 1-9: (1) EP4S100G2F40C2ES1 devices do not have 10G ATX PLL blocks. Use the CMU PLL to generate transceiver clocks for channels configured at 11.3 Gbps. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Channel Locations 1-15 Figure 1-10 shows the transceiver channel, PLL, and PCIe hard IP block locations for the EP4S100G5F45 Stratix IV GT devices. Figure 1-10. Transceiver Channel, PLL, and PCIe Hard IP Block Locations in EP4S100G5F45 Stratix IV GT Devices (1) EP4S100G5F45 Transceiver Block QR3 Transceiver Block QL3 Channel 3 (10G) Channel 3 (10G) Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (10G) Transceiver Block QL2 PCIe hard IP CMU Channel 0 Channel 1 (10G) PCIe hard IP CMU Channel 0 ATX PLL (10G) Transceiver Block QR2 Channel 3 (10G) Channel 3 (10G) Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (6G) ATX PLL (6G) Transceiver Block QR1 Transceiver Block QL1 Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) ATX PLL (6G) Transceiver Block QL0 PCIe hard IP Channel 3 (10G) PCIe hard IP Channel 3 (10G) ATX PLL (6G) Transceiver Block QR0 Channel 3 (10G) Channel 3 (10G) Channel 2 (10G) Channel 2 (10G) CMU Channel 1 CMU Channel 1 CMU Channel 0 CMU Channel 0 Channel 1 (10G) Channel 1 (10G) Channel 0 (10G) Channel 0 (10G) Note to Figure 1-10: (1) EP4S100G5F45 devices are the same as EP4S100G3F45 and EP4S100G4F45 devices except that the GXBR0 transceiver block is 10G instead of 8G. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-16 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Transceiver Block Architecture Figure 1-11 shows the transceiver block architecture of Stratix GX and GT devices. Figure 1-11. Top-Level View of a Transceiver Block Transceiver Block Transceiver Block QL1 Transceiver Block QR1 Transceiver Channel 3 1 Transceiver Channel 2 1 Channel 3 Channel 2 Central Control Unit (CCU) CMU1 Channel 2 Channel 3 Channel 2 Channel 1 Channel 0 Channel 1 Channel 0 3 CMU0 Channel 2 Transceiver Block QL0 Transceiver Block QR0 Channel 3 Channel 2 Channel 3 Channel 2 Channel 1 Channel 0 Channel 1 Channel 0 Transceiver Channel 1 1 Transceiver Channel 0 1 Each transceiver block has the following components: 1. Four full-duplex (transmitter and receiver) transceiver channels that support serial data rates from 600 Mbps to 8.5 Gbps in Stratix IV GX devices and 600 Mbps to 11.3 Gbps in Stratix IV GT devices. For more information, refer to "Transceiver Channel Architecture" on page 1-17. 2. Two CMU channels--CMU0 and CMU1 channels--that provide the high-speed serial and low-speed parallel clock to the transceiver channels. For more information, refer to "CMU Channel Architecture" on page 1-100. 3. Central control unit (CCU) that implements the XAUI state machine for XGMII-to-PCS code group conversion, XAUI deskew state machine, shared control signal generation block, PCIe rateswitch controller block, and reset control logic Stratix IV Device Handbook Volume 2: Transceivers The shared control signal generation block provides control signals to the transceiver channels in bonded functional modes, such as XAUI, PCIe, and Basic x4. The PCIe rateswitch controller block controls the rateswitch circuit in the CMU0 channel in x4 configurations. In PCIe x8 configuration, the PCIe rateswitch controller block of the CCU in the master transceiver block is active. For more information, refer to "PCIe Gen2 (5 Gbps) Support" on page 1-140. September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-17 The Stratix IV GT transceiver architecture has the following components: Regular transceiver channels with PMA and PCS support CMU channels with PMA-only support ATX PLL blocks Four transceiver channels and two CMU channels are located in each transceiver block on the left and right sides of the device. Each Stratix IV GT device also has two 10G ATX PLLs that support data rates between 9.9 Gbps and 11.3 Gbps. Additionally, each Stratix IV GT device has two 6G ATX PLLs that support data rates between 600 Mbps and 6.5 Gbps, except the EP4S100G5F45 device that has four 6G ATX PLLs. 1 The 6G ATX PLL does not support all data rates between 600 Mbps and 6.5 Gbps. Transceiver Channel Architecture Figure 1-12 shows the Stratix IV GX and GT transceiver channel datapath. Figure 1-12. Stratix IV GX and GT Transceiver Datapath wrclk rdclk CDR Word Aligner Deskew FIFO Receiver Channel PMA Rate Match FIFO 8B/10 Decoder Byte Deserializer RX Phase Compensation FIFO Receiver Channel PCS Deserializer rdclk 8B/10B Encoder rx_datain wrclk Serializer Byte Serializer Byte Ordering PIPE Interface PCIe hard IP FPGA Fabric TX Phase Compensation FIFO tx_dataout Transmitter Channel PMA Transmitter Channel PCS Transmitter Channel Datapath Receiver Channel Datapath Each transceiver channel consists of the: Transmitter channel, further divided into: Transmitter channel PCS Transmitter channel PMA Receiver channel, further divided into: Receiver channel PCS Receiver channel PMA Each transceiver channel interfaces to either the PCIe hard IP block (PCIe hard IP-transceiver interface) or directly to the FPGA fabric (FPGA fabric-transceiver interface). The transceiver channel interfaces to the PCIe hard IP block if the hard IP block is used to implement the PCIe PHY MAC, data link layer, and transaction layer. Otherwise, the transceiver channel interfaces directly to the FPGA fabric. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-18 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Each regular Stratix IV GT transceiver channel can be categorized into: 1 8G transceiver channel--supports data rates between 600 Mbps and 8.5 Gbps 10G Transceiver Channel--supports data rates between 600 Mbps and 11.3 Gbps The PCIe hard IP-transceiver interface is beyond the scope of this chapter. This chapter describes the FPGA fabric-transceiver interface. f For more information about the PCIe hard IP block, refer to the PCI Express Compiler User Guide. Figure 1-13 shows the FPGA fabric-transceiver interface and transceiver PMA-PCS interface. Figure 1-13. FPGA Fabric-Transceiver Interface and Transceiver PMA-PCS Interface FPGA Fabric-Transceiver Interface PMA-PCS Interface FPGA Fabric FPGA Fabric CDR Deserializer Word Aligner Deskew FIFO Receiver Channel PMA Rate Match FIFO 8B/10 Decoder Byte Deserializer Receiver Channel PCS Byte Ordering tx_clkout 8B/10B Encoder rdclk rx_datain wrclk rdclk Serializer Byte Serializer wrclk RX Phase Compensation FIFO PCIe hard IP PIPE Interface tx_clkout TX Phase Compensation FIFO tx_dataout Transmitter Channel PMA Transmitter Channel PCS PMA-PCS Interface The transceiver channel datapath can be divided into the following two modes based on the FPGA fabric-transceiver interface width (channel width) and the transceiver channel PMA-PCS width (serialization factor): Single-width mode Double-width mode Table 1-6 lists the FPGA fabric-transceiver interface widths (channel width) and transceiver PMA-PCS widths (serialization factor) allowed in single-width and double-width modes. Table 1-6. FPGA Fabric-Transceiver Interface Width and Transceiver PMA-PCS Widths (Part 1 of 2) Name PMA-PCS interface widths FPGA fabric-transceiver interface width Stratix IV Device Handbook Volume 2: Transceivers Single-Width Double-Width 8/10 bit 16/20 bit 8/10 bit 16/20 bit 16/20 bit 32/40 bit September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-19 Table 1-6. FPGA Fabric-Transceiver Interface Width and Transceiver PMA-PCS Widths (Part 2 of 2) Name Single-Width Supported functional modes Double-Width PCIe Gen1 and Gen2 (OIF) CEI PHY Interface XAUI SONET/SDH OC96 GIGE Basic double-width Serial RapidIO SONET/SDH OC12 and OC48 SDI Basic single-width Data rate range in Basic functional mode 0.6 Gbps to 3.75 Gbps 1 Gbps to 8.5 Gbps Transmitter Channel Datapath The transmitter channel datapath, shown in Figure 1-12 on page 1-17, consists of the following blocks: TX phase compensation FIFO Byte serializer 8B/10B encoder Transmitter output buffer The Stratix IV GX and GT transceiver provides the Enable low latency PCS mode option in the ALTGX MegaWizardTM Plug-In Manager. If you select this option, the 8B/10B encoder in the datapath is disabled. TX Phase Compensation FIFO The TX phase compensation FIFO interfaces the transmitter channel PCS and the FPGA fabric PCIe interface. It compensates for the phase difference between the low-speed parallel clock and the FPGA fabric interface clock. The TX phase compensation FIFO operates in low-latency and high-latency modes. Figure 1-14 shows the datapath and clocking of the TX phase compensation FIFO. Figure 1-14. TX Phase Compensation FIFO Data Path from the FPGA Fabric or PIPE Interface TX Phase Compensation FIFO wr_clk Data Path to the Byte Serializer or the 8B/10B Encoder or Serializer rd_clk (1) tx_coreclk tx_clkout coreclkout Note to Figure 1-14: (1) The clock used to clock the write side of the compensation FIFO is based on whether or not you enable the tx_coreclk option in the ALTGX MegaWizard Plug-In Manager. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-20 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-7 lists the TX phase compensation FIFO modes. Table 1-7. TX Phase Compensation FIFO Modes Modes Description The FIFO is four words deep. Latency through the FIFO is two to three FPGA fabric parallel clock cycles (pending characterization). Low-Latency The default setting for every mode. High-Latency The FIFO is eight words deep. The latency through the FIFO is four to five FPGA parallel cycles (pending characterization). Non-Bonded Functional For example, in GIGE mode, the read port of the phase compensation FIFO is clocked by the low-speed parallel clock. The write clock is fed by the tx_clkout port of the associated channel. Bonded Functional For example, in XAUI mode, the write clock of the FIFO is clocked by coreclkout provided by the CMU0 clock divider block. You can clock the write side using tx_coreclk provided from the FPGA fabric by enabling the tx_coreclk port in the ALTGX MegaWizard Plug-In Manager. If you use this port, ensure that there is 0 parts-per-million (PPM) difference in frequency between the write and read side. The Quartus(R) II software requires that you provide a 0 PPM assignment in the Assignment Editor. f For more information about the TX phase compensation FIFO, refer to the "Limitations of the Quartus II Software-Selected Transmitter Phase Compensation FIFO Write Clock" section of the Transceiver Clocking in Stratix IV Devices chapter. Input Data In PCIe functional mode, the input data comes from the PCIe interface. In all other functional modes, the input data comes directly from the FPGA fabric. Output Data Destination Block The output from the TX phase compensation FIFO is used by the byte serializer block, 8B/10B encoder, or serializer block. Table 1-8 lists the conditions under which the TX phase compensation FIFO outputs are provided to these blocks. Table 1-8. Output Data Destination Block for the TX Phase Compensation FIFO Output Data Byte Serializer If you select: single-width mode and channel width = 16 or 20 If you select: double-width mode and channel width = 32 or 40 Stratix IV Device Handbook Volume 2: Transceivers 8B/10B Encoder Serializer If you select: If you select: single-width mode and channel width = 8 and 8B/10B encoder enabled low-latency PCS bypass mode enabled or single-width mode and channel width = 8 or 10 If you select: If you select: double-width mode and channel width = 16 and 8B/10B encoder enabled low-latency PCS bypass mode enabled or double-width mode and channel width = 16 or 20 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-21 TX Phase Compensation FIFO Status Signal An optional tx_phase_comp_fifo_error port is available in all functional modes to indicate a receiver phase compensation FIFO overflow or under-run condition. The tx_phase_comp_fifo_error signal is asserted high when the TX phase compensation FIFO either overflows or under-runs due to any frequency PPM difference between the FIFO read and write clocks. If the tx_phase_comp_fifo_error flag is asserted, verify the FPGA fabric-transceiver interface clocking to ensure that there is 0 PPM difference between the TX phase compensation FIFO read and write clocks. Byte Serializer The byte serializer divides the input datapath by two. This allows you to run the transceiver channel at higher data rates while keeping the FPGA fabric interface frequency within the maximum limit stated in the "Interface Frequency" section in the DC and Switching Characteristics for Stratix IV Devices chapter. In single-width mode, it converts the two-byte-wide datapath to a one-byte-wide datapath. In double-width mode, it converts the four-byte-wide datapath to a two-byte-wide datapath. It is optional in configurations that do not exceed the FPGA fabrictransceiver interface maximum frequency limit. For example, if you want to run the transceiver channel at 6.25 Gbps, without the byte serializer in double-width mode, the FPGA fabric interface clock frequency must be 312.5 MHz (6.25/20). This violates the FPGA fabric interface frequency limit. When you use the byte serializer, the FPGA fabric interface frequency is 156.25 MHz (6.25G/40). You can enable the byte serializer in single-width or double-width mode. 1 The byte deserializer is required in configurations that exceed the FPGA fabric-transceiver interface maximum frequency limit. f For more information about the maximum frequency limit for the transceiver interface, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Single-Width Mode Figure 1-15 shows the byte serializer datapath in single-width mode. For data port width, refer to Table 1-9. Figure 1-15. Byte Serializer Datapath in Single-Width Mode (1), datain[] Byte Serializer /2 (2) dataout[] Low-Speed Parallel Clock Notes to Figure 1-15: (1) For the datain[] and dataout[] port widths, refer to Table 1-9. (2) The datain signal is the input from the FPGA fabric that has already passed through the TX phase compensation FIFO. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-22 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The byte serializer forwards the LSByte first, followed by the MSByte. The input data width to the byte serializer depends on the channel width option that you selected in the ALTGX MegaWizard Plug-In Manager. For example, in single-width mode, assuming a channel width of 20, the byte serializer sends out the least significant word datain[9:0] of the parallel data from the FPGA fabric, followed by datain[19:10]. Table 1-9 lists the input and output data widths of the byte serializer in single-width mode. Table 1-9. Input and Output Data Width of the Byte Serializer in Single-Width Mode Deserialization Width Single-width mode Input Data Width to the Byte Serializer Output Data Width from the Byte Serializer 16 8 20 10 Double-Width Mode Figure 1-16 shows the byte serializer datapath in double-width mode. For data port width, refer to Table 1-10. Figure 1-16. Byte Serializer Datapath in Double-Width Mode (1), datain[] Byte Serializer /2 (2) dataout[] Low-Speed Parallel Clock Notes to Figure 1-16: (1) For the datain[] and dataout[] port width, refer to Table 1-10. (2) The datain signal is the input from the FPGA fabric that has already passed through the TX phase compensation FIFO. The operation in double-width mode is similar to that of single-width mode. For example, assuming a channel width of 40, the byte serializer forwards datain[19:0] first, followed by datain[39:20]. Table 1-10 lists the input and output data widths of the byte serializer in double-width mode. Table 1-10. Input and Output Data Width of the Byte Serializer in Double-Width Mode Deserialization Width Double-width mode Input Data Width to the Byte Serializer Output Data Width from the Byte Serializer 32 16 40 20 Asserting the tx_digitalreset signal resets the byte serializer block. If you select the 8B/10B Encoder option in the ALTGX MegaWizard Plug-In Manager, the 8B/10B encoder uses the output from the byte serializer. Otherwise, the byte serializer output is forwarded to the serializer. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-23 8B/10B Encoder The 8B/10B encoder generates 10-bit code groups from the 8-bit data and 1-bit control identifier. The 8B/10B encoder operates in two modes: single-width and double-width. Figure 1-17 shows the 8B/10B encoder in single-width and double-width mode. Figure 1-17. 8B/10B Encoder in Single-Width Mode Single Width Double Width From the Byte Serializer To the Serializer From the Byte Serializer datain[7:0] control_code 8B/10B Encoder 8B/10B Encoder dataout[9:0] tx_forcedisp To the Serializer datain[15:8] tx_dispval control_code[1] tx_forcedisp[1] MSB Encoding dataout[19:10] tx_dispval[1] datain[7:0] control_code[0] dataout[9:0] tx_forcedisp[0] LSB Encoding tx_dispval[0] Single-Width Mode The left side of Figure 1-17 shows the 8B/10B encoder in single-width mode. In this mode, the 8B/10B encoder translates the 8-bit data to a 10-bit code group (control word or data word) with proper disparity. If the control_code input is high, the 8B/10B encoder translates the input data[7:0] to a 10-bit control word. If the control_code input is low, the 8B/10B encoder translates the input data[7:0] to a 10-bit data word. You can use the tx_forcedisp and tx_dispval ports to control the running disparity of the generated output data. For more information, refer to "Controlling Running Disparity" on page 1-27. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-24 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-18 shows the conversion format. The LSB is transmitted first. Figure 1-18. 8B/10B Conversion Format 7 6 5 4 3 2 1 control_code 0 H G F E D C B A 8B/10B Conversion j h g f i e d c b a 9 8 6 5 4 1 7 3 2 0 LSB MSB Control Code Encoding The ALTGX MegaWizard Plug-In Manager provides the tx_ctrlenable port to indicate whether the 8-bit data at the tx_datain port should be encoded as a control word (Kx.y). When tx_ctrlenable is low, the 8B/10B encoder block encodes the byte at the tx_datain port (the user-input port) as data (Dx.y). When tx_ctrlenable is high, the 8B/10B encoder encodes the input data as a Kx.y code group. The waveform in Figure 1-19 shows the second 0 x BC encoded as a control word (K28.5). The rest of the tx_datain bytes are encoded as a data word (Dx.y). Figure 1-19. Control Word and Data Word Transmission clock tx_datain[7:0] 83 78 D3.4 D24.3 BC BC 0F 00 K28.5 D15.0 D0.0 BF 3C tx_ctrlenable code group D28.5 D31.5 D28.1 The IEEE 802.3 8B/10B encoder specification identifies only a set of 8-bit characters for which tx_ctrlenable should be asserted. If you assert tx_ctrlenable for any other set of bytes, the 8B/10B encoder might encode the output 10-bit code as an invalid code (it does not map to a valid Dx.y or Kx.y code), or unintended valid Dx.y code, depending on the value entered. It is possible for a downstream 8B/10B decoder to decode an invalid control word into a valid Dx.y code without asserting code error flags. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1 1-25 For example, depending on the current running disparity, the invalid code K24.1 (tx_datain = 8'h38 + tx_ctrl = 1'b1) can be encoded to 10'b0110001100 (0 x 18C), which is equivalent to a D24.6+ (8'hD8 from the RD+ column). Altera recommends that you do not assert tx_ctrlenable for unsupported 8-bit characters. Reset Condition The tx_digitalreset signal resets the 8B/10B encoder. During reset, running disparity and data registers are cleared. Also, the 8B/10B encoder outputs a K28.5 pattern from the RD- column continuously until tx_digitalreset is de-asserted. The input data and control code from the FPGA fabric is ignored during the reset state. After reset, the 8B/10B encoder starts with a negative disparity (RD-) and transmits three K28.5 code groups for synchronization before it starts encoding and transmitting the data on its output. 1 While tx_digitalreset is asserted, the downstream 8B/10B decoder that receives the data might observe synchronization or disparity errors. Figure 1-20 shows the reset behavior of the 8B/10B encoder. When in reset (tx_digitalreset is high), a K28.5- (K28.5 10-bit code group from the RD- column) is sent continuously until tx_digitalreset is low. Due to some pipelining of the transmitter channel PCS, some "don't cares" (10'hxxx) are sent before the three synchronizing K28.5 code groups. User data follows the third K28.5 code group. Figure 1-20. 8B/10B Encoder Output During tx_digitalreset Assertion clock tx_digitalreset dataout[9:0] K28.5- K28.5- K28.5- xxx ... xxx K28.5- K28.5+ K28.5- Dx.y+ Double-Width Mode In double-width mode, the 8B/10B encoder operates in a cascaded mode, as shown on the right side of Figure 1-20 on page 1-25. The LSByte of the input data is encoded and transmitted prior to the MSByte. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-26 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture In double-width mode, the cascaded 8B/10B encoder generates two 10-bit code groups from two 8-bit data and their respective control code identifiers. Figure 1-21 shows the conversion format. The LSB shown in Figure 1-21 is transmitted first. Figure 1-21. 8B/10B Conversion Format in Double-Width Mode CTRL[1:0] H ' G ' F' E' D ' C ' B' A ' H G F E D C B A 15 6 5 4 3 2 1 0 14 13 12 11 10 9 8 7 Parallel Data Cascaded 8B/10B Conversion j' h ' 19 18 g' f' i' e' d' c' b' a' j h g f i e d c b a 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 LSB MSB Control Code Encoding In double-width mode, the tx_ctrlenable[1:0] port is used to identify which 8-bit data is to be encoded as a control word. The lower bit, tx_ctrlenable[0], is associated with the LSByte; the upper bit, tx_ctrlenable[1], is associated with the MSByte. When tx_ctrlenable is low, the byte at the tx_datain port of the transceiver is encoded as data (Dx.y); otherwise, it is encoded as a control code (Kx.y). Figure 1-22 shows that only the lower byte of the tx_datain[15:0] port is encoded as a control code because tx_ctrlenable[0] is high in the second clock cycle. Figure 1-22. Encoded Control Word and Data Word Transmission clock tx_datain[15:0] 8378 tx_ctrlenable[1:0] code group BCBC 0 D3.4 0F00 1 D24.3 D28.5 BF3C 0 K28.5 D15.0 D0.0 D31.5 D28.1 The 8B/10B encoder does not check to see if the code word entered is one of the 12 valid control code groups specified in the IEEE 802.3 8B/10B encoder specification. If an invalid control code is entered, the resulting 10-bit code may be encoded as an invalid code (it does not map to a valid Dx.y or Kx.y code), or unintended valid Dx.y code, depending on the value entered. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-27 The following is an example of an invalid control word encoded into a valid Dx.y code. With an encoding invalid code K24.1 (tx_datain = 8'h38 + tx_ctrl = 1'b1), depending on the current running disparity, the K24.1 can be encoded as 10'b0110001100 (0 x 18C), which is equivalent to a D24.6+ (8'hD8 from the RD+ column). An 8B/10B decoder can decode this and not assert a code error flag. 1 Altera does not recommend sending invalid control words to the 8B/10B encoder. Reset Condition The tx_digitalreset signal resets the 8B/10B encoder. During reset, the running disparity and data registers are cleared. Also, the 8B/10B encoder outputs a K28.5 pattern with proper disparity continuously until tx_digitalreset goes low. The inputs from the tx_datain and tx_ctrlenable ports are ignored during the reset state. After reset, the 8B/10B encoder starts the LSByte with a negative disparity (RD-) and the MSByte with a positive disparity (RD+) and transmits six K28.5 code groups (three on the LSByte and three on the MSByte encoder) for synchronizing before it starts encoding and transmitting data. 1 If the tx_digitalreset signal is asserted, the downstream 8B/10B decoder receiving the data might get synchronization or disparity errors. Figure 1-23 shows the reset behavior of the 8B/10B encoder. When in reset (tx_digitalreset is high), a K28.5- on LSB and K28.5+ on MSB is sent continuously until tx_digitalreset is low. Due to pipelining of the TX channel, there will be some "don't cares" (10'hxxx) until the first K28.5 is sent (Figure 1-23 shows six "don't cares", but the number of "don't cares" can vary). Both the LSByte and MSByte transmit three K28.5s before the data at the tx_datain port is encoded and sent out. Figure 1-23. Transmitted Output Data When tx_digitalreset is Asserted clock tx_digitalreset dataout[19:10] dataout[9:0] k28.5+ k28.5- k28.5+ k28.5- k28.5+ xxx xxx xxx k28.5+ k28.5+ k28.5+ Dx.y+ k28.5- xxx xxx xxx k28.5- k28.5- k28.5- Dx.y- Controlling Running Disparity After power on or reset, the 8B/10B encoder has a negative disparity and chooses the 10-bit code from the RD- column (refer to the 8B/10B encoder specification for the RD+ and RD- column values). The ALTGX MegaWizard Plug-In Manager provides the tx_forcedisp and tx_dispval ports to control the running disparity of the output from the 8B/10B encoder. These ports are available only in Basic single-width and Basic double-width modes. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-28 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture A high value on the tx_forcedisp port is the control signal to the disparity value of the output data. The disparity value (RD+ or RD-) is indicated by the value on the tx_dispval port. If the tx_forcedisp port is low, tx_dispval is ignored and the current running disparity is not altered. Forcing disparity can either maintain the current running disparity calculations if the forced disparity value (on the tx_dispval bit) matches the current running disparity, or flip the current running disparity calculations if it does not. If the forced disparity flips the current running disparity, the downstream 8B/10B decoder might detect a disparity error. Table 1-11 lists the tx_forcedisp and tx_dispval port values. Table 1-11. tx_forcedisp and tx_dispval Port Values tx_forcedisp tx_dispval Disparity Value 0 X Current running disparity has no change 1 0 Encoded data has positive disparity 1 1 Encoded data has negative disparity Figure 1-24 shows the current running disparity being altered in Basic single-width mode by forcing a positive disparity K28.5 when it was supposed to be a negative disparity K28.5. In this example, a series of K28.5 code groups are continuously being sent. The stream alternates between a positive running disparity (RD+) K28.5 and a negative running disparity (RD-) K28.5 to maintain a neutral overall disparity. The current running disparity at time n + 3 indicates that the K28.5 in time n + 4 should be encoded with a negative disparity. Because tx_forcedisp is high at time n + 4, and tx_dispval is low, the K28.5 at time n + 4 is encoded as a positive disparity code group. Figure 1-24. 8B/10B Encoder Force Running Disparity Operation in Single-Width Mode n n+1 n+2 n+3 n+4 n+5 n+6 n+7 BC BC BC BC BC BC BC BC RD- RD+ RD- RD+ RD+ RD- RD+ RD- 17C 283 283 17C 283 17C clock tx_in[7:0] tx_ctrlenable tx_forcedisp tx_dispval Current Running Disparity dataout[9:0] 17C 283 Figure 1-25 shows the current running disparity being altered in Basic double-width mode by forcing a positive disparity on a negative disparity K28.5. In this example, a series of K28.5 are continuously being sent. The stream alternates between a positive ending running disparity (RD+) K28.5 and a negative ending running disparity (RD-) K28.5 as governed by the 8B/10B encoder specification to maintain a neutral overall disparity. The current running disparity at the end of time n + 2 indicates that the Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-29 K28.5 at the low byte position in time n + 4 should be encoded with a positive disparity. Because tx_forcedisp is high at time n + 4, the low signal level of tx_dispval is used to convert the lower byte K28.5 to be encoded as a positive disparity code word. As the upper bit of tx_forcedisp is low at n + 4, the high byte K28.5 takes the current running disparity from the low byte. Figure 1-25. 8B/10B Encoder Force Current Running Disparity in Double-Width Mode n+2 n n+4 clock tx_datain[15:0] BC BC BC BC tx_ctrlenable[1:0] BC BC 01 00 11 tx_forcedisp[1:0] 00 tx_dispval[1:0] Current Running Disparity tx_dataout[19:0] BC BC 00 RD- RD + RD- RD+ RD+ RD- RD+ RD- 17C 283 17C 283 283 17C 283 17C Transmitter Polarity Inversion The positive and negative signals of a serial differential link might accidentally be swapped during board layout. Solutions like a board re-spin or major updates to the logic in the FPGA fabric can be expensive. The transmitter polarity inversion feature is provided to correct this situation. An optional tx_invpolarity port is available in all functional modes except (OIF) CEI PHY to dynamically enable the transmitter polarity inversion feature. In single-width mode, a high value on the tx_invpolarity port inverts the polarity of every bit of the 8-bit or 10-bit input data word to the serializer in the transmitter datapath. In double-width mode, a high value on the tx_invpolarity port inverts the polarity of every bit of the 16-bit or 20-bit input data word to the serializer in the transmitter datapath. Because inverting the polarity of each bit has the same effect as swapping the positive and negative signals of the differential link, correct data is seen by the receiver. tx_invpolarity is a dynamic signal and might cause initial disparity errors at the receiver of an 8B/10B encoded link. The downstream system must be able to tolerate these disparity errors. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-30 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-26 shows the transmitter polarity inversion feature in a single-width and double-width datapath configuration. Figure 1-26. Transmitter Polarity Inversion in Single-Width and Double-Width Mode Double-Width Configuration Single-Width Configuration Output from transmitter PCS 0 MSB 1 1 0 1 0 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 0 0 1 0 1 0 1 1 0 1 0 1 0 0 1 1 0 tx_invpolarity = high 1 Stratix IV Device Handbook Volume 2: Transceivers Converted data output to the transmitter serializer LSB 0 MSB LSB September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-31 Transmitter Bit Reversal Table 1-12 lists the transmission bit order with and without the transmitter bit reversal enabled. Table 1-12. Transmission Bit Order for the Bit Reversal Feature Transmitter Bit Reversal Feature Not enabled (default) Single-Width Mode (8- or 10-Bit) Double-Width Mode (16- or 20-Bit) LSB to MSB LSB to MSB MSB to LSB MSB to LSB For example: Enabled For example: 8-bit--D[7:0]rewired to D[0:7] 16-bit--D[15:0]rewired to D[0:15] 10-bit-- D[9:0]rewired to D[0:9] 20-bit--D[19:0]rewired to D[0:19] Figure 1-27 shows the transmitter bit reversal feature in Basic single-width for a 10-bit wide datapath configuration. Figure 1-27. Transmitter Bit Reversal Operation in Basic Single-Width Mode Output from transmitter PCS TX bit reversal option enabled in the ALTGX MegaWizard September 2012 Altera Corporation Converted data output to the transmitter serializer D[9] D[0] D[8] D[1] D[7] D[2] D[6] D[3] D[5] D[4] D[4] D[5] D[3] D[6] D[2] D[7] D[1] D[8] D[0] D[9] Stratix IV Device Handbook Volume 2: Transceivers 1-32 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-28 shows the transmitter bit reversal feature in Basic double-width mode for a 20-bit wide datapath configuration. Figure 1-28. Transmitter Bit Reversal Operation in Basic Double-Width Mode Output from transmitter PCS TX bit reversal option enabled in the ALTGX MegaWizard Stratix IV Device Handbook Volume 2: Transceivers Converted data output to the transmitter serializer D[19] D[0] D[18] D[1] D[17] D[2] D[16] D[3] D[15] D[4] D[14] D[5] D[13] D[6] D[12] D[7] D[11] D[8] D[10] D[9] D[9] D[10] D[8] D[11] D[7] D[12] D[6] D[13] D[5] D[14] D[4] D[15] D[3] D[16] D[2] D[17] D[1] D[18] D[0] D[19] September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-33 Serializer The serializer converts the incoming low-speed parallel signal from the transceiver PCS to high-speed serial data and sends it to the transmitter buffer. The serializer supports an 8-bit or 10-bit serialization factor in single-width mode and a 16-bit or 20-bit serialization factor in double-width mode. The serializer block drives the serial data to the output buffer, as shown in Figure 1-29. The serializer block sends out the LSB of the input data. Figure 1-29. Serializer Block in 8-Bit PCS-PMA Interface D7 D7 D6 D6 D5 D5 D4 D4 D3 D3 D2 D2 D1 D1 D0 D0 8 To Output Buffer parallel clock from local divider block parallel clock from CMU0 clock divider Low-Speed Parallel Clock parallel clock from master transceiver block (1) serial clock from local divider block High-Speed Serial Clock serial clock from CMU0 clock divider serial clock from master transceiver block (1) Note to Figure 1-29: (1) The CMU0 clock divider of the master transceiver block provides the clocks. It is used only in bonded modes (for example, Basic x8, PCIe x8 mode). Figure 1-30 shows the serial bit order of the serializer block output. In this example, a constant 8'h6A (01101010) value is serialized and the serial data is transmitted from LSB to MSB. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-34 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-30. Serializer Bit Order (1) Low-speed parallel clock High-speed serial clock tx_datain[7..0] tx_dataout[0] 01101010 00000000 0 1 0 1 0 1 1 0 Note to Figure 1-30: (1) It is assumed that the input data to the serializer is 8 bits (channel width = 8 bits or 16 bits with the 8B/10B encoder disabled). Transmitter Output Buffer The Stratix IV GX and GT transmitter buffers are architecturally similar to each other. They both support programmable output differential voltage (VOD), pre-emphasis, and on-chip termination (OCT) settings. The transmitter buffer power supply only provides voltage to the transmitter output buffers in the transceiver channels. The transmitter output buffer, shown in Figure 1-31, has additional circuitry to improve signal integrity, such as VOD, programmable three-tap pre-emphasis circuit, internal termination circuitry, and receiver detect capability to support PCIe functional mode. Figure 1-31. Transmitter Output Buffer 42.5, 50 , 60 , 75 +VTT- Programmable Pre-emphasis and VOD Transmitter Output Pins 42.5, 50 , 60 , 75 Receiver Detect Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-35 Table 1-13 and Table 1-14 list the supported settings of the transmitter buffers in the Stratix IV GX and GT devices, respectively. Table 1-13. Supported Settings for the Stratix IV GX Transmitter Buffer Parameter Setting 600 Mbps to 8.5 Gbps (1.4 V) Data rate 600 Mbps to 6.5 Gbps (1.5 V) Transmitter buffer power (VCCH_GXBL/Rn) 1.4 V or 1.5 V 1.4-V and 1.5-V pseudo current mode logic (PCML) Transmitter buffer I/O standard Transmitter buffer VCM 0.65 V Table 1-14. Supported Settings for the Stratix IV GT Transmitter Buffer Parameter Setting Data rate 600 Mbps--11.3 Gbps Transmitter buffer power (VCCH_GXBL/Rn) 1.4 V Transmitter buffer I/O standard 1.4-V PCML Transmitter buffer VCM 0.65 V Programmable Transmitter Termination The Stratix IV GX and GT transmitter buffers includes programmable on-chip differential termination of 85, 100, 120, or 150 . The resistance is adjusted by the on-chip calibration circuit in the calibration block (for more information, refer to "Calibration Blocks" on page 1-201), which compensates for temperature, voltage, and process changes. The Stratix IV GX and GT transmitter buffers in the transceiver are current mode drivers. Therefore, the resultant VOD is a function of the transmitter termination value. For more information about resultant VOD values, refer to "Programmable Output Differential Voltage" on page 1-36. You can disable OCT and use external termination. If you select external termination, the transmitter common mode is tri-stated. You can set the transmitter termination in the ALTGX MegaWizard Plug-In Manager. You can also set the OCT through the Assignment Editor. Set the assignment shown in Table 1-15 to the transmitter serial output pin. Table 1-15. Stratix IV GX and GT OCT Assignment Settings Assign To September 2012 Transmitter Serial Output Data Pin Assignment Name Output termination Available Values OCT 85 , OCT 100 , OCT 120 , OCT 150 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-36 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Programmable Output Differential Voltage The Stratix IV GX and GT devices allow you to customize the differential output voltage to handle different trace lengths, various backplanes, and receiver requirements, as shown in Figure 1-32. You can change the VOD values using the dynamic reconfiguration controller. Set the VOD value through the tx_vodctrl[2:0] port of the dynamic reconfiguration controller. For example, to set VOD to a value of 3, set the tx_vodctrl[2:0] to 011. Figure 1-32. VOD (Differential) Signal Level Single-Ended Waveform VA +VOD VB Differential Waveform +700 +VOD 0-V Differential VOD (Differential) = VA - VB VOD (Differential) -VOD -700 f For more information about Stratix IV GX and GT VOD values, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Programmable Pre-Emphasis The programmable pre-emphasis module in each transmit buffer boosts high frequencies in the transmit data signal, which might be attenuated in the transmission media. Using pre-emphasis can maximize the data opening at the far-end receiver. The transmission line's transfer function can be represented in the frequency domain as a low-pass filter. Any frequency components below -3dB can pass through with minimal loss. Frequency components greater than -3dB are attenuated. This variation in frequency response yields data-dependant jitter and other inter-symbol interference (ISI) effects. By applying pre-emphasis, the high-frequency components are boosted; that is, pre-emphasized. Pre-emphasis equalizes the frequency response at the receiver so the difference between the low-frequency and high-frequency components is reduced, which minimizes the ISI effects from the transmission medium. Pre-emphasis requirements increase as data rates through legacy backplanes increase. You set the pre-emphasis settings in the ALTGX MegaWizard Plug-In Manager. The Stratix IV GX and GT transceivers provide three pre-emphasis taps--pre tap, first post tap, and second post tap. The ALTGX MegaWizard Plug-In Manager provides options to select the different values on these three taps. The pre tap sets the pre-emphasis on the data bit before the transition. The first post tap and second post tap set the pre-emphasis on the transition bit and the successive bit, respectively. The pre tap and second post tap also provide inversion control, shown by negative values on the corresponding tap settings in the ALTGX MegaWizard Plug-In Manager. The ALTGX MegaWizard Plug-In Manager only shows the valid pre-emphasis tap values for a selected VOD and transmitter termination resistance setting. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-37 Programmable Transmitter Output Buffer Power (VCCH) The ALTGX MegaWizard Plug-In Manager provides an option to select VCCH. Table 1-16 lists the data rates for the two VCCH options. Table 1-16. VCCH Option Data Rates VCCH Options Stratix IV GX Data Rate Stratix IV GT Data Rate 1.4 V 600 Mbps to 8.5 Gbps 600 Mbps to 11.3 Gbps 1.5 V 600 Mbps to 6.5 Gbps -- Link Coupling for Stratix IV GX and GT Devices A high-speed serial link can be AC-coupled or DC-coupled, depending on the serial protocol being implemented. AC-Coupled Links In an AC-coupled link, the AC-coupling capacitor blocks the transmitter DC VCM. The on-chip or off-chip receiver termination and biasing circuitry automatically restores the selected VCM. Figure 1-33 shows an AC-coupled link. Figure 1-33. AC-Coupled Link AC Coupling Capacitor Transmitter Receiver Physical Medium Physical Medium AC Coupling Capacitor RX Termination TX Termination TX VCM RX VCM The following protocols supported by Stratix IV GX and GT devices mandate AC-coupled links: PCIe Gigabit Ethernet Serial RapidIO XAUI SDI Stratix IV GT devices allow the high-speed links to be AC-coupled for the entire data rate range between 600 Mbps and 11.3 Gbps. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-38 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture DC-Coupled Links In a DC-coupled link, the transmitter DC VCM is seen unblocked at the receiver buffer. The link VCM depends on the transmitter VCM and the receiver VCM. The on-chip or off-chip receiver termination and biasing circuitry must ensure compatibility between the transmitter and receiver VCM. Figure 1-34 shows a DC-coupled link. Figure 1-34. DC-Coupled Link Transmitter Receiver Physical Medium Physical Medium TX Termination RX Termination TX VCM RX VCM The Stratix IV GX and GT transmitter can be DC-coupled to a Stratix IV GX and GT receiver for the entire operating data rate range of Stratix IV GX, from 600 Mbps to 8.5 Gbps. The Stratix IV GT transmitter can be DC-coupled to the Stratix IV GT receiver for the entire data rate range of 600 Mbps to 11.3 Gbps with Tx Vcm = 0.65 V and Rx Vcm = 0.82 V. For more information on the DC coupling capabilities of the Stratix IV GT device, refer to Table 1-23 on page 1-48. PCIe Receiver Detect The Stratix IV GX and GT transmitter buffers have a built-in receiver detection circuit for use in the PCIe mode for Gen1 and Gen2 data rates. This circuit detects if there is a receiver downstream by sending out a pulse on the common mode of the transmitter and monitoring the reflection. This mode requires the transmitter buffer to be tri-stated (in Electrical Idle mode), OCT utilization, and a 125 MHz fixedclk signal. You can enable this feature in PCIe mode by setting the tx_forceelecidle and tx_detectrxloopback ports to 1'b1. Receiver detect circuitry is active only in the P1 power state. f For more information about power states, refer to the PCIe 2.0 specification. In the P1 power state, the transmitter output buffer is tri-stated because the transmitter output buffer is in electrical idle. A high on the tx_detectrxloopback port triggers the receiver detect circuitry to alter the transmitter output buffer VCM. The sudden change in VCM effectively appears as a step voltage at the tri-stated transmitter buffer output. If a receiver (that complies with PCIe input impedance Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-39 requirements) is present at the far end, the time constant of the step voltage is higher. If a receiver is not present or is powered down, the time constant of the step voltage is lower. The receiver detect circuitry snoops the transmitter buffer output for the time constant of the step voltage to detect the presence of the receiver at the far end. A high pulse is driven on the pipephydonestatus port and 3'b011 is driven on the pipestatus port to indicate that a receiver has been detected. There is some latency after asserting the tx_detectrxloopback signal, before the receiver detection is indicated on the pipephydonestatus port. For signal timing to perform the receiver detect operation, refer to Figure 1-109 on page 1-134. 1 The tx_forceelecidle port must be asserted at least 10 parallel clock cycles prior to the tx_detectrxloopback port to ensure that the transmitter buffer is tri-stated. PCIe Electrical Idle The Stratix IV GX and GT transmitter output buffers support transmission of PCIe Electrical Idle (or individual transmitter tri-state). The tx_forceelecidle port puts the transmitter buffer in Electrical Idle mode. This port has a specific functionality in each power state. For the signal timing to perform the electrical idle transmission in PCIe mode, refer to Figure 1-108 on page 1-133. f For more information about using the tx_forceelecidle signal under different power states, refer to the PCIe specification 2.0. Transmitter Local Clock Divider Block Each transmitter channel contains a local clock divider block. It receives the high-speed clock from the CMU0 PLL or CMU1 PLL and generates the high-speed serial clock for the serializer and the low-speed parallel clock for the transmitter PCS datapath. The low-speed parallel clock is also forwarded to the FPGA fabric (tx_clkout). The local clock divider block allows each transmitter channel to run at /1, /2, or /4 of the CMU PLL data rate. The local clock divider block is used only in non-bonded functional modes (for example, GIGE, SONET/SDH, and SDI mode). Figure 1-35 shows the transmitter local clock divider block. Figure 1-35. Transmitter Local Clock Divider Block High-Speed Serial Clock CMU0 PLL High-Speed Clock /n CMU1 PLL High-Speed Clock September 2012 Altera Corporation / 4, 5, 8, or 10 Low-Speed Parallel Clock /1, 2, or 4 Stratix IV Device Handbook Volume 2: Transceivers 1-40 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Receiver Channel Datapath This section describes the Stratix IV GX and GT receiver channel datapath architecture. The sub-blocks in the receiver datapath are described in order from the serial receiver input buffer to the receiver phase compensation FIFO buffer at the FPGA fabric-transceiver interface. Figure 1-12 on page 1-17 shows the receiver channel datapath in Stratix IV GX and GT devices. The receiver channel PMA datapath consists of the following blocks: Receiver input buffer Clock and data recovery (CDR) unit Deserializer The receiver channel PCS datapath consists of the following blocks: Word aligner Deskew FIFO Rate match (clock rate compensation) FIFO 8B/10B decoder Byte deserializer Byte ordering Receiver phase compensation FIFO PCIe interface The receiver datapath is very flexible and allows multiple configurations, depending on the selected functional mode. You can configure the receiver datapath using the ALTGX MegaWizard Plug-In Manager. Receiver Input Buffer The Stratix IV GX and GT receiver input buffers are architecturally similar to each other. They both support programmable common mode voltage (Rx VCM), equalization, DC gain, and on-chip termination (OCT) settings. Table 1-17 lists the supported settings of the receiver input buffers in Stratix IV GX and GT devices. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-41 The receiver input buffer receives serial data from the rx_datain port and feeds it to the CDR unit. In the reverse serial loopback (pre-CDR) configuration, it also feeds the received serial data to the transmitter output buffer. Figure 1-36 shows the receiver input buffer. Figure 1-36. Receiver Input Buffer To the Transmitter Output Buffer in the Reverse Serial Loopback (Pre-CDR) Configuration Receiver Input Buffer Equalization and DC Gain Circuitry From Serial Data Input Pins (rx_datain) To CDR 85/100/ 120/150 RX Vcm 0.82/1.1-V Signal Threshold Detection Circuitry Signal Detect Table 1-17 lists the electrical features supported by the Stratix IV GX and GT receiver input buffer. Table 1-17. Electrical Features Supported by the Receiver Input Buffer for Stratix IV GX and GT Devices (1) I/O Standard Differential OCT with Calibration () VCM (V) Coupling Programmable DC Gain (dB) 1.4 V PCML 85, 100, 120, 150 0.82 AC, DC up to 16 Stratix IV GX 1.5 V PCML 85, 100, 120, 150 0.82 AC, DC up to 16 0.6 Gbps to 8.5 Gbps 2.5 V PCML 85, 100, 120, 150 0.82 AC up to 16 LVPECL 85, 100, 120, 150 0.82 AC up to 16 LVDS 85, 100, 120, 150 1.1 AC, DC up to 16 Stratix IV GT 1.4 V PCML 85, 100, 120, 150 0.82 AC, DC up to 16 0.6 Gbps to 11.3 Gbps LVDS 85, 100, 120, 150 1.1 AC, DC up to 16 Data Rate Note to Table 1-17: (1) Programmable equalization settings are 1 to 16 dB for Stratix IV GX and GT devices; for example, rx_eqctrl = 4'h0 maps to 1 dB gain, rx_eqctrl = 4'h1 maps to 2 dB gain, and so on. The Stratix IV GX and GT receiver buffers support the following features: September 2012 Programmable differential OCT Programmable VCM AC and DC coupling Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-42 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Programmable equalization and DC gain Signal threshold detection circuitry Programmable Differential On-Chip Termination The Stratix IV GX and GT receiver buffers support optional differential OCT resistors of 85, 100, 120, and 150 . To select the desired receiver OCT resistor, make the assignments shown in Table 1-18 in the Quartus II software Assignment Editor. Table 1-18. Stratix IV GX and GT Receiver On-Chip Termination Assignment Settings Assign To 1 rx_datain (Receiver Input Data Pins) Assignment Name Input Termination Stratix IV GX Available Values OCT 85 , OCT 100 , OCT 120 , OCT 150 , Off Stratix IV GT Available Values OCT 85 ,OCT 100 , OCT 120 , OCT 150 , Off The Stratix IV GX and GT receiver OCT resistors have calibration support to compensate for process, voltage, and temperature variations. For more information about OCT calibration support, refer to "Calibration Blocks" on page 1-201. Programmable VCM The Stratix IV GX and GT receiver buffers have on-chip biasing circuitry to establish the required VCM at the receiver input. It supports VCM settings of 0.82 V and 1.1 V that you can select in the ALTGX MegaWizard Plug-In Manager. You must select 0.82 V as the receiver buffer VCM for the following receiver input buffer I/O standards: 1.4-V PCML 1.5-V PCML 2.5-V PCML LVPECL You must select 1.1 V as the receiver buffer VCM for the LVDS receiver input buffer I/O standard. 1 On-chip biasing circuitry is effective only if you select on-chip receiver termination. If you select external termination, you must implement off-chip biasing circuitry to establish the VCM at the receiver input buffer. Link Coupling for Stratix IV GX Devices A high-speed serial link can either be AC-coupled or DC-coupled, depending on the serial protocol being implemented. Most of the serial protocols require links to be AC-coupled, but protocols such as Common Electrical I/O (CEI) optionally allow DC coupling. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-43 AC-Coupled Links In an AC-coupled link, the AC coupling capacitor blocks the transmitter DC VCM. The on-chip or off-chip receiver termination and biasing circuitry automatically restores the selected VCM. Figure 1-37 shows an AC-coupled link. Figure 1-37. AC-Coupled Link AC Coupling Capacitor Transmitter Receiver Physical Medium Physical Medium AC Coupling Capacitor RX Termination TX Termination TX VCM RX VCM Note to Figure 1-37: (1) The receiver termination and biasing can be on-chip or off-chip. The following protocols supported by Stratix IV GX and GT devices mandate AC-coupled links: September 2012 PCIe Gigabit Ethernet Serial RapidIO XAUI SDI Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-44 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture DC-Coupled Links In a DC-coupled link, the transmitter DC VCM is seen unblocked at the receiver buffer. Link VCM depends on the transmitter VCM and the receiver VCM. The on-chip or off-chip receiver termination and biasing circuitry must ensure compatibility between the transmitter and the receiver VCM. Figure 1-38 shows a DC-coupled link. Figure 1-38. DC-Coupled Link Transmitter Receiver Physical Medium Physical Medium TX Termination RX Termination TX VCM RX VCM Note to Figure 1-38: (1) The receiver termination and biasing can be on-chip or off-chip. You might choose to use the DC-coupled high-speed link for these functional modes only: Basic single- and double-width (OIF) CEI PHY interface The following sections describe DC-coupling requirements for a high-speed link with a Stratix IV GX device used as the transmitter, receiver, or both. Specifically, the following link configurations are described: Stratix IV Device Handbook Volume 2: Transceivers Stratix IV GX Transmitter (PCML) to Stratix IV GX Receiver (PCML) Stratix II GX Transmitter (PCML) to Stratix IV GX Receiver (PCML) Stratix IV GX Transmitter (PCML) to Stratix II GX Receiver (PCML) LVDS Transmitter to Stratix IV GX Receiver (PCML) September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-45 Figure 1-39 shows a typical Stratix IV GX transmitter (PCML) to Stratix IV GX Receiver (PCML) DC-coupled link. Figure 1-39. Stratix IV GX Transmitter (PCML) to Stratix IV GX Receiver (PCML) VCCH = 1.4 V/1.5 V Physical Medium Stratix IV GX Transmitter Stratix IV GX Receiver Physical Medium 42.5/50/60/75- TX Termination 42.5/50/60/75- TX Termination TX VCM 42.5/50/60/75- RX Termination 42.5/50/60/75- RX Termination Rs (1) 0.65 V RX VCM 0.82 V Note to Figure 1-39: (1) RS is the parasitic resistance present in the on-chip RX termination and biasing circuitry. Table 1-19 lists the allowed transmitter and receiver settings in a Stratix IV GX transmitter (PCML) to Stratix IV GX receiver (PCML) DC-coupled link. Table 1-19. Settings for a Stratix IV GX Transmitter (PCML) to Stratix IV GX Receiver (PCML) DC-Coupled Link Transmitter (Stratix IV GX) Settings Data Rate VCCH 600-8500 Mbps (1) 1.4 V/1.5 V Receiver (Stratix IV GX) Settings TX VCM Differential Termination Data Rate RX VCM Differential Termination 0.65 V 85/100/120/150 600-8500 Mbps 0.82 V 85/100/120/150 Note to Table 1-19: (1) VCCH = 1.5 V can support data rates from 600 Mbps to 6500 Mbps. VCCH = 1.4 V can support data rates from 600 Mbps to 8500 Mbps. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-46 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-40 shows the Stratix II GX transmitter (PCML) to Stratix IV GX receiver (PCML) coupled link. Figure 1-40. Stratix II GX Transmitter (PCML) to Stratix IV GX Receiver (PCML) VCCH = 1.2 V/1.5 V Physical Medium Stratix II GX Transmitter Stratix IV GX Receiver Physical Medium 50/60/75- TX Termination 50/60/75- TX Termination TX VCM 42.5/50/60/75- RX Termination 42.5/50/60/75- RX Termination Rs (1) 0.6 V/0.7 V RX VCM 0.82 V Note to Figure 1-40: (1) RS is the parasitic resistance present in the on-chip RX termination and biasing circuitry. Table 1-20 lists the allowed transmitter and receiver settings in a Stratix II GX to Stratix IV GX DC-coupled link. Table 1-20. Settings for a Stratix II GX to Stratix IV GX DC-Coupled Link Transmitter (Stratix II GX) Settings Data Rate 600-6375 Mbps VCCH (1) 1.5 V (1.5 V PCML) TX VCM (1) 0.6 V/0.7 V Receiver (Stratix IV GX) Settings Differential Termination Data Rate RX VCM Differential Termination 100/120/150 600-6375 Mbps 0.82 V 100/120/150 Note to Table 1-20: (1) VCCH = 1.5 V with TX VCM = 0.7 V can support data rates from 600 Mbps to 3125 Mbps. VCCH = 1.5 V with TX VCM = 0.6 V can support data rates from 600 Mbps to 6375 Mbps. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-47 Figure 1-41 shows the Stratix IV GX transmitter (PCML) to Stratix II GX receiver (PCML) DC-coupled link. Figure 1-41. Stratix IV GX Transmitter (PCML) to Stratix II GX Receiver (PCML) VCCH = 1.4/1.5 V Physical Medium Stratix IV GX Transmitter Stratix II GX Receiver Physical Medium 42.5/50/60/75- TX Termination 42.5/50/60/75- TX Termination TX VCM 50/60/75- RX Termination 50/60/75- RX Termination Rs (1) 0.65 V RX VCM 0.85 V Note to Figure 1-41: (1) RS is the parasitic resistance present in the on-chip RX termination and biasing circuitry. Table 1-21 lists the allowed transmitter and receiver settings in a Stratix IV GX transmitter (PCML) to Stratix II GX receiver (PCML) DC-coupled link. Table 1-21. Settings for a Stratix IV GX to Stratix II GX DC-Coupled Link Transmitter (Stratix IV GX) Settings Data Rate VCCH 600-6375 Mbps (1) 1.4/1.5 V Receiver (Stratix II GX) Settings TX VCM Differential Termination Data Rate I/O Standard RX VCM Differential Termination 0.65 V 100/120/150 600-6375 Mbps 1.4/1.5 V PCML 0.85 V 100/120/150 Note to Table 1-21: (1) VCCH = 1.5 V can support data rates from 600 Mbps to 6500 Mbps. VCCH = 1.4 V can support data rates from 600 Mbps to 6375 Mbps. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-48 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-42 shows the LVDS transmitter to Stratix IV GX receiver (PCML) DC-coupled link. Figure 1-42. LVDS Transmitter to Stratix IV GX Receiver (PCML) Physical Medium LVDS Transmitter Stratix IV GX Receiver Physical Medium 50- RX Termination 50- RX Termination Rs RX VCM (1) 1.1 V Note to Figure 1-42: (1) RS is the parasitic resistance present in the on-chip RX termination and biasing circuitry. Table 1-22 lists the allowed transmitter and receiver settings in a LVDS transmitter to Stratix IV GX receiver DC-coupled link. Table 1-22. Settings for a LVDS transmitter to Stratix IV GX Receiver DC-Coupled Link (1) Receiver (Stratix IV GX) Settings RX VCM Differential Termination RS 1.1 V 100 (2) Notes to Table 1-22: (1) When DC-coupling an LVDS transmitter to the Stratix IV GX receiver, use RX VCM = 1.1 V and series resistance value RS to verify compliance with the LVDS specification. (2) Pending characterization. Link Coupling for Stratix IV GT Devices Stratix IV GT devices allow the high-speed links to be AC- or DC-coupled links (AC-coupling allowed for the entire data rate range between 600 Mbps and 11.3 Gbps). Table 1-23 lists the allowed DC-coupling scenarios for Stratix IV GT devices. Table 1-23. Allowed DC-Coupling Scenarios for Stratix IV GT Devices (Part 1 of 2) From (Transmitter I/O Standard) To (Receiver I/O Standard) Data Rate Range Stratix IV GT Transmitter (1.4-V PCML) Stratix IV GT Receiver (1.4-V PCML) 600 Mbps to 11.3 Gbps Stratix IV GX Transmitter (1.4-V PCML) Stratix IV GT Receiver (1.4-V PCML) 600 Mbps to 8.5 Gbps Stratix IV Device Handbook Volume 2: Transceivers Conditions TX VCM = 0.65 V RX VCM = 0.82 V TX VCM = 0.65 V RX VCM = 0.82 V September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-49 Table 1-23. Allowed DC-Coupling Scenarios for Stratix IV GT Devices (Part 2 of 2) From (Transmitter I/O Standard) Stratix II GX Transmitter (1.5-V PCML) To (Receiver I/O Standard) Stratix IV GT Receiver (1.4-V PCML) Third-Party LVDS Transmitter Stratix IV GT Receiver (LVDS) Data Rate Range 600 Mbps to 6.375 Gbps 600 Mbps to 6.5 Gbps Conditions TX VCM = 0.7 V (600 Mbps to 3.125 Gbps) TX VCM = 0.6 V (3.125 Gbps to 6.375 Gbps) RX VCM = 0.82 V RX VCM = 1.1 V Programmable Equalization and DC Gain The transfer function of the physical medium can be represented as a low-pass filter in the frequency domain. Frequency components below -3 dB frequency pass through with minimal loss. Frequency components greater than -3 dB frequency are attenuated as a function of frequency due to skin-effect and dielectric losses. This variation in frequency response yields data-dependent jitter and other ISI effects, which can cause incorrect sampling of the input data. Each Stratix IV GX and GT receiver buffer has independently programmable equalization circuitry that boosts the high-frequency gain of the incoming signal, thereby compensating for the low-pass filter effects of the physical medium. The amount of high-frequency gain required depends on the loss characteristics of the physical medium. Stratix IV GX and GT equalization circuitry supports 16 equalization settings that provide up to 16 dB of high-frequency boost. You can select the appropriate equalization setting in the ALTGX MegaWizard Plug-In Manager. Stratix IV GX and GT receiver buffers also support programmable DC gain circuitry. Unlike equalization circuitry, DC gain circuitry provides equal boost to the incoming signal across the frequency spectrum. The receiver buffer supports DC gain settings of 0, 3, 6, 9, and 12 dB. You can select the appropriate DC gain setting in the ALTGX MegaWizard Plug-In Manager. Signal Threshold Detection Circuitry In PCIe mode, you can enable the optional signal threshold detection circuitry by not selecting the Force signal detection option in the ALTGX MegaWizard Plug-In Manager. If enabled, this option senses whether the signal level present at the receiver input buffer is above the signal detect threshold voltage that you specified in the What is the signal detect and signal loss threshold? option in the ALTGX MegaWizard Plug-In Manager. 1 The appropriate signal detect threshold level that complies with the PCIe compliance parameter VRX-IDLE-DETDIFFp-p is available in the DC and Switching Characteristics for Stratix IV Devices chapter. Signal threshold detection circuitry has a hysteresis response that filters out any high-frequency ringing caused by inter-symbol interference or high-frequency losses in the transmission medium. If the signal threshold detection circuitry senses the signal level present at the receiver input buffer to be higher than the signal detect threshold, it asserts the rx_signaldetect signal high. Otherwise, the signal threshold detection circuitry de-asserts the rx_signaldetect signal low. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-50 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Adaptive Equalization (AEQ) Stratix IV GX and GT receivers offer an Adaptive Equalization feature that automatically compensates for losses on the receiver channels. High-speed interface systems are used at different data rates with multiple backplane environments. These systems require different equalization settings to compensate for changing data rates and back plane characteristics. Manually selecting optimal equalization settings is cumbersome under these changing system characteristics. The Adaptive Equalization feature solves this problem by enabling the Stratix IV device to continuously tune the receiver equalization settings based on the frequency content of the incoming signal and comparing it with internally generated reference signals. Without this feature, you would have to tune the receiver channel's equalization stages manually, finding the optimal settings through trial and error, then locking in those values at compile time. The AEQ block resides within the PMA of the receiver channel and is available on the four regular channels of a transceiver block. To use AEQ, you must first enable the AEQ hardware in the ALTGX MegaWizard Plug-In Manager and the AEQ control block in the ALTGX_RECONFIG MegaWizard Plug-In Manager. To enable the AEQ feature, in ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers, select the Enable adaptive equalizer control option. When you select AEQ, two ports, aeq_fromgxb[] and aeq_togxb[], become available on the ALTGX and ALTGX_RECONFIG instances. These ports provide an interface between the PMA of the receiver channel and the AEQ control block in the ALTGX_RECONFIG MegaWizard Plug-In Manager. 1 Stratix IV Device Handbook Volume 2: Transceivers AEQ hardware is not present in the CMU channels. September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-51 Figure 1-43 shows the receiver channel data path with the AEQ feature. Figure 1-43. Receiver Channel Data Path Showing AEQ ALTGX Instance 1 Receiver Channel 0 aeq_togxb[23:0] logical_channel_address = 0 aeq_fromgxb[7:0] rx_datain[0] AEQ Hardware ALTGX_RECONFIG busy User Logic AEQ Control Block error ALTGX Instance 2 Receiver Channel 1 logical_channel_address = 4 aeq_fromgxb[15:8] rx_datain[1] AEQ Hardware reconfig_mode_sel[3:0] aeq_togxb[47:24] Modes of Operation of the AEQ Depending on the value you set for reconfig_mode_sel[3:0], the AEQ has three modes of operation: Continuous mode--This feature is not supported. One-time mode--The AEQ finds a stable setting of the receiver equalizer and locks that value. After it is locked, the equalizer values are no longer changed. This mode is available in one channel or all channels of the receiver. The reconfig_mode_sel[3:0] = 1001 in this mode. Powerdown mode--This feature is not supported. f For more information about the AEQ port connections and various waveforms in all the above modes, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. EyeQ The EyeQ hardware is available in Stratix IV GX and GT transceivers to analyze the receiver data recovery path, including receiver gain, clock jitter and noise level. You can use EyeQ to monitor the width of the incoming data eye and assess the quality of the incoming signal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-52 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Normally, the receiver CDR samples the incoming signal at the center of the eye. When you enable the EyeQ hardware, it allows the CDR to sample across 32 different positions within one unit interval (UI) of a data eye. You can manually control the sampling points and check the bit-error rate (BER) at each of these 32 sampling points. At the center of the eye, the BER is 0. As the sampling point is moved away from the center of the eye towards an edge, the BER increases. By observing sampling points with 0 BER and sampling points with higher BER, you can determine the eye width. 1 The EyeQ hardware is available for both regular transceiver channels and CMU channels. The EyeQ block resides within the PMA of the receiver channel and is available for both the transceiver channels and CMU channels of a transceiver block. Figure 1-44 shows the EyeQ feature within a receiver channel datapath. 1 You must implement logic to check the bit error rate (BER). This includes a pattern generator and checker. Figure 1-44 shows the receiver channel data path using the EyeQ feature. Figure 1-44. Receiver Channel Data Path showing the EyeQ Feature ctrl_writedata[15:0] ctrl_address[15:0] ctrl_write ctrl_read ALTGX Instance ALTGX_RECONFIG Instance reconfig_fromgxb[17:0] rx_datain[0] EyeQ Hardware EyeQ Control Block reconfig_togxb[3:0] Receiver Channel 0 busy error ctrl_waitrequest reconfig_mode_sel[3:0] ctrl_readdata[15:0] f For more information about using the EyeQ feature, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-53 Clock and Data Recovery Unit Each Stratix IV GX and GT receiver channel has an independent CDR unit to recover the clock from the incoming serial data stream. The high-speed and low-speed recovered clocks are used to clock the receiver PMA and PCS blocks. Figure 1-45 shows the CDR block diagram. Figure 1-45. Clock and Data Recovery Unit (1) Clock and Data Recovery (CDR) Unit rx_locktorefclk rx_locktodata signal detect rx_freqlocked LTR/LTD Controller Recovered Clock /2 Phase Down Detector Up (PD) rx_datain Charge Pump + Loop Filter rx_cruclk /1, /2, /4 /2 Phase Frequency Detector (PFD) VCO /L rx_pll_locked Up Down /M Note to Figure 1-45: (1) The blue colored path is active in lock-to-reference mode; the red colored path is active in lock-to-data mode. The CDR operates either in LTR mode or LTD mode. In LTR mode, the CDR tracks the input reference clock. In LTD mode, the CDR tracks the incoming serial data. After the receiver power up and reset cycle, the CDR must be kept in LTR mode until it locks to the input reference clock. After it is locked to the input reference clock, the CDR output clock is trained to the configured data rate. The CDR can now switch to LTD mode to recover the clock from incoming data. The LTR/LTD controller controls the switch between LTR and LTD modes. Lock-to-Reference (LTR) Mode In LTR mode, the phase frequency detector in the CDR tracks the receiver input reference clock, rx_cruclk. The PFD controls the charge pump that tunes the VCO in the CDR. Depending on the data rate and the selected input reference clock frequency, the Quartus II software automatically selects the appropriate /M and /L divider values such that the CDR output clock frequency is half the data rate. An active high, the rx_pll_locked status signal is asserted to indicate that the CDR has locked to the phase and frequency of the receiver input reference clock. Figure 1-45 on page 1-53 shows the active blocks (in blue) when the CDR is in LTR mode. 1 September 2012 The phase detector (PD) is inactive in LTR mode. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-54 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture You can drive the receiver input reference clock with the following clock sources: Dedicated REFCLK pins (refclk0 and refclk1) of the associated transceiver block Inter-transceiver block (ITB) clock lines from other transceiver blocks on the same side of the device (up to six ITB clock lines, two from each transceiver block) Global PLD clock driven by a dedicated clock input pin Clock output from the left and right PLLs in the FPGA fabric Table 1-24 lists CDR /M and/L divider values. Table 1-24. CDR Divider Values Parameter Value /M Divider 4, 5, 8, 10, 16, 20, 25 /L Divider 1, 2, 4, 8 Note to Table 1-24: (1) The maximum reference clock frequency of 672 MHz is only applicable to speed grades -2 and -3. For speed grade -4, the maximum reference clock frequency is 637.5 MHz. For input reference clock frequencies greater than 325 MHz, the Quartus II software automatically selects the appropriate /1, /2, or /4 pre-divider to meet the PFD input frequency limitation of 325 MHz. Lock-to-Data (LTD) Mode The CDR must be in LTD mode to recover the clock from the incoming serial data during normal operation. In LTD mode, the phase detector (PD) in the CDR tracks the incoming serial data at the receiver buffer. Depending on the phase difference between the incoming data and the CDR output clock, the PD controls the CDR charge pump that tunes the VCO. Figure 1-45 on page 1-53 shows the active blocks (in red) when the CDR is in LTD mode. 1 The PFD is inactive in LTD mode. The rx_pll_locked signal toggles randomly and has no significance in LTD mode. After switching to LTD mode, it can take a maximum of 1 ms for the CDR to get locked to the incoming data and produce a stable recovered clock. The actual lock time depends on the transition density of the incoming data and the PPM difference between the receiver input reference clock and the upstream transmitter reference clock. The receiver PCS logic must be held in reset until the CDR produces a stable recovered clock. f For more information about receiver reset recommendations, refer to the Reset Control and Power Down in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-55 PCIe Clock Switch Circuitry The feedback path from the CDR VCO to the PD has a /2 divider that is used in PCIe mode configured at Gen2 (5 Gbps) data rate for the dynamic switch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rates. When the PHY-MAC layer instructs a Gen2-to-Gen1 signaling rateswitch, the /2 divider is enabled. When the PHY-MAC layer instructs a Gen1-to-Gen2 signaling rateswitch, the /2 divider is disabled. For more information about the PCIe signaling rateswitch, refer to "Dynamic Switch Between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) Signaling Rate" on page 1-141. 1 The /2 divider in the receiver CDR between the VCO and the PD is disabled in all other functional modes. LTR/LTD Controller The LTR/LTD controller controls whether the CDR is in LTR or LTD mode. You can configure the LTR/LTD controller either in automatic lock mode or manual lock mode. Two optional input ports (rx_locktorefclk and rx_locktodata) allow you to configure the LTR/LTD controller in either automatic lock mode or manual lock mode. Table 1-25 lists the relationship between these optional input ports and the LTR/LTD controller lock mode. Table 1-25. Optional Input Ports and LTR/LTD Controller Lock Mode 1 rx_locktorefclk rx_locktodata LTR/LTD Controller Lock Mode 1 0 Manual - LTR Mode x 1 Manual - LTD Mode 0 0 Automatic Lock Mode If you do not instantiate the optional rx_locktorefclk and rx_locktodata signals, the Quartus II software automatically configures the LTR/LTD controller in automatic lock mode. Automatic Lock Mode In automatic lock mode, the LTR/LTD controller initially sets the CDR to lock to the input reference clock (LTR mode). After the CDR locks to the input reference clock, the LTR/LTD controller automatically sets it to lock to the incoming serial data (LTD mode) when the following three conditions are met: Signal threshold detection circuitry indicates the presence of valid signal levels at the receiver input buffer Valid for PCIe mode only. This condition is defaulted to true for all other modes. The CDR output clock is within the configured PPM frequency threshold setting with respect to the input reference clock (frequency locked) The CDR output clock and the input reference clock are phase matched within approximately 0.08 UI (phase locked) The switch from LTR to LTD mode is indicated by the assertion of the rx_freqlocked signal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-56 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture In LTD mode, the CDR uses a phase detector to keep the recovered clock phase-matched to the data. If the CDR does not stay locked to data due to frequency drift or severe amplitude attenuation, the LTR/LTD controller switches the CDR back to LTR mode to lock to the input reference clock. In automatic lock mode, the LTR/LTD controller switches the CDR from LTD to LTR mode when the following conditions are met: Signal threshold detection circuitry indicates the absence of valid signal levels at the receiver input buffer Valid for PCIe mode only. This condition is defaulted to true for all other modes. The CDR output clock is not within the configured PPM frequency threshold setting with respect to the input reference clock The switch from LTD to LTR mode is indicated by the de-assertion of the rx_freqlocked signal. Manual Lock Mode In automatic lock mode, the LTR/LTD controller relies on the PPM detector and the phase relationship detector to set the CDR in LTR or LTD mode. The PPM detector and phase relationship detector reaction times can be too long for some applications that require faster CDR lock time. You can manually control the CDR to reduce its lock time using the rx_locktorefclk and rx_locktodata ports. In manual lock mode, the LTR/LTD controller sets the CDR in LTR or LTD mode depending on the logic level on the rx_locktorefclk and rx_locktodata signals. When the rx_locktorefclk signal is asserted high, the LTR/LTD controller forces the CDR to lock to the reference clock. When the rx_locktodata signal is asserted high, it forces the CDR to lock to data. When both signals are asserted, the rx_locktodata signal takes precedence over the rx_locktorefclk signal, forcing the CDR to lock to data. When the rx_locktorefclk signal is asserted high, the rx_freqlocked signal does not have any significance and is always driven low, indicating that the CDR is in LTR mode. When the rx_locktodata signal is asserted high, the rx_freqlocked signal is always driven high, indicating that the CDR is in LTD mode. If both signals are deasserted, the CDR is in automatic lock mode. 1 The Altera-recommended transceiver reset sequence varies depending on the CDR lock mode. f For more information about reset sequence recommendations, refer to the Reset Control and Power Down in Stratix IV Devices chapter. Offset Cancellation in the Receiver Buffer and Receiver CDR As silicon progresses towards smaller process nodes, the performance of circuits at these smaller nodes depends more on process variations. These process variations result in analog voltages that can be offset from the required ranges. Offset cancellation logic corrects these offsets. The receiver buffer and receiver CDR require offset cancellation. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-57 Offset cancellation is executed automatically once each time a Stratix IV GX and GT device is powered on (after the device has finished programming and switches to user mode as indicated by CONFIG_DONE=1). The control logic for offset cancellation is integrated into the ALTGX_RECONFIG megafunction. The reconfig_fromgxb and reconfig_togxb buses and the necessary clocks must be connected between the ALTGX instance and the ALTGX_RECONFIG instance. 1 You must reprogram your device to restart the Offset Cancellation process. f For more information about offset cancellation control logic connectivity, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. 1 During offset cancellation, signified by a high on the busy signal, rx_analogreset is not relevant until the busy signal goes low. Offset cancellation logic requires a separate clock. In PCIe mode, you must connect the clock input to the fixedclk port provided by the ALTGX MegaWizard Plug-In Manager. The frequency of this clock input must be 125 MHz. For all other functional modes, connect the clock input to the reconfig_clk port provided by the ALTGX MegaWizard Plug-In Manager. The frequency of the clock connected to the reconfig_clk port must be within the range of 37.5 to 50 MHz. Figure 1-46 shows the interface of the offset cancellation control logic (ALTGX_RECONFIG instance) and the ALTGX instance. Figure 1-46. Interface of Offset Cancellation Control Logic to the ALTGX Instance ALTGX_RECONFIG Instance ALTGX Instance with 4 Channels Transceiver Block TX Dynamic Re-config logic reconfig_togxb busy Buffer reconfig_clk Offset Cancellation Logic RX CDR reconfig_fromgxb TX Buffer RX CDR TX Buffer RX CDR TX Buffer RX CDR reconfig_clk [ The offset cancellation process begins by disconnecting the path from the receiver input buffer to the receiver CDR. It then sets the receiver CDR into a fixed set of dividers to guarantee a VCO clock rate that is within the range necessary to provide proper offset cancellation. Subsequently, the offset cancellation process goes through various states and culminates in the offset cancellation of the receiver buffer and the receiver CDR. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-58 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture After offset cancellation is complete, the divider settings are restored. Then the reconfiguration block sends and receives data to the ALTGX instance using the reconfig_togxb and reconfig_fromgxb buses. Connect the buses between the ALTGX_RECONFIG and ALTGX instances. The de-assertion of the busy signal from the offset cancellation control logic indicates the offset cancellation process is complete. f Due to the offset cancellation process, the transceiver reset sequence has changed. For more information about the offset cancellation process, refer to the Reset Control and Power Down in Stratix IV Devices chapter. Deserializer The deserializer block clocks in serial input data from the receiver buffer using the high-speed serial recovered clock and deserializes it using the low-speed parallel recovered clock. It forwards the deserialized data to the receiver PCS channel. In single-width mode, the deserializer supports 8-bit and 10-bit deserialization factors. In double-width mode, the deserializer supports 16-bit and 20-bit deserialization factors. Figure 1-47 shows the deserializer operation in single-width mode with a 10-bit deserialization factor. Figure 1-47. Deserializer Operation in Single-Width Mode Received Data D9 D9 D8 D8 D7 D7 D6 D6 D5 D5 10 Clock Recovery Unit D4 D4 D3 D3 D2 D2 D1 D1 D0 D0 To Word Aligner High-Speed Serial Recovered Clock Low-Speed Parallel Recovered Clock Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-59 Figure 1-48 shows the serial bit order of the deserializer block input and the parallel data output of the deserializer block in single-width mode with a 10-bit deserialization factor. The serial stream (0101111100) is deserialized to a value 10'h17C. The serial data is assumed to be received LSB to MSB. Figure 1-48. Deserializer Bit Order in Single-Width Mode Low-Speed Parallel Clock High-Speed Serial Clock datain 0 0 1 1 1 1 1 0 dataout 1 0 1 1 0 0 0 0 0 1 0 1 0101111100 1010000011 Word Aligner Because the data is serialized before transmission and then deserialized at the receiver, it loses the word boundary of the upstream transmitter upon deserialization. The word aligner receives parallel data from the deserializer and restores the word boundary based on a pre-defined alignment pattern that must be received during link synchronization. Serial protocols such as PCIe, XAUI, Gigabit Ethernet, Serial RapidIO, and SONET/SDH, specify a standard word alignment pattern. For proprietary protocols, the Stratix IV GX and GT transceiver architecture allows you to select a custom word alignment pattern specific to your implementation. In addition to restoring the word boundary, the word aligner also implements the following features: Synchronization state machine in functional modes such as PCIe, XAUI, GIGE, Serial RapidIO, and Basic single-width Programmable run length violation detection in all functional modes Receiver polarity inversion in all functional modes except PCIe Receiver bit reversal in Basic single-width and Basic double-width modes Receiver byte reversal in Basic double-width modes Depending on the configured functional mode, the word aligner operates in one of the following three modes: September 2012 Manual alignment mode Automatic synchronization state machine mode Bit-slip mode Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-60 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-49 shows the word aligner operation in all supported configurations. Figure 1-49. Word Aligner in All Supported Configurations PMA-PCS Interface Width Single-Width Double-Width 10-Bit Wide 8-Bit Wide Manual Bit-Slip Alignment (Basic (OC-12, OC-48, Single-Width) Basic Single-Width) Manual Alignment (Basic Single-Width) 20-Bit Wide 16-Bit Wide Bit-Slip (Basic Single-Width, SDI) Automatic Synchronization State Machine (PCIe XAUI, GIGE, Basic Single-Width, Serial RapidIO) Manual Alignment (Basic Double-Width, OC-96) Bit-Slip (Basic Double-Width) Manual Alignment (Basic Double-Width) Bit-Slip (Basic Double-Width) Word Aligner in Single-Width Mode In single-width mode, the PMA-PCS interface is either 8 or 10 bits wide. In 8-bit wide PMA-PCS interface modes, the word aligner receives 8-bit wide data from the deserializer. In 10-bit wide PMA-PCS interface modes, the word aligner receives 10-bit wide data from the deserializer. Depending on the configured functional mode, you can configure the word aligner in manual alignment mode, automatic synchronization state machine mode, or bit-slip mode. Word Aligner in Single-Width Mode with 8-Bit PMA-PCS Interface Modes The following functional modes support the 8-bit PMA-PCS interface: SONET/SDH OC-12 SONET/SDH OC-48 Basic single-width Table 1-26 lists the word aligner configurations allowed in functional modes with an 8-bit PMA-PCS interface. Table 1-26. Word Aligner Configurations with an 8-Bit PMA-PCS Interface Allowed Word Configurations Allowed Word Alignment Pattern Length Manual Alignment 16 bits SONET/SDH OC-48 Manual Alignment 16 bits Basic single-width Manual Alignment, Bit-Slip 16 bits Functional Mode SONET/SDH OC-12 Manual Alignment Mode Word Aligner with 8-Bit PMA-PCS Interface Modes In manual alignment mode, the word aligner operation is controlled by the input signal rx_enapatternalign. The word aligner operation is edge-sensitive to the rx_enapatternalign signal. After de-assertion of rx_digitalreset, a rising edge on the rx_enapatternalign signal triggers the word aligner to look for the word alignment pattern in the received data stream. In SONET/SDH OC-12 and OC-48 modes, the word aligner looks for 16'hF628 (A1A2) or 32'hF6F62828 (A1A1A2A2), Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-61 depending on whether the input signal rx_a1a2size is driven low or high, respectively. In Basic single-width mode, the word aligner looks for the 16-bit word alignment pattern programmed in the ALTGX MegaWizard Plug-In Manager. The word aligner aligns the 8-bit word boundary to the first word alignment pattern received after the rising edge on the rx_enapatternalign signal. Two status signals, rx_syncstatus and rx_patterndetect, with the same latency as the datapath, are forwarded to the FPGA fabric to indicate word aligner status. On receiving the first word alignment pattern after the rising edge on the rx_enapatternalign signal, both the rx_syncstatus and rx_patterndetect signals are driven high for one parallel clock cycle synchronous to the MSByte of the word alignment pattern. Any word alignment pattern received thereafter in the same word boundary causes only the rx_patterndetect signal to go high for one clock cycle. 1 For the word aligner to re-synchronize to a new word boundary, you must de-assert rx_enapatternalign and re-assert it again to create a rising edge. After a rising edge on the rx_enapatternalign signal, if the word alignment pattern is found in a different word boundary, the word aligner re-synchronizes to the new word boundary and asserts the rx_syncstatus and rx_patterndetect signals for one parallel clock cycle. Figure 1-50 shows word aligner behavior in SONET/SDH OC-12 functional mode. The LSByte (8'hF6) and the MSByte (8'h28) of the 16-bit word alignment pattern are received in parallel clock cycles n and n + 1, respectively. The rx_syncstatus and rx_patterndetect signals are both driven high for one parallel clock cycle synchronous to the MSByte (8'h28) of the word alignment pattern. After initial word alignment, the 16-bit word alignment pattern is again received across the word boundary in clock cycles m, m + 1, and m + 2. The word aligner does not re-align to the new word boundary because of the lack of a preceding rising edge on the rx_enapatternalign signal. If you create a rising edge on the rx_enapatternalign signal before the word alignment pattern is received across clock cycles m, m + 1, and m + 2, the word aligner re-aligns to the new word boundary, causing both the rx_syncstatus and rx_patterndetect signals to go high for one parallel clock cycle. Figure 1-50. Bit-Slip Mode in 8-Bit PMA-PCS Interface Mode n rx_dataout[7:0] 11110110 F6 n+1 00101000 28 m m+1 m+2 0110xxxx 10001111 xxxx0010 8F x2 6x rx_enapatternalign rx_patterndetect rx_syncstatus September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-62 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Bit-Slip Mode Word Aligner with 8-Bit PMA-PCS Interface Modes Basic single-width mode with 8-bit PMA-PCS interface width allows the word aligner to be configured in bit-slip mode. The word aligner operation is controlled by the input signal rx_bitslip in bit-slip mode. At every rising edge of the rx_bitslip signal, the bit-slip circuitry slips one bit into the received data stream, effectively shifting the word boundary by one bit. In bit-slip mode, the word aligner status signal rx_patterndetect is driven high for one parallel clock cycle when the received data after bit-slipping matches the 16-bit word alignment pattern programmed in the ALTGX MegaWizard Plug-In Manager. You can implement a bit-slip controller in the FPGA fabric that monitors either the rx_dataout signal and/or the rx_patterndetect signal and controls the rx_bitslip signal to achieve word alignment. Figure 1-51 shows an example of the word aligner configured in bit-slip mode. For this example, consider that 8'b11110000 is received back-to-back and 16'b0000111100011110 is specified as the word alignment pattern. A rising edge on the rx_bitslip signal at time n + 1 slips a single bit 0 at the MSB position, forcing the rx_dataout to 8'b01111000. Another rising edge on the rx_bitslip signal at time n + 5 forces rx_dataout to 8'b00111100. Another rising edge on the rx_bitslip signal at time n + 9 forces rx_dataout to 8'b00011110. Another rising edge on the rx_bitslip signal at time n + 13 forces the rx_dataout to 8'b00001111. At this instance, rx_dataout in cycles n + 12 and n + 13 is 8'b00011110 and 8'b00001111, respectively, which matches the specified 16-bit alignment pattern 16'b0000111100011110. This results in the assertion of the rx_patterndetect signal. Figure 1-51. Word Aligner Configured in Bit-Slip Mode n n+1 n+2 n+3 n+4 n+5 n+6 n+7 n+8 n + 9 n + 10 n + 11 n + 12 n + 13 n + 14 rx_clkout 11110000 rx_datain rx_dataout[7:0] 11110000 01111000 00111100 00011110 00001111 rx_bitslip rx_patterndetect Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-63 Word Aligner in Single-Width Mode with 10-Bit PMA-PCS Interface Modes The following functional modes support the 10-bit PMA-PCS interface: PCIe Gen1 and Gen2 Serial RapidIO XAUI GIGE SDI Basic single-width mode This section describes the following word aligner 10-bit PMA-PCS interface modes: Automatic synchronization state machine mode with 10-bit PMA-PCS interface mode Manual alignment mode with 10-bit PMA-PCS interface mode Bit-slip mode with 10-bit PMA-PCS interface mode Table 1-27 lists the word aligner configurations allowed in functional modes with a 10-bit PMA-PCS interface. Table 1-27. Word Aligner Configurations with a 10-Bit PMA-PCS Interface Functional Mode PCIe Allowed Word Aligner Configurations Automatic synchronization state machine Allowed Word Alignment Pattern Length 10 bits Serial RapidIO Automatic synchronization state machine 10 bits XAUI Automatic synchronization state machine 7 bits, 10 bits GIGE Automatic synchronization state machine 7 bits, 10 bits SDI Bit-slip Basic single-width mode Manual alignment, Automatic synchronization state machine, Bit-slip N/A 7 bits, 10 bits Automatic Synchronization State Machine Mode Word Aligner with 10-Bit PMA-PCS Interface Mode Protocols such as PCIe, XAUI, Gigabit Ethernet, and Serial RapidIO require the receiver PCS logic to implement a synchronization state machine to provide hysteresis during link synchronization. Each of these protocols defines a specific number of synchronization code groups that the link must receive to acquire synchronization and a specific number of erroneous code groups that it must receive to fall out of synchronization. In PCIe, XAUI, Gigabit Ethernet, and Serial RapidIO functional modes, the Quartus II software configures the word aligner in automatic synchronization state machine mode. It automatically selects the word alignment pattern length and pattern as specified by each protocol. In each of these functional modes, the protocol-compliant synchronization state machine is implemented in the word aligner. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-64 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture In Basic single-width functional mode with a 10-bit PMA-PCS interface, you can configure the word aligner in automatic synchronization state machine mode by selecting the Use the built-in synchronization state machine option in the ALTGX MegaWizard Plug-In Manager. It also allows you to program a custom 7-bit or 10-bit word alignment pattern that the word aligner uses for synchronization. 1 The 10-bit input data to the word aligner configured in automatic synchronization state machine mode must be 8B/10B encoded. Table 1-28 lists the synchronization state machine parameters that the Quartus II software allows in supported functional modes. The synchronization state machine parameters are fixed for PCIe, XAUI, GIGE, and Serial RapidIO modes as specified by the respective protocol. For Basic single-width mode, you can program these parameters as suited to your proprietary protocol implementation. Table 1-28. Synchronization State Machine Functional Modes Functional Mode PCIe XAUI GIGE Serial RapidIO Basic Single-Width Mode Number of valid synchronization code groups or ordered sets received to achieve synchronization 4 4 3 127 1 to 256 Number of erroneous code groups received to lose synchronization 17 4 4 3 1 to 64 Number of continuous good code groups received to reduce the error count by one 16 4 4 255 1 to 256 After de-assertion of the rx_digitalreset signal in automatic synchronization state machine mode, the word aligner starts looking for the word alignment pattern or synchronization code groups in the received data stream. When the programmed number of valid synchronization code groups or ordered sets is received, the rx_syncstatus signal is driven high to indicate that synchronization is acquired. The rx_syncstatus signal is constantly driven high until the programmed number of erroneous code groups is received without receiving intermediate good groups; after which the rx_syncstatus is driven low. The word aligner indicates loss of synchronization (rx_syncstatus remains low) until the programmed number of valid synchronization code groups are received again. Manual Alignment Mode Word Aligner with 10-Bit PMA-PCS Interface Mode In Basic single-width mode with a 10-bit PMA-PCS interface, you can configure the word aligner in manual alignment mode by selecting the Use manual word alignment mode option in the ALTGX MegaWizard Plug-In Manager. In manual alignment mode, the word aligner operation is controlled by the input signal rx_enapatternalign. The word aligner operation is level-sensitive to the rx_enapatternalign signal. If the rx_enapatternalign signal is held high, the word aligner looks for the programmed 7-bit or 10-bit word alignment pattern in the received data stream. It updates the word boundary if it finds the word alignment pattern in a new word boundary. If the rx_enapatternalign signal is de-asserted low, the word aligner maintains the current word boundary even when it sees the word alignment pattern in a new word boundary. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-65 Two status signals, rx_syncstatus and rx_patterndetect, with the same latency as the datapath, are forwarded to the FPGA fabric to indicate the word aligner status. After receiving the first word alignment pattern after the rx_enapatternalign signal is asserted high, both the rx_syncstatus and rx_patterndetect signals are driven high for one parallel clock cycle. Any word alignment pattern received thereafter in the same word boundary causes only the rx_patterndetect signal to go high for one clock cycle. Any word alignment pattern received thereafter in a different word boundary causes the word aligner to re-align to the new word boundary only if the rx_enapatternalign signal is held high. The word aligner asserts the rx_syncstatus signal for one parallel clock cycle whenever it re-aligns to the new word boundary. Figure 1-52 shows the manual alignment mode word aligner operation with 10-bit PMA-PCS interface mode. In this example, a /K28.5/ (10'b0101111100) is specified as the word alignment pattern. The word aligner aligns to the /K28.5/ alignment pattern in cycle n because the rx_enapatternalign signal is asserted high. The rx_syncstatus signal goes high for one clock cycle, indicating alignment to a new word boundary. The rx_patterndetect signal also goes high for one clock cycle to indicate initial word alignment. At time n + 1, the rx_enapatternalign signal is de-asserted to instruct the word aligner to lock the current word boundary. The alignment pattern is detected again in a new word boundary across cycles n + 2 and n + 3. The word aligner does not align to this new word boundary because the rx_enapatternalign signal is held low. The /K28.5/ word alignment pattern is detected again in the current word boundary during cycle n + 5, causing the rx_patterndetect signal to go high for one parallel clock cycle. Figure 1-52. Word Aligner with 10-Bit PMA-PCS Manual Alignment Mode n n+1 n+2 n+3 n+4 n+5 1111001010 1000000101 111110000 0101111100 rx_clkout rx_dataout[10..0] 111110000 0101111100 111110000 rx_enapatternalign rx_patterndetect rx_syncstatus 1 September 2012 If the word alignment pattern is known to be unique and does not appear between word boundaries, you can constantly hold the rx_enapatternalign signal high because there is no possibility of false word alignment. If there is a possibility of the word alignment pattern occurring across word boundaries, you must control the rx_enapatternalign signal to lock the word boundary after the desired word alignment is achieved to avoid re-alignment to an incorrect word boundary. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-66 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Bit-Slip Mode Word Aligner with 10-Bit PMA-PCS Interface Mode In some Basic single-width configurations with a 10-bit PMA-PCS interface, you can configure the word aligner in bit-slip mode by selecting the Use manual bit slipping mode option in the ALTGX MegaWizard Plug-In Manager. The word aligner operation for Basic single-width with a 10-bit PMA-PCS interface is similar to the word aligner operation in Basic single-width mode with an 8-bit PMA-PCS interface. For word aligner operation in bit-slip mode, refer to "Manual Alignment Mode Word Aligner with 8-Bit PMA-PCS Interface Modes" on page 1-60. The only difference is that the bit-slip word aligner with 10-bit PMA-PCS interface modes allows 7-bit and 10-bit word alignment patterns, whereas the one with 8-bit PMA-PCS interface modes allows only 16-bit word alignment patterns. Word Aligner in Double-Width Mode In double-width mode, the PMA-PCS interface is either 16 or 20 bits wide. In 16-bit PMA-PCS interface modes, the word aligner receives 16 bit wide data from the deserializer. In 20-bit PMA-PCS interface modes, the word aligner receives 10-bit wide data from the deserializer. Depending on the configured functional mode, you can configure the word aligner in manual alignment mode or bit-slip mode. Automatic synchronization state machine mode is not supported for word aligner in double-width mode. Word Aligner in Double-Width Mode with 16-Bit PMA-PCS Interface Modes The following functional modes support the 16-bit PMA-PCS interface: SONET/SDH OC-96 (OIF) CEI PHY interface Basic double-width Table 1-29 lists the word aligner configurations allowed in functional modes with a 16-bit PMA-PCS interface. Table 1-29. Word Aligner Configurations with 16-Bit PMA-PCS Interface (1) Allowed Word Aligner Configurations Allowed Word Alignment Pattern Length SONET/SDH OC-96 Manual alignment 16 bits, 32 bits Basic double-width Manual alignment, Bit-slip 8 bits, 16 bits, 32 bits Functional Mode Note to Table 1-29: (1) The word aligner is bypassed in (OIF) CEI PHY interface mode. Manual Alignment Mode Word Aligner with 16-Bit PMA-PCS Interface Modes In manual alignment mode, the word aligner starts looking for the programmed 8-bit, 16-bit, or 32-bit word alignment pattern in the received data stream as soon as rx_digitalreset is de-asserted low. It aligns to the first word alignment pattern received regardless of the logic level driven on the rx_enapatternalign signal. Any word alignment pattern received thereafter in a different word boundary does not cause the word aligner to re-align to this new word boundary. After the initial word alignment following de-assertion of the rx_digitalreset signal, if a word re-alignment is required, you must use the rx_enapatternalign signal. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-67 Word aligner operation is controlled by the input signal rx_enapatternalign and is edge-sensitive to the rx_enapatternalign signal. A rising edge on the rx_enapatternalign signal triggers the word aligner to look for the word alignment pattern in the received data stream. The word aligner aligns the 16-bit word boundary to the first word alignment pattern received after the rising edge on the rx_enapatternalign signal. Any word alignment pattern received thereafter in a different word boundary does not cause the word aligner to re-align to this new word boundary. If another word re-alignment is required, you must de-assert and re-assert the rx_enapatternalign signal to create a rising edge on this signal. Two status signals, rx_syncstatus and rx_patterndetect, with the same latency as the datapath, are forwarded to the FPGA fabric to indicate word aligner status. After receiving the first word alignment pattern, the rx_patterndetect signal is driven high for one parallel clock cycle synchronous to the data that matches the MSByte of the word alignment pattern. Any word alignment pattern received thereafter in the same word boundary causes rx_patterndetect to go high for one parallel clock cycle. After receiving the first word alignment pattern, the rx_syncstatus signal is constantly driven high until the word aligner sees another rising edge on the rx_enapatternalign signal. The rising edge on the rx_enapatternalign signal re-triggers the word alignment operation. Figure 1-53 shows the manual alignment mode word aligner operation in 16-bit PMA-PCS interface mode. In this example, a 16'hF628 is specified as the word alignment pattern. The word aligner aligns to the 16'hF628 pattern received in cycle n after de-assertion of rx_digitalreset. The rx_patterndetect[1] signal is driven high for one parallel clock cycle. The rx_syncstatus[1] signal is driven high constantly until cycle n + 2, after which it is driven low because of the rising edge on the rx_enapatternalign signal that re-triggers the word aligner operation. The word aligner receives the word alignment pattern again in cycle n + 4, causing the rx_patterndetect[1] signal to be driven high for one parallel clock cycle and the rx_syncstatus[1] signal to be driven high constantly. Figure 1-53. Manual Alignment Mode Word Aligner in 16-Bit PMA-PCS Interface Modes n n+1 n+2 F628 xxxx xxxx 00 10 11 00 10 xxxx rx_dataout n+3 xxxx n+4 F628 xxxx xxxx rx_digitalreset rx_enapatternalign rx_syncstatus[1:0] rx_patterndetect[1:0] September 2012 Altera Corporation 00 00 10 11 10 00 Stratix IV Device Handbook Volume 2: Transceivers 1-68 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Bit-Slip Mode Word Aligner with 16-Bit PMA-PCS Interface Modes In some Basic double-width configurations with 16-bit PMA-PCS interface, you can configure the word aligner in bit-slip mode by selecting the Use manual bit slipping mode option in the ALTGX MegaWizard Plug-In Manager. The word aligner operation for Basic double-width with 16-bit PMA-PCS interface is similar to the word aligner operation in Basic single-width mode with 8-bit PMA-PCS interface. For word aligner operation in bit-slip mode, refer to "Word Aligner in Single-Width Mode with 8-Bit PMA-PCS Interface Modes" on page 1-60. The only difference is that the bit-slip word aligner in 16-bit PMA-PCS interface modes allows 8-bit and 16-bit word alignment patterns, whereas the bit-slip word aligner in 8-bit PMA-PCS interface modes allows only a 16-bit word alignment pattern. Word Aligner in Double-Width Mode with 20-Bit PMA-PCS Interface Modes A 20-bit PMA-PCS interface is supported only in Basic double-width mode. Table 1-30 lists the word aligner configurations allowed in functional modes with a 20-bit PMA-PCS interface. Table 1-30. Word Aligner in 20-Bit PMA-PCS Interface Modes Functional Mode Basic double-width Allowed Word Aligner Configurations Allowed Word Alignment Pattern Length Manual alignment, Bit-slip 7 bits, 10 bits, 20 bits Manual Alignment Mode Word Aligner with 20-Bit PMA-PCS Interface Modes The word aligner operation in Basic double-width mode with 20-bit PMA-PCS interface is similar to the word aligner operation in Basic double-width mode with a 16-bit PMA-PCS interface. For word aligner operation in manual alignment mode, refer to "Word Aligner in Double-Width Mode with 16-Bit PMA-PCS Interface Modes" on page 1-66. The only difference is that the manual alignment mode word aligner in 20-bit PMA-PCS interface modes allows 7-, 10-, and 20-bit word alignment patterns, whereas the manual alignment mode word aligner in 16-bit PMA-PCS interface modes allows only 8-, 16-, and 32-bit word alignment patterns. Bit-Slip Mode Word Aligner with 20-Bit PMA-PCS Interface Modes In some Basic single-width configurations with 20-bit PMA-PCS interface, you can configure the word aligner in bit-slip mode by selecting the Use manual bit slipping mode option in the ALTGX MegaWizard Plug-In Manager. The word aligner operation for Basic double-width with 20-bit PMA-PCS interface is similar to the word aligner operation in Basic single-width mode with an 8-bit PMA-PCS interface. For word aligner operation in bit-slip mode, refer to "Word Aligner in Single-Width Mode with 8-Bit PMA-PCS Interface Modes" on page 1-60. The difference is that the bit-slip word aligner in 20-bit PMA-PCS interface modes allows only 7-, 10-, and 20-bit word alignment patterns, whereas the bit-slip word aligner in 8-bit PMA-PCS interface modes allows only a 16-bit word alignment pattern. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-69 Table 1-31 lists the word aligner options available in Basic single-width and double-width modes. Table 1-31. Word Aligner Options Available in Basic Single-Width and Double-Width Modes Functional Mode PMA-PCS Interface Width Word Alignment Mode Manual Alignment Word Alignment Pattern Length 16-bit (1) (Part 1 of 2) rx_enapatternalign Sensitivity rx_syncstatus Behavior rx_patterndetect Behavior Rising Edge Sensitive Asserted high for one parallel clock cycle when the word aligner aligns to a new word boundary. Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. N/A N/A Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. Level Sensitive Asserted high for one parallel clock cycle when the word aligner aligns to a new word boundary. Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. N/A Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. 8-bit Bit-Slip Basic Single-Width Manual Alignment 10-bit Bit-Slip Automatic Synchronization State Machine September 2012 Altera Corporation 16-bit 7- and 10-bit 7- and 10-bit 7- and 10-bit N/A N/A Stays high as long as the synchronization conditions are satisfied. Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. Stratix IV Device Handbook Volume 2: Transceivers 1-70 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-31. Word Aligner Options Available in Basic Single-Width and Double-Width Modes Functional Mode PMA-PCS Interface Width Word Alignment Mode Manual Alignment Word Alignment Pattern Length 8-, 16-, and 32-bit 8-, 16-, and 32-bit rx_syncstatus Behavior rx_patterndetect Behavior Rising Edge Sensitive Stays high after the word aligner aligns to the word alignment pattern. Goes low on receiving a rising edge on rx_enapattern align until a new word alignment pattern is received. Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. N/A N/A Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. Rising Edge Sensitive Stays high after the word aligner aligns to the word alignment pattern. Goes low on receiving a rising edge on rx_enapattern align until a new word alignment pattern is received. Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. N/A Asserted high for one parallel clock cycle when the word alignment pattern appears in the current word boundary. Basic DoubleWidth Manual Alignment 7-, 10-, and 20-bit 20-bit Bit-Slip 7-, 10-, and 20-bit (Part 2 of 2) rx_enapatternalign Sensitivity 16-bit Bit-Slip (1) N/A Note to Table 1-31: (1) For more information about word aligner operation, refer to "Word Aligner in Single-Width Mode" on page 1-60 and "Word Aligner in DoubleWidth Mode" on page 1-66. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-71 Programmable Run Length Violation Detection The programmable run length violation circuit resides in the word aligner block and detects consecutive 1s or 0s in the data. If the data stream exceeds the preset maximum number of consecutive 1s or 0s, the violation is signified by the assertion of the rx_rlv signal. The run length violation status signal on the rx_rlv port has lower latency when compared with the parallel data on the rx_dataout port. The rx_rlv signal in each channel is clocked by its parallel recovered clock. The FPGA fabric clock might have phase difference and/or PPM difference (in asynchronous systems) with respect to the recovered clock. To ensure that the FPGA fabric clock can latch the rx_rlv signal reliably, the run length violation circuitry asserts the rx_rlv signal for a minimum of two recovered clock cycles in single-width modes and a minimum of three recovered clock cycles in double-width modes. The rx_rlv signal can be asserted longer, depending on the run length of the received data. In single-width mode, the run length violation circuit detects up to a run length of 128 (for an 8-bit deserialization factor) or 160 (for a 10-bit deserialization factor). The settings are in increments of four or five for the 8-bit or 10-bit deserialization factors, respectively. In double-width mode, the run length violation circuit maximum run length detection is 512 (with a run length increment of eight) and 640 (with a run length increment of 10) for the 16-bit and 20-bit deserialization factors, respectively. Table 1-32 lists the detection capabilities of the run length violation circuit. Table 1-32. Detection Capabilities of the Run Length Violation Circuit Mode PMA-PCS Interface Width Single-width mode Double-width mode Run Length Violation Detector Range Minimum Maximum 8-bit 4 128 10-bit 5 160 16-bit 8 512 20-bit 10 640 Receiver Polarity Inversion The positive and negative signals of a serial differential link are often erroneously swapped during board layout. Solutions like board re-spin or major updates to the PLD logic can be expensive. The receiver polarity inversion feature is provided to correct this situation. An optional rx_invpolarity port is available in all single-width and double-width modes except (OIF) CEI PHY and PCIe modes to dynamically enable the receiver polarity inversion feature. In single-width modes, a high value on the rx_invpolarity port inverts the polarity of every bit of the 8-bit or 10-bit input data word to the word aligner in the receiver datapath. In double-width modes, a high value on the rx_invpolarity port inverts the polarity of every bit of the 16-bit or 20-bit input data word to the word aligner in the receiver datapath. Because inverting the polarity of each bit has the same effect as swapping the positive and negative signals of the differential link, correct data is seen by the receiver. rx_invpolarity is a dynamic signal and can cause initial disparity errors in an 8B/10B encoded link. The downstream system must be able to tolerate these disparity errors. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-72 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The generic receiver polarity inversion feature is different from the PCIe 8B/10B polarity inversion feature. The generic receiver polarity inversion feature inverts the polarity of the data bits at the input of the word aligner and is not available in PCIe mode. The PCIe 8B/10B polarity inversion feature inverts the polarity of the data bits at the input of the 8B/10B decoder and is available only in PCIe mode. Figure 1-54 shows the receiver polarity inversion feature in single-width and double-width datapath configurations. Figure 1-54. Receiver Polarity Inversion in Single-Width and Double Width Mode Double-Width Configuration Single-Width Configuration 1 0 0 1 1 0 0 1 0 1 0 1 0 1 rx_invpolarity = High 1 1 0 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 0 0 1 0 Output from Deserializer Stratix IV Device Handbook Volume 2: Transceivers To the Word Aligner 0 1 Input to Word Aligner September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-73 Receiver Bit Reversal By default, the Stratix IV GX and GT receiver assumes a LSB-to-MSB transmission. If the transmission order is MSB-to-LSB, the receiver forwards the bit-flipped version of the parallel data to the FPGA fabric on the rx_dataout port. The receiver bit reversal feature is available to correct this situation. The receiver bit reversal feature is available through the rx_revbitordwa port only in Basic single-width and double-width modes with the word aligner configured in bit-slip mode. When the rx_revbitordwa signal is driven high in Basic single-width mode, the 8-bit or 10-bit data D[7:0] or D[9:0] at the output of the word aligner gets rewired to D[0:7] or D[0:9], respectively. When the rx_revbitordwa signal is driven high in Basic double-width mode, the 16-bit or 20-bit data D[15:0] or D[19:0] at the output of the word aligner gets rewired to D[0:15] or D[0:19], respectively. Flipping the parallel data using this feature allows the receiver to forward the correct bit-ordered data to the FPGA fabric on the rx_dataout port in the case of MSB-to-LSB transmission. Figure 1-55 shows the receiver bit reversal feature in Basic single-width 10-bit wide datapath configurations. Figure 1-55. Receiver Bit Reversal in Single-Width Mode D[9] D[0] D[8] D[1] D[7] D[2] D[6] D[3] D[5] rx_revbitordwa = high D[4] D[5] D[3] D[6] D[2] D[7] D[1] D[8] D[0] D[9] Output of Word Aligner before RX bit reversal September 2012 Altera Corporation D[4] Output of Word Aligner after RX bit reversal Stratix IV Device Handbook Volume 2: Transceivers 1-74 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-56 shows the receiver bit reversal feature in Basic double-width 20-bit wide datapath configurations. Figure 1-56. Receiver Bit Reversal in Double-Width Mode D[19] D[0] D[18] D[1] D[17] D[2] D[16] D[3] D[15] D[4] D[14] D[5] D[13] D[6] D[12] D[7] D[11] D[8] D[10] rx_revbitordwa = high D[9] D[9] D[10] D[8] D[11] D[7] D[12] D[6] D[13] D[5] D[14] D[4] D[15] D[3] D[16] D[2] D[17] D[1] D[18] D[0] D[19] Output of Word Aligner before RX bit reversal Output of Word Aligner after RX bit reversal Because receiver bit reversal is done at the output of the word aligner, a dynamic bit reversal also requires a reversal of the word alignment pattern. As a result, the Receiver Bit Reversal feature is dynamic only if the receiver is dynamically reconfigurable (it allows changing the word alignment pattern dynamically) or uses manual bit slip alignment mode (no word alignment pattern). The Receiver Bit Reversal feature is static in all other Basic mode configurations. You can enable this feature using the MegaWizard Plug-In Manager. In configurations where the Receiver Bit Reversal feature is dynamic, an rx_revbitordwa port is available to control the bit reversal dynamically. A high on the rx_revbitordwa port reverses the bit order at the input of the word aligner. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-75 Receiver Byte Reversal in Basic Double-Width Modes The MSByte and LSByte of the input data to the transmitter may be erroneously swapped. The receiver byte reversal feature is available to correct this situation. An optional port, rx_revbyteordwa, is available only in Basic double-width mode to enable receiver byte reversal. In 8B/10B enabled mode, a high value on rx_revbyteordwa exchanges the 10-bit MSByte for the LSByte of the 20-bit word at the output of the word aligner in the receiver datapath. In non-8B/10B enabled mode, a high value on rx_revbyteordwa exchanges the 8-bit MSByte for the LSByte of the 16-bit word at the output of the word aligner in the receiver datapath. This compensates for the erroneous exchanging at the transmitter and corrects the data received by the downstream systems. rx_revbyteorderwa is a dynamic signal and can cause an initial disparity error at the receiver of an 8B/10B encoded link. The downstream system must be able to tolerate this disparity error. Figure 1-57 shows the receiver byte reversal feature. Figure 1-57. Receiver Byte Reversal Feature MSByte 00 02 04 06 08 0A LSByte 01 03 05 07 09 0B MSByte 01 03 05 07 09 0B LSByte 00 02 MSByte 00 LSByte 01 Data without Receiver Byte Reversal enabled Data with Receiver Byte Reversal enabled 04 06 08 0A 02 07 09 0B 03 06 08 0A rx_revbyteordwa Word Aligner Output with rx_revbyteordwa asserted Deskew FIFO Code groups received across four lanes in a XAUI link can be misaligned with respect to one another because of skew in the physical medium or differences between the independent clock recoveries per lane. The XAUI protocol allows a maximum skew of 40 UI (12.8 ns) as seen at the receiver of the four lanes. XAUI protocol requires the physical layer device to implement a deskew circuitry to align all four channels. To enable the deskew circuitry at the receiver to align the four channels, the transmitter sends a /A/ (/K28.3/) code group simultaneously on all four channels during inter-packet gap (IPG). The skew introduced in the physical medium and the receiver channels can cause the /A/ code groups to be received misaligned. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-76 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Deskew circuitry performs the deskew operation by the XAUI functional mode. Deskew circuitry consists of: 1 A 16-word deep deskew FIFO in each of the four channels Control logic in the CMU0 channel of the transceiver block that controls the deskew FIFO write and read operations in each channel Deskew circuitry is only available in XAUI mode. The deskew FIFO in each channel receives data from its word aligner. The deskew operation begins only after link synchronization is achieved on all four channels as indicated by a high level on the rx_syncstatus signal from the word aligner in each channel. Until the first /A/ code group is received, the deskew FIFO read and write pointers in each channel are not incremented. After the first /A/ code group is received, the write pointer starts incrementing for each word received but the read pointer is frozen. If the /A/ code group is received on each of the four channels within 10 recovered clock cycles of each other, the read pointer for all four deskew FIFOs is released simultaneously, aligning all four channels. Figure 1-58 shows lane skew at the receiver input and how the deskew FIFO uses the /A/ code group to align the channels. Figure 1-58. Deskew FIFO--Lane Skew at the Receiver Input Lane 0 K K Lane 1 Lane 2 K Lane 3 R A K R R K K R K R K K R A K R R K K R K K R A K R R K K R K R K K R A K R R K K R K R Lane Skew at Receiver Input R Lane 0 K K R A K R R K K R K R Lane 1 K K R A K R R K K R K R Lane 2 K K R A K R R K K R K R Lane 3 K K R A K R R K K R K R Lanes are Deskewed by Lining up the "Align"/A/, Code Groups After alignment of the first ||A|| column, if three additional aligned ||A|| columns are observed at the output of the deskew FIFOs of the four channels, the rx_channelaligned signal is asserted high, indicating channel alignment is acquired. After acquiring channel alignment, if four misaligned ||A|| columns are seen at the output of the deskew FIFOs in all four channels with no aligned ||A|| columns in between, the rx_channelaligned signal is de-asserted low, indicating loss-of-channel alignment. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-77 The deskew operation in XAUI functional mode is compliant to the PCS deskew state machine diagram specified in clause 48 of IEEE P802.3ae, as shown in Figure 1-59. (1) Figure 1-59. Deskew FIFO Operation in XAUI Functional Mode reset + (sync_status=FAIL * SUDI) LOSS_OF_ALIGNMENT SUDI(![/||A||/]) align_status FAIL enable_deskew TRUE AUDI sync_status OK * SUDI(![/||A||/]) ALIGN_DETECT_1 !deskew_error * SUDI(![/||A||/]) enable_deskew FALSE AUDI deskew_error * SUDI SUDI(![/||A||/]) ALIGN_DETECT_2 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI SUDI(![/||A||/]) ALIGN_DETECT_3 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI 1 SUDI(![/||A||/]) ALIGN_ACQUIRED_1 !deskew_error * SUDI(![/||A||/]) enable_deskew FALSE AUDI deskew_error * SUDI 2 SUDI(![/||A||/]) ALIGN_ACQUIRED_2 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI 3 SUDI(![/||A||/]) 1 ALIGN_ACQUIRED_3 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI SUDI(![/||A||/]) 2 ALIGN_ACQUIRED_4 AUDI deskew_error * SUDI !deskew_error * SUDI(![/||A||/]) SUDI(![/||A||/]) 3 Note to Figure 1-59: (1) This figure is from IEEE P802.3ae. Rate Match (Clock Rate Compensation) FIFO In asynchronous systems, the upstream transmitter and local receiver can be clocked with independent reference clocks. Frequency differences in the order of a few hundred PPM can corrupt the data when latching from the recovered clock domain (the same clock domain as the upstream transmitter reference clock) to the local receiver reference clock domain. The rate match FIFO compensates for small clock frequency differences between the upstream transmitter and the local receiver clocks by inserting or removing SKP symbols or ordered sets from the IPG or idle streams. It deletes SKP symbols or ordered sets when the upstream transmitter reference clock frequency is higher than the local receiver reference clock frequency. It inserts SKP symbols or ordered-sets when the local receiver reference clock frequency is higher than the upstream transmitter reference clock frequency. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-78 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The rate match FIFO consists of a 20-word deep FIFO and necessary logic that controls insertion and deletion of a skip character or ordered set, depending on the PPM difference. The rate match FIFO is mandatory and cannot be bypassed in the following functional modes: PCIe XAUI GIGE The rate match FIFO is optional in the following functional modes: Basic single-width Basic double-width SRIO The rate match FIFO receives data from the word aligner (non-XAUI functional modes) or deskew FIFO (XAUI functional mode) in the receiver datapath. It provides the following status signals forwarded to the FPGA fabric: 1 rx_rmfifodatainserted--indicates insertion of a skip character or ordered set rx_rmfifodatadeleted--indicates deletion of a skip character or ordered set rx_rmfifofull--indicates rate match FIFO full condition rx_rmfifoempty--indicates rate match FIFO empty condition The rate match FIFO status signals are not available in PCIe mode. These signals are encoded on the pipestatus[2:0] signal in PCIe mode as specified in the PCIe specification. Rate Match FIFO in PCIe Mode In PCIe mode, the rate match FIFO is capable of compensating up to 300 PPM (total 600 PPM) difference between the upstream transmitter and the local receiver. The PCIe protocol requires the transmitter to send SKP ordered sets during IPGs, adhering to rules listed in the base specification. The SKP ordered set is defined as a /K28.5/ COM symbol followed by three consecutive /K28.0/ SKP symbol groups. The PCIe protocol requires the receiver to recognize a SKP ordered set as a /K28.5/ COM symbol followed by one to five consecutive /K28.0/ SKP symbols. The rate match FIFO operation is compliant to PCIe Base Specification 2.0. The rate match operation begins after the synchronization state machine in the word aligner indicates synchronization is acquired by driving the rx_syncstatus signal high. The rate match FIFO looks for the SKP ordered set and deletes or inserts SKP symbols as necessary to prevent the rate match FIFO from overflowing or under-running. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-79 The rate match FIFO inserts or deletes only one SKP symbol per SKP ordered set received. Rate match FIFO insertion and deletion events are communicated to the FPGA fabric on the pipestatus[2:0] port from each channel. The pipestatus[2:0] signal is driven to 3'b001 for one clock cycle synchronous to the /K28.5/ COM symbol of the SKP ordered set in which the /K28.0/ SKP symbol is inserted. The pipestatus[2:0] signal is driven to 3'b010 for one clock cycle synchronous to the /K28.5/ COM symbol of the SKP ordered set from which the /K28.0/ SKP symbol is deleted. Figure 1-60 shows an example of rate match deletion in the case where two /K28.0/ SKP symbols are required to be deleted. Only one /K28.0/ SKP symbol is deleted per SKP ordered set received. Figure 1-60. Rate Match Deletion in PCIe Mode SKIP Symbol Deleted First Skip Ordered Set Second Skip Ordered Set datain K28.5 K28.0 Dx.y K28.5 K28.0 dataout K28.5 Dx.y K28.5 K28.0 K28.0 K28.0 pipestatus[2:0] 3'b010 xxx 3'b010 xxx xxx xxx K28.0 K28.0 Figure 1-61 shows an example of rate match insertion in the case where two SKP symbols are required to be inserted. Only one /K28.0/ SKP symbol is inserted per SKP ordered set received. Figure 1-61. Rate Match Insertion in PCIe Mode SKIP Symbol Inserted First Skip Ordered Set Second Skip Ordered Set datain K28.5 K28.0 Dx.y K28.5 K28.0 K28.0 K28.0 K28.0 dataout K28.5 K28.0 K28.0 Dx.y K28.5 K28.0 K28.0 K28.0 K28.0 K28.0 pipestatus[2:0] 3'b001 xxx xxx xxx 3'b001 xxx xxx xxx xxx xxx The rate match FIFO full and empty conditions are communicated to the FPGA fabric on the pipestatus[2:0] port from each channel. The rate match FIFO in PCIe mode automatically deletes the data byte that causes the FIFO to go full and drives pipestatus[2:0] = 3'b101 synchronous to the subsequent data byte. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-80 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-62 shows the rate match FIFO full condition in PCIe mode. The rate match FIFO becomes full after receiving data byte D4. Figure 1-62. Rate Match FIFO Full Condition in PCIe Mode datain D1 D2 D3 D4 D5 D6 D7 dataout D1 D2 D3 D4 D6 D7 D8 xx pipestatus[2:0] xxx xxx xxx xxx xxx xxx xxx 3'b101 D8 xx xx The rate match FIFO automatically inserts /K30.7/ (9'h1FE) after the data byte that causes the FIFO to go empty and drives PCIestatus[2:0] = 3'b110 flag synchronous to the inserted /K30.7/ (9'h1FE). Figure 1-63 shows rate match FIFO empty condition in PCIe mode. The rate match FIFO becomes empty after reading out data byte D3. Figure 1-63. Rate Match FIFO Empty Condition in PCIe Mode datain D1 D2 D3 D4 D5 D6 dataout D1 D2 D3 /K30.7/ D4 D5 xxx xxx 3'b110 xxx xxx pipestatus[2:0] 1 xxx You can configure the rate match FIFO in low latency mode by turning off the Enable Rate Match FIFO option in the ALTGX MegaWizard Plug-In Manager. Rate Match FIFO in XAUI Mode In XAUI mode, the rate match FIFO is capable of compensating for up to 100 PPM (200 PPM total) difference between the upstream transmitter and the local receiver reference clock. The XAUI protocol requires the transmitter to send /R/ (/K28.0/) code groups simultaneously on all four lanes (denoted as ||R|| column) during inter-packet gaps, adhering to rules listed in the IEEE P802.3ae specification. The rate match FIFO operation in XAUI mode is compliant to the IEEE P802.3ae specification. The rate match operation begins after: The synchronization state machine in the word aligner of all four channels indicates synchronization was acquired by driving its rx_syncstatus signal high The deskew FIFO block indicates alignment was acquired by driving the rx_channelaligned signal high The rate match FIFO looks for the ||R|| column (simultaneous /R/ code group on all four channels) and deletes or inserts the ||R|| column to prevent the rate match FIFO from overflowing or under-running. It can insert or delete as many ||R|| columns as necessary to perform the rate match operation. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-81 Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, that indicate rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. If an ||R|| column is deleted, the rx_rmfifodeleted flag from each of the four channels goes high for one clock cycle per deleted ||R|| column. If an ||R|| column is inserted, the rx_rmfifoinserted flag from each of the four channels goes high for one clock cycle per inserted ||R|| column. Figure 1-64 shows an example of rate match deletion in the case where three ||R|| columns must be deleted. Figure 1-64. Rate Match Deletion in XAUI Mode First ||R|| Column Before Rate Match: Second ||R|| Column Third ||R|| Column Fourth ||R|| Column datain[3] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[2] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[1] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[0] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 dataout[3] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[2] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[1] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[0] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 After Rate Match: rx_rmfifodatadeleted September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-82 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-65 shows an example of rate match insertion in the case where two ||R|| columns are required to be inserted. Figure 1-65. Rate Match Insertion in XAUI Mode First ||R|| Column Second ||R|| Column datain[3] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 datain[2] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 datain[1] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 datain[0] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 dataout[3] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 dataout[2] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 dataout[1] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 dataout[0] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 rx_rmfifodatainserted Two flags, rx_rmfifofull and rx_rmfifoempty, are forwarded to the FPGA fabric to indicate rate match FIFO full and empty conditions. In XAUI mode, the rate match FIFO does not automatically insert or delete code groups to overcome FIFO empty and full conditions, respectively. It asserts the rx_rmfifofull and rx_rmfifoempty flags for at least three recovered clock cycles to indicate rate match FIFO full and empty conditions, respectively. 1 In the case of rate match FIFO full and empty conditions, you must assert the rx_digitalreset signal to reset the receiver PCS blocks. Rate Match FIFO in GIGE Mode In GIGE mode, the rate match FIFO is capable of compensating for up to 100 PPM (200 PPM total) difference between the upstream transmitter and the local receiver reference clock. The GIGE protocol requires the transmitter to send idle ordered sets /I1/ (/K28.5/D5.6/) and /I2/ (/K28.5/D16.2/) during inter-packet gaps, adhering to rules listed in the IEEE 802.3 specification. The rate match operation begins after the synchronization state machine in the word aligner indicates synchronization is acquired by driving the rx_syncstatus signal high. The rate match FIFO is capable of deleting or inserting the /I2/ (/K28.5/D16.2/) ordered set to prevent the rate match FIFO from overflowing or under running during normal packet transmission. The rate match FIFO is also capable of deleting or inserting the first two bytes of the /C2/ ordered set (/K28.5/D2.2/Dx.y/Dx.y/) to prevent the rate match FIFO from overflowing or under running during the auto negotiation phase. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-83 The rate match FIFO can insert or delete as many /I2/ or /C2/ (first two bytes) as necessary to perform the rate match operation. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, that indicate rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. Both the rx_rmfifodatadeleted and rx_rmfifodatainserted flags are asserted for two clock cycles for each deleted and inserted /I2/ ordered set, respectively. Figure 1-66 shows an example of rate match FIFO deletion in the case where three symbols are required to be deleted. Because the rate match FIFO can only delete /I2/ ordered set, it deletes two /I2/ ordered sets (four symbols deleted). Figure 1-66. Rate Match Deletion in GIGE Mode /I2/ Ordered Set Deleted First /I2/ Ordered Set Second /I2/ Ordered Set datain Dx.y K28.5 D16.2 K28.5 dataout Dx.y K28.5 D16.2 Dx.y Third /I2/ Ordered Set D16.2 Dx.y D16.2 K28.5 rx_rmfifodatadeleted Figure 1-67 shows an example of rate match FIFO insertion in the case where one symbol is required to be inserted. Because the rate match FIFO can only insert a /I2/ ordered set, it inserts one /I2/ ordered set (two symbols inserted). Figure 1-67. Rate Match Insertion in GIGE Mode First /I2/ Ordered Set Second /I2/ Ordered Set datain Dx.y K28.5 D16.2 K28.5 D16.2 dataout Dx.y K28.5 D16.2 K28.5 D16.2 K28.5 D16.2 Dx.y rx_rmfifodatainserted Two flags, rx_rmfifofull and rx_rmfifoempty, are forwarded to the FPGA fabric to indicate rate match FIFO full and empty conditions. In GIGE mode, the rate match FIFO does not insert or delete code groups automatically to overcome FIFO empty and full conditions, respectively. It asserts the rx_rmfifofull and rx_rmfifoempty flags for at least two recovered clock cycles to indicate rate match FIFO full and empty conditions, respectively. 1 September 2012 In the case of rate match FIFO full and empty conditions, you must assert the rx_digitalreset signal to reset the receiver PCS blocks. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-84 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Rate Match FIFO in Basic Single-Width Mode In Basic single-width mode, the rate match FIFO is capable of compensating for up to 300 PPM (600 PPM total) difference between the upstream transmitter and the local receiver reference clock. 1 To enable the rate match FIFO in Basic single-width mode, the transceiver channel must have both the transmitter and receiver channel instantiated. You must select the Receiver and Transmitter option in the What is the operation mode? field in the ALTGX MegaWizard Plug-In Manager. You must also enable the 8B/10B encoder/decoder in Basic single-width mode with rate match FIFO enabled. Depending on your proprietary protocol implementation, you can select two 20-bit rate match patterns in the ALTGX MegaWizard Plug-In Manager under the What is the rate match pattern1 and What is the rate match pattern2 fields. Each of the two programmed 20-bit rate match patterns consists of a 10-bit skip pattern and a 10-bit control pattern. You must choose 10-bit code groups that have a neutral disparity as the skip patterns. The rate match FIFO operation begins after the word aligner synchronization status rx_syncstatus goes high. When the rate matcher receives either of the two 10-bit control patterns followed by the respective 10-bit skip pattern, it inserts or deletes the 10-bit skip pattern as necessary to avoid the rate match FIFO from overflowing or under running. The rate match FIFO can delete a maximum of four skip patterns from a cluster, if there is one skip pattern left in the cluster after deletion. The rate match FIFO can insert a maximum of four skip patterns in a cluster, if there are no more than five skip patterns in the cluster after insertion. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, indicating rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. Figure 1-68 shows an example of rate match FIFO deletion in the case where three skip patterns are required to be deleted. In this example, /K28.5/ is the control pattern and neutral disparity /K28.0/ is the skip pattern. The first skip cluster has a /K28.5/ control pattern followed by two /K28.0/ skip patterns. The second skip cluster has a /K28.5/ control pattern followed by four /K28.0/ skip patterns. The rate match FIFO deletes only one /K28.0/ skip pattern from the first skip cluster to maintain at least one skip pattern in the cluster after deletion. Two /K28.0/ skip patterns are deleted from the second cluster for a total of three skip patterns deletion requirement. Figure 1-68. Rate Match Deletion in Basic Single-Width Mode Three Skip Patterns Deleted Second Skip Cluster First Skip Cluster datain K28.5 K28.0 K28.0 K28.5 K28.0 K28.0 dataout K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.0 K28.0 rx_rmfifodatadeleted Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-85 Figure 1-69 shows an example of rate match FIFO insertion in the case where three skip patterns are required to be inserted. In this example, /K28.5/ is the control pattern and neutral disparity /K28.0/ is the skip pattern. The first skip cluster has a /K28.5/ control pattern followed by three /K28.0/ skip patterns. The second skip cluster has a /K28.5/ control pattern followed by one /K28.0/ skip pattern. The rate match FIFO inserts only two /K28.0/ skip patterns into the first skip cluster to maintain a maximum of five skip patterns in the cluster after insertion. One /K28.0/ skip pattern is inserted into the second cluster for a total of three skip patterns to meet the insertion requirement. Figure 1-69. Rate Match Insertion in Basic Single-Width Mode Three Skip Patterns Inserted First Skip Cluster Second Skip Cluster datain K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 K28.0 Dx.y dataout K28.5 K28.0 K28.0 K28.0 K28.0 K28.0 K28.5 K28.0 K28.0 K28.0 Dx.y rx_rmfifoinserted Two flags, rx_rmfifofull and rx_rmfifoempty, are forwarded to the FPGA fabric to indicate rate match FIFO full and empty conditions. The rate match FIFO in Basic single-width mode automatically deletes the data byte that causes the FIFO to go full and asserts the rx_rmfifofull flag synchronous to the subsequent data byte. Figure 1-70 shows the rate match FIFO full condition in Basic single-width mode. The rate match FIFO becomes full after receiving data byte D4. Figure 1-70. Rate Match FIFO Full Condition in Basic Single-Width Mode datain D1 D2 D3 D4 D5 D6 D7 dataout D1 D2 D3 D4 D6 D7 D8 D8 xx xx xx rx_rmfifofull The rate match FIFO automatically inserts /K30.7/ (9'h1FE) after the data byte that causes the FIFO to go empty and asserts the rx_fifoempty flag synchronous to the inserted /K30.7/ (9'h1FE). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-86 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-71 shows the rate match FIFO empty condition in Basic single-width mode. The rate match FIFO becomes empty after reading out data byte D3. Figure 1-71. Rate Match FIFO Empty Condition in Basic Single-Width Mode datain D1 D2 D3 D4 D5 D6 dataout D1 D2 D3 /K30.7/ D4 D5 rx_rmfifoempty Rate Match FIFO in Basic Double-Width Mode In Basic double-width mode, the rate match FIFO is capable of compensating up to 300 PPM (total 600 PPM total) difference between the upstream transmitter and the local receiver reference clock. 1 To enable the rate match FIFO in Basic double-width mode, the transceiver channel must have both the transmitter and receiver channel instantiated. You must select the Receiver and Transmitter option in the What is the operation mode? field in the ALTGX MegaWizard Plug-In Manager. You must also enable the 8B/10B encoder/decoder in Basic double-width mode with rate match FIFO enabled. Depending on your proprietary protocol implementation, you can select two 20-bit rate match patterns in the ALTGX MegaWizard Plug-In Manager under the What is the rate match pattern1 and What is the rate match pattern2 fields. Each of the two programmed 20-bit rate match patterns consists of a 10-bit skip pattern and a 10-bit control pattern. You must choose 10-bit code groups that have a neutral disparity as the skip patterns. The rate match FIFO operation begins after the word aligner synchronization status rx_syncstatus goes high. When the rate matcher receives either of the two 10-bit control patterns followed by the respective 10-bit skip pattern, it inserts or deletes a pair of 10-bit skip patterns as necessary to avoid the rate match FIFO from overflowing or under running. The rate match FIFO can delete as many pairs of skip patterns from a cluster necessary to avoid the rate match FIFO from overflowing. The rate match FIFO can delete a pair of skip patterns only if the two 10-bit skip patterns appear in the same clock cycle on the LSByte and MSByte of the 20-bit word. If the two skip patterns appear straddled on the MSByte of a clock cycle and the LSByte of the next clock cycle, the rate match FIFO cannot delete the pair of skip patterns. The rate match FIFO can insert as many pairs of skip patterns into a cluster necessary to avoid the rate match FIFO from under running. The 10-bit skip pattern can appear on MSByte or LSByte, or both, of the 20-bit word. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, indicating rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. Figure 1-72 shows an example of rate match FIFO deletion in the case where three skip patterns are required to be deleted. In this example, /K28.5/ is the control pattern and neutral disparity /K28.0/ is the skip pattern. The first skip cluster has a /K28.5/ control pattern in the LSByte and /K28.0/ skip pattern in the MSByte of a clock cycle followed by one /K28.0/ skip pattern in the LSByte of the next clock cycle. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-87 The rate match FIFO cannot delete the two skip patterns in this skip cluster because they do not appear in the same clock cycle. The second skip cluster has a /K28.5/ control pattern in the MSByte of a clock cycle followed by two pairs of /K28.0/ skip patterns in the next two cycles. The rate match FIFO deletes both pairs of /K28.0/ skip patterns (for a total of four skip patterns deleted) from the second skip cluster to meet the three skip pattern deletion requirement. Figure 1-72. Rate Match Deletion in Basic Double-Width Mode Two Pairs of Skip Patterns Deleted Second Skip Cluster First Skip Cluster datain[19:10] Dx.y K28.0 Dx.y K28.5 K28.0 K28.0 Dx.y datain[9:0] Dx.y K28.5 K28.0 Dx.y K28.0 K28.0 Dx.y dataout[19:0] Dx.y K28.0 Dx.y K28.5 Dx.y dataout[9:0] Dx.y K28.5 K28.0 Dx.y Dx.y rx_rmfifodatadeleted Figure 1-73 shows an example of rate match FIFO insertion in the case where three skip patterns are required to be inserted. In this example, /K28.5/ is the control pattern and neutral disparity /K28.0/ is the skip pattern. The first skip cluster has a /K28.5/ control pattern in the LSByte and /K28.0/ skip pattern in the MSByte of a clock cycle followed by one /K28.0/ skip pattern in the LSByte of the next clock cycle. The rate match FIFO inserts pairs of skip patterns in this skip cluster to meet the three skip pattern insertion requirement. Figure 1-73. Rate Match Insertion in Basic Double-Width Mode Second Skip Cluster First Skip Cluster dataout[19:0] Dx.y K28.0 Dx.y K28.5 K28.0 K28.0 dataout[9:0] Dx.y K28.5 Dx.y Dx.y K28.0 K28.0 datain[19:10] Dx.y K28.0 K28.0 K28.0 Dx.y K28.5 K28.0 K28.0 datain[9:0] Dx.y K28.5 K28.0 K28.0 Dx.y Dx.y K28.0 K28.0 rx_rmfifodatainserted September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-88 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Two flags, rx_rmfifofull and rx_rmfifoempty, are forwarded to the FPGA fabric to indicate rate match FIFO full and empty conditions. The rate match FIFO in Basic double-width mode automatically deletes the pair of data byte that causes the FIFO to go full and asserts the rx_rmfifofull flag synchronous to the subsequent pair of data bytes. Figure 1-74 shows the rate match FIFO full condition in Basic double-width mode. The rate match FIFO becomes full after receiving the 20-bit word D5D6. Figure 1-74. Rate Match FIFO Full Condition in Basic Double-Width Mode datain[19:10] D2 D4 D6 D8 D10 D12 datain[9:0] D1 D3 D5 D7 D9 D11 dataout[19:0] D2 D4 D6 D10 D12 xx dataout[9:0] D1 D3 D5 D9 D11 xx rx_rmfifofull The rate match FIFO automatically inserts a pair of /K30.7/ ({9'h1FE,9'h1FE}) after the data byte that causes the FIFO to go empty and asserts the rx_fifoempty flag synchronous to the inserted pair of /K30.7/ ({9'h1FE,9'h1FE}). Figure 1-75 shows the rate match FIFO empty condition in Basic double-width mode. The rate match FIFO becomes empty after reading out the 20-bit word D5D6. Figure 1-75. Rate Match FIFO Empty Condition in Basic Double-Width Mode datain[19:10] D2 D4 D6 D8 D10 D12 datain[9:0] D1 D3 D5 D7 D9 D11 dataout[19:0] D2 D4 D6 /K30.7/ D8 D10 dataout[9:0] D1 D3 D5 /K30.7/ D7 D9 rx_rmfifoempty Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-89 8B/10B Decoder Protocols such as PCIe, XAUI, GIGE, and Serial RapidIO require the serial data sent over the link to be 8B/10B encoded to maintain the DC balance in the serial data transmitted. These protocols require the receiver PCS logic to implement an 8B/10B decoder to decode the data before forwarding it to the upper layers for packet processing. The Stratix IV GX and GT receiver channel PCS datapaths implement the 8B/10B decoder after the rate matcher. In functional modes with rate matcher enabled, the 8B/10B decoder receives data from the rate matcher. In functional modes with rate matcher disabled, the 8B/10B decoder receives data from the word aligner. The 8B/10B decoder operates in two modes (Figure 1-76): Single-width mode Double-width mode Figure 1-76. 8B/10B Decoder in Single-Width and Double-Width Mode Single-Width Mode Double-Width Mode rx_dataout [15:8] datain [19:10] rx_dataout[15:8] datain[19:10] rx_ctrldetect[1] rx_ctrldetect[1] 8B/10B Decoder (MSB Byte) 8B/10B Decoder (MSB Byte) rx_errdetect[1] rx_disperr[1] rx_disperr[1] recovered clock or tx_clkout[0] Current Running Disparity datain[9:0] rx_dataout[7:0] recovered clock or tx_clkout[0] Current Running Disparity rx_dataout[7:0] datain[9:0] rx_ctrldetect rx_ctrldetect 8B/10B Decoder (LSB Byte) rx_errdetect[1] 8B/10B Decoder (LSB Byte) rx_errdetect rx_errdetect rx_disperr rx_disperr recovered clock or tx_clkout[0] recovered clock or tx_clkout[0] 8B/10B Decoder in Single-Width Mode The left side of Figure 1-76 shows the 8B/10B decoder in single-width mode. In this mode, the 8B/10B decoder receives 10-bit data from the rate matcher or word aligner (when rate matcher is disabled) and decodes it into an 8-bit data + 1-bit control identifier. The decoded data is fed to the byte deserializer or the receiver phase compensation FIFO (if byte deserializer is disabled). 1 September 2012 The 8B/10B decoder is compliant to Clause 36 in the IEEE802.3 specification. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-90 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The 8B/10B decoder operates in single-width mode in the following functional modes: PCIe XAUI GIGE Serial RapidIO Basic single-width For PCIe, XAUI, GIGE, and Serial RapidIO functional modes, the ALTGX MegaWizard Plug-In Manager forces selection of the 8B/10B decoder in the receiver datapath. In Basic single-width mode, it allows you to enable or disable the 8B/10B decoder depending on your proprietary protocol implementation. Figure 1-77 shows a 10-bit code group decoded into an 8-bit data and a 1-bit control identifier by the 8B/10B decoder in single-width mode. Figure 1-77. 8B/10B Decoder in Single-Width Mode j h g f i e d c b a 9 8 7 6 5 4 3 2 1 0 MSB Received Last LSB Received First 8B/10B Conversion ctrl 7 6 5 4 3 2 1 0 H G F E D C B A Parallel Data Control Code Group Detection The 8B/10B decoder indicates whether the decoded 8-bit code group is a data or control code group on the rx_ctrldetect port. If the received 10-bit code group is one of the 12 control code groups (/Kx.y/) specified in the IEEE802.3 specification, the rx_ctrldetect signal is driven high. If the received 10-bit code group is a data code group (/Dx.y/), the rx_ctrldetect signal is driven low. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-91 Figure 1-78 shows the 8B/10B decoder decoding the received 10-bit /K28.5/ control code group into an 8-bit data code group (8'hBC) driven on the rx_dataout port. The rx_ctrldetect signal is asserted high synchronous with 8'hBC on the rx_dataout port, indicating that it is a control code group. The rest of the codes received are data code groups /Dx.y/. Figure 1-78. 8B/10B Decoder in Control Code Group Detection clock datain[9..0 ] D3.4 D24.3 D28.5 K28.5 78 BC D15.0 D0.0 D31.5 0F 00 BF rx_ctrldetect rx_dataout[7..0] 83 BC 8B/10B Decoder in Double-Width Mode The left side of Figure 1-76 on page 1-89 shows the 8B/10B decoder in double-width mode. In this mode, two 8B/10B decoders are cascaded for decoding the 20-bit encoded data, as shown in Figure 1-79. The 10-bit LSByte of the received 20-bit encoded data is decoded first and the ending running disparity is forwarded to the 8B/10B decoder responsible for decoding the 10-bit MSByte. The cascaded 8B/10B decoder decodes the 20-bit encoded data into 16-bit data + 2-bit control identifier. The MSB and LSB of the 2-bit control identifier corresponds to the MSByte and LSByte of the 16-bit decoded data code group. The decoded data is fed to the byte deserializer or the receiver phase compensation FIFO (if byte deserializer is disabled). 1 Each of the two cascaded 8B/10B decoders is compliant to Clause 36 in the IEEE802.3 specification. The 8B/10B decoder operates in double-width mode only in Basic double-width functional mode. You can enable or disable the 8B/10B decoder depending on your proprietary protocol implementation. Figure 1-79 shows a 20-bit code group decoded into 16-bit data and 2-bit control identifier by the 8B/10B decoder in double-width mode. Figure 1-79. 8B/10 Decoder in 20-Bit Double-Width Mode j1 h1 g1 f1 i1 e1 d1 c1 b1 a1 j h g f i e d c b a 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 MSB LSB Cascaded 8B/10B Conversion CTRL[1..0] September 2012 15 14 13 13 11 10 9 8 7 6 5 4 3 2 1 0 H1 G1 F1 E1 D1 C1 B1 A1 H G F E D C B A Altera Corporation Parallel Data Stratix IV Device Handbook Volume 2: Transceivers 1-92 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Control Code Group Detection The cascaded 8B/10B decoder indicates whether the decoded 16-bit code group is a data or control code group on the 2-bit rx_ctrldetect[1:0] port. The rx_ctrldetect[0] signal is driven high or low depending on whether decoded data on the rx_dataout[7:0] port (LSByte) is a control or data code group, respectively. The rx_ctrldetect[1] signals are driven high or low depending on whether decoded data on the rx_dataout[15:8] port (MSByte) is a control or data code group, respectively. Figure 1-80 shows the 8B/10B decoding of the received 10-bit /K28.5/ control code group into 8-bit data code group (8'hBC) driven on the rx_dataout port. The rx_ctrldetect signal is asserted high synchronous with 8'hBC on the rx_dataout port, indicating that it is a control code group. The rest of the codes received are data code groups /Dx.y/. Figure 1-80. 8B/10B Decoder 10-Bit Control Code Group clock datain[19:10] D3.4 D28.5 D15.0 D3.4 datain[9:0] D24.3 D28.5 D15.0 D3.4 rx_ctrldetect[1:0] 00 01 rx_dataout[15:0] 16'h8378 16'hBCBC 00 16'h0F0F 16'h8383 Byte Deserializer The FPGA fabric-transceiver interface frequency has an upper limit that is stated in the "Interface Frequency" section in the DC and Switching Characteristics chapter. In functional modes that have a receiver PCS frequency greater than the upper limit stated in the DC and Switching Characteristics chapter, the parallel received data and status signals cannot be forwarded directly to the FPGA fabric because it violates this upper limit for the FPGA fabric-transceiver interface frequency. In such configurations, the byte deserializer is required to reduce the FPGA fabric-transceiver interface frequency to half while doubling the parallel data width. For example, at 3.2 Gbps data rate with a deserialization factor of 10, the receiver PCS datapath runs at 320 MHz. The 10-bit parallel received data and status signals at 320 MHz cannot be forwarded to the FPGA fabric because it violates the upper limit for the FPGA fabric-transceiver interface frequency. The byte serializer converts the 10-bit parallel received data at 320 MHz into 20-bit parallel data at 160 MHz before forwarding to the FPGA fabric. 1 The byte deserializer is required in configurations that exceed the FPGA fabric-transceiver interface clock upper frequency limit. It is optional in configurations that do not exceed the FPGA fabric-transceiver interface clock upper frequency limit. The byte deserializer operates in two modes: Stratix IV Device Handbook Volume 2: Transceivers Single-width mode Double-width mode September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-93 Byte Deserializer in Single-Width Mode In single-width mode, the byte deserializer receives 8-bit wide data from the 8B/10B decoder or 10-bit wide data from the word aligner (if the 8B/10B decoder is disabled) and deserializes it into 16-bit or 20-bit wide data at half the speed. Figure 1-81 shows the byte deserializer in single-width mode. Figure 1-81. Byte Deserializer in Single-Width Mode datain[7:0] or datain[9:0] D1 D2 D3 D4 Byte Deserializer D2 D4 D1 D3 dataout[15:0] or dataout[19:0] /2 Receiver PCS Clock Byte Deserializer in Double-Width Mode In double-width mode, the byte deserializer receives 16-bit wide data from the 8B/10B decoder or 20-bit wide data from the word aligner (if the 8B/10B decoder is disabled) and deserializes it into 32-bit or 40-bit wide data at half the speed. Figure 1-82 shows the byte deserializer in double-width mode. Figure 1-82. Byte Deserializer in Double-Width Mode dataout[15:0] or dataout[19:0] D1D2 D3D4 D5D6 D7D8 Byte Deserializer D3D4 D7D8 D1D2 D5D6 dataout[31:0] or dataout[39:0] /2 Receiver PCS Clock September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-94 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Byte Ordering Block In single-width modes with the 16-bit or 20-bit FPGA fabric-transceiver interface, the byte deserializer receives one data byte (8 or 10 bits) and deserializes it into two data bytes (16 or 20 bits). Depending on when the receiver PCS logic comes out of reset, the byte ordering at the output of the byte deserializer may or may not match the original byte ordering of the transmitted data. The byte misalignment resulting from byte deserialization is unpredictable because it depends on which byte is being received by the byte deserializer when it comes out of reset. Figure 1-83 shows a scenario in which the MSByte and LSByte of the two-byte transmitter data appears straddled across two word boundaries after getting byte deserialized at the receiver. Figure 1-83. MSByte and LSByte of the Two-Byte Transmitter Data Straddled Across Two Word Boundaries Transmitter tx_datain[15:8] (MSByte) D2 D4 D6 tx_datain[7:0] (LSByte) D1 D3 D5 Byte Serializer Receiver xx D1 D2 D3 D4 D5 D6 xx Byte Deserializer D1 D3 D5 xx rx_dataout[15:8] (MSByte) xx D2 D4 D6 rx_dataout[7:0] (LSByte) In double-width modes with the 32-bit or 40-bit FPGA fabric-transceiver interface, the byte deserializer receives two data bytes (16 or 20 bits) and deserializes it into four data bytes (32 or 40 bits). Figure 1-84 shows a scenario in which the two MSBytes and LSBytes of the four-byte transmitter data appears straddled across two word boundaries after getting byte deserialized at the receiver. Figure 1-84. MSByte and LSByte of the Four-Byte Transmitter Data Straddled Across Two Word Boundaries Transmitter tx_datain[31:16] D3D4 (MSBytes) D7D8 tx_datain[15:0] D1D2 (LSBytes) D5D6 Receiver Byte Serializer xx D1D2 D2D4 D5D6 D7D8 xx Byte Deserializer D1D2 D5D6 xx rx_dataout[31:16] (MSBytes) xx D3D4 D7D8 rx_dataout[15:0] (LSBytes) Stratix IV GX and GT transceivers have an optional byte ordering block in the receiver datapath that you can use to restore proper byte ordering before forwarding the data to the FPGA fabric. The byte ordering block looks for the user-programmed byte ordering pattern in the byte-deserialized data. You must select a byte ordering pattern that you know appears at the LSByte(s) position of the parallel transmitter data. If the byte ordering block finds the programmed byte ordering pattern in the MSByte(s) position of the byte-deserialized data, it inserts the appropriate number of user-programmed PAD bytes to push the byte ordering pattern to the LSByte(s) position, thereby restoring proper byte ordering. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-95 Byte Ordering Block in Single-Width Modes Table 1-33 lists the single-width byte ordering block functional modes. Table 1-33. Single Width Functional Modes for the Byte Ordering Block Functional Modes Descriptions SONET/SDH OC-48 Basic single-width mode with: Basic single-width mode with: 1 -- 16-bit FPGA fabric-transceiver interface No 8B/10B decoder (8-bit PMA-PCS interface) Word aligner in manual alignment mode 16-bit FPGA fabric-transceiver interface 8B/10B decoder Word aligner in automatic synchronization state machine mode For more information about configurations that allow the byte ordering block in the receiver datapath, refer to "Basic Single-Width Mode Configurations" on page 1-113. The Quartus II software automatically configures the byte ordering pattern and byte ordering PAD pattern for SONET/SDH OC-48 functional mode. For more information, refer to "OC-48 Byte Ordering" on page 1-177. In Basic single-width mode, you can program a custom byte ordering pattern and byte ordering PAD pattern in the ALTGX MegaWizard Plug-In Manager. Table 1-34 lists the byte ordering pattern length allowed in Basic single-width mode. Table 1-34. Byte Ordering Pattern Length in Basic Single-Width Mode Functional Mode Byte Ordering Pattern Length Byte Ordering PAD Pattern Length 8 bits 8 bits Basic single-width mode with: 16-bit FPGA fabric-transceiver interface No 8B/10B decoder Word aligner in manual alignment mode Basic single-width mode with: 16-bit FPGA fabric-transceiver interface 8B/10B decoder Word aligner in automatic synchronization state machine mode 9 bits (1) 9 bits Note to Table 1-34: (1) If a /Kx.y/ control code group is selected as the byte ordering pattern, the MSB of the 9-bit byte ordering pattern must be 1'b1. If a /Dx.y/ data code group is selected as the byte ordering pattern, the MSB of the 9-bit byte ordering pattern must be 1'b0. The least significant 8 bits must be the 8B/10B decoded version of the code group used for byte ordering. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-96 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Byte Ordering Block in Double-Width Modes Table 1-35 lists the double-width byte ordering block functional modes. Table 1-35. Double Width Functional Modes for the Byte Ordering Block Functional Modes Basic double-width mode with: Basic double-width mode with: Basic double-width mode with: 1 Descriptions 32-bit FPGA fabric-transceiver interface No 8B/10B decoder (16-bit PMA-PCS interface) Word aligner in manual alignment mode 32-bit FPGA fabric-transceiver interface 8B/10B decoder (20-bit PMA-PCS interface) Word aligner in manual alignment mode 40-bit FPGA fabric-transceiver interface No 8B/10B decoder (20-bit PMA-PCS interface) Word aligner in manual alignment mode For more information about configurations that allow the byte ordering block in the receiver datapath, refer to "Basic Double-Width Mode Configurations" on page 1-117. In Basic double-width modes, you can program a custom byte ordering pattern and byte ordering PAD pattern in the ALTGX MegaWizard Plug-In Manager. Table 1-36 lists the byte ordering pattern length allowed in Basic double-width mode. Table 1-36. Byte Ordering Pattern Length in Basic Double-Width Mode Functional Mode Byte Ordering Pattern Length Byte Ordering PAD Pattern Length 16 bits, 8 bits 8 bits Basic double-width mode with: 32-bit FPGA fabric-transceiver interface No 8B/10B decoder (16-bit PMA-PCS interface) Word aligner in manual alignment mode Basic double-width mode with: 32-bit FPGA fabric-transceiver interface 8B/10B decoder (20-bit PMA-PCS interface) Word aligner in manual alignment mode 18 bits, 9 bits (1) 9 bits Basic double-width mode with: 40-bit FPGA fabric-transceiver interface No 8B/10B decoder (20-bit PMA-PCS interface) Word aligner in manual alignment mode 20 bits, 10 bits 10 bits Note to Table 1-36: (1) The 18-bit byte ordering pattern D[17:0] consists of MSByte D[17:9] and LSByte D[8:0]; D[17] corresponds to rx_ctrldetect[1] and D[16:9] corresponds to rx_dataout[15:8]. Similarly, D[9] corresponds to rx_ctrldetect[0] and D[7:0] corresponds to rx_dataout[7:0]. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-97 The byte ordering block modes of operation in both single-width and double-width modes are: Word-alignment-based byte ordering User-controlled byte ordering Word-Alignment-Based Byte Ordering In word-alignment-based byte ordering, the byte ordering block starts looking for the byte ordering pattern in the byte-deserialized data every time it sees a rising edge on the rx_syncstatus signal. After a rising edge on the rx_syncstatus signal, if the byte ordering block finds the first data byte that matches the programmed byte ordering pattern in the MSByte position of the byte-deserialized data, it inserts one programmed PAD pattern to push the byte ordering pattern in the LSByte position. If the byte ordering block finds the first data byte that matches the programmed byte ordering pattern in the LSByte position of the byte-deserialized data, it considers the data to be byte ordered and does not insert any PAD pattern. In either case, the byte ordering block asserts the rx_byteorderalignstatus signal. 1 You can choose word-alignment-based byte ordering in the Rate match/Byte order tab of the ALTGX MegaWizard Plug-In Manager. For the What do you want the byte ordering to be based on? question, select the The sync status signal from the word aligner option. Figure 1-85 shows an example of the byte ordering operation in single-width modes. In this example, A is the programmed byte ordering pattern and PAD is the programmed PAD pattern. The byte deserialized data places the byte ordering pattern A in the MSByte position, resulting in incorrect byte ordering. Assuming that a rising edge on the rx_syncstatus signal had occurred before the byte ordering block sees the byte ordering pattern A in the MSByte position, the byte ordering block inserts a PAD byte and pushes the byte ordering pattern A in the LSByte position. The data at the output of the byte ordering block has correct byte ordering as reflected on the rx_byteorderalignstatus signal. Figure 1-85. Byte Ordering in Single-Width Modes Transmitter tx_datain[15:8] tx_datain[7:0] D2 D1 D3 A D5 D4 Byte Serializer Channel Receiver Byte Deserializer D1 A xx D2 D4 D3 Byte Ordering D1 PAD D3 D5 rx_dataout[15:8] xx D2 A D4 rx_dataout[7:0] rx_byteorderalignstatus September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-98 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture User-Controlled Byte Ordering Unlike word-alignment-based byte ordering, user-controlled byte ordering provides control to the user logic to restore correct byte ordering at the receiver. When enabled, an rx_enabyteord port is available that you can use to trigger the byte ordering operation. A rising edge on the rx_enabyteord port triggers the byte ordering block. After a rising edge on the rx_enabyteord signal, if the byte ordering block finds the first data byte that matches the programmed byte ordering pattern in the MSByte position of the byte-deserialized data, it inserts one programmed PAD pattern to push the byte ordering pattern in the LSByte position. If the byte ordering block finds the first data byte that matches the programmed byte ordering pattern in the LSByte position of the byte-deserialized data, it considers the data to be byte ordered and does not insert any PAD byte. In either case, the byte ordering block asserts the rx_byteorderalignstatus signal. Figure 1-86 shows user-controlled byte ordering in Basic double-width Mode. Figure 1-86. User-Controlled Byte Ordering in Basic Double-Width Mode Transmitter tx_datain[31:16] (MSByte) D2D3 tx_datain[15:0] (LSByte) D0D1 D4D5 D8D9 D6D7 Receiver D0D1 xxxx Byte Serializer B1B2 Channel B1B2 B1B2 D6D7 B1B2 Byte Deserializer xxxx D2D3 D4D5 Byte Ordering D8D9 D0D1 P1P2 D4D5 D8D9 xxxx rx_dataout [31:16] (MSByte) xxxx D2D3 B1B2 D6D7 B1B2 rx_dataout[15:0] (LSByte) rx_enabyteord rx_byteorderalignstatus Receiver Phase Compensation FIFO The receiver phase compensation FIFO in each channel ensures reliable transfer of data and status signals between the receiver channel and the FPGA fabric. The receiver phase compensation FIFO compensates for the phase difference between the parallel receiver PCS clock (FIFO write clock) and the FPGA fabric clock (FIFO read clock). The receiver phase compensation FIFO operates in one of the following two modes: Stratix IV Device Handbook Volume 2: Transceivers Low latency mode--The Quartus II software automatically configures the receiver phase compensation FIFO in low latency mode in all functional modes. In this mode, the FIFO is four words deep and the latency through the FIFO is two to three parallel clock cycles (pending characterization). High latency mode--In this mode, the FIFO is eight words deep and the latency through the FIFO is four to five parallel clock cycles (pending characterization). September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-99 The receiver phase compensation FIFO write clock source varies with the receiver channel configuration. Table 1-37 lists the receiver phase compensation FIFO write clock source in different configurations. Table 1-37. Receiver Phase Compensation FIFO Write Clock Source Receiver Phase Compensation FIFO Write Clock Configuration Without Byte Serializer With Byte Serializer Non-bonded channel configuration with rate matcher Parallel transmitter PCS clock from the local clock divider in the associated channel (tx_clkout) Divide-by-two version of the parallel transmitter PCS clock from the local clock divider in the associated channel (tx_clkout) Non-bonded channel configuration without rate matcher Parallel recovered clock from the receiver PMA in the associated channel (rx_clkout) Divide-by-two version of the parallel recovered clock from the receiver PMA in the associated channel (rx_clkout) x4 bonded channel configuration Parallel transmitter PCS clock from the central clock divider in the CMU0 of the associated transceiver block (coreclkout) Divide-by-two version of the parallel transmitter PCS clock from the central clock divider in CMU0 of the associated transceiver block (coreclkout) x8 bonded channel configuration Parallel transmitter PCS clock from the central clock divider in CMU0 of the master transceiver block (coreclkout from master transceiver block) Divide-by-two version of the parallel transmitter PCS clock from the central clock divider in CMU0 of the master transceiver block (coreclkout from master transceiver block) The receiver phase compensation FIFO read clock source varies depending on whether or not you instantiate the rx_coreclk port in the ALTGX MegaWizard Plug-In Manager. Table 1-38 lists the receiver phase compensation FIFO read clock source in different configurations. Table 1-38. Receiver Phase Compensation FIFO Read Clock Source Receiver Phase Compensation FIFO Read Clock Configuration rx_coreclk Port Not Instantiated rx_coreclk Port Instantiated (1) Non-bonded channel configuration with rate matcher FPGA fabric clock driven by the clock signal on the tx_clkout port FPGA fabric clock driven by the clock signal on the rx_coreclk port Non-bonded channel configuration without rate matcher FPGA fabric clock driven by the clock signal on the rx_clkout port FPGA fabric clock driven by the clock signal on the rx_coreclk port x4 bonded channel configuration FPGA fabric clock driven by the clock signal on the coreclkout port FPGA fabric clock driven by the clock signal on the rx_coreclk port x8 bonded channel configuration FPGA fabric clock driven by the clock signal on the coreclkout port FPGA fabric clock driven by the clock signal on the rx_coreclk port Note to Table 1-38: (1) The clock signal driven on the rx_coreclk port must have 0 PPM frequency difference with respect to the receiver phase compensation FIFO write clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-100 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Receiver Phase Compensation FIFO Error Flag An optional rx_phase_comp_fifo_error port is available in all functional modes to indicate a receiver phase compensation FIFO under-run or overflow condition. The rx_phase_comp_fifo_error signal is asserted high when the phase compensation FIFO gets either full or empty. This feature is useful to verify a phase compensation FIFO under-run or overflow condition as a probable cause of link errors. CMU Channel Architecture Stratix IV GX and GT devices contain two CMU channels--CMU0 and CMU1--within each transceiver block that you can configure as a transceiver channel or as a clock generation block. In addition, each CMU channel contains a CMU PLL that provides clocks to the transmitter channels within the same transceiver block. Figure 1-87 shows the two CMU channels in a transceiver block. Figure 1-87. CMU Channels in a Transceiver Block Stratix IV GX Transceiver Block High-speed serial clock from xN top High-speed serial clock from xN bottom To Transmitter PMA High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider Block To Transmitter PCS High-speed serial clock from xN top High-speed serial clock from xN bottom Transmitter Channel 2 Transmitter Channel 3 Input Reference Clocks (2) CMU1 Channel CMU1 PLL High-Speed Clock CMU0 PLL High-Speed Clock Input Reference Clocks (2) CMU0 Channel High-Speed SERIAL Clock (1) Low-Speed PARALLEL Clock (1) Low-speed parallel clock from xN top Low-speed parallel clock from xN bottom To Transmitter PMA High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider Block To Transmitter PCS Low-speed parallel clock from xN top Low-speed parallel clock from xN bottom Transmitter Channel 0 Transmitter Channel 1 Notes to Figure 1-87: (1) Clocks are provided to support bonded channel functional mode. (2) For more information, refer to the Transceiver Clocking in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-101 The following sections describe the CMU channel building blocks. Configuring CMU Channels for Clock Generation Each CMU channel has a CMU PLL that generates high-speed serial transceiver clocks when the CMU channel is configured as a CMU. The CMU0 clock divider block also generates the low-speed parallel transceiver clock for x4, x8, and xN bonded mode configurations such as XAUI, Basic x4, Basic x8, and Basic (PMA-Direct) xN. The CMU0 channel has additional capabilities to support bonded protocol functional modes such as Basic x4, XAUI, and PCIe. Use the ALTGX MegaWizard Plug-In Manager to select these functional modes (to enable Basic x4 functional mode, select the x4 option in Basic mode). For more information, refer to "Functional Modes" on page 1-110. 1 For Stratix IV GT devices, you can use the CMU PLL to generate transceiver clocks at data rates between 600 Mbps and 11.3 Gbps. CMU0 Channel The CMU0 channel, shown in Figure 1-88, contains the following blocks: CMU0 PLL (refer to "CMU0 Clock Divider" on page 1-103) CMU0 clock divider (refer to "CMU Clock Divider" on page 1-108) Figure 1-88. CMU0 Channel with the CMU0 PLL and CMU0 Clock Divider pll_locked CMU0 Channel pll_powerdown CMU0 PLL High-Speed Clock (1) PLL Cascade Clock Global Clock Line Dedicated refclk0 Dedicated refclk1 CMU0 PLL input reference clock High-Speed Serial Clock for Bonded Modes (2) CMU0 PLL CMU1 PLL High-Speed Clock CMU0 Clock Divider Low-Speed Parallel Clock for Bonded Modes ITB Clock Lines 6 PCIE_gen2switch (to PCIe rate switch controller block in the CCU) PCIE_gen2switch_done (to PCIe rate switch controller block in the CCU) Notes to Figure 1-88: (1) To provide clocks for its PMA and PCS blocks in non-bonded functional modes (for example, GIGE functional mode), the transmitter channel uses the transmitter local clock divider to divide this high-speed clock output. (2) Used in XAUI, Basic x4, and PCIe x4 functional modes. In PCIe x8 functional mode, only the CMU0 channel of the master transceiver block provides clock output to all eight transceiver channels configured in PCIe functional mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-102 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture CMU0 PLL Figure 1-89 shows the CMU0 PLL. Figure 1-89. CMU0 PLL CMU0 PLL /M PLL Cascade Clock Global Clock Line CMU0 PLL Input Reference Clock Dedicated refclk0 /1, /2, /4, /8 PFD Charge Pump + Loop Filter VCO /L CMU0 High-Speed Clock Dedicated refclk1 ITB Clock Lines (1) 6 Note to Figure 1-89: (1) The inter transceiver block (ITB) clock lines are the maximum value. The actual number of ITB lines in your device depends on the number of transceiver blocks on one side of your device. You can select the input reference clock to the CMU0 PLL from multiple clock sources: PLL cascade clock--the output from the general purpose PLLs in the FPGA fabric Global clock line--the input reference clock from the dedicated CLK pins are connected to the global clock line refclk0--dedicated REFCLK in the transceiver block refclk1--dedicated REFCLK in the transceiver block Inter transceiver block lines--the ITB lines connect refclk0 and refclk1 of all other transceiver blocks on the same side of the device The CMU0 PLL generates the high-speed clock from the input reference clock. The PFD tracks the VCO output with the input reference clock. f For more information about transceiver input reference clocks, refer to the Transceiver Clocking in Stratix IV Devices chapter. The VCO in the CMU0 PLL is half rate and runs at half the serial data rate. To generate the high-speed clock required to support a native data rate range of 600 Mbps to 8.5 Gbps, the CMU0 PLL uses two multiplier blocks (/M and /L) in the feedback path (shown in Figure 1-89). 1 The ALTGX MegaWizard Plug-In Manager provides the list of input reference clock frequencies based on the data rate you select. The Quartus II software automatically selects the /M and /L settings based on the input reference clock frequency and serial data rate. f The CMU0 and CMU1 PLLs have a dedicated pll_locked signal that is asserted to indicate that the CMU PLL is locked to the input reference clock. You can use the pll_locked signal in your transceiver reset sequence, as described in the Reset Control and Power Down in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-103 PLL Bandwidth Setting The bandwidth of a PLL is the measure of its ability to track input clock and jitter. It is determined by the -3 dB frequency of the closed-loop gain of the PLL. There are three bandwidth settings: high, medium, and low. You can program the PLL bandwidth setting using the ALTGX MegaWizard Plug-In Manager. The high bandwidth setting filters out internal noise from the VCO because it tracks the input clock above the frequency of the internal VCO noise. With the low bandwidth setting, if the noise on the input reference clock is greater than the internal noise of the VCO, the PLL filters out the noise above the -3 dB frequency of the closed-loop gain of the PLL. The medium bandwidth setting is a compromise between the high and low bandwidth settings. The -3 dB frequencies for these settings can vary because of the non-linear nature and frequency dependencies of the circuit. Power Down CMU0 PLL You can power down the CMU0 PLL by asserting the pll_powerdown signal. f For more information, refer to the Reset Control and Power Down in Stratix IV Devices chapter. CMU0 Clock Divider The high-speed clock output from the CMU0 PLL is forwarded to two clock dividers: the CMU0 clock divider and the transmitter channel local clock divider. Use the clock divider only in bonded channel functional modes. In non-bonded functional modes (such as GIGE functional mode), the local clock divider divides the high-speed clock to provide clocks for its PCS and PMA blocks. This section only describes the CMU0 clock divider. f For more information about the local clock divider, refer to the "Transceiver Channel Datapath Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. You can configure the CMU0 clock divider shown in Figure 1-90 to select the high-speed clock output from the CMU0 or CMU1 PLLs. The CMU1 PLL is present in the CMU1 channel. Figure 1-90. CMU0 Clock Divider PCIE_gen2switch_done CMU0 Clock Divider Block Low-Speed Parallel Clock for Transmitter Channel PCS (for Bonded Modes) PCIE_gen2switch CMU0 High-Speed Clock Output /N (1, 2, 4) CMU1 High-Speed Clock Output PCIE rateswitch circuit 0 /S (4, 5, 8, 10) coreclkout to FPGA Fabric (for Bonded Modes) /2 1 High-Speed Serial Clock (for Bonded Modes) September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-104 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture High-Speed Serial Clock Generation The /N divider receives the high-speed clock output from one of the CMU PLLs and produces a high-speed serial clock. Use this clock for bonded functional modes such as Basic x4/x8, XAUI, and PCIe x4/x8 configurations. In XAUI and Basic x4/x8 modes, the Quartus II software chooses the path (shown by "1" in the MUX) and provides the high-speed serial clock to all the transmitter channels within the transceiver block. In PCIe x4 mode, the clock path through the PCIe rateswitch circuit block is selected. This high-speed serial clock is provided to all the transmitter channels. In PCIe x8 mode and Basic x8 mode, only the CMU0 clock divider of the master transceiver block provides the high-speed serial clock to all eight channels. In PCIe x1 mode, the CMU0 clock divider does not provide a high-speed serial clock. Instead, the local clock divider block in the transmitter channel receives the CMU0 or CMU1 PLL high-speed clock output and generates the high-speed serial clock to its serializer. f For more information about the clock from the master transceiver block, refer to the Transceiver Clocking in Stratix IV Devices chapter. PCIE Rateswitch Circuit The PCIE rateswitch circuit is enabled only in PCIe x4 mode. In PCIe x8 mode, the PCIE rateswitch circuit of the CMU0 clock divider of the master transceiver block is active. There are two paths in the PCIE rateswitch circuit. One path divides the /N output by two. The other path forwards the /N divider output. 1 When you set the rateswitch port to 0, the PCIe rateswitch controller (in the CCU) signals the PCIE rateswitch circuit to select the divide by /2 to provide a high-speed serial clock for the Gen1 (2.5 Gbps) data rate. When you set the rateswitch port to 1, the /N divider output is forwarded, providing a high-speed serial clock for the Gen2 (5 Gbps) data rate to the transmitter channels. The PCIE rateswitch circuit performs the rateswitch operation only for the transmitter channels. For the receiver channels, the rateswitch circuit within the receiver CDR performs the rateswitch operation. The PCIE rateswitch circuit is controlled by the PCIe rateswitch controller in the CCU. The PCIe rateswitch controller asserts the pipephydonestatus signal for one clock cycle after the rateswitch operation is completed for both the transmit and receive channels. Figure 1-91 shows the timing diagram for the rateswitch operation. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-105 For more information about PCIe functional mode rate switching, refer to "PCIe Gen2 (5 Gbps) Support" on page 1-140. Figure 1-91. Rateswitch in PCIe Mode (1) 250 MHz (Gen 1) 250 MHz (Gen 1) 500 MHz (Gen 2) Low-Speed Parallel Clock rateswitch pipephydonestatus T1 T1 Note to Figure 1-91: (1) Time T1 is pending characterization. 1 When creating a PCIe Gen2 configuration, configure the CMU PLL to 5 Gbps. This helps to generate the 2.5 Gbps and 5 Gbps high-speed serial clock using the rateswitch circuit. Low-Speed Parallel Clock Generation The /S divider receives the clock output from the /N divider or PCIE rateswitch circuit (only in PCIe mode) and generates the low-speed parallel clock for the PCS block of all transmitter channels and coreclkout for the FPGA fabric. If the byte serializer block is enabled in bonded channel modes, the /S divider output is divided by the /2 divider and sent out as coreclkout to the FPGA fabric. The Quartus II software automatically selects the /S values based on the deserialization width setting (single-width or double-width mode) that you select in the ALTGX MegaWizard Plug-In Manager. For more information about single-width or double-width mode, refer to "Transceiver Channel Architecture" on page 1-17. 1 September 2012 The Quartus II software automatically selects all the divider settings based on the input clock frequency, data rate, deserialization width, and channel width settings. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-106 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture CMU1 Channel The CMU1 channel, shown in Figure 1-92, contains the CMU1 PLL that provides the high-speed clock to the transmitter channels within the transceiver block. The CMU1 PLL is similar to the CMU0 PLL (refer to "CMU0 PLL" on page 1-102). Figure 1-92. CMU1 Channel (Grayed Area Shows the Inactive Block) pll_locked CMU1 PLL High-Speed Clock pll_powerdown PLL Cascade Clock Global Clock Line Dedicated refclk0 Dedicated refclk1 CMU1 PLL Input Reference Clock CMU1 PLL CMU1 Clock Divider ITB Clock Lines 6 CMU1 Channel The CMU1 PLL generates the high-speed clock that is only used in non-bonded functional modes. The transmitter channels within the transceiver block can receive a high-speed clock from either of the two CMU PLLs and uses local dividers to provide clocks to its PCS and PMA blocks. f For more information about using two CMU PLLs to configure transmitter channels, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Power Down CMU1 PLL You can power down the CMU1 PLL by asserting the pll_powerdown signal. f For more information, refer to the Reset Control and Power Down in Stratix IV Devices chapter. Configuring CMU Channels as Transceiver Channels You can configure the two CMU channels in the transceiver block of Stratix IV GX and GT devices as full-duplex PMA-only channels to run between 600 Mbps and 6.5 Gbps. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-107 Figure 1-93 shows the functional blocks that are enabled to support the transceiver channel functionality. Figure 1-93. Functional Blocks Enabled to Support Transceiver Channel Functionality From xN From xN top bottom clock line clock line High-speed clock from the adjacent CMU channel x4 clock line From the FPGA fabric To the FPGA fabric CMU Channel CMU Clock divider block (/1, /2, /4) serializer tx_data_out CMU PLL CONFIGURED AS RX CDR rx_data_in 1 deserializer The CMU PLL is configured as a CDR to recover data. The dedicated input reference clock pin is configured to receive serial data. When configured as a full-duplex or receiver-only channel, the CMU PLL performs the functionality of the receiver CDR and recovers clock from the incoming serial data. The high-speed serial and low-speed parallel recovered clocks are used by the deserializer in the CMU channel and the deserialized data is forwarded directly to the FPGA fabric. When configured as a full-duplex or transmitter-only channel, the serializer in the CMU channel serializes the parallel data from the FPGA fabric and drives the serial data to the transmitter buffer. Table 1-39 lists the pins that are used as transmit and receive serial pins. Table 1-39. Transmit and Receive Serial Pins (Part 1 of 2) Pins (1) REFCLK_[L,R][0,2,4,6]P, GXB_CMURX_[L_R][0,2,4,6]P GXB_TX_[L,R][0,2,4,6] (2) REFCLK_[L,R][1,3,5,7]P, GXB_CMURX_[L_R][1,3,5,7]P September 2012 (2) Altera Corporation (3) When a CMU Channel is Configured as a Transceiver Channel When a CMU Channel is Configured for Clock Generation Receive serial input for CMU Channel0 Input reference clocks Transmit serial output for CMU Channel0 Not available for use Receive serial input for CMU Channel1 Input reference clocks Stratix IV Device Handbook Volume 2: Transceivers 1-108 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-39. Transmit and Receive Serial Pins (Part 2 of 2) Pins (1) GXB_TX_[L,R][1,3,5,7]P (3) When a CMU Channel is Configured as a Transceiver Channel When a CMU Channel is Configured for Clock Generation Transmit serial output for CMU Channel1 Not available for use Notes to Table 1-39: (1) These indexes are for the Stratix IV GX and GT device with the maximum number of transceiver blocks. For exact information about how many of these pins are available for a specific device family, refer to the Overview for the Stratix IV Device Family chapter. (2) Pins 0,2,4,6 are hardwired to CMU channel0 in the corresponding transceiver blocks. (3) Pins 1,3,5,7 are hardwired to CMU channel1 in the corresponding transceiver blocks. Interpret the pins column as follows: For pins REFCLK_[L,R][0,2,4,6]P and GXB_CMURX_[L_R][0,2,4,6], the "L, R" indicates the left and right side and the "0, 2, 4, 6" indicates the different pins. For example, a pin on the left side with index 0 is named: REFCLK_L0P, GXB_CMURX_L0P. 1 The receiver serial input pins are hardwired to their corresponding CMU channels. For more information, refer to the notes to Table 1-39. Serializer and Deserializer The serializer and deserializer convert the serial-to-parallel data on the transmitter and receiver side, respectively. The ALTGX MegaWizard Plug-In Manager provides the Basic (PMA Direct) functional mode (with a none and xN option) to configure a transceiver channel to enable the transmitter serializer and receiver deserializer. To configure a CMU channel as a transceiver channel, you must use this functional mode. The input data width options to serializer/from deserializer for a channel configured in this mode are 8, 10, 16, and 20. f To understand the maximum FPGA fabric-transceiver interface clock frequency limits, refer to the "Transmitter Channel Datapath Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. CMU Clock Divider When you configure a CMU channel in Basic (PMA Direct) x1 mode, the CMU clock divider divides the high-speed clock from the other CMU channel (used as a clock generation unit) within the same transceiver block and provides the high-speed serial clock and low-speed parallel clocks to the transmitter side of the CMU channel. The CMU clock divider can divide the high-speed clock by /1, /2, and /4. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-109 Clocks for the Transmitter Serializer When you configure the CMU channel as a transceiver channel, the clocks for the transmitter side is provided by one of these sources: The other CMU channel in the same transceiver block that is configured as a clock multiplication unit From CMU channel0 on the other transceiver block on the same side of the device through the xN clock line (the xN_Top or xN_Bottom clock line). If you configure a CMU channel in Basic (PMA Direct) xN mode, you can use this clocking option From one of the ATX PLL blocks on the same side of the device through the xN clock line (the xN_Top or xN_Bottom clock line) Input Reference Clocks for the Receiver CDR When you configure a CMU Channel as a transceiver channel, there are multiple sources of input reference clocks for the receiver CDR: From adjacent REFCLKs within the same transceiver block, if the adjacent CMU Channel is not used as a transceiver channel From the REFCLK of adjacent transceiver blocks on the same side of the device, if the corresponding CMU channels are not used as transceiver channels. For REFCLK connections to the CMU channel from the global clock lines and PLL cascade network, refer to Table 1-6 on page 1-18. f For more information, refer to the "Input Reference Clocking" section of the Transceiver Clocking in Stratix IV Devices chapter. Clocks for the Receiver Deserializer The CDR provides high-speed serial and low-speed parallel clocks to the receiver deserializer from the recovered data. Other CMU Channel Features The CMU channels provide the following features: Analog control options--Differential output voltage (VOD), pre-emphasis, equalization, and DC gain settings present in the regular channels are also available in the CMU channels. OCT--CMU channels can have an OCT feature. The allowed termination values are the same as regular channels (85, 100, 120, and 150 ). Loopback--The available loopback options are serial, reverse serial (pre-CDR), and reverse serial (CDR) loopback. For more information about analog controls and OCT, refer to "Transmitter Output Buffer" on page 1-34 and "Receiver Input Buffer" on page 1-40. For information about loopback, refer to "Loopback Modes" on page 1-190. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-110 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Dynamic Reconfiguration of the CMU Channel Analog Controls f For the dynamic reconfiguration capabilities of the CMU channels in Basic (PMA Direct) x1/xN configurations, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. Functional Modes Table 1-40 lists the transceiver functional modes you can use to configure the Stratix IV GX and GT devices using the ALTGX MegaWizard Plug-In Manager. Table 1-40. Functional Modes for the Stratix IV GX and GT Devices Functional Mode Data Rate Refer To Basic Single Width 600 Mbps to 3.75 Gbps "Basic Single-Width Mode Configurations" on page 1-113 Basic Double Width 1 Gbps to 8.5 Gbps "Basic Double-Width Mode Configurations" on page 1-117 PCIe Gen1 at 2.5 Gbps Gen2 at 5 Gbps "PCIe Mode Configurations" on page 1-128 XAUI 3.125 Gbps up to HiGig at 3.75 Gbps "XAUI Mode Datapath" on page 1-157 GIGE 1.25 Gbps "GIGE Mode Datapath" on page 1-167 Serial RapidIO SONET/SDH SDI (OIF) CEI PHY Interface 1.25 Gbps 2.5 Gbps 3.125 Gbps OC-12 OC-48 HD at 1.485/1.4835 Gbps 3G at 2.97/2.967 Gbps "Serial RapidIO Mode" on page 1-182 "SONET/SDH Frame Structure" on page 1-172 >4.976 Gbps to 6.375 Gbps "SDI Mode Datapath" on page 1-180 "(OIF) CEI PHY Interface Mode Datapath" on page 1-182 Table 1-41 lists the transceiver functional modes you can use to configure the Stratix IV GT devices using the ALTGX MegaWizard Plug-In Manager. Table 1-41. Functional Modes for the Stratix IV GT Devices (Part 1 of 2) Functional Mode Data Rate Refer To Basic Single Width 600 Mbps to 3.75 Gbps "Basic Single-Width Mode Configurations" on page 1-113 Basic Double Width 1.0 to 11.3 Gbps "Basic Double-Width Mode Configurations" on page 1-117 Basic (PMA-Direct) single-width 600 Mbps to 3.25 Gbps "Basic (PMA Direct) x1 Configuration" on page 1-189 Basic (PMA-Direct) double-width 1.0 to 6.5 Gbps "Basic (PMA Direct) xN Configuration" on page 1-189 XAUI 3.125 Gbps up to HiGig at 3.75 Gbps "XAUI Mode Datapath" on page 1-157 GIGE 1.25 Gbps "GIGE Mode Datapath" on page 1-167 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-111 Table 1-41. Functional Modes for the Stratix IV GT Devices (Part 2 of 2) Functional Mode Data Rate Refer To 2.5 Gbps 3.125 Gbps OC-48 OC-96 SDI 3G at 2.97/2.967 Gbps (OIF) CEI PHY Interface > 4.976 Gbps to 6.375 Gbps Serial RapidIO SONET/SDH "Serial RapidIO Mode" on page 1-182 "SONET/SDH Frame Structure" on page 1-172 "SDI Mode Datapath" on page 1-180 "(OIF) CEI PHY Interface Mode Datapath" on page 1-182 Basic Functional Mode The Stratix IV GX and GT transceiver datapaths are extremely flexible in Basic functional mode. To configure the transceivers in Basic functional mode, you must select Basic in the Which protocol will you be using? option of the ALTGX MegaWizard Plug-In Manager. Basic functional mode can be further sub-divided into the following two functional modes: Basic single-width mode Basic double-width mode You can configure the transceiver in Basic single-width mode by selecting Single in the What is the deserializer block width? option in the ALTGX MegaWizard Plug-In Manager. You can configure the transceiver in Basic double-width mode by selecting Double in the What is the deserializer block width? option in the ALTGX MegaWizard Plug-In Manager. Table 1-42 lists the Stratix IV GX and GT PCS-PMA interface widths and data rates supported in Basic single-width and double-width modes. Table 1-42. PCS-PMA Interface Widths and Data Rates in Basic Single-Width and Double-Width Modes for Stratix IV GX and GT Devices Basic Functional Mode Supported Data Rate Range (1) PMA-PCS Interface Width Basic single-width mode 600 Mbps to 3.75 Gbps 8 bit, 10 bit Basic double-width mode 1 Gbps to 8.5 Gbps 16 bit, 20 bit Note to Table 1-42: (1) The data rate range supported in Basic single-width and double-width modes varies depending on whether or not you use the byte serializer/deserializer. For more information, refer to "Basic Single-Width Mode Configurations" on page 1-113 and "Basic Double-Width Mode Configurations" on page 1-117. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-112 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-43 lists the Stratix IV GT PCS-PMA interface widths and data rates supported in Basic single-width and double-width modes. Table 1-43. PCS-PMA Interface Widths and Data Rates Supported in Basic Single-Width and Double-Width Modes for Stratix IV GT Devices (1) Basic Functional Mode Supported Data Rate Range PMA-PCS Interface Width Basic single-width mode 600 Mbps to 3.75 Gbps 8-bit, 10-bit Basic double-width mode 1.0 to 11.3 Gbps 16-bit, 20-bit Note to Table 1-43: (1) The data rate range supported in Basic single-width and double-width modes varies depending on whether or not you use the byte serializer/deserializer. For more information, refer to"Basic Single-Width Mode Configurations" on page 1-113 and "Basic Double-Width Mode Configurations" on page 1-117. Low Latency PCS Datapath The ALTGX MegaWizard Plug-In Manager provides an Enable low latency PCS mode option when configured in Basic single-width or Basic double-width mode. If you select this option, the following transmitter and receiver channel PCS blocks are bypassed to yield a low latency PCS datapath: 8B/10B encoder and decoder Word aligner Deskew FIFO Rate match (clock rate compensation) FIFO Byte ordering In low latency PCS modes, the transmitter and receiver phase compensation FIFOs are always enabled. Depending on the targeted data rate, you can optionally bypass the byte serializer and deserializer blocks. For more information, refer to "Basic SingleWidth Mode Configurations" on page 1-113 and "Basic Double-Width Mode Configurations" on page 1-117. 1 The PCS latency in Basic single-width and Basic double-width modes with and without the low latency PCS mode option is pending characterization. 1 Basic double-width mode configurations at data rates of > 6.5 Gbps are only allowed in low-latency PCS bypass mode. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-113 Basic Single-Width Mode Configurations Figure 1-94 shows Stratix IV GX transceiver configurations allowed in Basic single-width functional mode with an 8-bit PMA-PCS interface. Figure 1-95 shows Stratix IV GT transceiver configurations allowed in Basic single-width functional mode with an 8-bit PMA-PCS interface. Figure 1-94. Transceiver Configurations in Basic Single-Width Mode with an 8-Bit PMA-PCS Interface for Stratix IV GX Devices Stratix IV GX Configurations Protocol Basic Functional Modes PMA-PCS Interface Width (1) Double Width Single Width 8-bit 10-Bit 16-Bit 20-Bit XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-Bit 10-Bit 10-Bit 10-Bit 8-Bit 16-Bit 10-Bit PMA-PCS Interface Width Basic Single-Width 8-Bit PMA-PCS Interface Width Data Rate (Gbps) 0.6 - 3.2 Channel Bonding x1, x4, x8 Low-Latency PCS Word Aligner (Pattern Length) Deterministic Latency PIPE 10-Bit Disabled Enabled Bit-Slip (16-Bit) Disabled Disabled Disabled Manual Alignment (16-Bit) 8B/10B Encoder/Decoder Disabled Disabled Rate Match FIFO 20-Bit Disabled Disabled Byte SerDes Disabled Enabled Disabled Enabled Disabled Enabled Data Rate (Gbps) 0.6 2.0 0.6 3.125 0.6 2.0 0.6 3.125 0.6 2.0 0.6 3.2 Byte Ordering Disabled Disabled Disabled Disabled Disabled Disabled FPGA Fabric-Transceiver Interface Width 8-Bit 16-Bit 16-Bit 8-Bit 16-Bit 8-Bit 16-Bit FPGA FPGA Fabric Fabric-Transceiver Transceiver Interface Frequency Interface Frequency (MHz) 75 250 37.5 195.3125 37.5 195.3125 75 250 37.5 195.3125 75 250 37.5 200 TX PCS Latency FPGA Fabric (FPGA Fabric-Transceiver Transceiver Interface Clock Cycles) Interface Frequency 5 -6 4 - 5.5 4 - 5.5 5 -6 4 - 5.5 4-5 4 - 5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 11-13 7-9 7-9 11 - 13 7-9 3-4 3 - 4.5 (1) Enabled Note to Figure 1-94: (1) The maximum data rate specification shown in Figure 1-94 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-114 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-95. Transceiver Configurations in Basic Single-Width Mode with an 8-Bit PMA-PCS Interface for Stratix IV GT Devices Stratix IV GT Configurations Protocol Basic Functional Modes PMA-PCS/Fabric Interface Width Single Width 8-bit Double Width 10-bit 16-bit 20-bit SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Data Rate 600 Mbps to 3.2 Gbps Channel Bonding x1, x4, x8 Low-Latency PCS (1) XAUI Disabled Manual Alignment (16-bit) Bit-Slip (16-bit) Disabled 8B/10B Encoder/Decoder Disabled Disabled Disabled Rate Match FIFO Disabled Disabled Disabled Byte SerDes Enabled Enabled Enabled 0.6 3.125 0.6 3.125 0.6 - 3.2 Data Rate (Gbps) Byte Ordering Disabled Enabled Disabled Disabled FPGA FabricTransceiver Interface Width 16-bit 16-bit 16-bit 16-bit 155.5 - 195 .3125 155.5 - 195.3125 155.5 - 195.3125 155.5 200 4 - 5.5 4 - 5.5 4 - 5.5 4 - 5.5 7-9 7-9 7-9 FPGATX Fabric PCS Latency (FPGA Fabric-Transceiver 10-Bit 20-Bit Enabled Word Aligner (Pattern Length) FPGA FPGA FabricFabricTransceiver Transceiver Interface Interface Frequency Frequency (MHz) Deterministic Latency Basic Single Width 8-bit PMA-PCS Interface Width PMA-PCS/Fabric Interface Width (1) PIPE Interface Clock Cycles) Interface Frequency RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 3 - 4.5 Note to Figure 1-95: (1) The maximum data rate specification shown in Figure 1-95 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-115 Figure 1-96 shows Stratix IV GX transceiver configurations allowed in Basic single-width functional mode with a 10-bit PMA-PCS interface. Figure 1-97 shows Stratix IV GT transceiver configurations allowed in Basic single-width functional mode with a 10-bit PMA-PCS interface. Figure 1-96. Transceiver Configurations in Basic Single-Width Mode with a 10-Bit PMA-PCS Interface for Stratix IV GX Devices Stratix IV GX Configurations Protocol Basic Functional Modes PMA-PCS Interface Width Single Width 8-Bit Double Width 10-Bit 16-Bit 20-Bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-Bit 10-Bit 10-Bit 10-Bit 8-Bit 16-Bit 10-Bit 10-Bit 20-Bit Basic Single-Width 10-Bit PMA-PCS Interface Width PMA-PCS Interface Width (1) Deterministic Latency 0.6 - 3.75 Data Rate (Gbps) x1, x4, x8 Channel Bonding Low-Latency PCS Disabled Word Aligner (Pattern Length) Enabled Automatic Synchronization State Machine (7-Bit, 10-Bit) Bit-Slip (7-Bit, 10-Bit) Manual Alignment (7-Bit, 10-Bit) 8B/10B Encoder/Decoder Disabled Enabled Disabled Enabled Disabled Rate Match FIFO Disabled Disabled Disabled Disabled Disabled Disabled Disabled Enabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled 0.6 3.75 0.6 2.5 0.6 3.75 0.6 2.5 0.6 3.75 Enabled Disabled Disabled Disabled Disabled 16-Bit 8-Bit 16-Bit 10-Bit 20-Bit 60 250 30 187.5 4-5 4 - 5.5 3-4 3 - 4. 5 Byte SerDes Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Data Rate (Gbps) 0.6 2.5 0.6 3.75 0.6 2.5 0.6 3.75 0.6 2.5 0.6 3.75 0.6 2.5 0.6 3.75 0.6 2.5 0.6 3.75 Byte Ordering Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled FPGA Fabric-Transceiver Interface Width 10-Bit 20-Bit 8-Bit 16-Bit 10-Bit 20-Bit 8-Bit 16-Bit 10-Bit 20-Bit 8-Bit 16-Bit 60 250 30 187.5 60 250 30 187.5 60 250 30 187.5 60 250 30 187.5 60 250 30 187.5 60 250 30 187.5 30 187.5 60 250 30 187.5 TX PCS Latency FPGA Fabric (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 5-6 4 - 5.5 5-6 4 - 5.5 5-6 4 - 5.5 5-6 4-5 5-6 4 - 5.5 5-6 4 - 5.5 4 - 5.5 5-6 4 - 5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 9 - 11 6-8 9 - 11 6-8 9 - 11 6-8 9 - 11 6-8 9 - 11 6-8 9 - 11 6-8 6-8 20 24 11.5 14.5 (1) FPGA FPGA Fabric Fabric-Transceiver Transceiver Interface Frequency Frequency Interface (MHz) Enabled Disabled Disabled 0.6 2.5 Note to Figure 1-96: (1) The maximum data rate specification shown in Figure 1-96 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-116 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-97. Transceiver Configurations in Basic Single-Width Mode with a 10-Bit PMA-PCS Interface for Stratix IV GT Devices Stratix IV GT Configurations Protocol Basic Functional Modes PMA-PCS/Fabric Interface Width (1) Single Width 8-bit 10-bit Double Width 16-bit 20-bit 8 XAUI SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit PMA-PCS Interface Width Basic Single Width 10-bit PMA-PCS Interface Width Data Rate 600 Mbps - 3.75 Gbps Channel Bonding x1, x4, x8 Low-Latency PCS Deterministic Latency 10-Bit 20-Bit Enabled Disabled Word Aligner (Pattern Length) Manual Alignment (7-bit, 10-bit) Automatic Synchronization State Machine (7-bit, 10-bit) Bit-Slip (7-bit, 10-bit) 8B/10B Encoder/Decoder Disabled Enabled Disabled Enabled Disabled Rate Match FIFO Disabled Disabled Disabled Disabled Disabled Disabled Disabled Enabled Enabled Disabled Disabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Data Rate (Gbps) 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 0.6 - 2.5 0.6 - 3.75 Byte Ordering Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled Enabled Disabled Disabled Disabled Disabled FPGA FabricTransceiver Interface Width 10-bit 20-bit 8-bit 16-bit 10-bit 20-bit 8-bit 16-bit 10-bit 20-bit 8-bit 16-bit 16-bit 8-bit 16-bit 10-bit 20-bit 248 .8 250 124.4 - 187.5 248 .8 250 124 .4 - 187.5 248.8 250 124.4 - 187.5 248.8 250 124.4 - 187.5 248.8 250 124 .4 - 187.5 248.8 250 124.4 - 187 .5 124.4 - 187.5 248.8 250 124.4 - 187 .5 248.8 250 5-6 4 - 5.5 5-6 4 - 5.5 5-6 4-5 5-6 4 - 5.5 5-6 4 - 5.5 4 - 5.5 5-6 4 - 5.5 4-5 6-8 9 - 11 6-8 9 - 11 6-8 9 - 11 6-8 6-8 20 24 Byte SerDes (1) PIPE FPGA FPGA FabricFabricTransceiver Transceiver Interface Frequency Interface Frequency (MHz) TX PCS Latency FPGA Fabric (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 5-6 4 - 5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 9 - 11 6-8 4-5 9 - 11 6-8 9 - 11 11.5 14.5 44 -- 55 3-4 124.4 - 187.5 4 - 5.5 4 - 5.5 5.5 44 -- 5.5 3 - 4. 5 Note to Figure 1-97: (1) The maximum data rate specification shown in Figure 1-97 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-117 Basic Double-Width Mode Configurations Figure 1-98 shows Stratix IV GX transceiver configurations allowed in Basic double-width functional mode with a 16-bit PMA-PCS interface. Figure 1-99 shows Stratix IV GT transceiver configurations allowed in Basic double-width functional mode with a 16-bit PMA-PCS interface. Figure 1-98. Transceiver Configurations in Basic Double-Width Mode with a 16-Bit PMA-PCS Interface for Stratix IV GX Devices Stratix IV GX Configurations Protocol Basic Functional Modes PMA-PCS Interface Width SingleWidth 8-Bit DoubleWidth 10-Bit 16-Bit 20-Bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-Bit 10-Bit 10-Bit 10-Bit 8-Bit 16-Bit 10-Bit 20-Bit 1.0 - 8.5 Data Rate (Gbps) Channel Bonding x1, x4, x8 Disabled Low-Latency PCS Enabled Word Aligner (Pattern Length) Manual Alignment (8-, 16-, 32-Bit) 8B/10B Encoder/Decoder Disabled Disabled Disabled Rate Match FIFO Disabled Disabled Disabled Byte SerDes Disabled Data Rate (Gbps) 1.0 4.0 Byte Ordering Disabled Disabled FPGA Fabric-Transceiver Interface Width 16-Bit 32-Bit FPGA FPGA Fabric Fabric-Transceiver Transceiver Interface Frequency Interface Frequency (MHz) 62.5 250 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) (1) 10-Bit Basic Double-Width 16-Bit PMA-PCS Interface Width PMA-PCS Interface Width (1) Deterministic Latency Bit-Slip (8-, 16-, 32-Bit) Enabled Disabled Disabled Enabled Disabled Enabled 1.0 4.0 1.0 6.5 1.0 4.0 1.0 8.5 Disabled Disabled Disabled Disabled 32-Bit 16-Bit 32-Bit 16-Bit 32-Bit 31.25 203.125 31.25 203.125 62.5 250 31.25 203.125 62.5 250 31.25 265.625 5-6 4 - 5.5 4 - 5.5 5-6 4 - 5.5 4-5 4 - 5.5 11 - 13 6.5 - 8.5 6.5 - 8.5 11 - 13 6.5 - 8.5 3-4 3 - 4.5 1.0 6.5 Enabled (2) Notes to Figure 1-98: (1) The maximum data rate specification shown in Figure 1-98 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) The byte ordering block is available only if you select the word alignment pattern length of 16 or 32 bits. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-118 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-99. Transceiver Configurations in Basic Double-Width Mode with a 16-Bit PMA-PCS Interface for Stratix IV GT Devices Stratix IV GT Configurations Protocol Basic Functional Modes PMA-PCS/Fabric Interface Width (1) Single Width 8-bit Double Width 10-bit 16-bit 20-bit PIPE XAUI SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit PMA-PCS Interface Width Basic Double Width 16-bit PMA-PCS Interface Width Data Rate (Gbps) 1.0 - 8.5 Channel Bonding x1, x4, x8 Low-Latency PCS Disabled Word Aligner (Pattern Length) Manual Alignment (8-, 16-, 32-bit) Bit-Slip (8-, 16-, 32-bit) Disabled 8B/10B Encoder/Decoder Disabled Disabled Disabled Rate Match FIFO Disabled Disabled Disabled Disabled Enabled Disabled Enabled Disabled Enabled Data Rate (Gbps) 1.0 - 4.0 1.0 - 6.5 1.0 - 4.0 1.0 - 6.5 1.0 - 4.0 1.0 - 8.5 Byte Ordering Disabled Disabled Enabled (Note 1) Disabled Disabled Disabled Disabled FPGA FabricTransceiver Interface Width 16-bit 32-bit 32-bit 16-bit 32-bit 16-bit 32-bit 155 .5 250 77.75 - 203.125 77.75 - 203.125 155 .5 250 77.75 - 203.125 155.5 250 77.75 - 265.625 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 5-6 4 - 5.5 4 - 5.5 5-6 4 - 5.5 4-5 4 - 5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 11 - 13 6.5 - 8.5 6.5 - 8.5 11 - 13 6.5 - 8.5 3-4 3 - 4.5 FPGA FPGA FabricFabricTransceiver Transceiver Interface Frequency Interface Frequency (MHz) 10-Bit 20-Bit Enabled Byte SerDes (1) Deterministic Latency Note to Figure 1-99: (1) The maximum data rate specification shown in Figure 1-99 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-119 Figure 1-100 shows Stratix IV GX transceiver configurations allowed in Basic double-width functional mode with a 20-bit PMA-PCS interface. Figure 1-101 shows Stratix IV GT transceiver configurations allowed in Basic double-width functional mode with a 20-bit PMA-PCS interface. Figure 1-100. Transceiver Configurations in Basic Double-Width Mode with a 20-Bit PMA-PCS Interface for Stratix IV GX Devices Stratix IV GX Configurations Protocol Basic Functional Modes PMA-PCS Interface Width Single Width 8-bit Double Width 10-bit 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Data Rate (Gbps ) x1, x4, x8 Low-Latency PCS Enabled Disabled Manual Alignment (7-, 10-, 20-bit) Word Aligner (Pattern Length ) 8B/10B Encoder /Decoder Disabled Rate Match FIFO Disabled Bit-Slip (7-, 10-, 20-bit) Enabled Disabled Enabled Disabled Disabled Disabled Disabled Disabled Enabled 1.0 - 8.5 1.0 - 5.0 1.0 - 8.5 1.0 6.5 1.0 - 8.5 Disabled Disabled Disabled Disabled Disabled Disabled 32-bit 20-bit 40-bit 16-bit 32-bit 20-bit 40-bit 25 212.5 50 325 25 212.5 50 250 25 212.5 50 325 25 212.5 4 - 5.5 5-6 4 - 5.5 5-6 4 - 5.5 10 - 12 6.5 - 8. 5 10 - 12 Enabled Disabled Enabled Data Rate (Gbps) 1.0 6.5 1.0 - 8.5 1.0 - 5.0 1.0 - 8.5 1.0 - 5.0 1.0 - 8.5 Byte Ordering Disabled Disabled Enabled (2) Disabled Disabled Enabled (2) Disabled Disabled FPGA Fabric Transceiver Interface Width 20-bit 40-bit 40-bit 16-bit 32-bit 32-bit 16-bit 50 325 25 212.5 25 212.5 50 250 25 212.5 25 212.5 50 250 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 5-6 4 - 5.5 4 - 5.5 5-6 4 - 5.5 4 - 5.5 5-6 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 10 - 12 6.5 - 8.5 6.5 - 8.5 10 - 12 22 - 26 Enabled Enabled Disabled 6.5 - 8. 5 6.5 - 8. 5 Disabled Disabled Enabled FPGA FPGA Fabric Fabric -Transceiver Transceiver Interface Frequency Interface Frequency ( MHz) Disabled Enabled Disabled (1) 20-Bit 1.0 - 8.5 Channel Bonding Byte SerDes 10-Bit Basic Double Width 20-bit PMA-PCS Interface Width PMA-PCS Interface Width (1) Deterministic Latency 13 - 16 Disabled 1.0 6.5 6.5 - 8. 5 4-5 3-4 4 - 5.5 3 - 4.5 Notes to Figure 1-100: (1) The maximum data rate specification shown in Figure 1-100 is valid only for the -2 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) The byte ordering block is available only if you select the word alignment pattern length of 20 bits. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-120 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-101. Transceiver Configurations in Basic Double-Width Mode with a 20-Bit PMA-PCS Interface for Stratix IV GT Devices Stratix IV GT Configurations Protocol Basic Functional Modes PMA PCS/Fabric Interface Width (1) Single Width 8-bit Double Width 10-bit 16-bit 20-bit Deterministic Latency PIPE XAUI SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit PMA-PCS Interface Width Basic Double Width 20-bit PMA-PCS Interface Width Data Rate 600 Mbps - 11.3 Gbps Channel Bonding 10-Bit 20-Bit x1, x4, x8 Low-Latency PCS Enabled Disabled (2) Manual Alignment (7-, 10-, 20-bit) Word Aligner (Pattern Length ) 8B/10B Encoder /Decoder Disabled Rate Match FIFO Disabled Enabled Disabled Enabled 1.0 8.5 1.0 5.0 1.0 8.5 2.488 6.5 2.488 11.3 Disabled Disabled Disabled Disabled Disabled Disabled Disabled 16-bit 32-bit 20-bit 40-bit 16-bit 32-bit 20-bit 40-bit 62.2 212.5 124.4 250 62.2 212.5 124.4 325 62.2 212.5 124.4 250 62.2 212.5 124.4 325 4 - 5.5 5-6 5-6 4 - 5.5 5-6 10 - 12 6.5 - 8. 5 10 - 12 1.0 8.5 1.0 5.0 1.0 8.5 1.0 5.0 Byte Ordering Disabled Disabled Enabled Disabled Disabled Enabled Disabled FPGA Fabric Transceiver Interface Width 20-bit 40-bit 40-bit 16-bit 32-bit 32-bit 124.4 325 62.2 212.5 62.2 212.5 124.4 250 62.2 212.5 4 - 5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 10 - 12 6.5 - 8.5 6.5 - 8.5 10 - 12 Disabled 1.0 6.5 1.0 6.5 5-6 Disabled Enabled Data Rate (Gbps) 4 - 5.5 Disabled Disabled Disabled 4 - 5.5 Disabled Enabled Enabled 5-6 Enabled Disabled Disabled TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency Disabled Enabled Enabled FPGA FPGA Fabric Fabric -Transceiver Transceiver Interface Frequency Interface Frequency ( MHz) Disabled Disabled Disabled Byte SerDes (1) Bit-Slip (7-, 10-, 20-bit) 6.5 - 8. 5 6.5 - 8. 5 22 - 26 Enabled 1.0 8.5 4 - 5.5 13 - 16 4 - 5.5 4-5 6.5 - 8. 5 3-4 62.2 282.5 4 - 5.5 3 - 4.5 Notes to Figure 1-101: (1) The maximum data rate specification shown in Figure 1-101 is valid only for the -1 (fastest) speed grade devices. For data rate specifications for other speed grades offered, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) The circled configuration supports data rates up to 11.3 Gbps per channel to implement 40G/100G links. f For more information about 40G/100G transceivers, refer to: Stratix IV Device Handbook Volume 2: Transceivers Enabling 40G/100G Solutions with FPGAs with 11.3-Gbps Transceivers web cast Stratix IV FPGA 40G/100G IP Solutions website AN 570: Implementing the 40G/100G Ethernet Protocol in Stratix IV Devices September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-121 SATA and SAS Options Serial advanced technology attachment (SATA) and serial attached SCSI (SAS) are computer bus standards used in computers to transfer data between a mother board and mass storage devices. Stratix IV GX and GT devices offer options to implement a transceiver that satisfies SATA and SAS protocols. These options are: Transmitter in electrical idle mode Receiver signal detect functionality These options and their selections are described in the following sections. Transmitter Buffer Electrical Idle In Basic functional mode, you can enable the optional input signal tx_forceelecidle. When this input signal of a channel is asserted high, the transmitter buffer in that channel is placed in the electrical idle state. During electrical idle, the output of the transmitter buffer is tri-stated. This signal is used in applications such as SATA and SAS for generating out of band (OOB) signals. An OOB signal is a pattern of idle times and burst times. Different OOB signals are distinguished by their different idle times. 1 Manual CDR lock mode is required because you must be in lock-to-reference mode during OOB signaling. 1 For more information about the transmitter buffer in the Electrical Idle state, refer to the "Transmitter Buffer Electrical Idle" section in "PCIe Mode" on page 1-127. Receiver Input Signal Detect In Basic functional mode, you can enable the optional rx_signaldetect signal (used for protocols such as SATA and SAS) only if you select the 8B/10B block. When you select the optional rx_signaldetect signal, an option is available to set the desired threshold level of the signal being received at the receiver's input buffer. If the signal threshold detection circuitry senses the signal level present at the receiver input buffer to be higher than the chosen signal detect threshold, it asserts the rx_signaldetect signal high. Otherwise, the signal threshold detection circuitry de-asserts the rx_signaldetect signal low. This signal is useful in applications such as SATA and SAS for detecting OOB signals. f For more information on the signal threshold detection circuitry, refer to the ""Signal Threshold Detection Circuitry" section. f For information about other protocols supported using Basic functional mode, refer to AN 577: Recommended Protocol Configurations for Stratix IV FPGAs. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-122 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Deterministic Latency Mode Stratix IV GX and GT devices have a deterministic latency option available for use in high-speed serial interfaces such as CPRI (Common Public Radio Interface) and Open Base Station Architecture Initiative Reference Point3 (OBSAI RP3). Both CPRI and OBSAI RP3 protocols place stringent requirements on the amount of latency variation that is permissible through a link implementing these protocols. Figure 1-102 shows the transceiver datapath when using deterministic latency mode. Figure 1-102. Transceiver Datapath When in Deterministic Latency Mode wrclk rdclk CDR Word Aligner Deskew FIFO Receiver Channel PMA Rate Match FIFO 8B/10 Decoder Byte Deserializer Byte Ordering Receiver Channel PCS Deserializer rdclk 8B/10B Encoder rx_datain wrclk Serializer Byte Serializer RX Phase Compensation FIFO PIPE Interface PCIe hard IP FPGA Fabric TX Phase Compensation FIFO tx_dataout Transmitter Channel PMA Transmitter Channel PCS Transmitter Channel Datapath Receiver Channel Datapath To implement this mode, select the Deterministic Latency option under the Which Protocol will you be using? section in the ALTGX MegaWizard Plug-In Manager. When you select this option, the transmitter channel is automatically placed in bit-slip mode and Enable TX Phase Comp FIFO in register mode is automatically selected as well. The receiver's phase compensation FIFO is automatically placed in the register mode. In addition, an output port (rx_bitslipboundaryselectout[4:0]) from the receiver's word aligner and an input port (tx_bitslipboundaryselect[4:0]) for the transmitter bit-slip circuitry are instantiated. The option for placing the transmitter phase compensation FIFO in register mode is also available. Transmitter Bit Slipping The transmitter is bit slipped to achieve deterministic latency. Use the tx_bitslipboundaryselect[4:0] port to set the number of bits that the transmitter block needs to slip. Table 1-44 lists the number of bits that are allowed to be slipped under different channel widths. Table 1-44. Number of Transmitter Bits Allowed to be Slipped in Deterministic Latency Mode Stratix IV Device Handbook Volume 2: Transceivers Channel Width Slip Zero 8/10 bit 9 bits 16/20 bit 19 bits September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-123 Receiver Bit Slipping The number of bits slipped in the receiver's word aligner is given out on the rx_bitslipboundaryselectout[4:0] output port. The information on this output depends on your deserializer block width. In single-width mode with 8/10-bit channel width, the number of bits slipped in the receiver path is given out sequentially on this output. For example, if zero bits are slipped, the output on rx_bitslipboundaryselectout[4:0] shows a value of 0(00000); if two bits are slipped, the output on rx_bitslipboundaryselectout[4:0] shows a value of 2 (00010). In double-width mode with 16/20-bit channel width, the output is 19 minus the number of bits slipped. For example, if zero bits are slipped, the output on rx_bitslipboundaryselectout[4:0] shows a value of 19 (10011); if two bits are slipped, the output on rx_bitslipboundaryselectout[4:0] shows a value of 17 (10001). The information about the rx_bitslipboundaryselectout[4:0] output port helps in calculating the latency through the receiver datapath. You can use the information on rx_bitslipboundaryselectout[4:0] to set up the tx_bitslipboundaryselect[4:0] appropriately to cancel out the latency uncertainty. Receiver Phase Comp FIFO in Register Mode To remove the latency uncertainty through the receiver's phase compensation FIFO, select the Enable the RX phase comp FIFO in register mode option in the ALTGX MegaWizard Plug-In Manager. In register mode, the phase compensation FIFO acts as a register and thereby removes the uncertainty in latency. The latency through the phase compensation FIFO in register mode is one clock cycle. This mode is available in: Basic single-width mode with 8-bit channel width and 8B/10B Encoder enabled or 10-bit channel width with 8B/10B disabled. Basic double-width mode with 16-bit channel width and 8B/10B encoder enabled or 20-bit channel width with 8B/10B disabled. Transmitter Phase Compensation FIFO in Register Mode In register mode, the phase compensation FIFO acts as a register and thereby removes the uncertainty in latency. The latency through the transmitter and receiver phase compensation FIFO in register mode is one clock cycle. CMU PLL Feedback To implement deterministic latency functional mode, the phase relationship between the low-speed parallel clock and CMU PLL input reference clock must be deterministic. You can achieve this by selecting the Enable PLL phase frequency detector (PFD) feedback to compensate latency uncertainty in Tx dataout and Tx clkout paths relative to the reference clock option in the ALTGX MegaWizard Plug-In Manager. By selecting this option, a feedback path is enabled that ensures a deterministic relationship between the low-speed parallel clock and CMU PLL input reference clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-124 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture In order to achieve deterministic latency through the transceiver, the reference clock to the CMU PLL must be the same as the low-speed parallel clock. For example, if you need a data rate of 1.2288 Gbps to be implemented for the CPRI protocol that places stringent requirements on the amount of latency variation, you must choose a reference clock of 122.88 MHz to allow for a feedback path from the CMU PLL to be used. This feedback path reduces the variations in latency. When selecting this option, you must provide an input reference clock to the CMU PLL that is of the same frequency as the low-speed parallel clock. 1 In a CPRI implementation, the input reference clock to the CMU PLL must be the same as the low-speed parallel clock. Each CPRI channel uses one CMU PLL; therefore, each transceiver block can implement two CPRI x1 channels only. ATX PLLs do not have the feedback path enabled; therefore, they cannot be used for implementing the CPRI configuration. In the deterministic latency x4 option, up to four CPRI TX channels can be bundled in an x4 group so that they all have the same TX uncertainty and just require one TX PLL to compensate for it. This is allowed in cases where the data rates are multiples of a single PLL output frequency; for example, 0.6144 Gbps, 1.228 Gbps, 2.4576 Gbps, and 4.9152 Gbps. For x4 bundled channels to maintain PLL lock during auto-negotiation, the IP must use over-sampling (sending the same bit multiple times) to output lower auto-negotiated line rates. Do not use the hard 8B/10B for oversampled channels. CPRI and OBSAI You can use deterministic latency functional mode to implement protocols such as CPRI and OBSAI. The CPRI interface defines a digital point-to-point interface between the Radio Equipment Control (REC) and the Radio Equipment (RE) allowing flexibility in either co-locating the REC and the RE or remote location of the RE. Figure 1-103 shows various CPRI topologies. In most cases, CPRI links are between REC and RE modules or between two RE modules in a chain configuration. Figure 1-103. CPRI Topologies RE RE RE Ring RE RE Tree and Branch RE REC Radio Equipment Control RE RE Chain Piont-to-Point RE RE Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-125 If the destination for high-speed serial data leaving the REC is the first RE, it is a single-hop connection. If serial data from the REC has to traverse through multiple REs before reaching the destination RE, it is a multi-hop connection. Remotely locating the RF transceiver from the main base station introduces a complexity with overall system delay. CPRI specification requires that the accuracy of measurement of round-trip delay on single-hop and multi-hop connections be within 16.276 ns in order to properly estimate the cable delay. For a single-hop system, this allows a variation in round-trip delay of up to 16.276 ns. For multi-hop systems however, the allowed delay variation is divided among number of hops in the connection-- typically equal to 16.276 ns/ (# of hops), but not always equally divided among the hops. Deterministic latency on a CPRI link also enables highly accurate triangulation of a caller's location. The OBSAI was established by several OEM's for developing a set of specifications that can be used for configuring and connecting common modules into base transceiver stations (BTS). The BTS has four main modules--radio frequency (RF), baseband, control and transport. Figure 1-104 shows a typical BTS. The radio frequency module (RFM) receives signals using portable devices and converts them to digital data. The baseband module processes the encoded signal and brings it back to baseband before transmitting it to the terrestrial network using the transport module. Coordination between these three functions is maintained by a control module. Figure 1-104. BTS in OSBAL System Software Baseband Module Transport Module RF Module RP3 (1) RP2 (1) Interface Proprietary Module(s) BB Switch Control & Clock RFM Clock and Sync Control Module RP1 (1) Power System Note to Figure 1-104: (1) "RP" means Reference Point. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-126 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Under the deterministic latency option, CPRI data rates can be implemented in single-width mode with 8/10-bit channel width and double-width mode with 16/20-bit channel width options only. Figure 1-105 shows the block diagram of the deterministic latency option. Figure 1-105. Block Diagram of the Deterministic Latency Option Stratix IV GX and GT Configurations Protocol Basic Functional Modes PMA-PCS Interface Width Single Width 8-bit 10-Bit Double Width 16-Bit 20-Bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-Bit 10-Bit 10-Bit 10-Bit 8-Bit 16-Bit 10-Bit Deterministic Latency 10-Bit 20-Bit Functional Mode Deterministic Latency Deterministic Latency Data Rate 0.6 - 3.75 Gbps - GX 600 Mbps -3.75 Gbps - GT 0.6 - 8.5 Gbps - GX 600 Mbps - 8.5 Gbps - GT Channel Bonding x1, x4 x1, x4 Low-Latency PCS Disabled Disabled Word Aligner (Pattern Length) Manual Alignment (10-Bit) Manual Alignment (10-Bit, 20-bit) 8B/10B Encoder/Decoder Rate Match FIFO Disabled Enabled Disabled Enabled Disabled Disabled Disabled Disabled Byte SerDes Disabled Enabled Disabled Enabled Disabled Enabled Disabled Enabled Data Rate (Gbps) 0.6 - 2.5 GX 2.488 -2.5 GT 0.6 - 3.75 GX 2.488 - 3.75 GT 0.6 - 2.5 GX 2.488 -2.5 GT 0.6 - 3.75 GX 2.488 - 3.75 GT 1.0 - 6.5 GX 2.488 - 6.5 GT 1.0 - 8.5 GX 2.488 - 8.5 GT 1.0 - 5.0 GX 2.488 - 5.0 GT 1.0 - 8.5 GX 2.488 - 8.5 GT Byte Ordering Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled FPGA Fabric-Transceiver Interface Width 10-Bit 20-Bit 8-Bit 16-Bit 20-Bit 40-Bit 16-Bit 32-Bit FPGA FPGA Fabric Fabric-Transceiver Transceiver Interface Frequency Interface Frequency (MHz) 60 - 250 - GX 30 - 187.5 - GX 248.8 - 250 - GT 124.4 - 187.5 - GT 60 - 250 - GX 30 - 187.5 - GX 248.8 - 250 - GT 124.4 - 187.5 - GT 50 - 325 - GX 124.4 - 325 - GT 25 - 212.5 - GX 62.2 - 212.5 - GT 50 - 250 - GX 124.4 - 250 - GT TX PCS Latency (FPGA Fabric-Transceiver Interface Frequency Interface Clock Cycles) 4 4 4 4 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 7 7 8 8 1 Stratix IV Device Handbook Volume 2: Transceivers 25 - 212.5 - GX 62.2 - 212.5 - GT To implement CPRI/OBSAI using deterministic latency mode, Altera recommends using configurations with the byte serializer/deserializer disabled. September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-127 Table 1-45 lists the PMA-PCS interface widths, CPRI and OSBAI data rates in deterministic latency mode. Table 1-45. PMA-PCS Interface Widths, CPRI and OSBAI Data Rates in Deterministic latency Mode Supported Data Rate Range Deterministic Latency Mode Single-width mode PMA-PCS Interface Width for CPRI & OBSAI 600 Mbps to 3.75 Gbps 8 bit/10 bit CPRI Data Rate (GBPS) PCS Clock Frequency (MHz) OBSAI Data Rate (Gbps) PCS Clock Frequency (MHz) 0.6144 61.44 768 76.8 1.2288 122.88 1.536 153.6 245.76 -- -- 2.4576 16 bit/20 bit Double-width mode > 1 Gbps 16 bit/20 bit 32 bit/40 bit (1) 3.072 4.915 (2), (3) 6.144 (2), (3), (4) 153.6 245.76 307.2 1.536 6.144 76.8 (3) 153.6 (3), (4) 307.2 3.072 Notes to Table 1-45: (1) When configured in double-width mode for the same data rate, the core clock frequency is halved. (2) Requires double-width mode. (3) When configured for 32/40-bit channel width requiring byte serializer/deserializer, the core clock is halved. (4) Requires the byte serializer/deserializer. PCIe Mode Intel Corporation has developed a PHY interface for the PCIe Architecture specification to enable implementation of a PCIe-compliant physical layer device. The PCIe specification also defines a standard interface between the physical layer device and the media access control layer (MAC). Version 2.0 of the PCIe specification provides implementation details for a PCIe-compliant physical layer device at both Gen1 (2.5 GT/s) and Gen2 (5 GT/s) signaling rates. To implement a Version 2.0 PCIe-compliant PHY, you must configure the Stratix IV GX and GT transceivers in PCIe functional mode. Stratix IV GX and GT devices have built-in PCIe hard IP blocks that you can use to implement the PHY-MAC layer, data link layer, and transaction layer of the PCIe protocol stack. You can also bypass the PCIe hard IP blocks and implement the PHY-MAC layer, data link layer, and transaction layer in the FGPA fabric using a soft IP. If you enable the PCIe hard IP blocks, the Stratix IV transceivers interface with these hard IP blocks. Otherwise, the Stratix IV transceivers interface with the FPGA fabric. You can configure the Stratix IV GX and GT transceivers in PCIe functional mode using one of the following two methods: 1 September 2012 ALTGX MegaWizard Plug-In Manager--if you do not use the PCIe hard IP block PCIe Compiler--if you use the PCIe hard IP block Description of PCIe hard IP architecture and PCIe mode configurations allowed when using the PCIe hard IP block are beyond the scope of this chapter. For more information about the PCIe hard IP block, refer to the PCI Express Compiler User Guide. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-128 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture PCIe Mode Configurations Stratix IV GX and GT transceivers support both Gen1 (2.5 Gbps) and Gen2 (5 Gbps) data rates in PCIe functional mode. When configured for the Gen2 (5 Gbps) data rate, the Stratix IV GX and GT transceivers allow dynamic switching between Gen2 (5 Gbps) and Gen1 (2.5 Gbps) signaling rates. Dynamic switch capability between the two PCIe signaling rates is critical for speed negotiation during link training. Stratix IV GX and GT transceivers support x1, x4, and x8 lane configurations in PCIe functional mode at both 2.5 Gbps and 5 Gbps data rates. In PCIe x1 configuration, the PCS and PMA blocks of each channel are clocked and reset independently. PCIe x4 and x8 configurations support channel bonding for four-lane and eight-lane PCIe links. In these bonded channel configurations, the PCS and PMA blocks of all bonded channels share common clock and reset signals. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-129 Figure 1-106 shows the Stratix IV GX and GT transceiver configurations allowed in PCIe functional mode. Figure 1-106. Stratix IV GX and GT Transceivers in PCIe Functional Mode Stratix IV GX and GT Configurations Protocol Basic Functional Modes PMA-PCS Interface Width Single Width 8-bit Double Width 10-bit 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Deterministic Latency 10-Bit PIPE Functional Mode 2.5 Gbps (Gen1) 5 Gbps (Gen2) Channel Bonding x1, x4, x8 x1, x4, x8 PMA-PCS Interface Width 10-Bit 10-Bit Data Rate Automatic Synchronization State Machine (/K28.5+/,/K28.5-/) Word Aligner (Pattern) Automatic Synchronization State Machine (/K28.5+/,/K28.5-/) 8B/10B Encoder/ Decoder Enabled Enabled Rate Match FIFO Enabled Enabled PCI Express hardIP 20-Bit Enabled Disabled Enabled Disabled Byte SerDes Disabled Disabled Enabled Disabled Enabled PCS-hardIP or PCS-FPGA Fabric Interface Width 8-Bit 8-Bit 16-Bit 8-Bit 16-Bit PCS-hardIP or PCS-FPGA Fabric Interface Frequency TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) September 2012 Altera Corporation 250 MHz 250 MHz 125 MHz 500 MHz 250 MHz 5-6 5-6 4 - 5.5 5-6 4 - 5.5 20 - 24 20 - 24 11.5 14.5 20 - 24 11.5 14.5 Stratix IV Device Handbook Volume 2: Transceivers 1-130 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture PCIe Mode Datapath Figure 1-107 shows the Stratix IV GX and GT transceiver datapath when configured in PCIe functional mode. Figure 1-107. Stratix IV GX and GT Transceiver Datapath in PCIe x1 Mode Transmitter Channel PCS rdclk wrclk rdclk 8B/10B High-Speed Serial Clock Word Aligner 8B/10B Decoder Byte Ordering Receiver Channel PMA Rate Match FIFO Receiver Channel PCS Byte Deserializer rx_coreclk[0] Low-Speed Parallel Clock tx_clkout[0] RX Phase Compensation FIFO PIPE Interface PCIe hard IP /2 FPGA Fabric Transceiver Interface Clock Serializer 8B/10B Encoder CDR wrclk Byte Serializer Deskew FIFO tx_coreclk[0] TX Phase Compensation FIFO Transmitter Channel PMA Deserializer FPGA Fabric Low -Speed Parallel Clock /2 Parallel Recovery Clock Transmitter Channel Datapath FPGA Fabric Receiver Channel Datapath For more information, refer to "Rate Match (Clock Rate Compensation) FIFO" on page 1-77. Table 1-46 lists the transceiver datapath clock frequencies in PCIe functional mode configured using the ALTGX MegaWizard Plug-In Manager. Table 1-46. Stratix IV GX and GT Transceiver Datapath Clock Frequencies in PCIe Mode Functional Mode PCIe x1, x4, and x8 (Gen1) PCIe x1, x4, and x8 (Gen2) Data Rate High-Speed Serial Clock Frequency Parallel Recovered Clock and Low-Speed Parallel Clock Frequency 2.5 Gbps 1.25 GHz 250 MHz 5 Gbps 2.5 GHz 500 MHz FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Serializer/ Deserializer (8 Bit Wide) With Byte Serializer/ Deserializer (16 Bit Wide) 250 MHz 125 MHz N/A (1) 250 MHz Note to Table 1-46: (1) In PCIe functional mode at Gen2 (5 Gbps) data rate, the byte serializer/deserializer cannot be bypassed. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-131 Transceiver datapath clocking varies between non-bonded (x1) and bonded (x4 and x8) configurations in PCIe mode. f For more information about transceiver datapath clocking in different PCIe configurations, refer to the Transceiver Clocking in Stratix IV Devices chapter. Table 1-47 lists the transmitter and receiver datapaths in PCIe mode. Table 1-47. Datapaths in PCIe Mode Transmitter Datapath Receiver Datapath PCIe interface Y Y Transmitter phase compensation FIFO Y -- Optional byte serializer (enabled for 16-bit and disabled for 8-bit FPGA fabric-transceiver interface Y -- 8B/10B encoder Y -- 10:1 serializer Y -- Transmitter buffer with receiver detect circuitry Y -- Receiver buffer with signal detect circuitry -- Y 1:10 deserializer -- Y Word aligner that implements PCIe-compliant synchronization state machine -- Y Optional rate match FIFO (clock rate compensation) that can tolerate up to 600 PPM frequency difference -- Y 8B/10B decoder -- Y Optional byte deserializer (enabled for 16-bit and disabled for 8-bit FPGA fabric-transceiver interface) -- Y Receiver phase compensation FIFO -- Y Table 1-48 lists the features supported in PCIe functional mode for 2.5 Gbps and 5 Gbps data rate configurations. For more information, refer to "Rate Match FIFO in PCIe Mode" on page 1-78. Table 1-48. Supported Features in PCIe Mode (Part 1 of 2) 2.5 Gbps (Gen1) 5 Gbps (Gen2) x1, x4, x8 link configurations Y Y PCIe-compliant synchronization state machine Y Y 300 PPM (total 600 PPM) clock rate compensation Y Y 8-bit FPGA fabric-transceiver interface Y -- 16-bit FPGA fabric-transceiver interface Y Y Transmitter buffer electrical idle Y Y Receiver Detection Y Y 8B/10B encoder disparity control when transmitting compliance pattern Y Y Power state management Y Y Feature September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-132 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-48. Supported Features in PCIe Mode (Part 2 of 2) 2.5 Gbps (Gen1) 5 Gbps (Gen2) Receiver status encoding Y Y Dynamic switch between 2.5 Gbps and 5 Gbps signaling rate -- Y Dynamically selectable transmitter margining for differential output voltage control -- Y Dynamically selectable transmitter buffer de-emphasis of -3.5 db and -6 dB -- Y Feature PCIe Interface In PCIe mode, each channel has a PCIe interface block that transfers data, control, and status signals between the PHY-MAC layer and the transceiver channel PCS and PMA blocks. The PCIe interface block is compliant to version 2.0 of the PCIe specification. If you use the PCIe hard IP block, the PHY-MAC layer is implemented in the hard IP block. Otherwise, the PHY-MAC layer can be implemented using soft IP in the FPGA fabric. 1 The PCIe interface block is only used in PCIe mode and cannot be bypassed. Besides transferring data, control, and status signals between the PHY-MAC layer and the transceiver, the PCIe interface block implements the following functions required in a PCIe-compliant physical layer device: Forces the transmitter buffer in electrical idle state Initiates the receiver detect sequence 8B/10B encoder disparity control when transmitting compliance pattern Manages the PCIe power states Indicates the completion of various PHY functions; for example, receiver detection and power state transitions on the pipephydonestatus signal Encodes the receiver status and error conditions on the pipestatus[2:0] signal as specified in the PCIe specification Transmitter Buffer Electrical Idle When the input signal tx_forceelecidle is asserted high, the PCIe interface block puts the transmitter buffer in that channel in the electrical idle state. During electrical idle, the transmitter buffer differential and common mode output voltage levels are compliant to the PCIe Base Specification 2.0 for both PCIe Gen1 and Gen2 data rates. Figure 1-108 shows the relationship between the assertion of the tx_forceelecidle signal and the transmitter buffer output on the tx_dataout port. Time T1 taken from the assertion of the tx_forceelecidle signal to the transmitter buffer reaching electrical idle voltage levels is pending characterization. Once in the electrical idle state, the PCIe protocol requires the transmitter buffer to stay in electrical idle for a minimum of 20 ns for both Gen1 and Gen2 data rates. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1 1-133 The minimum period of time for which the tx_forceelecidle signal must be asserted high such that the transmitter buffer stays in electrical idle state for at least 20 ns is pending characterization. Figure 1-108. Transmitter Buffer Electrical Idle State tx_forcelecidle tx_dataout T1 >20 ns The PCIe specification requires the transmitter buffer to be in electrical idle in certain power states. For more information about the tx_forceelecidle signal levels required in different PCIe power states, refer to Table 1-50 on page 1-137. Receiver Detection During the detect substate of the link training and status state machine (LTSSM), the PCIe protocol requires the transmitter channel to perform a receiver detect sequence to detect if a receiver is present at the far end of each lane. The PCIe specification requires the receiver detect operation to be performed during the P1 power state. The PCIe interface block in Stratix IV GX and GT transceivers provide an input signal tx_detectrxloopback for the receiver detect operation. When the input signal tx_detectrxloopback is asserted high in the P1 power state, the PCIe interface block sends a command signal to the transmitter buffer in that channel to initiate a receiver detect sequence. In the P1 power state, the transmitter buffer must always be in the electrical idle state. After receiving this command signal, the receiver detect circuitry creates a step voltage at the output of the transmitter buffer. If an active receiver (that complies with the PCIe input impedance requirements) is present at the far end, the time constant of the step voltage on the trace is higher when compared with the time constant of the step voltage when the receiver is not present. The receiver detect circuitry monitors the time constant of the step signal seen on the trace to determine if a receiver was detected. The receiver detect circuitry monitor requires a 125-MHz clock for operation that you must drive on the fixedclk port. 1 For the receiver detect circuitry to function reliably, the AC-coupling capacitor on the serial link and the receiver termination values used in your system must be compliant to the PCIe Base Specification 2.0. Receiver detect circuitry communicates the status of the receiver detect operation to the PCIe interface block. If a far-end receiver is successfully detected, the PCIe interface block asserts pipephydonestatus for one clock cycle and synchronously drives the pipestatus[2:0] signal to 3'b011. If a far-end receiver is not detected, the PCIe interface block asserts pipephydonestatus for one clock cycle and synchronously drives the pipestatus[2:0] signal to 3'b000. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-134 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-109 and Figure 1-110 show the receiver detect operation where a receiver was successfully detected and where a receiver was not detected, respectively. Figure 1-109. Receiver Detect, Successfully Detected 2'b10(P1) powerdown[1:0] tx_detectrxloopback pipephydonestatus pipestatus[2:0] 3'b000 3'b011 Figure 1-110. Receiver Detect, Unsuccessfully Detected 2'b10 (P1) powerdown[1:0] tx_detectrxloopback pipephydonestatus 3'b000 pipestatus[2:0] Compliance Pattern Transmission Support The LTSSM state machine can enter the polling.compliance substate where the transmitter is required to transmit a compliance pattern as specified in the PCIe Base Specification 2.0. The polling.compliance substate is intended to assess if the transmitter is electrically compliant with the PCIe voltage and timing specifications. The compliance pattern is a repeating sequence of the following four code groups: Stratix IV Device Handbook Volume 2: Transceivers /K28.5/ /D21.5/ /K28.5/ /D10.2/ September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-135 The PCIe protocol requires the first /K28.5/ code group of the compliance pattern to be encoded with negative current disparity. To satisfy this requirement, the PCIe interface block provides the input signal tx_forcedispcompliance. A high level on tx_forcedispcompliance forces the associated parallel transmitter data on the tx_datain port to transmit with negative current running disparity. For 8-bit transceiver channel width configurations, you must drive tx_forcedispcompliance high in the same parallel clock cycle as the first /K28.5/ of the compliance pattern on the tx_datain port. For 16-bit transceiver channel width configurations, you must drive the tx_forcedispcompliance high in the same parallel clock cycle as /K28.5/D21.5/ of the compliance pattern on the tx_datain port. Figure 1-111 and Figure 1-112 show the required level on the tx_forcedispcompliance signal while transmitting the compliance pattern in 8-bit and 16-bit channel width configurations, respectively. Figure 1-111. Compliance Pattern Transmission Support, 8-Bit Channel Width Configurations tx_datain[7:0] K28.5 D21.5 K28.5 D10.2 K28.5 D21.5 K28.5 D10.2 BC B5 BC 4A BC B5 BC 4A tx_ctrlenable tx_forcedispcompliance Figure 1-112. Compliance Pattern Transmission Support, 16-Bit Wide Channel Configurations tx_datain[15:0] /K28.5/D21.5/ /K28.5/D10.2/ /K28.5/D21.5/ /K28.5/D10.2/ B5BC BC4A B5BC BC4A tx_ctrlenable[1:0] 01 tx_forcedispcompliance Power State Management The PCIe specification defines four power states--P0, P0s, P1, and P2--that the physical layer device must support to minimize power consumption. September 2012 P0 is the normal operating state during which packet data is transferred on the PCIe link. P0s, P1, and P2 are low-power states into which the physical layer must transition as directed by the PHY-MAC layer to minimize power consumption. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-136 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The PCIe specification provides the mapping of these power states to the LTSSM states specified in the PCIe Base Specification 2.0. The PHY-MAC layer is responsible for implementing the mapping logic between the LTSSM states and the four power states in the PCIe-compliant PHY. The PCIe interface in Stratix IV GX and GT transceivers provides an input port, powerdn[1:0], for each transceiver channel configured in PCIe mode. Table 1-49 lists mapping between the logic levels driven on the powerdn[1:0] port and the resulting power state that the PCIe interface block puts the transceiver channel into. Table 1-49. Power State Functions and Descriptions Power State powerdn Function P0 2'b00 Transmits normal data, transmits electrical idle, or enters into loopback mode Normal operation mode P0s 2'b01 Only transmits electrical idle Low recovery time saving state P1 2'b10 Transmitter buffer is powered down and can do a receiver detect while in this state High recovery time power saving state P2 2'b11 Transmits electrical idle or a beacon to wake up the downstream receiver Lowest power saving state 1 Description When transitioning from the P0 power state to lower power states (P0s, P1, and P2), the PCIe specification requires the physical layer device to implement power saving measures. Stratix IV GX and GT transceivers do not implement these power saving measures except putting the transmitter buffer in electrical idle in the lower power states. The PCIe interface block indicates successful power state transition by asserting the pipephydonestatus signal for one parallel clock cycle as specified in the PCIe specification. The PHY-MAC layer must not request any further power state transition until the pipephydonestatus signal has indicated the completion of the current power state transition request. Figure 1-113 shows an example waveform for a transition from the P0 to P2 power state. Figure 1-113. Power State Transition from the P0 to P2 Power State Parallel Clock powerdn[1:0] 2'b00 (P0) 2'b11 (P2) pipephydonestatus Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-137 The PCIe specification allows the PCIe interface to perform protocol functions; for example, receiver detect, loopback, and beacon transmission, in specified power states only. This requires the PHY-MAC layer to drive the tx_detectrxloopback and tx_forceelecidle signals appropriately in each power state to perform these functions. Table 1-50 lists the logic levels that the PHY-MAC layer must drive on the tx_detectrxloopback and tx_forceelecidle signals in each power state. Table 1-50. Logic Levels for tx_detectrxloopback and tx_forceelecidle in Different Power States Power State P0 P0s P1 P2 tx_detectrxloopback tx_forceelecidle 0: normal mode 0: Must be de-asserted 1: datapath in loopback mode 1: Illegal mode 0: Illegal mode Don't care 1: Must be asserted in this state 0: Electrical Idle 0: Illegal mode 1: receiver detect 1: Must be asserted in this state Don't care De-asserted in this state for sending beacon. Otherwise asserted. Receiver Status The PCIe specification requires the PHY to encode the receiver status on a 3-bit RxStatus[2:0] signal. This status signal is used by the PHY-MAC layer for its operation. The PCIe interface block receives status signals from the transceiver channel PCS and PMA blocks and encodes the status on the 3-bit output signal pipestatus[2:0] to the FPGA fabric. The encoding of the status signals on pipestatus[2:0] is compliant with the PCIe specification and is listed in Table 1-51. Table 1-51. Encoding of the Status Signals on pipestatus[2:0] pipestatus[2:0] Description Error Condition Priority 3'b000 Received data OK N/A 3'b001 One SKP symbol added 5 3'b010 One SKP symbol deleted 6 3'b011 Receiver detected N/A 3'b100 8B/10B decode error 1 3'b101 Elastic buffer (rate match FIFO) overflow 2 3'b110 Elastic buffer (rate match FIFO) underflow 3 3'b111 Received disparity error 4 Two or more of the error conditions (for example, 8B/10B decode error [code group violation], rate match FIFO overflow or underflow, and receiver disparity error), can occur simultaneously. The PCIe interface follows the priority listed in Table 1-51 while encoding the receiver status on the pipestatus[2:0] port. For example, if the PCIe interface receives an 8B/10B decode error and disparity error for the same symbol, it drives 3'b100 on the pipestatus[2:0] signal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-138 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Fast Recovery Mode The PCIe Base specification fast training sequences (FTS) are used for bit and byte synchronization to transition from L0s to L0 (PCIe P0s to P0) power states. When transitioning from the L0s to L0 power state, the PCIe Base Specification requires the physical layer device to acquire bit and byte synchronization after receiving a maximum of 255 FTS (~4 us at Gen1 data rate and ~2 us at Gen2 data rate). If you have configured the Stratix IV GX and GT receiver CDR in Automatic Lock mode, the receiver cannot meet the PCIe specification of acquiring bit and byte synchronization within 4 s (Gen1 data rate) or 2 s (Gen2 data rate) due to the signal detect and PPM detector time. To meet this specification, each Stratix IV GX and GT transceiver has a built-in Fast Recovery circuitry that you can optionally enable. 1 To enable the Fast Recovery circuitry, select the Enable fast recovery mode option in the ALTGX MegaWizard Plug-In Manager. If you enable the Fast Recovery mode option, the Fast Recovery circuitry controls the receiver CDR rx_locktorefclk and rx_locktodata signals to force the receiver CDR in LTR or LTD mode. It relies on the Electrical Idle Ordered Sets (EIOS), N_FTS sequences received in the L0 power state, and the signal detect signal from the receiver input buffer to control the receiver CDR lock mode. 1 The Fast Recovery circuitry is self-operational and does not require control inputs from you. When enabled, the rx_locktorefclk and rx_locktodata ports are not available in the ALTGX MegaWizard Plug-In Manager. Electrical Idle Inference The PCIe protocol allows inferring the electrical idle condition at the receiver instead of detecting the electrical idle condition using analog circuitry. Clause 4.2.4.3 in the PCIe Base Specification 2.0 specifies conditions to infer electrical idle at the receiver in various substates of the LTSSM state machine. In all PCIe modes (x1, x4, and x8), each receiver channel PCS has an optional Electrical Idle Inference module designed to implement the electrical idle inference conditions specified in the PCIe Base Specification 2.0. You can enable the Electrical Idle Inference module by selecting the Enable electrical idle inference functionality option in the ALTGX MegaWizard Plug-In manager. If enabled, this module infers electrical idle depending on the logic level driven on the rx_elecidleinfersel[2:0] input signal. The Electrical Idle Inference module in each receiver channel indicates whether the electrical idle condition is inferred or not on the pipeelecidle signal of that channel. The Electrical Idle Interface module drives the pipeelecidle signal high if it infers an electrical idle condition; otherwise, it drives it low. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-139 Table 1-52 lists electrical idle inference conditions specified in the PCIe Base Specification 2.0 and implemented in the Electrical Idle Inference module to infer electrical idle in various substates of the LTSSM state machine. For the Electrical Idle Inference Module to correctly infer an electrical idle condition in each LTSSM substate, you must drive the rx_elecidleinfersel[2:0] signal appropriately, as shown in Table 1-52. Table 1-52. Electrical Idle Inference Conditions LTSSM State Gen1 (2.5 Gbps) Gen2 (5 Gbps) rx_elecidleinfersel[2:0] L0 Absence of skip ordered set in 128 s window Absence of skip ordered set in 128 s window 3'b100 Recovery.RcvrCfg Absence of TS1 or TS2 Absence of TS1 or TS2 ordered set in 1280 UI interval ordered set in 1280 UI interval 3'b101 Recovery.Speed when successful speed negotiation = 1'b1 Absence of TS1 or TS2 Absence of TS1 or TS2 ordered set in 1280 UI interval ordered set in 1280 UI window 3'b101 Recovery.Speed when successful speed negotiation = 1'b0 Absence of an exit from Electrical Idle in 2000 UI interval Absence of an exit from Electrical Idle in 16000 UI interval 3'b110 Loopback.Active Absence of an exit from Electrical Idle in 128 s window N/A 3'b111 (as slave) In the Recovery.Speed substate of the LTSSM state machine with unsuccessful speed negotiation (rx_elecidleinfersel[2:0] = 3'b110), the PCIe Base Specification requires the receiver to infer an electrical idle condition (pipeelecidle = high) if absence of an exit from Electrical Idle is detected in a 2000 UI interval for Gen1 data rate and 16000 UI interval for Gen2 data rate. The electrical idle inference module detects an absence of exit from Electrical Idle if four /K28.5/ COM code groups are not received in the specified interval. In other words, when configured for Gen1 data rate and rx_elecidleinfersel[2:0] = 3'b110, the Electrical Idle Inference module asserts pipeelecidle high if it does not receive four /K28.5/ COM code groups in a 2000 UI interval. When configured for Gen1 data rate and rx_elecidleinfersel[2:0] = 3'b111 in the Loopback.Active substate of the LTSSM state machine, the Electrical Idle Inference module asserts pipeelecidle high if it does not receive four /K28.5/ COM code groups in a 128 s interval. When configured for Gen2 data rate and rx_elecidleinfersel[2:0] = 3'b110, the Electrical Idle Inference module asserts pipeelecidle high if it does not receive four /K28.5/ COM code groups in a 16000 UI interval. 1 The Electrical Idle Inference module does not have the capability to detect the electrical idle exit condition based on reception of the electrical idle exit ordered set (EIEOS), as specified in the PCIe Base Specification. If you select the Enable Electrical Idle Inference Functionality option in the ALTGX MegaWizard Plug-In Manager and drive rx_elecidleinfersel[2:0] = 3'b0xx, the Electrical Idle Inference block uses the EIOS detection from the Fast Recovery circuitry to drive the pipeelecidle signal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-140 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture If you do not select the Enable electrical idle inference functionality option in the ALTGX MegaWizard Plug-In Manager, the Electrical Idle Inference module is disabled. In this case, the rx_signaldetect signal from the signal detect circuitry in the receiver buffer is inverted and driven as the pipeelecidle signal. Recommendation When Using the Electrical Idle Inference Block In a PCIe link, when operating at Gen2 data rate, the downstream device can go into the Disable state after instruction from the upper layer. Once in the Disable state, the downstream device must detect an Electrical Idle Exit condition to go into the Detect state. At this same time, the upstream device can be directed by the upper layer to go into the Detect state and start transmitting the COM symbols to the downstream device at Gen 1 data rate. 1 The Disable and Detect states are different states of the Link Training and Status State Machine as described by the PCIe Base Specification Rev 2.0. 1 The COM symbol is an 8B/10B encoded value of K28.5 and is part of the training sequences TS1 and TS2 as described by the PCIe Base Specification Rev 2.0. When the Stratix IV GX and GT device is operating as a downstream device at PCIe Gen 2 data rates and if it goes into the Disable State, the Stratix IV GX and GT receiver must receive an Electrical Idle Exit condition in order to move out of the Disable state. For the Stratix IV GX and GT receiver, the Electrical Idle Exit condition is achieved when COM symbols are received from the upstream device. However, after the Disable state is achieved by the Stratix IV GX and GT receiver (the downstream device) during Gen 2 data rate operation, and if at the same time the upstream device is directed to transition to the Detect state, the upstream device starts to send COM symbols at Gen 1 data rate. Consequently, the Stratix IV GX and GT receiver (the downstream device) does not recognize the COM symbols as it is operating at Gen 2 data rate. To avoid this scenario, the Link Training Status State Machine (LTSSM) in the FPGA fabric of Stratix IV GX and GT receiver (the downstream device) must be implemented in such a way that whenever the downstream device goes into the Disable state and the upstream device is directed to go into the Detect state, the rateswitch signal must be transitioned from high to low. This allows the Stratix IV GX and GT receiver (the downstream device) to move from Gen 2 to Gen 1 data rate. Subsequently, the Stratix IV GX and GT receiver (the downstream device) recognizes the COM symbols being sent by the upstream device at Gen 1 data rates and moves from the Disable state to the Detect state. PCIe Gen2 (5 Gbps) Support The PCIe functional mode supports the following additional features when configured for 5 Gbps data rate: Stratix IV Device Handbook Volume 2: Transceivers Dynamic switch between 2.5 Gbps and 5 Gbps signaling rate Dynamically selectable transmitter margining for differential output voltage control Dynamically selectable transmitter buffer de-emphasis of -3.5 db and -6 dB September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-141 Dynamic Switch Between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) Signaling Rate During link training, the upstream and downstream PCIe ports negotiate the speed (2.5 Gbps or 5 Gbps) at which the link operates. Because the upstream and downstream PCIe ports do not know the speed capabilities of their link partner, the PCIe protocol requires each port to start with a Gen1 (2.5 Gbps) signaling rate. One of the ports capable of supporting the Gen2 (5 Gbps) signaling rate might initiate a speed change request by entering the Recovery state of the LTSSM. In the Recovery state, each port advertises its speed capabilities by transmitting training sequences as specified in the PCIe Base Specification 2.0. If both ports are capable of operating at the Gen2 (5 Gbps) signaling rate, the PHY-MAC layer instructs the physical layer device to operate at the Gen2 (5 Gbps) signaling rate. To support speed negotiation during link training, the PCIe specification requires a PCIe-compliant physical layer device to provide an input signal (Rate) to the PHY-MAC layer. When this input signal is driven low, the physical layer device must operate at the Gen1 (2.5 Gbps) signaling rate; when driven high, this input signal must operate at the Gen2 (5 Gbps) signaling rate. The PCIe specification allows the PHY-MAC layer to initiate a signaling rateswitch only in power states P0 and P1 with the transmitter buffer in the Electrical Idle state. The PCIe specification allows the physical layer device to implement the signaling rateswitch using either of the following approaches: Change the transceiver datapath clock frequency, keeping the transceiver interface width constant Change the transceiver interface width between 8 bit and 16 bit, keeping the transceiver clock frequency constant When configured in PCIe functional mode at Gen2 (5 Gbps) data rate, the ALTGX MegaWizard Plug-In Manager provides the input signal rateswitch. The rateswitch signal is functionally equivalent to the Rate signal specified in the PCIe specification. The PHY-MAC layer can use the rateswitch signal to instruct the Stratix IV GX and GT device to operate at either Gen1 (2.5 Gbps) or Gen2 (5 Gbps) data rate, depending on the negotiated speed between the upstream and downstream ports. A low-to-high transition on the rateswitch signal initiates a data rateswitch from Gen1 (2.5 Gbps) to Gen2 (5 Gbps). A high-to-low transition on the rateswitch signal initiates a data rateswitch from Gen2 (5 Gbps) to Gen1 (2.5 Gbps). The signaling rateswitch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) is achieved by changing the transceiver datapath clock frequency between 250 MHz and 500 MHz, while maintaining a constant transceiver interface width of 16-bit. The dedicated PCIe rateswitch circuitry performs the dynamic switch between the Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rate. The PCIe rateswitch circuitry consists of: September 2012 PCIe rateswitch controller PCIe clock switch circuitry Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-142 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture PCIe Rateswitch Controller The rateswitch signal serves as the input signal to the PCIe rateswitch controller. After seeing a transition on the rateswitch signal from the PHY-MAC layer, the PCIe rateswitch controller performs the following operations: Controls the PCIe clock switch circuitry to switch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rate depending on the rateswitch signal level Disables and resets the transmitter and receiver phase compensation FIFO pointers until the PCIe clock switchover circuitry indicates successful rateswitch completion Communicates completion of rateswitch to the PCIe interface module, which in turn communicates completion of the rateswitch to the PHY-MAC layer on the pipephydonestatus signal PCIe rateswitch controller location: In PCIe x1 mode, the PCIe rateswitch controller is located in the transceiver PCS of each channel. In PCIe x4 mode, the PCIe rateswitch controller is located in CMU0_Channel within the transceiver block. In PCIe x8 mode, the PCIe rateswitch controller is located in CMU0_Channel within the master transceiver block. 1 When operating at the Gen 2 data rate, asserting the rx_digitalreset signal causes the PCIe rateswitch circuitry to switch the transceiver to Gen 1 data rate. 1 When switching from Gen1 to Gen2 using the dynamic reconfiguration controller, you must set the two ports of the dynamic reconfiguration controller, tx_preemp_0t and tx_preemp_2t, to zero to meet the Gen2 de-emphasis specifications. When switching from Gen2 to Gen1, if your system requires specific settings on tx_preemp_01 and tx_preemp_2t, those values must be set at the respective two ports of the dynamic reconfiguration controller to meet your system requirements. PCIe Clock Switch Circuitry When the PHY-MAC layer instructs a rateswitch between the Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rates, both the transmitter high-speed serial and low-speed parallel clock and the CDR recovered clock must switch to support the instructed data rate. Stratix IV GX and GT transceivers have dedicated PCIe clock switch circuitry located in the following blocks: Stratix IV Device Handbook Volume 2: Transceivers Local clock divider in transmitter PMA of each transceiver channel CMU0 clock divider in CMU0_Channel of each transceiver block Receiver CDR in receiver PMA of each transceiver channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-143 PCIe transmitter high-speed serial and low-speed parallel clock switch occurs: In PCIe x1 mode, the CMU_PLL clock switch occurs in the local clock divider in each transceiver channel. In PCIe x4 mode, the CMU_PLL clock switch occurs in the CMU0 clock divider in the CMU0_Channel within the transceiver block. In PCIe x8 mode, the CMU_PLL clock switch occurs in the CMU0 clock divider in the CMU0_Channel within the master transceiver block. In PCIe x1, x4, and x8 modes, the recovered clock switch happens in the receiver CDR of each transceiver channel. Table 1-53 lists the locations of the PCIe rateswitch controller and the PCIe clock switch circuitry in PCIe x1, x4, and x8 modes. Table 1-53. PCIe Rateswitch Controller and Clock Switch Circuitry Location of PCIe Clock Switch Circuitry Channel Bonding Option Location of PCIe Rateswitch Controller Module Transmitter High-Speed Serial and Low-Speed Parallel Clock Switch Circuitry Recovered Clock Switch Circuitry x1 Individual channel PCS block Local clock divider in transmitter PMA of each channel CDR block in receiver PMA of each channel x4 CMU0_Channel CMU0 clock divider in CMU0_Channel CDR block in receiver PMA of each channel x8 CMU0_Channel of the master transceiver block CMU0 clock divider in CMU0_Channel of the master transceiver block CDR block in receiver PMA of each channel September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-144 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Dynamic Switch Between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) Signaling Rates in PCIe x1 Mode Figure 1-114 shows the PCIe rateswitch circuitry in PCIe x1 mode configured at Gen2 (5 Gbps) data rate. Figure 1-114. Dynamic Switch Signaling in PIPE x1 Mode Transceiver Channel Transceiver PCS Receiver Phase Compensation FIFO Clock and Data Recovery (CDR) Unit rx_locktorefclk rx_locktodata signal detect rx_freqlocked Phase Detector (PD) rx_datain rx_cruclk Transmitter Phase Compensation FIFO FPGA Fabric Parallel Recovered Clock 1 0 reset_int PIPE Interface Charge Pump + Loop Filter /1, /2, /4 /2 pipephydonestatus /2 pcie_gen2switch VCO /L Phase Frequency Detector (PD) reset_int PCI Express Rate Switch Controller Serial Recovered Clock LTR/LTD Controller /M Local Clock Divider pcie_gen2switch pcie_gen2switch_done CMU0_PLL Output Clock PCI Express Clock Switch Circuitry High-Speed Serial Clock /4, /5, /8, /10 Low-Speed Parallel Clock /1, /2, /4 rateswitch CMU1_PLL Output Clock PCI Express Clock Switch Circuitry In PCIe x1 mode configured at Gen2 (5 Gbps) data rate, when the PCIe rateswitch controller sees a transition on the rateswitch signal, it sends control signal pcie_gen2switch to the PCIe clock switch circuitry in the local clock divider block and the receiver CDR to switch to the instructed signaling rate. A low-to-high transition on the rateswitch signal initiates a Gen1 (2.5 Gbps) to Gen2 (5 Gbps) signaling rateswitch. A high-to-low transition on the rateswitch signal initiates a Gen2 (5 Gbps) to Gen1 (2.5 Gbps) signaling rateswitch. Table 1-54 lists the transceiver clock frequencies when switching between 2.5 Gbps and 5 Gbps signaling rates. Table 1-54. Transceiver Clock Frequencies Signaling Rates in PCIe x1 Mode Gen1 (2.5 Gbps) to Gen2 (5 Gbps) Switch (Low-to-High Transition on the rateswitch Signal) Gen2 (5 Gbps) to Gen1 (2.5 Gbps) Switch (High-to-Low Transition on the rateswitch Signal) High-Speed Serial Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Low-Speed Parallel Clock 250 MHz to 500 MHz 500 MHz to 250 MHz Serial Recovered Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Parallel Recovered Clock 250 MHz to 500 MHz 500 MHz to 250 MHz FPGA Fabric-Transceiver Interface Clock 125 MHz to 250 MHz 250 MHz to 125 MHz Transceiver Clocks Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-145 The PCIe clock switch circuitry in the local clock divider block performs the clock switch between 250 MHz and 500 MHz on the low-speed parallel clock when switching between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rates. It indicates successful completion of clock switch on the pcie_gen2switchdone signal to the PCIe rateswitch controller. The PCIe rateswitch controller forwards the clock switch completion status to the PCIe interface block. The PCIe interface block communicates the clock switch completion status to the PHY-MAC layer by asserting the pipephydonestatus signal for one parallel clock cycle. Figure 1-115 shows the low-speed parallel clock switch between Gen1 (250 MHz) and Gen2 (500 MHz) in response to the change in the logic level on the rateswitch signal. The rateswitch completion is shown marked with a one clock cycle assertion of the pipephydonestatus signal. 1 Time T1 from a transition on the rateswitch signal to the assertion of pipephydonestatus is pending characterization. Figure 1-115. Low-Speed Parallel Clock Switching in PCIe x1 Mode 250 MHz (Gen1) 500 MHz (Gen2) 250 MHz (Gen1) Low-Speed Parallel Clock rateswitch pipephydonestatus T1 T1 As a result of the signaling rateswitch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps), the FPGA fabric-transceiver interface clock switches between 125 MHz and 250 MHz. The FPGA fabric-transceiver interface clock clocks the read side and write side of the transmitter phase compensation FIFO and the receiver phase compensation FIFO, respectively. It is also routed to the FPGA fabric on a global or regional clock resource and looped back to clock the write port and read port of the transmitter phase compensation FIFO and the receiver phase compensation FIFO, respectively. Due to the routing delay between the write and read clock of the transmitter and receiver phase compensation FIFOs, the write pointers and read pointers might collide during a rateswitch between 125 MHz and 250 MHz. To avoid collision of the phase compensation FIFO pointers, the PCIe rateswitch controller automatically disables and resets the pointers during clock switch. When the PCIe clock switch circuitry in the local clock divider indicates successful clock switch completion, the PCIe rateswitch controller releases the phase compensation FIFO pointer resets. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-146 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Dynamic Switch Between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) Signaling Rates in PCIe x4 Mode Figure 1-116 shows the PCIe rateswitch circuitry in PCIe x4 mode configured at Gen2 (5 Gbps) data rate. Figure 1-116. Dynamic Switch Signaling in PCIe x4 Mode Transceiver Block Transceiver PCS Receiver Phase Compensation FIFO LTR/LTD Controller 0 Phase Detector (PD) Transmitter Phase Compensation FIFO pipephydonestatus[3:0] /2 /2 pcie_gen2switch Charge Pump + Loop Filter /1, /2, /4 rx_cruclk FPGA Fabric Parallel Recovered Clock 1 rx_datain reset_int PIPE Interface Serial Recovered Clock Clock and Data Recovery (CDR) Unit rx_locktorefclk rx_locktodata signal detect rx_freqlocked VCO /L Phase Frequency Detector (PD) reset_int /M CMU0_Channel High-Speed Serial Clock to the Four (PIPE x4) Bonded Channels CMU0 Clock Divider PCI Express Rate Switch Controller pcie_gen2switch PCI Express Clock Switch Circuitry pcie_gen2switch_done CCU CMU0 PLL /4, /5, /8, /10 /1, /2, /4 rateswitch CMU1_Channel CMU1 Clock Divider CMU1 PLL /1, /2, /4 Low-Speed Parallel Clock to the Four (PIPE x4) Bonded Channels /4, /5, /8, /10 PCI Express Clock Switch Circuitry In PCIe x4 mode configured at Gen2 (5 Gbps) data rate, when the PCIe rateswitch controller sees a transition on the rateswitch signal, it sends the pcie_gen2switch control signal to the PCIe clock switch circuitry in the CMU0 clock divider block and the receiver CDR to switch to the instructed signaling rate. A low-to-high transition on the rateswitch signal initiates a Gen1 (2.5 Gbps) to Gen2 (5 Gbps) signaling rateswitch. A high-to-low transition on the rateswitch signal initiates a Gen2 (5 Gbps) to Gen1 (2.5 Gbps) signaling rateswitch. Table 1-55 lists the transceiver clock frequencies when switching between the 2.5 Gbps and 5 Gbps signaling rates. Table 1-55. Transceiver Clock Frequencies Signaling Rates in PCIe x4 Mode (Part 1 of 2) Gen1 (2.5 Gbps) to Gen2 (5 Gbps) Switch (Low-to-High Transition on the rateswitch Signal) Gen2 (5 Gbps) to Gen1 (2.5 Gbps) Switch (High-to-Low Transition on the rateswitch Signal) High-Speed Serial Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Low-Speed Parallel Clock 250 MHz to 500 MHz 500 MHz to 250 MHz Serial Recovered Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Transceiver Clocks Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-147 Table 1-55. Transceiver Clock Frequencies Signaling Rates in PCIe x4 Mode (Part 2 of 2) Gen1 (2.5 Gbps) to Gen2 (5 Gbps) Switch (Low-to-High Transition on the rateswitch Signal) Gen2 (5 Gbps) to Gen1 (2.5 Gbps) Switch (High-to-Low Transition on the rateswitch Signal) Parallel Recovered Clock 250 MHz to 500 MHz 500 MHz to 250 MHz FPGA Fabric-Transceiver Interface Clock 125 MHz to 250 MHz 250 MHz to 125 MHz Transceiver Clocks The PCIe clock switch circuitry in the CMU0 clock divider block performs the clock switch between 250 MHz and 500 MHz on the low-speed parallel clock when switching between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rates. It indicates successful completion of clock switch on the pcie_gen2switchdone signal to the PCIe rateswitch controller. The PCIe rateswitch controller forwards the clock switch completion status to the PCIe interface block. The PCIe interface block communicates the clock switch completion status to the PHY-MAC layer by asserting the pipephydonestatus signal of all bonded channels for one parallel clock cycle. Figure 1-117 shows the low-speed parallel clock switch between Gen1 (250 MHz) and Gen2 (500 MHz) in response to the change in the logic level on the rateswitch signal. The rateswitch completion is shown marked with a one clock cycle assertion of the pipephydonestatus signal of all bonded channels. 1 Time T1 from a transition on the rateswitch signal to the assertion of pipephydonestatus is pending characterization. Figure 1-117. Low-Speed Parallel Clock Switching in PCIe x4 Mode 250 MHz (Gen1) 500 MHz (Gen2) 250 MHz (Gen1) Low-Speed Parallel Clock rateswitch pipephydonestatus[3] pipephydonestatus[0] T1 T1 As a result of the signaling rateswitch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps), the FPGA fabric-transceiver interface clock switches between 125 MHz and 250 MHz. The FPGA fabric-transceiver interface clock clocks the read side and write side of the transmitter phase compensation FIFO and the receiver phase compensation FIFO of all bonded channels, respectively. It is also routed to the FPGA fabric on a global or regional clock resource and looped back to clock the write port and read port of the transmitter phase compensation FIFO and the receiver phase compensation FIFO, respectively. Due to the routing delay between the write and read clock of the transmitter and receiver phase compensation FIFOs, the write pointers and read pointers might collide during a rateswitch between 125 MHz and 250 MHz. To avoid September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-148 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture collision of the phase compensation FIFO pointers, the PCIe rateswitch controller automatically disables and resets the phase compensation FIFO pointers of all bonded channels during clock switch. When the PCIe clock switch circuitry in the local clock divider indicates successful clock switch completion, the PCIe rateswitch controller releases the phase compensation FIFO pointer resets. Dynamic Switch Between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) Signaling Rates in PCIe x8 Mode Figure 1-118 shows the PCIe rateswitch circuitry in PCIe x8 mode configured at Gen2 (5 Gbps) data rate. Figure 1-118. Dynamic Switch Signaling in PCIe x8 Mode Slave Transceiver Block Transceiver PCS Receiver Phase Compensation FIFO rx_cruclk Parallel Recovered Clock 1 0 Transmitter Phase Compensation FIFO pipephydonestatus[7:4] LTR/LTD Controller Phase Detector (PD) rx_datain reset_int PIPE Interface Serial Recovered Clock Clock and Data Recovery (CDR) Unit rx_locktorefclk rx_locktodata signal detect rx_freqlocked Charge Pump + Loop Filter /1, /2, /4 /2 /2 rateswitch_asn /L VCO Phase Frequency Detector (PD) rateswitch_asn reset_int /M Master Transceiver Block FPGA Fabric Transceiver PCS Receiver Phase Compensation FIFO rx_cruclk Charge Pump + Loop Filter /L VCO Phase Frequency Detector (PD) reset_int /M pcie_gen2switch CMU0_Channel PCI Express Rate Switch Controller High-Speed Serial Clock to the Eight Bonded Channels in the Master and Slave Transceiver Blocks CMU0 Clock Divider pcie_gen2switch pcie_gen2switch_done CCU rateswitch /2 pcie_gen2switch /1, /2, /4 /2 Parallel Recovered Clock 1 0 Transmitter Phase Compensation FIFO pipephydonestatus[3:0] LTR/LTD Controller Phase Detector (PD) rx_datain reset_int PIPE Interface Serial Recovered Clock Clock and Data Recovery (CDR) Unit rx_locktorefclk rx_locktodata signal detect rx_freqlocked CMU0 PLL PCI Express Clock Switch Circuitry /4, /5, /8, /10 /1, /2, /4 CMU1_Channel CMU1 Clock Divider CMU1 PLL /1, /2, /4 Low-Speed Parallel Clock to the Eight Bonded Channels in the Master and Slave Transceiver Blocks /4, /5, /8, /10 PCI Express Clock Switch Circuitry Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-149 In PCIe x8 mode configured at 5 Gbps data rate, when the PCIe rateswitch controller sees a transition on the rateswitch signal, it sends the pcie_gen2switch control signal to the PCIe clock switch circuitry in the CMU0 clock divider of the master transceiver block and the receiver CDR in all eight bonded channels to switch to the instructed signaling rate. A low-to-high transition on the rateswitch signal initiates a Gen1 (2.5 Gbps) to Gen2 (5 Gbps) signaling rateswitch. A high-to-low transition on the rateswitch signal initiates a Gen2 (5 Gbps) to Gen1 (2.5 Gbps) signaling rateswitch. Table 1-56 lists the transceiver clock frequencies when switching between the 2.5 Gbps and 5 Gbps signaling rates. Table 1-56. Transceiver Clock Frequencies Signaling Rates in PCIe x8 Mode Transceiver Clocks Gen1 (2.5 Gbps) to Gen 2 (5 Gbps) Switch (Low-to-High Transition on the rateswitch Signal) Gen2 (5 Gbps) to Gen1 (2.5 Gbps) Switch (High-to-Low Transition on the rateswitch Signal) High-Speed Serial Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Low-Speed Parallel Clock 250 MHz to 500 MHz 500 MHz to 250 MHz Serial Recovered Clock 1.25 GHz to 2.5 GHz 2.5 GHz to 1.25 GHz Parallel Recovered Clock 250 MHz to 500 MHz 500 MHz to 250 MHz FPGA Fabric-Transceiver Interface Clock 125 MHz to 250 MHz 250 MHz to 125 MHz The PCIe clock switch circuitry in the CMU0 clock divider of the master transceiver block performs the clock switch between 250 MHz and 500 MHz on the low-speed parallel clock when switching between Gen1 (2.5 Gbps) and Gen2 (5 Gbps) signaling rates. It indicates successful completion of clock switch on the pcie_gen2switchdone signal to the PCIe rateswitch controller. The PCIe rateswitch controller forwards the clock switch completion status to the PCIe interface block. The PCIe interface block communicates the clock switch completion status to the PHY-MAC layer by asserting the pipephydonestatus signal of all eight bonded channels for one parallel clock cycle. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-150 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-119 shows the low-speed parallel clock switch between Gen1 (250 MHz) and Gen2 (500 MHz) in response to the change in the logic level on the rateswitch signal. The rateswitch completion is shown marked with a one clock cycle assertion of the pipephydonestatus signal of all eight bonded channels. 1 Time T1 from a transition on the rateswitch signal to the assertion of pipephydonestatus is pending characterization. Figure 1-119. Low-Speed Parallel Clock Switching in PCIe x8 Mode 250 MHz (Gen1) 500 MHz (Gen2) 250 MHz (Gen1) Low-Speed Parallel Clock rateswitch pipephydonestatus[7] pipephydonestatus[0] T1 T1 As a result of the signaling rateswitch between Gen1 (2.5 Gbps) and Gen2 (5 Gbps), the FPGA fabric-transceiver interface clock switches between 125 MHz and 250 MHz. The FPGA fabric-transceiver interface clock clocks the read side and write side of the transmitter phase compensation FIFO and the receiver phase compensation FIFO of all eight bonded channels, respectively. It is also routed to the FPGA fabric on a global or regional clock resource and looped back to clock the write port and read port of the transmitter phase compensation FIFO and the receiver phase compensation FIFO, respectively. Due to the routing delay between the write and read clock of the transmitter and receiver phase compensation FIFOs, the write pointers and read pointers might collide during a rateswitch between 125 MHz and 250 MHz. To avoid collision of the phase compensation FIFO pointers, the PCIe rateswitch controller automatically disables and resets the phase compensation FIFO pointers of all eight bonded channels during clock switch. When the PCIe clock switch circuitry in the local clock divider indicates successful clock switch completion, the PCIe rateswitch controller releases the phase compensation FIFO pointer resets. PCIe Cold Reset Requirements The PCIe Base Specification 2.0 defines the following three types of conventional resets to the PCIe system components: Stratix IV Device Handbook Volume 2: Transceivers Cold reset--fundamental reset after power up Warm reset--fundamental reset without removal and re-application of power Hot reset--In-band conventional reset initiated by the higher layer by setting the Hot Reset bit in the TS1 or TS2 training sequences September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-151 Fundamental reset is provided by the system to the component or adapter card using the auxiliary signal PERST#. The PCIe Base Specification 2.0 specifies that PERST# must be kept asserted for a minimum of 100 ms (TPVPERL) after the system power becomes stable in a cold reset situation. Additionally, all system components must enter the LTSSM Detect state within 20 ms and the link must become active within 100 ms after de-assertion of the PERST# signal. This implies that each PCIe system component must become active within 100 ms after PERST# is deasserted. 1 The link being active is interpreted as the physical layer device coming out of electrical idle in the L0 state of the LTSSM state machine. Figure 1-120 lists the PCIe cold reset timing requirements. Figure 1-120. PCIe Cold Reset Requirements 1 2 3 4 Power Rail Marker 1: Power becomes stable PERST# Marker 2: PERST# gets de-asserted TPVPERL 100 ms T2-3 d"20 ms T2-4 d"100 ms Marker 3: Maximum time for Marker 2 for the LTSSM to enter the Detect state Marker 4: Maximum time for Marker 2 for the link to become active The time taken by a PCIe port implemented using the Stratix IV GX and GT device to go from power up to link active state is described below: Power on reset (POR)--begins after power rails become stable. Typically takes 12 ms FPGA configuration/programming--begins after POR. Configuration time depends on the FPGA density Time taken from de-assertion of PERST# to link active--typically takes 40 ms (pending characterization and verification of PCIe soft IP and hard IP) To meet the PCIe specification of 200 ms from power on to link active, the Stratix IV GX and GT device configuration time must be less than 148 ms (200 ms -12 ms for power on reset and -40 ms for the link to become active after PERST# de-assertion). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-152 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-57 lists the typical configuration times for Stratix IV GX devices when configured using the Fast Passive Parallel (FPP) configuration scheme at 125 MHz. Table 1-57. Typical Configuration Times for Stratix IV GX Devices Configured with Fast Passive Parallel Stratix IV GX Stratix IV GT Configuration Time (ms) EP4SGX70 -- 48 EP4SGX110 -- 48 EP4SGX230 EP4S(40/100)G2 95 EP4SGX290 EP4S100G3 128 EP4SGX360 EP4S100G4 128 EP4SGX530 EP4S(40/100)G5 172 f For more information about the FPP configuration scheme, refer to the Configuration, Design Security, Remote System Upgrades in Stratix IV Devices chapter. 1 Most flash memories available can run up to 100 MHz. To configure the Stratix IV GX and GT device at 125 MHz, Altera recommends using a MAX II device to convert the 16-bit flash memory output at 62.5 MHz to 8-bit configuration data input to the Stratix IV GX and GT device at 125 MHz. PCI Express Electrical Gold Test with Compliance Base Board (CBB) The PCI Express Electrical Gold Test requires the v2.0 CBB to be connected to the Device Under Test (DUT). The CBB sends out a 100 MHz signal for 1 ms to indicate the Link Training and Status State Machine (LTSSM) of the downstream device Under Test (DUT) to transition to several polling compliance states. Under these states, the DUT sends out data at Gen1, Gen2 (with -3.5db de-emphasis), and Gen2 (with -6 db de-emphasis) rates, which can be observed in the scope to confirm electrical signal compliance. The CBB is DC-coupled to the downstream receiver. When you use the Stratix IV GX and GT device as DUT, because of being DC-coupled to CBB with a different common mode level, the Stratix IV GX and GT receiver does not receive the required VCM (0.85 V) to detect the signal. The logic in the FPGA fabric that implements LTSSM cannot transition to the multiple polling compliance states to complete the test. Therefore, when testing with the CBB, force the LTSSM implemented in the FPGA fabric to transfer to different polling compliance states using an external push button or user logic. If you use the Stratix IV GX and GT PCIe hard IP block, assert the test_in[6] port of the PCIe Compiler-generated wrapper file in your design. Asserting this port forces the LTSSM within the hard IP block to transition to these states. The test_in[6] port must be asserted for a minimum of 16 ns and less than 24 ms. f For more information about the PCIe hard IP block, refer to the PCI Express Compiler User Guide. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-153 XAUI Mode XAUI is an optional, self-managed interface that you can insert between the reconciliation sublayer and the PHY layer to transparently extend the physical reach of the XGMII. XAUI addresses several physical limitations of the XGMII. XGMII signaling is based on the HSTL Class 1 single-ended I/O standard, which has an electrical distance limitation of approximately 7 cm. Because XAUI uses a low-voltage differential signaling method, the electrical limitation is increased to approximately 50 cm. Another advantage of XAUI is simplification of backplane and board trace routing. XGMII is composed of 32 transmit channels, 32 receive channels, 1 transmit clock, 1 receive clock, 4 transmitter control characters, and 4 receive control characters for a 74-pin wide interface. XAUI, on the other hand, only consists of 4 differential transmitter channels and 4 differential receiver channels for a 16-pin wide interface. This reduction in pin count significantly simplifies the routing process in the layout design. Figure 1-121 shows the relationships between the XGMII and XAUI layers. Figure 1-121. XAUI and XGMII Layers LAN Carrier Sense Multiple Access/Collision Detect (CSMA/CD) Layers Higher Layers Logical Link Control (LLC) OSI Reference Model Layers MAC Control (Optional) Media Access Control (MAC) Application Reconciliation Presentation Session Transport 10 Gigabit Media Independent Interface Optional XGMII Extender XGMII Extender Sublayer 10 Gigabit Attachment Unit Interface XGMII Extender Sublayer 10 Gigabit Media Independent Interface Network PCS Data Link Physical PMA Physical Layer Device PMD Medium Dependent Interface Medium 10 Gb/s September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-154 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The XGMII interface consists of four lanes of 8 bits. At the transmit side of the XAUI interface, the data and control characters are converted within the XGXS into an 8B/10B encoded data stream. Each data stream is then transmitted across a single differential pair running at 3.125 Gbps (3.75 Gbps for HiGig). At the XAUI receiver, the incoming data is decoded and mapped back to the 32-bit XGMII format. This provides a transparent extension of the physical reach of the XGMII and also reduces the interface pin count. In Stratix IV GX and GT XAUI functional mode, the interface between the transceiver and FPGA fabric is 64 bits wide (four channels of 16 bits each) at single data rate. XAUI functions as a self-managed interface because code group synchronization, channel deskew, and clock domain decoupling is handled with no upper layer support requirements. This functionality is based on the PCS code groups that are used during the IPG time and idle periods. PCS code groups are mapped by the XGXS to XGMII characters, as listed in Table 1-58. Table 1-58. XGMII Character to PCS Code-Group Mapping XGMII TXC XGMII TXD (1) PCD Code Group Description 0 00 through FF Dxx,y Normal data transmission 1 07 K28.0 or K28.3 or K28.5 Idle in ||I|| 1 07 K28.5 Idle in ||T|| 1 9C K28.4 Sequence 1 FB K27.7 Start 1 FD K29.7 Terminate 1 FE K30.7 Error 1 Any other value K30.7 Invalid XGMII character Note to Table 1-58: (1) The values in the XGMII TXD column are in hexadecimal. Figure 1-122 shows an example of mapping between XGMII characters and the PCS code groups that are used in XAUI. The idle characters are mapped to a pseudo-random sequence of /A/, /R/, and /K/ code groups. Figure 1-122. Example of Mapping XGMII Characters to PCS Code Groups XGMII T/RxD<7..0> | | S Dp D D D --- D D D D | | | | | | T/RxD<15..8> | | Dp Dp D D D --- D D D T | | | | | | T/RxD<23..16> | | Dp Dp D D D --- D D D | | | | | | | T/RxD<31..24> | | Dp Dp D D D --- D D D | | | | | | | R S Dp D D D --- D D D D A R R K K R Dp D D D --- D D D T A R R K K R PCS Lane 0 K Lane 1 K R Dp Lane 2 K R Dp Dp D D D --- D D D K A R R K K R Lane 3 K R Dp Dp D D D --- D D D K A R R K K R Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-155 PCS code groups are sent via PCS ordered sets. PCS ordered sets consist of combinations of special and data code groups defined as a column of code groups. These ordered sets are composed of four code groups beginning in lane 0. Table 1-59 lists the defined idle ordered sets (||I||) that are used for the self-managed properties of XAUI. Table 1-59. Defined Idle Ordered Set Code Ordered Set ||I|| Encoding Idle Number of Code Groups Substitute for XGMII Idle ||K|| Synchronization column 4 /K28.5/K28.5/K28.5/K28.5/ ||R|| Skip column 4 /K28.0/K28.0/K28.0/K28.0/ ||A|| Align column 4 /K28.3/K28.3/K28.3/K28.3/ Stratix IV GX and GT transceivers configured in XAUI mode provide the following protocol features: September 2012 XGMII-to-PCS code conversion at the transmitter PCS-to-XGMII code conversion at the receiver 8B/10B encoding and decoding IEEE P802.3ae-compliant synchronization state machine 100 PPM clock rate compensation Channel deskew of four lanes of the XAUI link Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-156 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-123 shows the XAUI mode configuration supported in Stratix IV GX and GT devices. Figure 1-123. Stratix IV GX and GT XAUI Mode Configuration Stratix IV GX and GT Configurations Protocol Basic Single Width Functional Modes PMA-PCS Interface Width 8-bit 10-bit Double Width 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Functional Mode XAUI Data Rate (Gbps) 3.125 - 3.75 Channel Bonding x4 Low-Latency PCS Disabled Word Aligner (Pattern Length) Automatic Synchronization State Machine (10-Bit/K28.5/) Deskew FIFO Enabled 8B/10B Encoder/Decoder Enabled Rate Match FIFO Enabled Byte Ordering Disabled FPGA Fabric-Transceiver Interface Width 16-Bit FPGA Fabric-Transceiver Interface Frequency (MHz) 156.25187.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Stratix IV Device Handbook Volume 2: Transceivers 10-Bit 20-Bit Enabled Byte SerDes TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency Deterministic Latency 4.5 - 6 14.5 18 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-157 XAUI Mode Datapath Figure 1-124 shows the ALTGX megafunction transceiver datapath when configured in XAUI mode. Figure 1-124. Transceiver Datapath in XAUI Mode Channel 3 Channel 2 Transmitter Channel PCS TX Phase Compensation FIFO wrclk Byte Serializer rdclk 8B/10B Encoder Transmitter Channel PMA Serializer rdclk wrclk tx_coreclk[3:2] /2 Low-Speed Parallel Clock from CMU 0 Click Divider Channel 3 Channel 2 Receiver Channel PCS Byte Deserializer RX Phase Compensation FIFO 8B/10B Decoder Rate Match FIFO Deskew FIFO Word Aligner Receiver Channel PMA Deserializer CDR rx_coreclk[3:2] Ch0 Parallel Recovered Clock Ch2 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU 0 Clock Divider coreclkout FPGA Fabric-Transceiver Interface Clock CMU1_Channel /2 Input Reference Clock CMU1_PLL CMU1 Clock Divider CMU0_PLL CMU0 Clock Divider FPGA Fabric CMU0_Channel Input Reference Clock Low-Speed Parallel Clock High-Speed Serial Clock Channel 1 Channel 0 Transmitter Channel PCS Byte Serializer TX Phase Compensation FIFO wrclk Serializer rdclk wrclk rdclk 8B/10B Encoder Transmitter Channel PMA tx_coreclk[1:0] /2 Low-Speed Parallel Clock from CMU 0 Clock Divider Channel 1 Receiver Channel PCS Channel 0 RX Phase Compensation FIFO Byte Deserializer rx_coreclk[1:0] 8B/10B Decoder Rate Match FIFO Deskew FIFO Word Aligner Ch0 Parallel Recovered Clock Receiver Channel PMA DeSerializer CDR Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU 0 Clock Divider September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-158 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture XGMII-To-PCS Code Conversion at the Transmitter In XAUI mode, the 8B/10B encoder in the Stratix IV GX and GT transmitter datapath is controlled by a transmitter state machine that maps various 8-bit XGMII codes to 10-bit PCS code groups. This state machine complies with the IEEE P802.3ae PCS transmit source state diagram shown in Figure 1-125. Figure 1-125. XGMII-To-PCS Code Conversion in XAUI Mode (1) !reset !(TX=||IDLE|| + TX=||Q|| SEND_Q IF TX=||T|| THEN cvtx_terminate tx_code_group<39:0> ENCODE(TX) PUDR (next_ifg + A_CNT0) next_ifg = A_CNT0 reset SEND_A SEND_K tx_code_group<39:0> ||A|| next_ifg K PUDR Q_det tx_code_group<39:0> ||K|| next_ifg A PUDR !Q_det UCT B SEND_Q tx_code_group<39:0> TQMSG Q_det K PUDR A A_CNT0 * cod_sel=1 B UCT SEND_RANDOM_K tx_code_group<39:0> ||K|| PUDR SEND_RANDOM_R tx_code_group<39:0> ||R|| A_CNT0 * cod_sel=1 B A_CNT0 * cod_sel=1 A A_CNT=0 A_CNT=0 A SEND_RANDOM_A A_CNT0 * cod_sel=1 tx_code_group<39:0> ||A|| PUDR Q_det B !Q_det * cod_sel=1 SEND_RANDOM_Q tx_code_group<39:0> TQMSG Q_det FALSE PUDR B A A !Q_det * cod_set=1 cod_set=1 cod_set=1 Note to Figure 1-125: (1) This figure is from IEEE P802.3ae. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-159 Table 1-60 lists the XGMII-to-PCS code group conversion in XAUI functional mode. The XGMII TXC control signal is equivalent to the tx_ctrlenable signal; the XGMII TXD control signal is equivalent to the tx_datain[7:0] signal. Table 1-60. XGMII Character to PCS Code-Group Mapping XGMII TXC XGMII TXD (1) PCD Code Group Description 0 00 through FF Dxx,y Normal data transmission 1 07 K28.0 or K28.3 or K28.5 Idle in ||I|| 1 07 K28.5 Idle in ||T|| 1 9C K28.4 Sequence 1 FB K27.7 Start 1 FD K29.7 Terminate 1 FE K30.7 Error 1 Any other value K30.7 Invalid XGMII character Note to Table 1-58: (1) The values in the XGMII TXD column are in hexadecimal. PCS-To-XGMII Code Conversion at the Receiver In XAUI mode, the 8B/10B decoder in the Stratix IV GX and GT receiver datapath is controlled by a XAUI receiver state machine that converts received PCS code groups into specific 8-bit XGMII codes. This state machine complies with the IEEE P802.3ae specifications. Table 1-61 lists the PCS-to-XGMII code group conversion in XAUI functional mode. The XGMII RXC control signal is equivalent to the rx_ctrldetect signal; the XGMII RXD control signal is equivalent to the rx_dataout[7:0] signal. Table 1-61. PCS Code Group to XGMII Character Mapping XGMII RXC XGMII RXD (1) PCD Code Group Description Normal data transmission 0 00 through FF Dxx,y 1 07 K28.0 or K28.3 or K28.5 Idle in ||I|| 1 07 K28.5 Idle in ||T|| 1 9C K28.4 Sequence 1 FB K27.7 Start 1 FD K29.7 Terminate 1 FE K30.7 Error 1 FE Invalid code group Received code group Note to Table 1-58: (1) The values in the XGMII RXD column are in hexadecimal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-160 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Word Aligner The word aligner in XAUI functional mode is configured in automatic synchronization state machine mode. The Quartus II software automatically configures the synchronization state machine to indicate synchronization when the receiver receives four /K28.5/ comma code groups without intermediate invalid code groups. The synchronization state machine implemented in XAUI mode is compliant to the PCS synchronization state diagram specified in Clause 48 of the IEEE P802.3ae specification and is shown in Figure 1-126. Figure 1-126. IEEE 802.3ae PCS Synchronization State Diagram (1) power_on=TRUE+mr_main_rest=TRUE + (signal_detectCHANGE=TRUE + mr_loopback=FALSE +PUDI) LOSS_OF_SYNC [PUDI * signal_detect=FAIL + mr_loopback=FALSE] + PUDI(![/COMMA/]) sync_status FAIL rx_even ! rx_even SUDI (signal_detect=OK+mr_loopback=TRUE)* * PUDI([/COMMA/] COMMA_DETECT_1 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] ACQUIRE_SYNC_1 PUDI(![/COMMA/] *[/INVALID/] rx_even ! rx_even SUDI cgbad rx_even=FALSE+PUDI([/COMMA/] COMMA_DETECT_2 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] ACQUIRE_SYNC_2 PUDI(![/COMMA/] *[/INVALID/] rx_even ! rx_even SUDI cgbad rx_even=FALSE+PUDI([/COMMA/] COMMA_DETECT_3 SYNC_ACQUIRED_1 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] sync_status OK rx_even ! rx_even SUDI cgbad cggood 2 cggood SYNC_ACQUIRED_2 SYNC_ACQUIRED_2A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cgbad cggood *good_cgs = 3 cgbad cggood *good_cgs = 3 3 cggood SYNC_ACQUIRED_3 SYNC_ACQUIRED_3A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cggood *good_cgs = 3 cgbad cgbad 2 cggood *good_cgs = 3 cggood SYNC_ACQUIRED_4 SYNC_ACQUIRED_4A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cggood *good_cgs = 3 cgbad cgbad 3 cggood *good_cgs = 3 Note to Figure 1-126: (1) This figure is from IEEE P802.3ae. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-161 Receiver synchronization is indicated on the rx_syncstatus port of each channel. A high on the rx_syncstatus port indicates that the lane is synchronized; a low on the rx_syncstatus port indicates that it has fallen out of synchronization. The receiver loses synchronization when it detects four invalid code groups separated by less than four valid code groups or when it is reset. Deskew FIFO Code groups received across four lanes in a XAUI link can be misaligned with respect to one another because of skew in the physical medium or differences between the independent clock recoveries per lane. The XAUI protocol allows a maximum skew of 40 UI (12.8 ns) as seen at the receiver of the four lanes. The XAUI protocol requires the physical layer device to implement a deskew circuitry to align all four channels. To enable the deskew circuitry at the receiver to align the four channels, the transmitter sends a /A/ (/K28.3/) code group simultaneously on all four channels during inter-packet gap. The skew introduced in the physical medium and the receiver channels can be /A/ code groups to be received misaligned with respect to each other. The deskew operation is performed by the deskew FIFO in XAUI functional mode. The deskew FIFO in each channel receives data from its word aligner. The deskew operation begins only after link synchronization is achieved on all four channels as indicated by a high on the rx_syncstatus signal from the word aligner in each channel. Until the first /A/ code group is received, the deskew FIFO read and write pointers in each channel are not incremented. After the first /A/ code group is received, the write pointer starts incrementing for each word received but the read pointer is frozen. If the /A/ code group is received on each of the four channels within 10 recovered clock cycles of each other, the read pointer of all four deskew FIFOs is released simultaneously, aligning all four channels. Figure 1-127 shows lane skew at the receiver input and how the deskew FIFO uses the /A/ code group to align the channels. Figure 1-127. Receiver Input Lane Skew in XAUI Mode Lane 0 K Lane 2 K R A K R R K K R K R Lane 1 K K R A K R R K K R K K R A K R R K K R K R K K R A K R R K K R K K Lane 3 September 2012 R Lane Skew at Receiver Input R Lane 0 K K R A K R R K K R K R Lane 1 K K R A K R R K K R K R Lane 2 K K R A K R R K K R K R Lane 3 K K R A K R R K K R K R Altera Corporation Lanes are Deskewed by Lining up the "Align"/A/, Code Groups Stratix IV Device Handbook Volume 2: Transceivers 1-162 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture After alignment of the first ||A|| column, if three additional aligned ||A|| columns are observed at the output of the deskew FIFOs of the four channels, the rx_channelaligned signal is asserted high, indicating channel alignment is acquired. After acquiring channel alignment, if four misaligned ||A|| columns are seen at the output of the deskew FIFOs in all four channels with no aligned ||A|| columns in between, the rx_channelaligned signal is de-asserted low, indicating loss-of-channel alignment. The deskew FIFO operation in XAUI functional mode is compliant with the PCS deskew state machine diagram specified in clause 48 of the IEEE P802.3ae, as shown in Figure 1-128. Figure 1-128. Deskew FIFO in XAUI Mode (1) reset + (sync_status=FAIL * SUDI) LOSS_OF_ALIGNMENT SUDI(![/||A||/]) align_status FAIL enable_deskew TRUE AUDI sync_status OK * SUDI(![/||A||/]) ALIGN_DETECT_1 !deskew_error * SUDI(![/||A||/]) enable_deskew FALSE AUDI deskew_error * SUDI SUDI(![/||A||/]) ALIGN_DETECT_2 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI SUDI(![/||A||/]) ALIGN_DETECT_3 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI A SUDI(![/||A||/]) ALIGN_ACQUIRED_1 !deskew_error * SUDI(![/||A||/]) enable_deskew FALSE AUDI deskew_error * SUDI B SUDI(![/||A||/]) ALIGN_ACQUIRED_2 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI C SUDI(![/||A||/]) A ALIGN_ACQUIRED_3 AUDI !deskew_error * SUDI(![/||A||/]) deskew_error * SUDI SUDI(![/||A||/]) B ALIGN_ACQUIRED_4 AUDI deskew_error * SUDI !deskew_error * SUDI(![/||A||/]) SUDI(![/||A||/]) C Note to Figure 1-128: (1) This figure is from IEEE P802.3ae. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-163 Rate Match FIFO In XAUI mode, the rate match FIFO is capable of compensating for up to 100 PPM (200 PPM total) difference between the upstream transmitter and the local receiver reference clock. The XAUI protocol requires the transmitter to send /R/ (/K28.0/) code groups simultaneously on all four lanes (denoted as ||R|| column) during inter-packet gaps, adhering to rules listed in the IEEE P802.3ae specification. The rate match FIFO operation in XAUI mode is compliant to the IEEE P802.3ae specification. The rate match operation begins after: The synchronization state machine in the word aligner of all four channels indicates synchronization has been acquired by driving the rx_syncstatus signal high The deskew FIFO indicates alignment has been acquired by driving the rx_channelaligned signal high The rate match FIFO looks for the ||R|| column (simultaneous /R/ code group on all four channels) and deletes or inserts ||R|| column to prevent the rate match FIFO from overflowing or under-running. The rate match FIFO can insert or delete as many ||R|| columns as necessary to perform the rate match operation. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, indicating rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. If an ||R|| column is deleted, the rx_rmfifodeleted flag from each of the four channels goes high for one clock cycle per deleted ||R|| column. If an ||R|| column is inserted, the rx_rmfifoinserted flag from each of the four channels goes high for one clock cycle per inserted ||R|| column. Figure 1-129 shows an example of rate match deletion in the case where three ||R|| columns are required to be deleted. For more information, refer to "Rate Match FIFO in XAUI Mode" on page 1-80. Figure 1-129. Rate Match Deletion in XAUI Mode First ||R|| Column Second ||R|| Column Third ||R|| Column Fourth ||R|| Column datain[3] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[2] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[1] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 datain[0] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 K28.0 K28.0 K28.5 dataout[3] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[2] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[1] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 dataout[0] K28.5 K28.3 K28.5 K28.5 K28.0 K28.5 rx_rmfifodatadeleted September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-164 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-130 shows an example of rate match insertion in the case where two ||R|| columns are required to be inserted. Figure 1-130. Rate Match Insertion in XAUI Mode First ||R|| Column Second ||R|| Column dataout[3] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 dataout[2] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 dataout[1] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 dataout[0] K28.5 K28.3 K28.5 K28.0 K28.5 K28.0 datain[3] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 datain[2] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 datain[1] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 datain[0] K28.5 K28.3 K28.5 K28.0 K28.0 K28.0 K28.5 K28.0 rx_rmfifodatainserted For more information, refer to "Rate Match (Clock Rate Compensation) FIFO" on page 1-77. GIGE Mode IEEE 802.3 defines the 1000 Base-X PHY as an intermediate, or transition, layer that interfaces various physical media with the media access control (MAC) in a gigabit ethernet system. It shields the MAC layer from the specific nature of the underlying medium. The 1000 Base-X PHY is divided into three sub-layers: Physical coding sublayer Physical media attachment Physical medium dependent (PMD) The PCS sublayer interfaces with the MAC through the gigabit medium independent interface (GMII). The 1000 Base-X PHY defines a physical interface data rate of 1 Gbps. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-165 Figure 1-131 shows the 1000 Base-X PHY position in a Gigabit Ethernet OSI reference model. Figure 1-131. 1000 Base-X PHY in a Gigabit Ethernet OSI Reference Model LAN CSMA/CD Layers OSI Reference Model Layers Higher Layers LLC Application MAC (Optional) Presentation MAC Session Transport Reconciliation GMII Network Data Link PCS PMA PMD 1000 Base-X PHY Physical Medium Stratix IV GX and GT transceivers, when configured in GIGE functional mode, have built-in circuitry to support the following PCS and PMA functions defined in the IEEE 802.3 specification: 1 September 2012 8B/10B encoding and decoding Synchronization Upstream transmitter and local receiver clock frequency compensation (rate matching) Clock recovery from the encoded data forwarded by the receiver PMD Optional rx_recovclkout port enables recovered clock at the pin level (use with VCXO) Serialization and deserialization Stratix IV GX and GT transceivers do not have built-in support for other PCS functions; for example, auto-negotiation state machine, collision-detect, and carrier-sense. If required, you must implement these functions in a PLD logic array or external circuits. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-166 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-132 shows the GIGE mode configuration supported in Stratix IV GX devices. Figure 1-132. GIGE Mode for Stratix IV GX Devices Stratix IV GX Configurations Protocol Basic Single Width Functional Modes PMA-PCS Interface Width 8-bit 10-bit Double Width 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Functional Mode GIGE Data Rate (Gbps) 1.25 Channel Bonding x1 Low-Latency PCS Disabled Word Aligner (Pattern Length) Automatic Synchronization State Machine (7-Bit Comma, 10-Bit /K28.5/) 8B/10B Encoder/Decoder Enabled Rate Match FIFO Enabled Byte SerDes Disabled Byte Ordering Disabled FPGA Fabric-Transceiver Interface Width 8-Bit FPGA Fabric-Transceiver Interface Frequency (MHz) 125 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 5-6 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 20 - 24 Stratix IV Device Handbook Volume 2: Transceivers Deterministic Latency 10-Bit 20-Bit September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-167 GIGE Mode Datapath Figure 1-133 shows the transceiver datapath when configured in GIGE functional mode. Figure 1-133. GIGE Mode Datapath Transmitter Channel PCS FPGA Fabric TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA 8B/10B Encoder Serializer tx_coreclk[0] High-Speed Serial Clock Local Clock Divider Low-Speed Parallel Clock tx_clkout[0] FPGA Fabric-Transceiver Interface Clock Receiver Channel PCS RX Phase Compensation FIFO Rate Match FIFO 8B/10B 8B/10B Decoder Decoder Word Aligner Receiver Channel PMA DeDeSerializer Serializer CDR Parallel Recovered Clock rx_coreclk[0] Low-Speed Parallel Clock Table 1-62 lists the transceiver datapath clock frequencies in GIGE functional mode. Table 1-62. Transceiver Datapath Clock Frequencies in GIGE Mode Functional Mode Data Rate High-Speed Serial Clock Frequency GIGE 1.25 Gbps 625 MHz Parallel Recovered Clock and Low-Speed Parallel Clock Frequency FPGA Fabric-Transceiver Interface Clock Frequency 125 MHz 125 MHz 8B/10B Encoder In GIGE mode, the 8B/10B encoder clocks in 8-bit data and 1-bit control identifiers from the transmitter phase compensation FIFO and generates 10-bit encoded data. The 10-bit encoded data is fed to the serializer. For more information about 8B/10B encoder functionality, refer to "8B/10B Encoder" on page 1-23. GIGE Protocol--Ordered Sets and Special Code Groups Table 1-63 lists ordered sets and special code groups specified in the IEEE 802.3 specification. Table 1-63. GIGE Ordered Sets (Part 1 of 2) Code September 2012 Ordered Set Number of Code Groups Encoding /C/ Configuration -- Alternating /C1/ and /C2/ /C1/ Configuration 1 4 /K28.5/D21.5/Config_Reg /C2/ Configuration 2 4 /K28.5/D2.2/Config_Reg Altera Corporation (1) (1) Stratix IV Device Handbook Volume 2: Transceivers 1-168 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-63. GIGE Ordered Sets (Part 2 of 2) Code Ordered Set Number of Code Groups Encoding /I/ IDLE -- Correcting /I1/, Preserving /I2/ /I1/ IDLE 1 2 /K28.5/D5.6 /I2/ IDLE 2 2 /K28.5/D16.2 Encapsulation -- -- /R/ Carrier_Extend 1 /K23.7/ /S/ Start_of_Packet 1 /K27.7/ /T/ End_of_Packet 1 /K29.7/ /V/ Error_Propagation 1 /K30.7/ Note to Table 1-63: (1) Two data code groups representing the Config_Reg value. Idle Ordered-Set Generation The IEEE 802.3 specification requires the GIGE PHY to transmit idle ordered sets (/I/) continuously and repetitively whenever the GMII is idle. This ensures that the receiver maintains bit and word synchronization whenever there is no active data to be transmitted. In GIGE functional mode, any /Dx.y/ following a /K28.5/ comma is replaced by the transmitter with either a /D5.6/ (/I1/ ordered set) or a /D16.2/ (/I2/ ordered set), depending on the current running disparity. The exception is when the data following the /K28.5/ is /D21.5/ (/C1/ ordered set) or /D2.2/ (/C2/) ordered set. If the running disparity before the /K28.5/ is positive, an /I1/ ordered set is generated. If the running disparity is negative, a /I2/ ordered set is generated. The disparity at the end of a /I1/ is the opposite of that at the beginning of the /I1/. The disparity at the end of a /I2/ is the same as the beginning running disparity (right before the idle code). This ensures a negative running disparity at the end of an idle ordered set. A /Kx.y/ following a /K28.5/ is not replaced. 1 Note that /D14.3/, /D24.0/, and /D15.8/ are replaced by /D5.6/ or /D16.2/ (for /I1/, /I2/ ordered sets). /D21.5/ (part of the /C1/ order set) is not replaced. Figure 1-134 shows the automatic idle ordered set generation. Figure 1-134. Automatic Ordered Set Generation clock tx_datain [ ] K28.5 D14.3 K28.5 D24.0 K28.5 D15.8 K28.5 D21.5 Dx.y tx_dataout Dx.y K28.5 D5.6 K28.5 D16.2 K28.5 D16.2 K28.5 D21.5 Ordered Set Stratix IV Device Handbook Volume 2: Transceivers /I1/ /I2/ /I2/ /C2/ September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-169 Reset Condition After de-assertion of tx_digitalreset, the GIGE transmitter automatically transmits three /K28.5/ comma code groups before transmitting user data on the tx_datain port. This could affect the synchronization state machine behavior at the receiver. Depending on when you start transmitting the synchronization sequence, there could be an even or odd number of /Dx.y/ code groups transmitted between the last of the three automatically sent /K28.5/ code groups and the first /K28.5/ code group of the synchronization sequence. If there is an even number of /Dx.y/ code groups received between these two /K28.5/ code groups, the first /K28.5/ code group of the synchronization sequence begins at an odd code group boundary (rx_even = FALSE). An IEEE802.3-compliant GIGE synchronization state machine treats this as an error condition and goes into the loss of sync state. Figure 1-135 shows an example of even numbers of /Dx.y/ between the last automatically sent /K28.5/ and the first user-sent /K28.5/. The first user-sent /K28.5/ code group received at an odd code group boundary in cycle n + 3 takes the receiver synchronization state machine in the loss of sync state. The first synchronization ordered set /K28.5/Dx.y/ in cycles n + 3 and n + 4 is discounted and three additional ordered sets are required for successful synchronization. Figure 1-135. Reset Condition in GIGE Mode n n+1 n+2 n+3 n+4 K28.5 Dx.y Dx.y K28.5 Dx.y clock tx_digitalreset tx_dataout K28.5 xxx K28.5 K28.5 K28.5 Dx.y K28.5 Dx.y Word Aligner The word aligner in GIGE functional mode is configured in automatic synchronization state machine mode. The Quartus II software automatically configures the synchronization state machine to indicate synchronization when the receiver receives three consecutive synchronization ordered sets. A synchronization ordered set is a /K28.5/ code group followed by an odd number of valid /Dx.y/ code groups. The fastest way for the receiver to achieve synchronization is to receive three continuous {/K28.5/, /Dx.y/} ordered sets. Receiver synchronization is indicated on the rx_syncstatus port of each channel. A high on the rx_syncstatus port indicates that the lane is synchronized; a low on the rx_syncstatus port indicates that the lane has fallen out of synchronization. The receiver loses synchronization when it detects four invalid code groups separated by less than three valid code groups or when it is reset. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-170 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Table 1-64 lists the synchronization state machine parameters when configured in GIGE mode. Table 1-64. Synchronization State Machine Parameters in GIGE Functional Mode Synchronization State Machine Parameters Setting Number of valid {/K28.5/, /Dx,y/} ordered sets received to achieve synchronization 3 Number of errors received to lose synchronization 4 Number of continuous good code groups received to reduce the error count by 1 4 Figure 1-136 shows the synchronization state machine implemented in GIGE mode. Figure 1-136. Synchronization State Machine in GIGE Mode (1) power_on=TRUE+mr_main_rest=TRUE + (signal_detectCHANGE=TRUE + mr_loopback=FALSE +PUDI) LOSS_OF_SYNC [PUDI * signal_detect=FAIL + mr_loopback=FALSE] + PUDI(![/COMMA/]) sync_status FAIL rx_even ! rx_even SUDI (signal_detect=OK+mr_loopback=TRUE)* * PUDI([/COMMA/] COMMA_DETECT_1 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] ACQUIRE_SYNC_1 PUDI(![/COMMA/] *[/INVALID/] rx_even ! rx_even SUDI cgbad rx_even=FALSE+PUDI([/COMMA/] COMMA_DETECT_2 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] ACQUIRE_SYNC_2 PUDI(![/COMMA/] *[/INVALID/] rx_even ! rx_even SUDI cgbad rx_even=FALSE+PUDI([/COMMA/] COMMA_DETECT_3 SYNC_ACQUIRED_1 rx_even TRUE SUDI PUDI(![/|DV|/] PUDI([/|DV|/] sync_status OK rx_even ! rx_even SUDI cgbad cggood 2 cggood SYNC_ACQUIRED_2 SYNC_ACQUIRED_2A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cgbad cggood *good_cgs = 3 cgbad cggood *good_cgs = 3 3 cggood SYNC_ACQUIRED_3 SYNC_ACQUIRED_3A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cggood *good_cgs = 3 cgbad cgbad 2 cggood *good_cgs = 3 cggood SYNC_ACQUIRED_4 SYNC_ACQUIRED_4A rx_even ! rx_even SUDI good_cgs 0 rx_even ! rx_even SUDI good_cgs good_cgs + 1 cggood *good_cgs = 3 cgbad cgbad 3 cggood *good_cgs = 3 Note to Figure 1-136: (1) This figure is from IEEE P802.3ae. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-171 Rate Match FIFO In GIGE mode, the rate match FIFO is capable of compensating for up to 100 PPM (200 PPM total) difference between the upstream transmitter and the local receiver reference clock. The GIGE protocol requires the transmitter to send idle ordered sets /I1/ (/K28.5/D5.6/) and /I2/ (/K28.5/D16.2/) during inter-packet gaps adhering to the rules listed in the IEEE 802.3 specification. The rate match operation begins after the synchronization state machine in the word aligner indicates synchronization is acquired by driving the rx_syncstatus signal high. The rate matcher deletes or inserts both symbols (/K28.5/ and /D16.2/) of the /I2/ ordered sets even if it requires deleting only one symbol to prevent the rate match FIFO from overflowing or under-running. It can insert or delete as many /I2/ ordered sets as necessary to perform the rate match operation. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, indicating rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. Both the rx_rmfifodatadeleted and rx_rmfifodatainserted flags are asserted for two clock cycles for each deleted and inserted /I2/ ordered set, respectively. Figure 1-137 shows an example of rate match FIFO deletion where three symbols are required to be deleted. Because the rate match FIFO can only delete /I2/ ordered set, it deletes two /I2/ ordered sets (four symbols deleted). Figure 1-137. Rate Match Deletion in GIGE Mode /I2/ SKIP Symbol Deleted First /I2/ Skip Ordered Set Second /I2/ Skip Ordered Set Third /I2/ Skip Ordered Set datain Dx.y K28.5 D16.2 K28.5 dataout Dx.y K28.5 D16.2 Dx.y D16.2 K28.5 D16.2 Dx.y rx_rmfifodatadeleted Figure 1-138 shows an example of rate match FIFO insertion in the case where one symbol is required to be inserted. Because the rate match FIFO can only delete /I2/ ordered set, it inserts one /I2/ ordered set (two symbols inserted). Figure 1-138. Rate Match Insertion in GIGE Mode First /I2/ Ordered Set Second /I2/ Ordered Set datain Dx.y K28.5 D16.2 K28.5 D16.2 dataout Dx.y K28.5 D16.2 K28.5 D16.2 K28.5 D16.2 Dx.y rx_rmfifodatainserted For more information, refer to "Rate Match (Clock Rate Compensation) FIFO" on page 1-77. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-172 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture SONET/SDH Mode SONET/SDH is one of the most common serial-interconnect protocols used in backplanes deployed in communications and telecom applications. SONET/SDH defines various optical carrier (OC) sub-protocols for carrying signals of different capacities through a synchronous optical hierarchy. SONET/SDH Frame Structure Base OC-1 frames are byte-interleaved to form SONET/SDH frames. For example, 12 OC-1 frames are byte-interleaved to form one OC-12 frame; 48 OC-1 frames are byte-interleaved to form one OC-48 frame, and so on. SONET/SDH frame sizes are constant, with a frame transfer rate of 125 s. Figure 1-139 shows the SONET/SDH frame structure. Figure 1-139. SONET/SDH Mode Nx3 Bytes Transport Overhead NxA1 Nx3 Bytes Payload NxA2 NxJ0/Z0 9 Rows Transport overhead bytes A1 and A2 are used for restoring frame boundary from the serial data stream. Frame sizes are fixed, so the A1 and A2 bytes appear within the serial data stream every 125 s. In an OC-12 system, 12 A1 bytes are followed by 12 A2 bytes. Similarly, in an OC-48 system, 48 A1 bytes are followed by 48 A2 bytes. In SONET/SDH systems, byte values of A1 and A2 are fixed as follows: A1 = 11110110 or 8'hF6 A2 = 00101000 or 8'h28 You can employ Stratix IV GX and GT transceivers as physical layer devices in a SONET/SDH system. These transceivers provide support for SONET/SDH protocol-specific functions and electrical features; for example, alignment to A1A2 or A1A1A2A2 pattern. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-173 Stratix IV transceivers are designed to support the following three SONET/SDH sub-protocols: September 2012 OC-12 at 622 Mbps with 8-bit channel width (not supported in Stratix IV GT devices) OC-48 at 2488.32 Mbps with 16-bit channel width OC-96 at 4976 Mbps with 32-bit channel width Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-174 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-140 shows SONET/SDH mode configurations supported in Stratix IV GX and GT devices. Figure 1-140. SONET/SDH Mode Configurations in Stratix IV GX and GT Devices Stratix IV GX and GT Configurations Protocol Basic Single Width Functional Modes PMA-PCS Interface Width 8-bit 10-bit Double Width 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Deterministic Latency 10-Bit 20-Bit SONET/ SDH Functional Mode Data Rate (Gbps) 0.622 (OC-12) (1) 2.488 (OC-48) 4.976 (OC-96) Channel Bonding x1 x1 x1 Low-Latency PCS Disabled Disabled Disabled Word Aligner (Pattern Length) Manual Alignment (16-Bit A1A2, 32-Bit A1A1A2A2) Manual Alignment (16-Bit A1A2, 32-Bit A1A1A2A2) Manual Alignment (32-Bit A1A1A2A2) 8B/10B Encoder/Decoder Disabled Disabled Disabled Rate Match FIFO Disabled Disabled Disabled Byte SerDes Disabled Enabled Enabled Byte Ordering Disabled Enabled Disabled FPGA Fabric-Transceiver Interface Width 8-Bit 16-Bit 32-Bit FPGA Fabric-Transceiver Interface Frequency (MHz) 77.75 155.5 155.5 5-6 4 - 5.5 4 - 5.5 11 - 13 7-9 6.5 8.5 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Note to Figure 1-41: (1) This is not supported in Stratix IV GT devices. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-175 SONET/SDH OC-12 Datapath Figure 1-141 shows the transceiver datapath when configured in SONET/SDH OC-12 mode. Figure 1-141. SONET/SDH OC-12 Datapath Transmitter Channel PCS FPGA Fabric Transmitter Channel PMA TX Phase Compensation FIFO wrclk Serializer rdclk tx_coreclk High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider tx_clkout FPGA Fabric-Transmitter Interface Clock Receiver Channel PCS RX Phase Compensation FIFO Word Aligner Receiver Channel PMA DeSerializer CDR rx_coreclk Parallel Recovered Clock rx_clkout FPGA Fabric-Receiver Interface Clock SONET/SDH OC-48 Datapath Figure 1-142 shows the transceiver datapath when configured in SONET/SDH OC-48 mode. Figure 1-142. SONET/SDH OC-48 Datapath FPGA Fabric Transmitter Channel PCS TX Phase Compensation FIFO wrclk Transmitter Channel PMA Byte Serializer Serializer rdclk rdclk wrclk tx_coreclk High-Speed Serial Clock /2 Low-Speed Parallel Clock tx_clkout FPGA Fabric-Transmitter Interface Clock Receiver Channel PCS RX Phase Compensation FIFO rx_coreclk Byte Ordering Byte Deserializer Word Aligner Local Clock Divider Receiver Channel PMA DeSerializer CDR /2 Parallel Recovered Clock rx_clkout FPGA Fabric-Receiver Interface Clock September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-176 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture SONET/SDH OC-96 Datapath Figure 1-143 shows the transceiver datapath when configured in SONET/SDH OC-96 mode. Figure 1-143. SONET/SDH OC-96 Datapath Transmitter Channel PCS FPGA Fabric TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA Byte Serializer Serializer rdclk wrclk tx_coreclk High-Speed Serial Clock /2 Low-Speed Parallel Clock tx_clkout FPGA Fabric-Transmitter Interface Clock Receiver Channel PCS RX Phase Compensation FIFO Byte Deserializer rx_coreclk Word Aligner Local Clock Divider Receiver Channel PMA DeSerializer CDR /2 Parallel Recovered Clock rx_clkout FPGA Fabric-Receiver Interface Clock SONET/SDH Transmission Bit Order Unlike Ethernet, where the LSB of the parallel data byte is transferred first, SONET/SDH requires the MSB to be transferred first and the LSB to be transferred last. To facilitate the MSB-to-LSB transfer, you must enable the following options in the ALTGX MegaWizard Plug-In Manager: Flip transmitter input data bits Flip receiver output data bits Depending on whether data bytes are transferred MSB-to-LSB or LSB-to-MSB, you must select the appropriate word aligner settings in the ALTGX MegaWizard Plug-In Manager. Table 1-65 on page 1-177 lists the correct word aligner settings for each bit transmission order. Word Alignment The word aligner in SONET/SDH OC-12, OC-48, and OC-96 modes is configured in manual alignment mode, as described in "Word Aligner in Single-Width Mode with 8-Bit PMA-PCS Interface Modes" on page 1-60. In OC-12 and OC-48 configurations, you can configure the word aligner to either align to a 16-bit A1A2 pattern or a 32-bit A1A1A2A2 pattern. This is controlled by the rx_a1a2size input port to the transceiver. A low level on the rx_a1a2size port configures the word aligner to align to a 16-bit A1A2 pattern; a high level on the rx_a1a2size port configures the word aligner to align to a 32-bit A1A1A2A2 pattern. In OC-96 configuration, the word aligner is only allowed to align to a A1A1A2A2 pattern, so the input port rx_ala2size is unavailable. Barring this difference, the OC-96 word alignment operation is similar to that of the OC-12 and OC-48 configurations. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-177 You can configure the word aligner to flip the alignment pattern bits programmed in the MegaWizard Plug-In Manager and compare them with the incoming data for alignment. This feature offers flexibility to the SONET backplane system for either a MSB-to-LSB or LSB-to-MSB data transfer. Table 1-65 lists word alignment patterns that you must program in the ALTGX MegaWizard Plug-In Manager based on the bit-transmission order and the word aligner bit-flip option. Table 1-65. Word Aligner Settings Serial Bit Transmission Order Word Alignment Bit Flip Word Alignment Pattern MSB-to-LSB On 1111011000101000 (16'hF628) MSB-to-LSB Off 0001010001101111 (16'h146F) LSB-to-MSB Off 0010100011110110 (16'h28F6) The behavior of the SONET/SDH word aligner control and status signals, along with an operational timing diagram, are explained in "Word Aligner in Single-Width Mode with 8-Bit PMA-PCS Interface Modes" on page 1-60. OC-48 and OC-96 Byte Serializer and Deserializer The OC-48 and OC-96 transceiver datapath includes the byte serializer and deserializer to allow the PLD interface to run at a lower speed. The OC-12 configuration does not use the byte serializer and deserializer blocks. The byte serializer and deserializer blocks are explained in "Byte Serializer" on page 1-21 and "Byte Deserializer" on page 1-92, respectively. The OC-48 byte serializer converts 16-bit data words from the FPGA fabric and translates the 16-bit data words into two 8-bit data bytes at twice the rate. The OC-48 byte deserializer takes in two consecutive 8-bit data bytes and translates them into a 16-bit data word to the FPGA fabric at half the rate. The OC-96 byte serializer converts 32-bit data words from the FPGA fabric and translates them into two 16-bit data words at twice the rate. The OC-96 byte deserializer takes in two consecutive 16-bit data words and translates them into a 32-bit data word to the FPGA fabric at half the rate. OC-48 Byte Ordering Because of byte deserialization, the MSByte of a word might appear at the rx_dataout port along with the LSByte of the next word. In an OC-48 configuration, the byte ordering block is built into the datapath and can be leveraged to perform byte ordering. Byte ordering in an OC-48 configuration is automatic, as explained in "Word-Alignment-Based Byte Ordering" on page 1-97. In automatic mode, the byte ordering block is triggered by the rising edge of the rx_syncstatus signal. As soon as the byte ordering block sees the rising edge of the rx_syncstatus signal, it compares the LSByte coming out of the byte deserializer with the A2 byte of the A1A2 alignment pattern. If the LSByte coming out of the byte deserializer does not match the A2 byte set in the ALTGX MegaWizard Plug-In Manager, the byte ordering block inserts a PAD character, as seen in Figure 1-144. Insertion of this PAD character enables the byte ordering block to restore the correct byte order. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-178 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1 The PAD character is defaulted to the A1 byte of the A1A2 alignment pattern. Figure 1-144. OC-48 Byte Ordering in Automatic Mode From Byte Deserializer To PLD Core rx_dataout (MSB) A1 A1 A2 A2 D0 D2 rx_dataout (LSB) X A1 A1 A2 A2 D1 Byte Ordering Block A1 A1 Pad A2 D1 D3 X A1 A1 A2 D0 D2 rx_clkout rx_syncstatus rx_syncstatus rx_byteorderalignstatus SDI Mode The Society of Motion Picture and Television Engineers (SMPTE) defines various SDI standards for transmission of uncompressed video. The following three SMPTE standards are popular in video broadcasting applications: SMPTE 259M standard--more popularly known as the standard-definition (SD) SDI, is defined to carry video data at 270 Mbps SMPTE 292M standard--more popularly known as the high-definition (HD) SDI, is defined to carry video data at either 1485 Mbps or 1483.5 Mbps SMPTE 424M standard--more popularly known as the third-generation (3G) SDI, is defined to carry video data at either 2970 Mbps or 2967 Mbps You can configure Stratix IV GX and GT transceivers in HD-SDI or 3G-SDI configuration using the ALTGX MegaWizard Plug-In Manager. Table 1-66 lists the ALTGX configurations supported by Stratix IV transceivers in SDI mode. Table 1-66. ALTGX Configurations in SDI Mode Configuration HD (1) 3G (2) Data Rate (Mbps) REFCLK Frequencies (MHz) FPGA Fabric-Transceiver Interface Width 1485 74.25, 148.5 10 bit and 20 bit 1483.5 74.175, 148.35 10 bit and 20 bit 2970 148.5, 297 Only 20-bit interface allowed in 3G 2967 148.35, 296.7 Only 20-bit interface allowed in 3G Notes to Table 1-66: (1) Not supported by Stratix IV GT devices. (2) Stratix IV GT devices only support the 3G configuration. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-179 Figure 1-145 shows SDI mode configurations supported in Stratix IV GX and GT devices. Figure 1-145. SDI Mode Stratix IV GX and GT Configurations Protocol Basic Single Width Functional Modes PMA-PCS Interface Width 8-bit 10-bit Double Width 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Functional Mode HD-SDI (1.485/1.4835) (1) 3G-SDI (2.97/2.967) Channel Bonding x1 x1 Low-Latency PCS Disabled Disabled Word Aligner (Pattern Length) Bit-Slip Bit-Slip 8B/10B Encoder/Decoder Disabled Disabled Rate Match FIFO Disabled Disabled Byte SerDes Disabled Enabled Enabled Byte Ordering Disabled Disabled Disabled FPGA Fabric-Transceiver Interface Width 10-Bit 20-Bit 20-bit 148.5/ 148.35 74.25/ 74.175 148.5/ 148.35 5-6 4 - 5.5 4 - 5.5 9 - 11 6-8 6-8 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 10-Bit 20-Bit SDI Data Rate (Gbps) FPGA Fabric-Transceiver Interface Frequency (MHz) Deterministic Latency Note to Figure 1-45: (1) Not supported in Stratix IV GT devices. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-180 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture SDI Mode Datapath Figure 1-146 shows the transceiver datapath when configured in SDI mode. Figure 1-146. SDI Mode Datapath Transmitter Channel PCS FPGA Fabric TX Phase Compensation FIFO wrclk rdclk Byte Serializer wrclk tx_coreclk Transmitter Channel PMA Serializer rdclk /2 High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider tx_clkout FPGA Fabric-Transmitter Interface Clock Receiver Channel PCS RX Phase Compensation FIFO rx_coreclk Byte Deserializer /2 Word Aligner Receiver Channel PMA DeSerializer CDR Parallel Recovered Clock rx_clkout FPGA Fabric-Receiver Interface Clock Transmitter Datapath The transmitter datapath, in HD-SDI configuration with 10-bit wide FPGA fabric-transceiver interface, consists of the transmitter phase compensation FIFO and the 10:1 serializer. The transmitter datapath, in HD-SDI and 3G-SDI configurations with 20-bit wide FPGA fabric-transceiver interface, also includes the byte serializer. 1 In SDI mode, the transmitter is purely a parallel-to-serial converter. SDI transmitter functions, such as scrambling and cyclic redundancy check (CRC) code generation, must be implemented in the FPGA logic array. Receiver Datapath In the 10-bit channel width SDI configuration, the receiver datapath is comprised of the clock recovery unit (CRU), 1:10 deserializer, word aligner in bit-slip mode, and receiver phase compensation FIFO. In the 20-bit channel width SDI configuration, the receiver datapath also includes the byte deserializer. 1 SDI receiver functions, such as de-scrambling, framing, and CRC checker, must be implemented in the FPGA logic array. Receiver Word Alignment and Framing In SDI systems, the word aligner in the receiver datapath is not useful because word alignment and framing happens after de-scrambling. Altera recommends driving the ALTGX megafunction rx_bitslip signal low to avoid having the word aligner insert bits in the received data stream. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-181 (OIF) CEI PHY Interface Mode Stratix IV GX and GT transceivers support a data rate between 4.976 Gbps and 6.375 Gbps in (OIF) CEI PHY interface mode. Figure 1-147 shows (OIF) CEI PHY interface mode configurations supported in Stratix IV GX and GT devices. Figure 1-147. (OIF) CEI PHY Interface Mode for Stratix IV GX and GT Devices Stratix IV GX and GT Configurations Protocol Basic Single Width Functional Modes PMA-PCS Interface Width September 2012 8-bit 10-bit Double Width 16-bit 20-bit PIPE XAUI GIGE SRIO SONET /SDH (OIF) CEI SDI 10-bit 10-bit 10-bit 10-bit 8-bit 16-bit 10-bit Functional Mode (OIF) CEI PHY Interface Mode Data Rate (Gbps) 3.125-6.375 Channel Bonding x1, x4 (Transmitter PMA-Only Bonding) Low-Latency PCS Disabled Word Aligner (Pattern Length) Disabled 8B/10B Encoder/Decoder Disabled Rate Match FIFO Disabled Byte SerDes Enabled Byte Ordering Disabled FPGA Fabric-Transceiver Interface Width 32-Bit FPGA Fabric-Transceiver Interface Frequency (MHz) 97.65625199.21875 TX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) Interface Frequency 4 -5.5 RX PCS Latency (FPGA Fabric-Transceiver Interface Clock Cycles) 6.5 8.5 Altera Corporation Deterministic Latency 10-Bit 20-Bit Stratix IV Device Handbook Volume 2: Transceivers 1-182 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture (OIF) CEI PHY Interface Mode Datapath Figure 1-148 shows the ALTGX megafunction transceiver datapath when configured in (OIF) CEI PHY interface mode. Figure 1-148. (OIF) CEI PHY Interface Mode Datapath Transmitter Channel PCS FPGA Fabric TX Phase Compensation FIFO wrclk Byte Serializer rdclk wrclk tx_coreclk /2 Serializer rdclk High-Speed Serial Clock Low-Speed Parallel Clock tx_clkout FPGA Fabric-Transmitter Interface Clock Receiver Channel PCS Byte DeSerializer RX Phase Compensation FIFO rx_coreclk Transmitter Channel PMA Local Clock Divider Receiver Channel PMA DeSerializer CDR /2 Parallel Recovered Clock rx_clkout FPGA Fabric-Receiver Interface Clock Figure 1-149 shows transceiver clocking in (OIF) CEI PHY interface mode. Figure 1-149. Transceiver Clocking in (OIF) CEI PHY Interface Mode Transceiver Block Clocking with the Use central clock divider to improve transmitter jitter option disabled CMU PLL Ch 3 Local Clock Divider Block Channel 3 Ch 2 Local Clock Divider Block Channel 2 Ch 1 Local Clock Divider Block Channel 1 Ch 0 Local Clock Divider Block Channel 0 Serial RapidIO Mode The RapidIO Trade Association defines a high-performance, packet-switched interconnect standard to pass data and control information between microprocessors, digital signal, communications, and network processors, system memories, and peripheral devices. Serial RapidIO physical layer specification defines three line rates: Stratix IV Device Handbook Volume 2: Transceivers 1.25 Gbps 2.5 Gbps 3.125 Gbps September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-183 It also defines two link widths--single-lane (1x) and bonded four-lane (4x) at each line rate. Stratix IV GX and GT transceivers support only single-lane (1x) configuration at all three line rates. Four 1x channels configured in Serial RapidIO mode can be instantiated to achieve a 4x Serial RapidIO link. The four transmitter channels in this 4x Serial RapidIO link are not bonded. The four receiver channels in this 4x Serial RapidIO link do not have lane alignment or deskew capability. Figure 1-150 shows the ALTGX transceiver datapath when configured in Serial RapidIO mode. Figure 1-150. Serial RapidIO Mode Datapath Transmitter Channel PCS FPGA Fabric TX Phase Compensation FIFO wrclk rdclk 8B/10B Encoder Byte Serializer tx_coreclk[0] Transmitter Channel PMA Serializer High-Speed Serial Clock Local Clock Divider /2 Low-Speed Parallel Clock tx_clkout[0] FPGA Fabric-Transceiver Interface Clock Receiver Channel PCS RX Phase Compensation FIFO Byte DeSerializer Rate Match FIFO 8B/10B Decoder Word Aligner Receiver Channel PMA DeSerializer CDR rx_coreclk[0] Parallel Recovered Clock /2 Low-Speed Parallel Clock Stratix IV GX and GT transceivers, when configured in Serial RapidIO functional mode, provide the following PCS and PMA functions: 1 September 2012 8B/10B encoding/decoding Word alignment Lane synchronization state machine Clock recovery from the encoded data Serialization/deserialization Stratix IV GX and GT transceivers do not have built-in support for other PCS functions; for example, pseudo-random idle sequence generation and lane alignment in 4x mode. Depending on your system requirements, you must implement these functions in the logic array or external circuits. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-184 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Synchronization State Machine In Serial RapidIO mode, the ALTGX MegaWizard Plug-In Manager defaults the word alignment pattern to K28.5. The word aligner has a synchronization state machine that handles the receiver lane synchronization. The ALTGX MegaWizard Plug-In Manager automatically defaults the synchronization state machine to indicate synchronization when the receiver receives 127 K28.5 (10'b0101111100 or 10'b1010000011) synchronization code groups without receiving an intermediate invalid code group. After synchronization, the state machine indicates loss of synchronization when it detects three invalid code groups separated by less than 255 valid code groups or when it is reset. Receiver synchronization is indicated on the rx_syncstatus port of each channel. A high on the rx_syncstatus port indicates that the lane is synchronized and a low indicates that it has fallen out of synchronization. Table 1-67 lists the ALTGX megafunction synchronization state machine parameters when configured in Serial RapidIO mode. Table 1-67. Synchronization State Machine Parameters in Serial RapidIO Mode Parameters Number of valid K28.5 code groups received to achieve synchronization. Number of errors received to lose synchronization. Number of continuous good code groups received to reduce the error count by one. Stratix IV Device Handbook Volume 2: Transceivers Number 127 3 255 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-185 Figure 1-151 shows a conceptual view of the synchronization state machine implemented in Serial RapidIO functional mode. Figure 1-151. Synchronization State Machine in Serial RapidIO Mode Loss of Sync Data = Comma Data = !Valid Comma Detect if Data == Comma kcntr++ else kcntr=kcntr Data = valid; kcntr < 127 kcntr = 127 Synchronized Data=Valid Data = !Valid ecntr = 3 Synchronized Error Detect if Data == !Valid ecntr++ gcntr=0 else if gcntr==255 ecntr-gcntr=0 else gcntr++ ecntr = 0 Rate Match FIFO in Serial RapidIO Mode In Serial RapidIO mode, the rate match FIFO is capable of compensating for up to 100 PPM (200 PPM total) difference between the upstream transmitter and the local receiver reference clock. September 2012 1 To enable the rate match FIFO in Serial RapidIO mode, the transceiver channel must have both the transmitter and receiver channel instantiated. You must select the Receiver and Transmitter option in the What is the operation mode? field in the ALTGX MegaWizard Plug-In Manager. The 8B/10B encoder/decoder is always enabled in Serial RapidIO mode. 1 Rate matcher is an optional block available for selection in SRIO functional mode. However, this block is not fully compliant to the SRIO specification. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-186 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Depending on your implementation, you can select two 20-bit rate match patterns in the ALTGX MegaWizard Plug-In Manager under the What is the rate match pattern1 and What is the rate match pattern2 fields. Each of the two programmed 20-bit rate match patterns consists of a 10-bit skip pattern and a 10-bit control pattern. For Serial RapidIO mode in the ALTGX MegaWizard Plug-In Manager, the control pattern1 defaults to K28.5 with positive disparity and the skip pattern1 defaults to K29.7 with positive disparity. The control pattern2 defaults to K28.5 with negative disparity and the skip pattern2 defaults to K29.7 with negative disparity. The rate match FIFO operation begins after the word aligner synchronization status rx_syncstatus goes high. When the rate matcher receives either of the two 10-bit control patterns followed by the respective 10-bit skip pattern, it inserts or deletes the 10-bit skip pattern as necessary to avoid the rate match FIFO from overflowing or under-running. In Serial RapidIO mode, the rate match FIFO can delete/insert a maximum of one skip pattern from a cluster. Two flags, rx_rmfifodatadeleted and rx_rmfifodatainserted, indicate that rate match FIFO deletion and insertion events, respectively, are forwarded to the FPGA fabric. Figure 1-152 shows an example of rate match FIFO deletion in the case where one skip pattern is required to be deleted. In this example, the first skip cluster has a /K28.5/ control pattern followed by two /K29.7/ skip patterns. The second skip cluster has a /K28.5/ control pattern followed by four /K29.7/ skip patterns. The rate match FIFO deletes only one /K29.7/ skip pattern from the first skip cluster. One /K29.7/ skip pattern is deleted from the second cluster. Figure 1-152. Rate Match FIFO Deletion with One Skip Pattern Deleted One Skip Pattern Deleted First Skip Cluster datain dataout Second Skip Cluster K28.5 K29.7 K29.7 K28.5 K29.7 K29.7 K29.7 K29.7 K28.5 K29.7 K28.5 K29.7 K29.7 K29.7 K29.7 Dx.y Dx.y rx_rmfifodatadeleted Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-187 Figure 1-153 shows an example of rate match FIFO insertion in the case where one skip pattern is required to be inserted. In this example, the first skip cluster has a /K28.5/ control pattern followed by three /K29.7/ skip patterns. The second skip cluster has a /K28.5/ control pattern followed by two /K29.7/ skip patterns. The rate match FIFO inserts only one /K29.7/ skip pattern into the first skip cluster. One /K29.7/ skip pattern is inserted into the second cluster. Figure 1-153. Rate Match FIFO Deletion with One Skip Pattern Inserted One Skip Pattern Inserted First Skip Cluster datain dataout Second Skip Cluster K28.5 K29.7 K29.7 K29.7 K28.5 K29.7 K29.7 Dx.y K28.5 K29.7 K29.7 K29.7 K29.7 K28.5 K29.7 K29.7 K29.7 Dx.y Two flags, rx_rmfifofull and rx_rmfifoempty, are forwarded to the FPGA fabric to indicate rate match FIFO full and empty conditions. For more information about the behavior of these two signals, refer to "Rate Match FIFO in Basic Single-Width Mode" on page 1-84. Basic (PMA Direct) Functional Mode In Basic (PMA Direct) functional mode, the Stratix IV GX and GT transceiver datapath contains only PMA blocks. Parallel data is transferred directly between the FPGA fabric and the serializer/deserializer inside the transmitter/receiver PMA. Because all PCS blocks are bypassed in Basic (PMA Direct) mode, you must implement the required PCS logic in the FPGA fabric. You can configure four regular transceiver channels inside each transceiver block in Basic (PMA Direct) functional mode. You can configure two CMU channels inside each transceiver block only in Basic (PMA Direct) functional mode, as they do not support PCS circuitry. In PMA Direct mode, you must create your own logic to support PCS functionality. There are specific reset sequences to be followed in this mode. Use dynamic reconfiguration to dynamically reconfigure the various PMA controls to tailor the transceivers in PMA direct drive mode for a particular application. f For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. For more information about the reset sequence to follow in PMA-Direct mode, refer to the Reset Control and Power Down in Stratix IV Devices chapter. The term `PMA-Direct' is used to describe various configurations in this mode. 1 September 2012 In Basic (PMA Direct) mode, all the PCS blocks are bypassed; therefore, any PCS-type features (for example, phase compensation FIFOs, byte serializer, 8B/10B encoder/decoder, word aligner, deskew FIFO, rate match FIFO, byte deserializer, and byte ordering), must be implemented in the FPGA fabric. In Basic (PMA Direct) mode, you must create your own logic to support PCS functionality. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-188 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Figure 1-154 shows the Stratix IV GX and GT transceiver configured in Basic (PMA Direct) functional mode. The grayed out blocks indicate areas that are not active in this mode. Figure 1-154. Stratix IV GX and GT Transceiver Configured in Basic (PMA Direct) Mode Transmitter Channel PCS wrclk rdclk Word Aligner Deskew FIFO Receiver Channel PMA Rate Match FIFO 8B/10B Decoder Byte Deserializer Byte Ordering Receiver Channel PCS CDR rdclk Serializer 8B/10B Encoder Byte Serializer Deserializer wrclk RX Phase Compensation FIFO FPGA Fabric PIPE Interface PCI Express Hard IP TX Phase Compensation FIFO Transmitter Channel PMA Note to Figure 1-154: (1) The grayed out blocks shown in Figure 1-154 are not available in the CMU channels. Therefore, the CMU channels can be configured to operate as transceiver channels in PMA direct mode only. 1 The grayed out blocks shown in Figure 1-154 are not available in the CMU channels. Therefore, the CMU channels can be configured to operate as transceiver channels in PMA Direct mode only. In Basic (PMA Direct) Mode, you can configure the transceiver channel in two main configurations: Basic (PMA Direct) x1 configuration Basic (PMA Direct) xN configuration You can configure the transceiver in Basic (PMA Direct) x1/ xN mode by setting the appropriate sub-protocol in the Which sub protocol will you be using? field. You can select single-width or double-width by selecting Single/Double in the What is the deserializer block width? field in the ALTGX MegaWizard Plug-In Manager. In single-width mode, the PMA-PLD interface is 8 bit/10 bit wide; whereas in double-width mode, the PMA-PLD interface is 16 bit/20 bit wide. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-189 Table 1-68 lists the Stratix IV GX and GT PLD-PMA interface widths and data rates supported in Basic (PMA Direct) x1/xN single-width and double-width modes. Table 1-68. FPGA Fabric-PMA Interface Widths and Data Rates Supported in Basic (PMA Direct) x1/xN Single-Width and Double-Width Modes for Stratix IV GX and GT Devices Supported Data Rate Range Basic (PMA Direct) Functional Mode x1/xN Single-width mode x1/xN Double-width mode FPGA Fabric-PMA Interface Width Stratix IV GX Stratix IV GT C2 Speed Grade C3, I3, and M3 Speed Grades C4 Speed Grade I1, I2, and I3 Speed Grades 8 bit 0.6 Gbps to 2.6 Gbps 0.6 Gbps to 2.6 Gbps 0.6 Gbps to 2.6 Gbps 600 Mbps to 2.6 Gbps 10 bit 0.6 Gbps to 3.25 Gbps 0.6 Gbps to 3.25 Gbps 0.6 Gbps to 3.25 Gbps 600 Mbps to 3.25 Gbps 16 bit 1.0 Gbps to 5.2 Gbps 1.0 Gbps to 5.2 Gbps 1.0 Gbps to 5.0 Gbps 1.0 Gbps to 5.2 Gbps 20 bit 1.0 Gbps to 6.5 Gbps 1.0 Gbps to 6.5 Gbps 1.0 Gbps to 5.0 Gbps 1.0 Gbps to 6.5 Gbps Basic (PMA Direct) x1 Configuration You can configure a transceiver channel in this mode by setting the which protocol will you be using? field to Basic (PMA Direct) and the which sub protocol will you be using? field to none. In this configuration, the Quartus II software requires one of the two CMU PLLs within the same transceiver block to provide high-speed clocks to the transmitter side of the channel. If the CMU0 or CMU1 channel is configured in Basic (PMA Direct) x1 configuration, use their local clock dividers to provide clock to their respective transmitter channels. f For information about clocking restrictions in Basic (PMA Direct) x1 mode, refer to the "Non-Bonded Basic (PMA Direct) Mode Channel Configurations" section in the Transceiver Clocking in Stratix IV Devices chapter. f For information about routing the clocks to transceiver channels in Basic (PMA Direct) x1 mode, refer to the Transceiver Clocking in Stratix IV Devices chapter. Basic (PMA Direct) xN Configuration You can configure a transceiver channel in this mode by setting the which protocol will you be using field to Basic (PMA Direct) and the which sub protocol will you be using field to xN. In this mode, all the transmitter channels can receive their high-speed clock from the CMU0 PLL from the transceiver blocks or the ATX PLL present on the same side of the device. These clocks are provided through the xN_Top or xN_Bottom clock line. In this mode, if you use a CMU PLL to generate the transceiver channel datapath interface clocks, only the CMU0 central clock divider of the transceiver block containing the CMU PLL is used. f For information about clocking restrictions in Basic (PMA Direct) xN mode, refer to the "Non-Bonded Basic (PMA Direct) Mode Channel Configurations" section in the Transceiver Clocking in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-190 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture f For more information about combining multiple transceiver channels, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Each receiver in a receiver channel has a dedicated CDR that provides a high-speed clock. f For more information about timing closure in Basic (PMA Direct) mode, refer to AN 580: Achieving Timing Closure in Basic (PMA Direct) Functional Mode. Loopback Modes Stratix IV GX and GT devices provide various loopback options that allow you to verify how different functional blocks work in the transceiver channel. The available loopback options are: "Serial Loopback" on page 1-190--available in all functional modes except PCIe mode "Parallel Loopback" on page 1-191--available in either single-width or double-width modes. "Reverse Serial Loopback" on page 1-193--available in Basic mode only "Reverse Serial Pre-CDR Loopback" on page 1-194--available in Basic mode only "PCIe Reverse Parallel Loopback" on page 1-194--available in PCIe mode Serial Loopback The serial loopback option is available for all functional modes except PCIe mode. Figure 1-155 shows the datapath for serial loopback. The data from the FPGA fabric passes through the transmitter channel and gets looped back to the receiver channel, bypassing the receiver buffer. The received data is available to the FPGA logic for verification. Using this option, you can check the working for all enabled PCS and PMA functional blocks in the transmitter and receiver channel. When you enable the serial loopback option, the ALTGX MegaWizard Plug-In Manager provides the rx_seriallpbken port to dynamically enable serial loopback on a channel-by-channel basis. Set the rx_seriallpbken signal to logic high to enable serial loopback. When serial loopback is enabled, the transmitter channel sends the data to both the tx_dataout output port and to the receiver channel. The differential output voltage on the tx_dataout ports is based on the selected VOD settings. The looped back data is received by the receiver CDR and is retimed through different clock domains. You must provide an alignment pattern for the word aligner to enable the receiver channel to retrieve the byte boundary. Suppose the device is not in serial loopback mode and is receiving data from a remote device. At this point, the receiver CDR's recovered clock is locked to the data from that source. If the device is placed in serial loopback mode, the data source to the receiver changes from the remote device to local transmitter channel. This prompts the receiver CDR to start tracking the phase of the new data source. During this time, the receiver CDR's recovered clock may be unstable. As the receiver PCS is running off of this recovered clock, you must place the receiver PCS under reset by asserting the rx_digitalreset signal during this time period. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1 1-191 When moving into or out of serial loopback, you must assert rx_digitalreset for a minimum of two parallel clock cycles. Figure 1-155. Serial Loopback Datapath Transmitter Channel PCS Transmitter Channel PMA BIST PRBS, High-Freq, Low-Freg pattern generator TX Phase Compensation FIFO Byte Serializer 8B/10B Encoder Serializer FPGA Fabric Serial loop back can be dynamically enabled Receiver Channel PCS Receiver Channel PMA BIST PRBS verifier RX Phase Compensation FIFO Byte Ordering Byte Deserializer 8B/10B Decoder Rate Match FIFO Deskew FIFO Word Aligner Deserializer Receiver CDR Parallel Loopback You can configure a transceiver channel in this mode by setting the which protocol will you be using? field to Basic and the which sub protocol will you be using? field to BIST. You can only configure a Receiver and Transmitter transceiver channel in this functional mode. You can configure a transceiver channel in this mode in either a single-width or double-width configuration. The BIST pattern generator and pattern verifier are located near the FPGA fabric in the PCS block of the transceiver channel. This placement allows for testing the complete transmitter PCS and receiver PCS datapaths for bit errors. This mode is primarily used for transceiver channel debugging, if needed. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-192 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture The parallel loopback mode is available only with a built-in 16-bit incremental pattern generator and verifier. The channel width is fixed to 16 bits in this mode. Also in this mode, the incremental pattern 00-FF is looped back to the receiver channel at the PCS functional block boundary before the PMA and is sent to the tx_dataout port. The received data is verified by the verifier. This loopback allows you to verify the complete PCS block. The differential output voltage of the transmitted serial data on the tx_dataout port is based on the selected VOD settings. The datapath for parallel loopback is shown in Figure 1-156. The incremental data pattern is not available to the FPGA logic for verification. Figure 1-156. Enabled PCS Functional Blocks in Parallel Loopback Transmitter Channel PCS BIST incremental pattern generator TX Phase Compensation FIFO Byte Serializer 8B/10B Encoder Transmitter Channel PMA Serializer FPGA Fabric Receiver Channel PCS BIST incremental pattern verifier Receiver Channel PMA Parallel loopback RX Compensation FIFO Byte Deserializer 8B/10B Decoder Word Aligner Deserializer Receiver CDR Table 1-69 lists the enabled PCS functional blocks for single-width and double-width mode. The last column in Table 1-69 lists the supported channel width setting for parallel loopback. Table 1-69. Enabled PCS Functional Blocks for Parallel Loopback 8B/10B Encoder Byte Serializer Data Rate Range Supported Channel Width Setting in the ALTGX MegaWizard Plug-In Manager for Parallel Loopback Single-width mode Enabled Enabled 600 Mbps to 3.125 Gbps 16 Double-width mode Enabled Disabled 1 Gbps to 5 Gbps 16 Configuration The status signals rx_bistdone and rx_bisterr indicate the status of the verifier. The rx_bistdone port is asserted and stays high when the verifier either receives one full cycle of incremental pattern or it detects an error in the receiver data. The rx_bisterr signal is asserted and stays high when the verifier detects an error. You can reset the incremental pattern generator and verifier by asserting the tx_digitalreset and rx_digitalreset signals, respectively. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture 1-193 Reverse Serial Loopback The Reverse Serial Loopback can be set by selecting the radial button under the Loopback tab in the ALTGX Megawizard. In reverse serial loopback mode, the data is received through the rx_datain port, retimed through the receiver CDR and sent out to the tx_dataout port. The received data is also available to the FPGA logic. No dynamic pin control is available to select or deselect reverse serial loopback. Figure 1-157 shows the transceiver channel datapath for reverse serial loopback mode. The active block of the transmitter channel is only the transmitter buffer. You can change the output differential voltage and the pre-emphasis first post tap values on the transmitter buffer through the ALTGX MegaWizard Plug-In Manager or through the dynamic reconfiguration controller. Reverse serial loopback is often implemented when using a bit error rate tester (BERT) on the upstream transmitter. Figure 1-157. Reverse Serial Loopback Datapath (Grayed-Out Blocks are Not Active in this Mode) Transmitter Channel PCS TX Phase Compensation FIFO Byte Serialzier 8B/10B Encoder Serializer FPGA Fabric Receiver Channel PMA Receiver Channel PCS RX Phase Compensation FIFO September 2012 Transmitter Channel PMA Byte Ordering Altera Corporation Byte De-Serializer 8B/10B Decoder Word Aligner Reverse Serial Loopback DeSerializer Receiver CDR Stratix IV Device Handbook Volume 2: Transceivers 1-194 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Block Architecture Reverse Serial Pre-CDR Loopback The reverse serial pre-CDR loopback is available as a subprotocol under Basic functional mode. In reverse serial pre-CDR loopback, the data received through the rx_datain port is looped back to the tx_dataout port before the receiver CDR. The received data is also available to the FPGA logic. Figure 1-158 shows the transceiver channel datapath for reverse serial pre-CDR loopback mode. The active block of the transmitter channel is only the transmitter buffer. You can change the output differential voltage on the transmitter buffer through the ALTGX MegaWizard Plug-In Manager. The pre-emphasis settings for the transmitter buffer cannot be changed in this configuration. Figure 1-158. Reverse Serial Pre-CDR Loopback Datapath Transmitter Channel PCS Transmitter Channel PMA Serializer FPGA Fabric Receiver Channel PCS RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Word Aligner Receiver Channel PMA DeSerializer Reverse Serial Pre-CDR Loopback Receiver CDR PCIe Reverse Parallel Loopback PCIe reverse parallel loopback is only available in PCIe functional mode for Gen1 and Gen2 data rates. As shown in Figure 1-159, the received serial data passes through the receiver CDR, deserializer, word aligner, and rate matching FIFO buffer. It is then looped back to the transmitter serializer and transmitted out through the tx_dataout port. The received data is also available to the FPGA fabric through the rx_dataout port. This loopback mode is compliant with the PCIe specification 2.0. To enable this loopback mode, assert the tx_detectrxloopback port. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block 1-195 In Figure 1-159, the grayed areas show the inactive paths when the PCIe reverse parallel loopback mode is enabled. Figure 1-159. PCIe Reverse Parallel Loopback Mode Datapath (Grayed-Out Blocks are Not Active in this Mode) Transmitter Channel PCS wrclk rdclk Reverse Parallel Loopback Path CDR Deserializer Word Aligner Deskew FIFO Receiver Channel PMA Rate Match FIFO 8B/10B Decoder Byte Deserializer RX Phase Compensation FIFO Receiver Channel PCS Byte Ordering PCIe hard IP FPGA Fabric 8B10B Encoder Byte Serializer rdclk PIPE Interface wrclk Serializer TX Phase Compensation FIFO Transmitter Channel PMA Auxiliary Transmit (ATX) PLL Block Stratix IV GX and GT transceivers contain the ATX PLL block that you can use to generate high-speed clocks for the transmitter channels on the same side of the device. Each: Stratix IV GX device has 6G ATX PLL Stratix IV GT device has 6G ATX PLL and 10G ATX PLLs f For data rates supported by these ATX PLLs, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. 6G ATX PLL Block Stratix IV GX can have either two (one on each side of the device) or four (two on each side of the device) 6G ATX PLLs, depending on the specific devices. f For data rates supported by 6G ATX PLLs, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-196 Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block 10G ATX PLL Block Each Stratix IV GT device has two 10G ATX PLL blocks, one located on each side of the device. The 10G ATX PLLs provide low-jitter transceiver clocks to implement 40G/100G Ethernet and SFI-S links specified by IEEE802.3ba and OIF specifications. In EP4S40G2F40 and EP4S40G5H40 devices, you can use each 10G ATX PLL to generate transceiver clocks for up to six channels at data rates of up to 11.3 Gbps each. In EP4S100G2F40, EP4S100G5H40, and EP4S100G5F45 devices, you can use each 10G ATX PLL to generate transceiver clocks for up to 12 channels at data rates of up to 11.3 Gbps each. Figure 1-163 and Figure 1-164 show transceiver channels that support data rates up to 11.3 Gbps in each Stratix IV GT device. The 10G ATX PLL block consists of: 10G ATX PLL--Synthesizes the input reference clock to generate the high-speed serial transceiver clock at frequency of half the configured data rate ATX clock divider block--Divides the high-speed serial clock from the 10G ATX PLL to generate the low-speed parallel transceiver clock The 10G ATX PLL architecture is functionally similar to the 6G ATX PLL architecture, except that it is optimized for the 10 Gbps data rate range. Figure 1-160 shows the location of the ATX PLL blocks in two transceiver block device families. Figure 1-160. Location of ATX PLL Blocks in a Four-Transceiver Block Stratix IV GX Device (Two on Each Side) GXBL1 GXBR1 ATX PLL L0 (6G) ATX PLL R0 (6G) GXBL0 Stratix IV Device Handbook Volume 2: Transceivers GXBR0 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block 1-197 Figure 1-161 shows the location of the ATX PLL blocks in three transceiver block device families (for 230K and 530K devices and all other devices). Figure 1-161. Location of ATX PLL Blocks with a Six Transceiver Block Stratix IV GX Device (Three on Each side) In 230K and 530K Stratix IV GX Devices GXBL2 GXBR2 GXBL1 GXBR1 ATX PLL L0 (6G) ATX PLL R0 (6G) GXBL0 GXBR0 In Stratix IV GX Devices Other than 230K and 530K GXBL2 GXBR2 ATX PLL L1 (6G) ATX PLL R1 (6G) GXBL1 GXBR1 ATX PLL L0 (6G) ATX PLL R0 (6G) GXBL0 GXBR0 Figure 1-162 shows the location of the ATX PLL blocks in four transceiver block device families. Figure 1-162. Location of ATX PLL Blocks in an Eight-Transceiver Block Stratix IV GX Device (Four on Each Side) September 2012 Altera Corporation GXBL3 GXBR3 GXBL2 GXBR2 ATX PLL L1 (6G) ATX PLL R1 (6G) GXBL1 GXBR1 ATX PLL L0 (6G) ATX PLL R0 (6G) GXBL0 GXBR0 Stratix IV Device Handbook Volume 2: Transceivers 1-198 Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block Figure 1-163 and Figure 1-164 show the locations of the 6G and 10G ATX PLLs in each Stratix IV GT device. Figure 1-163. Location of Transceiver Channel and PLL in Stratix IV GT Devices (EP4S40G2F40, EP4S40G5H40, EP4S100G2F40 and EP4S100G5H40) Transceiver Block GXBL2 ATX PLL L1 (10G) Transceiver Block GXBL1 ATX PLL L0 (6G) Transceiver Block GXBL0 Transceiver Block GXBR2 ATX PLL R1 (10G) Transceiver Block GXBR1 ATX PLL R0 (6G) Transceiver Block GXBR0 Figure 1-164. Location of Transceiver Channel and PLL in Stratix IV GT Devices (EP4S100G5F45) Transceiver Block GXBL3 Transceiver Block GXBR3 ATX PLL L2 (10G) ATX PLL R2 (10G) Transceiver Block GXBL2 Transceiver Block GXBR2 ATX PLL L1 (6G) Transceiver Block GXBL1 ATX PLL L0 (6G) Transceiver Block GXBL0 ATX PLL R1 (6G) Transceiver Block GXBR1 ATX PLL R0 (6G) Transceiver Block GXBR0 Input Reference Clocks for the ATX PLL Block The 6G ATX PLL block does not have a dedicated reference clock pin. The following are the possible input reference clock sources: REFCLKs from the transceiver blocks on the same side of the device if the corresponding CMU channels are not used as transceiver channels Input reference clock provided through the PLL cascade clock network Clock inputs connected through the global clock lines Altera recommends using the REFCLK pins from the adjacent transceiver block below the ATX PLL block to improve performance. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block 1-199 For the 10G ATX PLL, Stratix IV GT devices only allow driving the reference clock source from one of the dedicated reflck pins on the same side of the device. 1 For improved jitter performance, Altera strongly recommends using the REFCLK pins of the transceiver block located immediately below the 10G ATX PLL block to drive the input reference clock. f For more information about the input reference clocks for ATX PLLs, refer to the Transceiver Clocking for Stratix IV Devices chapter. Architecture of the ATX PLL Block Figure 1-165 shows the ATX PLL block components (the ATX PLL, ATX clock divider, and a shared control signal generation block). Figure 1-165. ATX PLL Block ATX PLL high-speed Clock (1) ATX PLL Block pll_powerdown cascaded PLL clock global clock line high-speed serial clock for bonded modes (2) ATX PLL input reference clock ATX clock divider block ATX PLL ITB clock lines 8 PCIE_gen2switch PCI Express rate switch controller PCIErateswitch PCIE_gen2switch_done Notes to Figure 1-165: (1) In non-bonded functional modes (for example, CEI functional mode), the transmitter channel uses the transmitter local clock divider to divide this high-speed clock output to provide clocks for its PMA and PCS blocks. (2) This is used in Basic x4, x8, and PCIe x4 and x8 functional modes. The functional blocks on the ATX PLL are similar to the blocks explained in "CMU0 PLL" on page 1-102. The values of the /M and /L divider settings in the ATX PLL are automatically selected by the Quartus II software based on the transceiver channel configuration. The ATX PLL high-speed clock output provides high-speed serial clocks for non-bonded functional modes such as CEI (with the "none" subprotocol). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-200 Chapter 1: Transceiver Architecture in Stratix IV Devices Auxiliary Transmit (ATX) PLL Block ATX Clock Divider The ATX clock divider divides the ATX PLL high-speed clock and provides high-speed serial and low-speed parallel clock for bonded functional modes such as PCIe (x4 and x8), Basic x4 and x8, and PMA-Direct mode with xN configuration. For PCIe functional mode support, the ATX clock divider consists of the PCIe rateswitch circuit to enable dynamic rateswitch between PCIe Gen1 and Gen2 data rates. For more information on this circuit, refer to "CMU0 Channel" on page 1-101. The clock outputs from the ATX PLL block are provided to the transmitter channels through the xN_Top or xN_bottom clock lines, as shown in Figure 1-166. f For more information, refer to the Transceiver Clocking for Stratix IV Devices chapter. Figure 1-166. ATX Clock Divider ATX clock divider block PCIE_gen2switch_done Low-Speed Parallel Clocks (for bonded modes) PCIE_gen2switch ATX PLL high-speed clock output PCIe clockswitch circuit 0 /S (4, 5, 8, 10) coreclkout to FPGA fabric (for bonded modes) /2 1 High-Speed Serial Clock (for bonded modes) The Differences Between 10G ATX PLL, 6G ATX PLL, and CMU PLL Table 1-70 lists the differences between the 10G ATX PLL, 6G ATX PLL, and CMU PLL. Table 1-70. Differences Between the 10G ATX PLL, 6G ATX PLL, and CMU PLL (Part 1 of 2) Difference Category/PLLs Available in 10G ATX PLL 6G ATX PLL CMU PLL Stratix IV GT device Stratix IV GX and GT devices Stratix IV GX and GT devices 4.8 to 5.4 and 6.0 and 6.5 Data rates (Gbps) Input reference clock options 2.4 to 2.7 and 3.0 and 3.25 (1) 1.2 to 1.35 and 1.5 to 1.625 (1) 9.9 to 11.3 Only dedicated refclk pins on the same side of the device (2), (3) Clock inputs connected through the inter transceiver block (ITB) lines. Clock inputs connected through the PLL cascade clock network. Clock inputs connected through the global clock lines. (3) Stratix IV Device Handbook Volume 2: Transceivers Up to 8.5 for Stratix IV GX devices Up to 11.3 for Stratix IV GT devices Clock inputs connected through the inter transceiver block (ITB) lines. clock inputs connected through the PLL cascade clock network. Clock inputs connected through the global clock lines, refclk0 and refclk1 clock input, dedicated refclks in the transceiver block. (3) September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks 1-201 Table 1-70. Differences Between the 10G ATX PLL, 6G ATX PLL, and CMU PLL (Part 2 of 2) Difference Category/PLLs 10G ATX PLL Power Supply--VCCA_L/R (V) options for PLLs 6G ATX PLL 2.5 or 3.0 or 3.3 3.3 Lower when compared with the CMU PLL (5) Phase noise CMU PLL 3.0 or (4) 3.3 Lower when compared with the CMU PLL (5) (4) Higher when compared with the ATX PLLs (5) Notes to Table 1-70: (1) Using the L dividers available in ATX PLLs. (2) For improved jitter performance, Altera strongly recommends using the refclk pins of the transceiver block located immediately below the 10G ATX PLL block to drive the input reference clock. (3) For more information, refer to the Input Reference Clock Source table in the Stratix IV Transceiver Clocking chapter. (4) Option in Stratix IV GT devices. (5) For more information about phase noise and PLL bandwidths of ATX and CMU PLLs, refer to the characterization reports. Calibration Blocks Stratix IV GX and GT devices contain calibration circuits that calibrate the OCT resistors and the analog portions of the transceiver blocks to ensure that the functionality is independent of process, voltage, or temperature variations. Calibration Block Location Figure 1-167 shows the location and number of calibration blocks available for different Stratix IV GX and GT devices. In Figure 1-167 through Figure 1-172, the calibration block R0 and L0 refer to the calibration blocks on the right and left side of the devices, respectively. Figure 1-167. Calibration Block Locations in Stratix IV GX and GT Device with Two Transceiver Blocks (on Each Side) GXBL1 GXBR1 ATX PLL L0 ATX PLL R0 GXBL0 Calibration Block L0 2K September 2012 Altera Corporation Stratix IV GX and GT Device GXBR0 Calibration Block R0 2K Stratix IV Device Handbook Volume 2: Transceivers 1-202 Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks Figure 1-168 shows Stratix IV GX 230K and 530K devices that have three transceiver blocks each on the left and right side and one ATX PLL block on each side. Figure 1-168. Calibration Block Locations in Stratix IV GX 230K and 530K Devices with Three Transceiver Blocks (on Each Side) 2K Calibration Block L1 Calibration Block R1 GXBL2 GXBR2 GXBL1 Stratix IV Device Handbook Volume 2: Transceivers GXBR1 ATX PLL L0 ATX PLL R0 GXBL0 GXBR0 Calibration Block L0 2K Stratix IV GX Device 2K Calibration Block R0 2K September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks 1-203 Figure 1-169 shows Stratix IV GX devices other than 230K and 530K that have three transceiver blocks each on the left and right side and two ATX PLL blocks on each side. Figure 1-169. Calibration Block Locations in Stratix IV GX Devices Other than 230K and 530K with Three Transceiver Blocks (on Each Side) 2K 2K Calibration Block L1 Calibration Block R1 GXBL2 GXBR2 ATX PLL L1 ATX PLL R0 GXBL1 September 2012 Altera Corporation GXBR1 ATX PLL L0 ATX PLL R0 GXBL0 GXBR0 Calibration Block L0 2K Stratix IV GX Device Calibration Block R0 2K Stratix IV Device Handbook Volume 2: Transceivers 1-204 Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks Figure 1-170 shows Stratix IV GX devices that have four transceiver blocks each on the left and right side and two ATX PLL blocks on each side. Figure 1-170. Calibration Block Locations in Stratix IV GX Devices with Four Transceiver Blocks (on Each Side) 2K 2K Calibration Block L1 Calibration Block R1 GXBL3 GXBR3 GXBL2 ATX PLL L1 GXBR2 Stratix IV GX Device ATX PLL R1 GXBL1 GXBR1 ATX PLL L0 ATX PLL R0 GXBL0 GXBR0 Calibration Block L0 Calibration Block R0 2K 2K Figure 1-171 shows Stratix IV GX devices that have two transceiver blocks only on the right side of the device. Figure 1-171. Calibration Block Locations in Stratix IV GX Devices with Two Transceiver Blocks (Right Side Only) GXBR1 Stratix IV GX Device ATX PLL R0 GXBR0 Calibration Block R0 2K The Quartus II software automatically selects the appropriate calibration block based on the assignment of the transceiver tx_dataout and rx_datain pins. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks 1-205 Calibration The calibration block internally generates a constant internal reference voltage, independent of process, voltage, or temperature variations. It uses the internal reference voltage and external reference resistor (you must connect the resistor to the RREF pin) to generate constant reference currents. These reference currents are used by the analog block calibration circuit to calibrate the transceiver blocks. The OCT calibration circuit calibrates the OCT resistors present in the transceiver channels. You can enable the OCT resistors in the transceiver channels through the ALTGX MegaWizard Plug-In Manager. You must connect a separate 2 k (tolerance max 1%) external resistor on each RREF pin in the Stratix IV GX and GT device to ground. To ensure proper operation of the calibration block, the RREF resistor connection in the board must be free from external noise. Input Signals to the Calibration Block The ALTGX MegaWizard Plug-In Manager provides the cal_blk_clk and cal_blk_powerdown ports to control the calibration block: September 2012 cal_blk_clk--you must use the cal_blk_clk port to provide input clock to the calibration clock. The frequency of cal_blk_clk must be within 10 MHz to 125 MHz (this range is preliminary. Final values will be available after characterization). You can use dedicated clock routes such as the global or regional clock. If you do not have suitable input reference clock or dedicated clock routing resources available, use divide-down logic from the FPGA fabric to generate a slow clock and use local clocking routing. Drive the cal_blk_clk port of all ALTGX instances that are associated with the same calibration block from the same input pin or logic. cal_blk_powerdown--you can perform calibration multiple times by using the cal_blk_powerdown port available through the ALTGX MegaWizard Plug-In Manager. Assert this signal for approximately 500 ns. Following de-assertion of cal_blk_powerdown, the calibration block restarts the calibration process. Drive the cal_blk_powerdown port of all ALTGX instances that are associated with the same calibration block from the same input pin or logic. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-206 Chapter 1: Transceiver Architecture in Stratix IV Devices Calibration Blocks Figure 1-172 shows the required inputs to the calibration block. Figure 1-172. Input Signals to the Calibration Blocks RREF pin cal_blk_clk cal_blk_powerdown (1) Calibration Block Internal Reference Voltage Generator Reference Signal OCT Calibration Control OCT Calibration Circuit Analog Block Calibration Circuit Analog Block Calibration Control Note to Figure 1-172: (1) The transceiver on-chip termination calibration process takes approximately 33,000 cal_blk_clk cycles from the de-assertion of the cal_blk_powerdown signal. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Built-In Self Test Modes 1-207 Built-In Self Test Modes This section describes Built-In Self Test (BIST) modes. BIST Mode Pattern Generators and Verifiers Each transceiver channel in the Stratix IV GX and GT devices contain a different BIST pattern generator and verifier. Using these BIST patterns, you can verify the functionality of the functional blocks in the transceiver channel without requiring user logic. The BIST functionality is provided as an optional mechanism for debugging transceiver channels. Figure 1-173 shows the enabled input and output ports when you select BIST mode (except incremental patterns). Figure 1-173. Input and Output Ports for BIST Modes tx_datain[] tx_dataout tx_digitalreset rx_digitalreset rx_seriallpbken[] (1) rx_bisterr (2) Built-In Self Test (BIST) rx_bistdone (2) pll_inclk Notes to Figure 1-173: (1) rx_serilalpbken is required in PRBS. (2) rx_bisterr and rx_bistdone are only available in PRBS and BIST modes. Three types of pattern generators and verifiers are available: September 2012 BIST incremental data generator and verifier--This is only available in parallel loopback mode. For more information, refer to "Serial Loopback" on page 1-190. High frequency and low frequency pattern generator--The high frequency patterns generate alternate ones and zeros and the low frequency patterns generate five ones and five zeroes in single-width mode and ten ones and ten zeroes in double-width mode. These patterns do not have a corresponding verifier. You can enable the serial loopback option to dynamically loop the generated pattern to the receiver channel using the rx_seriallpbken port. Therefore, the 8B/10B encoder/decoder blocks are bypassed in the Basic PRBS mode. Pseudo Random Binary Sequence (PRBS) generator and verifier--The PRBS generator and verifier interface with the serializer and deserializer in the PMA blocks. The advantage of using a PRBS data stream is that the randomness yields an environment that stresses the transmission medium. In the data stream, you can observe both random jitter and deterministic jitter using a time interval analyzer, bit error rate tester, or oscilloscope. The PRBS repeats after completing an iteration. The number of bits the PRBSx pattern sends before repeating the pattern is (2 ^x -1) bits. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-208 Chapter 1: Transceiver Architecture in Stratix IV Devices Built-In Self Test Modes Different PRBS patterns are available as a subprotocol under Basic functional mode for single-width and double-width mode, as shown in the following sections. You can enable the serial loopback option in Basic PRBS mode to loop the generated pattern to the receiver channel. This creates a rx_seriallpbken port that you can use to dynamically control the serial loopback. The 8B/10B encoder/decoder blocks are bypassed in Basic PRBS mode. Figure 1-174 shows the datapath for the PRBS patterns. The generated PRBS pattern is sent to the transmitter serializer. The verifier checks the data from the word aligner. Figure 1-174. BIST PRBS, High Frequency, and Low Frequency Pattern Datapath Transmitter Channel PCS Transmitter Channel PMA BIST PRBS, High-Freq, Low-Freg pattern generator TX Phase Compensation FIFO Byte Serializer 8B/10B Encoder Serializer FPGA Fabric Serial loop back can be dynamically enabled Receiver Channel PCS Receiver Channel PMA BIST PRBS verifier RX Phase Compensation FIFO Byte Ordering Byte Deserializer 8B/10B Decoder Rate Match FIFO Deskew FIFO Word Aligner Deserializer Receiver CDR PRBS in Single-Width Mode Table 1-71 lists the various PRBS patterns and corresponding word alignment patterns for PRBS in single-width mode configuration. Table 1-71. Available PRBS, High Frequency, and Low Frequency Patterns in Single-Width Mode Patterns Polynomial Channel Width of 8 Bit (1) Word Alignment Pattern with Channel Width 8 Bit Maximum Data Rate With Channel Width 8 Bit (Gbps) Channel Width of 10 Bit (1) Word Alignment Pattern Maximum Data Rate with Channel Width 10 Bit (Gbps) PRBS 7 X7 + X 6 + 1 Y 16'h3040 2.5 N NA N/A PRBS 8 X8 + X 7 + 1 Y 16'hFF5A 2.5 N NA N/A PRBS 10 X10 PRBS 23 X23 High frequency (2) + X7 +1 N NA N/A Y 10'h3FF 3.125 + X18 +1 Y 16'hFFFF 2.5 N NA N/A 1010101010 Y NA 2.5 Y NA 3.125 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Built-In Self Test Modes 1-209 Table 1-71. Available PRBS, High Frequency, and Low Frequency Patterns in Single-Width Mode Patterns Polynomial Low Frequency (2) 0000011111 Channel Width of 8 Bit (1) Word Alignment Pattern with Channel Width 8 Bit Maximum Data Rate With Channel Width 8 Bit (Gbps) Channel Width of 10 Bit (1) Word Alignment Pattern Maximum Data Rate with Channel Width 10 Bit (Gbps) N NA N/A Y NA 3.125 Notes to Table 1-71: (1) Channel width refers to the What is the channel width? option in the General screen of the ALTGX MegaWizard Plug-In Manager. Based on the selection, an 8 or 10 bits wide pattern is generated as indicated by a Yes (Y) or No (N). (2) A verifier is not available for the specified patterns. The status signals rx_bistdone and rx_bisterr indicate the status of the verifier. The rx_bistdone port gets asserted and stays high when the verifier either receives one full cycle of incremental pattern or it detects an error in the receiver data. The rx_bisterr signal gets asserted and stays high when the verifier detects an error. You can reset the PRBS pattern generator and verifier by asserting the tx_digitalreset and rx_digitalreset signals, respectively. PRBS in Double-Width Mode Table 1-72 lists the various PRBS patterns and corresponding word alignment patterns for PRBS in double-width mode configuration. Table 1-72. Available PRBS, High Frequency, and Low Frequency Patterns in Double-Width Mode Patterns Polynomial Channel Width of 16-Bit (1) Word Alignment Pattern with Channel Width of 16-Bit Maximum Data Rate with Channel Width of 16-Bit (Gbps) Channel Width of 20-Bit (1) Word Alignment Pattern Maximum Data Rate with Channel Width of 20-Bit (Gbps) PRBS 7 X7 + X6 + 1 Y 16'h3040 5 Y 20'h43040 6.375 PRBS 23 X23 + X18 + 1 Y 32'h007FFFF F 5 Y 40'h00007FF FFF 6.375 (2) 1010101010 Y NA 5 Y N/A 6.375 (2) 0000011111 N NA N/A Y N/A 6.375 High frequency Low Frequency Notes to Table 1-72: (1) Channel width refers to the what is the channel width? option in the General screen of the ALTGX MegaWizard Plug-In Manager. Based on the selection, A 16 or 20 bits wide pattern is generated as indicated by a Yes (Y) or No (N). (2) Verifier is not available for the specified patterns. The status signals rx_bisterr and rx_bistdone are available to indicate the status of the verifier. For more information about the behavior of these status signals, refer to "Single-Width Mode" on page 1-21. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-210 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Transceiver Port Lists Instantiate the Stratix IV GX and GT transceivers using the ALTGX megafunction instance in the Quartus II MegaWizard Plug-In Manager. The ALTGX megafunction instance allows you to configure transceivers for your intended protocol and select optional control and status ports to and from the instantiated transceiver channels. Table 1-73 through Table 1-79 list a brief description of the ALTGX megafunction ports. Table 1-73 lists the ALTGX megafunction transmitter ports. Table 1-73. Stratix IV GX and GT ALTGX Megafunction Ports: Transmitter Ports (Part 1 of 3) Port Name Input/ Output Clock Domain Description Scope Transmitter Phase Compensation FIFO tx_datain Input Synchronous to tx_clkout or coreclkout. tx_clkout for non-bonded modes. coreclk for bonded modes. Parallel data input from the FPGA fabric to the transmitter. Bus width--depends on the channel width multiplied by the number of channels per instance. Channel FPGA fabric-transceiver interface clock. tx_clkout Output Clock signal Bonded channel configurations--not available. Non-bonded channel configurations--each channel has a tx_clkout signal. Use this clock signal to clock the parallel data tx_datain from the FPGA fabric into the transmitter. Channel Optional write clock port for the transmitter phase compensation FIFO. tx_coreclk Input Clock signal tx_phase_comp_fifo_ error Stratix IV Device Handbook Volume 2: Transceivers Output Synchronous to tx_clkout/ coreclkout clock signal. If not selected--the Quartus II software automatically selects tx_clkout/coreclkout as the write clock for transmitter phase compensation FIFO. If selected--you must drive this port with a clock that is frequency locked to tx_clkout/coreclkout. Transmitter phase compensation FIFO full or empty indicator. Channel A high level--the transmitter phase compensation FIFO is either full or empty. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-211 Table 1-73. Stratix IV GX and GT ALTGX Megafunction Ports: Transmitter Ports (Part 2 of 3) Port Name Input/ Output Clock Domain Description Scope 8B/10B Encoder 8B/10B encoder /Kx.y/ or /Dx.y/ control. tx_ctrlenable Input Synchronous to tx_clkout/ coreclkout clock signal. When asserted high--the 8B/10B encoder encodes the data on the tx_datain port as a /Kx.y/ control code group. When de-asserted low--it encodes the data on the tx_datain port as a /Dx.y/ data code group. Channel Channel Width: 8--tx_ctrlenable = 1 16--tx_ctrlenable = 2 32--tx_ctrlenable = 4 8B/10B encoder force disparity control. tx_forcedisp Input Asynchronous signal. Minimum pulse width is one parallel clock cycles. When asserted high--forces the 8B/10B encoder to encode the data on the tx_datain port with a positive or negative disparity depending on the tx_dispval signal level. When de-asserted low--the 8B/10B encoder encodes the data on the tx_datain port according to the 8B/10B running disparity rules. Channel Width: 8--tx_forcedisp = 1 16--tx_forcedisp = 2 32--tx_forcedisp = 4 Channel 8B/10B encoder force disparity value. Input tx_dispval Asynchronous signal. Minimum pulse width is one parallel clock cycles. A high level--when the tx_forcedisp signal is asserted high, it forces the 8B/10B encoder to encode the data on the tx_datain port with a negative starting running disparity. A low level--when the tx_forcedisp signal is asserted high, it forces the 8B/10B encoder to encode the data on the tx_datain port with a positive starting running disparity. September 2012 Altera Corporation Channel Channel Width: 8--tx_dispval = 1 16--tx_dispval = 2 32--tx_dispval = 4 Stratix IV Device Handbook Volume 2: Transceivers 1-212 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-73. Stratix IV GX and GT ALTGX Megafunction Ports: Transmitter Ports (Part 3 of 3) Port Name tx_invpolarity Input/ Output Input Clock Domain Asynchronous signal. Minimum pulse width is two parallel clock cycles. Description Scope Transmitter polarity inversion control. This feature is useful for correcting situations in which the positive and negative signals of the differential serial link are accidentally swapped during board layout. When asserted high in single-width modes--the polarity of every bit of the 8-bit or 10-bit input data to the serializer gets inverted. Channel When asserted high in double-width modes--the polarity of every bit of the 16-bit or 20-bit input data to the serializer gets inverted. Transmitter Physical Media Attachment tx_dataout fixedclk Stratix IV Device Handbook Volume 2: Transceivers Output N/A Input Clock signal Transmitter serial data output port. Channel 125-MHz clock for receiver detect and offset cancellation in PCIe mode. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-213 Table 1-74 lists the ALTGX megafunction receiver ports. Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 1 of 7) Port Name Input/ Output Clock Domain Description Scope Word alignment synchronization status indicator. rx_syncstatus Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Automatic synchronization state machine mode--this signal is driven high if the conditions required to remain in synchronization are met. Driven low if the conditions required to lose synchronization are met. Manual alignment mode--the behavior of this signal depends on whether the transceiver is configured in single-width or double-width mode. rx_ala2size September 2012 Altera Corporation Channel width: 8/10--rx_syncstatus = 1 16/20--rx_syncstatus = 2 32/40-- rx_syncstatus = 4 Bit-slip control for the word aligner configured in bit-slip mode. Input Asynchronous signal. Minimum pulse width is two parallel clock cycles. Available only in SONET OC-12 and OC-48 modes. Select between these options: Input Asynchronous Signal. Minimum pulse width is two parallel clock cycles. Output rx_rlv Bit-Slip mode--not available. For more information, refer to "Word Aligner in Single-Width Mode" on page 1-60 and "Word Aligner in Double-Width Mode" on page 1-66. rx_bitslip Asynchronous signal. Driven for a minimum of two recovered clock cycles in configurations without byte serializer and a minimum of three recovered clock cycles in configurations with byte serializer. Channel At every rising edge, word aligner slips one bit into the received data stream, effectively shifting the word boundary by one bit. 0 = 16-bit A1A2 1 = 32-bit A1A1A2A2 Channel Channel Run-length violation indicator. A high pulse is driven when the number of consecutive 1s or 0s in the received data stream exceeds the programmed run length violation threshold. Channel Stratix IV Device Handbook Volume 2: Transceivers 1-214 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 2 of 7) Port Name rx_invpolarity Input/ Output Input Clock Domain Asynchronous Signal. Minimum pulse width is two parallel clock cycles. Description Scope Generic receiver polarity inversion control. Useful feature for correcting situations where the positive and negative signals of the differential serial link are accidentally swapped during board layout. When asserted high in single-width modes--the polarity of every bit of the 8-bit or 10-bit input data word to the word aligner gets inverted. Channel When asserted high in double-width modes--the polarity of every bit of the 16-bit or 20-bit input data to the word aligner gets inverted. Receiver bit reversal control. This is a useful feature where the link transmission order is MSB to LSB. rx_revbitorderwa Input Asynchronous Signal. Minimum pulse width is two parallel clock cycles. Available only in Basic single-width and double-width modes with the word aligner configured in bit-slip mode. When asserted high in Basic single-width modes--the 8-bit or 10-bit data D[7:0] or D[9:0] at the output of the word aligner gets rewired to D[0:7] or D[0:9], respectively. rx_revbyteorderwa Stratix IV Device Handbook Volume 2: Transceivers Input Asynchronous Signal. Minimum pulse width is two parallel clock cycles. Channel When asserted high in Basic double-width modes--the 16-bit or 20-bit data D[15:0] or D[19:0] at the output of the word aligner gets rewired to D[0:15] or D[0:19], respectively. Receiver byte reversal control. This is a useful feature in situations where the MSByte and LSByte of the transmitted data are erroneously swapped. Available only in Basic double-width mode. When asserted high, the MSByte and LSByte of the 16- and 20-bit data at the output of the word aligner get swapped. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-215 Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 3 of 7) Port Name Input/ Output Clock Domain Description Scope Deskew FIFO XAUI deskew FIFO channel aligned indicator. rx_channelaligned Output Available only in XAUI mode. A high level--the XAUI deskew state machine is either in ALIGN_ACQUIRED_1, ALIGN_ACQUIRED_2, ALIGN_ACQUIRED_3, or ALIGN_ACQUIRED_4 state, as specified in the PCS deskew state diagram in the IEEE P802.3ae specification. Synchronous to coreclkout clock signal Transceiver block A low level--the XAUI deskew state machine is either in LOSS_OF_ALIGNMENT, ALIGN_DETECT_1, ALIGN_DETECT_2, or ALIGN_DETECT_3 state, as specified in the PCS deskew state diagram in the IEEE P802.3ae specification. Rate Match (Clock Rate Compensation) FIFO rx_ rmfifodatainserted rx_rmfifodatadeleted rx_rmfifofull rx_rmfifoempty September 2012 Altera Corporation Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes. Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Rate match FIFO insertion status indicator. A high level--the rate match pattern byte has inserted to compensate for the PPM difference in reference clock frequencies between the upstream transmitter and the local receiver. Channel Rate match FIFO deletion status indicator. A high level--the rate match pattern byte got deleted to compensate for the PPM difference in reference clock frequencies between the upstream transmitter and the local receiver. Channel Rate match FIFO full status indicator. A high level indicates that the rate match FIFO is full. Without byte serializer --driven a minimum of two recovered clock cycles. With byte serializer--driven a minimum of three recovered clock cycles. Channel Rate match FIFO empty status indicator. A high level--the rate match FIFO is empty. Without byte serializer--driven a minimum of two recovered clock cycles. With byte serializer--driven a minimum of three recovered clock cycles. Channel Stratix IV Device Handbook Volume 2: Transceivers 1-216 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 4 of 7) Port Name Input/ Output Clock Domain Description Scope 8B/10B Decoder Receiver control code indicator. rx_ctrldetect Output Synchronous to rx_clkout, tx_clkout, and coreclkout clock signals Available in configurations with 8B/10B decoder. A high level--the associated received code group is a control (/Kx.y/) code group. A low level--the associated received code group is a data (/Dx.y/) code group. Channel The width of this signal depends on the following channel width: Channel Width: 8--rx_ctrldetect = 1 16--rx_ctrldetect = 2 32--rx_ctrldetect = 4 8B/10B code group violation or disparity error indicator. rx_errdetect Output Available in configurations with 8B/10B decoder. A high level--a code group violation or disparity error was detected on the associated received code group. Use with the rx_disperr signal to differentiate between a code group violation and/or a disparity error as follows: Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes [rx_errdetect: rx_disperr] 2'b00--no error 2'b10--code group violation 2'b11--disparity error or both Channel Channel Width: 8--rx_errdetect = 1 16--rx_errdetect = 2 32--rx_errdetect = 4 8B/10B disparity error indicator port. rx_disperr Stratix IV Device Handbook Volume 2: Transceivers Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Available in configurations with 8B/10B decoder. A high level--a disparity error was detected on the associated received code group. Channel Channel Width: 8--rx_disperr = 1 16--rx_disperr = 2 32--rx_disperr = 4 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-217 Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 5 of 7) Port Name Input/ Output Clock Domain Description Scope 8B/10B running disparity indicator. rx_runningdisp Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Available in configurations with the 8B/10B decoder. A high level--the data on the rx_dataout port was received with a negative running disparity. A low level--the data on the rx_dataout port was received with a positive running disparity. Channel Width: 8--rx_runningdisp = 1 16--rx_runningdisp = 2 32--rx_runningdisp = 4 Channel Byte Ordering Block Enable byte ordering control. rx_enabyteord Input Asynchronous signal rx_ byteorderalignstatus Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes Available in configurations with the byte ordering block enabled. The byte ordering block is rising-edge sensitive to this signal. Channel A low-to-high transition triggers the byte ordering block to restart the byte ordering operation. Byte ordering status indicator. Available in configurations with the byte ordering block enabled. A high level--the byte ordering block has detected the programmed byte ordering pattern in the LSByte of the received data from the byte deserializer. Channel Receiver Phase Compensation FIFO rx_dataout Output Synchronous to rx_clkout or coreclkout. rx_clkout for non-bonded modes. coreclkout for bonded modes rx_clkout Output Clock signal Parallel data output from the receiver to the FPGA fabric. The bus width depends on the channel width multiplied by the number of channels per instance. Channel Recovered clock from the receiver channel. September 2012 Altera Corporation Available only when the rate match FIFO is not used in the receiver datapath. Channel Stratix IV Device Handbook Volume 2: Transceivers 1-218 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 6 of 7) Port Name Input/ Output Clock Domain Description Scope Optional read clock port for the receiver phase compensation FIFO. rx_coreclk Input Clock signal Synchronous to tx_clkout or coreclkout. rx_phase_comp_fifo_ error Output tx_clkout for non-bonded modes. If not selected--the Quartus II software automatically selects rx_clkout/tx_clkout/ coreclkout as the read clock for the receiver phase compensation FIFO. If selected--drive this port with a clock that has 0 PPM difference with respect to rx_clkout/tx_clkout/ coreclkout. Receiver phase compensation FIFO full or empty indicator. Channel A high level--the receiver phase compensation FIFO is either full or empty. Channel coreclkout for bonded modes. Receiver Physical Media Attachment (PMA) rx_datain Input N/A rx_cruclk Input Clock signal Receiver serial data input port. Channel Input reference clock for the receiver clock and data recovery. Channel Receiver CDR lock-to-reference indicator. rx_pll_locked Output Asynchronous signal A high level--the receiver CDR is locked to the input reference clock. Channel A low level--the receiver CDR is not locked to the input reference clock. Receiver CDR lock mode indicator. rx_freqlocked Output Asynchronous signal A high level--the receiver CDR is in lock-to-data mode. Channel A low level--the receiver CDR is in lock-to-reference mode. Receiver CDR lock-to-data mode control signal. rx_locktodata Stratix IV Device Handbook Volume 2: Transceivers Input Asynchronous signal When asserted high--the receiver CDR is forced to lock-to-data mode. When de-asserted low--the receiver CDR lock mode depends on the rx_locktorefclk signal level. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-219 Table 1-74. Stratix IV GX and GT ALTGX Megafunction Ports: Receiver Ports (Part 7 of 7) Input/ Output Port Name Clock Domain Description Scope Receiver CDR lock-to-reference mode control signal. Input rx_locktorefclk Asynchronous signal The rx_locktorefclk signal, along with the rx_locktodata signal, controls whether the receiver CDR is in automatic (0/0), lock-to-reference (0/1), or lock-to-data (1/x) mode. Channel Signal threshold detect indicator. Output rx_signaldetect Asynchronous signal Available in Basic functional mode when the 8B/10B Encoder/Decoder is selected. Available in PCIe mode. A high level--that the signal present at the receiver input buffer is above the programmed signal detection threshold value. Channel If the electrical idle inference block is disabled in PCIe mode, the rx_signaldetect signal is inverted and driven on the pipeelecidle port. Serial loopback control port. Input rx_seriallpbken Asynchronous signal 0-normal datapath, no serial loopback 1-serial loopback Channel Table 1-75 lists the ALTGX megafunction CMU ports. Table 1-75. Stratix IV GX and GT ALTGX Megafunction Ports: CMU (Part 1 of 2) Port Name Input/ Output Clock Domain Input Clock signal pll_inclk Description Scope Input reference clock for the CMU phase-locked loop. Transceiver block CMU PLL lock indicator. pll_locked September 2012 Output Altera Corporation Asynchronous signal A high level--the CMU PLL is locked to the input reference clock. A low level--the CMU PLL is not locked to the input reference clock. Transceiver block Stratix IV Device Handbook Volume 2: Transceivers 1-220 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-75. Stratix IV GX and GT ALTGX Megafunction Ports: CMU (Part 2 of 2) Port Name Input/ Output Clock Domain Asynchronous signal. pll_powerdown Input For minimum pulse width requirements, refer the device DC and Switching Characteristics chapter. Description Scope CMU PLL power down. Asserted high--the CMU PLL is powered down. De-asserted low--the CMU PLL is active and locks to the input reference clock. Transceiver block Note: Asserting the pll_powerdown signal does not power down the REFCLK buffers. FPGA fabric-transceiver interface clock. coreclkout Output Generated by the CMU0 clock divider in the transceiver block in x4 bonded channel configurations. Generated by the CMU0 clock divider in the master transceiver block in x8 bonded channel configurations. Not available in non-bonded channel configurations. Use to clock the write port of the transmitter phase compensation FIFOs in all bonded channels and to clock parallel data tx_datain from the FPGA fabric into the transmitter phase compensation FIFO of all bonded channels. Use to clock the read port of the receiver phase compensation FIFOs in all bonded channels with rate match FIFO enabled and to clock parallel data rx_dataout from the receiver phase compensation FIFOs of all bonded channels (with rate match FIFO enabled) into the FPGA fabric. Clock signal Transceiver block Table 1-76 lists the ALTGX megafunction dynamic reconfiguration ports. Table 1-76. Stratix IV GX and GT ALTGX Megafunction Ports: Dynamic Reconfiguration (Part 1 of 2) Port Name Input/ Output Clock Domain Description Scope Dynamic reconfiguration clock. reconfig_clk Stratix IV Device Handbook Volume 2: Transceivers Input Clock signal Also used for offset cancellation in all modes except PCIe mode. If configured in Transmitter only mode--the frequency range is 2.5 MHz to 50 MHz. If configured in Receiver only or Receiver and Transceiver mode--the frequency range of this clock is 37.5 MHz to 50 MHz. September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-221 Table 1-76. Stratix IV GX and GT ALTGX Megafunction Ports: Dynamic Reconfiguration (Part 2 of 2) Port Name reconfig_togxb reconfig_fromgxb Input/ Output Clock Domain Input Asynchronous signal From the dynamic reconfiguration controller. Output Asynchronous signal To the dynamic reconfiguration controller. Description Scope Table 1-77 lists the ALTGX megafunction PCIe interface ports. Table 1-77. Stratix IV GX and GT ALTGX Megafunction Ports: PCIe Interface (Part 1 of 4) Port Name Input/ Output Clock Domain Description Scope PCIe Interface (Available only in PCIe functional Mode) PCIe power state control. Input powerdn Functionally equivalent to the powerdown[1:0] signal defined in the PCIe specification revision 2.0. The width of this signal is 2 bits and is encoded as follows: Asynchronous signal Channel 2'b00: P0--Normal Operation 2'b01: P0s--Low Recovery Time Latency, Low Power State 2'b10: P1--Longer Recovery Time Latency, Lower Power State 2'b11: P2--Lowest Power State Force 8B/10B encoder to encode with a negative running disparity. tx_ forcedispcompliance Input Asynchronous signal Functionally equivalent to the txcompliance signal defined in PCIe specification revision 2.0. Must be asserted high only when transmitting the first byte of the PCIe compliance pattern to force the 8B/10B encode with a negative running disparity as required by the PCIe protocol. Channel Force transmitter buffer to PCIe electrical idle signal levels. tx_forceelecidle Input Asynchronous signal Functionally equivalent to the txelecidle signal defined in the PCIe specification revision 2.0. Available in the Basic mode. Channel PCIe rateswitch control. Input rateswitch September 2012 Altera Corporation Asynchronous signal 1'b0--Gen1 (2.5 Gbps) 1'b1--Gen2 (5 Gbps) Stratix IV Device Handbook Volume 2: Transceivers 1-222 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-77. Stratix IV GX and GT ALTGX Megafunction Ports: PCIe Interface (Part 2 of 4) Port Name Input/ Output Clock Domain Description Scope Transmitter differential output voltage level control. tx_pipemargin Input Functionally equivalent to the txmargin signal defined in the PCIe specification revision 2.0. Available only in PCIe Gen2 configuration. The width of this signal is 3 bits and is decoded as follows: Asynchronous signal 3'b000--Normal Operating Range 3'b001--Full Swing = 800 - 1200 mV 3'b010--TBD 3'b011--TBD 3'b100--If last value, full Swing = 200 to 400 mV 3'b101--If last value, full Swing = 200 to 400 mV 3'b110--If last value, full Swing = 200 to 400 mV 3'b111--If last value, full Swing = 200 to 400 mV Transmitter buffer de-emphasis level control. tx_pipedeemph Input Asynchronous signal Functionally equivalent to the txdeemph signal defined in the PCIe specification revision 2.0. Available only in PCIe Gen2 configuration. 1'b0: -6 dB de-emphasis 1'b1:-3.5 dB de-emphasis PCIe polarity inversion control. pipe8b10binvpolarity Stratix IV Device Handbook Volume 2: Transceivers Input Asynchronous signal Functionally equivalent to the rxpolarity signal defined in the PCIe specification revision 2.0. Available only in PCIe mode. When asserted high--the polarity of every bit of the 10-bit input data to the 8B/10B decoder gets inverted. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists 1-223 Table 1-77. Stratix IV GX and GT ALTGX Megafunction Ports: PCIe Interface (Part 3 of 4) Port Name Input/ Output Clock Domain Description Scope Receiver detect or PCIe loopback control. tx_detectrxloopback Input Functionally equivalent to the txdetectrx/loopback signal defined in the PCIe specification revision 2.0. When asserted high in the P1 power state with the tx_forceelecidle signal asserted--the transmitter buffer begins the receiver detection operation. After the receiver detect completion is indicated on the pipephydonestatus port, this signal must be de-asserted. Asynchronous signal Channel When asserted high in the P0 power state with the tx_forceelecidle signal de-asserted-- the transceiver datapath gets dynamically configured to support parallel loopback, as described in "PCIe Reverse Parallel Loopback" on page 1-194. PCIe receiver status port. Output pipestatus Functionally equivalent to the rxstatus[2:0] signal defined in the PCIe specification revision 2.0. Synchronized with tx_clock. The width of this signal is 3 bits per channel. The encoding of receiver status on the pipestatus port is as follows: N/A 000--Received data OK 001--1 skip added 010--1 skip removed 011--Receiver detected 100--8B/10B decoder error 101--Elastic buffer overflow 110--Elastic buffer underflow 111--Received disparity error Channel PHY function completion indicator. pipephydonestatus Output Functionally equivalent to the phystatus signal defined in the PCIe specification revision 2.0. Assert this signal high for one parallel clock cycle to communicate completion of several PHY functions, such as power state transition, receiver detection, and signaling rate change between Gen1 (2.5 Gbps) and Gen2 (5 Gbps). N/A September 2012 Altera Corporation Channel Synchronized with tx_clkout. Stratix IV Device Handbook Volume 2: Transceivers 1-224 Chapter 1: Transceiver Architecture in Stratix IV Devices Transceiver Port Lists Table 1-77. Stratix IV GX and GT ALTGX Megafunction Ports: PCIe Interface (Part 4 of 4) Port Name rx_pipedatavalid Input/ Output Output Clock Domain N/A Description Valid data and control on the rx_dataout and rx_ctrldetect ports indicator. Functionally equivalent to the rxvalid signal defined in the PCIe specification revision 2.0. Scope Channel Electrical idle detected or inferred at the receiver indicator. pipeelecidle Output Functionally equivalent to the rxelecidle signal defined in the PCIe specification revision 2.0. If the electrical idle inference block is enabled-- it drives this signal high when it infers an electrical idle condition, as described in "Electrical Idle Inference" on page 1-138. Otherwise, it drives this signal low. Asynchronous signal Channel If the electrical idle inference block is disabled-- the rx_signaldetect signal from the signal detect circuitry in the receiver buffer is inverted and driven on this port. Table 1-78 lists the ALTGX megafunction reset and power down ports. Table 1-78. Stratix IV GX and GT ALTGX Megafunction Ports: Reset and Power Down (Part 1 of 2) Port Name Input/ Output Clock Domain Description Scope Asynchronous signal. gxb_powerdown rx_digitalreset Stratix IV Device Handbook Volume 2: Transceivers Input Input For minimum pulse width requirements, refer the device DC and Switching Characteristics chapter. Asynchronous signal. Minimum pulse width is two parallel clock cycles. Transceiver block power down. When asserted high--all digital and analog circuitry within the PCS, PMA, CMU channels, and the CCU of the transceiver block, is powered down. Transceiver block Asserting the gxb_powerdown signal does not power down the REFCLK buffers. Receiver PCS reset. When asserted high--the receiver PCS blocks are reset. Refer to Reset Control and Power Down. Channel September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Reference Information 1-225 Table 1-78. Stratix IV GX and GT ALTGX Megafunction Ports: Reset and Power Down (Part 2 of 2) Port Name rx_analogreset tx_digitalreset Input/ Output Clock Domain Input Asynchronous signal. Minimum pulse width is two parallel clock cycles. Input Asynchronous signal. Minimum pulse width is two parallel clock cycles. Description Scope Receiver PMA reset. When asserted high--analog circuitry within the receiver PMA gets reset. Refer to Reset Control and Power Down. Channel Transmitter PCS reset. When asserted high, the transmitter PCS blocks are reset. Refer to Reset Control and Power Down. Channel Table 1-79 lists the ALTGX megafunction calibration block ports. Table 1-79. Stratix IV GX and GT ALTGX Megafunction Ports: Calibration Block Input/ Output Clock Domain cal_blk_clk Input Clock signal Clock for transceiver calibration blocks. Device cal_blk_powerdown Input Clock signal Calibration block power down control. Device Port Name Description Scope Reference Information Use the links listed in Table 1-80 for more information about some useful reference terms used in this chapter. Table 1-80. Reference Information (Part 1 of 3) September 2012 Terms Used in this Chapter Useful Reference Points (OIF) CEI PHY Interface Mode page 1-181 8B/10B Decoder page 1-89 8B/10B Encoder page 1-23 AEQ page 1-50 Auxiliary Transmit (ATX) PLL Block page 1-195 Basic (PMA Direct) Functional Mode page 1-187 Basic Functional Mode page 1-111 Built-In Self Test Modes page 1-207 Byte Ordering Block page 1-95 Byte Serializer page 1-94 Calibration Blocks page 1-201 Clock and Data Recovery Unit (CDR) page 1-53 CMU Channel Architecture page 1-100 CMU0 PLL page 1-102 CMU1 PLL page 1-102 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-226 Chapter 1: Transceiver Architecture in Stratix IV Devices Reference Information Table 1-80. Reference Information (Part 2 of 3) Stratix IV Device Handbook Volume 2: Transceivers Terms Used in this Chapter Useful Reference Points CPRI and OBSAI page 1-124 Deserializer page 1-58 Deskew FIFO page 1-75 Deterministic Latency Mode page 1-122 EyeQ page 1-51 GIGE Mode page 1-164 Lock-to-Data (LTD) page 1-54 Lock-to-Reference (LTR) page 1-53 Low Latency PCS Datapath page 1-112 Offset Cancellation in the Receiver Buffer and Receiver CDR page 1-56 Parallel loopback page 1-191 PCIe Clock Switch Circuitry page 1-55 PCIe Mode page 1-127 PCIe Reverse Parallel Loopback page 1-194 Programmable Common Mode Voltage page 1-40 Programmable Equalization and DC Gain page 1-49 Programmable Pre-Emphasis page 1-36 Programmable Differential On-Chip Termination page 1-42 Rate Match (Clock Rate Compensation) FIFO page 1-77 Receiver Bit Reversal page 1-73 Receiver Input Buffer page 1-40 Receiver Phase Compensation FIFO page 1-98 Receiver Polarity Inversion page 1-71 Reverse Serial Loopback page 1-193 Reverse serial Pre-CDR Loopback page 1-194 SATA and SAS options page 1-121 SDI Mode page 1-178 Serial Loopback page 1-190 Serial RapidIO Mode page 1-182 Signal Threshold Detection Circuitry page 1-49 SONET/SDH Mode page 1-172 Transceiver Block Architecture page 1-16 Transceiver Channel Locations page 1-4 Transceiver Port Lists page 1-210 Transmitter Bit Reversal page 1-31 Transmitter Local Clock Divider Block page 1-39 Transmitter Output Buffer page 1-34 Transmitter Polarity Inversion page 1-29 TX Phase Compensation FIFO page 1-19 September 2012 Altera Corporation Chapter 1: Transceiver Architecture in Stratix IV Devices Reference Information 1-227 Table 1-80. Reference Information (Part 3 of 3) Terms Used in this Chapter Useful Reference Points Word Aligner page 1-59 XAUI Mode page 1-153 Document Revision History Table 1-81 lists the revision history for this chapter. Table 1-81. Document Revision History (Part 1 of 2) Date Version September 2012 December 2011 June 2011 February 2011 March 2010 September 2012 4.5 4.4 4.3 Changes Updated Figure 1-1, Figure 1-4, Figure 1-14, Figure 1-87, and Figure 1-89 to close FB #65098. Updated Figure 1-45 to match Stratix V CDR Unit (to close FB #65098). Updated Note 1 of Figure 1-101 to close FB #31792. Updated Table 1-73 to close FB #53976. Updated the "Reverse Serial Loopback" section to close FB #44323. Updated Figure 1-9, Figure 1-10, Figure 1-112, and Figure 1-172. Updated Table 1-5 and Table 1-17. Updated the "Compliance Pattern Transmission Support" and "PCIe Cold Reset Requirements" sections. Added military speed grade to Table 1-68. Applied new template. Updated the "Overview", "Transceiver Block Architecture", "DC-Coupled Links", "Link Coupling for Stratix IV GX and GT Devices", "Configuring CMU Channels for Clock Generation", "Configuring CMU Channels as Transceiver Channels", "Offset Cancellation in the Receiver Buffer and Receiver CDR", "Modes of Operation of the AEQ", "Word-Alignment-Based Byte Ordering", "SATA and SAS Options", "GIGE Mode". "Loopback Modes", "Reverse Serial Loopback", "Input Signals to the Calibration Block", and "PCI Express Electrical Gold Test with Compliance Base Board (CBB)" sections. Updated Figure 1-1, Figure 1-2, Figure 1-3, Figure 1-4, Figure 1-5, Figure 1-6, Figure 1-7, Figure 1-8, Figure 1-9, Figure 1-10, Figure 1-11, Figure 1-56, Figure 1-57, Figure 1-64, Figure 1-76, Figure 1-83, Figure 1-84,Figure 1-95, Figure 1-69, Figure 1-97, Figure 1-99, Figure 1-100, Figure 1-101, Figure 1-105, Figure 1-111, Figure 1-112, and Figure 1-157. Updated Table 1-2, Table 1-3, Table 1-4, Table 1-5, Table 1-13, Table 1-14, Table 1-16, Table 1-17, Table 1-18, Table 1-19, Table 1-21, Table 1-23, Table 1-24, Table 1-41, Table 1-43, Table 1-45, Table 1-68, Table 1-70, Table 1-74, and Table 1-77. Updated chapter title. Applied new template. Minor text edits. Added two references to the beginning of the chapter. Updated the "Configuring CMU Channels for Clock Generation" section. Updated Figure 1-101. Minor text edits. 4.2 4.1 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 1-228 Chapter 1: Transceiver Architecture in Stratix IV Devices Reference Information Table 1-81. Document Revision History (Part 2 of 2) Date Version November 2009 June 2009 March 2009 4.0 Changes Added "Adaptive Equalization (AEQ)", "EyeQ", "SATA and SAS Options", "Deterministic Latency Mode", "CPRI and OBSAI", and "Reference Information" sections. Added Figure 1-91, Figure 1-93, Figure 1-95, and Figure 1-97. Added Stratix IV GT device information. Updated Figures. Updated Tables. Re-organized chapter information. Minor text edits. Updated the "Introduction", "Auxiliary Transmit (ATX) PLL Block", "Rate Match (Clock Rate Compensation) FIFO", "Transmitter Buffer Electrical Idle", "PCIe Gen2 (5 Gbps) Support", "Reverse Serial Loopback", and "Reverse Serial Pre-CDR Loopback" sections. Added new "PCI Express Electrical Gold Test with Compliance Base Board (CBB)", "Recommendation When Using the Electrical Idle Inference Block". and "Rate Match FIFO in Serial RapidIO Mode" sections. Added new Figure 1-165. Updated Table 1-2, Table 1-17, Table 1-32, Table 1-34, and Table 1-52. Updated Figure 1-7, and Figure 1-165 through Figure 1-168. Minor text edits. Reorganized sections. Added the section "Link Coupling". Updated the section "DC-Coupled Links". 3.1 3.0 November 2008 2.0 Added Offset Cancellation in the Receiver Buffer and Receiver CDR to the Receiver Channel Datapath section May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation 2. Transceiver Clocking in Stratix IV Devices September 2012 SIV52002-3.4 SIV52002-3.4 This chapter provides detailed information about the Stratix(R) IV transceiver clocking architecture. For this chapter, the term "Stratix IV devices" includes both Stratix IV GX and GT devices. Similarly, the term "Stratix IV transceivers" includes both Stratix IV GX and GT transceivers. The clocking architecture chapter is divided into three main sections: "Input Reference Clocking" on page 2-2--describes how the reference clock is provided to the clock multiplier unit (CMU)/auxiliary transmit phase-locked loop (ATX PLL) to generate the clocks required for transceiver operation. "Transceiver Channel Datapath Clocking" on page 2-20--describes the clocking architecture internal to the transceiver block. "FPGA Fabric-Transceiver Interface Clocking" on page 2-51--describes the clocking options available when interfacing the transceiver with the FPGA fabric. Other sections in this chapter include: "FPGA Fabric PLLs-Transceiver PLLs Cascading" on page 2-9 "Using the CMU/ATX PLL for Clocking User Logic in the FPGA Fabric" on page 2-71 "Configuration Examples" on page 2-72 Figure 2-1 shows an overview of the clocking architecture. Figure 2-1. Clocking Architecture Overview Transceivers Input Reference Clocks CMU/ATX PLL or CDR FPGA Fabric FPGA Fabric-Transciever Interface Clocks Transceiver Channels Transceiver Channel Datapath Clocks (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 2: Transceivers September 2012 Feedback Subscribe 2-2 Chapter 2: Transceiver Clocking in Stratix IV Devices Glossary of Terms Glossary of Terms Table 2-1 lists the terms used in the chapter. Table 2-1. Glossary of Terms Used in this chapter Convention Description ATX PLL Auxiliary transmit PLL block. For more information, refer to the "Auxiliary Transmit (ATX) PLL Block" section in the Transceiver Architecture in Stratix IV Devices chapter. CDR Clock data recovery block. For more information, refer to the "Clock and Data Recovery Unit" section in the Transceiver Architecture in Stratix IV Devices chapter. CMU Clock multiplier unit. For more information, refer to "CMU Channel Architecture" section in the Transceiver Architecture in Stratix IV Devices chapter. ITB lines The Inter-Transceiver block (ITB) clock lines provide an input reference clock path from the refclk pins of one transceiver block CMU PLLs and receiver CDRs of other transceiver blocks. They also provide input reference clock to ATX PLLs. For more information, refer to "InterTransceiver Block (ITB) Clock Lines" on page 2-8. Input Reference Clocking Each transceiver block has: Two clock multiplier unit channels--the CMU0_Channel and CMU1_Channel You can configure each as either a CMU to generate transceiver clocks or as a PMA-Only channel. f For more information, refer to the "CMU Channel Architecture" section in the Transceiver Architecture in Stratix IV Devices chapter. Four regular channels When the CMU channel is configured as a CMU, the CMU PLL synthesizes the input reference clock to generate the high-speed serial transceiver clock. When the CMU channel is configured as a Receiver Only or Receiver and Transmitter channel, the CMU PLL acts as a CDR and uses the input reference clock as a training clock when it is in lock-to-reference (LTR) mode. Each of the four regular channels also has a receiver CDR that uses the input reference clock as a training clock when it is in LTR mode. Each Stratix IV device also has ATX PLLs that you can use in addition to the CMU PLLs to generate the high-speed serial transceiver clock. The ATX PLLs also need an input reference clock for operation. 6G ATX PLLs are available in both Stratix IV GX and Stratix IV GT devices. 10G ATX PLLs are available only in Stratix IV GT devices. f For more information, refer to the "Auxiliary Transmit (ATX) PLL Block" and the "Transmitter Channel Datapath" sections in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking 2-3 Input Reference Clock Source Receiver clock data recoveries (CDRs), CMU PLLs (when the CMU channel is configured as a CMU) and ATX PLLs can derive the input reference clock from one of the sources listed in Table 2-2. Table 2-2. Input Reference Clock Source Jitter Performance Index Clock Source CMU PLL 6G ATX PLL 10G ATX PLL CDR 1 refclk0 and refclk1 pins of the same transceiver block Yes No (1) No (1) Yes 1 2 refclk0 and refclk1 pins of other transceiver blocks on the same side of the device using the ITB clock lines (2) Yes Yes Yes Yes 2 (3) 3 Clock output from the left and right PLLs in the FPGA fabric with voltage controlled oscillator (VCO) bypass mode (5), (6) Yes Yes No Yes 3 4 Clock output from the left and right PLLs in the FPGA fabric Yes Yes No Yes 4 5 Dedicated CLK input pins on the FPGA global clock network Yes Yes No Yes 4 (4) Notes to Table 2-2: (1) ATX PLLs do not have dedicated refclk pins. (2) For more information, refer to "Inter-Transceiver Block (ITB) Clock Lines" on page 2-8. (3) For better jitter performance, Altera strongly recommends using the refclk0 and refclk1 pins of the transceiver block located immediately below the ATX PLL. (4) Lowest number indicates best jitter performance. (5) For more information, refer to "Configuration Examples" on page 2-72. (6) When in VCO bypass mode, you can only divide the reference clock by the N integer. For more information, refer to "Left and Right, Left, or Right PLL in VCO Bypass Mode" on page 2-17. When a CMU channel is configured as a channel, its CMU PLL acts as a receiver CDR and can derive the input reference clock sources 2 through 5 listed in the Table 2-2. You can also use the refclk pin of the other CMU channel within the transceiver block as a clock source as long as the other CMU channel is not configured as a Receiver only or Receiver and Transmitter channel. For example, the CMU0 PLL can derive its input reference clock from the refclk1 pin if the CMU1 channel is not configured as a Receiver only or Receiver and Transmitter channel. 1 When a CMU channel is configured as a channel, its refclk pin is used to receive serial input data. As a result, the refclk pin is not available to provide the input reference clock. Table 2-3 lists the input reference clock frequencies allowed for the 10G ATX PLL. Table 2-3. Input Reference Clock Frequencies for the 10G ATX PLL Clock Data Rate (Gbps) 9.9 to 11.3 September 2012 Altera Corporation Allowed Divider Values Reference Clock Frequency (MHz) M = 16, N = 1 281.25 to 322 M = 16, N = 2 562.5 to 706.25 Stratix IV Device Handbook Volume 2: Transceivers 2-4 Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking Figure 2-2 shows the input reference clock sources for CMU PLLs and receiver CDRs within a transceiver block. One global clock line is available for each CMU PLL and receiver CDR in a transceiver block. This allows each CMU PLL and receiver CDR to derive its input reference clock from a separate FPGA CLK input pin. Figure 2-2. Input Reference Clock Sources in a Transceiver Block refclk0 2 Transceiver Block Channel 3 refclk1 2 ITB Clock Lines 6 CDR 6 CDR Global Clock Line PLL Cascade Clock Channel 2 ITB Clock Lines Global Clock Line PLL Cascade Clock CMU1 Channel ITB Clock Lines Global Clock Line PLL Cascade Clock 6 CMU1 PLL CMU0 Channel ITB Clock Lines Global Clock Line PLL Cascade Clock 6 CMU0 PLL 6 CDR 6 CDR Channel 1 ITB Clock Lines Global Clock Line PLL Cascade Clock Channel 0 ITB Clock Lines Global Clock Line PLL Cascade Clock Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking 2-5 Figure 2-3 shows the input reference clock sources for CMU PLLs, ATX PLLs, and receiver CDRs in four transceiver blocks on the right side of the EP4SGX530F45 device. In this figure, the input reference clock sources for four transceiver blocks are located only on the right side of the device but the EP4SGX530NF45 device has similar input reference clock resources available for the four transceiver blocks located on the left side of the device as well. Figure 2-3 also shows the ITB clock lines on the right side of the device. The number of ITB clock lines available in any Stratix IV GX device is equal to the number of refclk pins available in that device. Figure 2-3. Input Reference Clock Sources Across Transceiver Blocks ITB[7:0] Transceiver Block GXBR3 refclk0 2 8 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 Transceiver Block GXBR2 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 To FPGA Fabric ITB Clock Lines 8 Global Clock Line PLL Cascade Clock ATX PLL R1 (6G) Transceiver Block GXBR1 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 ITB Clock Lines 8 Global Clock Line PLL Cascade Clock ATX PLL R0 (6G) Transceiver Block GXBR0 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-6 Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking Figure 2-4 shows the input reference clock sources for CMU PLLs, ATX PLLs, and receiver CDRs in four transceiver blocks on the right side of the EP4S100G5F45 device. In this figure, the input reference clock sources for four transceiver blocks are located only on the right side of the EP4S100G5F45 device but the device has similar input reference clock resources available for the four transceiver blocks located on the left side of the device as well. Figure 2-4 also shows the ITB clock lines on the right side of the EP4S100G5F45 device. The number of ITB clock lines available in any Stratix IV GT device is equal to the number of refclk pins available in that device. Figure 2-4. Input Reference Clock Sources Across Transceiver Blocks for Stratix IV GT Devices ITB[7:0] 8 Transceiver Block GXBR3 refclk0 2 6 Two CMU PLLs and Four RX CDRs 8 ATX PLL R2 (10G) 6 Global Clock Line PLL Cascade Clock refclk1 2 ITB Clock Lines Transceiver Block GXBR2 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 To FPGA Fabric ITB Clock Lines 8 Global Clock Line PLL Cascade Clock ATX PLL R1 (6G) Transceiver Block GXBR1 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 ITB Clock Lines 8 Global Clock Line PLL Cascade Clock ATX PLL R0 (6G) Transceiver Block GXBR0 refclk0 2 6 Global Clock Line PLL Cascade Clock refclk1 6 Two CMU PLLs and Four RX CDRs 2 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking 2-7 refclk0 and refclk1 Pins Each transceiver block has two dedicated refclk pins that you can use to drive the CMU PLL, receiver CDR, or both, input reference clocks. Each of the two CMU PLLs and four receiver CDRs within a transceiver block can derive its input reference clock from either the refclk0 or refclk1 pin. 1 The refclk pins provide the cleanest input reference clock path to the CMU/ATX PLLs when compared with other input reference clock sources. Altera recommends using the refclk pins to drive the CMU PLL input reference clock for improved transmitter output jitter performance. Table 2-4 lists the electrical specifications for the input reference clock signal driven on the refclk pins. f For specifications regarding the input frequency supported by the refclk pins, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Table 2-4. Electrical Specifications for the Input Reference Clock Protocol GIGE XAUI RapidIO(R) I/O Standard 1.2-V PCML, 1.4 PCML 1.4-V PCML 1.5-V PCML Serial SONET/SDH 2.5-V PCML SDI Differential LVPECL (OIF) CEI PHY Interface LVDS Basic 1.2-V PCML, 1.4 PCML 1.4-V PCML 1.5-V PCML 2.5-V PCML Differential LVPECL LVDS HCSL (1) PCI Express(R) (PCIe) Coupling Termination AC On-chip (2) AC On-chip (2) DC Off-chip (3) Notes to Table 2-4: (1) In PCIe mode, you have the option of selecting the HCSL standard for the reference clock if compliance to the PCIe protocol is required. You can select this I/O standard option only if you configured the transceiver in PCIe functional mode. For more information, refer to Figure 2-5 on page 2-8. (2) Termination values supported are the same as the Receiver pin differential on-chip termination resistors specified in the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For an example termination scheme, refer to Figure 2-5 on page 2-8. 1 If you select the HCSL I/O standard for the PCIe reference clock, add the following assignment to your project quartus settings file (.qsf): set_instance_assignment -name INPUT_TERMINATION OFF -to <refclk_pin_name> September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-8 Chapter 2: Transceiver Clocking in Stratix IV Devices Input Reference Clocking Figure 2-5 shows an example termination scheme for a reference clock signal when configured as HCSL. Figure 2-5. Termination Scheme for a Reference Clock Signal When Configured as HCSL (Note 1) PCI Express (HCSL) refclk Source Stratix IV Rs (2) refclk + Rs (2) refclk - Rp = 50 Rp = 50 Notes to Figure 2-5: (1) No biasing is required if the reference clock signals are generated from a clock source that conforms to the PCIe specification. (2) Select resistor values as recommended by the PCIe clock source vendor. Inter-Transceiver Block (ITB) Clock Lines The refclk0 and refclk1 pins of other transceiver blocks using the ITB clock lines provide an input reference clock path from the refclk pins of one transceiver block to the CMU PLLs and receiver CDRs of the other transceiver blocks. In designs that have channels located in different transceiver blocks, the ITB clock lines eliminate the need to connect the on-board reference clock crystal oscillator to the refclk pin of each transceiver block. The ITB clock lines also drive the clock signal on the refclk pins to the clock logic in the FPGA fabric. The ITB clock lines also provide an input reference clock path from the refclk pins of any transceiver block to the ATX PLLs located on the same side of the device. Each refclk pin drives one ITB clock line for a total of up to eight ITB clock lines on each of the right and left sides of the device, as shown in Figure 2-3 on page 2-5. 1 The ITB clock lines provide input reference clock paths from the refclk pins of one transceiver block to the CMU PLLs and receiver CDRs of other transceiver blocks located on the same side of the device. Dedicated CLK Input Pins on the FPGA Global Clock Network Stratix IV devices provide up to eight differential clock input pins located in non-transceiver I/O banks that you can use to provide up to eight input reference clocks to the transceiver blocks. The Quartus(R) II software automatically chooses the global clock network to route the input reference clock signal from the CLK pins to the transceiver blocks. f For more information, refer to the "Dedicated Clock Input Pins" section in the Clock Networks and PLLs in Stratix IV Devices chapter. One global clock resource is available for each CMU PLL, 6G ATX PLL, and receiver CDR. This allows each CMU PLL, 6G ATX PLL, and receiver CDR to derive its input reference clock from a separate FPGA CLK input pin. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-9 Clock Output from Left and Right PLLs in the FPGA Fabric You can use the synthesized clock output from one of the left or right PLLs to provide the input reference clock to the CMU PLLs, 6G ATX PLLs, and receiver CDRs. Stratix IV devices provide a dedicated clock path from the left PLLs (PLL_L1, PLL_L2, PLL_L3, and PLL_L4) in the FPGA fabric to the PLL cascade network located on the left side of the device. Stratix IV devices also provide a dedicated clock path from the right PLLs (PLL_R1, PLL_R2, PLL_R3, and PLL_R4) in the FPGA fabric to the PLL cascade network located on the right side of the device. The additional clock multiplication factors available in the left and right PLLs allow more options for on-board crystal oscillator frequencies. FPGA Fabric PLLs-Transceiver PLLs Cascading The CMU PLL synthesizes the input reference clock to generate the high-speed serial clock used in the transmitter PMA. The receiver CDR synthesizes the input reference clock in lock-to-reference (LTR) mode to generate the high-speed serial clock. This high-speed serial clock output from the CMU PLL and the receiver CDR runs at a frequency that is half the configured data rate. The CMU PLLs and receiver CDRs only support multiplication factors (M) of 2, 4, 5, 8, 10, 16, 20, and 25. If you use an on-board crystal oscillator to provide the input reference clock through the dedicated refclk pins or ITB lines, the allowed crystal frequencies are limited by the CMU PLL and the receiver CDR multiplication factors. The input reference clock frequencies are also limited by the allowed phase frequency detector (PFD) frequency range. Example 1: Channel Configuration with a 4 Gbps Data Rate Consider a channel configured for a 4 Gbps data rate. The high-speed serial clock output from the CMU PLL and the receiver CDR must run at 2 Gbps. Table 2-5 lists the allowed input reference clock frequencies for Example 1. Table 2-5. Allowed Input Reference Clock Frequency for Example 1 Multiplication Factor (M) On-Board Crystal Reference Clock Frequency (MHz) Allowed With /N = 1 With /N = 2 2 1000 2000 No. Violates the PFD frequency limit. 4 500 1000 No. Violates the PFD frequency limit. 5 400 800 Yes but only for /N = 1. 8 250 500 Yes 10 200 400 Yes 16 125 250 Yes 20 100 200 Yes 25 80 160 Yes For a 4 Gbps data rate, the Quartus II software only allows an input reference clock frequency of 80, 100, 125, 160, 200, 250, 400, and 500 MHz. To overcome this limitation, Stratix IV devices allow the synthesized clock output from the left and right PLLs in the FPGA fabric to drive the CMU PLL and receiver CDR input reference clock. The additional clock multiplication factors available in the left and right PLLs allow more options for on-board crystal oscillator frequencies. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-10 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading Dedicated Left and Right PLL Cascade Network Stratix IV devices have a dedicated PLL cascade network on the left and right side of the device that connects to the input reference clock selection multiplexer of the CMU PLLs, 6G ATX PLLs, and receiver CDRs on the left and right side of the device, respectively. The dedicated PLL cascade networks are segmented by bidirectional tri-state buffers located along the clock line. Segmentation of the dedicated PLL cascade network allows two or more left and right PLLs to drive the cascade clock line simultaneously. Because the number of left and right PLLs and transceiver blocks vary from device to device, the capability of cascading a left and right PLL to the CMU PLLs, 6G ATX PLLs, and receiver CDRs also varies from device to device. The following sections describe the Stratix IV GX and GT FPGA fabric-Transceiver PLLs cascading for the various device packages. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-11 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 780-Pin Package Stratix IV GX devices in 780-pin packages do not support FPGA fabric PLLs-transceiver PLLs cascading. FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1152-Pin Package Figure 2-6 shows the FPGA fabric PLLs-Transceiver PLLs cascading options allowed in the EP4SGX110FF35 device (red), the EP4SGX230FF35, EP4SGX290FF35 and EP4SGX360FF35 devices (blue), and the EP4SGX530HH35 device (black). Figure 2-6. FPGA Fabric PLLs-Transceiver PLLs Cascading Options Allowed for 1152-Pin Package Devices PLL Cascade Network PLL Cascade Network EP4SGX530HH35 EP4SGX230FF35 EP4SGX290FF35 EP4SGX360FF35 Transceiver Block GXBL1 Channel 3 Channel 2 Channel 0 CDR CDR Channel 1 Channel 0 CDR CDR CDR CDR September 2012 Channel 1 Channel 0 Transceiver Block GXBR0 CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR Altera Corporation Channel 2 ATX PLL R0 (6G) CDR CDR Channel 3 CMU1 PLL CMU0 PLL Transceiver Block GXBL0 Channel 2 EP4SGX110FF35 CMU0 PLL ATX PLL L0 (6G) Channel 3 PLL_R2 CDR CMU1 PLL Channel 1 Transceiver Block GXBR1 PLL_L2 CDR CDR Channel 3 Channel 2 Channel 1 Channel 0 Stratix IV Device Handbook Volume 2: Transceivers 2-12 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1517-Pin Package Figure 2-7 shows the FPGA fabric PLLs-Transceiver PLLs cascading options allowed in the EP4SGX180HF40, EP4SGX230KF40, EP4SGX290KF40, EP4SGX360KF40, and EP4SGX530KF40 devices. 1 For Stratix IV GT devices, FPGA fabric PLLs-Transceiver PLLs cascading is not supported for the 10G ATX PLLs. For the EP4S40G2KF40, EP4S40G5KF40, EP4S100G2KF40, and EP4S100G5KF40 devices, FPGA fabric PLLs-Transceiver PLLs cascading for the 6G ATX PLLs and CMU PLLs is the same as the Stratix IV GX devices in the 1517-pin package. Figure 2-7. FPGA Fabric PLLs-Transceiver PLLs Cascading Options Allowed in the 1517-Pin Package Devices PLL Cascade Network PLL Cascade Network Transceiver Block GXBR2 Transceiver Block GXBL2 Channel 3 Channel 2 Channel 1 Channel 0 CDR CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR ATX PLL L1 (6G) (1) Channel 3 Channel 2 Channel 1 Channel 0 PLL_L3 CDR EP4SGX180HF40 EP4SGX230KF40 EP4SGX290KF40 EP4SGX360KF40 EP4SGX530KF40 Channel 1 Channel 0 Channel 0 ATX PLL R1 (6G) (1) PLL_R3 CDR CDR CDR CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR Transceiver Block GXBL0 Channel 2 Channel 1 Transceiver Block GXBR1 CMU1 PLL ATX PLL L0 (6G) Channel 3 Channel 2 PLL_R2 PLL_L2 Transceiver Block GXBL1 Channel 3 Channel 3 Channel 2 Channel 1 Channel 0 ATX PLL R0 (6G) Transceiver Block GXBR0 CDR CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR Channel 3 Channel 2 Channel 1 Channel 0 Note to Figure 2-7: (1) ATX PLL L1 and ATX PLL R1 are not present in the EP4SGX230KF40 device. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-13 FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1932-Pin Package Figure 2-8 shows the FPGA fabric PLLs-Transceiver PLLs cascading options allowed in the EP4SGX530NF45, EP4SGX360NF45, and EP4SGX290NF45 devices. 1 For Stratix IV GT devices, FPGA fabric PLLs-Transceiver PLLs cascading is not supported for the 10G ATX PLLs. For the EP4S100G3NF45, EP4S100G4N45, and EP4S100G5NF45 devices, FPGA fabric PLLs-Transceiver PLLs cascading for the 6G ATX PLLs and CMU PLLs is the same as the Stratix IV GX devices in the 1932-pin package. Figure 2-8. FPGA Fabric PLLs-Transceiver PLLs Cascading Options Allowed in the 1932-Pin Package Device Transceiver Block GXBL3 Channel 3 Channel 2 Channel 1 Channel 0 CDR PLL Cascade Network PLL Cascade Network PLL_R1 PLL_L1 Channel 2 CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR Channel 0 CDR CDR PLL_R3 Channel 1 Channel 0 September 2012 CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR Altera Corporation Channel 3 Channel 2 Channel 1 Channel 0 Channel 3 Channel 2 Channel 1 Channel 0 Transceiver Block GXBR0 CDR CDR Channel 0 ATX PLL R0 (6G) Transceiver Block GXBL0 Channel 2 Channel 1 Transceiver Block GXBR1 ATX PLL L0 (6G) Channel 3 Channel 2 ATX PLL R1 (6G) CDR PLL_L3 Channel 0 CDR EP4SGX530NF45 EP4SGX360NF45 EP4SGX290NF45 Transceiver Block GXBL1 Channel 1 PLL_R2 PLL_L2 ATX PLL L1 (6G) Channel 2 Channel 3 Transceiver Block GXBR2 CDR Channel 1 Channel 3 CDR CDR Transceiver Block GXBL2 Channel 3 Transceiver Block GXBR3 PLL_L4 PLL_R4 Channel 3 Channel 2 CDR Channel 1 CDR Channel 0 Stratix IV Device Handbook Volume 2: Transceivers 2-14 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading FPGA Fabric PLLs-Transceiver PLLs Cascading Rules You can only cascade the left PLLs (PLL_L1, PLL_L2, PLL_L3, and PLL_L4) to the transceiver blocks located on the left side of the device. Similarly, you can only cascade the right PLLs (PLL_R1, PLL_R2, PLL_R3, and PLL_R4) to the transceiver blocks located on the right side of the device. The PLL cascade networks are single clock lines segmented by bidirectional tri-state buffers located along the clock line. Segmentation of the PLL cascade network allows two left and right PLLs to drive the cascade clock line simultaneously and provides the input reference clock to the CMU PLLs and receiver CDRs in different transceiver blocks. When cascading two or more FPGA fabric PLLs to the CMU PLLs and receiver CDRs, there must be no crossover in the cascaded clock paths on the PLL cascade network (Figure 2-9). 1 For better noise rejection, ensure the bandwidth setting of the FPGA fabric PLL (the upstream PLL) is lower than the transceiver PLL (the downstream PLL). Example 2: Design Target--EP4SGX530NF45 Device If your design is targeted for a EP4SGX530NF45 device, it requires providing input reference clocks to the following CMU PLLs and receiver CDRs from two right PLLs in the FPGA fabric: CMU0 PLL in Transceiver Block GXBR1 Receiver CDRs in channel 2 and channel 3 in Transceiver Block GXBR1 Case 1: use PLL_R4 to provide the input reference clock to the receiver CDRs in channel 2 and channel 3 (shown in GREEN) and use PLL_R1 to provide the input reference clock to the CMU0 PLL (shown in BLUE) in transceiver block GXBR1. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-15 Figure 2-9 shows that this FPGA fabric-Transceiver PLL cascading configuration is illegal due to crossover (shown in RED) of the cascade clock paths on the PLL cascade network. Figure 2-9. Illegal FPGA Fabric-Transceiver PLL Cascading Configuration Transceiver Block GXBL3 Channel 3 Channel 2 Channel 1 Channel 0 PLL Cascade Network CDR PLL Cascade Network Transceiver Block GXBR3 CDR PLL_L1 Channel 3 PLL_R1 CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR Channel 2 Channel 1 Channel 0 Transceiver Block GXBR2 Transceiver Block GXBL2 Channel 3 CDR CDR Channel 3 Channel 2 CDR CDR Channel 2 Channel 1 Channel 0 CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR PLL_L2 ATX PLL L1 (6G) PLL_R2 CDR Transceiver Block GXBR1 CDR PLL_L3 Channel 2 CDR Channel 1 X CDR Channel 1 Channel 0 CDR CDR CDR CDR September 2012 Channel 0 ATX PLL R0 (6G) CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR Altera Corporation Channel 1 Transceiver Block GXBR0 CDR CDR Channel 2 CMU1 PLL CMU0 PLL Transceiver Block GXBL0 Channel 2 CDR CMU0 PLL ATX PLL L0 (6G) Channel 3 Channel 3 PLL_R3 CMU1 PLL Channel 0 Channel 0 ATX PLL R1 (6G) EP4SGX530NF45 Transceiver Block GXBL1 Channel 3 Channel 1 PLL_L4 PLL_R4 CDR Channel 3 Channel 2 Channel 1 Channel 0 Stratix IV Device Handbook Volume 2: Transceivers 2-16 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading Case 2: use PLL_R1 to provide the input reference clock to the receiver CDRs in channel 2 and channel 3 (shown in BLUE) and use PLL_R4 to provide the input reference clock to the CMU0 PLL (shown in GREEN) in transceiver block GXBR1. Figure 2-10 shows that this FPGA fabric-Transceiver PLL cascading configuration is legal as there is no crossover of the cascade clock paths on the PLL cascade network. Figure 2-10. Legal FPGA Fabric-Transceiver PLL Cascading Configuration Transceiver Block GXBL3 Channel 3 Channel 2 Channel 1 Channel 0 CDR PLL Cascade Network PLL Cascade Network PLL_R1 PLL_L1 Channel 2 Channel 1 Channel 0 CDR CDR CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR PLL_R2 PLL_L2 Channel 0 CDR PLL_R3 Channel 1 Channel 0 CDR CDR CDR CMU1 PLL CMU0 PLL CMU0 PLL CDR CDR CDR CDR ATX PLL L0 (6G) Channel 2 CDR CMU1 PLL CDR CDR CDR CMU1 PLL CMU1 PLL CMU0 PLL CMU0 PLL CDR Stratix IV Device Handbook Volume 2: Transceivers Channel 3 Channel 2 Channel 1 Channel 0 Channel 3 Channel 2 Channel 1 Channel 0 Transceiver Block GXBR0 CDR CDR Channel 0 ATX PLL R0 (6G) Transceiver Block GXBL0 Channel 3 Channel 1 Transceiver Block GXBR1 PLL_L3 Channel 1 Channel 2 ATX PLL R1 (6G) EP4SGX530NF45 Transceiver Block GXBL1 Channel 2 Channel 3 Transceiver Block GXBR2 CDR ATX PLL L1 (6G) Channel 3 CDR CMU1 PLL Transceiver Block GXBL2 Channel 3 Transceiver Block GXBR2 CDR PLL_L4 PLL_R4 CDR Channel 3 Channel 2 Channel 1 Channel 0 September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-17 Left and Right, Left, or Right PLL in VCO Bypass Mode If all CMU channels on the same side of the device are configured as channels, all refclk pins are used as receiver serial input data pins. All CMU PLLs are also used as receiver CDRs. In such designs, you must use the 6G ATX PLLs to generate the high-speed serial and low-speed parallel transceiver clocks provided that the configured data rate is supported by the 6G ATX PLLs. Additionally, Altera recommends providing the input reference clock to the 6G ATX PLL using the left or right PLL cascade clock line because none of the refclk pins are available. To avoid jitter amplification because of cascading of the left or right PLL to the 6G ATX PLL, you must place the left or right PLL in VCO bypass mode. When in VCO bypass mode, you can only divide the reference clock by the N integer. f For more information about CMU PLLs, refer to "Configuring CMU Channels as Transceiver Channels" in the Transceiver Architecture in Stratix IV Devices chapter. Figure 2-11 shows that in VCO bypass mode, the input reference clock from the dedicated FPGA CLK pins to the inclk port of the left and right, left, or right PLL bypasses the PLL loop and is driven directly on the PLL output clock port. Figure 2-11. Left and Right, Left, or Right PLL in VCO Bypass Mode Left and Right PLL /M Reference clock from the dedicated CLK pin September 2012 Altera Corporation /N Phase Frequency Detector Charge Pump + Loop Filter Voltage Controlled Oscillator C1 Input reference clock to the 6G ATX PLL Stratix IV Device Handbook Volume 2: Transceivers 2-18 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading Figure 2-12 shows 24 channels on the right side of the EP4SGX530NF45 device configured in Basic (PMA Direct) xN mode running at 6.5 Gbps with a 20-bit FPGA fabric-PMA interface width. Because all 24 channels on the right side of the device are configured in Basic (PMA Direct) xN mode, use the right PLL_R1 configured in VCO bypass mode to provide the input reference clock to the 6G ATX PLL. Because the data rate of 6.5 Gbps requires a left and right, left, or right PLL to meet FPGA fabric-Transmitter PMA interface timing, the tx_clkout from one of the 24 channels is phase shifted using PLL_R2. Use the phase-shifted output clock from PLL_R2 to clock the FPGA fabric logic that generates the transmitter parallel data and control signals. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric PLLs-Transceiver PLLs Cascading 2-19 Figure 2-12. Input Reference Clocking Using Left and Right, Left, or Right PLL in VCO Bypass Mode PLL Cascade Clock Line (3) xN_Bottom (1) Transceiver Block GXBR 3 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 FPGA Fabric (Transmitter Data Generation Logic) Transceiver Block GXBR 2 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 Dedicated FPGA CLK Pin PLL_R1 (VCO Bypass Mode) Reference Clock PLL_R2 (Phase Shift 45 to meet interface timing) (2) ATX PLL Block R1 (6G) tx_clkout Transceiver Block GXBR1 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 FPGA Fabric (Transmitter Data Generation Logic) ATX PLL Block R1 (6G) Transceiver Block GXBR0 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 xN_Top (1) Notes to Figure 2-12: (1) For more information, refer to "Transceiver Channel Datapath Clocking" on page 2-20. (2) For more information, refer to AN 580: Achieving Timing Closure in Basic (PMA Direct) Functional Mode. (3) The green line represents the PLL cascade clock line and the blue lines represent 6G ATX PLL R1. For more information about configuring left or right PLLs in VCO bypass mode, refer to "Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode" on page 2-78. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-20 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Transceiver Channel Datapath Clocking This section describes the transmitter channel and receiver channel datapath clocking in various configurations. Datapath clocking varies with physical coding sublayer (PCS) configurations in different functional modes as well as channel bonding options. This section contains: 1 "Transmitter Channel Datapath Clocking" on page 2-20 "Receiver Channel Datapath Clocking" on page 2-39 Clocking described in this section is internal to the transceiver and clock routing is primarily performed by the Quartus II software. f For more information about manually picking and placing CMU and ATX PLLs, refer to AN 578: Manual Placement of CMU PLLs and ATX PLLs in Stratix IV GX and GT Devices. Transmitter Channel Datapath Clocking This section describes the transmitter channel PMA and PCS datapath clocking in non-bonded and bonded channel configurations: Stratix IV Device Handbook Volume 2: Transceivers "Non-Bonded Channel Configurations" on page 2-24 "Bonded Channel Configurations" on page 2-27 "Non-Bonded Basic (PMA Direct) Mode Channel Configurations" on page 2-34 "Bonded Basic (PMA Direct) xN Mode Channel Configurations" on page 2-36 September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-21 Transmitter Channel-to-Channel Skew Optimization for Modes Other than Basic (PMA Direct) Mode High-speed serial clock and low-speed parallel clock skew between channels and unequal latency in the transmitter phase compensation FIFO contribute to transmitter channel-to-channel skew. Transmitter datapath clocking is set up to provide low channel-to-channel skew when compared with non-bonded channel configurations. In bonded channel configurations--the high-speed serial clock and low-speed parallel clock for all bonded channels are generated by the CMU0 clock divider or the ATX clock divider block, resulting in lower channel-to-channel clock skew. The transmitter phase compensation FIFO in all bonded channels (except in Basic [PMA Direct] xN mode) share common pointers and control logic generated in the central control unit (CCU), resulting in equal latency in the transmitter phase compensation FIFO of all bonded channels. The lower transceiver clock skew and equal latency in the transmitter phase compensation FIFOs in all channels provides lower channel-to-channel skew in bonded channel configurations. In non-bonded channel configurations--the high-speed serial clock and low-speed parallel clock in each channel are generated independently by its local clock divider. This results in higher channel-to-channel clock skew. The transmitter phase compensation FIFO in each non-bonded channel (except in Basic [PMA Direct] mode) has its own pointers and control logic that can result in unequal latency in the transmitter phase compensation FIFO of each channel. The higher transceiver clock skew and unequal latency in the transmitter phase compensation FIFO in each channel can result in higher channel-to-channel skew. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-22 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Transmitter Channel Datapath Clocking Resources The Stratix IV transceivers support various non-bonded and bonded transceiver clocking configurations through the dedicated x1, x4, and xN high-speed serial and low-speed parallel clock lines. Figure 2-13 shows the transceiver clock distribution in x1, x4, x8, and xN bonded modes. Figure 2-13. Transceiver Clock Distribution in the Stratix IV GT EP4S100G5F45 and Stratix IV GX EP4SGX530KF40 Devices Transceiver Block GXBR3 x1 CMU1 GXBR3 x4 GXBR3 xN_Bottom Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR3 ATX PLL Block (10G) (1) Transceiver Block GXBR2 x1 CMU1 GXBR2 x4 GXBR2 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR2 ATX PLL Block (6G) Transceiver Block GXBR1 x1 CMU1 GXBR1 x4 GXBR1 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR1 ATX PLL Block (6G) Transceiver Block GXBR0 x1 CMU1 GXBR0 x4 GXBR0 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR0 xN_Top Note to Table 2-14: (1) The 10G ATX PLL block is not available for the EP4SGX530KF40 device. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-23 Non-bonded and bonded configurations use the following: x1 non-bonded configurations use the x1 clock lines to distribute only the high-speed serial transceiver clock synthesized by the CMU0 PLL or CMU1 PLL to the clock transmitter channels located in the same transceiver block. The low-speed parallel transceiver clock is generated in the transceiver channels using the local clock dividers. x4 bonded configurations use the x4_GXB clock lines to distribute both the high-speed serial and low-speed parallel transceiver clocks generated by the CMU0_Channel to clock the bonded transmitter channels located in the same transceiver block. x8 and xN bonded configurations use the xN_Top or xN_Bottom clock lines to distribute both the high-speed serial and low-speed parallel transceiver clocks generated by the CMU0 channel block to all bonded transmitter channels located across transceiver blocks. ATX PLLs always use xN lines to distribute the high-speed serial and low-speed parallel transceiver clocks. Use the xN_Top line if the CMU0 PLL or ATX PLL that generates the transceiver clocks is located at the top of the transmitter channel. Use the xN_Bottom line if the CMU0 PLL or ATX PLL is located at the bottom of the transmitter channel. Because there is only one xN_Top and xN_Bottom line on each side of the device, using an ATX PLL limits the use of the xN clock lines to distribute the transceiver clocks to other transmitter channels in the design. Transmitter Channel Clocking Configurations Figure 2-14 shows various transmitter channel clocking configurations. Figure 2-14. Transmitter Channel Clocking Configurations Transmitter Channel Clocking Configurations Bonded Non-Bonded x4 Bonded Non-Bonded Basic (PMA Direct) Bonded Basic (PMA Direct) xN x8 Bonded Transmitter channels configured in modes other than Basic (PMA Direct) mode use both the transmitter channel PCS and PMA blocks. As a result, Stratix IV devices allow placing these transmitter channels only in the four regular channels of a transceiver block. Stratix IV devices do not allow configuring the CMU channels in any mode other than Basic (PMA Direct) mode because of the absence of PCS blocks in the CMU channels. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-24 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking The transmitter channel datapath clocking in modes other than Basic (PMA Direct) mode depends on whether the transmitter channel is configured in non-bonded or bonded mode. Non-Bonded Channel Configurations The following modes other than Basic (PMA Direct) functional mode have a non-bonded transmitter channel configuration: PCIe x1--Gen1 and Gen2 Gigabit Ethernet (GIGE) Serial RapidIO SONET/SDH SDI (OIF) CEI PHY Interface Basic (except Basic x4 and Basic x8 modes) Deterministic Latency Use the CMU channels to generate transceiver clocks for all the non-bonded functional modes listed above. Additionally, you may use the ATX PLLs if the configured data rate falls within the ATX PLL data rate range specified in the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-25 Figure 2-15 shows transmitter channel datapath clocking in non-bonded channel configurations when clocked using the CMU PLLs. Figure 2-15. Transmitter Datapath Clocking in a Non-Bonded Configuration Clocked by CMU PLLs (1) Channel 3 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Byte Serializer wrclk tx_coreclk[3] Transmitter Channel PMA 8B/10B Encoder Serializer rdclk x1 High-Speed Serial Clock /2 Low-Speed Parallel Clock FPGA Fabric-Transciever Interface Clock Local Clock Divider Block tx_clkout[3] Channel 2 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA Byte Serializer 8B/10B Encoder wrclk tx_coreclk[2] Serializer rdclk x1 High-Speed Serial Clock /2 Low-Speed Parallel Clock FPGA Fabric-Transciever Interface Clock Local Clock Divider Block tx_clkout[2] Input Reference Clock CMU1 Clock Divider CMU1 PLL CMU1 Channel FPGA Fabric Input Reference Clock CMU0 Clock Divider CMU0 PLL CMU0 Channel Channel 1 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA Byte Serializer 8B/10B Encoder tx_coreclk[1] Serializer rdclk wrclk x1 High-Speed Serial Clock /2 Low-Speed Parallel Clock FPGA Fabric-Transceiver Interface Clock Local Clock Divider Block tx_clkout[1] Channel 0 Transmitter Channel PMA Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk tx_coreclk[0] rdclk Byte Serializer wrclk 8B/10B Encoder x1 High-Speed Serial Clock /2 Low-Speed Parallel Clock FPGA Fabric-Transceiver Interface Clock Serializer rdclk Local Clock Divider Block tx_clkout[0] Note to Figure 2-15: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, and the blue lines represent the x1 high-speed serial clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-26 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking In non-bonded channel configurations clocked by the CMU PLL, each channel can derive its clock independently from either CMU0 PLL or CMU1 PLL within the same transceiver block. The CMU PLL synthesizes the input reference clock to generate a clock that is distributed to the local clock divider block in each channel using the x1 high-speed serial clock line. Depending on the configured functional mode, the local clock divider block in each channel generates the low-speed parallel clock and high-speed serial clock. The serializer in the transmitter channel PMA uses both the low-speed parallel clock and high-speed serial clock for its parallel-in-serial-out operation. The low-speed parallel clock clocks the 8B/10B encoder (if enabled) and the read port of the byte serializer (if enabled) in the transmitter channel PCS. Depending on whether you use the byte serializer or not, the low-speed parallel clock (when you do not use the byte serializer) or a divide-by-two version of the low-speed parallel clock (when you use the byte serializer) from the local clock divider block clocks the read port of the transmitter phase compensation FIFO in all four bonded channels. This clock is driven directly on the tx_clkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to clock the transmitter data and control logic in the FPGA fabric for all four bonded channels. 1 If you configure the ATX PLL to clock the transmitter channel, the ATX PLL block drives the high-speed serial clock and low-speed parallel clock to the transmitter channel on the xN_Top or xN_Bottom lines. f For more information, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Table 2-2 lists the transmitter channel datapath clock frequencies in non-bonded functional modes that have a fixed data rate. Table 2-6. Transmitter Channel Datapath Clock Frequencies in Non-Bonded Functional Modes Data Rate High-Speed Serial Clock Frequency Low-Speed Parallel Clock Frequency (MHz) PCIe x1 (Gen 1) 2.5 Gbps 1.25 GHz PCIe x1 (Gen 2) 5 Gbps Functional Mode FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Serializer (MHz) With Byte Serializer (MHz) 250 250 125 2.5 GHz 500 N/A 250 1.25 Gbps 625 MHz 125 125 N/A 1.25 Gbps 625 MHz 125 N/A 62.5 2.5 Gbps 1.25 GHz 250 N/A 125 3.125 Gbps 1.5625 GHz 312.5 N/A 156.25 SONET/SDH OC12 622 Mbps 311 MHz 77.75 77.75 N/A SONET/SDH OC48 2.488 Gbps 1.244 GHz 311 N/A 155.5 1.485 Gbps 742.5 MHz 148.5 148.5 74.25 1.4835 Gbps 741.75 MHz 148.35 148.35 74.175 GIGE Serial RapidIO HD-SDI 3G-SDI Stratix IV Device Handbook Volume 2: Transceivers 2.97 Gbps 1.485 GHz 297 N/A 148.5 2.967 Gbps 1.4835 GHz 296.7 N/A 148.35 September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-27 Bonded Channel Configurations In PCS and PMA bonded channel configurations, the PCS and PMA blocks of all bonded channels are clocked by the same low-speed parallel clock and high-speed serial clock from the CMU0 clock divider or the ATX PLL block. The phase compensation FIFOs of all bonded channels also share common read and write pointers and enable signals generated in the CCU. Stratix IV devices support x4 PCS and PMA channel bonding that allows bonding of four channels within the same transceiver block. Stratix IV devices also support x8 channel bonding that allows bonding of eight PCS and PMA channels across two transceiver blocks on the same side of the device. x4 PCS and PMA Bonded Channel Configuration The following functional modes support x4 PCS and PMA bonded transmitter channel configuration: PCIe x4--Gen1 and Gen2 XAUI Basic x4 Use the CMU channels to generate the transceiver clocks for all x4 bonded functional modes listed above. Additionally, you may use the ATX PLLs to generate the transceiver clocks for PCIe x4 Gen 2 and Basic x4 functional mode. 1 September 2012 You must assign tx_dataout[0] of the x4 bonded link (XAUI or PCIe x4) to physical channel 0 of the transceiver block, tx_dataout[1] to physical channel 1 of the transceiver block, tx_dataout[2] to physical channel 2 of the transceiver block, and tx_dataout[3] to physical channel 3 of the transceiver block. Otherwise, the Quartus II compilation errors out. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-28 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-16 shows the transmitter channel datapath clocking in x4 channel bonding configurations when clocked using the CMU0 channel. Figure 2-16. Transmitter Datapath Clocking in x4 Bonded Configurations (1) Channel 3 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA Byte Serializer 8B/10B Encoder tx_coreclk[3] x4 High-Speed Serial Clock /2 FPGA Fabric-Transceiver Interface Clock Serializer rdclk wrclk x4 Low-Speed Parallel Clock Channel 2 Transmitter Channel PMA Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Byte Serializer 8B/10B Encoder tx_coreclk[2] Serializer rdclk wrclk x4 High-Speed Serial Clock /2 FPGA Fabric-Transceiver Interface Clock x4 Low-SPeed Parallel Clock coreclkout /2 Input Reference Clock CMU1 Channel CMU1 PLL FPGA Fabric Input Reference Clock CMU0 Channel CMU0 Clock Divider CMU0 PLL x4 Low-Speed Parallel Clock x4 High-Speed Serial Clock Channel 1 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk rdclk Transmitter Channel PMA Byte Serializer 8B/10B Encoder Serializer rdclk wrclk tx_coreclk[1] x4 High-Speed Serial Clock /2 FPGA Fabric-Transceiver Interface Clock x4 Low-Speed Parallel Clock Channel 0 Transmitter Channel PCS PCIe hard IP PIPE Interface TX Phase Compensation FIFO wrclk tx_coreclk[0] FPGA Fabric-Transceiver Interface Clock rdclk Byte Serializer Transmitter Channel PMA 8B/10B Encoder Serializer rdclk wrclk x4 High-Speed Serial Clock /2 x4 Low-Speed Parallel Clock Note to Figure 2-16: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, and the blue lines represent the high-speed serial clock. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-29 The transceiver clocks are distributed to the four bonded channels on the x4 high-speed serial and x4 low-speed parallel clock lines. The serializer in the transmitter channel PMA of the four bonded channels uses the same low-speed parallel clock and high-speed serial clock from CMU0 Channel for their parallel-in-serial-out operation. The low-speed parallel clock clocks the 8B/10B encoder and the write port of the byte serializer (if enabled) in the transmitter channel PCS. Depending on whether the you use the byte serializer or not, the low-speed parallel clock (when you do not use the byte serializer) or a divide-by-two version of the low-speed parallel clock (when you use the byte serializer) from the CMU0 clock divider block clocks the read port of the transmitter phase compensation FIFO in all four bonded channels. This clock is driven directly on the coreclkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to clock the transmitter data and control logic in the FPGA fabric for all four bonded channels. 1 The ATX PLL block drives the high-speed serial clock and low-speed parallel clock to the transmitter channels on the xN_Top or xN_Bottom lines. f For more information, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. In x4 PCS and PMA bonded channel configurations, the transmitter phase compensation FIFOs in all four bonded channels share common read and write pointers and enable signals generated in the CMU0 channel of the transceiver block. This ensures equal transmitter phase compensation FIFO latency across all four bonded channels, resulting in low transmitter channel-to-channel skew. Table 2-3 lists the transmitter datapath clock frequencies in x4 bonded functional modes that have a fixed data rate. Table 2-7. Transmitter Datapath Clock Frequencies in x4 Bonded Functional Modes Data Rate (Gbps) High-Speed Serial Clock Frequency (GHz) Low-Speed Parallel Clock Frequency (MHz) PCIe x4 (Gen 1) 2.5 1.25 PCIe x4 (Gen 2) 5 3.125 Functional Mode XAUI FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Serializer (MHz) With Byte Serializer (MHz) 250 250 125 2.5 500 N/A 250 1.5625 312.5 N/A 156.25 x8 PCS and PMA Bonded Channel Configuration The following functional modes support x8 PCS and PMA bonded transmitter channel configuration: PCIe x8--Gen1 and Gen2 Basic x8 Use either the CMU PLL or the ATX PLL to generate the transceiver clocks in Basic x8 functional modes. Use the ATX PLL in PCIe x8 Gen2 mode in order to meet the transmitter jitter compliance. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-30 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking The eight bonded channels are located in two transceiver blocks, referred to as the master transceiver block and the slave transceiver block, with four channels each. When clocked using a CMU PLL, the CMU0 clock divider in CMU0 channel of the master transceiver block drives the high-speed serial clock and low-speed parallel clock on the xN_Top clock line. The serializer in the transmitter channel PMA of all eight bonded channels uses the same low-speed parallel clock and high-speed serial clock driven by the CMU0 channel of the master transceiver block on the xN_Top clock line. The low-speed parallel clock from CMU0 channel of the master transceiver block clocks the 8B/10B encoder and the write port of the byte serializer (if enabled) in the transmitter channel PCS of all eight channels. Depending on whether you use the byte serializer or not, the low-speed parallel clock (when you do not use the byte serializer) or a divide-by-two version of the low-speed parallel clock (when you use the byte serializer) from the CMU0 clock divider block clocks the read port of the transmitter phase compensation FIFO in all eight bonded channels. This clock is driven directly on the coreclkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to clock the transmitter data and control logic in the FPGA fabric for all eight bonded channels. 1 If you choose the ATX PLL to generate the transceiver clocks for the x8 bonded channels, Altera recommends placing the ATX PLL between the master and slave transceiver block to minimize transmitter channel-to-channel skew. In this configuration, the ATX PLL block drives the high-speed serial clock and low-speed parallel clock to the master transceiver block on the xN_Bottom lines. It drives the high-speed serial clock and low-speed parallel clock to the slave transceiver block on the xN_Top lines. f For more information, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. In PCIe x8 and Basic x8 bonded channel configurations, the transmitter phase compensation FIFOs in all eight bonded channels share common read and write pointers and enable signals generated in the CCU of the master transceiver block. This ensures equal transmitter phase compensation FIFO latency across all eight bonded channels, resulting in low transmitter channel-to-channel skew. 1 Stratix IV Device Handbook Volume 2: Transceivers The difference in clock routing delays between the x4 clock lines and the xN clock lines can result in higher transmitter channel-to-channel skew. To compensate for this difference in clock routing delays between the x4 and the xN clock lines, the Stratix IV transceivers introduce a fixed amount of delay in the x4 clock lines of the transceiver block whose CMU0 channel generates the transceiver clocks in Basic x8 bonded channel configuration. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-31 Figure 2-17 shows the transmitter datapath clocking in PCIe x8 channel bonding configurations when clocked using the CMU channel in the master transceiver block. Figure 2-17. Transmitter Datapath Clocking in x8 Bonded Configuration (1) Slave Transceiver Block Transmitter Channel PCS PCIe hard IP TX Phase Compensation FIFO PIPE Interface wrclk tx_coreclk[7:4] rdclk Byte Serializer 8B/10B Encoder Transmitter Channel PMA Serializer rdclk wrclk /2 FPGA Fabric-Transceiver Interface Clock Low-Speed Parallel Clock from CMU0 of the Master Transceiver Block CMU1 Channel CMU1 PLL CMU1 Clock Divider CMU0 PLL CMU0 Clock Divider CMU0 Channel FPGA Fabric Master Transceiver Block Transmitter Channel PCS PCIe hard IP tx_coreclk[3:0] TX Phase Compensation FIFO PIPE Interface wrclk rdclk Byte Serializer wrclk Serializer rdclk /2 FPGA Fabric-Transceiver Interface Clock 8B/10B Encoder Transmitter Channel PMA Low-Speed Parallel Clock from CMU0 of the Master Transceiver Block coreclkout /2 Input Reference Clock CMU1 Channel CMU1 PLL CMU1 Clock Divider CMU0 Channel Input Reference Clock CMU0 PLL CMU0 Clock Divider Low-Speed Parallel Clock High-Speed Serial Clock Note to Figure 2-17: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, and the blue lines represent the high-speed serial clock. Figure 2-18 through Figure 2-20 show the allowed master and slave transceiver block locations and PCIe logical lane to physical transceiver channel mapping in all Stratix IV devices. 1 September 2012 The Quartus II compilation errors out if you do not map the PCIe logical lanes to the physical transceiver channels, as shown in Figure 2-18 through Figure 2-20. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-32 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-18 shows one PCIe x8 link in two transceiver block devices and two PCIe x8 links in four transceiver block devices. Figure 2-18. One PCIe x8 Link in Two Transceiver Block Devices and Two PCIe x8 Links in Four Transceiver Block Devices (1) Two PCIe x8 Links in Four Transceiver Block Devices EP4SGX290FH29, EP4SGX360FH29, EP4SGX110FF35, EP4SGX230FF35, EP4SGX290FF35, EP4SGX360FF35, EP4SGX230HF35, EP4SGX290HF35, EP4SGX360HF35, EP4SGX530HH35 One PCIe x8 Link in Two Transceiver Block Devices EP4SGX70DF29 EP4SGX110DF29 EP4SGX230DF29 Transceiver Block GXBL1 (Slave) Transceiver Block GXBR1 (Slave) PCIe Lane 7 Channel3 Channel3 PCIe Lane 7 PCIe Lane 6 Channel2 Channel2 PCIe Lane 6 PCIe Lane 5 Channel1 Channel1 PCIe Lane 5 PCIe Lane 4 Channel0 Channel0 PCIe Lane 4 Second PCIe x8 Link Transceiver Block GXBL0 (Master) Channel3 First PCIe x8 Link Transceiver Block GXBR0 (Master) Channel3 PCIe Lane 3 PCIe Lane 2 Channel2 Channel2 PCIe Lane 2 PCIe Lane 1 Channel1 Channel1 PCIe Lane 1 PCIe Lane 0 Channel0 Channel0 PCIe Lane 0 PCIe Lane 3 Note to Figure 2-18: (1) You can use a x4 PCIe configuration in either a master or slave block. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-33 Figure 2-19 shows two PCIe x8 links in six transceiver block devices. Figure 2-19. Two PCIe x8 Links in Six Transceiver Block Devices (Note 1), (2) EP4SGX230KF40, EP4SGX290KF40, EP4SGX360KF40, EP4SGX530KF40 Transceiver Block GXBR2 Channel3 Transceiver Block GXBL2 Channel3 PCIe Lane 7 PCIe Lane 6 Channel2 Channel1 Channel0 Channel0 Transceiver Block GXBR1 (Slave) Channel3 Transceiver Block GXBL1 (Slave) Channel3 Channel2 Channel1 PCIe Lane 5 PCIe Lane 4 PCIe Lane 3 PCIe Lane 2 Channel2 Channel1 Channel0 Channel2 Channel1 Second PCIe x8 Link First PCIe x8 Link Channel0 Transceiver Block GXBL0(Master) Channel3 Transceiver Block GXBR0 (Master) Channel3 Channel2 Channel1 Channel2 Channel1 Channel0 Channel0 PCIe Lane 1 PCIe Lane 0 PCIe Lane 7 PCIe Lane 6 PCIe Lane 5 PCIe Lane 4 PCIe Lane 3 PCIe Lane 2 PCIe Lane 1 PCIe Lane 0 Notes to Figure 2-19: (1) Stratix IV devices with six transceiver blocks allow a maximum of two PCIe x8 links occupying four transceiver blocks. You can configure the other two transceiver blocks to implement other functional modes. (2) You can use a x4 PCIe configuration in either a master or slave block. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-34 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-20 shows four PCIe x8 links in eight transceiver block devices. Figure 2-20. Four PCIe x8 Links in Eight Transceiver Block Devices (1) EP4SGX530NF45 Transceiver Block GXBL3 (Slave) Transceiver Block GXBR3 (Slave) Channel3 Channel2 Channel3 Channel2 PCIe Lane 7 PCIe Lane 6 PCIe Lane 5 Channel1 Channel1 PCIe Lane 5 PCIe Lane 4 Channel0 Channel0 PCIe Lane 4 PCIe Lane 7 Fourth PCIe x8 Link Second PCIe x8 Link PCIe Lane 6 Transceiver Block GXBR2 (Master) Channel3 Channel2 PCIe Lane 3 PCIe Lane 2 Transceiver Block GXBL2 (Master) Channel3 Channel2 PCIe Lane 1 Channel1 Channel1 PCIe Lane 1 PCIe Lane 0 Channel0 Channel0 PCIe Lane 0 PCIe Lane 6 Transceiver Block GXBL1 (Slave) Channel3 Channel2 Transceiver Block GXBR1 (Slave) Channel3 Channel2 PCIe Lane 5 Channel1 Channel1 PCIe Lane 5 PCIe Lane 4 Channel0 Channel0 PCIe Lane 4 PCIe Lane 3 PCIe Lane 7 Third PCIe x8 Link First PCIe x8 Link PCIe Lane 2 PCIe Lane 7 PCIe Lane 6 PCIe Lane 2 Transceiver Block GXBL0 (Master) Channel3 Channel2 Transceiver Block GXBR0 (Master) Channel3 Channel2 PCIe Lane 1 Channel1 Channel1 PCIe Lane 1 PCIe Lane 0 Channel0 Channel0 PCIe Lane 0 PCIe Lane 3 PCIe Lane 3 PCIe Lane 2 Note to Figure 2-20: (1) You can use a x4 PCIe configuration in either a master or slave block. Non-Bonded Basic (PMA Direct) Mode Channel Configurations Figure 2-21 shows four regular channels and the CMU1 channel in a transceiver block configured in non-bonded Basic (PMA Direct) mode. Each channel derives its clock independently from either the CMU0 PLL or CMU1 PLL within the same transceiver block if the CMU channel is configured as a CMU PLL. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-35 f For more information about Basic (PMA Direct) mode, refer to the Transceiver Architecture in Stratix IV Devices chapter. Figure 2-21. Transmitter Channel PMA Directly Interfacing to the User Logic in the FPGA Fabric (1) Channel 3 Transmitter Channel PCS Transmitter Channel PMA Serializer tx_clkout[3] x1 High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider Block Channel 2 Transmitter Channel PCS Transmitter Channel PMA Serializer tx_clkout[2] x1 High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider Block Transmitter Channel PMA CMU1_Channel Serializer CMU1 Clock Divider FPGA Fabric CMU0_PLL CMU0 Clock Divider CMU0_Channel Channel 1 Transmitter Channel PCS Transmitter Channel PCS Transmitter Channel PMA Serializer x1 High-Speed Serial Clock tx_clkout[1] Low-Speed Parallel Clock Local Clock Divider Block Channel 0 TransmitterChannel ChannelPCS PCS Transmitter Transmitter Channel PMA Serializer tx_clkout[0] x1 High-Speed Serial Clock Low-Speed Parallel Clock Local Clock Divider Block Note to Figure 2-21: (1) The green lines represent the low-speed parallel clock and the blue lines represent the high-speed serial clock. 1 September 2012 Stratix IV devices do not allow the 6G ATX PLL to generate transceiver clocks in non-bonded Basic (PMA Direct) mode. The transmitter clock for channels configured in non-bonded Basic (PMA Direct) mode must be generated by one of the CMU PLLs in the transceiver block containing the channels. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-36 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking The CMU0 PLL synthesizes the input reference clock to generate a clock that is distributed to the local clock divider block in each of the four regular channels using the x1 high-speed serial clock line. It is also forwarded to the CMU1 clock divider in the CMU1 channel configured as a non-bonded Basic (PMA-Direct) channel. The local clock divider block in each regular channel and the CMU1 clock divider in the CMU1 channel generate the low-speed parallel clock and high-speed serial clock. The serializer in the transmitter channel PMA of each channel uses both the low-speed parallel clock and high-speed serial clock for its parallel-in-serial-out operation. The low-speed parallel clock is also driven directly on the tx_clkout port as the FPGA fabric-Transceiver interface clock. You can use the tx_clkout port to clock transmitter data and control logic in the FPGA fabric. Bonded Basic (PMA Direct) xN Mode Channel Configurations Bonded Basic (PMA Direct) xN mode offers low transmitter channel-to-channel skew in addition to the flexibility of implementing custom PCS logic in the FPGA fabric. Stratix IV devices allow bonding all regular channels and CMU channels on one side of the device in Basic (PMA Direct) xN mode. For example, devices such as EP4SGX530NF45 or EP4S100G5F45 allow bonding of up to 24 channels placed in four transceiver blocks on each side of the device. 1 The coreclkout port is not available in Basic (PMA Direct) xN mode. In bonded channel configurations, the CMU0 clock divider of all the transceiver blocks is used, as shown in Figure 2-17. Unlike bonded channel configurations, in Basic (PMA Direct) xN configuration: Stratix IV Device Handbook Volume 2: Transceivers If you use the ATX PLL to generate the transceiver datapath interface clocks, only the clock divider of the ATX PLL is used. If you use the CMU PLL to generate the transceiver datapath interface clocks, only the CMU0 clock divider block of the transceiver block containing the CMU PLL is used. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-37 Figure 2-22 shows transmitter channel clocking for 17 channels configured in Basic (PMA Direct) xN mode. Figure 2-22. Transmitter Channel Clocking for 17 Channels Configured in Basic (PMA Direct) xN Mode Transceiver Block GXBR2 Regular Channel TX PCS Regular Channel TX PMA Serializer CMU0 Channel TX PMA Serializer xN_Bottom High-Speed Serial and Low-Speed Parallel clock Transceiver Block GXBR1 Regular Channel TX PCS Regular Channel TX PMA Serializer x4 High-Speed Serial and Low-Speed Parallel clock CMU1 Channel TX PMA FPGA Fabric Serializer CMU1_Channel CMU0_PLL Transmitter Channel PMA CMU0 Clock Divider Transceiver Block GXBR0 Regular Channel TX PCS Regular Channel TX PMA Serializer xN_Top High-Speed Serial and Low-Speed Parallel clock CMU0 Channel TX PMA Serializer Figure 2-22 shows 17 channels configured in Basic (PMA Direct) xN mode and located across three transceiver blocks on the right side of the Stratix IV device. Each of the two transceiver blocks, GXBR0 and GXBR2, contain six of the 17 xN bonded channels located in four regular channels and two CMU channels. The remaining five of the 17 xN bonded channels are located in four regular channels and the CMU1 channel of the transceiver block GXBR1. 1 September 2012 Stratix IV devices allow both CMU channels and 6G ATX PLL blocks to generate the high-speed serial and low-speed parallel transceiver clocks when configured in Basic (PMA Direct) xN mode. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-38 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking f For more examples regarding this clocking scheme, refer to: "Example 1: Channel Configuration with a 4 Gbps Data Rate" on page 2-9 AN 571: Implementing the SERDES Framer Interface Level 5 (SFI-5.1) Protocol in Stratix IV Devices AN 572: Implementing the Scalable SERDES Framer Interface (SFI-S) Protocol in Stratix IV GT Devices Transmitter Channel-to-Channel Skew Optimization in Basic (PMA Direct) xN Mode In Basic (PMA-Direct) xN mode, the CMU0 channel distributes the transceiver clocks to the channels placed in the same transceiver block using the x4 clock lines. The x4 clock lines drive the xN_Top and xN_Bottom clock lines to distribute the transceiver clocks to the transmitter channels located in transceiver blocks on the bottom and top. The difference in clock routing delays between the x4 clock lines and the xN clock lines can result in higher transmitter channel-to-channel skew. To compensate for this difference in clock routing delays between the x4 and the xN clock lines, the Stratix IV transceivers introduce a fixed amount of delay in the x4 clock lines of the transceiver block whose CMU0 channel generates the transceiver clocks. 1 The delay compensation mechanism engaged in Basic (PMA Direct) mode only compensates for the clock routing delays between the transceiver block whose CMU0 channel generates the transceiver clocks and its adjacent transceiver block located above and below. To minimize transmitter channel-to-channel skew in xN bonded channels, use the recommended placement shown in Table 2-8. Table 2-8. Recommended Placement of Channels and CMU in Bonded Modes 1 Channel Placement CMU Placement 2 adjacent transceiver blocks In either of the two transceiver blocks. 3 adjacent transceiver blocks In the middle transceiver block. 4 adjacent transceiver blocks In either of the middle transceiver blocks. If you use the ATX PLL to generate the transceiver clocks, Altera recommends placing the channels in the transceiver blocks adjacent to the ATX PLL on both sides of the ATX PLL. f For manual placement of the CMU and ATX PLLs, if the Quartus II software does not automatically pick the most optimal location for skew, refer to AN 578: Manual Placement of CMU PLLs and ATX PLLs in Stratix IV GX and GT Devices. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-39 Meeting Timing in Basic (PMA Direct) Mode Timing may not be met for higher data rates when transceiver channels are configured in Basic (PMA Direct) functional mode. To meet FPGA fabric-Transmitter PMA interface timing above certain data rates, you may need to phase shift the interface clock tx_clkout used to clock the transmitter user logic. To meet FPGA fabric-Receiver hold time violations, you may have to modify the way data is captured in the FPGA fabric. f For more information, refer to AN 580: Achieving Timing Closure in Basic (PMA Direct) Functional Mode. Receiver Channel Datapath Clocking This section describes the receiver PMA and PCS datapath clocking in supported configurations. Receiver datapath clocking varies between non-bonded and bonded channel configurations. It also varies with the use of PCS blocks, such as deskew FIFO and rate matcher. This section describes the following: "Non-Bonded Channel Configurations" "Bonded Channel Configurations" on page 2-43 "Basic (PMA Direct) Mode Channel Configurations" on page 2-49 Non-Bonded Channel Configurations In non-bonded channel configurations, receiver PCS blocks of each channel are clocked independently. Each non-bonded channel also has separate rx_analogreset and rx_digitalreset signals that allow independent reset of the receiver PCS logic in each channel. 1 For more information about transceiver reset and power down signals, refer to the Reset Control and Power Down in Stratix IV Devices chapter. In non-bonded channel configurations, receiver channel datapath clocking has two scenarios: "Non-Bonded Receiver Clocking Without Rate Matcher" "Non-Bonded Receiver Clocking with Rate Matcher" on page 2-41 Non-Bonded Receiver Clocking Without Rate Matcher The following functional modes have non-bonded receiver channel configuration without rate matcher: September 2012 SONET/SDH SDI (OIF) CEI PHY Interface Basic without rate matcher Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-40 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-23 shows receiver datapath clocking in non-bonded channel configurations without rate matcher. Figure 2-23. Receiver Datapath Clocking in Non-Bonded Configurations Without Rate Matcher (1) Channel 3 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Receiver Channel PMA Input Reference Clock Byte Ordering Byte De-Serializer 8B/10B Decoder Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[3] Ch3 Parallel Recovered Clock FPGA Fabric-Transceiver Interface Clock /2 rx_clkout[3] Channel 2 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte De-Serializer 8B/10B Decoder Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[2] Ch2 Parallel Recovered Clock FPGA Fabric-Transceiver Interface Clock /2 rx_clkout[2] FPGA Fabric Channel 1 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte De-Serializer 8B/10B Decoder Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[1] Ch1 Parallel Recovered Clock FPGA Fabric-Transceiver Interface Clock /2 rx_clkout[1] Channel 0 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte De-Serializer 8B/10B Decoder Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[0] Ch0 Parallel Recovered Clock FPGA Fabric-Transceiver Interface Clock rx_clkout[0] /2 Note to Figure 2-23: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the parallel recovered clock, and the blue lines represent the serial recovered clock. In non-bonded configurations without rate matcher, the CDR in each receiver channel recovers the serial clock from the received data. The serial recovered clock is divided within the receiver PMA to generate the parallel recovered clock. The deserializer uses the serial recovered clock in the receiver PMA. The parallel recovered clock and deserialized data is forwarded to the receiver PCS. The parallel recovered clock in each channel clocks the word aligner and 8B/10B decoder (if enabled). Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-41 Depending on whether you use the byte deserializer or not, the parallel recovered clock (when you do not use the byte deserializer) or a divide-by-two version of the parallel recovered clock (when you use the byte deserializer) clocks the write port of the receiver phase compensation FIFO. This clock is driven directly on the rx_clkout port as the FPGA fabric-Transceiver interface clock. You can use the rx_clkout signal to capture the receiver data and status signals in the FPGA fabric. Table 2-9 lists the receiver datapath clock frequencies in non-bonded functional modes without rate matcher. Table 2-9. Receiver Datapath Clock Frequencies in Non-Bonded Functional Modes Without Rate Matcher FPGA Fabric-Transceiver Interface Clock Frequency Data Rate Serial Recovered Clock Frequency Parallel Recovered Clock Frequency (MHz) SONET/SDH OC12 622 Mbps 311 MHz SONET/SDH OC48 2.488 Gbps Functional Mode HD-SDI 3G-SDI Without Byte Deserializer (MHz) With Byte Deserializer (MHz) 77.75 77.75 N/A 1.244 GHz 311 N/A 155.5 1.485 Gbps 742.5 MHz 148.5 148.5 74.25 1.4835 Gbps 741.75 MHz 148.35 148.35 74.175 2.97 Gbps 1.485 GHz 297 N/A 148.5 2.967 Gbps 1.4835 GHz 296.7 N/A 148.35 Non-Bonded Receiver Clocking with Rate Matcher The following functional modes have non-bonded receiver channel configuration with rate-matcher: September 2012 PCIe x1 GIGE Serial RapidIO Basic with rate matcher Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-42 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-24 shows the receiver datapath clocking in non-bonded channel configurations with rate matcher. Figure 2-24. Receiver Datapath Clocking in Non-Bonded Configurations with Rate Matcher (1) Channel 3 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[3] Ch3 Parallel Recovered Clock FPGA Fabric_Transceiver Interface Clock Local Clock Divider /2 Low-Speed Parallel Clock tx_clkout[3] From CMU0 PLL From CMU1 PLL Transmitter Channel PMA Channel 2 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[2] Ch2 Parallel Recovered Clock FPGA Fabric_Transceiver Interface Clock /2 tx_clkout[2] Local Clock Divider Low-Speed Parallel Clock From CMU0 PLL From CMU1 PLL Transmitter Channel PMA FPGA Fabric Channel 1 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[1] FPGA Fabric_Transceiver Interface Clock Ch1 Parallel Recovered Clock /2 Local Clock Divider Low-Speed Parallel Clock tx_clkout[1] From CMU0 PLL From CMU1 PLL Transmitter Channel PMA Channel 0 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[0] FPGA Fabric_Transceiver Interface Clock /2 tx_clkout[0] Ch0 Parallel Recovered Clock Low-Speed Parallel Clock Local Clock Divider From CMU0 PLL From CMU1 PLL Transmitter Channel PMA Note to Figure 2-24: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, the dark red lines represent the parallel recovered clock, and the blue lines represent the serial recovered clock. In non-bonded configurations with rate matcher, the CDR in each receiver channel recovers the serial clock from the received data. The serial recovered clock is divided within the receiver PMA to generate the parallel recovered clock. The deserializer uses the serial recovered clock in the receiver PMA. The parallel recovered clock and deserialized data are forwarded to the receiver PCS. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-43 The parallel recovered clock from the receiver PMA in each channel clocks the word aligner and the write port of the rate match FIFO. The low-speed parallel clock from the transmitter local clock divider block in each channel clocks the read port of the rate match FIFO, the 8B/10B decoder, and the write port of the byte deserializer (if enabled). The parallel transmitter PCS clock or its divide-by-two version (if byte deserializer is enabled) clocks the write port of the receiver phase compensation FIFO. It is also driven on the tx_clkout port as the FPGA fabric-Transceiver interface clock. You can use the tx_clkout signal to latch the receiver data and status signals in the FPGA fabric. Table 2-10 lists the receiver datapath clock frequencies in non-bonded functional modes with rate matcher. Table 2-10. Receiver Datapath Clock Frequencies in Non-Bonded Functional Modes with Rate Matcher Data Rate (Gbps) Serial Recovered Clock Frequency Parallel Recovered Clock and Parallel Transmitter PCS Clock Frequency (MHz) PCIe x1 (Gen 1) 2.5 1.25 GHz PCIe x1 (Gen 2) 5 Functional Mode GIGE FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Deserializer (MHz) With Byte Deserializer (MHz) 250 250 125 2.5 GHz 500 N/A 250 1.25 625 MHz 125 125 N/A 1.25 625 MHz 125 N/A 62.5 2.5 1.25 GHz 250 N/A 125 3.125 1.5625 GHz 312.5 N/A 156.25 Serial RapidIO Bonded Channel Configurations The Stratix IV device supports x4 channel bonding that allows bonding of four channels within the same transceiver block. It also supports x8 channel bonding that allows bonding of eight channels across two transceiver blocks on the same side of the device. In bonded channel configurations, the low-speed parallel clock for all bonded channels are generated by the same CMU0 clock divider or the ATX clock divider block, resulting in lower channel-to-channel clock skew. The receiver phase compensation FIFO in all bonded channels (except in Basic [PMA Direct] xN mode) share common pointers and control logic generated in the CCU, resulting in equal latency in the receiver phase compensation FIFO of all bonded channels. 1 Bonding is not supported on the receive side for Basic x4 and Basic x8 functional modes. If you use rate matcher, the clocking scheme for Basic x4 and Basic x8 functional modes, the clocking is similar to PCIe x4 mode, as shown in Figure 2-26 on page 2-46 and PCIe x8 mode, as shown in Figure 2-27 on page 2-48. x4 Bonded Channel Configuration The following functional modes support x4 receiver channel bonded configuration: September 2012 XAUI ("x4 Bonded Channel Configuration with Deskew FIFO" on page 2-44) PCIe ("x4 Bonded Channel Configuration Without Deskew FIFO" on page 2-46) Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-44 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking x4 Bonded Channel Configuration with Deskew FIFO XAUI functional mode has x4 bonded channel configuration with deskew FIFO. Figure 2-25 shows the receiver datapath clocking in x4 channel bonding configurations with deskew FIFO. Figure 2-25. Receiver Datapath Clocking in x4 Bonded Channel Configuration with Deskew FIFO (1) Channel 3 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Receiver Channel PMA Input Reference Clock Rate Match FIFO Deskew FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[3] Ch3 Parallel Recovered Clock Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 2 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Receiver Channel PMA Input Reference Clock Rate Match FIFO Deskew FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[2] Ch2 Parallel Recovered Clock Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider coreclkout FPGA Fabric-Transceiver Interface Clock /2 CMU1 Channel Input Reference Clock FPGA Fabric CMU1 Clock Divider CMU1 PLL CMU0 Channel Input Reference Clock CMU0 Clock Divider CMU0 PLL Low-Speed Parallel Clock Channel 1 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Deskew FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[1] Ch1 Parallel Recovered Clock Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 0 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Deskew FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[0] Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Note to Figure 2-25: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, the dark red lines represent the Ch0 parallel recovered clock, and the blue lines represent the serial recovered clock. In x4 bonded channel configurations with deskew FIFO, the CDR in each receiver channel recovers the serial clock from the received data. The serial recovered clock is divided within each channel's receiver PMA to generate the parallel recovered clock. The deserializer uses the serial recovered clock in the receiver PMA. The parallel recovered clock and deserialized data is forwarded to the receiver PCS in each channel. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-45 The parallel recovered clock from the receiver PMA in each channel clocks the word aligner in that channel. The parallel recovered clock from Channel 0 clocks the deskew FIFO and the write port of the rate match FIFO in all four bonded channels. The low-speed parallel clock from the CMU0 clock divider block in CMU0_Channel clocks the read port of the rate match FIFO, the 8B/10B decoder, and the write port of the byte deserializer (if enabled) in all four bonded channels. The low-speed parallel clock or its divide-by-two version (if byte deserializer is enabled) clocks the write port of the receiver phase compensation FIFO. It is also driven on the coreclkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to latch the receiver data and status signals in the FPGA fabric for all four bonded channels. Table 2-11 lists the receiver datapath clock frequencies in x4 bonded functional modes with deskew FIFO. Table 2-11. Receiver Datapath Clock Frequencies in x4 Bonded Functional Modes with Deskew FIFO Parallel Recovered Clock and Parallel Transmitter PCS Clock Frequency (MHz) FPGA-Fabric Transceiver Interface Clock Frequency Data Rate (Gbps) Serial Recovered Clock Frequency PCIe x4 (Gen 1) 2.5 1.25 GHz 250 250 125 PCIe x4 (Gen 2) 5 2.5 GHz 500 N/A 250 3.125 1.5625 MHz 312.5 N/A 156.25 Functional Mode XAUI September 2012 Altera Corporation Without Byte Deserializer (MHz) With Byte Deserializer (MHz) Stratix IV Device Handbook Volume 2: Transceivers 2-46 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking x4 Bonded Channel Configuration Without Deskew FIFO PCIe x4 functional modes supports the x4 bonded channel configuration without deskew FIFO. Figure 2-26 shows the receiver datapath clocking in x4 channel bonding configurations without deskew FIFO. Figure 2-26. Receiver Datapath Clocking in x4 Bonded Channel Configuration Without Deskew FIFO (1) Channel 3 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[3] Ch3 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 2 Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Receiver Channel PMA Input Reference Clock Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[2] Ch2 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider coreclkout FPGA Fabric_Transceiver Interface Clock /2 CMU1 Channel Input Reference Clock CMU1 PLL Input Reference Clock CMU0 PLL CMU1 Clock Divider FPGA Fabric CMU0 Channel CMU0 Clock Divider Low-Speed Parallel Clock Channel 1 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[1] Ch1 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 0 Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[0] Ch0 Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider Note to Figure 2-26: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, the dark red lines represent the parallel recovered clock, and the blue lines represent the serial recovered clock. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-47 In x4 bonded channel configurations without deskew FIFO, the CDR in each receiver channel recovers the serial clock from the received data. The serial recovered clock is divided within each channel's receiver PMA to generate the parallel recovered clock. The deserializer uses the serial recovered clock in the receiver PMA. The parallel recovered clock and deserialized data is forwarded to the receiver PCS in each channel. The parallel recovered clock from the receiver PMA in each channel clocks the word aligner and the write side of the rate matcher FIFO in that channel. The low-speed parallel clock from the CMU0 clock divider block in CMU0_Channel clocks the read port of the rate match FIFO, the 8B/10B decoder, and the write port of the byte deserializer (if enabled). The low-speed parallel clock or its divide-by-two version (if byte deserializer is enabled) clocks the receiver phase compensation FIFO. It is also driven on the coreclkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to latch the receiver data and status signals in the FPGA fabric for all four bonded channels. Table 2-12 lists the receiver datapath clock frequencies in x4 bonded functional modes without deskew FIFO. Table 2-12. Receiver Datapath Clock Frequencies in x4 Bonded Functional Modes without Deskew FIFO Data Rate (Gbps) Serial Recovered Clock Frequency (GHz) Parallel Recovered Clock and Parallel Transmitter PCS Clock Frequency (MHz) PCIe x4 (Gen 1) 2.5 1.25 PCIe x4 (Gen 2) 5 2.5 Functional Mode September 2012 Altera Corporation FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Deserializer (MHz) With Byte Deserializer (MHz) 250 250 125 500 N/A 250 Stratix IV Device Handbook Volume 2: Transceivers 2-48 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking x8 Bonded Channel Configuration PCIe x8 functional mode supports the x8 receiver channel bonding configuration. The eight bonded channels are located in two transceiver blocks, referred to as the master transceiver block and slave transceiver block, with four channels each. Figure 2-27 shows the receiver datapath clocking in PCIe x8 bonded channel configuration. Figure 2-27. Receiver Datapath Clocking in x8 Bonded Channel Configuration (1) Slave Transceiver Block Receiver Channel PCS PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer Receiver Channel PMA Input Reference Clock 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[7:4] Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider CMU1 Channel CMU1 PLL CMU1 Clock Divider CMU0 PLL CMU0 Clock Divider CMU0 Channel FPGA Fabric Master Transceiver Block Receiver Channel PCS Receiver Channel PMA Input Reference Clock PCIe hard IP PIPE Interface RX Phase Compensation FIFO Byte Ordering Byte DeSerializer 8B/10B Decoder Rate Match FIFO Word Aligner DeSerializer CDR Serial Recovered Clock rx_coreclk[3:0] Parallel Recovered Clock /2 Low-Speed Parallel Clock from CMU0 Clock Divider coreclkout FPGA Fabric_Transceiver Interface Clock /2 Input Reference Clock CMU1 Channel CMU1 PLL CMU1 Clock Divider CMU0 Channel Input Reference Clock CMU0 PLL CMU0 Clock Divider Note to Figure 2-27: (1) The red lines represent the FPGA fabric-Transceiver interface clock, the green lines represent the low-speed parallel clock, the dark red lines represent the parallel recovered clock, and the blue lines represent the serial recovered clock. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking 2-49 The CDR in each of the eight receiver channels recovers the serial clock from the received data on that channel. The serial recovered clock is divided within each channel's receiver PMA to generate the parallel recovered clock. The deserializer uses the serial recovered clock in the receiver PMA. The parallel recovered clock and deserialized data from the receiver PMA in each channel is forwarded to the receiver PCS in that channel. The parallel recovered clock from the receiver PMA in each channel clocks the word aligner and the write side of the rate matcher FIFO in that channel. The low-speed parallel clock from the CMU0 clock divider of the master transceiver block clocks the read port of the rate match FIFO, the 8B/10B decoder, and the write port of the byte deserializer (if enabled) in all eight channels. The low-speed parallel clock or its divide-by-two version (if byte deserializer is enabled) clocks the write port of the receiver phase compensation FIFO in all eight channels. It is also driven on the coreclkout port as the FPGA fabric-Transceiver interface clock. You can use the coreclkout signal to latch the receiver data and status signals in the FPGA fabric for all eight bonded channels. Table 2-13 lists the receiver datapath clock frequencies in PCIe x8 functional mode. Table 2-13. Receiver Datapath Clock Frequencies PCIe x8 Functional Mode Data Rate (Gbps) Serial Recovered Clock Frequency (GHz) Parallel Recovered Clock and Parallel Transmitter PCS Clock Frequency (MHz) PCIe x8 (Gen 1) 2.5 1.25 PCIe x8 (Gen 2) 5 2.5 Functional Mode FPGA Fabric-Transceiver Interface Clock Frequency Without Byte Deserializer (MHz) With Byte Deserializer (MHz) 250 250 125 500 N/A 250 Basic (PMA Direct) Mode Channel Configurations Figure 2-28 shows six channels in a transceiver block configured in Basic (PMA Direct) functional mode with two of the channels being CMU channels. The receiver channel PMA directly interfaces to the user logic in the FPGA fabric. The CDR recovers the high-speed serial clock and low-speed parallel clock for the deserializer. The low-speed parallel clock is forwarded to the FPGA fabric as rx_clkout. 1 September 2012 Bonded mode is not available for receivers configured in Basic (PMA Direct) functional mode. Data registers to capture the receiver data in the FPGA fabric for each channel must be clocked by rx_clkout forwarded by that channel's CDR. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-50 Chapter 2: Transceiver Clocking in Stratix IV Devices Transceiver Channel Datapath Clocking Figure 2-28. Receiver Channel PMA Directly Interfacing to the User Logic in the FPGA Fabric Channel 3 (1) Receiver Channel PMA Receiver Channel PCS DeSerializer CDR High-Speed Serial Clock rx_clkout[5] Low-Speed Parallel Clock Channel 2 Receiver Channel PCS Receiver Channel PMA DeSerializer CDR High-Speed Serial Clock rx_clkout[4] Low-Speed Parallel Clock CMU1_Channel Receiver Channel PMA DeSerializer CDR High-Speed Serial Clock rx_clkout[3] Low-Speed Parallel Clock FPGA Fabric CMU0_Channel Receiver Channel PMA DeSerializer CDR High-Speed Serial Clock rx_clkout[2] Low-Speed Parallel Clock Channel 1 Receiver Channel PMA Receiver Channel PCS DeSerializer CDR High-Speed Serial Clock rx_clkout[1] Low-Speed Parallel Clock Channel 0 Receiver Channel PMA Receiver Channel PCS DeSerializer CDR High-Speed Serial Clock rx_clkout[0] Low-Speed Parallel Clock Note to Figure 2-28: (1) The green lines represent the low-speed parallel clock and the blue lines represent the serial recovered clock. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-51 FPGA Fabric-Transceiver Interface Clocking The FPGA fabric-Transceiver interface clocks consist of clock signals from the FPGA fabric to the transceiver blocks and clock signals from the transceiver blocks to the FPGA fabric. These clock resources use the clock networks in the FPGA core that include the global, regional, and periphery clock networks. The FPGA fabric-Transceiver interface clocks can be subdivided into the following three categories: 1 Input Reference Clocks--Refer to "Input Reference Clock Source" on page 2-3. Transceiver Datapath Interface Clocks--are used to transfer data, control, and status signals between the FPGA fabric and the transceiver channels. The transceiver channel forwards the tx_clkout signal (in non-bonded modes) or the coreclkout signal (in bonded channel modes) to the FPGA fabric to clock the data and control signals into the transmitter. The transceiver channel also forwards the recovered rx_clkout clock (in configurations without rate matcher) or tx_clkout/coreclkout (in configurations with rate matcher) to the FPGA fabric to clock the data and status signals from the receiver into the FPGA fabric. Other Transceiver Clocks--The following transceiver clocks form a part of the FPGA fabric-Transceiver interface clocks: cal_blk_clk--calibration block clock fixed_clk--125 MHz fixed-rate clock used in the PCIe receiver detect circuitry and for the adaptive equalization (AEQ) block reconfig_clk--clock used for transceiver dynamic reconfiguration (for more information, refer to Table 2-5 on page 2-9) In Basic (PMA Direct) functional mode, only tx_clkout and rx_clkout are available to clock the logic in the core. In bonded mode, you may use tx_clkout of one of the channels to clock all of the channels. For receivers in bonded mode, you must use separate rx_clkout for each channel. Table 2-14 lists the FPGA fabric-Transceiver interface clocks. Table 2-14. FPGA Fabric-Transceiver Interface Clocks (Note 1) (Part 1 of 2) Clock Name Clock Description Interface Direction FPGA Fabric Clock Resource Utilization (1) pll_inclk CMU PLL input reference clock when driven from an FPGA CLK input pin FPGA fabric-to-transceiver Global clock rx_cruclk Receiver CDR input reference clock when driven from an FPGA CLK input pin FPGA fabric-to-transceiver Global clock, Regional clock tx_clkout Phase compensation FIFO clock Transceiver-to-FPGA fabric Global clock, Regional clock, Periphery clock coreclkout Phase compensation FIFO clock Transceiver-to-FPGA fabric Global clock, Regional clock, Periphery clock rx_clkout Phase compensation FIFO clock Transceiver-to-FPGA fabric Global clock, Regional clock, Periphery clock fixed_clk PCIe receiver detect clock FPGA fabric-to-transceiver Global clock, Regional clock September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-52 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Table 2-14. FPGA Fabric-Transceiver Interface Clocks (Note 1) (Part 2 of 2) Clock Name reconfig_clk (2) cal_blk_clk Clock Description Interface Direction FPGA Fabric Clock Resource Utilization (1) Transceiver dynamic reconfiguration clock FPGA fabric-to-transceiver Global clock Transceiver calibration block clock FPGA fabric-to-transceiver Global clock, Regional clock Notes to Table 2-11: (1) For more information about global, regional, and periphery clock resources available in each device, refer to the Clock Networks and PLLs in Stratix IV Devices chapter. (2) Ensure that the reconfig_clk is a free-running clock that is not derived from the transceiver blocks. "FPGA Fabric-Transmitter Interface Clocking" on page 2-52 and "FPGA Fabric-Receiver Interface Clocking" on page 2-61 describe the criteria and methodology to share transmitter and receiver phase compensation FIFO clocks in order to reduce the global, regional, and periphery clock resource usage in your design. FPGA Fabric-Transmitter Interface Clocking The transmitter phase compensation FIFO compensates for the phase difference between the FPGA fabric clock (phase compensation FIFO write clock) and the parallel transmitter PCS clock (phase compensation FIFO read clock). The transmitter phase compensation FIFO write clock forms the FPGA fabric-Transmitter interface clock. The phase compensation FIFO write clock and read clocks must have exactly the same frequency (0 parts-per-million [PPM] frequency difference). Stratix IV transceivers provide the following two options for selecting the transmitter phase compensation FIFO write clock: 1 "Quartus II-Selected Transmitter Phase Compensation FIFO Write Clock" "User-Selected Transmitter Phase Compensation FIFO Write Clock" on page 2-58 User-selection is provided to share transceiver datapath interface clocks in order to reduce the global, regional, and periphery clock resource usage in your design. Quartus II-Selected Transmitter Phase Compensation FIFO Write Clock If you do not select the tx_coreclk port in the ALTGX MegaWizardTM Plug-In Manager, the Quartus II software automatically selects the transmitter phase compensation FIFO write clock for each channel in that ALTGX instance. The Quartus II software selects the FIFO write clock depending on the channel configuration. Non-Bonded Channel Configuration In a non-bonded channel configuration, the transmitter channels may or may not be identical. Identical transmitter channels are defined as channels that have exactly the same CMU PLL input reference clock source, exactly the same CMU PLL configuration, and exactly the same transmitter PMA and PCS configuration. 1 Stratix IV Device Handbook Volume 2: Transceivers Identical transmitter channels may have different transmitter voltage output differential (VOD), transmitter common mode voltage (VCM), or pre-emphasis setting. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-53 Example 3: Two Groups of Two Identical Channels in a Transceiver Block Example 3 assumes channels 0 and 1, driven by CMU0_PLL in a transceiver block, are identical. Also, channels 2 and 3, driven by CMU1_PLL in the same transceiver block, are identical. In this case, the Quartus II software automatically drives the write port of the transmitter phase compensation FIFO in channels 0 and 1 with the tx_clkout[0] signal. It also drives the write port of the transmitter phase compensation FIFO in channels 2 and 3 with the tx_clkout[2] signal. Use the tx_clkout[0] signal to clock the transmitter data and control logic for channels 0 and 1 in the FPGA fabric. Use the tx_clkout[2] signal to clock the transmitter data and control logic for channels 2 and 3 in the FPGA fabric. 1 September 2012 This configuration uses two FPGA global and/or regional clock resources, one for the tx_clkout[0] signal and the other for the tx_clkout[2] signal. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-54 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Figure 2-29 shows the FPGA fabric-Transmitter interface clocking for Example 3. Figure 2-29. FPGA Fabric-Transmitter Interface Clocking for Example 3 (1) Channel 3 Transmitter Channel PCS Transmitter Channel PMA TX Phase Compensation FIFO Channel 3 TX Data and Control Logic wrclk rdclk tx_coreclk[3] Local Clock Divider Block /2 Low-Speed Parallel Clock Channel 2 Transmitter Channel PCS Transmitter Channel PMA TX Phase Compensation FIFO Channel 2 TX Data and Control Logic wrclk rdclk tx_coreclk[2] Local Clock Divider Block /2 Low-Speed Parallel Clock tx_clkout[2] FPGA Fabric Input Reference Clock CMU1 PLL Input Reference Clock CMU0 PLL High-Speed Serial Clock CMU1 Clock DIvider CMU1 Channel CMU0 Clock DIvider CMU0 Channel Channel 1 Transmitter Channel PCS Transmitter Channel PMA TX Phase Compensation FIFO Channel 1 TX Data and Control Logic wrclk rdclk tx_coreclk[1] Local Clock Divider Block /2 Low-Speed Parallel Clock Channel 0 Transmitter Channel PCS Transmitter Channel PMA TX Phase Compensation FIFO Channel 0 TX Data and Control Logic wrclk tx_coreclk[0] rdclk /2 Low-Speed Parallel Clock Local Clock Divider Block tx_clkout[0] Note to Figure 2-29: (1) The green lines represent the low-speed parallel clock and the blue lines represent the high-speed serial clock. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-55 Bonded Channel Configuration In x4 and x8 bonded channel configurations, all channels within the transceiver block are identical. The Quartus II software automatically drives the write port of the transmitter phase compensation FIFO in all channels with the coreclkout signal. Use the coreclkout signal to clock the transmitter data and control logic for all four channels in the FPGA fabric. Figure 2-30 shows the FPGA fabric-Transmitter interface clocking in a x4 bonded channel configuration. Figure 2-30. FPGA Fabric-Transmitter Interface Clocking in a x4 Bonded Channel Configuration (1) Channel 3 Transmitter Channel PCS TX Phase Compensation FIFO Channel 3 TX Data and Control Logic wrclk rdclk tx_coreclk[3] /2 Parallel PCS Clock Channel 2 Transmitter Channel PCS TX Phase Compensation FIFO Channel 2 TX Data and Control Logic wrclk rdclk tx_coreclk[2] /2 Parallel PCS Clock CMU1 Channel CMU1 PLL FPGA Fabric coreclkout /2 Input Reference Clock CMU0 Channel CMU0 Clock Divider CMU0 PLL Channel 1 Transmitter Channel PCS Channel 1 TX Data and Control Logic TX Phase Compensation FIFO wrclk rdclk tx_coreclk[1] /2 Parallel PCS Clock Channel 0 Transmitter Channel PCS Channel 0 TX Data and Control Logic tx_coreclk[0] TX Phase Compensation FIFO wrclk rdclk /2 Parallel PCS Clock Note to Figure 2-30: (1) The green lines represent the parallel PCS clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-56 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Limitations of the Quartus II Software-Selected Transmitter Phase Compensation FIFO Write Clock The Quartus II software uses a single tx_clkout signal to clock the transmitter phase compensation FIFO write port of all identical channels within a transceiver block. This results in one global and/or regional clock resource being used for each group of identical channels within a transceiver block. For identical channels located across the transceiver blocks, the Quartus II software does not use a single tx_clkout signal to clock the write port of the transmitter phase compensation FIFOs for all channels. It uses one tx_clkout signal for each group of identical channels per transceiver block. This results in higher global and regional clock resource usage. Example 4: Sixteen Identical Channels Across Four Transceiver Blocks Figure 2-31 shows 16 identical transmitter channels located across four transceiver blocks. The Quartus II software uses tx_clkout from Channel 0 in each transceiver block to clock the write port of the transmitter phase compensation FIFO in all four channels in that transceiver block. This results in four global and/or regional clock resources being used, one for each transceiver block. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-57 Figure 2-31. Sixteen Identical Channels Across Four Transceiver Blocks for Example 4 (1) Transceiver Block GXBR3 Channel [15:12] TX Data tx_coreclk[15:12] and Control Logic Channel 3 Channel 2 Channel 1 Channel 0 tx_clkout[12] Transceiver Block GXBR2 Channel [11:8] TX Data and Control Logic Channel 3 tx_coreclk[11:8] Channel 2 Channel 1 Channel 0 tx_clkout[8] FPGA Fabric Transceiver Block GXBR1 Channel [7:4] TX Data and Control Logic Channel 3 tx_coreclk[7:4] Channel 2 Channel 1 Channel 0 tx_clkout[4] Transceiver Block GXBR0 Channel [3:0] TX Data and Control Logic Channel 3 tx_coreclk[3:0] Channel 2 Channel 1 Channel 0 tx_clkout[0] Note to Figure 2-31: (1) The red lines represent tx_clkout[12], the blue lines represent tx_clkout[8], the green lines represent tx_clkout[4], and the brown lines represent tx_clkout[0]. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-58 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Because all 16 channels are identical, using a single tx_clkout to clock the transmitter phase compensation FIFO in all 16 channels results in only one global or regional clock resource being used instead of four. To achieve this, you must choose the transmitter phase compensation FIFO write clocks instead of the Quartus II software automatic selection, as described in "User-Selected Transmitter Phase Compensation FIFO Write Clock" on page 2-58. User-Selected Transmitter Phase Compensation FIFO Write Clock The ALTGX MegaWizard Plug-In Manager provides an optional port named tx_coreclk for each instantiated transmitter channel. If you enable this port, the Quartus II software does not automatically select the transmitter phase compensation FIFO write clock source. Instead, the signal that you drive on the tx_coreclk port of the channel clocks the write side of its transmitter phase compensation FIFO. Use the flexibility of selecting the transmitter phase compensation FIFO write clock to reduce global and regional clock resource usage. You can connect the tx_coreclk ports of all identical channels in your design and drive them using a common clock driver that has 0 PPM frequency difference with respect to the FIFO read clocks of these channels. Use the common clock driver to clock the transmitter data and control logic in the FPGA fabric for all identical channels. This FPGA fabric-Transceiver interface clocking scheme uses only one global or regional clock resource for all identical channels in your design. Example 5: Sixteen Identical Channels Across Four Transceiver Blocks Figure 2-32 shows 16 identical transmitter channels located across four transceiver blocks. The tx_coreclk ports of all 16 transmitter channels are connected together and driven by a common clock driver. This common clock driver also drives the transmitter data and control logic of all 16 transmitter channels in the FPGA fabric. You use only one global or regional clock resource with this clocking scheme, compared to four global and regional clock resources needed without the tx_coreclk ports (the Quartus II software-selected transmitter phase compensation FIFO write clock). Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-59 Figure 2-32. Sixteen Identical Channels Across Four Transceiver Blocks for Example 5 Common Clock Driver Transceiver Block GXBR3 Channel [15:12] TX Data and Control Logic Channel 3 tx_coreclk[15:12] Channel 2 Channel 1 Channel 0 tx_clkout[12] Transceiver Block GXBR2 Channel [11:8] TX Data and Control Logic Channel 3 tx_coreclk[11:8] Channel 2 Channel 1 Channel 0 FPGA Fabric tx_clkout[8] Transceiver Block GXBR1 Channel [7:4] TX Data and Control Logic Channel 3 tx_coreclk[7:4] Channel 2 Channel 1 Channel 0 tx_clkout[4] Transceiver Block GXBR0 Channel [3:0] TX Data and Control Logic Channel 3 tx_coreclk[3:0] Channel 2 Channel 1 Channel 0 tx_clkout[0] Common Clock Driver Selection Rules The common clock driver driving the tx_coreclk ports of all identical channels must have 0 PPM frequency difference with respect to the transmitter phase compensation FIFO read clocks of these channels. If there is any frequency difference between the FIFO write clock (tx_coreclk) and the FIFO read clock, the FIFO overflows or under-runs, resulting in corrupted data transfer between the FPGA fabric and the transmitter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-60 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Table 2-15 lists the transmitter phase compensation FIFO read clocks that the Quartus II software selects in various configurations. Table 2-15. Transmitter Phase Compensation FIFO Read Clocks Transmitter Phase Compensation FIFO Read Clock Configuration Without Byte Serializer Parallel transmitter PCS clock from the local Non-Bonded Channel clock divider in the associated channel Configuration (tx_clkout) With Byte Serializer Divide-by-two version of the parallel transmitter PCS clock from the local clock divider in the associated channel (tx_clkout) Divide-by-two version of the low-speed parallel clock from the CMU0 clock divider of the associated transceiver block (coreclkout) x4 Bonded Channel Configuration Low-speed parallel clock from the CMU0 clock divider of the associated transceiver block (coreclkout) x8 Bonded Channel Configuration Divide-by-two version of the low-speed parallel Low-speed parallel clock from the CMU0 clock from the CMU0 clock divider of the master clock divider of the master transceiver block transceiver block (coreclkout from master (coreclkout from master transceiver block) transceiver block) To ensure that you understand the 0 PPM clock driver rule, the Quartus II software expects the following set of user assignments whenever you use the tx_coreclk port to drive the transmitter phase compensation FIFO write clock: 1 GXB 0 PPM Core Clock Setting Failing to make this assignment correctly when using the tx_coreclk port results in a Quartus II compilation error. The GXB 0 PPM core clock setting allows the following clock drivers to drive the tx_coreclk ports: 1 tx_clkout in non-bonded channel configurations coreclkout in bonded channel configurations FPGA_CLK input pins Transceiver refclk pins Clock output from left and right and top and bottom PLLs (PLL_L, PLL_R, and PLL_T, PLL_B) The Quartus II software does not allow gated clocks or clocks generated in FPGA logic to drive the tx_coreclk ports. Because the GXB 0 PPM core clock setting allows the FPGA CLK input pins and transceiver refclk pins as the clock driver, the Quartus II compiler cannot determine if there is a 0 PPM difference between the FIFO write clock and read clock for each channel. 1 Stratix IV Device Handbook Volume 2: Transceivers You must ensure that the clock driver for all connected tx_coreclk ports has a 0 PPM difference with respect to the FIFO read clock in those channels. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-61 Table 2-16 lists the Quartus II assignments that you must make in the assignment editor. Table 2-16. Quartus II Assignments Full design hierarchy name of one of the following clock drivers that you choose to drive the tx_coreclk ports of all identical channels (1): From tx_clkout coreclkout FPGA CLK input pins Transceiver refclk pins Clock output from the left and right or top and bottom PLLs tx_dataout port of one of the identical channels To tx_dataout pins of all identical channels whose tx_coreclk ports are connected together and driven by the 0 PPM clock driver. Assignment Name GXB 0 PPM Core Clock Setting Value ON Note to Table 2-13: (1) You can find the full hierarchy name of the 0 PPM clock driver using the Node Finder feature in the Quartus II Assignment Editor. For more implementation information, refer to "Configuration Example 2: Configuring Sixteen Identical Channels Across Four Transceiver Blocks" on page 2-75. Basic (PMA Direct) mode In Basic (PMA Direct) mode, each channel must be clocked by its own tx_clkout. As a result, the number of global and/or regional clock resources required is significantly higher. In Basic (PMA Direct) xN mode, to save on global and/or regional clock resources, you may use tx_clkout from centrally located channels to clock all the channels. The coreclkout port is not available in Basic (PMA Direct) xN mode. FPGA Fabric-Receiver Interface Clocking The receiver phase compensation FIFO compensates for the phase difference between the parallel receiver PCS clock (FIFO write clock) and the FPGA fabric clock (FIFO read clock). The receiver phase compensation FIFO read clock forms the FPGA fabric-Receiver interface clock. The FIFO write clock and read clock must have exactly the same frequency (0 PPM frequency difference). Stratix IV transceivers provide the following two options for selecting the receiver phase compensation FIFO read clock: 1 September 2012 "Quartus II Software-Selected Receiver Phase Compensation FIFO Read Clock" on page 2-62 "User-Selected Receiver Phase Compensation FIFO Read Clock" on page 2-68 User-selection is provided to share transceiver datapath interface clocks in order to reduce the global, regional, and periphery clock resource usage in your design. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-62 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Quartus II Software-Selected Receiver Phase Compensation FIFO Read Clock If you do not select the rx_coreclk port in the ALTGX MegaWizard Plug-In Manager, the Quartus II software automatically selects the receiver phase compensation FIFO read clock for each channel in that ALTGX instance. The Quartus II software selects the FIFO read clock depending on the channel configuration. In non-bonded channel configurations, the FPGA fabric-receiver interface clocking has two scenarios: Receivers that do not use a rate matcher block (refer to "Non-Bonded Receiver Clocking Without Rate Matcher" on page 2-39) Receivers that use a rate matcher block (refer to "Non-Bonded Receiver Clocking with Rate Matcher" on page 2-41) Non-Bonded Channel Configuration with Rate Matcher In non-bonded channel configuration, the transceiver channels may or may not be identical. Identical transceiver channels are defined as channels that have exactly the same CMU PLL and receiver CDR input reference clock sources, exactly the same CMU PLL and receiver CDR configuration, and exactly the same PMA and PCS configuration. Example 6: Two Groups of Two Identical Channels in a Transceiver Block Example 6 assumes channels 0 and 1, driven by the CMU0 PLL in a transceiver block, are identical. Also, channels 2 and 3, driven by the CMU1 PLL in the same transceiver block, are identical. In this case, the Quartus II software automatically drives the read port of the receiver phase compensation FIFO in channels 0 and 1 with the tx_clkout[0] signal. It also drives the read port of the receiver phase compensation FIFO in channels 2 and 3 with the tx_clkout[2] signal. Use the tx_clkout[0] signal to latch the receiver data and status signals from channels 0 and 1 in the FPGA fabric. Use the tx_clkout[2] signal to latch the receiver data and status signals from channels 2 and 3 in the FPGA fabric. 1 Stratix IV Device Handbook Volume 2: Transceivers This configuration uses two FPGA global and/or regional clock resources, one for the tx_clkout[0] signal and the other for the tx_clkout[2] signal. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-63 Figure 2-33 shows the FPGA fabric-Receiver interface clocking for Example 6. Figure 2-33. FPGA Fabric-Receiver Interface Clocking for Example 6 (1) Channel 3 Receiver Channel PCS Receiver Channel PMA RX Phase Compensation FIFO Channel 3 RX Data and Status Logic rdclk wrclk rx_coreclk[3] Transmitter Channel PMA Local Clock Divider Block /2 Low-Speed Parallel Clock Channel 2 Receiver Channel PCS Receiver Channel PMA RX Phase Compensation FIFO Channel 2 RX Data and Status Logic rdclk wrclk Transmitter Channel PMA rx_coreclk[2] Local Clock Divider Block /2 Low-Speed Parallel Clock tx_clkout[2] Reference Clock CMU1 Channel CMU1 PLL High-Speed Serial Clock FPGA Fabric Reference Clock CMU0 Channel CMU0 PLL Channel 1 Receiver Channel PCS Receiver Channel PMA RX Phase Compensation FIFO Channel 1 RX Data and Status Logic rdclk wrclk Transmitter Channel PMA rx_coreclk[1] Local Clock Divider Block /2 Low-Speed Parallel Clock Channel 0 Receiver Channel PCS Receiver Channel PMA RX Phase Compensation FIFO Channel 0 RX Data and Status Logic rdclk rx_coreclk[0] wrclk Transmitter Channel PMA /2 Low-Speed Parallel Clock Local Clock Divider Block tx_clkout[0] Note to Figure 2-33: (1) The green lines represent the low-speed parallel clock and the blue lines represent the high-speed serial clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-64 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Non-Bonded Channel Configuration Without Rate Matcher In non-bonded channel configuration without rate matcher, the Quartus II software cannot determine if the incoming serial data in all channels have a 0 PPM frequency difference. The Quartus II software automatically drives the read port of the receiver phase compensation FIFO in each channel with the recovered clock driven on the rx_clkout port of that channel. Use the rx_clkout signal from each channel to latch its receiver data and status signals in the FPGA fabric. 1 Stratix IV Device Handbook Volume 2: Transceivers This configuration uses one FPGA global, regional clock, or both, resource per channel for the rx_clkout signal. September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-65 Figure 2-34 shows the FPGA fabric-Receiver interface clocking for non-bonded channel configurations without rate matcher. Figure 2-34. FPGA Fabric-Receiver Interface Clocking for Non-Bonded Channel Configurations Without Rate Matcher (1) Channel 3 Receiver Channel PCS Receiver Channel PMA RX Phase Compensation FIFO Channel 3 RX Data and Status Logic rdclk wrclk rx_coreclk[3] /2 rx_datain[3] CDR Parallel Recovered Clock rx_clkout[3] Channel 2 Receiver Channel PCS Receiver PMA Receiver Channel PMA RX Phase Compensation FIFO Channel 2 RX Data and Status Logic rdclk wrclk Input Reference Clock rx_coreclk[2] /2 CDR Parallel Recovered Clock rx_datain[2] rx_clkout[2] FPGA Fabric Channel 1 Receiver Channel PCS Receiver PMA Receiver Channel PMA RX Phase Compensation FIFO Channel 1 RX Data and Status Logic rdclk wrclk Input Reference Clock rx_coreclk[1] /2 CDR rx_datain[1] Parallel Recovered Clock rx_clkout[1] Channel 0 Receiver Channel PCS Receiver PMA Receiver Channel PMA RX Phase Compensation FIFO Channel 0 RX Data and Status Logic rdclk wrclk Input Reference Clock rx_coreclk[0] /2 Parallel Recovered Clock CDR rx_datain[0] rx_clkout[0] Note to Figure 2-34: (1) The red lines represent rx_clkout[3], the blue lines represent rx_clkout[2], the green lines represent rx_clkout[1], and the brown lines represent rx_clkout[0]. Bonded Channel Configuration All bonded transceiver channel configurations have rate matcher in the receiver data path. In x4 and x8 bonded channel configurations, the Quartus II software automatically drives the read port of the receiver phase compensation FIFO in all channels with the coreclkout signal (from the master transceiver block in the case of x8 bonded mode). Use the coreclkout signal to latch the receiver data and status signals from all channels in the FPGA fabric. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-66 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 1 This configuration uses one FPGA global and/or regional clock resource per bonded link for the coreclkout signal. Figure 2-35 shows the FPGA fabric-Receiver interface clocking in x4 bonded channel configuration. Figure 2-35. FPGA Fabric-Receiver Interface Clocking in a x4 Bonded Channel Configuration (1) Channel 3 Transmitter Channel PCS RX RXPhase Phase Compensation Compensation FIFO FIFO Channel 3 RX Data and Status Logic rdclk rdclk wrclk wrclk rx_coreclk[3] /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 2 Transmitter Channel PCS RX Phase Compensation FIFO Channel 2 RX Data and Status Logic rdclk wrclk rx_coreclk[2] /2 Low-Speed Parallel Clock from CMU0 Clock Divider CMU1 Channel CMU1 PLL FPGA Fabric coreclkout /2 Reference Clock CMU0 Channel CMU0 Clock Divider CMU0 PLL Channel 1 Transmitter Channel PCS Channel 1 RX Data and Status Logic RX Phase Compensation FIFO rdclk wrclk rx_coreclk[1] /2 Low-Speed Parallel Clock from CMU0 Clock Divider Channel 0 Transmitter Channel PCS Channel 0 RX Data and Status Logic RX Phase Compensation FIFO rdclk wrclk rx_coreclk[0] /2 Low-Speed Parallel Clock from CMU0 Clock Divider Note to Figure 2-35: (1) The green lines represent low-speed parallel clock from the CMU0 clock divider. Limitations of the Quartus II Software-Selected Receiver Phase Compensation FIFO Read Clock In non-bonded channel configurations without rate matcher, the Quartus II software cannot determine if the incoming serial data in all channels has a 0 PPM frequency difference. The Quartus II software uses the recovered clock rx_clkout signal from each channel to clock the read port of its receiver phase compensation FIFO. This results in one global, regional, or global and regional clock resource being used per channel for the rx_clkout signal. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-67 Example 7: Sixteen Channels Across Four Transceiver Blocks Figure 2-36 shows 16 non-bonded receiver channels without rate matcher, located across four transceiver blocks. The incoming serial data to all 16 channels have a 0 PPM frequency difference with respect to each other. The Quartus II software uses rx_clkout from each channel to clock the read port of its receiver phase compensation FIFO. This results in 16 global, regional, or global and regional clock resources being used, one for each channel. Figure 2-36. Sixteen Non-Bonded Receiver Channels without Rate Match for Example 7 Transceiver Block GXBR3 rx_coreclk[15] Channel 3 Channel [15:12] RX Data and Status Logic rx_coreclk[14] rx_clkout[15] rx_coreclk[13] rx_clkout[14] rx_coreclk[12] rx_clkout[13] Channel 2 Channel 1 Channel 0 rx_clkout[12] Transceiver Block GXBR2 rx_coreclk[11] Channel 3 Channel [11:8] RX Data and Status Logic rx_coreclk[10] rx_clkout[11] rx_coreclk[9] rx_clkout[10] rx_coreclk[8] rx_clkout[9] Channel 2 Channel 1 Channel 0 rx_clkout[8] FPGA Fabric Transceiver Block GXBR1 rx_coreclk[7] Channel 3 Channel [7:4] RX Data and Status Logic rx_coreclk[6] rx_clkout[7] Channel 2 rx_coreclk[5] rx_clkout[6] rx_coreclk[4] rx_clkout[5] Channel 1 Channel 0 rx_clkout[4] Transceiver Block GXBR0 rx_coreclk[3] Channel 3 Channel [3:0] RX Data and Status Logic rx_coreclk[2] rx_clkout[3] rx_coreclk[1] rx_clkout[2] rx_coreclk[0] rx_clkout[1] Channel 2 Channel 1 Channel 0 rx_clkout[0] September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-68 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Because the recovered clock rx_clkout signals from all 16 channels have a 0 PPM frequency difference, you can use a single rx_clkout to clock the receiver phase compensation FIFO in all 16 channels. This results in only one global, regional, or global and regional clock resource being used instead of 16. To achieve this, you must select the receiver phase compensation FIFO read clocks instead of the Quartus II software default selection, as described in "User-Selected Receiver Phase Compensation FIFO Read Clock" on page 2-68. User-Selected Receiver Phase Compensation FIFO Read Clock The ALTGX MegaWizard Plug-In Manager provides an optional port named rx_coreclk for each instantiated receiver channel. If you enable this port, the Quartus II software does not automatically select the receiver phase compensation FIFO read clock source. Instead, the signal that you drive on the rx_coreclk port of the channel clocks the read side of its receiver phase compensation FIFO. You can use the flexibility of selecting the receiver phase compensation FIFO read clock to reduce the global, regional, or global and regional clock resource usage. You can connect the rx_coreclk ports of all the receiver channels in your design and drive them using a common clock driver that has a 0 PPM frequency difference with respect to the FIFO write clocks of these channels. Use this common clock driver to latch the receiver data and status signals in the FPGA fabric for these channels. This FPGA fabric-Transceiver interface clocking scheme uses only one global, regional, or global and regional clock resource for all channels. Example 8: Sixteen Identical Channels Across Four Transceiver Blocks Figure 2-37 shows 16 channels located across four transceiver blocks. The incoming serial data to all 16 channels has a 0 PPM frequency difference with respect to each other. The rx_coreclk ports of all 16 channels are connected together and driven by a common clock driver. This common clock driver also latches the receiver data and status logic of all 16 receiver channels in the FPGA fabric. Only one global, regional, or global and regional clock resource is used with this clocking scheme, compared to 16 global, regional, or global and regional clock resources needed without the rx_coreclk ports (the Quartus II software-selected receiver phase compensation FIFO read clock). Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking 2-69 Figure 2-37. Sixteen Identical Channels Across Four Transceiver Blocks for Example 8 Common Clock Driver Transceiver Block GXBR3 Channel [15:12] RX Data and Status Logic Channel 3 rx_coreclk[15:12] Channel 2 Channel 1 rx_clkout[15:12] Channel 0 Transceiver Block GXBR2 Channel [11:8] RX Data and Status Logic Channel 3 rx_coreclk[11:8] Channel 2 Channel 1 rx_clkout[11:8] Channel 0 FPGA Fabric Transceiver Block GXBR1 Channel [7:4] RX Data and Status Logic Channel 3 rx_coreclk[7:4] Channel 2 Channel 1 rx_clkout[7:4] Channel 0 Transceiver Block GXBR0 Channel [3:0] RX Data and Status Logic Channel 3 rx_coreclk[3:0] Channel 2 Channel 1 rx_clkout[3:0] Channel 0 Common Clock Driver Selection Rules The common clock driver driving the rx_coreclk ports of all channels must have a 0 PPM frequency difference with respect to the receiver phase compensation FIFO write clocks of these channels. If there is any frequency difference between the FIFO read clock (rx_coreclk) and the FIFO write clock, the FIFO overflows or under-runs, resulting in corrupted data transfer between the FPGA fabric and the receiver. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-70 Chapter 2: Transceiver Clocking in Stratix IV Devices FPGA Fabric-Transceiver Interface Clocking Table 2-17 lists the receiver phase compensation FIFO write clocks that the Quartus II software selects in various configurations. Table 2-17. Receiver Phase Compensation FIFO Write Clocks Receiver Phase Compensation FIFO Write Clock Configuration Without Byte Serializer With Byte Serializer Non-Bonded Channel Configuration with rate matcher Low-speed parallel clock from the local clock divider in the associated channel (tx_clkout) Divide-by-two version of the low-speed parallel clock from the local clock divider in the associated channel (tx_clkout) Non-Bonded Channel Configuration without rate matcher Parallel recovered clock from the receiver PMA in the associated channel (rx_clkout) Divide-by-two version of the parallel recovered clock from the receiver PMA in the associated channel (rx_clkout) x4-Bonded Channel Configuration Low-speed parallel clock from the CMU0 clock divider of the associated transceiver block (coreclkout) Divide-by-two version of the low-speed parallel clock from the CMU0 clock divider of the associated transceiver block (coreclkout) x8-Bonded Channel Configuration Low-speed parallel clock from the CMU0 clock divider of the master transceiver block (coreclkout from the master transceiver block) Divide-by-two version of the low-speed parallel clock from the CMU0 clock divider of the master transceiver block (coreclkout from the master transceiver block) To ensure that you understand the 0 PPM clock driver rule, the Quartus II software expects the following user assignment whenever you use the rx_coreclk port to drive the receiver phase compensation FIFO read clock: 1 GXB 0 PPM Core Clock Setting Failing to make this assignment correctly when using the rx_coreclk port results in a Quartus II compilation error. The GXB 0 PPM core clock setting user assignment allows the following clock drivers to drive the rx_coreclk ports: 1 tx_clkout in non-bonded channel configurations with rate matcher tx_clkout and rx_clkout in non-bonded configurations without rate matcher coreclkout in bonded channel configurations FPGA CLK input pins Transceiver refclk pins Clock output from left and right and top and bottom PLLs (PLL_L, PLL_R, and PLL_T, PLL_B) The Quartus II software does not allow gated clocks or clocks generated in FPGA logic to drive the tx_coreclk ports. Because the 0 PPM clock group assignment allows the FPGA CLK input pins and transceiver refclk pins as the clock driver, the Quartus II compiler cannot determine if there is a 0 PPM difference between the FIFO write clock and read clock for each channel. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Using the CMU/ATX PLL for Clocking User Logic in the FPGA Fabric 1 2-71 You must ensure that the clock driver for all the connected rx_coreclk ports has a 0 PPM difference with respect to the FIFO write clock in those channels. Table 2-18 lists the Quartus II assignments that you must make. Table 2-18. Quartus II Assignments Full design hierarchy name of one of the following clock drivers that you choose to drive the rx_coreclk ports of all identical channels (1): From tx_clkout rx_clkout coreclkout FPGA CLK input pins Transceiver refclk pins Clock output from the left and right or top and bottom PLLs tx_dataout port of one of the identical channels To rx_datain pins of all channels whose rx_coreclk ports are connected together and driven by the 0 PPM clock driver. Assignment Name GXB 0 PPM Core Clock Setting Value ON Note to Table 2-18: (1) You can find the full hierarchy name of the 0 PPM clock driver using the Node Finder feature in the Quartus II Assignment Editor. For more implementation details, refer to "Configuration Example 3: Configuring Sixteen Channels Across Four Transceiver Blocks" on page 2-76. Basic (PMA Direct) Mode In Basic (PMA Direct) mode, each channel must be clocked by its own rx_clkout. As a result, the number of global and/or regional clock resources required is significantly higher. Bonding is not supported for receivers configured in Basic (PMA Direct) functional mode. Using the CMU/ATX PLL for Clocking User Logic in the FPGA Fabric Some designs that use multiple clock domains may run out of PLLs in the FPGA fabric. In such a scenario, if your design has CMU or ATX PLLs that are not being used, it may be possible to use them for clocking user logic in the FPGA fabric. However, the CMU PLLs and ATX PLLs do not have many features that are supported by the PLLs in the FPGA fabric. The following are the supported features on CMU PLLs and ATX PLLs used as PLLs for clocking user logic in the FPGA fabric: September 2012 Single clock output Programmable PLL bandwidth PLL PFD power down control Lock status signal Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-72 Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples To use this feature, you must create an ALTGX instance with a single channel in Transmitter Only mode that uses the required CMU PLL or ATX PLL. To create the ALTGX instance, follow these steps: 1. Choose Basic (PMA Direct) xN mode as the protocol. 2. Select Transmitter Only operation mode. 3. Select the input clock frequency. 4. Select the appropriate values of data rate and channel width based on the desired output clock frequency. To generate a 250 MHz clock using an input clock frequency of 50 MHz, select a channel width of 10 and a data rate of 2500 Mbps (Equation 2-1). Equation 2-1. f out = data rate channel width 5. You can select the PLL bandwidth by choosing Tx PLL bandwidth mode. 6. You can instantiate the pll_locked port to indicate the PLL lock status. 7. You can instantiate pll_powerdown or gxb_powerdown to enable the PLL PFD power down control. 8. Use tx_clkout of the ALTGX instance as the clock source for clocking user logic in the FPGA fabric. Configuration Examples This section describes the following examples: Stratix IV Device Handbook Volume 2: Transceivers "Configuration Example 1: Configuring 24 Channels in Basic (PMA Direct) xN Mode in the EP4S100G5F45 Device" on page 2-73 "Configuration Example 2: Configuring Sixteen Identical Channels Across Four Transceiver Blocks" on page 2-75 "Configuration Example 3: Configuring Sixteen Channels Across Four Transceiver Blocks" on page 2-76 "Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode" on page 2-78 September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 2-73 Configuration Example 1: Configuring 24 Channels in Basic (PMA Direct) xN Mode in the EP4S100G5F45 Device Each transceiver block has four regular channels and two CMU channels that you can configure in Basic (PMA Direct) xN mode. The EP4S100G5F45 device has four transceiver blocks located on each side of the device allowing configuration of up to 24 channels in Basic (PMA Direct) xN mode. When all 24 channels on one side of the device are configured in Basic (PMA Direct) xN mode, all eight CMU channels (two in each transceiver block) are configured as PMA-Only channels. Use the refclk pins in each of the four transceiver blocks as receiver serial data input pins and configure the CMU PLLs as receiver CDRs when the CMU channel is configured as a PMA-Only channel. Due to the non-availability of CMU PLLs, you must use the 6G ATX PLL to generate the high-speed serial and low-speed parallel transceiver clocks for all 24 channels. Due to the non-availability of a refclk pin, you must use the left and right, or left or right PLL in VCO bypass mode to provide the reference clock through the PLL cascade clock line. For more information about left and right PLL VCO bypass mode, refer to "Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode" on page 2-78. Figure 2-38 shows 24 channels on the right side of the EP4S100G5F45 device configured in Basic (PMA Direct) xN mode running at 6.5 Gbps with a 20-bit FPGA fabric-PMA interface width. Because all 24 channels on the right side of the device are configured in Basic (PMA Direct) xN mode, the right PLL_R1 configured in VCO bypass mode is used to provide the input reference clock to the 6G ATX PLL. The 6G ATX PLL generates the high-speed serial and low-speed parallel transceiver clocks that are distributed to the 24 channels though the xN_Top and xN_Bottom clock network. Because the data rate of 6.5 Gbps requires a left and right, or left or right PLL to meet FPGA fabric-Transmitter PMA interface timing, tx_clkout from one of the 24 channels is phase shifted by 315 using PLL_R2. The phase shifted output clock from PLL_R2 is used to clock the FPGA fabric logic that generates the transmitter parallel data and control signals. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-74 Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples Figure 2-38. Twenty-Four Channels on the Right Side of the EP4S100G5F45 Device Configured in Basic (PMA Direct) xN Mode for Configuration Example 1 (1) PLL Cascade Clock Line xN_Bottom Transceiver Block GXBR3 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 FPGA Fabric (Transmitter Data Generation Logic) ATX PLL R2 (10G) Transceiver Block GXBR2 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 Dedicated FPGA CLK Pin PLL_R1 (VCO Bypass Mode) Reference Clock PLL_R2 (Phase Shift 315 to meet interface timing) ATX PLL R1 (6G) tx_clkout Transceiver Block GXBR1 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 FPGA Fabric (Transmitter Data Generation Logic) ATX PLL R0 (6G) Transceiver Block GXBR0 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 xN_Top Note to Figure 2-38: (1) The green line represents the PLL cascade clock line and the blue lines represent the 6G ATX PLL block. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 2-75 Configuration Example 2: Configuring Sixteen Identical Channels Across Four Transceiver Blocks 1 This example relates to "User-Selected Receiver Phase Compensation FIFO Read Clock" on page 2-68. Figure 2-39 shows 16 identical transmitter channels located across four transceiver blocks. The tx_coreclk ports of all 16 transmitter channels are connected together and driven by the tx_clkout[4] signal from channel 0 in transceiver block GXBR1. The tx_clkout[4] signal also drives the transmitter data and control logic of all 16 transmitter channels in the FPGA fabric. With this clocking scheme, only one global clock resource is used by the tx_clkout[4] signal. Figure 2-39. Sixteen Identical Channels Across Four Transceiver Blocks for Configuration Example 2 Transceiver Block GXBR3 Channel [15:12] TX Data and Control Logic Channel 3 tx_coreclk[15:12] Channel 2 Channel 1 Channel 0 tx_clkout[12] Transceiver Block GXBR2 Channel [11:8] TX Data and Control Logic Channel 3 tx_coreclk[11:8] Channel 2 Channel 1 Channel 0 FPGA Fabric tx_clkout[8] Transceiver Block GXBR1 Channel [7:4] TX Data and Control Logic Channel 3 tx_coreclk[7:4] Channel 2 Channel 1 Channel 0 tx_clkout[4] Transceiver Block GXBR0 Channel [3:0] TX Data and Control Logic Channel 3 tx_coreclk[3:0] Channel 2 Channel 1 Channel 0 tx_clkout[0] September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-76 Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples Table 2-19 lists the Quartus II assignments that you must make for the clocking scheme shown in Figure 2-38. Table 2-19. Quartus II Assignments From top_level/top_xcvr_instance1/altgx_component/tx_clkout[4] To tx_dataout[15..0] Assignment Name GXB 0 PPM Core Clock Setting Value ON (1) Note to Table 2-19: (1) This is an example design hierarchy path for the tx_clkout[4] signal. Configuration Example 3: Configuring Sixteen Channels Across Four Transceiver Blocks 1 This example relates to "User-Selected Receiver Phase Compensation FIFO Read Clock" on page 2-68. Figure 2-40 shows 16 non-bonded channels without rate matcher located across four transceiver blocks. The incoming serial data to all 16 channels has a 0 PPM frequency difference with respect to each other. The rx_coreclk ports of all 16 channels are connected together and driven by rx_clkout[9] in transceiver block GXBR2. rx_clkout[9] also clocks the receiver data and status signals of all 16 channels in the FPGA fabric. With this clocking scheme, only one global, regional, or global and regional clock resource is used by rx_clkout[9]. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 2-77 Figure 2-40. Sixteen Channels Across Four Transceiver Blocks for Configuration Example 3 Transceiver Block GXBR3 Channel [15:12] RX Data and Status Logic Channel 3 rx_coreclk[15:12] Channel 2 Channel 1 rx_clkout[15:12] Channel 0 Transceiver Block GXBR2 Channel [11:8] RX Data and Status Logic Channel 3 rx_coreclk[11:8] Channel 2 rx_clkout[11:8] Channel 1 Channel 0 rx_clkout[9] FPGA Fabric Transceiver Block GXBR1 Channel [7:4] RX Data and Status Logic Channel 3 rx_coreclk[7:4] Channel 2 rx_clkout[7:4] Channel 1 Channel 0 Transceiver Block GXBR0 Channel [3:0] RX Data and Status Logic Channel 3 rx_coreclk[3:0] Channel 2 rx_clkout[3:0] Channel 1 Channel 0 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-78 Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples Table 2-20 lists the Quartus II assignments that you must make for the clocking scheme shown in Figure 2-40. Table 2-20. Quartus II Assignments for Appendix Example 4 From top_level/top_xcvr_instance1/altgx_component/rx_clkout[9] (1) To rx_datain[15..0] Assignment Name GXB 0 PPM Core Clock Setting Value ON Note to Table 2-20: (1) This is an example design hierarchy path for the rx_clkout[9] signal. Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode 1 This example relates to "Left and Right, Left, or Right PLL in VCO Bypass Mode" on page 2-17. To configure the left and right, left, or right PLL in VCO bypass mode, follow these steps: 1. Under the General/Modes tab, enter the desired input reference clock frequency. a. Under PLL Type, select Left_Right_PLL. b. Under Operation mode, select the With no compensation option (Figure 2-41). Figure 2-41. No Compensation Option Used for Configuration Example 4 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 2-79 2. Under the Inputs/Lock tab, select Create output file(s) using the 'Advanced' PLL parameters (Figure 2-42). Figure 2-42. Create Output File(s) Using the `Advanced' PLL Parameters Option Use for Configuration Example 4 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-80 Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 3. Under the Output Clocks tab turn off Use this clock for clk c0. 4. Turn on Use this clock for clk c1 (Figure 2-43). 1 The VCO bypass option is only enabled for clock output c1. Figure 2-43. Use This Clock Option Used for Configuration Example 4 5. Click Finish for the MegaWizard Plug-In Manager to generate the verilog .v file for the ALTPLL instantiation. 6. Next, from the command line, go to the directory where you have the ALTPLL instance files (.v or .vhdl) and type the following command: qmegawiz -silent -wiz_override="c1_test_source=1,c1_mode=BYPASS,clk1_counter=C1" pll0.v This command places your ALTPLL instance in VCO bypass mode. Revisit the .v or .vhdl file associated with the ALTPLL instance. Examine the file which is automatically updated to incorporate the PLL in a VCO bypass mode. 1 VCO bypass mode is not supported in the .mif file. Therefore, you can not manually modify the .mif file to set the PLL in VCO bypass mode. 7. Finally, connect clk c1output of the left and right, left, or right PLL to the input reference clock port of the ATX PLL used to generate the transceiver clocks. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples 2-81 Document Revision History Table 2-21 lists the revision history for this chapter. Table 2-21. Document Revision History Date Version September 2012 3.4 December 2011 3.3 February 2011 Changes Updated the "Non-Bonded Channel Configurations" section to close FB #65105. Updated Table 2-2. Updated the "Left and Right, Left, or Right PLL in VCO Bypass Mode" section. Updated Table 2-4. Updated Figure 2-7, Figure 2-18, Figure 2-19, Figure 2-20. Updated the "Configuration Example 4: Configuring Left and Right, Left, or Right PLL in VCO Bypass Mode" section. Applied new template. Updated chapter title. Applied new template. Updated Table 2-4. Updated Figure 2-7, Figure 2-8, Figure 2-16, and Figure 2-21. Updated the "Transceiver Channel Datapath Clocking" and "Configuration Example 3: Configuring Sixteen Channels Across Four Transceiver Blocks" sections. Added a note to the "refclk0 and refclk1 Pins" section. Changed "datapath clocks" to "datapath interface clocks". Minor text edits. Added Figure 2-1, Figure 2-12,and Figure 2-13. Added Table 2-1, Table 2-2, Table 2-8, and Table 2-2. Updated Table 2-5 and Table 2-14. Updated all graphics. Updated all sections. Added Stratix IV GT information. Re-organized information. Minor text edits. Updated Figure 2-5 and Figure 2-7. Updated the "Transceiver Data Rates Supported in Basic (PMA Direct) Mode", "FPGA Fabric PLLs-Transceiver PLLs Cascading in the 780-Pin Package", "FPGA Fabric PLLs-Transceiver PLLs Cascading in the 1152-Pin Package", sections. Removed Table 2-5, Table 2-6, Table 2-7 Removed Figure 2-17 and Figure 2-18. Minor text edits. 3.2 March 2010 3.1 November 2009 June 2009 3.0 2.2 March 2009 2.1 Minor updates. November 2008 20 Update to chapter. May 2008 1.0 Initial release. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 2-82 Stratix IV Device Handbook Volume 2: Transceivers Chapter 2: Transceiver Clocking in Stratix IV Devices Configuration Examples September 2012 Altera Corporation 3. Configuring Multiple Protocols and Data Rates in Stratix IV Devices September 2012 SIV52003-4.2 SIV52003-4.2 This chapter describes the procedure for merging; for example, when combining multiple protocols and data rates within a transceiver block. The instances you can combine include Receiver Only and Transmitter and Receiver channels as well as channels configured in Protocol Functional modes, channels using PLL cascade clocks, channels in multiple transceiver blocks, and channels with a Basic (PMA Direct) configuration. This chapter also offers several examples of sharing the clock multiplier unit phase-locked loops (CMU PLLs). f For information about the supported data rate range for the auxiliary transmit (ATX) PLL, refer to the "Transceiver Performance Specifications" section in the DC and Switching Characteristics for Stratix IV Devices chapter. Overview Each transceiver channel in a Stratix(R) IV GX and GT device can run at an independent data rate or in an independent protocol mode. Within each transceiver channel, the transmitter and receiver channels can run at different data rates. Each transceiver block consists of two CMU PLLs that provide clocks to all the transmitter channels within the transceiver block. Each receiver channel contains a dedicated clock data recovery (CDR) unit. In addition to the CMU PLLs, the ATX PLLs are available to provide clocks to the transmitter channels that are configured for a specific data rate range. This chapter includes the following sections: "Glossary of Terms" on page 3-2 "Creating Transceiver Channel Instances" on page 3-3 "General Requirements to Combine Channels" on page 3-3 "Sharing CMU PLLs" on page 3-5 "Sharing ATX PLLs" on page 3-10 "Combining Receiver Only Channels" on page 3-10 "Combining Transmitter Channel and Receiver Channel Instances" on page 3-11 "Combining Transceiver Instances in Multiple Transceiver Blocks" on page 3-13 "Combining Transceiver Instances Using PLL Cascade Clocks" on page 3-16 "Combining Channels Configured in Protocol Functional Modes" on page 3-17 "Combining Transceiver Channels in Basic (PMA Direct) Configurations" on page 3-25 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 2: Transceivers September 2012 Feedback Subscribe 3-2 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Glossary of Terms "Combination Requirements When You Enable Channel Reconfiguration" on page 3-42 "Combining Transceiver Channels When You Enable the Adaptive Equalization (AEQ) Feature" on page 3-47 "Combination Requirements for Stratix IV Devices" on page 3-49 "Summary" on page 3-49 Each transmitter channel has a local divider (/1, /2, or /4) that divides the high-speed clock output of the CMU PLL to provide high-speed serial and low-speed parallel clocks for its physical coding sublayer (PCS) and physical medium attachment (PMA) functional blocks. You can configure the RX CDR present in the receiver channel to a distinct data rate and provide separate input reference clocks. Each receiver channel also contains a local divider that divides the high-speed clock output of the RX CDR and provides clocks for its PCS and PMA functional blocks. To enable transceiver channel settings, the Quartus(R) II software provides the ALTGX MegaWizardTM Plug-In Manager interface. The ALTGX MegaWizard Plug-In Manager allows you to instantiate a single transceiver channel or multiple transceiver channels in Receiver and Transmitter, Receiver only, and Transmitter only configurations. Glossary of Terms Table 3-1 lists the terms used in the chapter. Table 3-1. Glossary of Terms Used in this Chapter Configuration Description Regular Channels This refers to the four transceiver channels in each transceiver block that contain PCS. Basic (PMA Direct) This refers to the Basic (PMA Direct) configuration that you can use for both regular and CMU channels. Basic (PMA Direct) mode has two variations, x1 and xN. The term "Basic (PMA Direct)" used in this chapter refers to both x1 and xN and to regular/CMU Channels. Any specific reference to x1 and xN or regular/CMU channels is stated explicitly. Non-Basic (PMA Direct) This term refers to all single channel non-bonded configurations (for example, GIGE, PCI Express(R) [PCIe] x1) or bonded channel configurations that have PCS enabled (for example, Basic x4 and x8, XAUI, PCIe x4 and x8). Also, any reference to a channel in non-Basic (PMA Direct) mode indicates that the channel is a regular transceiver channel. Basic (PMA Direct) x1 A transceiver channel set up in this configuration uses the high-speed serial clock from the CMU PLL that is present within the same transceiver block. You can select this configuration by setting the Which protocol you will be using? option to Basic (PMA Direct) and the Which sub protocol you will be using? option to none. A transceiver channel set up in this configuration uses the xN high-speed clock lines. You can Basic (PMA Direct) xN select this configuration by setting the Which protocol you will be using? option to Basic (PMA direct) and the Which sub protocol you will be using? option to xN. f For more information about transceiver channel set up using a Basic (PMA Direct) xN configuration, refer to the Transceiver Clocking in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Creating Transceiver Channel Instances 3-3 Creating Transceiver Channel Instances The two ways you can instantiate multiple transceiver channels in the General screen of the ALTGX MegaWizard Plug-In Manager are: In the What is the number of channels? option, select the required value. This method creates all the transceiver channels with identical configurations. For an example, refer to "Combining Transceiver Instances in Multiple Transceiver Blocks" on page 3-13. In the What is the number of channels? option, select 1 and create a single channel transceiver instance. To instantiate additional transceiver channels with an identical configuration, select the created ALTGX instance multiple times. If you need additional transceiver channels with different configurations, create separate ALTGX megafunction instances with different settings and use them in your design. When you create instances using the above methods, you can force the placement of up to four transceiver channels within the same transceiver block. Do this by assigning the tx_dataout and rx_datain ports of the channel instances to a single transceiver bank. If you do not assign pins to the tx_dataout and rx_datain ports, the Quartus II software chooses default pin assignments. When you compile the design, the Quartus II software combines multiple channel instances within the same transceiver block if the instances meet specific requirements. The following sections explain these requirements for different transceiver configurations. General Requirements to Combine Channels When you create multiple ALTGX instances, the Quartus II software requires that you to set identical values for the following parameters and signals to combine the ALTGX instances within the same transceiver block or in transceiver blocks on the same side of the device. The following sections describe these requirements. Transmitter Buffer Voltage (VCCH) The Stratix IV GX device provides you the option to select 1.4 V or 1.5 V for the VCCH supply through the ALTGX MegaWizard Plug-In Manager. The Stratix IV GT device only allows 1.4 V for the VCCH supply. To combine the channel instances within the same transceiver block, the Quartus II software requires that you to set the same VCCH value in all the channel instances. 1 The data rate of the transmitter channel is limited based on the VCCH value selected. f For the data rate restrictions, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Transceiver Analog Power (VCCA_L/R) The Stratix IV GX and GT device contains two different power supply pins, VCCA_L and VCCA_R that provide power to the PMA blocks in all the transceiver channels on the left and right sides of the device, respectively. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-4 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices General Requirements to Combine Channels The Stratix IV GX and GT device provides you the option to select 2.5 V or 3.0 V for the VCCA_L/R supply through the ALTGX MegaWizard Plug-In Manager. The Stratix IV GT device only allows 3.3 V for the VCCA_L/R supply. You must set the same VCCA_L/R value for all the transceiver channel instances to enable the Quartus II software to place them in the transceiver blocks on the same side of the device. For example, if you have two ALTGX instances that you would like to place on the left side transceiver banks GXBL0 and GXBL1, the VCCA_L/R values in the two ALTGX instances must be the same. 1 The data rate of the transceiver channel is limited based on the VCCA_L/R value selected. f For the data rate restrictions, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Control Signals This section contains information about the gxb_powerdown, reconfig_fromgxb, and reconfig_togxb ports. gxb_powerdown Port The gxb_powerdown port is an optional port that you can enable in the ALTGX MegaWizard Plug-In Manager. If enabled, you must drive the gxb_powerdown port in the ALTGX instances from the same logic or the same input pin to enable the Quartus II software to assign them in the same transceiver block. reconfig_fromgxb and reconfig_togxb Ports In the ALTGX MegaWizard Plug-In Manager, the reconfig_fromgxb and reconfig_togxb ports are enabled if you select one of the following options in the Reconfig screen: 1 Analog Controls (VOD, Pre-emphasis, Manual Equalization, and EyeQ) Enable Channel and Transmitter PLL reconfiguration Offset cancellation for receiver channels (always enabled if the configuration is Transmitter and Receiver or Receiver only) To combine multiple instances within the same transceiver block: The reconfig_fromgxb ports must be enabled in each instance AND These ports must be connected to the same reconfig controller For example, consider that you want to place a Receiver only and Transmitter only instance in the same transceiver block. For the Receiver only instance, the Quartus II software automatically enables the reconfig_fromgxb port. For the Transmitter only instance, you must select the options in the Reconfig screen (mentioned above) to enable the reconfig_fromgxb port. In the design, connect these ports from the Transmitter only and Receiver only instance to the same reconfig controller. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing CMU PLLs 3-5 f For more information about connecting these ports to the dynamic reconfiguration controller, refer to the "Connecting the ALTGX and ALTGX_RECONFIG Instances" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. Calibration Clock and Power Down Each calibration block in a Stratix IV GX and GT device is shared by multiple transceiver blocks. If your design uses multiple transceiver blocks, depending on the transceiver banks selected, you must connect the cal_blk_clk and cal_blk_powerdown ports of all channel instances to the same input pin or logic. f For more information about the calibration block and transceiver banks that are connected to a specific calibration block, refer to the "Calibration Blocks" section in the Transceiver Architecture in Stratix IV Devices chapter. 1 Asserting the cal_blk_powerdown port affects calibration on all transceiver channels connected to the calibration block. Sharing CMU PLLs When you create multiple transceiver channel instances using CMU PLLs and intend to combine these instances in the same transceiver block, the Quartus II software checks whether a single CMU PLL can be used to provide clock outputs for the transmitter side of the channel instances. If a single CMU PLL is not sufficient, the Quartus II software attempts to combine the channel instances using two CMU PLLs. Otherwise, the Quartus II software issues a Fitter error. The following two sections describes the ALTGX instance requirements to enable the Quartus II software to share the CMU PLL. Multiple Channels Sharing a CMU PLL To enable the Quartus II software to share the same CMU PLL for multiple channels, the following parameters in the channel instantiations must be identical: September 2012 "Base data rate" (the CMU PLL is configured for this data rate) CMU PLL bandwidth setting Reference clock frequency Input reference clock pin pll_powerdown port of the ALTGX instances must be driven from the same logic GXB_TX_PLL_Reconfig_Group assignment (refer to Table 3-14 on page 3-42) If the selected functional mode in one instance is (OIF) CEI Phy Interface or PCIe, the other instance must have the same functional mode to share the CMU PLL. For example, if you have two channels, one configured in Basic mode and the other configured in (OIF) CEI Phy Interface mode at the same data rate, the Quartus II software does not share the same PLL because the internal parameters for these two functional modes are different. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-6 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing CMU PLLs Each channel instance can have a different local divider setting. This is a useful option when you intend to run each channel within the transceiver block at different data rates that are derived from the same base data rate using the local divider values /1, /2, and /4. Example 1 shows this design configuration. Example 1 Consider an example design with four instances of a Receiver and Transmitter configuration in the same transceiver block at various serial data rates. Assume that each instance contains a channel and is driven from the same clock source and has the same CMU PLL bandwidth settings. Table 3-2 lists the configuration for Example 1. Table 3-2. Configuration for Example 1 ALTGX MegaWizard Plug-In Manager Settings User-Created Instance Name Number of Channels Configuration Effective Data Rate (Gbps) inst0 1 Receiver and Transmitter 4.25 inst1 1 Receiver and Transmitter 2.125 inst2 1 Receiver and Transmitter 1.0625 inst3 1 Receiver and Transmitter 4.25 For Example 1, you can share a single CMU PLL for all four channels because: One CMU PLL can be configured to run at 4.25 Gbps. Each channel can divide the CMU PLL clock output using the local divider and achieve the required data rates of 4.25 Gbps, 2.125 Gbps, and 1.0625 Gbps. Because each receiver channel has a dedicated CDR, the receiver side in each instance can be set up for these three data rates without any restrictions. To enable the Quartus II software to share a single CMU PLL for all four channels, set the values listed in Table 3-3 in the General screen of the ALTGX MegaWizard Plug-In Manager. Table 3-3. ALTGX MegaWizard Plug-In Manager Settings for Example 1 Instance inst0 inst1 inst2 Stratix IV Device Handbook Volume 2: Transceivers General Screen Option What is the effective data rate? Setting (Gbps) 4.25 Specify base data rate 4.25 (1) What is the effective data rate? 2.125 Specify base data rate 4.25 (1) What is the effective data rate? 1.0625 Specify base data rate 4.25 (1) September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing CMU PLLs 3-7 Table 3-3. ALTGX MegaWizard Plug-In Manager Settings for Example 1 Instance inst3 General Screen Option What is the effective data rate? Specify base data rate Setting (Gbps) 4.25 4.25 (1) Note to Table 3-3: (1) The Specify base data rate option is 4.25 Gbps for all four instances. Given that the CMU PLL bandwidth setting and input reference clock are the same and that the pll_powerdown ports are driven from the same logic or pin, the Quartus II software shares a single CMU PLL that runs at 4.25 Gbps. You can force the placement of the transceiver channels to a specific transceiver block by assigning pins to tx_dataout and rx_datain. Otherwise, the Quartus II software selects a transceiver bank. Figure 3-1 and Figure 3-2 show the scenario before and after the Quartus II software combines the transceiver channel instances. Because the RX CDR is not shared between channels, only the CMU PLL is shown. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-8 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing CMU PLLs 1 Each of the ALTGX instances has a pll_powerdown port. You must drive the pll_powerdown ports for all the instances from the same logic to enable the Quartus II software to share the same CMU PLL. Figure 3-1. ALTGX Instances Before Compilation for Example 1 ALTGX Effective Data Rate: 4.25 Gbps CMU PLL Base Data Rate: 4.25 Gbps Inst 0 ALTGX Effective Data Rate: 2.125 Gbps CMU PLL Base Data Rate: 4.25 Gbps Inst 1 ALTGX Effective Data Rate: 1.0625 Gbps CMU PLL Base Data Rate: 4.25 Gbps Inst 2 ALTGX Effective Data Rate: 4.25 Gbps CMU PLL Base Data Rate: 4.25 Gbps Inst 3 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing CMU PLLs 3-9 Figure 3-2 shows the scenario after the Quartus II software combines the transceiver channel instances. Figure 3-2. Combined Instances after Compilation for Example 1 Transceiver Block Inst 0 Effective Data Rate: 4.25 Gbps TX Loc Div: /1 Inst 1 Effective Data Rate: 2.125 Gbps TX Loc Div: /2 Inst 2 Effective Data Rate: 1.0625 Gbps TX Loc Div: /4 Inst 3 Effective Data Rate: 4.25 Gbps TX Loc Div: /1 CMU PLL Base Data Rate: 4.25 Gbps Example 2 Consider the example design listed in Table 3-4. When you have two instances with the same serial data rate but with different CMU PLL data rates, the Quartus II software creates a separate CMU PLL for the two instances. Table 3-4. Configuration for Example 2 ALTGX MegaWizard Plug-In Manager Settings User-Created Instance Name Number of Channels Configuration Effective Data Rate (Gbps) Base Data Rate (Gbps) inst0 1 Receiver and Transmitter 2.5 2.5 inst1 1 Receiver and Transmitter 2.5 5 inst2 1 Receiver and Transmitter 1 1 1 September 2012 Even though the effective data rate of inst1 is 2.5 Gbps (5 Gbps/2 = 2.5 Gbps), the same as inst0, when you compile the design, the Quartus II software requires two CMU PLLs to provide clocks for the transmitter side of the two instances because their base data rates are different. In this example, you have the third instance, inst2, that requires a third CMU PLL. Therefore, the Quartus II software cannot combine the above three instances within the same transceiver block. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-10 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Sharing ATX PLLs Sharing ATX PLLs The Quartus II software allows you to share the same ATX PLL for multiple transceiver instances if the following requirements are met: The ATX PLL bandwidth in both instances are the same If the selected functional mode in one instance is (OIF) CEI Phy Interface or PCIe, the other functional modes must be the same to share the ATX PLL. For example, if you have two channels, one configured in Basic mode and the other configured in (OIF) CEI Phy Interface mode at the same data rate, the Quartus II software does not share the same PLL because the internal parameters for these two functional modes are different. The base data rate and effective data rate values are the same. The pll_powerdown port in the instances are connected to the same logic. The instances are placed on the same side of the device. There is no contention on the xN clock lines from the ATX PLL and the two instances. 1 For more information about xN clocking, refer to the "Transmitter Channel Data Path Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. Combining Receiver Only Channels You can selectively use the receiver in the transceiver channel by selecting the Receiver only configuration in the What is the Operating Mode? option on the General screen of the ALTGX MegaWizard Plug-In Manager. You can combine Receiver only channel instances of different configurations and data rates into the same transceiver block. Because each receiver channel contains its own dedicated CDR, each Receiver only instance (assuming one receiver channel per instance) can have a different data rate. 1 For the Quartus II software to combine the Receiver only instances within the same transceiver block, you must connect gxb_powerdown (if used) for all the channel instances to the same logic or input pin. For more information, refer to "General Requirements to Combine Channels" on page 3-3. 1 If your design contains a Receiver only instance, the Quartus II software disables all the settings for the unused transmitter channel present in the same physical transceiver channel. Therefore, the unused transmitter channel is always powered down in the hardware. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transmitter Channel and Receiver Channel Instances 3-11 Combining Transmitter Channel and Receiver Channel Instances You can create separate transmitter and receiver channel instances and assign the tx_dataout and rx_datain pins of the transmitter and receiver instances, respectively, to the same physical transceiver channel. This configuration is useful when you intend to run the transmitter and receiver channel at different serial data rates. To create separate transmitter and receiver channel instances, select the Transmitter only and Receiver only options in the operating mode (General screen) of the ALTGX MegaWizard Plug-In Manager. Multiple Transmitter Channel and Receiver Channel Instances The Quartus II software allows you to combine multiple Transmitter only and Receiver only channel instances within the same transceiver block. Based on the pin assignments, the Quartus II software combines the corresponding Transmitter only and Receiver only channels in the same physical channel. To enable the Quartus II software to combine the transmitter channel and receiver channel instances in the same transceiver block, follow the rules and requirements outlined in: "General Requirements to Combine Channels" on page 3-3 "Multiple Channels Sharing a CMU PLL" on page 3-5 "Combining Receiver Only Channels" on page 3-10 Example 3 Consider the example design listed in Table 3-5 with four ALTGX instances. Table 3-5. Four ALTGX Instances for Example 3 Instance Name Configuration Serial Data Rate (Gbps) Input Reference Clock Frequency (MHz) inst0 Transmitter only 3.125 156.25 inst1 Receiver only 2.5 156.25 inst2 Transmitter only 1.25 125 inst3 Receiver only 2 125 After you create the above instances, if you force the placement of the instances, as listed in Table 3-6, the Quartus II software combines inst0 and inst1 to physical channel 0, and inst2 and inst3 to physical channel 1. Table 3-6. Forced Placement of the Instances for Example 3 Instance Name September 2012 Altera Corporation Physical Channel Pin Assignments in the Same Transceiver Block inst0 TX pin of channel 0 inst1 RX pin of channel 0 inst2 TX pin of channel 1 inst3 RX pin of channel 1 Stratix IV Device Handbook Volume 2: Transceivers 3-12 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transmitter Channel and Receiver Channel Instances Figure 3-3 and Figure 3-4 show the transceiver channel instances before and after compilation. Figure 3-3. ALTGX Transceiver Channel Instances Before Compilation for Example 3 ALTGX Effective Data Rate: 3.125 Gbps CMU PLL Base Data Rate: 3.125 Gbps Inst 0 ALTGX RX CDR Base Data Rate: 2.5 Gbps Inst 1 ALTGX Effective Data Rate: 1.25 Gbps CMU PLL Base Data Rate: 1.25 Gbps Inst 2 ALTGX RX CDR Base Data Rate: 2 Gbps Inst 3 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Instances in Multiple Transceiver Blocks 3-13 Figure 3-4 shows the transceiver channel instances after compilation. Figure 3-4. Combined Transceiver Instances After Compilation for Example 3 Transceiver Block Transceiver Channel0 TX Channel Inst 0 Effective Data Rate: 3.25 Gbps TX Loc Div: /1 Inst 1 RX Channel Effective Data Rate: 2.5 Gbps CMU PLL Base Data Rate: 3.125 Gbps CMU PLL Base Data Rate: 1.25 Gbps Transceiver Channel1 TX Channel Inst 2 Effective Data Rate: 1.25 Gbps TX Loc Div: /1 RX Channel Inst 3 Effective Data Rate: 2 Gbps Combining Transceiver Instances in Multiple Transceiver Blocks The method to instantiate multiple transceiver channels using a single ALTGX instance is described in "Creating Transceiver Channel Instances" on page 3-3. The following section describes the method to instantiate multiple transceiver channels using multiple transceiver blocks. When you create a transceiver instance that has more than four transceiver channels (assuming that the instance is created in non-Basic (PMA Direct) functional mode which requires regular channels), the Quartus II software attempts to combine the transceiver channels in multiple transceiver blocks. This is shown in the following examples. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-14 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Instances in Multiple Transceiver Blocks Example 4 Consider the design example configuration listed in Table 3-7 with two ALTGX instances. Table 3-7. Two ALTGX Instances for Example 4 Instance Name Number of Transceiver Channels Configuration Serial Data Rate (Gbps) Input Reference Clock (MHz) inst0 7 Receiver and Transmitter 4.25 125 from refclk0 inst1 1 Receiver and Transmitter 4.25 125 from refclk0 (same as inst0) In this case, assuming that all the required parameters specified in "Multiple Channels Sharing a CMU PLL" on page 3-5 are identical for inst0 and inst1, the Quartus II software fits inst0 and inst1 in two transceiver blocks. Figure 3-5 and Figure 3-6 show the transceiver instances before and after compilation. Figure 3-5. Transceiver Channel Instances Before Compilation for Example 4 Inst0 Effective Data Rate: 4.25 Gbps Input Clock Frequency: 125 MHz Number of Channels: 7 Inst1 Effective Data Rate: 4.25 Gbps Input Clock Frequency: 125 MHz Number of Channels: 1 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Instances in Multiple Transceiver Blocks 3-15 Figure 3-6 shows the transceiver instances after compilation. Figure 3-6. Combined Transceiver Instances After Compilation for Example 4 Transceiver Block 0 Ch3 of inst0 Effective Data Rate: 4.25 Gbps Ch2 of inst0 Effective Data Rate: 4.25 Gbps CMU PLL Base Data Rate: 4.25 Gbps Ch1 of inst0 Effective Data Rate: 4.25 Gbps Ch0 of inst0 Effective Data Rate: 4.25 Gbps Transceiver Block 1 Ch0 of inst1 Effective Data Rate: 4.25 Gbps Ch6 of inst0 Effective Data Rate: 4.25 Gbps CMU PLL Base Data Rate: 4.25 Gbps Ch5 of inst0 Effective Data Rate: 4.25 Gbps Ch4 of inst0 Effective Data Rate: 4.25 Gbps You can force the placement of the transceiver channels in specific transceiver banks by assigning pins to the tx_dataout and rx_datain ports of inst0 and inst1. Even though inst0 instantiates seven transceiver channels, the ALTGX MegaWizard Plug-In Manager provides only a one-bit wide pll_inclk port for inst0. In your design, provide only one clock input for the pll_inclk port. The Quartus II software uses two transceiver blocks to fit the seven channels and internally connects the input reference clock (connected to the pll_inclk port in your design) to the CMU PLLs of two transceiver blocks. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-16 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Instances Using PLL Cascade Clocks 1 For inst1, the ALTGX MegaWizard Plug-In Manager provides a pll_inclk port. In this example, it is assumed that a single reference clock is provided for inst0 and inst1. Therefore, connect the pll_inclk port of inst0 and inst1 to the same input reference clock pin. This enables the Quartus II software to share a single CMU PLL in transceiver block 1 that has three channels of inst0 and one channel of inst1 (shown as ch4, ch5, and ch6 in transceiver block 1 in Figure 3-6). For the RX CDRs in inst0, the ALTGX MegaWizard Plug-In Manager provides seven bits for the rx_cruclk port (if you do not select the Train Receiver CDR from pll_inclk option in the PLL/Ports screen). This allows separate input reference clocks to the RX CDRs of each channel. Combining Transceiver Instances Using PLL Cascade Clocks The Stratix IV GX and GT transceiver has the ability to cascade the output of the general purpose PLLs (PLL_L and PLL_R) to the CMU PLLs, ATX PLLs, and receiver CDRs. The left side PLLs can only be cascaded with the transceivers on the left side of the device. Similarly, the right side PLLs can only be cascaded with the transceivers on the right side of the device. Each side of the Stratix IV GX and GT device contains a PLL cascade clock network; a single line network that connects the PLL cascade clock to the transceiver block. This clock line is segmented to allow different PLL cascade clocks to drive the transceiver CMU PLLs, ATX PLLs, and RX CDRs. Within the same segment, only a single PLL_L/PLL_R can drive these transceiver PLLs/CDRs. Therefore, if you create two instances that use different PLLs for cascading, you cannot place these instances within the transceiver block. The segmentation locations differ based on the device family. f For more information about using the PLL cascade clock and segmentation, refer to the "Dedicated Left and Right PLL Cascade Lines Network" section in the Transceiver Clocking in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes 3-17 Combining Channels Configured in Protocol Functional Modes This section describes how to combine channels for various protocol functional modes. Combining Channels in Bonded Functional Modes This section describes the combination requirements in the two variations of bonded functional modes using transceiver PCS blocks. The two bonded functional modes are: "Bonded x4 Functional Mode"--Examples of bonded x4 mode: Basic mode with the sub protocol set to x4 XAUI PCIe mode with the sub protocol set to Gen1 x4 or Gen2 x4. "Bonded x8 Functional Mode" on page 3-20--Examples of bonded x8 mode: Basic mode with the sub protocol x8 PCIe mode with the sub protocol x8 Bonded x4 Functional Mode The combination requirements for Basic x4, Deterministic Latency x4, and PCIe x4 functional modes (if you do not use the PCIe hard IP block) are similar. In this mode, the transmitter channels are synchronized to reduce skew. The Quartus II software shares the control from physical transmitter channel 0 with the other transmitter channels in the transceiver block. Therefore, when you an create an instance in this mode, the logical transmit channel 0 (tx_dataout[0] in the instance) must be assigned by the physical channel location 0 in the transceiver block. The central clock divider block in the CMU0 channel forwards the high-speed serial and low-speed parallel clocks to the transmitter channels. f This clocking scheme is described in the "Bonded Channel Configurations" section of the Transceiver Clocking in Stratix IV Devices chapter. Because you used the central clock divider, the are two restrictions on the channel combinations: 1. If you configure channels in bonded x4 functional mode, the remaining transmitter channels (regular or CMU channels) within the transceiver block can be used only in Basic (PMA Direct) x1 or xN mode. 1 If PCIe functional mode uses the PCIe hard IP block, the combination requirements are different. For more information, refer to "Combining Channels Using the PCIe hard IP Block with Other Channels" on page 3-24. The receiver channels are clocked independently. Therefore, you can configure the unused receiver channels within a transceiver block in any allowed configuration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-18 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes Figure 3-7 shows examples of supported and unsupported combinations. Figure 3-7. Examples of Supported and Unsupported Configurations to Combine Instances in Basic x4 Mode Transceiver Block Transceiver Block Instance 0 2 Regular Channels Instance 0 3 Regular Channels Instance 0 2 Regular Channels Receiver and Transmitter Basic mode and x4 sub-protocol Receiver and Transmitter Basic mode and x4 sub-protocol Receiver and Transmitter Basic mode and x4 sub-protocol Instance 1 1 CMU Channel Instance 1 2 Regular Channels Transceiver Block Instance 1 2 Regular Channels Receiver only (any functional mode) Transmitter and Receiver (Basic [PMA Direct] xN Functional mode) Transmitter only (any mode other than Basic [PMA Direct] mode) Red: Unsupported Supported Configuration Supported Configuration Unsupported Configuration The CMU0 PLL or CMU1 PLL can drive the central clock divider block in the CMU0 channel. In cases where you use CMU1 PLL for bonded x4 mode, the Quartus II software does not allow you to use CMU0 PLL for any other configuration because part of the CMU0 channel (the central clock divider) is already used by the bonded x4 functional mode. Using the remaining channels in Basic (PMA Direct) x1 or xN mode depends on the following conditions. 1. If CMU1 PLL is available for clock generation, you can use the remaining transmitter channels in the transceiver block in Basic (PMA Direct) x1 configuration. 2. If you want to configure the remaining transmitter channels at the same data rate as the bonded x4 functional mode, you can configure the remaining transmitter channels in Basic (PMA Direct) x1 mode. The requirements are specified in "Sharing CMU PLLs" on page 3-5 and "General Requirements to Combine Channels" on page 3-3. 3. If all the regular channels are configured in bonded x4 functional mode, you can configure the transmitter side of the CMU0 channel in Basic (PMA Direct) xN mode in single-width configuration only (double-width configuration is not supported). You can use the CMU1 channel in Basic (PMA Direct) xN single-width or double-width configuration. 1 Stratix IV Device Handbook Volume 2: Transceivers This only applies to the transmitter side and not the receiver side of the CMU channel. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes 3-19 4. Using the receiver side of the CMU channels depends on whether you use CMU1 PLL or CMU0 PLL to generate clocks for the bonded x4 functional mode. If the CMU PLL within the corresponding CMU channel is not available to perform CDR functionality, you cannot configure it as a receiver. 5. If you use ATX PLL to generate clocks for the x4 bonded functional mode, you can use both the Transmitter and Receiver side of the CMU0 and CMU1 channels. You must satisfy the requirements specified in number 3. Figure 3-8 shows a configuration in which all the transmitter channels in the transceiver block are used. 1 For XAUI, the option to select ATX PLL is not available. Figure 3-8 shows the combination of Basic/PCIe x4 functional mode with Basic (PMA Direct) xN mode within the same transceiver block. Figure 3-8. Basic x4 Functional Mode Configuration when Combining Channels (4) ATX PLL xN Top Clock Line (3) TX3 - Basic x4/ PCIe x4 TX2 - Basic x4/ PCIe x4 CMU1 Channel (PMA Direct xN mode) (1) CMU0 Channel TX - Basic (Basic [PMA Direct] xN mdoe) (2) Central Clock Divider x4 Clock Line (3) TX1 - Basic x4/ PCIe x4 TX0 - Basic x4/ PCIe x4 xN Bottom Clock Line ATX PLL Notes to Figure 3-8: (1) You can configure this channel in Basic (PMA Direct) single-width or double-width mode. (2) You can configure this channel only in Basic (PMA Direct) single-width mode. (3) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black line represents the xN bottom clock line. (4) To simplify the illustration, only the transmitter side is shown. PCIe x4 refers to PCIe with the sub protocol set to Gen1 x4 and Gen2 x4. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-20 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes Bonded x8 Functional Mode Bonded x8 functional mode is similar to bonded x4 functional mode except that the controls are shared from the physical channel 0 of the master transceiver block. The master is the lower of the two adjacent transceiver blocks selected for the x8 configuration. Therefore, when you an create an instance in this mode, you must assign the logical transmit channel 0 (tx_dataout[0] in the instance to the physical channel location 0 in the master transceiver block. f There are specific transceiver blocks that can be paired as master-slave in the x8 configuration. f The master is the adjacent lower transceiver block. For more information about location requirements, refer to the "Bonded Channel Configurations" section of the Transceiver Clocking in Stratix IV Devices chapter. In Basic x8 functional mode, you can select the number of channels to be less than 8 by setting the What is the number of channels? option on the General screen. In this instance, you can use the remaining transmitter channels only in Basic (PMA Direct) x1 or xN mode. In PCIe Gen1 x8 and Gen2 x8 functional modes, the number of regular channels used is always 8. The number of remaining transmitter channels (CMU channels or regular channels) in the two transceiver blocks available for use in Basic (PMA Direct) x1 or xN mode depends on whether the x8 functional mode uses CMU PLL or ATX PLL, as described below. 1 Stratix IV Device Handbook Volume 2: Transceivers If PCIe functional mode uses the PCIe hard IP block, the combination requirements are different. For more information, refer to "Combining Channels Using the PCIe hard IP Block with Other Channels" on page 3-24. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes 1 3-21 Each receiver channel configured in Basic x8 functional mode is clocked independently by the recovered clock from its receiver CDR. You can use the available receiver channels in any configuration. Figure 3-9 shows examples of supported and unsupported configuration in Basic x8 mode. Figure 3-9. Examples of Supported and Unsupported Configurations to Combine Instances in Basic x8 Mode Two Adjacent Transceiver Blocks Two Adjacent Transceiver Blocks Two Adjacent Transceiver Blocks Instance 0 6 Channels Instance 0 6 Channels Instance 0 6 Channels Receiver and Transmitter Basic mode and x8 sub-protocol Receiver and Transmitter Basic mode and x4 sub-protocol Receiver and Transmitter Basic mode and x8 sub-protocol Instance 1 2 Channels Instance 1 1 Channel Instance 1 2 Channels Receiver only (any functional mode) Transmitter and Receiver (Basic [PMA Direct] xN Functional mode) Transmitter only (any mode other than Basic [PMA Direct] function mode) Red: Unsupported Supported Configuration Supported Configuration Unsupported Configuration When the eight regular channels are used up in bonded x8 functional mode: September 2012 If the ATX PLL is used to generate clocks for the x8 functional mode shown in Figure 3-10, you can use the four CMU channels (two from the master and slave transceiver block) in Basic (PMA Direct) xN mode. Within Basic (PMA Direct) xN mode, you can configure the CMU0 channels in the master and slave transceiver block only in single-width mode (use the single-width mode option in the General screen). If a CMU1 channel or regular channels are available for use, you can use them in Basic (PMA Direct) xN mode in single-width or double-width configuration. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-22 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes Figure 3-10. Basic x8/PCIe x8 Functional Mode Configuration when Combining Channels (ATX PLL) (4) ATX PLL Slave Transceiver Block xN Top Clock Line (3) TX7 - Basic x8/ PCIe x8 TX6 - Basic x8/ PCIe x8 x4 Clock Line (3) CMU1 Channel (PMA Direct xN mode) (1) CMU0 Channel TX - (Basic [PMA Direct] xN mode) (2) Central Clock Divider TX5 - Basic x8/ PCIe x8 xN Bottom Clock Line (3) TX4 - Basic x8/ PCIe x8 Master Transceiver Block TX3 - Basic x8/ PCIe x8 TX2 - Basic x8/ PCIe x8 x4 Clock Line (3) CMU1 Channel (Basic [PMA Direct] xN mode) (1) CMU0 Channel TX - (Basic [PMA Direct] xN mode) (2) Central Clock Divider TX1 - Basic x8/ PCIe x8 xN Bottom Clock Line (3) TX0 - Basic x8/ PCIe x8 ATX PLL Notes to Figure 3-10: (1) You can configure this channel in Basic (PMA Direct) single-width or double-width mode. (2) You can configure this channel only in Basic (PMA Direct) single-width mode. (3) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black line represents the xN bottom clock line. (4) To simplify the illustration, only the transmitter side is shown. PCIe x8 refers to PCIe with the sub protocol set to Gen1 x8 and Gen2 x8. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes 3-23 If the CMU PLL is used to generate clocks for the x8 bonded functional mode, you can use the CMU0 channel in the slave transceiver block only in Basic (PMA Direct) xN mode in the single-width configuration. You can use the CMU1 channels in both the master and slave transceiver blocks in Basic (PMA Direct) xN mode in single-width or double-width configuration. Figure 3-11 shows the Basic x8 functional mode configuration for these combination restrictions when using CMU PLL. Figure 3-11. Basic x8/PCIe x8 Functional Mode Configuration when Combining Channels (CMU PLL) (4) ATX PLL xN Top Clock Line (3) Slave Transceiver Block TX7 - Basic x8/ PCIe x8 TX6 - Basic x8/ PCIe x8 CMU1 Channel (Basic [PMA Direct] xN mode) (1) x4 Clock Line (3) CMU0 Channel TX - (Basic [PMA Direct] xN mode) (2) Central Clock Divider TX5 - Basic x8/ PCIe x8 xN Bottom Clock Line (3) TX4 - Basic x8/ PCIe x8 Master Transceiver Block TX3 - Basic x8/ PCIe x8 TX2 - Basic x8/ PCIe x8 x4 Clock Line (3) CMU1 Channel (Basic [PMA Direct] xN mode) (1) CMU0 Channel (Used for clock generation) Central Clock Divider TX1 - Basic x8/ PCIe x8 TX0 - Basic x8/ PCIe x8 Notes to Figure 3-11: (1) You can configure this channel in Basic (PMA Direct) single-width or double-width mode. (2) You can configure this channel only in Basic (PMA Direct) single-width mode. (3) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black line represents the xN bottom clock line. (4) To simplify the illustration, only the transmitter side is shown. PCIe x8 refers to PCIe with the sub protocol set to Gen1 x8 and Gen2 x8. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-24 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Channels Configured in Protocol Functional Modes Combining Channels Configured in Deterministic Latency Mode The ALTGX MegaWizard Plug-In Manager provides Deterministic Latency mode with two variations (x1 and x4) to eliminate uncertainty in the transceiver data path. This functional mode provides the Enable Phase Frequency Detector (PFD) feed back ... option in the PLL/Ports screen. If you select this option for x1, the low-speed parallel clock from the transmitter serializer is fed back to the PFD input of the CMU PLL; for x4, the output of the low-speed parallel clock from the central clock divider is provided as feed back. For the x1 variation, one CMU PLL is required for each transmitter channel in the instance. As a result, in x1 mode, you can configure only two channels within the transceiver block in this mode. The restrictions for Deterministic Latency mode in x4 mode are the same as that of the bonded x4 functional mode. For more information, refer to "Bonded x4 Functional Mode" on page 3-17. Combining Channels Using the PCIe hard IP Block with Other Channels The Stratix IV GX and GT device contains an embedded PCIe hard IP block that performs the phyMAC, datalink, and transaction layer functionality specified by PCIe base specification 2.0. Each PCIe hard IP block is shared by two transceiver blocks. The PCI Express Compiler Wizard provides you the options to configure the PCIe hard IP block. When enabled, the transceiver channels associated with this block are also enabled. There are restrictions on combining transceiver channels with different functional and/or protocol modes (for example, Basic mode) within two contiguous transceiver blocks with the channels that use the PCIe hard IP block. The restrictions depend on the number of channels used (x1 or x4) and the number of virtual channels (VCs) selected in the PCI Express Compiler MegaWizard Plug-In Manager. Table 3-8 lists the restrictions. 1 When you use the PCIe hard IP block, there are placement restrictions on the locations of the transceiver channels. f For these channel placement restrictions, refer to the PCI Express Compiler User Guide. Table 3-8. PCIe Hard IP Block Restrictions When Combining Transceiver Channels with Different Functional and/or Protocol Modes (Part 1 of 2) (1), (2), (7) PCIe Configuration (PCIe hard IP Options Enabled in the PCI Express Compiler MegaWizard Plug-In Manager) Transceiver Block 0 (4) Transceiver Block 1 (5) (3) Link Width Lane (Data Interface Width) x1 64-bit x4 64-bit Stratix IV Device Handbook Volume 2: Transceivers Virtual Channel (VC) Ch0 (6) Ch1 Ch2 Ch3 Ch4 Ch5 Ch6 Ch7 1 PCIe x1 Avail. Avail. Avail. Avail. Avail. Avail. Avail. 2 PCIe x1 -- -- -- Avail. Avail. Avail. Avail. 1 PCIe x4 Avail. Avail. Avail. Avail. 2 PCIe x4 Avail. Avail. Avail. Avail. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-25 Table 3-8. PCIe Hard IP Block Restrictions When Combining Transceiver Channels with Different Functional and/or Protocol Modes (Part 2 of 2) (1), (2), (7) PCIe Configuration (PCIe hard IP Options Enabled in the PCI Express Compiler MegaWizard Plug-In Manager) Transceiver Block 0 (4) Transceiver Block 1 (5) (3) Link Width Lane (Data Interface Width) x4 128-bit x8 -- Virtual Channel (VC) Ch0 (6) Ch1 Ch2 Ch3 Ch4 Ch5 Ch6 Ch7 1 PCIe x4 Avail. Avail. Avail. Avail. 2 PCIe x4 -- -- Avail. Avail. -- PCIe x8 Notes to Table 3-8: (1) Avail. indicates that the channels can be used in other configurations. (2) An em-dash (--) indicates that the channels are NOT available for use. (3) The CMU PLL is used for the transmitter side of the channels in this table. (4) Transceiver block 0--the master transceiver block that provides high-speed serial and low-speed parallel clocks in a PCIe x4 or x8 configuration. (5) Transceiver block 1--the adjacent transceiver block that shares the same PCIe hard IP block with transceiver block 0. (6) The physical channel 0 in the transceiver block. For more information about physical-to-logical channel mapping in PCIe functional mode, refer to the "x8 Channel Configuration" section in the Transceiver Clocking in Stratix IV Devices chapter. (7) When you the use PCIe hard IP Block, you cannot configure the CMU channels within the transceiver block as transceiver channels. f For more information about the PCI Express Compiler MegaCore functions and hard IP implementation, refer to the PCI Express Compiler User Guide. If you configure a transceiver channel in PCIe configuration and if an ATX PLL is used to provide clocks for the transmitter side of the channel, you can use the remaining transmitter channels within the same transceiver block only in Basic (PMA Direct) x1 or xN mode. Combining Transceiver Channels in Basic (PMA Direct) Configurations In this configuration, the transmitter and receiver PCS blocks of a transceiver channel are bypassed and the transceiver channel can run at a maximum of 6.5 Gbps. f For the data rate restrictions in Basic (PMA Direct) mode, refer to the "Transceiver Performance Specifications" section in the DC and Switching Characteristics for Stratix IV Devices chapter. Using the Quartus II software, you can configure the two CMU channels and regular transceiver channels in Basic (PMA Direct) mode. The following sections describes the different scenarios for combining Basic (PMA Direct) mode with other transceiver configurations. f For information about the FPGA fabric-transceiver interface, refer to the "Non-Bonded Basic (PMA Direct) Mode Channel Configurations" section in the Transceiver Clocking in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-26 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations Combining Multiple Channels Configured in Basic (PMA Direct) x1 Configurations When you configure a transceiver channel in Basic (PMA Direct) x1 configuration, the Quartus II software requires one of the two CMU PLLs within the same transceiver block to provide high-speed clocks to the transmitter side of the channel (you cannot use the ATX PLL). Therefore, within a transceiver block, you can only combine a maximum of five transceiver channels (using both the Transmitter and Receiver) configured in Basic (PMA Direct) x1 mode (one CMU channel to perform the clock multiplication unit functionality). You can configure the transmitter side of the CMU channel that uses its CMU PLL for clock generation in Basic (PMA Direct) x1 in single-width or double-width configuration, as shown in Figure 3-17 on page 3-32. There are multiple ways you can combine channels in Basic (PMA Direct) x1 mode within the same transceiver block: "Multiple Basic (PMA Direct) x1 Configuration Instances with One Channel per Instance" on page 3-26 "One Instance in a Basic (PMA Direct) x1 Configuration with Multiple Transceiver Channels" on page 3-26 "Combining Multiple Instances of Transmitter Only and Receiver Only Configurations in Basic (PMA Direct) x1 Mode" on page 3-29 "Combining Channels Configured in Basic (PMA Direct) x1 with Non-Basic (PMA Direct) Modes" on page 3-29 Multiple Basic (PMA Direct) x1 Configuration Instances with One Channel per Instance If you create multiple instances of Basic (PMA Direct) x1 with one channel per instance, you can combine them within the same transceiver block. To achieve this combination, refer to the requirements specified in "Multiple Channels Sharing a CMU PLL" on page 3-5 and "General Requirements to Combine Channels" on page 3-3. Note that one CMU PLL within the transceiver block must provide a high-speed clock for the transmitter side of the channels. Therefore, at least one CMU channel (that contains the CMU PLL) within the transceiver block must be available to generate high-speed serial and low-speed parallel clocks for the channels configured in this mode. You can also place the individual instances in this configuration in separate transceiver blocks. For this placement, the Quartus II software enables one CMU PLL per instance. One Instance in a Basic (PMA Direct) x1 Configuration with Multiple Transceiver Channels In this case, if the number of channels selected in the instance is less than six, the Quartus II software, by default, combines these channels within the same transceiver block and uses one CMU PLL to provide the high-speed clocks. If the number of channels is six or more, the Quartus II software requires two transceiver blocks and two CMU PLLs, one from each transceiver block. The following two examples show the combinations of channels under two different conditions. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-27 Example 5 Consider a design example configuration with a Basic (PMA Direct) x1 instance with the number of channels set to 7 in the ALTGX MegaWizard Plug-In Manager. With this setting, the ALTGX MegaWizard Plug-In Manager provides 7 bits of gxb_powerdown, rx_analogreset, and pll_powerdown ports. In this case, the Quartus II software attempts to combine the five channels in the instance to one transceiver block and the remaining two channels to the second transceiver block, assuming that the gxb_powerdown and pll_powerdown ports for the five channels are driven from the same logic. Figure 3-12 and Figure 3-13 show the conditions before and after compilation. Figure 3-12. Logical View of the Instance with Seven Channels Before Compilation for Example 5 Inst0 Number of channels = 7 Receiver and Transmitter configuration: Basic (PMA Direct) x1 mode September 2012 Altera Corporation CMU PLL Stratix IV Device Handbook Volume 2: Transceivers 3-28 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations Figure 3-13 shows the conditions after compilation. In this example, the gxb_powerdown and pll_powerdown ports for channels 0 to 4 and channels 5 and 6 are driven from the same logic. Figure 3-13. Combined Channels After Compilation for Example 5 Transceiver Block0 Inst0: Channel 0 RX TX CMU PLL Inst0: Channel 1 RX TX Inst0: Channel 2 RX TX Inst0: Channel 3 RX TX Inst0: Channel 4 RX TX Transceiver Block1 Inst0: Channel 5 RX TX CMU PLL Inst0: Channel 6 RX TX If you connect each of the seven bits of the gxb_powerdown and pll_powerdown ports to different reset control logic, the Quartus II software requires seven transceiver blocks to combine the seven channels in the instance. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-29 Combining Multiple Instances of Transmitter Only and Receiver Only Configurations in Basic (PMA Direct) x1 Mode The Quartus II software allows you to combine instances of Transmitter Only and Receiver Only configurations in Basic (PMA Direct) x1 mode. You can also combine Transmitter Only instances in non-Basic (PMA Direct) x1 configuration (non-bonded only) and the Receiver Only instance in Basic (PMA Direct) configurations (and vice versa) in the same physical channel. The combination requirements of the instances in Basic (PMA Direct) x1 configuration are similar to that of non-Basic (PMA Direct) configuration. For more information, refer to the "Combining Transmitter Channel and Receiver Channel Instances" on page 3-11. Combining Channels Configured in Basic (PMA Direct) x1 with Non-Basic (PMA Direct) Modes You can combine a transceiver channel instance configured in Basic (PMA Direct) x1 configuration with instances set up in non-Basic (PMA Direct) configurations (for example, GIGE and SDI within the same transceiver block). If the CMU PLL configuration for the Basic (PMA Direct) x1 configuration and the non-Basic (PMA Direct) configuration instances meet the requirements specified in "Multiple Channels Sharing a CMU PLL" on page 3-5 and "General Requirements to Combine Channels" on page 3-3, the Quartus II software uses a single CMU PLL for these two instances. In addition, to share the same CMU PLL between the two instances, you cannot enable the channel reconfiguration option in the instance setup in non-Basic (PMA Direct) configuration. Example 6 Consider the example design listed in Table 3-9 for Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations at the same data rate. Table 3-9. Basic (PMA Direct) x1 and Non-Basic (PMA Direct) Configurations at the Same Data Rate for Example 6 Configuration CMU PLL Base Data Rate (Gbps) Transmitter Channel Effective Data Rate (Gbps) Input Reference Clock Frequency (MHz) inst0 GIGE 1.25 1.25 125 (assume refclk0) inst1 Basic (PMA Direct) x1 (four channels) 1.25 1.25 Same as inst0 Instances September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-30 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations Figure 3-14 shows Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations before compilation. Figure 3-14. Logical View of the Instances in Basic (PMA Direct) x1 and Non-Basic (PMA Direct) Configurations Before Compilation for Example 6 Inst0: GIGE functional mode Number of channels = 1 CMU PLL (1.25 G) Inst1: Basic (PMA Direct) x1 mode Number of channels = 4 CMU PLL (1.25 G) Figure 3-15 shows Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations after compilation. Figure 3-15. Combining Basic (PMA-Direct) x1 and Non-Basic (PMA Direct) Configurations After Compilation for Example 6 Transceiver Block Inst0: GIGE channel RX TX Inst1: Channel 0 RX TX Inst1: Channel 1 RX TX Inst1: Channel 2 RX TX Inst1: Channel 3 RX Stratix IV Device Handbook Volume 2: Transceivers TX CMU PLL (1.25 Gbps) September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-31 Example 7 Consider the example design listed in Table 3-10 for Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations at different data rates. Table 3-10. Basic (PMA Direct) x1 and Non-Basic (PMA Direct) Configurations at Different Data Rates for Example 7 Configuration CMU PLL Base Data Rate (Gbps) Transmitter Channel Effective Data Rate (Gbps) Input Reference Clock Frequency (MHz) inst0 GIGE 1.25 1.25 125 (assume refclk0) inst1 Basic (PMA Direct) x1 (three channels in Receiver and Transmitter configuration) 2 2 refclk1 inst2 Basic (PMA Direct) x1 (one channel in Transmitter Only configuration) 2 2 refclk1 Instances Figure 3-16 shows Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations before compilation. 1 The data rate configurations of the two CMU PLLs are different. Figure 3-16. Logical View of the Basic (PMA Direct) x1 and Non-Basic (PMA Direct) Configurations Before Compilation for Example 7 Inst0: GIGE functional mode Number of channels = 1 Inst1: Basic (PMA Direct) x1 mode Number of channels = 3 Inst2: Basic (PMA Direct) x1 mode Number of channels = 1 September 2012 Altera Corporation CMU PLL (1.25 Gbps) CMU PLL (2 Gbps) CMU PLL (2 Gbps) Stratix IV Device Handbook Volume 2: Transceivers 3-32 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations Figure 3-17 shows Basic (PMA Direct) x1 and non-Basic (PMA Direct) configurations after compilation. Figure 3-17. Combining Basic (PMA Direct) x1 and Non-Basic (PMA Direct) Instances in a Transceiver Block After Compilation for Example 7 GXBR CMU PLL (1.25 Gbps) Inst0: Channel (GIGE) RX TX Inst1: Channel 0 RX TX Inst1: Channel 1 RX TX Inst1: Channel 2 RX TX inst3: TX CMU PLL (2 Gbps) Key Observations 1 Stratix IV Device Handbook Volume 2: Transceivers To combine the these instances, two CMU PLLs are required due to the different data rates. Therefore, two CMU channels must be available to enable their respective CMU PLLs. Note that inst3 uses the transmit side of the CMU channel that uses the CMU PLL for clock generation. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-33 Basic (PMA Direct) xN Configurations When you configure a transceiver channel in Basic (PMA Direct) xN configuration, you can enable the Quartus II software to use the xN lines to provide clocks to the transmitter channels, as shown in Figure 3-18. The following are the possible sources driving the xN clock lines: The CMU0 central divider within the CMU0 channel. Only the CMU0 clock divider block can drive the xN clock lines. Either the CMU0 PLL or CMU1 PLL can drive the central clock divider block. f To understand the input clock connections to the central clock divider block, refer to the "CMU0 Channel" section in the Transceiver Architecture in Stratix IV Devices chapter. The ATX PLL block. Channel Placement in a Basic (PMA Direct) xN Mode Instance If you compile a design with a transceiver instance configured in Basic (PMA Direct) xN mode, the Quartus II software, by default, places these channels contiguously. You can force the placement of the transceiver channels across multiple transceiver blocks on the same side of the device by assigning pins to the transmitter and receiver serial ports. The logical channel 0 of the Basic (PMA Direct) xN mode instance does not have to be assigned to the physical channel 0 of a transceiver block. The logical channel 0 of an instance with multiple channels is tx_dataout[0] or rx_datain[0], which are the serial transmit and receive ports provided by the ALTGX MegaWizard Plug-In Manager. When you assign pins, you are not required to assign tx_dataout[0] to the location of physical channel 0 in the transceiver block to compile your design. This is not the case if you have a PCIe x4 configuration where tx_dataout[0]and rx_datain[0] must be assigned to physical channel 0 of the transceiver block. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-34 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations Figure 3-18 shows the different drivers of the xN_Top and xN_Bottom clock lines. Figure 3-18. The xN_Top and xN_Bottom Clock Line Connections Transceiver Block GXBR2 x1 CMU1 GXBR2 x4_GXBR2 xN_Bottom Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR2 ATX R1 PLL Block Transceiver Block GXBR1 x1 CMU1 GXBR1 x4_GXBR1 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR1 ATX R0 PLL Block Transceiver Block GXBR0 x1 CMU1 GXBR0 x4_GXBR0 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 Stratix IV Device Handbook Volume 2: Transceivers x1 CMU0 GXBR0 xN_Top September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-35 Examples of Combining Multiple Instances of Basic (PMA Direct) xN Modes The following section describes combining multiple transceiver channel instances in Basic (PMA Direct) xN mode. Configuration examples include transceiver channels with different data rates, configurations in Basic (PMA Direct) xN mode with non-Basic (PMA Direct) and ATX PLL, and unsupported configurations. Example 8 Consider the configuration for the two instances listed in Table 3-11 when combining transceiver channels in Basic (PMA Direct) xN mode with different data rates. Table 3-11. Combining Transceiver Channels in Basic (PMA Direct) xN Configuration with Different Data Rates for Example 8 User Defined Instance Name Number of Channels Effective Data Rate (Gbps) Configuration inst0 6 1.5 Receiver and transmitter inst1 5 1.25 Receiver and transmitter September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-36 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations You can place channels within a given instance non-contiguously, as shown in Figure 3-19. Figure 3-19. Non-Contiguous Placements of Channels Using Different CMU PLLs for Example 8 GXBR1 Inst0: Channel 0 RX TX Inst0: Channel 1 RX TX Inst0: Channel 2 RX TX x4 Clock Line (1) Inst0: Channel 3 RX TX Inst0: Channel 4 RX TX CMU0 Channel CMU0 PLL Base data rate 1.5 Gbps Central Clock Divider GXBR0 Inst1: Channel 0 RX TX Inst1: Channel 1 RX TX xN Top Clock Line (1) Inst1: Channel 2 RX TX Inst0: Channel 5 RX TX Inst1: Channel 3 RX TX xN Bottom Clock Line (1) Inst1: Channel 4 RX TX Transceiver Block 2 CMU0 Channel CMU0 PLL Base data rate 1.25 Gbps Central Clock Divider Note to Figure 3-19: (1) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black lines represent the xN bottom clock line. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-37 Key Observations Note that channel 5 in inst0 is placed in transceiver block 1 and receives the high-speed clock through the xN_Top clock line. Some of the channels in transceiver block 1 receive their high-speed clock from the xN_Bottom clock line. Because the xN_Top and xN_Bottom lines are separate, this scenario is allowed. To understand the clock multiplexer on the xN clock lines, refer to Figure 3-18 on page 3-34. Combining Channels Configured in Basic (PMA Direct) xN Configuration with Non-Basic (PMA Direct) Configurations The Quartus II software only allows a combination of a transceiver channel instances configured in Basic (PMA Direct) xN mode with instances in non-Basic (PMA Direct) configurations; for example, GIGE and SDI. Example 9 Consider the example design listed in Table 3-12 for the two instances when combining a Basic (PMA Direct) xN configuration with a non-Basic (PMA Direct) configuration using a CMU PLL. Table 3-12. Combining Basic (PMA Direct) xN Configuration with Non-Basic (PMA Direct) Configuration Using CMU PLL for Example 9 User Defined Instance Name Number of Channels Effective Data Rate (Gbps) Configuration Functional Mode inst0 9 1.5 Receiver and Transmitter Basic (PMA Direct) xN inst1 1 1.25 Receiver and Transmitter GIGE September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-38 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations You can place these two instances in two transceiver blocks, as shown in Figure 3-20. Figure 3-20. Combining Basic (PMA Direct) xN Configuration with Non-Basic (PMA Direct) Configuration Using CMU PLL in Two Transceiver Blocks For Example 9 GXBR1 Inst0:Channel 0 RX TX Inst0:Channel 1 TX 1 RX Inst0:Channel Inst0:Channel 2 RX TX Inst0:Channel 3 RX x4 Clock Line (1) TX Inst0:Channel 4 RX TX CMU0 Channel Central Clock Divider CMU0 PLL GXBR0 Inst0:Channel 5 RX TX Inst0:Channel 6 xN Top Clock Line (1) RX TX 1 Inst0:Channel Inst0:Channel 7 RX TX Inst0:Channel 8 RX TX Inst1:Channel 0 RX TX x1 Clock Line CMU PLL (1.25 Gbps) Note to Figure 3-20: (1) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black line represents the xN bottom clock line. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-39 Example 10 Consider the example design listed in Table 3-13 when combining a Basic (PMA Direct) xN configuration with a non-Basic (PMA Direct) configuration using an ATX PLL. Table 3-13. Combining Basic (PMA Direct) xN Configuration with Non-Basic (PMA Direct) Configuration Using ATX PLL for Example 10 User Defined Instance Name Number of Channels Effective Data Rate (Gbps) Configuration Functional Mode inst0 10 1.5 Receiver and Transmitter Basic (PMA Direct) xN configuration inst1 1 5 Receiver and Transmitter Basic mode (PCS+PMA) using ATX PLL September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-40 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations In this case, the ATX PLL provides the high-speed clock to the transmitter channel of inst1. Therefore, you can combine 10 channels of inst0 and one channel of inst1 in two transceiver blocks, as shown in Figure 3-21. Figure 3-21. Combining Basic (PMA Direct) xN Configuration with Non-Basic (PMA Direct) Configuration Using an ATX PLL for Example 10 (1) GXBR1 Inst0: Channel 0 RX TX Inst0: Channel 1 RX TX Inst0: Channel 2 RX TX x4 Clock Line (2) Inst0: Channel 3 RX TX Inst0: Channel 4 RX TX CMU0 Channel Central Clock Divider CMU0 PLL GXBR0 xN Top Clock Line (2) Inst0: Channel 5 RX TX Inst0: Channel 6 RX TX Inst0: Channel 7 RX TX Inst0: Channel 8 RX TX Inst0: Channel 9 RX TX Inst1: Channel 0 RX TX xN Bottom Clock Line (2) ATX PLL block ATX PLL Base data rate 5 Gbps Central Clock Divider Notes to Figure 3-21: (1) The ATX PLL provides the high-speed clock to channel 0 of inst1. (2) The red lines represent the xN top clock line, the blue lines represent the x4 clock line, and the black line represents the xN bottom clock line. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels in Basic (PMA Direct) Configurations 3-41 You can also combine channels configured in Basic (PMA Direct) xN mode with bonded x4 and x8 functional modes. For example scenarios, refer to Figure 3-8 on page 3-19 and Figure 3-10 on page 3-22. Example 11 Consider the unsupported placement design example shown Figure 3-22. The placement is unsupported because of the xN_Top clock line contention between the ATX PLL and the CMU0 PLL in transceiver block 0. Figure 3-22. Unsupported Placement Due to xN Clock Line Contention for Example 11 GXBR1 Inst0:Channel 0 RX TX Inst0:Channel 1 RX TX 1 Inst0:Channel Inst0:Channel 2 RX TX Inst0:Channel 3 RX x4 Clock Line (1) TX Inst0:Channel 4 RX TX CMU0 Channel CMU0 PLL Central Clock Divider ATX PLL R1 (5 Gbps) GXBR0 Contention in using the clock line xN Top Clock Line (1) Inst0:Channel 5 RX TX Inst0:Channel 6 RX TX 1 Inst0:Channel Inst0:Channel 7 RX TX Inst0:Channel 8 RX TX Inst0:Channel 9 RX TX Inst1:Channel 0 RX TX This channel cannot get its clock from the ATX PLL due to clock contention Note to Figure 3-22: (1) The red lines represent the xN top clock line and the blue lines represent the x4 clock line. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-42 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements When You Enable Channel Reconfiguration Combination Requirements When You Enable Channel Reconfiguration You can configure a transmitter channel to: Switch to an alternate CMU PLL present within the same transceiver block. Switch to multiple TX PLLs (CMU or ATX PLLs) that are present outside the transceiver block. f For more information about setting up the transceiver to enable one of these three options, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. The section describes the combination requirements for an instance that is configured in one of the three options mentioned above with other instances. Combination Requirements When You Enable the Use Alternate CMU PLL Option If you create a transceiver instance by selecting the use alternate CMU PLL option, the Quartus II software uses two CMU PLLs. If you intend to combine other transmitter instances within the same transceiver block, the CMU PLLs must be shared between multiple instances. To enable the Quartus II software to share the CMU PLLs between these instances, you must: 1. Select the user alternate CMU PLL option in all the instances. 2. Set the same PLL logical reference index value for the similar PLLs between the two instances (similar PLLs are the ones that have the same data rate, input clock frequency, and bandwidth setting). 3. Create a GXB TX PLL Reconfiguration group setting in the assignment editor and assign the same value for both instances. This setting is required for all instances and channels controlled by the shared ALT_GX_Reconfig. Table 3-14 lists the assignment for the first instance (Instance 1). Table 3-14. Assignment for the First Instance--Instance 1 Assignment Setting To <provide the tx_dataout port name of first instance> Assignment Name GXB TX PLL Reconfiguration group setting Value <any number> Table 3-15 lists the assignment for the second instance (Instance 2). Table 3-15. Assignment for the Second Instance--Instance 2 Assignment 1 Stratix IV Device Handbook Volume 2: Transceivers Setting To <provide the tx_dataout port name of second instance> Assignment Name GXB TX PLL Reconfiguration group setting Value <same number as that of the first instance> Ensure that the requirements specified in the "General Requirements to Combine Channels" on page 3-3 and "Sharing CMU PLLs" on page 3-5 sections are met. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements When You Enable Channel Reconfiguration 3-43 Example 12 shows the requirements. Example 12 Consider that you intend to run four channels within the transceiver block to switch between GIGE and SONET OC48 data rates. Assume that by default two channels run at GIGE data rates and the other two channels run at SONET OC48 data rates. Assume that Instance1 with two channels running at GIGE data rate is created with the configuration, as listed in Table 3-16. Table 3-16. Combining Requirements with the Use Alternate CMU PLL Option Enabled-- Instance 1 for Example 12 PLL Data Rate (Gbps) Input Reference Clock (MHz) PLL Logical Reference Index Main PLL 1.25 125 0 Alternate PLL 2.488 155.5 1 Create Instance 2 with the following parameters to enable the Quartus II software to share CMU PLLs between the two instances. Table 3-17 lists the required parameters to be set for Instance 2. Table 3-17. Combining Requirements with the Use Alternate CMU PLL Option Enabled-- Instance 2 for Example 12 PLL Data Rate (Gbps) Input Reference Clock (MHz) PLL Logical Reference Index Main PLL 2.488 155.5 1 Alternate PLL 1.25 125 0 Table 3-18 lists the assignment for the GXB TX PLL Reconfiguration group for Instance 1 when you compile the design. Table 3-18. Assignment for the GXB TX PLL Reconfiguration Group for Instance 1 Assignment Setting tx_dataout_instance1[0] September 2012 To Note that the number of channels in this instance is 2. You can use any one of the channel port names within this instance for this assignment. Assignment Name GXB TX PLL Reconfiguration group setting Value 6 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-44 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements When You Enable Channel Reconfiguration Table 3-19 lists the assignment for the GXB TX PLL Reconfiguration group for Instance 2 when you compile the design. Table 3-19. Assignment for the GXB TX PLL Reconfiguration Group for Instance 2 Assignment Setting tx_dataout_instance2[1] To You can use any one of the channel port names within this Instance for this assignment. Assignment Name GXB TX PLL Reconfiguration group setting Value 6 Key Observations The Main PLL in Instance 1 is configured for GIGE data rates because this is the data rate that you intend to run the first instance. The main PLL in Instance 2 is configured for SONET OC48 data rates because this is the data rate that you intend to run the second channel. Note that the PLL logical reference index values for similar PLLs in Instance 1 and Instance 2 are the same. The GXB TX PLL Reconfiguration group setting value for tx_dataout of Instance 1 and Instance 2 are the same. Combination Requirements When You Use Multiple TX PLLs This scenario describes transceiver configurations that have the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option in the reconfig screen enabled. If you create a transceiver instance using the above option and would like to share the CMU PLLs or ATX PLL with other instances, ensure that you have met the following requirements: Select the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option in other instances. 1 The PLL logical reference index value of the similar PLLs that you intend to combine must be the same in all the instances considered. 1 Stratix IV Device Handbook Volume 2: Transceivers The number of additional PLLs selected (in the How many additional PLLs are used option) can be different between instances. Similar PLLs are the ones that have the same data rate, input clock frequency, and bandwidth setting. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements When You Enable Channel Reconfiguration 3-45 Assign the same GXB TX PLL Reconfiguration group setting value for the tx_dataout ports of all the instances. This is explained in "Combination Requirements When You Enable the Use Alternate CMU PLL Option" on page 3-42. Ensure that the requirements specified in "General Requirements to Combine Channels" on page 3-3, "Sharing CMU PLLs" on page 3-5, and "Sharing ATX PLLs" on page 3-10 are met. If you create an instance using the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option and place your transmitter channels in one transceiver block (for example, QL1) and you use a CMU/ATX PLL from another transceiver block (for example, QL0), the channels (if used) in QL0 must be connected to the same reconfiguration controller as that of QL1. Example 13 shows an instance using multiple PLLs. Example 13 Consider that the following 12-channel design is targeted for a THREE transceiver block per side device. The requirements for this design are: 1. Four transceiver channels to switch independently between four protocol data rates (SONET OC48, FC 2G, GIGE, and OTU1). 2. Four transceiver channels to operate in SONET OC48 data rate. 3. Four transceiver channels to operate in GIGE data rate. To implement step 1, you need four TX PLLs. Place the four channels in the middle transceiver block (QL1--the left side was chosen for illustration purposes), and provide one CMU PLL from QL0 for the SONET OC48 data rate and one CMU PLL from QL2 for the GIGE data rate. Use the two CMU PLLs from QL1 for the FC 2G and OTU1 data rates. To implement step 2, note that the CMU PLL in QL0 already provides the SONET OC48 data rate. Therefore, use the four channels in QL0 to run the SONET OC48 data rate. To implement step 3, note that the CMU PLL in QL2 already provides the GIGE data rate. Therefore, use the four channels in QL2 to run the GIGE data rate. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-46 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements When You Enable Channel Reconfiguration Figure 3-23 shows the configuration for Example 13. Figure 3-23. Three Transceiver Block Configuration for Example 13 QL2 TX3: GIGE TX2: GIGE x1 Clock Line (2) CMU0 Channel (GIGE) (1) TX1: GIGE TX0: GIGE QL1 xN_Top Clock Li TX3: four data rates TX2: four data rates CMU0 Channel (FC 2G) (1) CMU1 Channel (OTU1) (1) xN_Bottom Clock Line (2 TX1: four data rates TX0: four data rates QL0 TX3: SONET OC48 TX2: SONET OC48 x1 Clock Line (2) CMU0 Channel (SONET OC 48) (1) TX1: SONET OC48 TX0: SONET OC48 Notes to Figure 3-23: (1) CMU channels are used for clock generation. (2) The red lines represent the xN top clock line, the blue lines represent the x1 clock line, the black lines represent xN bottom clock, the green lines represents the CMU1 channel, and the brown lines represent the CMU0 channel. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels When You Enable the Adaptive Equalization (AEQ) Feature 3-47 Create three Instances for steps 1, 2, and 3 with the following parameters: Instance 1 Select the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option. Number of additional PLLs: 3 (Table 3-20) Table 3-20. Instance 1 for Example 13 PLL Data Rate PLL Logical Reference Index Main PLL FC 2G 0 PLL1 OTU1 1 PLL2 GIGE 2 PLL3 SONET OC48 3 Instance 2 Select the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option. Number of additional PLLs: 0 (Table 3-21) Table 3-21. Instance 2 for Example 13 PLL Data Rate PLL Logical Reference Index Main PLL SONET OC48 3 Instance 3 Select the use additional CMU/ATX Transmitter PLLs from outside the transceiver block option. Number of additional PLLs: 0 (Table 3-22) Table 3-22. Instance 3 for Example 13 PLL Data Rate PLL Logical Reference Index Main PLL GIGE 2 For more information, refer to "Combination Requirements When You Enable the Use Alternate CMU PLL Option" on page 3-42. Combining Transceiver Channels When You Enable the Adaptive Equalization (AEQ) Feature To enable the AEQ feature in a transceiver channel, select the Enable Adaptive Equalization option in the Reconfig screen of the ALTGX MegaWizard Plug-In Manager. When you select this option, the aeq_fromgxb and aeq_togxb ports are enabled. f For more information about initiating the AEQ feature, refer to the "Adaptive Equalization (AEQ)" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-48 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combining Transceiver Channels When You Enable the Adaptive Equalization (AEQ) Feature This section describes the requirements to combine transceiver channels when you enable the AEQ feature. You are not required to enable AEQ in all instances to combine them within the same transceiver block. When you instantiate the reconfiguration controller (ALTGX_Reconfig), the aeq_fromgxb and aeq_togxb ports available depend on the setting in the what is the number of channels controlled by the reconfig controller option. In configurations where AEQ is enabled on some of the transceiver channels that are connected to the same reconfiguration controller, the reconfiguration controller instance has additional aeq_fromgxb ports. To compile the design successfully, connect the unused aeq_fromgxb ports to 0. Example 14 shows the configuration. Example 14 Consider that you have two ALTGX instances, Instance 1 and Instance 2 with one channel each. Assume that only Instance 1 has the Enable adaptive equalization option enabled. Because there are two instances, the starting channel numbers of Instance 1 and Instance 2 are spaced four apart (0 and 4, respectively). In the ALTGX_Reconfig Instance, set the what is the number of channels controlled by the reconfig controller option to 8. The ALTGX_Reconfig Instance has 48 bits for the aeq_fromgxb port (24 bits per 4 channels). Instance 1 has the aeq_fromgxb[23:0] port because AEQ is enabled. Instance 2 does not have this port. Because Instance 1 has the starting channel number of 0, connect aeq_fromgxb of instance 1 to aeq_fromgxb[23:0] of the ALTGX_Reconfig Instance and tie aeq_fromgxb[47:24] to 0. f For more information about setting this parameter, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. Figure 3-24 shows the required connection for the aeq_fromgxb port. Figure 3-24. Required Connection for the aeq_fromgxb Port reconfig_fromgxb[16:0] reconfig_fromgxb[33:0] Instance 1 aeq_fromgxb[23:0] aeq_fromgxb[47:0] 24'h0 reconfig_fromgxb[33:17] Instance 2 Reconfiguration Controller 1 Stratix IV Device Handbook Volume 2: Transceivers The top 24 bits of aeq_fromgxb are tied to 0. This is because the logical channel address of Instance 1 starts at 4. Therefore, the top 24 bits of aeq_fromgxb corresponds to Instance 2. September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Combination Requirements for Stratix IV Devices 3-49 Combination Requirements for Stratix IV Devices Stratix IV GT devices allow configuring multiple protocols or data rates in the same transceiver block. For common protocols supported by both Stratix IV GX and GT devices, as well as for Basic functional mode at data rates between 2.488 Gbps and 8.5 Gbps, Stratix IV GT devices follow the same transceiver channel placement rules as Stratix IV GX devices. Placement Rules for Transceiver Channels at 9.9 Gbps to 11.3 Gbps You can use either the CMU PLL or the 10G ATX PLL to generate transceiver clocks for channels configured at data rates between 9.9 Gbps and 10.3125 Gbps. If you use a 10G ATX PLL to generate transceiver clocks for any channel configured between 9.9 Gbps and 10.3125 Gbps within a transceiver block, the remaining channels in the same transceiver block must either be unused or must be configured at the same data rate and clocked by the same 10G ATX PLL. If you use a CMU PLL to generate transceiver clocks for any channel configured between 8.5 Gbps and 11.3 Gbps within a transceiver block, the remaining channels in the same transceiver block may be configured at a different data rate and clocked by another CMU PLL or 6G ATX PLL. In this case, Stratix IV GT devices follow the same transceiver channel placement rules as Stratix IV GX devices. Placing transceiver channels clocked by another PLL in the same transceiver block as a 10G channel can result in higher transmitter output jitter on the 10G channel. The amount of additional jitter is pending characterization. Summary The following is a summary for configuring multiple protocols and data rates in a transceiver block: September 2012 You can run each transceiver channel at independent data rates or in independent protocol functional modes. Each transceiver block consists of two CMU PLLs that provide clocks to run the transmitter channels within the transceiver block. To enable the Quartus II software to combine multiple instances of transceiver channels within a transceiver block, follow the rules specified in "General Requirements to Combine Channels" on page 3-3 and "Sharing CMU PLLs" on page 3-5. You can reset each CMU PLL within a transceiver block using a pll_powerdown signal. For each transceiver instance, the ALTGX MegaWizard Plug-In Manager provides an option to select the pll_powerdown port. If you want to share the same CMU PLL between multiple transceiver channels, connect the pll_powerdown ports of the instances and drive the signal from the same logic. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 3-50 Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Summary If you enable the PCIe hard IP block using the PCI Express Compiler, the Quartus II software has certain requirements for using the remaining transceiver channels within the transceiver block in the other configurations. For more information, refer to "Combining Channels Using the PCIe hard IP Block with Other Channels" on page 3-24. The Quartus II software supports two kinds of Basic (PMA Direct) configurations (x1 and xN). If you use Basic (PMA Direct) x1 configuration, you must use the CMU PLL within the same transceiver block. Document Revision History Table 3-23 lists the revision history for this chapter. Table 3-23. Document Revision History (Part 1 of 2) Date Version September 2012 February 2011 November 2009 June 2009 March 2009 Stratix IV Device Handbook Volume 2: Transceivers 4.2 Changes Updated the "Overview" section to close FB #65273. Updated Table 3-15. Updated the "Multiple Channels Sharing a CMU PLL", "Transmitter Buffer Voltage (VCCH)", "Transceiver Analog Power (VCCA_L/R)", "Calibration Clock and Power Down", "Combining Transceiver Instances Using PLL Cascade Clocks", "Combining Channels Using the PCIe hard IP Block with Other Channels", "Combination Requirements When You Enable the Use Alternate CMU PLL Option", and "Placement Rules for Transceiver Channels at 9.9 Gbps to 11.3 Gbps" sections. Updated chapter title. Applied new template. Minor text edits. Added "Sharing ATX PLLs" on page 3-9, "Combination Requirements When Channel Reconfiguration is Enabled" on page 3-42, "Combining Transceiver Channels When the Adaptive Equalization (AEQ) is Enabled" on page 3-47, and "Combination Requirements for Stratix IV GT Devices" on page 3-49. Added Figure 3-8, Figure 3-10, Figure 3-11, Figure 3-23, and Figure 3-24. Updated all other sections. Added Stratix IV GT information. Updated graphics. Minor text edits. Updated Table 3-7. Minor text edits. Updated sections "Combining Channels Using the PCI Express Hard IP Block with Other Channels" on page 3-17, "Convention Used" on page 3-21, "PMA Direct Mode Restrictions" on page 3-22, "Multiple `PMA Direct x1' Configuration Instances with One Channel per Instance" on page 3-22, "Combining Multiple Instances of TX Only and RX Only PMA-Direct x1 Configurations" on page 3-26, "Combining Transceiver Channels with PMA Direct Configuration" on page 3-21. Updated Table 3-7. Updated Figure 3-19. 4.1 4.0 3.1 3.0 September 2012 Altera Corporation Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Summary 3-51 Table 3-23. Document Revision History (Part 2 of 2) Date Version November 2008 May 2008 September 2012 2.0 1.0 Altera Corporation Changes Updated "Transmitter Buffer Voltage (VCCH)" on page 3-2 Added "reconfig_fromgxb and reconfig_togxb Ports" on page 3-3 Updated Figure 3-7 Added "Basic x8 Mode" on page 3-15 Added Figure 3-8 Updated Table 3-7 Initial release. Stratix IV Device Handbook Volume 2: Transceivers 3-52 Stratix IV Device Handbook Volume 2: Transceivers Chapter 3: Configuring Multiple Protocols and Data Rates in Stratix IV Devices Summary September 2012 Altera Corporation 4. Reset Control and Power Down in Stratix IV Devices September 2012 SIV52004-4.3 SIV52004-4.3 Stratix(R) IV devices offer multiple reset signals to control the transceiver channels and clock multiplier unit (CMU) phase-locked loops (PLLs) independently. The ALTGX Transceiver MegaWizardTM Plug-In Manager provides individual reset signals for each channel instantiated in your design. It also provides one power-down signal for each transceiver block. This chapter includes the following sections: "User Reset and Power-Down Signals" on page 4-2 "Transceiver Reset Sequences" on page 4-4 "PMA Direct Drive Mode Reset Sequences" on page 4-24 "Dynamic Reconfiguration Reset Sequences" on page 4-36 "Power Down" on page 4-38 "Simulation Requirements" on page 4-39 "Reference Information" on page 4-39 Figure 4-1 shows the reset control and power-down block for a Stratix IV device. Figure 4-1. Reset Control and Power-Down Block tx_digitalreset rx_digitalreset rx_analogreset Reset Controller pll_powerdown gxb_powerdown (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 2: Transceivers September 2012 Feedback Subscribe 4-2 Chapter 4: Reset Control and Power Down in Stratix IV Devices User Reset and Power-Down Signals User Reset and Power-Down Signals Each transceiver channel in the Stratix IV device has individual reset signals to reset its physical coding sublayer (PCS) and physical medium attachment (PMA) blocks. Each CMU PLL in the transceiver block has a dedicated reset signal. The transceiver block also has a power-down signal that affects all the channels and CMU PLLs in the transceiver block. 1 All reset and power-down signals are asynchronous. Table 4-1 lists the reset signals available for each transceiver channel. Table 4-1. Transceiver Channel Reset Signals ALTGX MegaWizard Plug-In Manager Configurations Signal tx_digitalreset (1) Transmitter Only Receiver and Transmitter Description Provides asynchronous reset to all digital logic in the transmitter PCS, including the XAUI transmit state machine. The minimum pulse width for this signal is two parallel clock cycles. Resets all digital logic in the receiver PCS, including: rx_digitalreset (1) Receiver Only Receiver and Transmitter XAUI receiver state machines GIGE receiver state machines XAUI channel alignment state machine BIST-PRBS verifier BIST-incremental verifier The minimum pulse width for this signal is two parallel clock cycles. rx_analogreset Receiver Only Receiver and Transmitter Resets the receiver CDR present in the receiver channel. The minimum pulse width is two parallel clock cycles. Note to Table 4-1: (1) Assert this signal until the clocks coming out of the transmitter PLL and receiver CDR are stabilized. Stable parallel clocks are essential for proper operation of the transmitter and receiver phase-compensation FIFOs in the PCS. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices User Reset and Power-Down Signals 4-3 Table 4-2 lists the power-down signals available for each CMU PLL transceiver block. Table 4-2. Transceiver Block Power-Down Signals Signal pll_powerdown Description Each transceiver block has two CMU PLLs. Each CMU PLL has this dedicated power-down signal. This signal powers down the CMU PLLs that provide high-speed serial and low-speed parallel clocks to the transceiver channels. (1) Powers down the entire transceiver block. When this signal is asserted, it powers down: gxb_powerdown (1) the PCS and PMA in all the transceiver channels the CMU PLLs This signal operates independently from the other reset signals and is common to the transceiver block. A status signal. Indicates the status of the transmitter PLL. pll_locked A high level--the transmitter PLL is locked to the incoming reference clock frequency. When pll_locked is low, tx_digitalreset must always be asserted. To de-assert tx_digitalreset, follow the initialization reset sequence for your specific mode. A status signal. rx_pll_locked A high level--the receiver CDR is locked to the incoming reference clock frequency. A status signal. Indicates the status of the receiver CDR lock mode. A high level--the receiver is in lock-to-data (LTD) mode. rx_freqlocked A low level--the receiver CDR is in lock-to-reference (LTR) mode. In automatic lock mode, when rx_freqlocked is low, rx_digitalreset must always be asserted. To de-assert rx_digitalreset, follow the initialization reset sequence for your specific mode. busy A status signal. An output from the ALTGX_RECONFIG block indicates the status of the dynamic reconfiguration controller. This signal remains low for the first reconfig_clk clock cycle after power up. It then is asserted from the second reconfig_clk clock cycle. Assertion on this signal indicates that the offset cancellation process is being executed on the receiver buffer as well as the receiver CDR. When this signal is de-asserted, it indicates that offset cancellation is complete. Note to Table 4-2: (1) The refclk (refclk0 or refclk1) buffer is not powered down by this signal. f For more information about offset cancellation, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. 1 If none of the channels is instantiated in a transceiver block, the Quartus(R) II software automatically powers down the entire transceiver block. Blocks Affected by the Reset and Power-Down Signals Table 4-3 lists the blocks that are affected by specific reset and power-down signals. Table 4-3. Blocks Affected by Reset and Power-Down Signals (Part 1 of 2) Transceiver Block rx_digitalreset rx_analogreset tx_digitalreset pll_powerdown gxb_powerdown CMU PLLs -- -- -- Y Y Transmitter Phase Compensation FIFO -- -- Y -- Y September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-4 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Table 4-3. Blocks Affected by Reset and Power-Down Signals (Part 2 of 2) Transceiver Block rx_digitalreset rx_analogreset tx_digitalreset pll_powerdown gxb_powerdown Byte Serializer -- -- Y -- Y 8B/10B Encoder -- -- Y -- Y Serializer -- -- Y -- Y Transmitter Buffer -- -- -- -- Y Transmitter XAUI State Machine -- -- Y -- Y Receiver Buffer -- -- -- -- Y Receiver CDR -- Y -- -- Y Receiver Deserializer -- -- -- -- Y Receiver Word Aligner Y -- -- -- Y Receiver Deskew FIFO Y -- -- -- Y Receiver Clock Rate Compensation FIFO Y -- -- -- Y Receiver 8B/10B Decoder Y -- -- -- Y Receiver Byte Deserializer Y -- -- -- Y Receiver Byte Ordering Y -- -- -- Y Receiver Phase Compensation FIFO Y -- -- -- Y Receiver XAUI State Machine Y -- -- -- Y BIST Verifiers Y -- -- -- Y Transceiver Reset Sequences You can configure transceiver channels in Stratix IV devices in various configurations. In all functional modes except XAUI functional mode, transceiver channels can be either bonded or non-bonded. In XAUI functional mode, transceiver channels must be bonded. In PCI Express(R) (PCIe) functional mode, transceiver channels can be either bonded or non-bonded and need to follow a specific reset sequence. The two categories of reset sequences for Stratix IV devices described in this chapter are: Stratix IV Device Handbook Volume 2: Transceivers "All Supported Functional Modes Except PCIe Functional Mode" on page 4-6-- describes the reset sequences in bonded and non-bonded configurations. "PCIe Functional Mode" on page 4-22--describes the reset sequence for the initialization/compliance phase and the normal operation phase in PCIe functional modes. September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-5 1 The busy signal remains low for the first reconfig_clk clock cycle. It then is asserted from the second reconfig_clk clock cycle. Subsequent de-assertion of the busy signal indicates the completion of the offset cancellation process. This busy signal is required in transceiver reset sequences except for Transmitter Only channel configurations. For more information, refer to the reset sequences shown in Figure 4-2 and the associated references listed in the figure notes. 1 Altera strongly recommends adhering to these reset sequences for proper operation of the Stratix IV transceiver. Figure 4-2 shows the transceiver reset sequences for Stratix IV devices. Figure 4-2. Transceiver Reset Sequences Chart Transceiver initialization reset sequences All supported functional modes except PCI Express (PIPE) and PMA Direct Drive Mode Dynamic Reconfiguration PMA Direct Drive PCI Express (PIPE) Initialization/ Compliance and Normal Operation Phases (1) Reset Sequence to change the TX PLL settings the transceiver channel Reset Sequence to change the data rate of the transceiver channel Bonded `Transmitter Only' channel (2) Non-Bonded `Receiver and Transmitter' channel Receiver CDR Receiver CDR in manual in automatic lock mode lock mode (3) (4) `Transmitter Only' channel (2) xN `Receiver Only' channel `Receiver and Transmitter' channel Receiver CDR Receiver CDR Receiver CDR Receiver CDR Receiver CDR in manual in automatic in manual in automatic in automatic lock mode lock mode lock mode lock mode lock mode (8) (5) (6) (9) (7) `Transmitter Only' channel (11) Receiver CDR in manual lock mode (10) x1 `Receiver and Transmitter' channel Receiver CDR in automatic lock mode (12) Receiver CDR in manual lock mode (13) `Receiver and Transmitter' channel Receiver CDR in automatic lock mode (14) Receiver CDR in manual lock mode (15) Notes to Figure 4-2: (1) Refer to the Timing Diagram in Figure 4-12. (2) Refer to the Timing Diagram in Figure 4-3. (3) Refer to the Timing Diagram in Figure 4-4. (4) Refer to the Timing Diagram in Figure 4-5. (5) Refer to the Timing Diagram in Figure 4-6. (6) Refer to the Timing Diagram in Figure 4-7. (7) Refer to the Timing Diagram in Figure 4-8. (8) Refer to the Timing Diagram in Figure 4-9. (9) Refer to the Timing Diagram in Figure 4-10. (10) Refer to the Timing Diagram in Figure 4-11. (11) Refer to the Timing Diagram in Figure 4-13. (12) Refer to the Timing Diagram in Figure 4-16. (13) Refer to the Timing Diagram in Figure 4-17. (14) Refer to the Timing Diagram in Figure 4-18. (15) Refer to the Timing Diagram in Figure 4-19. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-6 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences All Supported Functional Modes Except PCIe Functional Mode This section describes reset sequences for transceiver channels in bonded and non-bonded configurations. Timing diagrams of some typical configurations are shown to facilitate proper reset sequence implementation. In these functional modes, you can set the receiver CDR either in automatic lock or manual lock mode. 1 In manual lock mode, the receiver CDR locks to the reference clock (lock-to-reference) or the incoming serial data (lock-to-data), depending on the logic levels on the rx_locktorefclk and rx_locktodata signals. With the receiver CDR in manual lock mode, you can either configure the transceiver channels in the Stratix IV device in a non-bonded configuration or a bonded configuration. In a bonded configuration, for example in XAUI mode, four channels are bonded together. Table 4-4 lists the lock-to-reference (LTR) and lock-to-data (LTD) controller lock modes for the rx_locktorefclk and rx_locktodata signals. Table 4-4. Lock-To-Reference and Lock-To-Data Modes rx_locktorefclk rx_locktodata LTR/LTD Controller Lock Mode 1 0 Manual, LTR Mode -- 1 Manual, LTD Mode 0 0 Automatic Lock Mode Bonded Channel Configuration In a bonded channel configuration, you can reset all the bonded channels simultaneously. Examples of bonded channel configurations are XAUI, PCIe, and Basic x4 functional modes. In Basic x4 functional mode, you can bond Transmitter Only channels together. In XAUI mode, the receiver and transmitter channels are bonded. Each of the receiver channels in this mode has its own output status signals, rx_pll_locked and rx_freqlocked. You must consider the timing of these signals in the reset sequence. Table 4-5 lists the reset and power-down sequences for bonded configurations under the stated functional modes. Table 4-5. Reset and Power-Down Sequences for Bonded Channel Configurations Channel Set Up Receiver CDR Mode Refer to Transmitter Only Basic x4 "Transmitter Only Channel" on page 4-7 Receiver and Transmitter Automatic lock mode for XAUI functional mode "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode" on page 4-8 Receiver and Transmitter Manual lock mode for XAUI functional mode "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode" on page 4-10 Receiver and Transmitter Automatic lock mode for Basic x8 functional mode "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode" on page 4-12 Receiver and Transmitter Manual lock mode for Basic x8 functional mode "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode" on page 4-14 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-7 Transmitter Only Channel This configuration contains only a transmitter channel. If you create a Transmitter Only instance in the ALTGX MegaWizard Plug-In Manager in Basic x4 functional mode, use the reset sequence shown in Figure 4-3. Figure 4-3. Sample Reset Sequence for Four Transmitter Only Channels Reset and Power-Down Signals t pll_powerdown (1) 1 2 pll_powerdown 4 tx_digitalreset Output Status Signals 3 pll_locked Note to Figure 4-3: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. As shown in Figure 4-3, for the Transmitter Only channel configuration, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset signal asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert the tx_digitalreset signal (marker 4). At this point, the transmitter is ready for transmitting data. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-8 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode This configuration contains both a transmitter and receiver channel. For XAUI functional mode, with the receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-4. Figure 4-4. Sample Reset Sequence for Four Receiver and Transmitter Channels--Receiver CDR in Automatic Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll _ powerdown 4 tx _digitalreset 6 rx _analogreset 8 rx _digitalreset Output Status Signals 3 pll _locked Minimum of Two Parallel Clock Cycles 5 busy 7 rx_freqlocked[0] 7 rx_freqlocked[3] t LTD_Auto (2) Notes to Figure 4-4: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-9 As shown in Figure 4-4, for the receiver CDR in automatic lock mode configuration, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, and rx_digitalreset signals asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high, de-assert the tx_digitalreset signal. At this point, the transmitter is ready for data traffic. 4. For the receiver operation, after de-assertion of busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal. After rx_analogreset is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock. 5. Wait for the rx_freqlocked signal from each channel to go high. The rx_freqlocked signal of each channel may go high at different times (indicated by the slashed pattern at marker 7). 6. In a bonded channel group, when the rx_freqlocked signals of all the channels has gone high, from that point onwards, wait for at least tLTD_Auto for the receiver parallel clock to be stable, then de-assert the rx_digitalreset signal (marker 8). At this point, all the receivers are ready for data traffic. Note that rx_digitalreset must not be released if there is no data present at the receiver pins to avoid overflow/underflow of the phase compensation FIFOs. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-10 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode This configuration contains both a transmitter and receiver channel. For XAUI functional mode, with the receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-5. Figure 4-5. Sample Reset Sequence for Four Receiver and Transmitter Channels--Receiver CDR in Manual Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll _ powerdown 4 tx _digitalreset 6 rx _analogreset 9 rx _digitalreset t LTD_Manual (3) CDR Control Signals 8 rx_locktorefclk[0] 8 rx_locktorefclk[3] 8 rx_locktodata[0] 8 rx_locktodata[3] Output Status Signals 3 pll _ locked Minimum of Two Parallel Clock Cycles 5 busy 7 rx_pll_locked[0] 7 rx_pll_locked[3] t LTR_LTD_Manual (2) Notes to Figure 4-5: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTR_LTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-11 As shown in Figure 4-5, for the receiver CDR in manual lock mode configuration, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, rx_digitalreset, and rx_locktorefclk signals asserted and the rx_locktodata signal de-asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert the tx_digitalreset signal (marker 4). For the receiver operation, after de-assertion of the busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal. After the rx_analogreset signal is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock because rx_locktorefclk is asserted. 4. Wait for the rx_pll_locked signal from each channel to go high. The rx_pll_locked signal of each channel may go high at different times with respect to each other (indicated by the slashed pattern at marker 7). 5. In a bonded channel group, when the rx_pll_locked signal of all the channels have gone high, from that point onwards, wait for at least tLTR_LTD_Manual, then de-assert rx_locktorefclk and assert rx_locktodata (marker 8). At this point, the receiver CDR of all the channels enters into lock-to-data mode and starts locking to the received data. 6. After asserting the rx_locktodata signal, wait for at least tLTD_Manual before de-asserting rx_digitalreset (the time between markers 8 and 9). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-12 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode This configuration contains both a transmitter and a receiver channel. For Basic x8 functional mode, with the receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-6. Figure 4-6. Sample Reset Sequence for Eight Receiver and Transmitter Channels--Receiver CDR in Automatic Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll_ powerdown 4 tx_digitalreset 6 rx_analogreset 8 rx _digitalreset Output Status Signals 3 pll _ locked Minimum of Two Parallel Clock Cycles 5 busy 7 rx_freqlocked[0] 7 rx_freqlocked[8] t LTD_Auto (2) Notes to Figure 4-6: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-13 As shown in Figure 4-6, for the receiver CDR in automatic lock mode, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, and rx_digitalreset signals asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high, de-assert the tx_digitalreset signal. At this point, the transmitter is ready for data traffic. 4. For the receiver operation, after de-assertion of the busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal. After rx_analogreset is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock. 5. Wait for the rx_freqlocked signal from each channel to go high. The rx_freqlocked signal of each channel may go high at different times (indicated by the slashed pattern at marker 7). 6. In a bonded channel group, when the rx_freqlocked signals of all the channels have gone high, from that point onwards, wait for at least tLTD_Auto for the receiver parallel clock to stabilize, then de-assert the rx_digitalreset signal (marker 8). At this point, all the receivers are ready for data traffic. Note that rx_digitalreset must not be released if there is no data present at the receiver pins to avoid overflow/underflow of the phase compensation FIFOs. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-14 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode This configuration contains both a transmitter and receiver channel. For Basic x8 functional mode, with the receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-7. Figure 4-7. Sample Reset Sequence for Eight Receiver and Transmitter Channels--Receiver CDR in Manual Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll _ powerdown 4 tx _digitalreset 6 rx _analogreset 9 rx _digitalreset t LTD_Manual (3) CDR Control Signals 8 rx_locktorefclk[0] 8 rx_locktorefclk[3] 8 rx_locktodata[0] 8 rx_locktodata[3] Output Status Signals 3 pll _ locked Minimum of Two Parallel Clock Cycles 5 busy 7 rx_pll_locked[0] 7 rx_pll_locked[7] t LTR_LTD_Manual (2) Notes to Figure 4-7: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTR_LTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-15 As shown in Figure 4-7, for the receiver CDR in manual lock mode, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, rx_digitalreset, and rx_locktorefclk signals asserted and the rx_locktodata signal de-asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert the tx_digitalreset signal (marker 4). For the receiver operation, after de-assertion of the busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal. After the rx_analogreset signal is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock because rx_locktorefclk is asserted. 4. Wait for the rx_pll_locked signal from each channel to go high. The rx_pll_locked signal of each channel may go high at different times with respect to each other (indicated by the slashed pattern at marker 7). 5. In a bonded channel group, when the rx_pll_locked signal of all the channels has gone high, from that point onwards, wait for at least tLTR_LTD_Manual, then de-assert rx_locktorefclk and assert rx_locktodata (marker 8). At this point, the receiver CDR of all the channels enters into lock-to-data mode and starts locking to the received data. 6. De-assert rx_digitalreset at least tLTD_Manual (the time between markers 8 and 9) after asserting the rx_locktodata signal. Non-Bonded Channel Configuration In non-bonded channels, each channel in the ALTGX MegaWizard Plug-In Manager instance contains its own tx_digitalreset, rx_analogreset, rx_digitalreset, rx_pll_locked, and rx_freqlocked signals. You can reset each channel independently. For example, if there are four non-bonded channels, the ALTGX MegaWizard Plug-In Manager provides four each of the following signals: tx_digitalreset, rx_analogreset, rx_digitalreset, rx_pll_locked, and rx_freqlocked. Table 4-6 lists the reset and power-down sequences for one channel in a non-bonded configuration under the stated functional modes. Table 4-6. Reset and Power-Down Sequences for Bonded Channel Configurations (Part 1 of 2) Channel Set Up Receiver CDR Mode Refer to Transmitter Only Basic x4 "Transmitter Only Channel" on page 4-16 Receiver Only Automatic lock mode "Receiver Only Channel--Receiver CDR in Automatic Lock Mode" on page 4-16 Receiver Only Manual lock mode "Receiver Only Channel--Receiver CDR in Manual Lock Mode" on page 4-17 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-16 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Table 4-6. Reset and Power-Down Sequences for Bonded Channel Configurations (Part 2 of 2) Channel Set Up Receiver CDR Mode Refer to Receiver and Transmitter Automatic lock mode "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode" on page 4-18 Receiver and Transmitter Manual lock mode "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode" on page 4-20 1 Follow the same reset sequence for all the other channels in the non-bonded configuration. Transmitter Only Channel This configuration contains only a transmitter channel. If you create a Transmitter Only instance in the ALTGX MegaWizard Plug-In Manager, use the same reset sequence shown in Figure 4-3 on page 4-7. Receiver Only Channel--Receiver CDR in Automatic Lock Mode This configuration contains only a receiver channel. If you create a Receiver Only instance in the ALTGX MegaWizard Plug-In Manager with the receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-8. Figure 4-8. Sample Reset Sequence of Receiver Only Channel--Receiver CDR in Automatic Lock Mode Reset Signals 2 rx _ analogreset 4 rx _ digitalreset Two parallel clock cycles Output Status Signals 1 busy 3 rx _ freqlocked t LTD_Auto (1) Note to Figure 4-8: (1) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-17 As shown in Figure 4-8, for the receiver in CDR automatic lock mode, follow these reset steps: 1. After power up, wait for the busy signal to be de-asserted. 2. De-assert the rx_analogreset signal. 3. Keep the rx_digitalreset signal asserted during this time period. After you de-assert the rx_analogreset signal, the receiver CDR starts locking to the receiver input reference clock. 4. Wait for the rx_freqlocked signal to go high. 5. When rx_freqlocked goes high (marker 3), from that point onwards, wait for at least tLTD_Auto, then de-assert the rx_digitalreset signal (marker 4). At this point, the receiver is ready to receive data. Receiver Only Channel--Receiver CDR in Manual Lock Mode This configuration contains only a receiver channel. If you create a Receiver Only instance in the ALTGX MegaWizard Plug-In Manager with receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-9. Figure 4-9. Sample Reset Sequence of Receiver Only Channel--Receiver CDR in Manual Lock Mode Reset Signals 2 rx _ analogreset 5 rx _ digitalreset t LTD_Manual (2) CDR Control Signals 4 rx _ locktorefclk 4 rx _ locktodata Two parallel clock cycles Output Status Signals 1 busy 3 rx _ pll _ locked t LTR_LTD_Manual (1) Notes to Figure 4-9: (1) For tLTR_LTD duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-18 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences As shown in Figure 4-9, for the receiver CDR in manual lock mode, follow these reset steps: 1. After power up, wait for the busy signal to be asserted. 2. Keep the rx_digitalreset and rx_locktorefclk signals asserted and the rx_locktodata signal de-asserted during this time period. 3. After de-assertion of the busy signal, de-assert the rx_analogreset signal. The receiver CDR then starts locking to the receiver input reference clock because the rx_locktorefclk signal is asserted. 4. Wait for at least tLTR_LTD_Manual time (the time between markers 3 and 4) after the rx_pll_locked signal goes high and then de-assert the rx_locktorefclk signal. At the same time, assert the rx_locktodata signal (marker 4). At this point, the receiver CDR enters lock-to-data mode and the receiver PLL starts locking to the received data. 5. De-assert rx_digitalreset at least tLTD_Manual (the time between markers 4 and 5) after asserting the rx_locktodata signal. Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode This configuration contains both a transmitter and a receiver channel. If you create a Receiver and Transmitter instance in the ALTGX MegaWizard Plug-In Manager with the receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-10. Figure 4-10. Sample Reset Sequence of Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll _ powerdown 4 tx _ digitalreset 6 rx _ analogreset 8 rx _ digitalreset Two parallel clock cycles Output Status Signals 5 busy 3 pll _locked 7 rx _ freqlocked t LTD_Auto (2) Notes to Figure 4-10: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-19 As shown in Figure 4-10, for the receiver in CDR automatic lock mode, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, and rx_digitalreset signals asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert tx_digitalreset. For receiver operation, wait for the busy signal to be de-asserted, after which rx_analogreset is de-asserted. After you de-assert rx_analogreset, the receiver CDR starts locking to the receiver input reference clock. 4. Wait for the rx_freqlocked signal to go high (marker 7). 5. After the rx_freqlocked signal goes high, wait for at least tLTD_Auto, then de-assert the rx_digitalreset signal (marker 8). Note that rx_digitalreset must not be released if there is no data present at the receiver pins to avoid overflow/underflow of the phase compensation FIFOs. At this point, the transmitter and receiver are ready for data traffic. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-20 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode This configuration contains both a transmitter and receiver channel. If you create a Receiver and Transmitter instance in the ALTGX MegaWizard Plug-In Manager with the receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-11. Figure 4-11. Sample Reset Sequence of Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode t pll_powerdown (1) Reset Signals 1 2 pll _ powerdown 4 tx _digitalreset 6 rx _analogreset 9 rx _digitalreset t LTD_Manual (3) CDR Control Signals 8 rx _ locktorefclk 8 rx _ locktodata Two parallel clock cycles Output Status Signals 5 busy 3 pll _ locked 7 rx _ pll _ locked t LTR_LTD_Manual (2) Notes to Figure 4-11: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTR_LTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-21 As shown in Figure 4-11, perform the following reset procedure for the receiver in manual lock mode: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the tx_digitalreset, rx_analogreset, rx_digitalreset, and rx_locktorefclk signals asserted and the rx_locktodata signal de-asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. After the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert tx_digitalreset. For receiver operation, wait for the busy signal to be de-asserted. At this point rx_analogreset is de-asserted. When rx_analogreset is de-asserted, the receiver CDR starts locking to the receiver input reference clock because rx_locktorefclk is asserted. 4. Wait for at least tLTR_LTD_Manual (the time between markers 7 and 8) after the rx_pll_locked signal goes high, then de-assert the rx_locktorefclk signal. At the same time, assert the rx_locktodata signal (marker 8). At this point, the receiver CDR enters lock-to-data mode and the receiver CDR starts locking to the received data. 5. De-assert rx_digitalreset at least tLTD_Manual (the time between markers 8 and 9) after asserting the rx_locktodata signal. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-22 Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences PCIe Functional Mode You can configure PCIe functional mode with or without the receiver clock rate compensation FIFO in the Stratix IV device. The reset sequence remains the same whether or not you use the receiver clock rate compensation FIFO. PCIe Reset Sequence The PCIe protocol consists of an initialization/compliance phase and a normal operation phase. The reset sequences for these two phases are described based on the timing diagram in Figure 4-12. Figure 4-12. Reset Sequence of PCIe Functional Mode Initialization / Compliance Phase Normal Operation Phase t pll_powerdown (1) Reset / Power Down Signals 1 2 pll _ powerdown 4 tx _ digitalreset 6 rx _ analogreset 8 12 13 rx _ digitalreset T3 (3) Two parallel clock cycles Output Status Signals 5 busy 3 pll _locked 7 rx _ pll _locked 9 10 11 rx _ freqlocked T2 (2) T1 (2) Ignore receive data Notes to Figure 4-12: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) The minimum T1 and T2 period is 4 s. (3) The minimum T3 period is two parallel clock cycles. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Transceiver Reset Sequences 4-23 PCIe Initialization/Compliance Phase After the device is powered up, a PCIe-compliant device goes through the compliance phase during initialization. In this phase, the PCIe protocol requires the system to be operating at the Gen 1 data rate. The rx_digitalreset signal must be de-asserted during this compliance phase to achieve transitions on the pipephydonestatus signal, as expected by the link layer. The rx_digitalreset signal is de-asserted based on the assertion of the rx_freqlocked signal. During the initialization/compliance phase, do not use the rx_freqlocked signal to trigger a de-assertion of the rx_digitalreset signal. Instead, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). Keep the tx_digitalreset, rx_analogreset, and rx_digitalreset signals asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 2. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), de-assert tx_digitalreset. For a receiver operation, wait for the busy signal to be de-asserted. rx_analogreset is then de-asserted. After rx_analogreset is de-asserted, the receiver CDR starts locking to the receiver input reference clock. 3. When the receiver CDR locks to the input reference clock, as indicated by the rx_pll_locked signal going high at marker 7 in Figure 4-12, de-assert the rx_digitalreset signal (marker 8). After de-asserting rx_digitalreset, the pipephydonestatus signal transitions from the transceiver channel to indicate the status to the link layer. Depending on its status, pipephydonestatus helps with the continuation of the compliance phase. After successful completion of this phase, the device enters into the normal operation phase. PCIe Normal Phase For the normal PCIe phase, follow these steps: 1. After completion of the Initialization/Compliance phase, during the normal operation phase at the Gen 1 data rate, when the rx_freqlocked signal is de-asserted (marker 10 in Figure 4-12), wait for the rx_pll_locked signal assertion signifying the lock-to-reference clock. 2. Wait for the rx_freqlocked signal to go high again. In this phase, the received data is valid (not electrical idle) and the receiver CDR locks to the incoming data. Proceed with the reset sequence after assertion of the rx_freqlocked signal. 3. After the rx_freqlocked signal goes high, wait for at least 4 s before asserting rx_digitalreset (marker 12 in Figure 4-12) for two parallel receive clock cycles so that the receiver phase compensation FIFO is initialized. 4. During normal operation, after you speed-negotiate to the Gen 2 data rate, asserting the rx_digitalreset signal causes the PCIe rate switch circuitry to switch the transceiver to the Gen 1 data rate. Data from the transceiver block is not valid from the time the rx_freqlocked signal goes low (marker 10 in Figure 4-12) to the time rx_digitalreset is de-asserted (marker 13 in Figure 4-12). The PLD logic ignores the data during this period (between markers 10 and 13 in Figure 4-12). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-24 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 1 You can configure the Stratix IV device in x1, x4, and x8 PCIe configurations. The reset sequence described in "PCIe Reset Sequence" on page 4-22 applies to all these multi-lane configurations. PMA Direct Drive Mode Reset Sequences Stratix IV devices provide a PMA Direct mode in which all PCS blocks, including the phase compensation FIFOs, are bypassed in both the transmitter and receiver channel data paths. In this mode, the PMA block in the transmitter and receiver channels directly interface with the FPGA fabric. In PMA Direct drive mode, you can configure the transceiver channels as a single channel or in bonded configurations. Basic single- and double-width functional modes support bonding of PMA functional blocks across all transceiver channels on the same side of the device. 1 The tx_digitalreset and rx_digitalreset signals are not available because there are no PCS blocks available in this mode. Table 4-7 lists the reset and power-down sequences for PMA Direct drive xN functional mode. Table 4-7. Reset and Power-Down Sequences for PMA Direct Drive xN Configurations Channel Set Up Functional Mode Refer to Transmitter Only with no PLL_L/R Basic (PMA Direct) drive x4 "Transmitter Only Channel with No PLL_L/R" on page 4-25 Transmitter Only with a PLL_L/R Manual lock mode "Transmitter Only Channel with a PLL_L/R" on page 4-26 Receiver and Transmitter Automatic lock mode for Basic (PMA Direct) drive xN mode "Receiver and Transmitter Channel Set-up-- Receiver CDR in Automatic Lock Mode" on page 4-28 Receiver and Transmitter Manual lock mode for Basic (PMA Direct) drive xN mode "Receiver and Transmitter Channel Set-up-- Receiver CDR in Manual Lock Mode" on page 4-30 Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-25 Basic (PMA Direct) Drive xN Mode When bonding xN channels in a Basic (PMA Direct) drive mode configuration, you can reset all bonded channels simultaneously. Transmitter Only Channel with No PLL_L/R Figure 4-13 shows an example reset sequence timing diagram of four Transmitter Only channels in Basic (PMA Direct) drive x4 functional mode with no PLL_L/R. Figure 4-13. Reset Sequence Timing in Basic (PMA Direct) Drive x4 Mode Reset and Power-Down Signals t pll_powerdown (1) 1 2 pll_powerdown Ouput Status Signals 3 pll_locked Note to Figure 4-13: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. As shown in Figure 4-13, for the Transmitter Only channel in Basic (PMA Direct) drive functional x4 mode with no PLL_L/R, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum of tpll_powerdown (the time between markers 1 and 2). 2. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), the transmitter is ready to accept parallel data from the FPGA fabric and subsequently transmitting serial data reliably. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-26 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences Transmitter Only Channel with a PLL_L/R The Basic (PMA Direct) mode configuration that requires a PLL_L/R is one where each channel in PMA-Direct mode is identical. Figure 4-14 shows a simple set up of identical channels. Figure 4-14. Identical Channels Transmitter Channels in Basic (PMA Direct) Mode CH0 (2.5 Gbps) locked inclk0 c0 Left and Right PLL (ALTPLL) CH1 (2.5 Gbps) pll_locked tx_clkout High-Speed Serial Clock CH2 (2.5 Gbps) Transmitter Side User Logic in the FPGA Fabric tx_datain[9:0] CH3 (2.5 Gbps) refclk CMU Channel configured for clock generation (2.5 Gbps) Identical channels have the following same configuration: Same effective data rate Same transmitter local clock divider settings in each channel Same FPGA fabric-to-transceiver interface data path width The transmitter channels must receive the high-speed clock from the same PLL (either CMU PLL or ATX PLL). Figure 4-15 shows an example reset sequence timing diagram of four Transmitter Only channels in Basic (PMA Direct) Drive x4 functional mode with a PLL_L/R. As shown in Figure 4-15, for the Transmitter Only channel in Basic (PMA Direct) Drive functional mode with a PLL_L/R configuration, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum of tpll_powerdown (the time between markers 1 and 2). 2. After the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), wait for the locked signal to be asserted. The locked signal is an output of the PLL_L/R. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-27 3. After the PLL_L/R locks, as indicated by the locked signal going high (marker 4), the transmitter is ready to accept parallel data from the FPGA fabric and subsequently transmitting serial data reliably. Figure 4-15. Reset Sequence Timing Diagram of Four Transmitter-Only Channels in Basic (PMA Direct) Drive x4 Functional Mode Reset and Power-Down Signals t pll_powerdown (1) 1 2 pll_powerdown Output Status Signals 3 pll_locked 4 Locked (output of PLL_L/R) Keep the TX side user logic under reset until this point Note to Figure 4-15: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-28 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences Receiver and Transmitter Channel Set-up--Receiver CDR in Automatic Lock Mode This configuration contains both a transmitter and receiver channel. For PMA Direct drive xN mode, with the receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-16. In this example, N = 4. Figure 4-16. Reset Sequence with CDR in Automatic Lock Mode Reset and Power Down Signals t pll_powerdown (1) 1 2 pll_powerdown 5 rx_analogreset[0] 5 rx_analogreset[3] Ouput Status Signals 3 pll_locked Minimum of Two Parallel Clock Cycles 4 busy 6 rx_freqlocked[0] 6 rx_freqlocked[3] 7 valid parallel data into FPGA fabric rx_dataout[63:0] t LTD_Auto (2) Notes to Figure 4-16: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-29 As shown in Figure 4-16, for the receiver and transmitter channel in PMA Direct drive x4 double-width configuration with CDR in automatic lock mode, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown(the time between markers 1 and 2). 2. Keep the rx_analogreset signal asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), the transmitter is ready to accept parallel data from the FPGA fabric and transmitting serial data reliably. 4. For the receiver operation, after de-assertion of the busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signals of each channel. After rx_analogreset is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock. 5. Wait for the rx_freqlocked signal from each channel to go high. The rx_freqlocked signal of each channel may go high at different times (as indicated by the slashed pattern at marker 6). 6. In a PMA Direct drive x4 double-width configuration, when the rx_freqlocked signals of all the channels has gone high (marker 6), from that point onwards, wait for at least tLTD_Auto (marker 7) for the receiver parallel clock to become stable. At this point, all the receivers are ready for transferring valid parallel data into the FPGA fabric (until this time, Altera recommends that the user logic that processes this data be under reset). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-30 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences Receiver and Transmitter Channel Set-up--Receiver CDR in Manual Lock Mode This configuration contains both a transmitter and receiver channel. For PMA Direct drive xN mode, with receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-17. In this example, N = 4. Figure 4-17. Reset Sequence with CDR in Manual Lock Mode Reset and Power Down Signals t pll_powerdown (1) 1 2 pll_powerdown 5 rx_analogreset[0] 5 rx_analogreset[3] CDR Control Signals 7 rx_locktorefclk[0] 7 rx_locktorefclk[3] 7 rx_locktodata[0] 7 rx_locktodata[3] Ouput Status Signals 3 pll_locked Minimum of Two Parallel Clock Cycles 4 busy 6 rx_pll_locked[0] 6 rx_pll_locked[3] t LTR_LTD_Manual (2) 8 valid parallel data into FPGA fabric rx_dataout[63:0] t LTD_Manual (3) Notes to Figure 4-17: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTR_LTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-31 As shown in Figure 4-17, for the receiver and transmitter channel in PMA Direct drive x4 double-width configuration with CDR in manual lock mode, follow these reset steps: 1. After power up, assert pll_powerdown for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the rx_analogreset and rx_locktorefclk signals asserted and the rx_locktodata signal de-asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), the transmitter is ready to accept parallel data from the FPGA fabric and transmitting serial data reliably. 4. For the receiver operation, after de-assertion of the busy signal (marker 4), wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal. After the rx_analogreset signal is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock because rx_locktorefclk is asserted. 5. Wait for the rx_pll_locked signal from each channel to go high. The rx_pll_locked signal of each channel may go high at different times with respect to each other (indicated by the slashed pattern at marker 6). 6. In a PMA Direct drive x4 double-width configuration, when the rx_pll_locked signal of all the channels has gone high, from that point onwards, wait for at least tLTR_LTD_Manual, then de-assert rx_locktorefclk and assert rx_locktodata (marker 7). At this point, the receiver CDR of all the channels enters into lock-to-data mode and starts locking to the received data. 7. After assertion of the rx_locktodata signal, from that point onwards, wait for at least tLTD_Manual (marker 8) for the receiver parallel clock to become stable. At this point, all the receivers are ready for transferring valid parallel data into the FPGA fabric (until this time, Altera recommends that the user logic that processes this data be under reset). Basic (PMA Direct) Drive x1 Mode The following timing diagram examples are used to describe the reset and power down sequences for Basic (PMA Direct) drive mode without bonding between the transceiver channels. Table 4-8 lists the reset and power-down sequences for Basic (PMA Direct) drive x1 functional mode. Table 4-8. Reset and Power-Down Sequences for Basic (PMA Direct) Drive x1 Configurations Channel Set Up Functional Mode Refer to Receiver and Transmitter Automatic lock mode for Basic (PMA Direct) drive x1 mode "Receiver and Transmitter Channel Set-Up--Receiver CDR in Automatic Lock Mode" on page 4-32 Receiver and Transmitter Manual lock mode for Basic (PMA Direct) drive x1 mode "Receiver and Transmitter Channel Set-up--Receiver CDR in Manual Lock Mode" on page 4-34 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-32 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences Receiver and Transmitter Channel Set-Up--Receiver CDR in Automatic Lock Mode This configuration contains both a transmitter and receiver channel. For Basic (PMA Direct) drive x1 mode, with receiver CDR in automatic lock mode, use the reset sequence shown in Figure 4-18. In this example, four channels are configured in this mode. Figure 4-18. Reset Sequence with CDR in Automatic Lock Mode Reset and Power Down Signals t pll_powerdown (1) pll_powerdown[0] 1 2 pll_powerdown[3] 5 rx_analogreset[0] 5 rx_analogreset[3] Ouput Status Signals 3 pll_locked[0] 3 pll_locked[3] busy Minimum of Two Parallel Clock Cycles 4 6 rx_freqlocked[0] 6 rx_freqlocked[3] 7 valid parallel data into FPGA fabric rx_dataout[63:0] t TLD_Auto (2) Notes to Figure 4-18: (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-33 As shown in Figure 4-18, for the receiver and transmitter channel in Basic (PMA Direct) drive double-width configuration, non-bonded with CDR in automatic lock mode, follow these reset steps: 1. After power up, assert pll_powerdown of each channel for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the rx_analogreset signal of each channel asserted during this time period. After you de-assert the pll_powerdown signal on all channels, the transmitter PLL of each channel starts locking to the transmitter input reference clock. 3. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), the transmitters are ready for accepting parallel data from the FPGA fabric and subsequently transmitting serial data reliably. 4. For the receiver operation, after de-assertion of the busy signal, wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signals of each channel. After rx_analogreset is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock. 5. Wait for the rx_freqlocked signal from each channel to go high. The rx_freqlocked signal of each channel may go high at different times (indicated by the slashed pattern at marker 6). 6. In a Basic (PMA Direct) drive double-width configuration without bonding between channels, when the rx_freqlocked signals of all the channels have gone high (marker 6), from that point onwards, wait for at least tLTD_Auto (marker 7) for the receiver parallel clock to become stable. At this point, all the receivers are ready for transferring valid parallel data into the FPGA fabric (until this time, Altera recommends that the user logic that processes this data be under reset). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-34 Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences Receiver and Transmitter Channel Set-up--Receiver CDR in Manual Lock Mode This configuration contains both a transmitter and receiver channel. For Basic (PMA Direct) drive x1 mode, with receiver CDR in manual lock mode, use the reset sequence shown in Figure 4-19. In this example, four channels are configured in this mode. Figure 4-19. Reset Sequence with CDR in Manual Lock Mode Reset and Power Down Signals t pll_powerdown (1) 1 2 pll_powerdown[0] pll_powerdown[3] rx_analogreset[0] 5 rx_analogreset[3] 5 CDR Control Signals 7 rx_locktorefclk[0] 7 rx_locktorefclk[3] 7 rx_locktodata[0] 7 rx_locktodata[3] Ouput Status Signals 3 pll_locked Minimum of Two Parallel Clock Cycles 4 busy 6 rx_pll_locked[0] 6 rx_pll_locked[3] t LTR_LTD_Manual (2) 8 rx_dataout[63:0] valid parallel data into FPGA fabric t LTD_Manual (3) Notes to Figure 4-19 (1) For tpll_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTR_LTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (3) For tLTD_Manual duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices PMA Direct Drive Mode Reset Sequences 4-35 As shown in Figure 4-19, for the receiver and transmitter channel in Basic (PMA Direct) drive double-width configuration, non-bonded with CDR in manual lock mode, follow these reset steps: 1. After power up, assert pll_powerdown of each channel for a minimum period of tpll_powerdown (the time between markers 1 and 2). 2. Keep the rx_analogreset and rx_locktorefclk signals of each channel asserted and the rx_locktodata signals de-asserted during this time period. After you de-assert the pll_powerdown signal, the transmitter PLL starts locking to the transmitter input reference clock. 3. When the transmitter PLL locks, as indicated by the pll_locked signal going high (marker 3), the transmitters are ready to accept parallel data from the FPGA fabric and subsequently transmitting serial data reliably. 4. For the receiver operation, after de-assertion of the busy signal (marker 4), wait for a minimum of two parallel clock cycles to de-assert the rx_analogreset signal of each channel. After the rx_analogreset signal is de-asserted, the receiver CDR of each channel starts locking to the receiver input reference clock because rx_locktorefclk is asserted. 5. Wait for the rx_pll_locked signal from each channel to go high. The rx_pll_locked signal of each channel may go high at different times with respect to each other (indicated by the slashed pattern at marker 6). 6. In a Basic (PMA Direct) drive double-width configuration without bonding between channels, when the rx_pll_locked signal of all the channels has gone high, from that point onwards, wait for at least tLTR_LTD_Manual, then de-assert rx_locktorefclk and assert rx_locktodata (marker 7). At this point, the receiver CDR of all the channels enters into lock-to-data mode and starts locking to the received data. 7. After assertion of the rx_locktodata signal, from that point onwards, wait for at least tLTD_Manual (marker 8) for the receiver parallel clock to be stable. At this point, all the receivers are ready for transferring valid parallel data into the FPGA fabric (until this time, Altera recommends that the user logic that processes this data be reset). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-36 Chapter 4: Reset Control and Power Down in Stratix IV Devices Dynamic Reconfiguration Reset Sequences Dynamic Reconfiguration Reset Sequences When using dynamic reconfiguration in data rate divisions in TX or channel and TX CMU PLL select/reconfig modes, use the following reset sequences. Reset Sequence when Using Dynamic Reconfiguration with the `data rate division in TX' Option Use the example reset sequence shown in Figure 4-20 when you use the dynamic reconfiguration controller to change the data rate of the transceiver channel. In this example, dynamic reconfiguration is used to dynamically reconfigure the data rate of the transceiver channel configured in Basic x1 mode with the receiver CDR in automatic lock mode. Figure 4-20. Reset Sequence When Using the Dynamic Reconfiguration Controller to Change the Data Rate of the Transceiver Channel Reset and Control Signals 1 4 tx_digitalreset rate_switch_ctrl[1:0] New value 1 write_all Ouput Status Signals 2 3 busy As shown in Figure 4-20, when using the dynamic reconfiguration controller to change the configuration of the transmitter channel, follow these reset steps: 1. After power up and properly establishing that the transmitter is operating as desired, write the desired new value for the data rate in the appropriate register (in this example, rate_switch_ctrl[1:0]) and subsequently assert the write_all signal (marker 1) to initiate the dynamic reconfiguration. f For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. 2. Assert the tx_digitalreset signal. 3. As soon as write_all is asserted, the dynamic reconfiguration controller starts to execute its operation. This is indicated by the assertion of the busy signal (marker 2). 4. After the completion of dynamic reconfiguration, the busy signal is de-asserted (marker 3). 5. Lastly, tx_digitalreset can be de-asserted to continue with the transmitter operation (marker 4). Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Dynamic Reconfiguration Reset Sequences 4-37 Reset Sequence when Using Dynamic Reconfiguration with the `Channel and TX PLL select/reconfig' Option Use the example reset sequence shown in Figure 4-21 when you are using the dynamic reconfiguration controller to change the TX PLL settings of the transceiver channel. In this example, the dynamic reconfiguration is used to dynamically reconfigure the data rate of the transceiver channel configured in Basic x1 mode with receiver CDR in automatic lock mode. Figure 4-21. Reset Sequence When Using the Dynamic Reconfiguration Controller to Change the TX PLL Settings of the Transceiver Channel Reset and Control Signals 1 5 tx_digitalreset 1 6 rx_analogreset 1 8 rx_digitalreset reconfig_mode_sel[2:0] New value 1 write_all Ouput Status Signals Five parallel clock cycles 2 3 busy 4 channel_reconfig_done 7 rx_freqlocked t LTD_Auto (1) Note to Figure 4-21: (1) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. As shown in Figure 4-21, when using the dynamic reconfiguration controller to change the configuration of the transceiver channel, follow these reset steps: 1. After power up and establishing that the transceiver is operating as desired, write the desired new value in the appropriate registers (including reconfig_mode_sel[2:0]) and subsequently assert the write_all signal (marker 1) to initiate the dynamic reconfiguration. f For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. 2. Assert the tx_digitalreset, rx_analogreset, and rx_digitalreset signals. 3. As soon as write_all is asserted, the dynamic reconfiguration controller starts to execute its operation. This is indicated by the assertion of the busy signal (marker 2). 4. Wait for the assertion of the channel_reconfig_done signal (marker 4) that indicates the completion of dynamic reconfiguration in this mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-38 Chapter 4: Reset Control and Power Down in Stratix IV Devices Power Down 5. After assertion of the channel_reconfig_done signal, de-assert tx_digitalreset (marker 5) and wait for at least five parallel clock cycles to de-assert the rx_analogreset signal (marker 6). 6. Lastly, wait for the rx_freqlocked signal to go high. After rx_freqlocked goes high (marker 7), wait for tLTD_Auto to de-assert the rx_digitalreset signal (marker 8). At this point, the receiver is ready for data traffic. Power Down The Quartus II software automatically selects the power-down channel feature, which takes effect when you configure the Stratix IV device. All unused transceiver channels and blocks are powered down to reduce overall power consumption. The gxb_powerdown signal is an optional transceiver block signal. It powers down all transceiver channels and all functional blocks in the transceiver block. The minimum pulse width for this signal is 1 s. After power up, if you use the gxb_powerdown signal, wait for de-assertion of the busy signal, then assert the gxb_powerdown signal for a minimum of 1 s. Lastly, follow the sequence shown in Figure 4-22. The de-assertion of the busy signal indicates proper completion of the offset cancellation process on the receiver channel. Figure 4-22. Sample Reset Sequence of Four Receiver and Transmitter Channels-Receiver CDR in Automatic Lock Mode with the Optional gxb_powerdown Signal t gxb_powerdown (1) Reset/Power Down Signals 2 3 gxb_powerdown pll_powerdown 5 tx_digitalreset 6 rx_analogreset 8 rx_digitalreset Output Status Signals 1 busy 4 pll_locked 7 rx_freqlocked t LTD_Auto (2) Notes to Figure 4-22: (1) For tgxb_powerdown duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. (2) For tLTD_Auto duration, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 4: Reset Control and Power Down in Stratix IV Devices Simulation Requirements 4-39 Simulation Requirements The following are simulation requirements: The gxb_powerdown port is optional. In simulation, if the gxb_powerdown port is not instantiated, you must assert the tx_digitalreset, rx_digitalreset, and rx_analogreset signals appropriately for correct simulation behavior. If the gxb_powerdown port is instantiated, and the other reset signals are not used, you must assert the gxb_powerdown signal for at least one parallel clock cycle for correct simulation behavior. You can de-assert the rx_digitalreset signal immediately after the rx_freqlocked signal goes high to reduce the simulation run time. It is not necessary to wait for tLTD_Auto (as suggested in the actual reset sequence). The busy signal is de-asserted after about 20 parallel reconfig_clk clock cycles in order to reduce simulation run time. For silicon behavior in hardware, you can follow the reset sequences described in the previous pages. In PCIe mode simulation, you must assert the tx_forceelecidle signal for at least one parallel clock cycle before transmitting normal data for correct simulation behavior. Reference Information For more information about some useful reference terms used in this chapter, refer to the links listed in Table 4-9. Table 4-9. Reference Information (Part 1 of 2) September 2012 Terms Used in this Chapter Useful Reference Points Automatic Lock Mode page 4-8 Basic (PMA Direct) Drive x1 Mode page 4-31 Basic (PMA Direct) Drive xN Mode page 4-25 Bonded channel configuration page 4-6 busy page 4-3 Dynamic Reconfiguration Reset Sequences page 4-36 gxb_powerdown page 4-3 LTD page 4-6 LTR page 4-6 Manual Lock Mode page 4-10 Non-Bonded channel configuration page 4-15 PCIe page 4-22 pll_locked page 4-3 pll_powerdown page 4-3 rx_analogreset page 4-2 rx_digitalreset page 4-2 rx_freqlocked page 4-3 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 4-40 Chapter 4: Reset Control and Power Down in Stratix IV Devices Reference Information Table 4-9. Reference Information (Part 2 of 2) Terms Used in this Chapter Useful Reference Points rx_pll_locked page 4-3 tx_digitalreset page 4-2 Document Revision History Table 4-10 lists the revision history for this chapter. Table 4-10. Document Revision History Date Version Changes September 2012 4.3 Updated Table 4-2 to close FB #65274. December 2011 4.2 Updated Table 4-2. February 2010 November 2009 June 2009 Updated the "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode", "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode", "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode", "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode", "Receiver and Transmitter Channel--Receiver CDR in Automatic Lock Mode", "Receiver and Transmitter Channel--Receiver CDR in Manual Lock Mode", "Receiver and Transmitter Channel Set-up--Receiver CDR in Automatic Lock Mode", "Receiver and Transmitter Channel Set-up--Receiver CDR in Manual Lock Mode" Updated Figure 4-4, Figure 4-5, Figure 4-6, Figure 4-7, Figure 4-16, Figure 4-17, Figure 4-18, and Figure 4-19. Updated Table 4-2. Updated chapter title. Applied new template. Minor text edits. Updated all figures (except Figure 1, Figure 2, and Figure 14) and all sections so they use the same terms that are found in the DC and Switching Characteristics chapter in the Stratix IV Device Datasheet section. Added Table 4-1, Table 4-2, Table 4-5, Table 4-6, Table 4-7, and Table 4-8. Added the "Reference Information" section. Updated all figures (except Figure 1). Changed "PLL_powerdown" to "pll_powerdown" throughout. Minor text edits. Added new "Transmitter Only Channel with a PLL_L/R" section. Updated the "Transmitter Only Channel with no PLL_L/R" and "Transmitter Only Channel" sections. Minor text edits. 4.1 4.0 3.1 Added: March 2009 November 2008 Stratix IV Device Handbook Volume 2: Transceivers 3.0 2.0 "PMA Direct Drive Mode Reset Sequences" "Dynamic Reconfiguration Reset Sequences" Added chapter to the Stratix IV Device Handbook. September 2012 Altera Corporation 5. Dynamic Reconfiguration in Stratix IV Devices September 2012 SIV52005-3.4 SIV52005-3.4 Stratix(R) IV GX and GT transceivers allow you to dynamically reconfigure different portions of the transceivers without powering down any part of the device. This chapter describes and provides examples about the different modes available for dynamic reconfiguration. You can use the ALTGX_RECONFIG instance to reconfigure the physical medium attachment (PMA) controls, functional blocks, clock multiplier unit (CMU) phase-locked loops (PLLs), receiver clock data recovery (CDR), and input reference clocks of a transceiver channel. Additionally, you can monitor the receiver eye width, implement decision feedback control, and achieve adaptive equalization (AEQ) control with dynamic reconfiguration. This chapter contains the following sections: "Glossary of Terms" on page 5-1 "Dynamic Reconfiguration Controller Architecture" on page 5-3 "Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration" on page 5-4 "Dynamic Reconfiguration Modes Implementation" on page 5-12 "Dynamic Reconfiguration Controller Port List" on page 5-78 "Error Indication During Dynamic Reconfiguration" on page 5-90 "Dynamic Reconfiguration Duration" on page 5-91 "Dynamic Reconfiguration (ALTGX_RECONFIG Instance) Resource Utilization" on page 5-94 "Functional Simulation of the Dynamic Reconfiguration Process" on page 5-95 "Dynamic Reconfiguration Examples" on page 5-96 Glossary of Terms Table 5-1 lists the terms used in this chapter: Table 5-1. Glossary of Terms Used in this Chapter (Part 1 of 2) Term Description AEQ Control Logic AEQ control logic is soft IP that you can enable in the dynamic reconfiguration controller. AEQ Hardware AEQ hardware is circuitry that you can enable in the receiver portion of the transceivers. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 2: Transceivers September 2012 Feedback Subscribe 5-2 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Glossary of Terms Table 5-1. Glossary of Terms Used in this Chapter (Part 2 of 2) Term Description ALTGX_RECONFIG Instance Dynamic reconfiguration controller instance generated by the ALTGX_RECONFIG MegaWizardTM Plug-In Manager. ALTGX Instance Transceiver instance generated by the ALTGX MegaWizard Plug-In Manager. Alternate CMU Transmitter PLL Refers to one of the two CMU PLLs within a transceiver block. Refers to the following dynamic reconfiguration modes: Channel and Transmitter PLL Select/reconfig Mode CMU PLL reconfiguration Channel and CMU PLL reconfiguration Channel reconfiguration with transmitter PLL select Central control unit reconfiguration Logical Channel Addressing Used whenever the concept of logical channel addressing is explained. This term does not refer to the logical_channel_address port available in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Logical Reference Index Refers to the logical identification value that you must set up for the transmitter PLLs used in the design. You can use a set up value of 0, 1, 2 or 3 in the Reconfiguration Settings screen of the ALTGX MegaWizard Plug-In Manager. Logical tx pll Refers to the logical reference index value of the transmitter PLLs stored in the memory initialization file (.mif). Main PLL Refers to the transmitter PLL configured in the General screen of the ALTGX MegaWizard Plug-In Manager. Memory Initialization File, also known as .mif When you enable .mif generation in your design, a file with the .mif extension is generated. This file contains information about the various ALTGX MegaWizard Plug-In Manager options that you set. Each word in the .mif is 16 bits wide. The dynamic reconfiguration controller writes information from the .mif into the transceiver channel, but only when you use a reconfiguration mode that supports .mif-based reconfiguration. PMA Controls Represents analog controls (Voltage Output Differential [VOD], Pre-emphasis, and Manual Equalization) as displayed in both the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers. PMA-Only Channels Channels configured in Basic (PMA Direct) functional mode. Regular Transceiver Channel Refers to a transmitter channel, a receiver channel, or a duplex channel that has both PMA and physical coding sublayer (PCS) blocks. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Architecture 5-3 Dynamic Reconfiguration Controller Architecture The dynamic reconfiguration controller is a soft IP that utilizes FPGA-fabric resources. You can use only one controller per transceiver block. You cannot use the dynamic reconfiguration controller to control multiple Stratix IV devices or any off-chip interfaces. Figure 5-1 shows a conceptual view of the dynamic reconfiguration controller architecture. For a detailed description of the inputs and outputs of the ALTGX_RECONFIG instance, refer to "Dynamic Reconfiguration Controller Port List" on page 5-78. Figure 5-1. Dynamic Reconfiguration Controller ALTGX MegaWizard Plug-In Manager ALTGX_RECONFIG MegaWizard Plug-In Manager ALTGX_RECONFIG Instance (Dynamic Reconfiguration Controller) reconfig_clk read write_all ALTGX Instances PMA controls reconfig logic PMA control ports (1) Data Rate Switch control logic rate_switch_ctrl[1:0](TX only) reset_reconfig_address CMU PLL Reconfig control logic reconfig_data[15:0] reconfig_fromgxb[] Channel and CMU PLL reconfig control logic Channel reconfig with TX PLL select control logic logical_tx_pll_sel logical_tx_pll_sel_en reconfig_togxb[3:0] data valid addr Address Translation data reconfig_address[5:0] (2) logical_channel_address[] rx_tx_duplex_sel[] Central Control Unit reconfig logic Offset Cancellation control logic ctrl_write ctrl_read ctrl_address[15:0] ctrl_writedata[15:0] EyeQ control logic AEQ control logic reconfig_reset Parallel to Serial Converter busy error rate_switch_out_[1:0] (TX only) reconfig_address_out[6:0] reconfig_address_en channel_reconfig_done ctrl_readdata[15:0] ctrl_waitrequest aeq_togxb[] aeq_fromgxb[] reconfig_mode_sel[] Notes to Figure 5-1: (1) The PMA control ports consist of the VOD, pre-emphasis, DC gain, and manual equalization controls. (2) For more information, refer to Table 5-16 on page 5-78. 1 September 2012 You can use only one ALTGX_RECONFIG instance per transceiver block. You may use a single ALTGX_RECONFIG instance with multiple transceiver blocks. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-4 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration Stratix IV GX devices provide two MegaWizard Plug-In Manager interfaces to support dynamic reconfiguration--ALTGX and ALTGX_RECONFIG. ALTGX MegaWizard Plug-In Manager Use the ALTGX MegaWizard Plug-In manager to enable the dynamic reconfiguration settings for the transceiver instances. f For more information, refer to the "Reconfiguration Settings" section of the ALTGX Transceiver Setup Guide for Stratix IV Devices chapter. The reconfig_clk Clock Requirements for the ALTGX Instance You must connect the reconfig_clk port to the ALTGX instance in all the configurations using the dynamic reconfiguration feature. Table 5-2 lists the source clock for the offset cancellation circuit in the ALTGX instance, based on its configuration. Table 5-2. Source Clock for the Offset Cancellation Circuit in the ALTGX Instance Source Clock for the Offset Cancellation Circuit (1) ALTGX Configurations reconfig_clk Receiver only and Transmitter only reconfig_clk Receiver and Transmitter fixedclk PCI Express(R) (PCIe) Note to Figure 5-2: (1) The clock source used for offset cancellation must be a free running clock that is not derived from the PLL as this clock is required for offset cancellation at power up. Select the reconfig_clk frequency based on the ALTGX configuration shown in Table 5-3. This clock must be a free-running clock sourced from an I/O clock pin. Do not use dedicated transceiver REFCLK pins or any clocks generated by transceivers. 1 Altera recommends driving the reconfig_clk signal on a global clock resource. This clock must be a free-running clock sourced from an I/O clock pin. Do not use dedicated transceiver refclk pins or any clocks generated by transceivers. Table 5-3. reconfig_clk Settings for the ALTGX Instance ALTGX Instance Configuration reconfig_clk Frequency Range (MHz) Receiver and Transmitter 37.5 to 50 Receiver only 37.5 to 50 Transmitter only and PCIe 2.5 (1) to 50 Note to Figure 5-3: (1) The source clock for the offset cancellation circuit in the ALTGX instance must be faster than 37.5 MHz. Offset cancellation is not required for transmitters and is accomplished using a fixed clock in PCIe mode. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration 5-5 ALTGX_RECONFIG MegaWizard Plug-In Manager Use the ALTGX_RECONFIG MegaWizard Plug-In Manager to instantiate the dynamic reconfiguration controller. f For more information, refer to the Stratix IV ALTGX_RECONFIG Megafunction User Guide. The reconfig_clk Clock Requirements for the ALTGX_RECONFIG Instance You must connect the reconfig_clk input port of the ALTGX_RECONFIG instance to the same clock that is connected to the reconfig_clk input port of the ALTGX instance. Table 5-3 on page 5-4 lists the range of frequency values allowed for the reconfig_clk input port for the Receiver only, Receiver and Transmitter, and Transmitter only configuration modes of the ALTGX instance. Based on the ALTGX configurations (Receiver only, Transmitter only, and Receiver and Transmitter) controlled by the ALTGX_RECONFIG instance, select the fastest reconfig_clk frequency value. This satisfies both the offset cancellation control for the receiver channels and the dynamic reconfiguration of the transmitter and receiver channels. Interfacing ALTGX and ALTGX_RECONFIG Instances To dynamically reconfigure the transceiver channel, you must understand the concepts related to interfacing the transceivers with the dynamic reconfiguration controller. These concepts are: "Logical Channel Addressing" on page 5-5 "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10 "Connecting the ALTGX and ALTGX_RECONFIG Instances" on page 5-11 Logical Channel Addressing The dynamic reconfiguration controller identifies a transceiver channel by using the logical channel address. The What is the starting channel number? option in the ALTGX MegaWizard Plug-In Manager allows you to set the logical channel address of all the channels within the ALTGX instance. For channel reconfiguration with transmitter PLL select mode, the logical channel addressing concept extends to transmitter PLLs. For more information, refer to "Logical Channel Addressing When Using Additional PLLs" on page 5-52. The following sections describe the concept of logical channel addressing for ALTGX instances configured with: September 2012 Regular transceiver channels (PCS and PMA channels) PMA-only channels A combination of PMA-only channels and regular transceiver channels Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-6 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration Logical Channel Addressing of Regular Transceiver Channels For a single ALTGX instance connected to the dynamic reconfiguration controller, set the starting channel number to 0. The logical channel addresses of the first channel within the ALTGX instance is 0. The logical channel addresses of the remaining channels increment by one. For multiple ALTGX instances connected to the dynamic reconfiguration controller, set the starting channel number of the first instance to 0. For the starting channel number for the following ALTGX instances, you must set the next multiple of four. The logical channel address of channels within each ALTGX instance increment by one. Figure 5-2 shows how to set the starting channel number for multiple ALTGX instances controlled by a single dynamic reconfiguration controller, where both ALTGX instances have regular transceiver channels. Figure 5-2. Logical Channel Addressing of Regular Transceiver Channels ALTGX instance 1 Five regular transceiver channels Basic Functional Mode Starting channel number = 0 Set the What is the number of channels controlled by the reconfig controller? option to 12 (1) Channel 0 (logical channel address = 0) reconfig_fromgxb 1[33:0] Channel 1 (logical channel address = 1) reconfig_togx Channel 2 (logical channel address = 2 reconfig_fromgxb[50:0] (2) Channel 3 (logical channel address = 3) ALTGX_RECONFIG Instance 1 Channel 4 (logical channel address = 4) ALTGX instance 2 Two regular transceiver channels Basic Functional Mode Starting channel number = 8 Channel 0 (logical channel address = 8) reconfig_fromgxb 2[16:0] Channel 1 (logical channel address = 9) Notes to Figure 5-2: (1) For more information, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. (2) reconfig_fromgxb[50:0] = { reconfig_fromgxb 2[16:0], reconfig_fromgxb 1[33:0]}. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration 5-7 Logical Channel Addressing of PMA-Only Channels 1 CMU channels are always PMA-only channels. The regular transceiver channels can be optionally configured as PMA-only channels. Set the starting channel number for the PMA-only channels in the What is the starting channel number? option in the ALTGX MegaWizard Plug-In Manager. For a single ALTGX instance connected to the dynamic reconfiguration controller, set the starting channel number to 0. The logical channel address of the first channel in the ALTGX instance is 0. The logical channel addresses of the PMA-only channels within the same ALTGX instance increment in multiples of four (unlike the logical channel addressing of regular transceiver channels that are not configured in Basic [PMA Direct] functional mode, where the logical channel address increments in steps of one within the same ALTGX instance). For multiple ALTGX instances connected to the dynamic reconfiguration controller, set the starting channel number of the first instance to 0. You must set the next multiple of four as the starting channel number for the remaining ALTGX instances. 1 When PMA-only channel reconfiguration involves a transmitter PLL, you also must account for the logical channel address of the PLL used. If there are four channels in Basic [PMA Direct] xN functional mode, each channel requires a logical channel address (0, 4, 8, 12), and the transmitter PLL used requires an address (16). Figure 5-3 shows how to set the starting channel number for multiple ALTGX instances controlled by a single dynamic reconfiguration controller, where both ALTGX instances have PMA-only channels. For more information about the What is the number of channels controlled by the reconfig controller? option, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. For example, if you a transceiver configuration with Instance 1 (Inst1) with two channels that use Basic (PMA Direct) mode. Instance 2 (Inst2) has two channels and uses Basic mode, which uses the PCS block in the transceiver. In order to leave the needed gap in the reconfiguration controller signals, follow these steps: 1. Set Inst1 to have a starting channel number 0. The logical channel addresses 0 and 4 are then allocated to the two channels of Inst1. 2. Reserve logical channel address 8--do not use this address. 3. Set Inst2 to have a starting channel number 12. 4. Set the Number of channels option in the ALTGX_RECONFIG controller megafunction to the nearest multiple of four from the highest logical channel address. In this example, set the Number of channels option to 16. 5. When connecting the reconfig_fromgxb bus of the ALTGX_RECONFIG controller, connect the bits corresponding to the reserved address to 0. In this example, the ALTGX_RECONFIG controller provides 4*([16:0]) bits. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-8 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration 6. Connect the reconfig_fromgxb bus of the ALTGX_RECONIF controller and the ALTGX transceiver instances as follows: a. reconfig_fromgxb[33:0] of the ALTGX_RECONFIG instance reconfig_fromgxb[33:0] of Inst1. b. reconfig_fromgxb[50:34] of the ALTGX_RECONFIG instance - 17'h00000 c. reconfig_fromgxb[67:51] of the ALTGX_RECONFIG instance reconfig_fromgxb[17:0] of Inst2 Figure 5-3. Logical Channel Addressing of PMA-Only Channels ALTGX Instance 1 Basic (PMA Direct) Configuration Starting channel number = 0 Set the What is the number of channels controlled by the reconfig controller? option to 48 Channel 0 (logical channel address = 0) reconfig_fromgxb 1[135:0] Channel 1 (logical channel address = 4) reconfig_togxb[3:0] Channel 2 (logical channel address = 8) reconfig_fromgxb[203:0] (1) Channel 3 (logical channel address = 12) ALTGX_RECONFIG Instance 1 Channel 4 (logical channel address = 16) Channel 5 (logical channel address = 20) Channel 6 (logical channel address = 24) Channel 7 (logical channel address = 28) ALTGX Instance 2 Basic (PMA Direct) Configuration Starting channel number = 32 Channel 0 (logical channel address = 32) Channel 1 (logical channel address = 36 reconfig_fromgxb 2[67:0] Channel 2 (logical channel address = 40) Channel 3 (logical channel address = 44) Note to Figure 5-3: (1) reconfig_fromgxb[203:0] = { reconfig_fromgxb 2[67:0], reconfig_fromgxb 1[135:0]}. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration 5-9 Logical Channel Addressing--Combination of Regular Transceiver Channels and PMA-Only Channels For a combination of regular transceiver channels and PMA-only channels, there must be at least two different ALTGX instances connected to the same dynamic reconfiguration controller. This is because you cannot have a combination of regular transceiver channels and PMA-only channels within the same ALTGX instance. Set the starting channel number in the first ALTGX Instance 1 to 0. If you have configured ALTGX Instance 1 with regular transceiver channels, the logical channel addresses of the remaining channels increment in steps of one. Set the starting channel number of the following ALTGX Instance 2 as the next multiple of four. If you have configured ALTGX Instance 2 with PMA-only channels, the logical channel addresses of the remaining channels increment in steps of four. Figure 5-41 in "Example 1" on page 5-96 shows how to set the starting channel number for multiple ALTGX instances controlled by a single dynamic reconfiguration controller, where one ALTGX instance has PMA-only channels and the other ALTGX instance has regular transceiver channels. Table 5-18 in "Example 1" on page 5-96 lists an example scenario where the logical channel address of both the PMA-only channels and regular transceiver channels is set based on the starting channel number. For more information, refer to "Example 1" on page 5-96. Highest Possible Logical Channel Address Table 5-4 lists the highest possible logical channel address assigned to a transceiver channel in a Stratix IV device. The maximum number of transceiver channels in the largest Stratix IV device is 48 (24 transceiver channels located in four transceiver blocks on the right side of the device and 24 transceiver channels located in four transceiver blocks on the left side of the device). You can individually configure these 48 transceiver channels as 48 Transmitter only and 48 Receiver only channels. You achieve this by using 48 Transmitter only ALTGX instances and 48 Receiver only ALTGX instances in your design. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-10 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration Table 5-4. Highest Possible Logical Channel Address 96 ALTGX Instances ALTGX MegaWizard Plug-In Manager Setting What is the number of channels? in the General screen What is the starting channel number? in the Reconfig screen ALTGX_RECONFIG Instance ALTGX instance 1 ALTGX instance 2 48 48 TX instance 1: 0 RX instance 1: 192 TX instance 2: 4 RX instance 2: 196 . . . . . . . . . . . . TX instance 48: 188 ALTGX_RECONFIG instance 1: Controls all 96 ALTGX instances. RX instance 48: 380 The highest logical channel address is assigned to the Receiver only channel in the 96th ALTGX instance; therefore, the setting is 380. 1 The highest possible logical channel address assigned to a transceiver channel in a Stratix IV device is the same whether the channel is a regular transceiver channel or a PMA-only channel. Total Number of Channels Option in the ALTGX_RECONFIG Instance You can connect every dynamic reconfiguration controller in a design to either a single ALTGX instance or to multiple ALTGX instances. Depending on the number of channels within each of these ALTGX instances, you must set the total number of channels controlled by the dynamic reconfiguration controller in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Based on this information, the reconfig_fromgxb and logical_channel_address input ports vary in width. Use the following steps to determine the number of channels: 1. Determine the highest logical channel address among all the transceiver instances connected to the same dynamic reconfiguration controller. For more information, refer to "Logical Channel Addressing" on page 5-5. 2. Round the logical channel address value to the next higher multiple of four. 3. Use this value to set the What is the number of channels controlled by the reconfig controller? option. For more information, refer to "Example 1" on page 5-96. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Quartus II MegaWizard Plug-In Manager Interfaces to Support Dynamic Reconfiguration 5-11 Connecting the ALTGX and ALTGX_RECONFIG Instances There are two ways to connect the ALTGX_RECONFIG instance to the ALTGX instance in your design: Single dynamic reconfiguration controller--You can use a single ALTGX_RECONFIG instance to control all the ALTGX instances in your design. Figure 5-2 on page 5-6 shows a block diagram of a single dynamic reconfiguration controller in a design. Multiple dynamic reconfiguration controllers--Your design can have multiple ALTGX_RECONFIG instances but you can use only one ALTGX_RECONFIG instance per transceiver block, as shown in Figure 5-4. Figure 5-4. Multiple Dynamic Reconfiguration Controllers in a Design reconfig_fromgxb [n:0] ALTGX_RECONFIG Instance 1 ALTGX Instance 1 reconfig_togxb [3:0] reconfig_fromgxb [n:0] ALTGX_RECONFIG Instance 2 ALTGX Instance 2 reconfig_togxb [3:0] In the dynamic reconfiguration interface, you must connect the reconfig_fromgxb and reconfig_togxb signals between the ALTGX_RECONFIG instance and the ALTGX instance to successfully complete the dynamic reconfiguration process. Make the following connections: September 2012 Connect the reconfig_fromgxb input port of the ALTGX_RECONFIG instance to the reconfig_fromgxb output ports of all the ALTGX instances controlled by the ALTGX_RECONFIG instance. Connect the reconfig_fromgxb port of the ALTGX instance whose starting channel number is 0, to the lowest significant bit of the reconfig_fromgxb input port of the ALTGX_RECONFIG instance. Connect the reconfig_fromgxb port of the ALTGX instance with the next highest starting channel number to the following bits of the reconfig_fromgxb of the ALTGX_RECONFIG instance, and so on. Connect the same reconfig_togxb ports of all the ALTGX instances controlled by the ALTGX_RECONFIG instance to the reconfig_togxb output port of the ALTGX_RECONFIG instance. The reconfig_togxb output port is fixed to 4 bits. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-12 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Connecting reconfig_fromgxb for the Regular Transceiver Channels Figure 5-3 on page 5-8 shows how to connect the reconfig_fromgxb output port of the ALTGX instance to the reconfig_fromgxb input port of the ALTGX_RECONFIG instance for regular transceiver channels. Table 5-18 in "Example 1" on page 5-96 describes how to connect the reconfig_fromgxb port for regular transceiver channels. Connecting reconfig_fromgxb for the PMA-Only Channels Figure 5-3 on page 5-8 shows how to connect the reconfig_fromgxb output port of the ALTGX instance to the reconfig_fromgxb input port of the ALTGX_RECONFIG instance for PMA-only channels. Table 5-18 in "Example 1" on page 5-96 describes how to connect the reconfig_fromgxb port for PMA-Only channels. Dynamic Reconfiguration Modes Implementation The modes available for dynamically reconfiguring the Stratix IV transceivers are: "PMA Controls Reconfiguration Mode Details" on page 5-12 "Transceiver Channel Reconfiguration Mode Details" on page 5-19 Channel and CMU PLL reconfiguration (.mif based) Channel reconfiguration with transmitter PLL select (.mif based) CMU PLL reconfiguration (.mif based) Central control unit reconfiguration (.mif based) Data rate division in transmitter "Offset Cancellation Feature" on page 5-67 "EyeQ" on page 5-69 "Adaptive Equalization (AEQ)" on page 5-75 "Dynamic Reconfiguration Controller Port List" on page 5-78 The following sections describe each of these modes in detail. PMA Controls Reconfiguration Mode Details You can dynamically reconfigure the following PMA controls: Pre-emphasis settings DC gain settings Voltage output differential (VOD) settings Equalization settings (channel reconfiguration mode does not support equalization settings) The following section describes how to connect the transceiver channels (the ALTGX instance) to the dynamic reconfiguration controller (the ALTGX_RECONFIG instance) to dynamically reconfigure the PMA controls. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-13 The PMA control ports for the ALTGX_RECONFIG MegaWizard Plug-In Manager are available in the Analog controls screen. You can select the PMA control ports you want to reconfigure. For example, to use tx_vodctrl to write new VOD settings or to use tx_vodctrl_out to read the existing VOD settings. Dynamically Reconfiguring PMA Controls You can dynamically reconfigure the PMA controls of a transceiver channel using three methods: Reconfiguring the PMA controls of a specific transceiver channel. For more information, refer to "Method 1--Using the logical_channel_address Port". Dynamically reconfiguring the PMA controls of the transceiver channels without using the logical_channel_address port (where all transceiver channels are reconfigured). If you use this method, the PMA controls of all the transceiver channels connected to the dynamic reconfiguration controller are reconfigured. For more information, refer to "Method 2--Using the Same Control Signals for All Channels" on page 5-15. Dynamically reconfiguring the PMA controls of the transceiver channels without using the logical_channel_address port (where only the PMA controls of the transceiver channels are reconfigured). If you use this method, each channel has its own PMA control port. Based on the value set at the ports, the PMA controls of the corresponding transceiver channels are reconfigured. For more information, refer to "Method 3--Using Individual Control Signals for Each Channel" on page 5-17. For the above three methods, you can additionally use the rx_tx_duplex_sel[1:0] port transmitter and receiver parameters. For more information, refer to "Dynamic Reconfiguration Controller Port List" on page 5-78. Method 1--Using the logical_channel_address Port Using Method 1, you can dynamically reconfigure the PMA controls of a transceiver channel by using the logical_channel_address port without affecting the remaining active channels. Enable the logical_channel_address port by selecting the Use 'logical_channel_address' port for Analog controls reconfiguration option in the Analog controls screen of the ALTGX_RECONFIG MegaWizard Plug-In Manager. 1 This method is applicable only for a design where the dynamic reconfiguration controller controls more than one channel. When using Method 1, the selected PMA control write and read ports remain fixed in width, regardless of the number of channels controlled by the ALTGX_RECONFIG instance. To observe the width of the PMA control ports, refer to the ALTGX_RECONFIG MegaWizard Plug-In Manager. The value you set at the PMA control ports is only written into the specified transceiver channel. 1 September 2012 Ensure that the busy signal is low before you start a write or read transaction. The busy output status signal is asserted high when the dynamic reconfiguration controller is occupied writing or reading the PMA control values. When the write or read transaction has completed, the busy signal goes low. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-14 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Write Transaction Figure 5-5 shows the write transaction waveform when using Method 1. In this example, the number of channels connected to the dynamic reconfiguration controller is four. Therefore, the logical_channel_address port is 2 bits wide. Also, to initiate the write transaction, you must assert the write_all signal for one reconfig_clk cycle. Figure 5-5. Method 1--Write Transaction Waveform reconfig_clk write_all rx_tx_duplex_sel [1:0] 2'b00 2'b10 (transmitter portion only) logical_address_channel [1:0] 2'b00 2'b01 (second channel of the ALTGX instance) busy tx_vodctrl [2:0] Stratix IV Device Handbook Volume 2: Transceivers 3'b00 3'b11 September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-15 Read Transaction In this example, you want to read the existing VOD values from the transmit VOD control registers of the transmitter portion of a specific channel controlled by the ALTGX_RECONFIG instance. For this example, the number of channels connected to the dynamic reconfiguration controller is four. Therefore, the logical_channel_address port is 2 bits wide. Also, to initiate the read transaction, assert the read signal for one reconfig_clk clock cycle. After the read transaction has completed, the data_valid signal is asserted. Figure 5-6 shows the read transaction waveform. Figure 5-6. Method 1--Read Transaction Waveform reconfig_clk read rx_tx_duplex_sel [1:0] 2'b00 2'b10 (transmitter portion only) logical_address_channel [1:0] 2'b00 2'b01 (second channel of the ALTGX instance) busy data_valid tx_vodctrl_out [2:0] 1 3'b000 3'bXXX 3'b001 Simultaneous write and read transactions are not allowed. Method 2--Using the Same Control Signals for All Channels To use Method 2, enable the Use the same control signal for all channels option in the Analog controls screen of the ALTGX_RECONFIG MegaWizard Plug-In Manager. Using Method 2, you can write the same PMA control value into all the transceiver channels connected to the dynamic reconfiguration controller. The PMA control write ports remain fixed in width irrespective of the number of channels controlled by the ALTGX_RECONFIG instance. The PMA control read ports increase in width based on the number of channels controlled by the ALTGX_RECONFIG instance. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-16 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Write Transaction Assume that you have enabled tx_vodctrl in the ALTGX_RECONFIG MegaWizard Plug-In Manager to reconfigure the VOD of the transceiver channels. Figure 5-7 shows the write transaction to reconfigure the VOD. Figure 5-7. Method 2--Write Transaction Waveform reconfig_clk write_all rx_tx_duplex_sel [1:0] 2'b00 2'b10 (transmitter portion only) busy tx_vodctrl [2:0] 3'b00 3'b11 Read Transaction If you want to read the existing values from a specific channel connected to the ALTGX_RECONFIG instance, observe the corresponding byte positions of the PMA control output port after the read transaction is complete. For example, if the number of channels controlled by the ALTGX_RECONFIG instance is two, tx_vodctrl_out is 6 bits wide (tx_vodctrl_out[2:0] corresponds to channel 1 and tx_vodctrl_out[5:3] corresponds to channel 2). Figure 5-8 shows how to read the VOD values of the second channel. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-17 Figure 5-8 shows the read transaction waveform. The transmit VOD settings written in channels 1 and 2 prior to the read transaction are 3'b001 and 3'b010, respectively. Figure 5-8. Method 2--Read Transaction Waveform reconfig_clk read busy data_valid rx_tx_duplex_sel [1:0] tx_vodctrl [5:0] (1) 2'b00 2'b10 (transmitter portion only) 6'b000000 6'bXXXXXX 6'b010001 Note to Figure 5-8: (1) To read the current VOD values in channel 2, observe the values in tx_vodctrl_out[5:3]. 1 Simultaneous write and read transactions are not allowed. Method 3--Using Individual Control Signals for Each Channel You can optionally use Method 3 to individually reconfigure the PMA controls of each transceiver channel. When you disable the Use the same control signal for all channels option, the PMA control ports for the write transaction are also separate for each channel. For example, if you have two channels, tx_vodctrl is 6 bits wide (tx_vodctrl[2:0] corresponds to channel 1 and tx_vodctrl[5:3] corresponds to channel 2). The width of the PMA control ports for a read transaction are always separate for each channel (the same as the PMA control ports, as explained in "Method 2--Using the Same Control Signals for All Channels" on page 5-15.) September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-18 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Write Transaction In this method, the PMA controls are written into all the channels connected to the dynamic reconfiguration controller. Therefore, to write to a specific channel: 1. Retain the stored values of the other active channels using a read transaction. 2. Set the new value at the bits corresponding to the specific channel. 3. Perform a write transaction. For example, assume that the number of channels controlled by the ALTGX_RECONFIG instance is two, tx_vodctrl in this case is 6 bits wide (tx_vodctrl[2:0] corresponds to channel 1 and tx_vodctrl[5:3] corresponds to channel 2). Follow these steps: 1. If you want to dynamically reconfigure the PMA controls of only channel 2 with a new value, first perform a read transaction to retrieve the existing PMA control values from tx_vodctrl_out[5:0]. Take tx_vodctrl_out[2:0] and provide this value in tx_vodctrl[2:0] to the write in channel 1. By doing so, channel 1 is overwritten with the same value. 2. Perform a write transaction. This ensures that the new values are written only to channel 2, while channel 1 remains unchanged. Figure 5-9 shows a write transaction waveform using Method 3. Figure 5-9. Method 3--Write Transaction Waveform reconfig_clk write_all rx_tx_duplex_sel [1:0] 2'b00 2'b10 (transmitter portion only) busy tx_vodctrl [5:0] (1) 6'b000000 6'b000011 Note to Figure 5-9: (1) For this example, the number of channels controlled by the dynamic reconfiguration controller (ALTGX_RECONFIG instance) is two and the tx_vodctrl control port is enabled. 1 Stratix IV Device Handbook Volume 2: Transceivers Simultaneous write and read transactions are not allowed. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-19 Read Transaction The read transaction in Method 3 is identical to that in Method 2. Refer to "Read Transaction" on page 5-16. Transceiver Channel Reconfiguration Mode Details Table 5-5 lists the supported configurations for the various transceiver channel reconfiguration modes available in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Table 5-5. Transceiver Channel Reconfiguration Modes and .mif Requirements Supported Configurations Dynamic Reconfiguration Mode .mif Requirements Channel and CMU PLL reconfiguration Channel reconfiguration with transmitter PLL select Central control unit reconfiguration (2) Data rate division in transmitter To From All configurations of regular transceiver channels All configurations of regular transceiver channels Y Basic (PMA Direct) x1 configuration Basic (PMA Direct) x1 configuration Y Basic (PMA Direct) xN configuration Basic (PMA Direct) xN configuration Y Non-bonded configurations of regular transceiver channels Non-bonded configurations of regular transceiver channels Y Basic (PMA Direct) x1 configuration Basic (PMA Direct) x1 configuration Y Basic (PMA Direct) xN configuration Basic (PMA Direct) xN configuration Y x4 bonded mode x4 bonded mode Y x8 bonded mode x8 bonded mode Y All Transmitter only configurations of regular transceiver channels All Transmitter only configurations of regular transceiver channels (1) -- Note to Table 5-5: (1) Because the transmitter local divider is not available for bonded mode channels, data rate division is supported for non-bonded channels only. (2) Dynamic reconfiguration from a bonded mode with rate matcher to another bonded mode without rate matcher is not allowed. 1 You cannot dynamically reconfigure from Deterministic Latency mode to any other functional mode and vice-versa. Within Deterministic Latency mode, the following reconfigurations are not allowed: Phase Compensation FIFO register mode and a non-register mode PFD feedback mode and a non-PFD feedback mode For instance, you can dynamically reconfigure the data rate for CPRI mode. However, you cannot dynamically reconfigure from CPRI mode to a non-CPRI mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-20 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Memory Initialization File (.mif) As listed in Table 5-5, all the dynamic reconfiguration modes with a check mark in the ".mif Requirement" column use memory initialization files to reconfigure the transceivers. These .mifs contain the valid settings, in the form of words, required to reconfigure the transceivers. To understand using .mifs, it is helpful to understand these two concepts: How to generate a .mif?--The Quartus(R) II software generates .mifs when you provide the appropriate project settings and then compiles an ALTGX instance. For more information, refer to "Quartus II Settings to Enable .mif Generation" on page 5-20. How is a .mif used between the ALTGX_RECONFIG instance and the ALTGX instance?--The Quartus II software provides a design flow called the user memory initialization file flow. For more information, refer to ".mif-Based Design Flow" on page 5-22. Quartus II Settings to Enable .mif Generation The .mif is not generated by default in a Quartus II compilation. To generate a .mif, you must enable the following Quartus II software settings: 1. On the Assignments menu, select Settings (Figure 5-10). Figure 5-10. Step 1 to Enable .mif Generation Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-21 2. Select Fitter settings, then choose More Settings (Figure 5-11). Figure 5-11. Step 2 to Enable .mif Generation 3. In the Option box of the More Fitter Settings page, set the Generate GXB Reconfig MIF option to On (Figure 5-12). Figure 5-12. Step 3 to Enable .mif Generation The .mif is generated in the Assembler stage of the compilation process. However, for any change in the design or the above settings, the Quartus II software runs through the fitter stage before starting the Assembler stage. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-22 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation A .mif is generated for every ALTGX instance defined in the top-level RTL file. The Quartus II software creates the .mif under the <Project_DIR>/reconfig_mif folder. The file name is based on the ALTGX instance name (<instance name>.mif); for example, basic_gxb.mif. One design can have multiple .mifs (there is no limit) and you can use one .mif to reconfigure multiple channels. To generate a .mif, create a top-level design and connect the clock inputs in the RTL/schematic. Specifically, for the transceiver clock inputs pll_inclk_cruclk. 1 If you do not specify pins for tx_dataout and rx_datain for the transceiver channel, the Quartus II software selects a channel and generates a .mif for that channel. However, the .mif can still be used for any transceiver channel. You can generate multiple .mifs in the following two ways: Method 1: 1. Compile the design created and generate the first .mif. 2. Update the ALTGX instance with the alternate configuration. 3. Compile the design to get the second .mif. 1 If you have to generate .mifs for many configurations, Method 1 takes more time to complete. Method 2: 1. In the top-level design, instantiate all the different configurations of the ALTGX instantiation for which the .mif is required. 2. Connect the appropriate clock inputs of all the ALTGX instantiations. 3. Generate the .mif. The .mifs are generated for all the ALTGX configurations. 1 This method requires special attention when generating the .mif. Refer to the following: The different ALTGX instantiations must have the appropriate logical reference clock index option values. The clock inputs for each instance must be connected to the appropriate clock source. When you generate the .mif, use the proper naming convention for the files so you know the configuration supported by the .mif. .mif-Based Design Flow The .mif-based design flow involves writing the contents of the .mif to the transceiver channel or CMU PLL. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-23 To reconfigure the transceiver channel or CMU PLL, you must configure the required settings for the transceiver channel or CMU PLL in the ALTGX MegaWizard Plug-In Manager and compile the ALTGX instance. The dynamic reconfiguration controller requires that you write these configured settings through the .mif into the transceiver channel or CMU PLL (using the write_all and reconfig_data[15:0] signals). The maximum possible size of the .mif is 59 words. Each word contains legal register settings of the transceiver channel stored in 16 bits. reconfig_address_out[5:0] provides the address (location) of the 16-bit word in the .mif. Table 5-6 lists the .mif size depending on the ALTGX configuration. Table 5-6. .mif Size for the ALTGX Configuration ALTGX Configuration .mif Size in Words (1) PMA Direct Mode Duplex (Receiver and Transmitter) + Central control unit 60 33 Duplex (Receiver and Transmitter) 55 28 Receiver only 38 14 Transmitter only 19 15 Note to Table 5-6: (1) Each word in the .mif is 16 bits wide. You can store these .mifs in on-chip or off-chip memory. Applying a .mif in the User Design Store the .mif in on-chip or off-chip memory and connect it to the dynamic reconfiguration controller, as shown in Figure 5-13. Figure 5-13. .mif Instantiation in the User Design GIGE RAM ALTGX Instances GIGE .mif GIGE channel Reconfiguration User Logic SONET OC48 RAM ALTGX_RECONFIG Instance SONET OC48 channel SONET OC48 .mif When applying a .mif in the user design, be sure to: September 2012 Use the RAM: 1-PORT megafunction to instantiate a memory block. Choose the size of the memory block based on the size of the .mif generated. Instantiate the .mif in the memory block. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-24 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 1 Whenever a .mif is applied to a channel, the PMA controls for that channel are set to the default settings chosen in the ALTGX instance used for .mif generation. 1 The equalization settings of the receiver cannot be modified by a .mif. Reduced .mif Reconfiguration This mode is available only for the .mif-based transceiver channel reconfiguration modes. This is an optional feature that allows faster reconfiguration and faster simulation time. For example, if you intend to make minor changes to the transceiver channel, this might involve a change of only a few words in the .mif. Here is an example of changing only the termination setting: Assume that the only word difference is word address 32. Instead of loading the entire .mif, you can use altgx_diffmifgen.exe to generate a new .mif. This new .mif only has the modified words. The new .mif is 22 bits wide, compared with the 16 bits wide in the regular .mif. There are 6 bits of address in addition to 16 bits of data. <addr 6 bits> <data 16 bits> Enable the Use 'reconfig_address' to input address from the MIF in reduced MIF reconfiguration option in the Channel and TX PLL Reconfiguration screen of the ALTGX_RECONFIG MegaWizard Plug-In Manager. Use the reconfig_data[15:0] port to connect the 16 bits of data from the new .mif. Use the reconfig_address[5:0] port to connect the 6 bits of address from the new .mif. Using altgx_diffmifgen.exe Browse to the project directory where you have the Quartus II software installed. For example, altgx_diffmifgen.exe is available in the following path: \altera\91\quartus\bin The syntax for using this .exe is as follows: \altera\91\quartus\bin\altgx_diffmifgen.exe <a.mif> <b.mif> That is executed in the project directory with the .mifs. The altgx_diffmifgen.exe requires two or more ALTGX .mifs. Channel and CMU PLL Reconfiguration Mode Details Use this dynamic reconfiguration mode to reconfigure a transceiver channel to a different functional mode and data rate. To reconfigure a channel successfully, select the appropriate options in the ALTGX MegaWizard Plug-In Manager (described in the following sections) and generate a .mif. Connect the ALTGX_RECONFIG instance to the ALTGX instance. The dynamic reconfiguration controller reconfigures the transceiver channel by writing the .mif contents into the channel. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-25 1 Channel and CMU PLL reconfiguration mode only affects the channel involved in the reconfiguration (the transceiver channel specified by the logical_channel_address port), without affecting the remaining transceiver channels controlled by the dynamic reconfiguration controller. 1 You cannot reconfigure the auxiliary transmit (ATX) PLLs in Stratix IV transceivers. Channel Reconfiguration Classifications Table 5-7 lists the classification for channel and CMU PLL reconfiguration mode. Table 5-7. Channel Reconfiguration Classifications Data Rate Reconfiguration By reconfiguring the CMU PLL connected to the transceiver channel. By selecting the alternate CMU PLL in the transceiver block to supply clocks to the transceiver channel. Every transmitter channel has one local clock divider. Similarly, every receiver channel has one local clock divider. You can reconfigure the data rate of a transceiver channel by reconfiguring these local clock dividers to 1, 2, or 4. When you reconfigure these local clock dividers, ensure that the functional mode of the transceiver channel supports the reconfigured data rate. Functional Mode Reconfiguration Use this feature to reconfigure the existing functional mode of the transceiver channel to a totally different functional mode. There is no limit to the number of functional modes you can reconfigure the transceiver channel to if the various clocks involved support the transition. For more information about core clocks, refer to "Clocking/Interface Options" on page 5-30. 1 In addition to the categories mentioned, you can also choose to reconfigure both the data rate and functional mode of a transceiver channel. 1 For the following sections, assume that the transceiver channel has the Receiver and Transmitter configuration in the ALTGX MegaWizard Plug-In Manager, unless specified as Transmitter only or Receiver only. Blocks Reconfigured in Channel and CMU PLL Reconfiguration Mode The blocks that are reconfigured by this dynamic reconfiguration mode are the PCS and PMA blocks of a transceiver channel, the local divider settings of the transmitter and receiver channel, and the CMU PLL. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-26 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-14 shows the functional blocks that you can dynamically reconfigure using channel and CMU PLL reconfiguration mode. Figure 5-14. Channel and CMU PLL Reconfiguration in a Transceiver Block CMU Channel Full Duplex Transceiver Channel clock mux refclk0 TX CHANNEL CMU0 PLL Logical TX PLL select refclk1 clock mux LOCAL DIVIDER TX PMA + TX PCS CMU1 PLL RX CHANNEL clock mux RX CDR RX PMA + RX PCS Blocks that can be reconfigured in Channel and CMU PLL Reconfiguration mode 1 Channel reconfiguration from either a Transmitter only configuration to a Receiver only configuration or vice versa is not allowed. ALTGX MegaWizard Plug-In Manager Setup for Channel and CMU PLL Reconfiguration Mode To reconfigure the transceiver channel and CMU PLL, set up the ALTGX MegaWizard Plug-In Manager using the following steps: 1. Select the Channel and Transmitter PLL reconfiguration option in the Modes screen under the Reconfiguration Settings tab. 2. If you want to reconfigure the data rate of the transceiver channel by reconfiguring the CMU PLL, provide the new data rate you want the CMU PLL to run at in the General screen. 3. If you want to reconfigure the data rate of the transceiver channel by switching to the alternate CMU PLL within the same transceiver block, select the Use alternate CMU transmitter PLL option in the Modes screen. For more information, refer to the "Using the Alternate CMU Transmitter PLL" on page 5-27. 4. Provide the number of input reference clocks available for the CMU PLL in the How many input clocks? option of the corresponding PLL screen. The maximum number of input reference clocks allowed is 10. For more information, refer to "Guidelines for Specifying the Input Reference Clocks" on page 5-61. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-27 5. Provide the starting channel number in the Modes screen. For more information, refer to "Logical Channel Addressing" on page 5-5. 6. Provide the logical reference index of the CMU PLL in the What is the PLL logical reference index? option in the corresponding PLL screen. For more information, refer to "Selecting the Logical Reference Index of the CMU PLL" on page 5-29. 7. Provide the identification of the input reference clock used by the CMU PLL in the corresponding PLL screens. 8. Set up the Clocking/Interface options. For more information, refer to "Clocking/Interface Options" on page 5-30. 9. Set up the Channel Interface options. For more information, refer to "FPGA Fabric-Transceiver Channel Interface Selection" on page 5-36. Using the Alternate CMU Transmitter PLL To reconfigure the CMU PLL during run time, you need the flexibility to select one of the two CMU PLLs of a transceiver block. Consider that the transceiver channel is listening to CMU0 PLL and that you want to reconfigure CMU0 PLL, as shown in Figure 5-15. Figure 5-15. Reconfiguring the CMU0 PLL Main PLL logical_tx_pll value = 1 CMU Channels refclk0 refclk1 156.25 MHz Full Duplex Transceiver Channel clock mux 6.25 Gbps CMU0 PLL 125 MHz 1 0 clock mux TX CHANNEL Logical TX PLL select 6.25 Gbps TX PMA + TX PCS LOCAL DIVIDER 2.5 Gbps CMU1 PLL RX CHANNEL clock mux Active Connections 6.25 Gbps RX CDR 6.25 Gbps RX PMA + RX PCS Alternate PLL logical_tx_pll value = 0 Unused Connections September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-28 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation You can select CMU0 PLL by specifying its identity in the ALTGX MegaWizard Plug-In Manager. This identification is referred to as the logical tx pll value. This value provides a logical identification to CMU0 PLL and associates it with a transceiver channel without requiring the knowledge of its physical location. In the ALTGX MegaWizard Plug-In Manager, the transmitter PLL configuration set in the General screen is called the main PLL. When you provide the alternate PLL with a logical tx pll value (for example, 0), the main PLL automatically takes the complement value 1. The logical tx pll value for the main PLL is stored along with the other transceiver channel information in the generated .mif. 1 Stratix IV Device Handbook Volume 2: Transceivers The main PLL corresponds to the CMU PLL configuration set in the General screen of the ALTGX MegaWizard Plug-In Manager. The alternate PLL corresponds to the CMU PLL configuration set in the Alt PLL screen. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-29 Selecting the Logical Reference Index of the CMU PLL In Figure 5-16, transceiver channel 1 listens to CMU0 PLL of the transceiver block. Similarly, transceiver channel 2 listens to CMU1 PLL of the transceiver block. Figure 5-16. Logical Reference Index of CMU PLLs in a Transceiver Block (1) CMU Channels Full Duplex Transceiver Channel 1 clock mux 156.25 MHz refclk0 refclk1 TX CHANNEL 1 6.25 Gbps CMU0 PLL 125 MHz Logical TX PLL select clock mux LOCAL DIVIDER 6.25 Gbps TX PMA + TX PCS 2.5 Gbps CMU1 PLL RX CHANNEL 1 clock mux 6.25 Gbps RX CDR 6.25 Gbps RX PMA + RX PCS Full Duplex Transceiver Channel 2 TX CHANNEL 2 2.5 Gbps Logical TX PLL select LOCAL DIVIDER TX PMA + TX PCS RX CHANNEL 2 clock mux 2.5 Gbps RX CDR 2.5 Gbps RX PMA + RX PCS Note to Figure 5-16: (1) After the device powers up, the busy signal remains low for the first reconfig_clk cycle. To direct the ALTGX_RECONFIG instance to dynamically reconfigure CMU0 PLL, specify its logical reference index (the identity of a transmitter PLL). Similarly, to direct the ALTGX_RECONFIG instance to dynamically reconfigure CMU1 PLL instead, provide the logical reference index of CMU1 PLL. The allowed values for the logical reference index of the CMU PLLs within a transceiver block are 0 or 1. Similarly, the transmitter PLLs outside the transceiver block can also be assigned a logical reference index value. For more information, refer to "Selecting the PLL Logical Reference Index for Additional PLLs" on page 5-53. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-30 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 1 The logical reference index of the CMU0 PLL within a transceiver block is always the complement of the logical reference index of the CMU1 PLL within the same transceiver block. 1 This logical reference index value is stored as logical tx pll, along with the other transceiver channel settings in the .mif. Clocking/Interface Options The following describes the Clocking/Interface options. The core clocking setup describes the transceiver core clocks that are the write and read clocks of the transmit (TX) phase compensation FIFO and the receive (RX) phase compensation FIFO, respectively. Core clocking is classified as transmitter core clocking and receiver core clocking. Transmitter core clocking refers to the clock that is used to write the parallel data from the FPGA fabric into the Transmit Phase Compensation FIFO. You can use one of the following clocks to write into the Transmit Phase Compensation FIFO: 1 tx_coreclk--You can use a clock of the same frequency as tx_clkout from the FPGA fabric to provide the write clock to the Transmit Phase Compensation FIFO. If you use tx_coreclk, it overrides the tx_clkout options in the ALTGX MegaWizard Plug-In Manager. tx_clkout--The Quartus II software automatically routes tx_clkout to the FPGA fabric and back into the TX phase compensation FIFO. The Clocking/Interface screen is not available for PMA-only channels. Option 1: Share a Single Transmitter Core Clock Between Transmitters Enable this option if you want tx_clkout of the first channel (channel 0) of the transceiver block to provide the write clock to the TX phase compensation FIFOs of the remaining channels in the transceiver block. This option is typically enabled when all the channels of a transceiver block are of the same functional mode and data rate, and are reconfigured to the identical functional mode and data rate. Consider the following scenario: Four regular transceiver channels configured at 3 Gbps and in the same functional mode. Channel and CMU PLL reconfiguration mode is enabled in the ALTGX_RECONFIG MegaWizard Plug-In Manager. You want to reconfigure all four regular transceiver channels to 1.5 Gbps and vice versa. Option 1 is applicable in this scenario because it saves clock resources. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-31 Figure 5-17 shows the sharing of channel 0's tx_clkout between all four regular channels of a transceiver block. Figure 5-17. Option 1 for Transmitter Core Clocking (Channel and CMU PLL Reconfiguration Mode) FPGA Fabric Transceiver Block TX0 (3 Gbps/1.5 Gbps) RX0 TX1 (3 Gbps/1.5 Gbps) RX1 tx_clkout[0] TX2 (3 Gbps/1.5 Gbps) CMU1 PLL CMU0 PLL RX2 TX3 (3 Gbps/1.5 Gbps) RX3 Low-speed parallel clock generated by the TX0 local divider (tx_clkout[0]) High-speed serial clock generated by the CMU0 PLL Option 2: Use the Respective Channel Transmitter Core Clocks Enable this option if you want the individual transmitter channel tx_clkout signals to provide the write clock to their respective Transmit Phase Compensation FIFOs. This option is typically enabled when each transceiver channel is reconfigured to a different functional mode using channel reconfiguration. Consider the following scenario: Four regular transceiver channels configured at 3 Gbps and different functional modes. Channel and CMU PLL reconfiguration mode is enabled in the ALTGX_RECONFIG MegaWizard Plug-In Manager. You want to reconfigure each of the four regular transceiver channels to different data rates and different functional modes. Option 2 is applicable in this scenario because the design requires all four regular transceiver channels to be reconfigured to different data rates and functional modes. Each channel can be reconfigured to a different functional mode using the channel and CMU PLL reconfiguration mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-32 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-18 shows how each transmitter channel's tx_clkout signal provides a clock to the Transmit Phase Compensation FIFOs of the respective transceiver channels. Figure 5-18. Option 2 for Transmitter Core Clocking (Channel and CMU PLL Reconfiguration Mode) FPGA Fabric Transciever Block TX0 (3 Gbps/1.5 Gbps) tx_clkout[0] RX0 TX1 (3 Gbps/6 Gbps) tx_clkout[1] RX1 CMU1 PLL TX2 (3 Gbps/1.5 Gbps) tx_clkout[2] CMU0 PLL RX2 TX3 (3 Gbps) tx_clkout[3] RX3 High-speed serial clock generated by the CMU0 PLL Low-speed parallel clock generated by the local divider of the transceiver Receiver core clocking refers to the clock that is used to read the parallel data from the Receiver Phase Compensation FIFO into the FPGA fabric. You can use one of the following clocks to read from the Receive Phase Compensation FIFO: 1 Stratix IV Device Handbook Volume 2: Transceivers rx_coreclk--You can use a clock of the same frequency as rx_clkout from the FPGA fabric to provide the read clock to the Receive Phase Compensation FIFO. If you use rx_coreclk, it overrides the rx_clkout options in the ALTGX MegaWizard Plug-In Manager. rx_clkout--The Quartus II software automatically routes rx_clkout to the FPGA fabric and back into the Receive Phase Compensation FIFO. The Clocking/Interface screen is not available for PMA-only channels. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-33 Option 1: Share a Single Transmitter Core Clock Between Receivers Enable this option if you want tx_clkout of the first channel (channel 0) of the transceiver block to provide the read clock to the Receive Phase Compensation FIFOs of the remaining receiver channels in the transceiver block. This option is typically enabled when all the channels of a transceiver block are in a Basic or Protocol configuration with rate matching enabled and are reconfigured to another Basic or Protocol configuration with rate matching enabled. Consider the following scenario: Four regular transceiver channels configured to the Basic 2 Gbps functional mode with rate matching enabled. Channel and CMU PLL reconfiguration mode is enabled in the ALTGX_RECONFIG MegaWizard Plug-In Manager. You want to reconfigure all four regular transceiver channels to 3.125 Gbps configuration with rate matching enabled. Option 1 is applicable in this scenario. Figure 5-19 shows the sharing of channel 0's tx_clkout between all four channels of a transceiver block. Figure 5-19. Option 1 for Receiver Core Clocking (Channel and CMU PLL Reconfiguration Mode) FPGA Fabric Transceiver Block TX0 (2 Gbps) RX0 tx_clkout[0] TX1 (2 Gbps) RX1 CMU1 PLL TX2 (2 Gbps) CMU0 PLL RX2 TX3 (2 Gbps) Four regular transceiver channels configured at Basic 2G with Rate Matching and set up to switch to 3.125 Gbps with Rate Matching RX3 Low-speed parallel clock generated by the TX0 local divider (tx_clkout[0]) High-speed serial clock generated by the CMU0 PLL High-speed serial clock generated by the CMU1 PLL September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-34 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Option 2: Use the Respective Channel Transmitter Core Clocks Enable this option if you want the individual transmitter channel's tx_clkout signal to provide the read clock to its respective Receive Phase Compensation FIFO. This option is typically enabled when all the transceiver channels have rate matching enabled with different data rates and are reconfigured to another Basic or Protocol functional mode with rate matching enabled. Consider the following scenario: TX0/RX0: You want to dynamically reconfigure the Basic 1 Gbps configuration with rate matching enabled to the Basic 2 Gbps configuration with rate matching enabled. TX1/RX1: You want to dynamically reconfigure the Basic 4 Gbps configuration with rate matching enabled to the Basic 1 Gbps configuration with rate matching enabled. TX2/RX2 and TX3/RX3: You want to dynamically reconfigure the Basic 3.125 Gbps configuration with rate matching enabled to the 1 Gbps configuration with rate matching and vice versa. Channel and CMU PLL reconfiguration mode is enabled in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Option 2 is applicable because the design requires the individual transceiver channels to be reconfigured with different data rates to another Basic or Protocol functional mode with rate matching. Therefore, each channel can be reconfigured to another Basic or Protocol functional mode with rate matching enabled and a different data rate. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-35 Figure 5-20 shows the respective tx_clkout of each channel clocking the respective channels of a transceiver block. Figure 5-20. Option 2 for Receiver Core Clocking (Channel and CMU PLL Reconfiguration Mode) FPGA Fabric Transceiver Block TX0 (2 Gbps/1 Gbps) tx_clkout[0] RX0 TX1 (4 Gbps/1 Gbps) tx_clkout[1] RX1 CMU1 PLL TX2 (3.125 Gbps/1 Gbps) CMU0 PLL RX2 tx_clkout[2] TX3 (2 Gbps) RX3 High-speed serial clock generated by the CMU0 PLL High-speed serial clock generated by the CMU1 PLL Low-speed parallel clock generated by the local divider of the transceiver Option 3: Use the Respective Channel Receiver Core Clocks Enable this option if you want the individual channel's rx_clkout signal to provide the read clock to its respective Receive Phase Compensation FIFO. This option is typically enabled when the channel is reconfigured from a Basic or Protocol configuration with or without rate matching to another Basic or Protocol configuration with or without rate matching. Consider the following scenario: TX1/RX1: GIGE configuration to SONET/SDH OC48 configuration. TX2/RX2: Basic 2.5 Gbps configuration with rate matching disabled to Basic 1.244 Gbps configuration with rate matching disabled. Channel and CMU PLL reconfiguration mode is enabled in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Option 3 is applicable in this scenario. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-36 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-21 shows the respective rx_clkout of each channel clocking the respective receiver channels of a transceiver block. Figure 5-21. Option 3 for Receiver Core Clocking (Channel and CMU PLL Reconfiguration Mode) FPGA Fabric Transceiver Block TX0 (2 Gbps) RX0 rx_clkout[0] TX1 (2 Gbps) RX1 rx_clkout[1] High-speed serial clock generated by the CMU0 PLL CMU1 PLL CMU0 PLL High-speed serial clock generated by the CMU1 PLL Low-speed parallel clock generated by the local divider of the transceiver FPGA Fabric-Transceiver Channel Interface Selection This section describes the ALTGX MegaWizard Plug-In Manager settings related to the FPGA fabric-transceiver channel interface data width when you select and activate channel and CMU PLL reconfiguration mode. You must set up the FPGA fabric-transceiver channel interface data width when functional mode reconfiguration involves: changes in the FPGA fabric-transceiver channel data width OR enables and disables the static PCS blocks of the transceiver channel You can set up the FPGA fabric-transceiver channel interface data width by enabling the Channel Interface option in the Modes screen. Enable the Channel Interface option if the reconfiguration channel has: changed the FPGA fabric-transceiver channel interface data width OR Stratix IV Device Handbook Volume 2: Transceivers changed the input control signals and output status signals September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-37 There are two signals available when you enable the Channel Interface option: 1 tx_datainfull--The width of this input signal depends on the number of channels you set up in the General screen. It is 44 bits wide per channel. This signal is available only for Transmitter only and Receiver and Transmitter configurations. This port replaces the existing tx_datain port. rx_dataoutfull--The width of this output signal depends on the number of channels you set up in the General screen. It is 64 bits wide per channel. This signal is available only for Receiver only and Receiver and Transmitter configurations. This port replaces the existing rx_dataout port. In addition to these two ports, you can select the necessary control and status signals for the reconfigured channel in the Clocking/Interface screen. f For more information about control and status signals, refer to the "Transceiver Port Lists" section in the Transceiver Architecture in Stratix IV Devices chapter. These control and status signals are not applicable in Basic (PMA Direct) functional mode. Table 5-8 lists the signals not available when you enable the Channel Interface option. Table 5-8. Control and Status Signals Not Applicable in Basic (PMA Direct) Mode with the Channel Interface Option Enabled FPGA Fabric-Receiver Interface FPGA Fabric-Transmitter Interface rx_dataout tx_datain rx_syncstatus tx_ctrlenable rx_patterndetect tx_forcedisp rx_a1a2sizeout tx_dispval rx_ctrldetect rx_errdetect rx_disperr The Quartus II software has legal checks for the connectivity of tx_datainfull and rx_dataoutfull and the various control and status signals you enable in the Clocking/Interface screen. For example, the Quartus II software allows you to select and connect the pipestatus and powerdn signals. It assumes that you are planning to switch to and from PCIe functional mode. Table 5-9 lists the tx_datainfull[43:0] FPGA fabric-transceiver channel interface signals. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-38 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-9. tx_datainfull[43:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 1 of 3) FPGA Fabric-Transceiver Channel Interface Description (1) Transmit Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) tx_datainfull[7:0]: 8-bit data (tx_datain) The following signals are used only in 8B/10B modes: 8-bit FPGA fabric-transceiver Channel Interface tx_datainfull[8]: Control bit (tx_ctrlenable) tx_datainfull[9] Transmitter force disparity Compliance (PCIe) (tx_forcedisp) in all modes except PCIe. For PCIe mode, (tx_forcedispcompliance) is used. tx_datainfull[10]: Forced disparity value (tx_dispval) For Non-PIPE: 10-bit FPGA fabric-transceiver Channel Interface tx_datainfull[10]: Forced disparity value (tx_dispval) For PCIe: tx_datainfull[10]: Forced electrical idle (tx_forceelecidle) Two 8-bit Data (tx_datain) tx_datainfull[7:0] - tx_datain (LSByte) and tx_datainfull[18:11] - tx_datain (MSByte) The following signals are used only in 8B/10B modes: 16-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 16/20 bits tx_datainfull[8] - tx_ctrlenable (LSB) and tx_datainfull[19] - tx_ctrlenable (MSB) Force Disparity Enable tx_datainfull[9] - tx_forcedisp (LSB) and tx_datainfull[20] - tx_forcedisp (MSB) Force Disparity Value tx_datainfull[10] - tx_dispval (LSB) and tx_datainfull[21] - tx_dispval (MSB) Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-39 Table 5-9. tx_datainfull[43:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 2 of 3) FPGA Fabric-Transceiver Channel Interface Description (1) Transmit Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Two 8-bit Data (tx_datain) tx_datainfull[7:0] - tx_datain (LSByte) and tx_datainfull[29:22] - tx_datain (MSByte) The following signals are used only in 8B/10B modes: Two Control Bits (tx_ctrlenable) tx_datainfull[8] - tx_ctrlenable (LSB) and tx_datainfull[30] - tx_ctrlenable (MSB) Force Disparity Enable For non-PIPE: 16-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 8/10 bits tx_datainfull[9] - tx_forcedisp (LSB) and tx_datainfull[31] - tx_forcedisp (MSB) For PCIe: tx_datainfull[9] - tx_forcedispcompliance and tx_datainfull[31] - 0 Force Disparity Value For non-PIPE: tx_datainfull[10]: tx_dispval (LSB) and tx_datainfull[32] -tx_dispval (MSB) For PCIe: tx_datainfull[10] - tx_forceelecidle and tx_datainfull[32] - tx_forceelecidle 20-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 20 bits 20-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 10 bits September 2012 Altera Corporation Two 10-bit Data (tx_datain) tx_datainfull[9:0] - tx_datain (LSByte) and tx_datainfull[20:11] - tx_datain (MSByte) Two 10-bit Data (tx_datain) tx_datainfull[9:0] - tx_datain (LSByte) and tx_datainfull[31:22] - tx_datain (MSByte) Stratix IV Device Handbook Volume 2: Transceivers 5-40 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-9. tx_datainfull[43:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 3 of 3) FPGA Fabric-Transceiver Channel Interface Description (1) Transmit Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Four 8-bit Data (tx_datain) tx_datainfull[7:0]- tx_datain (LSByte) and tx_datainfull[18:11] tx_datainfull[29:22] tx_datainfull[40:33] - tx_datain (MSByte) The following signals are used only in 8B/10B modes: Four Control Bits (tx_ctrlenable) tx_datainfull[8] - tx_ctrlenable (LSB) and tx_datainfull[19] 32-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 16/20 bits tx_datainfull[30] tx_datainfull[41]- tx_ctrlenable (MSB) Force Disparity Enable (tx_forcedisp) tx_datainfull[9]- tx_forcedisp (LSB) and tx_datainfull[20] tx_datainfull[31] tx_datainfull[42]- tx_forcedisp (MSB) Force Disparity Value (tx_dispval) tx_datainfull[10]- tx_dispval (LSB) and tx_datainfull[21] tx_datainfull[32] tx_datainfull[43]- tx_dispval (MSB) Four 10-bit Data (tx_datain) 40-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 20 bits tx_datainfull[9:0] - tx_datain (LSByte) and tx_datainfull[20:11] tx_datainfull[31:22] tx_datainfull[42:33]- tx_datain (MSByte) Note to Table 5-9: (1) For all transceiver-related ports, refer to the "Transceiver Port Lists" section in the Transceiver Architecture for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-41 Table 5-10 lists the tx_dataoutfull[63:0] FPGA fabric-transceiver channel interface signals. Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 1 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) The following signals are used in 8-bit 8B/10B modes: rx_dataoutfull[7:0]: 8-bit decoded data (rx_dataout) rx_dataoutfull[8]: Control bit (rx_ctrldetect) rx_dataoutfull[9]: Code violation status signal (rx_errdetect) rx_dataoutfull[10]: rx_syncstatus rx_dataoutfull[11]: Disparity error status signal (rx_disperr) rx_dataoutfull[12]: Pattern detect status signal (rx_patterndetect) rx_dataoutfull[13]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes. 8-bit FPGA fabric-transceiver Channel Interface rx_dataoutfull[14]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes. rx_dataoutfull[14:13]: non-PCIe/PCIe mode (rx_pipestatus) rx_dataoutfull[15]: 8B/10B running disparity indicator (rx_runningdisp) The following signals are used in 8-bit SONET/SDH mode: rx_dataoutfull[7:0]: 8-bit un-encoded data (rx_dataout) rx_dataoutfull[8]: rx_a1a2sizeout rx_dataoutfull[10]: rx_syncstatus rx_dataoutfull[11]: Reserved rx_dataoutfull[12]: rx_patterndetect rx_dataoutfull[9:0]: 10-bit un-encoded data (rx_dataout) rx_dataoutfull[10]: rx_syncstatus rx_dataoutfull[11]: 8B/10B disparity error indicator (rx_disperr) 10-bit FPGA fabric-transceiver Channel Interface rx_dataoutfull[12]: rx_patterndetect rx_dataoutfull[13]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes rx_dataoutfull[15]: 8B/10B running disparity indicator (rx_runningdisp) September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-42 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 2 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Two 8-bit unencoded Data (rx_dataout) rx_dataoutfull[7:0] - rx_dataout (LSByte) and rx_dataoutfull[23:16]- rx_dataout (MSByte) The following signals are used in 16-bit 8B/10B modes: Two Control Bits rx_dataoutfull[8] - rx_ctrldetect (LSB) and rx_dataoutfull[24]- rx_ctrldetect (MSB) Two Receiver Error Detect Bits rx_dataoutfull[9] - rx_errdetect (LSB) and rx_dataoutfull[25]- rx_errdetect (MSB) Two Receiver Sync Status Bits rx_dataoutfull [10] - rx_syncstatus (LSB) and 16-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 16/20 bits rx_dataoutfull[26] - rx_syncstatus (MSB) Two Receiver Disparity Error Bits rx_dataoutfull [11] - rx_disperr (LSB) and rx_dataoutfull[27] - rx_disperr (MSB) Two Receiver Pattern Detect Bits rx_dataoutfull[12] - rx_patterndetect (LSB) and rx_dataoutfull[28]- rx_patterndetect (MSB) rx_dataoutfull[13] and rx_dataoutfull[45]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14] and rx_dataoutfull[46]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes Two 2-bit PCIe Status Bits rx_dataoutfull[14:13] - rx_pipestatus (LSB) and rx_dataoutfull[30:29] rx_pipestatus (MSB) rx_dataoutfull[15] and rx_dataoutfull[47]: 8B/10B running disparity indicator (rx_runningdisp) Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-43 Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 3 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Two 8-bit Data rx_dataoutfull[7:0] - rx_dataout (LSByte) and rx_dataoutfull[39:32] rx_dataout (MSByte) The following signals are used in 16-bit 8B/10B mode: Two Control Bits rx_dataoutfull[8] - rx_ctrldetect (LSB) and rx_dataoutfull[40] rx_ctrldetect (MSB) Two Receiver Error Detect Bits rx_dataoutfull[9] - rx_errdetect (LSB) and rx_dataoutfull[41]rx_errdetect (MSB) Two Receiver Sync Status Bits 16-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 8/10 bits rx_dataoutfull[10] - rx_syncstatus (LSB) and rx_dataoutfull[42]rx_syncstatus (MSB) Two Receiver Disparity Error Bits rx_dataoutfull[11] - rx_disperr (LSB) and rx_dataoutfull[43] rx_disperr (MSB) Two Receiver Pattern Detect Bits rx_dataoutfull[12] - rx_patterndetect (LSB) and rx_dataoutfull[44] rx_patterndetect (MSB) rx_dataoutfull[13] and rx_dataoutfull[45]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14] and rx_dataoutfull[46]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes Two 2-bit PCIe Status Bits rx_dataoutfull[14:13] - rx_pipestatus (LSB) and rx_dataoutfull[46:45]rx_pipestatus (MSB) rx_dataoutfull[15] and rx_dataoutfull[47]: 8B/10B running disparity indicator (rx_runningdisp) The following signals are used in 16-bit SONET/SDH mode: Two 8-bit Data rx_dataoutfull[7:0] - rx_dataout (LSByte) and rx_dataoutfull[39:32] rx_dataout (MSByte) Two Receiver Alignment Pattern Length Bits 16-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 8/10 bits (continued) rx_dataoutfull[8] - rx_a1a2sizeout (LSB) and rx_dataoutfull[40]rx_a1a2sizeout (MSB) Two Receiver Sync Status Bits rx_dataoutfull[10] - rx_syncstatus (LSB) and rx_dataoutfull[42] rx_syncstatus (MSB) Two Receiver Pattern Detect Bits rx_dataoutfull[12] - rx_patterndetect (LSB) and rx_dataoutfull[44] rx_patterndetect (MSB) September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-44 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 4 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Two 10-bit Data (rx_dataout) rx_dataoutfull[9:0] - rx_dataout (LSByte) and rx_dataoutfull[25:16] rx_dataout (MSByte) wo Receiver Sync Status Bits rx_dataoutfull[10] - rx_syncstatus (LSB) and rx_dataoutfull[26] rx_syncstatus (MSB) 20-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 20 bits rx_dataoutfull[11] and rx_dataoutfull[27]: 8B/10B disparity error indicator (rx_disperr) Two Receiver Pattern Detect Bits rx_dataoutfull[12] - rx_patterndetect (LSB) and rx_dataoutfull[28] rx_patterndetect (MSB) rx_dataoutfull[13] and rx_dataoutfull[29]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14] and rx_dataoutfull[30]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes rx_dataoutfull[15] and rx_dataoutfull[31]: 8B/10B running disparity indicator (rx_runningdisp) Two 10-bit Data rx_dataoutfull[9:0] - rx_dataout (LSByte) and rx_dataoutfull[41:32] rx_dataout (MSByte) Two Receiver Sync Status Bits rx_dataoutfull[10] - rx_syncstatus (LSB) and rx_dataoutfull[42] rx_syncstatus (MSB) 20-bit FPGA fabric-transceiver Channel Interface with PCS-PMA set to 10 bits rx_dataoutfull[11] and rx_dataoutfull[43]: 8B/10B disparity error indicator (rx_disperr) Two Receiver Pattern Detect Bits rx_dataoutfull[12] - rx_patterndetect (LSB) and rx_dataoutfull[44] rx_patterndetect (MSB) rx_dataoutfull[13] and rx_dataoutfull[45]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14] and rx_dataoutfull[46]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes rx_dataoutfull[15] and rx_dataoutfull[47]: 8B/10B running disparity indicator (rx_runningdisp) Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-45 Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 5 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) Four 8-bit un-encoded Data (rx_dataout) rx_dataoutfull[7:0]- rx_dataout (LSByte) rx_dataoutfull[23:16] rx_dataoutfull[39:32] rx_dataoutfull[55:48] - rx_dataout (MSByte) The following signals are used in 32-bit 8B/10B mode: Four Control Data Bits (rx_dataout) rx_dataoutfull[8] - rx_ctrldetect (LSB) rx_dataoutfull[24] rx_dataoutfull[40] rx_dataoutfull[56] - rx_ctrldetect (MSB) Four Receiver Error Detect Bits rx_dataoutfull[9]- rx_errdetect (LSB) rx_dataoutfull[25] rx_dataoutfull[41] rx_dataoutfull[57] - rx_errdetect (MSB) Four Receiver Pattern Detect Bits rx_dataoutfull[10]- rx_syncstatus (LSB) and 32-bit mode rx_dataoutfull[26] rx_dataoutfull[42] rx_dataoutfull[58] - rx_syncstatus (MSB) Four Receiver Disparity Error Bits rx_dataoutfull[11]- rx_disperr (LSB) rx_dataoutfull[27] rx_dataoutfull[43] rx_dataoutfull[59] - rx_disperr (MSB) Four Receiver Pattern Detect Bits rx_dataoutfull[12]- rx_patterndetect (LSB) rx_dataoutfull[28] rx_dataoutfull[44] rx_dataoutfull[60] - rx_patterndetect (MSB) rx_dataoutfull[13], rx_dataoutfull[29], rx_dataoutfull[45] and rx_dataoutfull[61]: Rate Match FIFO deletion status indicator (rx_rmfifodatadeleted) in non-PCIe/PCIe modes rx_dataoutfull[14], rx_dataoutfull[30], rx_dataoutfull[46], and rx_dataoutfull[62]: Rate Match FIFO insertion status indicator (rx_rmfifodatainserted) in non-PCIe/PCIe modes September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-46 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-10. rx_dataoutfull[63:0] FPGA Fabric-Transceiver Channel Interface Signal Descriptions (Part 6 of 6) FPGA Fabric-Transceiver Channel Interface Description Receive Signal Description (Based on Stratix IV GX Supported FPGA Fabric-Transceiver Channel Interface Widths) rx_dataoutfull[15], rx_dataoutfull[31], rx_dataoutfull[47], and rx_dataoutfull[63]: 8B/10B running disparity indicator (rx_runningdisp) The following signals are used in 32-bit SONET/SDH scrambled backplane mode: Four Control Data Bits (rx_dataout) rx_dataoutfull[7:0]- rx_dataout (LSByte) rx_dataoutfull[23:16] rx_dataoutfull[39:32] rx_dataoutfull[55:48] - rx_dataout (MSByte) rx_dataoutfull[8], rx_dataoutfull[24], rx_dataoutfull[40], and rx_dataoutfull[56]: four rx_a1a2sizeout 32-bit mode (continued) Four Receiver Sync Status Bits rx_dataoutfull[10]- rx_syncstatus (LSB) rx_dataoutfull[26] rx_dataoutfull[42] rx_dataoutfull[58] - rx_syncstatus (MSB) Four Receiver Pattern Detect Bits rx_dataoutfull[12]- rx_patterndetect (LSB) rx_dataoutfull[28] rx_dataoutfull[44] rx_dataoutfull[60] - rx_patterndetect (MSB) Four 10-bit Control Data Bits (rx_dataout) rx_dataoutfull[9:0]- rx_dataout (LSByte) rx_dataoutfull[25:16] rx_dataoutfull[41:32] rx_dataoutfull[57:48] - rx_dataout (MSByte) Four Receiver Sync Status Bits rx_dataoutfull[10]- rx_syncstatus (LSB) 40-bit mode rx_dataoutfull[26] rx_dataoutfull[42] rx_dataoutfull[58] - rx_syncstatus (MSB) Four Receiver Pattern Detect Bits rx_dataoutfull[12]- rx_patterndetect (LSB) rx_dataoutfull[28] rx_dataoutfull[44] rx_dataoutfull[60] - rx_patterndetect (MSB) Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-47 ALTGX_RECONFIG MegaWizard Plug-In Manager Setup for Channel and CMU PLL Reconfiguration Mode To setup channel and CMU PLL reconfiguration mode in the ALTGX_RECONFIG MegaWizard Plug-In Manager, follow these steps: 1. In the Reconfiguration settings screen, set the What is the number of channels controlled by the reconfig controller? option. For more information, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. 2. In the Reconfiguration settings screen, select the Channel and TX PLL select/reconfig option. The following control signals are always available when you enable the Channel and TX PLL select/reconfig option: channel_reconfig_done reconfig_address_out[5:0] The following ports are optional and available for selection in the Channel and TX PLL Reconfiguration screen: reset_reconfig_address reconfig_address_en logical_tx_pll_sel and logical_tx_pll_sel_en--For more information about these two ports, refer to "Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports" on page 5-59. rx_tx_duplex_sel[1:0] Channel and CMU PLL Reconfiguration Operation In channel reconfiguration, only a write transaction can occur; no read transactions are allowed. In the example shown in Figure 5-22, the ALTGX_RECONFIG controls two channels. Therefore, the logical_channel_address signal is 2 bits wide. Also, the transceiver channel is configured in Basic mode with the Receiver and Transmitter configuration. You can optionally choose to trigger write_all once by selecting the continuous write operation in the ALTGX_RECONFIG MegaWizard Plug-In Manager. The Quartus II software then continuously writes all the words required for reconfiguration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-48 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-22 shows a .mif write transaction when using channel and CMU PLL reconfiguration mode. Figure 5-22. .mif Write Transaction in Channel and CMU PLL Reconfiguration Mode reconfig_mode_sel[2:0] 3'b101 logical_channel_address[1:0] 2'b01 rx_tx_duplex_sel[1:0] 2'b00 reconfig_clk write_all busy reconfig_address_out[5:0] Addr0 Addr1 1st 16 bits Don't care 2nd 16 bits Addr54 Addr0 reconfig_address_en reconfig_data[15:0] 55th 16 bits Don't care channel_reconfig_done Notes to Figure 5-22: (1) The logical_channel_address port is set to 2'b01 to reconfigure the second transceiver channel. (2) The rx_tx_duplex_sel[1:0] port is set to 2'b00 to match the Receiver and Transmitter configuration of the specified transceiver channel. For guidelines regarding re-using .mifs, specifying input reference clocks, or using logical_tx_pll_sel ports, refer to "Special Guidelines" on page 5-57. f For more information about reset, refer to the "Reset Sequence when Using Dynamic Reconfiguration with the Channel and TX PLL select/reconfig Option" section in the Reset Control and Power Down in Stratix IV Devices chapter. Channel Reconfiguration with Transmitter PLL Select Mode Details You can reconfigure the data rate of a transceiver channel by switching between a maximum of four transmitter PLLs. You can select between the following transmitter PLLs: Stratix IV Device Handbook Volume 2: Transceivers CMU PLLs present in a transceiver block CMU PLLs present in other transceiver blocks ATX PLLs outside the transceiver block September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-49 You can use the channel reconfiguration with transmitter PLL select mode along with the CMU PLL reconfiguration mode only if it is a CMU PLL and not an ATX PLL. You can first reconfigure the second CMU PLL to the desired data rate using CMU PLL reconfiguration mode. Then use channel reconfiguration with transmitter PLL select mode to reconfigure the transceiver channel to listen to the second CMU PLL. For more information about supported configurations, refer to "Transceiver Channel Reconfiguration Mode Details" on page 5-19 and "Memory Initialization File (.mif)" on page 5-20. 1 Channel reconfiguration with transmitter PLL select mode is not applicable to regular transceiver channels in x4 and x8 bonded mode configurations. For guidelines regarding re-using .mifs, specifying input reference clocks, or using the logical_tx_pll_sel ports, refer to "Special Guidelines" on page 5-57. f For more information about reset, refer to the "Reset Sequence when Using Dynamic Reconfiguration with the Channel and TX PLL select/reconfig Option" section in the Reset Control and Power Down in Stratix IV Devices chapter. Blocks Reconfigured in the Channel Reconfiguration with Transmitter PLL Select Mode The blocks reconfigured in this mode have two types of multiplexers. When you switch between the CMU PLLs within the same transceiver block, the multiplexer that is reconfigured is within the transceiver block. It is located in the transmitter channel path. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-50 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-23 shows the multiplexers that you can dynamically reconfigure using channel reconfiguration with transmitter PLL select mode. Figure 5-23. Channel Reconfiguration with Transmitter PLL Select in a Transceiver Block CMU Channels Full Duplex Transceiver Channel clock mux refclk0 TX CHANNEL CMU0 PLL refclk1 Logical TX PLL select clock mux LOCAL DIVIDER TX PMA + TX PCS CMU1 PLL RX CHANNEL clock mux RX CDR RX PMA + RX PCS (1) Note to Figure 5-23: (1) Depending on the mode you select, PCS may or may not be present. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-51 Figure 5-24 shows the multiplexers that are reconfigured when you switch to an additional PLL that is outside the transceiver block. Figure 5-24. Multiplexers that are Reconfigured When you Switch to an Additional PLL Transceiver Block GXBR2 x1 CMU1 GXBR2 x4_GXBR2 xN_Bottom Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR2 ATX R1 PLL Block Transceiver Block GXBR1 x1 CMU1 GXBR1 x4_GXBR1 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 x1 CMU0 GXBR1 ATX R0 PLL Block Transceiver Block GXBR0 Channel 3 x1 CMU1 GXBR0 x4_GXBR0 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 September 2012 Altera Corporation x1 CMU0 GXBR0 xN_Top Stratix IV Device Handbook Volume 2: Transceivers 5-52 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation ALTGX MegaWizard Plug-In Manager Setup for Channel Reconfiguration with Transmitter PLL Select Mode Follow steps 1, 2, 4, 7, 8, and 9 described in "ALTGX MegaWizard Plug-In Manager Setup for Channel and CMU PLL Reconfiguration Mode" on page 5-26. In addition to these steps, you must also set up the following: Multi-PLL Settings The Use additional CMU/ATX Transmitter PLLs from outside the transceiver block option allows you select a maximum of four transmitter PLLs. Specify the number of additional PLLs required for the ALTGX instance in the Modes screen. Based on this number, the Quartus II software opens up the corresponding PLL screens (for example, PLL 1 and PLL 2). The PLL set up in the General screen is always the Main PLL and the settings are available in the Main PLL screen. Similarly, the PLL settings for the additional PLLs are available in the corresponding PLL1 screen, PLL 2 screen, and so on. Additional PLLs also include the CMU PLLs within the same transceiver block. For example, you can select the ATX PLL as the main PLL, and three additional PLLs as follows: PLL 1--CMU0 PLL of the same transceiver block PLL 2--CMU1 PLL of the same transceiver block PLL 3--CMU0 PLL/CMU1 PLL of another transceiver block. The Quartus II software differentiates between the CMU PLLs of the same transceiver block and the transmitter PLLs outside the transceiver block based on the Use central clock divider to drive the transmitter channels using x4/xN lines option. If you enable this option, the transmitter PLL is outside the transceiver block. Similarly, if you disable option, the transmitter PLL is one of the CMU PLLs within the same transceiver block. Logical Channel Addressing When Using Additional PLLs The logical channel addressing of the transceiver channel is the same as described in "Logical Channel Addressing" on page 5-5 so long as you are ONLY using the CMU PLLs within the same transceiver block. In the case of additional PLLs (when transmitter PLLs are outside the transceiver block), the additional PLLs also have their own logical channel address. This affects the starting channel number of the following ALTGX instances connected to the dynamic reconfiguration controller, if any. Therefore, you must take into account the logical channel address of transmitter PLLs outside the transceiver block when setting the Total number of channels controlled by the reconfig controller option in the ALTGX_RECONFIG instance. When you select the Use central clock divider to drive the transmitter channels using x4/xN lines option for an additional PLL, you can see its logical channel address value at the bottom of the corresponding PLL screen. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-53 Selecting the PLL Logical Reference Index for Additional PLLs The PLL logical reference index of additional PLLs outside the transceiver block can only be 2 or 3. When you enable the Use central clock divider to drive the transmitter channels using x4/xN lines option for an additional PLL, you can only select between 2 or 3 as the PLL logical reference index. When you disable the Use central clock divider to drive the transmitter channels using x4/xN lines option for an additional PLL, the additional PLL is one of the CMU PLLs within the same transceiver block. Therefore, the PLL logical reference index is either 0 or 1. For more information about the PLL logical reference index of CMU PLLs within the same transceiver block, refer to "Selecting the Logical Reference Index of the CMU PLL" on page 5-29. ALTGX_RECONFIG MegaWizard Plug-In Manager Setup for Channel Reconfiguration with Transmitter PLL Select Mode For more information, refer to the "ALTGX_RECONFIG MegaWizard Plug-In Manager Setup for Channel and CMU PLL Reconfiguration Mode" on page 5-47. Channel Reconfiguration with Transmitter PLL Select Operation Read transactions are not allowed in this mode. Figure 5-25 shows a .mif write transaction when dynamically reconfiguring a transceiver channel. The .mif write transaction in channel reconfiguration with transmitter PLL select mode remains the same except for the reconfig_mode_sel[2:0] value and the difference in the number of .mif words used. In this example, the transceiver channel is configured in Receiver and Transmitter configuration. Therefore, the .mif size is 8. You can optionally choose to trigger write_all once by selecting the continuous write operation in the ALTGX_RECONFIG MegaWizard Plug-In Manager. The Quartus II software then continuously writes all the words required for reconfiguration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-54 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-25. .mif write transaction in Channel and CMU PLL Reconfiguration Mode reconfig_mode_sel[2:0] 3'b101 logical_channel_address[1:0] 2'b01 rx_tx_duplex_sel[1:0] 2'b00 reconfig_clk write_all busy reconfig_address_out[5:0] Addr0 Addr1 1st 16 bits Don't care 2nd 16 bits Addr54 Addr0 reconfig_address_en reconfig_data[15:0] 55th 16 bits Don't care channel_reconfig_done For guidelines regarding re-using .mifs, specifying input reference clocks, or using logical_tx_pll_sel ports, refer to "Special Guidelines" on page 5-57. f For more information about reset, refer to the "Reset Sequence when Using Dynamic Reconfiguration with the Channel and TX PLL select/reconfig Option" section in the Reset Control and Power Down in Stratix IV Devices chapter. CMU PLL Reconfiguration Mode Details Use this mode to reconfigure only the CMU PLL without affecting the remaining blocks of the transceiver channel. When you reconfigure the CMU PLL of a transceiver block to run at a different data rate, all the transceiver channels listening to this CMU PLL also are reconfigured to the new data rate. 1 You cannot dynamically reconfigure a CMU PLL into a CMU channel and vice versa. For more information about the supported configurations in CMU PLL reconfiguration mode, refer to Table 5-5 on page 5-19. Transmitter PLL Powerdown During CMU PLL reconfiguration mode, the dynamic reconfiguration controller automatically powers down the selected CMU PLL until it completes reconfiguration. The ALTGX_RECONFIG instance does not provide external ports to control the CMU PLL power down. When you reconfigure the CMU PLL, the pll_locked signal goes low. Therefore, after reconfiguring the transceiver, wait for the pll_locked signal from the ALTGX instance before continuing normal operation. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-55 The dynamic reconfiguration controller powers down only the selected CMU PLL. The other CMU PLL is not affected. Blocks Reconfigured in CMU PLL Reconfiguration Mode Each transceiver block has two CMU PLLs--CMU0 PLL and CMU1 PLL.You can reconfigure each of these CMU PLLs to a different data rate in this mode. Figure 5-26 shows a view of the reconfigurable blocks using CMU PLL reconfiguration mode. Figure 5-26. CMU PLLs in a Transceiver Block in CMU PLL Reconfiguration Mode CMU Channels Full Duplex Transceiver Channel clock mux refclk0 TX CHANNEL CMU0 PLL refclk1 Logical TX PLL select clock mux LOCAL DIVIDER TX PMA + TX PCS CMU1 PLL RX CHANNEL clock mux RX CDR RX PMA + RX PCS (1) Note to Figure 5-26: (1) Depending on the mode you select, PCS may or may not be present. ALTGX MegaWizard Plug-In Manager Setup for CMU PLL Reconfiguration Mode If you want to reconfigure the CMU PLL to another data rate, enable .mif generation and set up the ALTGX MegaWizard Plug-In Manager, as described in the following steps. The dynamic reconfiguration controller reconfigures the CMU PLL with the new information stored in the .mif. 1. Select the Channel and Transmitter PLL reconfiguration option in the Modes screen. 2. Provide the new data rate you want the CMU PLL to run at in the General screen. 1 September 2012 The logical reference index of CMU0 PLL within a transceiver block is always the complement of the logical reference index of CMU1 PLL. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-56 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation ALTGX_RECONFIG Plug-In Manager Setup for CMU PLL Reconfiguration Mode For more information, refer to "ALTGX_RECONFIG MegaWizard Plug-In Manager Setup for Channel and CMU PLL Reconfiguration Mode" on page 5-47. CMU PLL Reconfiguration Operation Set the reconfig_mode_sel[2:0] signal to 3' b100 to activate this mode. Figure 5-27 shows a .mif write transaction in CMU PLL reconfiguration mode. The dynamic reconfiguration controller asserts the channel_reconfig_done signal to indicate that the CMU PLL reconfiguration is complete. In this example, the transceiver channel is configured in Receiver and Transmitter configuration. Therefore, the .mif size is 8. You can optionally choose to trigger write_all once by selecting the continuous write operation in the ALTGX_RECONFIG MegaWizard Plug-In Manager. The Quartus II software then continuously writes all the words required for reconfiguration. Figure 5-27. CMU PLL Reconfiguration .mif Write Transaction reconfig_mode_sel[2:0] 3'b100 reconfig_clk write_all busy reconfig_address_out[5:0] Addr0 Addr1 Addr7 Addr0 reconfig_address_en reconfig_data[15:0] 1st 16 bits Don't care 2nd 16 bits 8th 16 bits Don't care channel_reconfig_done For guidelines regarding re-using .mifs, specifying input reference clocks, or using logical_tx_pll_sel ports, refer to "Special Guidelines" on page 5-57. f For more information about reset, refer to the "Reset Sequence when Using Dynamic Reconfiguration with the Channel and TX PLL select/reconfig Option" section in the Reset Control and Power Down in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-57 Central Control Unit Reconfiguration Mode Details Central control unit reconfiguration mode is a .mif-based mode used to reconfigure the central control unit (CCU) of the transceiver. Use reconfig_mode_sel[] to activate this mode. Central control unit reconfiguration mode is applicable for bonded PCS configurations such as Basic x4 and x8, XAUI, and PCIe x4 and x8. For the allowed configurations, refer to Table 5-5 on page 5-19. For instance, to dynamically reconfigure an ALTGX instance in Basic x4 configuration to a XAUI configuration, you must first configure: 1. The transceiver channel and CMU PLL to run at the XAUI data rate and functional mode (use channel and CMU PLL reconfiguration mode). 2. Reconfigure the central control unit portion of the transceiver from Basic to XAUI functional mode (use central control unit reconfiguration mode). For more information about the central control unit reconfiguration mode, refer to "Example 2" on page 5-100. 1 Dynamic reconfiguration is not available if hard IP is used in PCIe mode. 1 To switch between one bonded PCS configuration and another, always use: 1) Channel and CMU PLL reconfiguration mode followed by 2) Central control unit reconfiguration mode Use the same .mif for both the these steps. In step 1, a partial .mif is written and the remaining contents of the .mif is written in step 2. In step 1, reconfigure all the channels one-by-one. In step-2, reconfiguration of the central control unit is transceiver-block based. Reconfigure any one of the four channels in the transceiver block. Special Guidelines The following section describes the special guidelines required for the transceiver channel reconfiguration modes previously described. This section includes the following: "Guidelines for Re-Using .mifs" on page 5-57 "Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports" on page 5-59 "Guidelines for Specifying the Input Reference Clocks" on page 5-61 Guidelines for Re-Using .mifs To configure the transceiver PLLs and receiver CDRs for multiple data rates, it is important to understand the input reference clock requirements. This helps you to efficiently create the clocking scheme for reconfiguration and to reuse the .mifs across all channels in the device. This section describes the clocking enhancements and the implications of using input clocks from various clock sources. The available clock inputs appear as a pll_inclk_rx_cruclk[] port and can be provided from the inter-transceiver block lines (also known as ITB lines), from the global clock networks that are driven by an input pin or by a PLL cascade clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-58 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation f For more information about input reference clocking, refer to the "Input Reference Clocking" section of the Transceiver Clocking in Stratix IV Devices chapter. The following section describes the clocking requirements to re-use .mifs. The .mif contains information about the input clock multiplexer settings and the functional blocks that you selected during the ALTGX MegaWizard Plug-In Manager instantiation. You can use a .mif to dynamically reconfigure any of the other transceiver channels in the device as long as the order of the clock inputs is consistent. For example, assume that a .mif is generated for a transceiver channel in transceiver block 0 and the input clock source is connected to the pll_inclk_rx_cruclk[0] port. When you use the generated .mif for a channel in other transceiver blocks (for example, transceiver block 1), the same clock source must be connected to the pll_inclk_rx_cruclk[0] port. Figure 5-28 and Figure 5-29 show the incorrect and correct order of input reference clocks, respectively. In Figure 5-28, the clocking is incorrect when re-using the .mif because the input reference clock is not connected to the corresponding pll_inclk_rx_cruclk[] ports in the two instances. Figure 5-28. Incorrect Input Reference Clock Connections When Reusing a .mif Stratix IV GX Device Transceiver Block 0 pll_inclk_rx_cruclk[0] 156.25 MHz (1) pll_inclk_rx_cruclk[1] ALTGX Instance 1 (1) Transceiver Block 1 pll_inclk_rx_cruclk[0] 125 MHz pll_inclk_rx_cruclk[1] ALTGX Instance 2 Note to Figure 5-28: (1) The red lines represent the alternate source of REFCLK. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-59 Figure 5-29 shows the correct input reference clock connections when re-using a .mif. Figure 5-29. Correct Input Reference Clock Connections When Reusing a .mif Stratix IV GX Device Transceiver Block 0 pll_inclk_rx_cruclk[0] 156.25 MHz pll_inclk_rx_cruclk[1] ALTGX Instance 1 (1) Transceiver Block 1 pll_inclk_rx_cruclk[0] 125 MHz pll_inclk_rx_cruclk[1] ALTGX Instance 2 (1) Note to Figure 5-29: (1) The red lines represent the alternate source of REFCLK. 1 You can re-use the .mif generated for a transceiver channel on one side of the device for a transceiver channel on the other side of the device, only if the input reference clock frequencies and order of the pll_inclk_rx_cruclk[] ports in the ALTGX instances on both sides are identical. In addition to the input reference clock requirements when re-using a .mif, refer to "Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports" on page 5-59 for additional ways to re-use a .mif. Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports This section describes when to enable the logical_tx_pll_sel and logical_tx_pll_sel_en ports and how to use them in the following dynamic reconfiguration modes: Channel and CMU PLL reconfiguration mode Channel reconfiguration with transmitter PLL select mode CMU PLL reconfiguration mode These are optional input ports to the ALTGX_RECONFIG instance. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-60 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-11 shows the conditions under which the dynamic reconfiguration controller uses either the logical_tx_pll_sel port value or the logical reference index value stored in the .mif. Figure 5-30 shows the logical_tx_pll_sel and logical_tx_pll_sel_en ports. Figure 5-30. Using logical_tx_pll_sel and logical_tx_pll_sel_en Ports ALTGX_RECONFIG instance logical tx pll (logical reference index value stored in the .mif) 0 selected logical reference index value 1 logical_tx_pll_sel (logical reference index specified on the port) logical_tx_pll_sel_en Table 5-11 lists how the dynamic reconfiguration controller selects between the logical reference index stored in the .mif (logical tx pll) and the logical reference index specified at the logical_tx_pll_sel port. Table 5-11. Various Combinations of the logical_tx_pll_sel and logical_tx_pll_sel_en Ports logical_tx_pll_sel logical_tx_pll_sel_en Logical Reference Index Value Selected by the ALTGX_RECONFIG Instance enabled enabled and value is 1 Value on the logical_tx_pll_sel port enabled enabled and value is 0 logical reference index value stored in the .mif (logical tx pll) enabled disabled Value on the logical_tx_pll_sel port disabled disabled logical reference index value stored in the .mif (logical tx pll) Altera recommends keeping track of the transmitter PLL that drives the channel when you configure a transceiver channel in the ALTGX MegaWizard Plug-In Manager. 1 The logical_tx_pll_sel port does not modify transceiver settings on the receiver side. If both the logical_tx_pll_sel and logical_tx_pll_sel_en ports are enabled, reconfigure the transmitter PLL. Keep the logical_tx_pll_sel and logical_tx_pll_sel_en signals at a constant logic level until the dynamic reconfiguration controller asserts the channel_reconfig_done signal. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-61 Table 5-12 lists the two conditions under which you can re-use .mifs when using the logical_tx_pll_sel and logical_tx_pll_sel_en ports. Table 5-12. Two Conditions Under Which You can Re-Use .mifs (logical_tx_pll_sel and logical_tx_pll_sel_en) Condition 1: Re-use the .mif created for one CMU PLL on the other CMU PLL of the same transceiver block. Condition 2: Re-use the .mif created for one transmitter PLL on the transmitter PLL of another transceiver block. Channel and CMU PLL Reconfiguration and CMU PLL Reconfiguration Channel Reconfiguration with Transmitter PLL Select Channel and CMU PLL Reconfiguration and CMU PLL Reconfiguration Consider that you create a .mif containing the desired ALTGX settings to reconfigure the CMU0 PLL. Assume that the logical reference index you assigned to CMU0 PLL is 0. Assume that the transceiver channel listens to CMU1 PLL and the logical reference index assigned to it is 0. You can re-use this .mif created for CMU0 PLL on CMU1 PLL of the same transceiver block if you want to reconfigure CMU1 PLL to the new data rate information stored in the .mif. You must set logical_tx_pll_ sel to the logical reference index of CMU1 PLL (1'b1) and logical_tx_pll_ sel_en to 1'b1 and then write this .mif into the transceiver channel. By doing so, the dynamic reconfiguration controller overwrites the logical tx pll value stored in the .mif with the logical reference index of CMU1 PLL. Generate a .mif for these settings. When you use channel reconfiguration with transmitter PLL select mode and reconfigure the transceiver channel with this .mif, the transceiver channel is reconfigured to listen to CMU1 PLL. If you want to reconfigure the transceiver channel to listen to CMU0 PLL instead, you can re-use this .mif. You must set logical_tx_pll_ sel to the logical reference index of CMU0 PLL (1'b1) and logical_tx_pll_ sel_en to 1'b1 and then write this .mif into the transceiver channel. Consider that you create a .mif containing the desired ALTGX settings to reconfigure the transmitter PLL of a transceiver block. Assume that the logical reference of the transmitter PLL is 1. You can re-use this .mif created to reconfigure the transmitter PLL of another transceiver block under the following condition: Channel Reconfiguration with Transmitter PLL Select Consider that you create a .mif containing the logical reference index of the transmitter PLL that the reconfigured transceiver channel needs to listen to. Assume that the transmitter PLL used is CMU0 PLL and the logical reference index assigned is 0. When you use channel reconfiguration with transmitter PLL select mode and reconfigure the transceiver channel with this .mif, the transceiver channel is reconfigured to listen to CMU0 PLL. If you want to reconfigure this transceiver channel to listen to another transmitter PLL outside the transceiver block, you can reuse this .mif, provided the intended data rate is the same. You want to reconfigure the transmitter PLL of the other transceiver block to exactly the same data rate information stored in the .mif. You must set logical_channel_ address to the logical channel address of the transmitter PLL you intend to reconfigure. Guidelines for Specifying the Input Reference Clocks The following are guidelines for setting up the input reference clocks in the Reconfiguration Settings screen of the ALTGX MegaWizard Plug-In Manager. September 2012 Assign the identification numbers to all input reference clocks that are used by the transmitter PLLs in their corresponding PLL screens. You can set up a maximum of 10 input reference clocks and assign identification numbers from 1 to 10. Keep the identification numbers consistent for all the .mifs generated in the design. Maintain the input reference clock frequencies settings for all the .mifs. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-62 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Figure 5-31 shows an example scenario where the input reference clock connections to the transceiver channels are based on what you set as the input clock source for each of the CMU transmitter PLLs within a transceiver block. Figure 5-31. Input Reference Clocks Connections to the Transceiver Channels Based on what you have set up as the input clock source for CMU0 PLL, this clock mux selects the corresponding input clock source for CMU0 PLL. CMU Channels Refclk0 (Identification number = 2) 156 .25 MHz clock mux Full Duplex Transceiver Channel 1 3.125 Gbps TX CHANNEL 1 CMU0 PLL 125MHz 3.125 Gbps Logical TX PLL select Refclk1 (Identification number = 1) LOCAL DIVIDER TX PMA + TX PCS (1) 1 Gbps clock mux CMU1 PLL RX CHANNEL 1 3.125 Gbps clock mux RX CDR 3.125 Gbps RX PMA + RX PCS Based on what you have set up as the input clock source for CMU1 PLL, this clock mux selects the corresponding input clock source for CMU1 PLL. Full Duplex Transceiver Channel 2 TX CHANNEL 2 1 Gbps Logical TX PLL select LOCAL DIVIDER TX PMA + TX PCS (1) RX CHANNEL 2 clock mux 1 Gbps RX CDR 1 Gbps RX PMA + RX PCS Note to Figure 5-31: (1) Depending on the mode you select, the PCS unit may or may not be present. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-63 Data Rate Division in Transmitter Mode Details You can use data rate division in transmitter mode to modify the data rate of the transmitter channel in multiples of 1, 2, and 4. This dynamic reconfiguration mode is available only for the transmit side and not for the receive side. Blocks Reconfigured in the Data Rate Division in Transmitter Mode The only block that is reconfigured by the data rate division in transmitter mode is the transmitter local divider block of a transmitter channel. You can set the transmitter local divider to a divide by value of /1, /2, or /4, as shown in Figure 5-32. Figure 5-32. Local Divider of a Transmitter Channel High-Speed Serial Clock High-Speed clock from TX PLL0 /n /4, /5, /8, or /10 Low-Speed Parallel Clock High-Speed clock from TX PLL1 /1, /2, or /4 You must be aware of the device operating range before you enable and use this feature. There are no legal checks that are imposed by the Quartus II software because it is an on-the-fly control feature. You must ensure that a specific functional mode supports the data rate range before dividing the clock when using this rate switch option. 1 Data rate division in transmitter mode is applicable only to channels configured in non-bonded mode clocked by the CMU0/CMU1 located within the same transceiver block. ALTGX MegaWizard Plug-In Manager Setup for Data Rate Division in Transmitter Mode Enable the following settings in the ALTGX MegaWizard Plug-In Manager: 1. Select the Channel and Transmitter PLL Reconfiguration option in the Reconfig screen to enable the ALTGX_RECONFIG instance to modify the transmitter channel local divider values dynamically. 2. Set the What is the starting channel number? option in the Reconfig screen. For more information, refer to "Logical Channel Addressing" on page 5-5. The alternate reference clock is not required because a single clock source is used. The /1, /2, or /4 data rates can be derived from the single input reference clock. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-64 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation ALTGX_RECONFIG MegaWizard Plug-In Manager Setup for Data Rate Division in Transmitter Mode Enable the following settings in the ALTGX_RECONFIG MegaWizard Plug-In Manager for data rate division in transmitter mode: 1. In the Reconfiguration settings screen, set the What is the number of channels controlled by the reconfig controller? option. For more information, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. 2. Specify the logical channel address of the transmitter channel at the logical_channel_address input port. 3. In the Reconfiguration settings screen, select the Data rate division in TX option. The rate_switch_ctrl[1:0] input port is available when you enable the Data rate division in TX option. The value you set at the rate_switch_ctrl[1:0] signal determines the transmitter local divider settings, as explained in "Dynamic Reconfiguration Controller Port List" on page 5-78. To read the existing local divider settings of the transmitter channel, select the Use 'rate_switch_out' port to read out the current data rate division option in the Error checks/Data rate switch screen. Decoding for the rate_switch_out[1:0] output signal is the same as the rate_switch_ctrl[1:0] input signal. 1 Dynamic rate switch has no effect on the dividers on the receive side of the transceiver channel. It can be used only for the transmitter. 1 Data rate division in transmitter mode does not require a .mif. Data Rate Division in Transmitter Operation The following sections describe the steps involved in write and read transactions for the data rate division in transmitter mode. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-65 For this example, the value set in the What is the number of channels controlled by the reconfig controller? option of the ALTGX_RECONFIG MegaWizard Plug-In Manager is 4. Therefore, the logical_channel_address input is 2 bits wide. Also, you must reconfigure the local divider settings of the transmitter channel whose logical channel address is 2'b01. Figure 5-33 shows a write transaction in data rate division in transmitter mode. Figure 5-33. Write Transaction in Data Rate Division in Transmitter Mode 3'bXXX 3'b011 rate_switch_ctrl[1:0] (1) 2'bXX 2'b10 2'bXX logical_channel_address 2'bXX 2'b01 2'bXX reconfig_mode_sel[2:0] reconfig_clk write_all busy Note to Figure 5-33: (1) For this example, you want to reconfigure the local divider settings of the transmitter channel to Divide by 4. Therefore, the value set at rate_switch_ctrl[1:0] is 2'b10. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-66 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation For this example, the value set in the What is the number of channels controlled by the reconfig controller? option of the ALTGX_RECONFIG MegaWizard Plug-In Manager is 4. Therefore, the logical_channel_address input is 2 bits wide. Also, you must read the existing local divider settings of the transmitter channel whose logical channel address is 2'b01. Figure 5-34 shows a read transaction waveform in data rate division in transmitter mode. Figure 5-34. Read Transaction in Data Rate Division in Transmitter Mode reconfig_mode_sel[2:0] 3'bXXX 3'b011 2'bXX 2'b01 reconfig_clk logical_channel_address (1) 2'bXX read busy rate_switch_out[1:0] 2'bXX 2'b01 Invalid output data_valid Note to Figure 5-34: (1) For this example, the existing local divider settings of the transmitter channel are Divide by 2. Therefore, the value read out at rate_switch_out[1:0] is 2'b01. 1 Do not perform a read transaction in date rate division in transmitter mode if rate_switch_out[1:0] is not selected in the ALTGX_RECONFIG MegaWizard Plug-In Manager. f For more information about reset, refer to the "Reset Sequence when Using Dynamic Reconfiguration with the Channel and TX PLL select/reconfig Option" section in the Reset Control and Power Down in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-67 Offset Cancellation Feature The Stratix IV GX and GT devices provide an offset cancellation circuit per receiver channel to counter the offset variations due to process, voltage, and temperature (PVT). These variations create an offset in the analog circuit voltages, pushing them out of the expected range. In addition to reconfiguring the transceiver channel, the dynamic reconfiguration controller performs offset cancellation on all receiver channels connected to it on power up. The Offset cancellation for Receiver channels option is automatically enabled in both the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers for Receiver and Transmitter and Receiver only configurations. It is not available for Transmitter only configurations. For Receiver and Transmitter and Receiver only configurations, you must connect the necessary interface signals between the ALTGX_RECONFIG and ALTGX (with receiver channels) instances. Offset cancellation is automatically executed once every time the device is powered on. The control logic for offset cancellation is integrated into the dynamic reconfiguration controller. You must connect the ALTGX_RECONFIG instance to the ALTGX instances (with receiver channels) in your design. You must connect the reconfig_fromgxb, reconfig_togxb, and necessary clock signals to both the ALTGX_RECONFIG and ALTGX (with receiver channels) instances. 1 The offset cancellation control functionality remains the same for both regular transceiver channels and PMA-only channels. Operation Every ALTGX instance for Receiver and Transmitter or Receiver only configurations require that the Offset cancellation for Receiver channels option is enabled in the Reconfig screen of the ALTGX MegaWizard Plug-In Manager. This option is enabled by default for the above two configurations. It is disabled for the Transmitter only configuration. Because this option is enabled by default, the ALTGX instance must be connected to an ALTGX_RECONFIG instance (dynamic reconfiguration controller). The offset cancellation controls are also enabled by default in the Reconfiguration settings screen of the ALTGX_RECONFIG instance. You must also set the starting channel number in the What is the starting channel number? option for every ALTGX instance connected to the ALTGX_RECONFIG instance. For more information, refer to: September 2012 "Logical Channel Addressing of Regular Transceiver Channels" on page 5-6 "Logical Channel Addressing of PMA-Only Channels" on page 5-7 "Logical Channel Addressing--Combination of Regular Transceiver Channels and PMA-Only Channels" on page 5-9 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-68 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation When the device powers up, the dynamic reconfiguration controller initiates offset cancellation on the receiver channel by disconnecting the receiver input pins from the receiver data path. It also sets the receiver CDR into a fixed set of dividers to guarantee a voltage controlled oscillator (VCO) clock rate within the range necessary to provide proper offset cancellation. Subsequently, the offset cancellation process goes through different states and culminates in the offset cancellation of the receiver buffer and receiver CDR. After offset cancellation is complete, the user divider settings are restored. The dynamic reconfiguration controller sends and receives data to the transceiver channel through the reconfig_togxb and reconfig_fromgxb signals. You must connect these signals between the ALTGX_RECONFIG instance and the ALTGX instance. You must also set the What is the number of channels controlled by the reconfig controller? option in the Reconfiguration settings screen of the ALTGX_RECONFIG MegaWizard Plug-In Manager. For more information, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. The Use 'logical_channel_address' port for Analog controls reconfiguration option in the Analog controls screen of the ALTGX_RECONFIG MegaWizard Plug-In Manager is not applicable for the receiver offset cancellation process. 1 If the design does not require PMA controls reconfiguration and uses optimum logic element (LE) resources, you can connect all the ALTGX instances in the design to a single dynamic reconfiguration controller (ALTGX_RECONFIG instance). 1 The gxb_powerdown signal must not be asserted during the offset cancellation sequence. To understand the impact on system start-up when you control all the transceiver channels using a single dynamic reconfiguration controller, refer to "PMA Controls Reconfiguration Duration" on page 5-91. ALTGX_RECONFIG Instance Signals Transition during Offset Cancellation Consider that the design has ALTGX instances with channels of both Transmitter only and Receiver only configurations. You must include the Transmitter only channels while setting the What is the starting channel number? option in the ALTGX instance and setting the What is the number of channels controlled by the reconfig controller? option in the ALTGX_RECONFIG instance for receiver offset cancellation. Stratix IV Device Handbook Volume 2: Transceivers After the device powers up, the busy signal remains low for the first reconfig_clk clock cycle. The busy signal then gets asserted for the second reconfig_clk clock cycle when the dynamic reconfiguration controller initiates the offset cancellation process. The de-assertion of the busy signal indicates the successful completion of the offset cancellation process. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-69 Figure 5-35 shows the dynamic reconfiguration signals transition during offset cancellation on the receiver channels. Figure 5-35. Dynamic Reconfiguration Signals Transition during Offset Cancellation on Receiver Channels reconfig_clk busy (1) Note to Figure 5-35: (1) After device power up, the busy signal remains low for the first reconfig_clk cycle. 1 Due to the offset cancellation process, the transceiver reset sequence has changed. For more information, refer to the Reset Control and Power Down in Stratix IV Devices chapter. EyeQ EyeQ hardware is available in Stratix IV transceivers to analyze and debug the receiver data recovery path (receiver gain, clock jitter, and noise level). You can use it to monitor the eye width and assess the quality of the incoming signal. Normally, the receiver CDR samples the incoming signal at the center of the eye. When you enable the EyeQ hardware, it allows the CDR to sample across 32 different positions across one unit interval (UI) of the incoming data. You can manually control the sampling points and check the bit-error rate (BER) at each of these 32 sampling points. These sampling points are also known as phase steps. The BER increases at the edge of the eye-opening. By observing the number of sampling points results in a desired BER value, you can determine the eye width. 1 The EyeQ hardware is available for both regular transceiver channels and CMU channels. f For more information about the supported data rates, phase step translation, and other specifications, refer to the DC and Switching Characterization for Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-70 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Enabling the EyeQ Control Logic and the EyeQ Hardware You must enable the EyeQ hardware in the ALTGX MegaWizard Plug-In Manager and the EyeQ control block in the ALTGX_RECONIG MegaWizard Plug-In Manager. EyeQ hardware is available for each transceiver channel in the receiver data path. Select the Analog Controls option in the Reconfiguration Settings screen of the ALTGX MegaWizard Plug-In Manager to enable the EyeQ hardware. EyeQ control logic is available in the dynamic reconfiguration controller. Select the EyeQ control option in the ALTGX_RECONFIG MegaWizard Plug-In Manager to enable the EyeQ control logic. f EyeQ uses an Avalon Memory Mapped interface. For more information about this interface, refer to the Avalon Interface Specification. Connections Between the ALTGX and ALTGX_RECONFIG Instances To enable the EyeQ options, follow these steps: 1. Enable the EyeQ options in the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers as explained in "Enabling the EyeQ Control Logic and the EyeQ Hardware" on page 5-70. 2. Connect the reconfig_{to/from}gxb ports between the ALTGX and ALTGX_RECONFIG instances. EyeQ control logic in the dynamic reconfiguration controller allows you to write to the registers in the EyeQ hardware. Therefore, you must have a state machine in the user design that communicates to the EyeQ control block of the ALTGX_RECONFIG instance. You can then access the internal registers of the EyeQ hardware indirectly through the EyeQ control logic. 1 Stratix IV Device Handbook Volume 2: Transceivers Altera recommends having an input pattern generator and checker to monitor the BER of the received data. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-71 Figure 5-36 shows the connections between the EyeQ hardware in the ALTGX instances and the EyeQ control logic in the dynamic reconfiguration controller. Figure 5-36. Connecting ALTGX and ALTGX_RECONFIG Instances with EyeQ Enabled ctrl_writedata[15:0] ctrl_address[15:0] ctrl_write ctrl_read ALTGX Instance ALTGX_RECONFIG Instance reconfig_fromgxb[17:0] EyeQ Hardware EyeQ Control Block rx_datain[0] reconfig_togxb[3:0] Receiver Channel 0 busy error ctrl_waitrequest reconfig_mode_sel[3:0] ctrl_readdata[15:0] Controlling the EyeQ Hardware The EyeQ hardware is controlled by writing to the EyeQ registers using EyeQ interface registers in the ALTGX_RECONFIG instance. Table 5-13 lists the register memory of the 16-bit EyeQ registers. Table 5-13. EyeQ Register Address Mapping Address 0x0 Description Bit[0]--0/1: Disable/Enable EyeQ feature Bit [15:1]--15'b000000000000000 Bits [5:0]--EyeQ phase step value. Refer to Table 5-14 for the EyeQ phase step encoding. Bits [15:6]--10'b0000000000 0x1 September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-72 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation Table 5-14 lists the EyeQ phase step encoding for the 32 phase steps spanning one unit interval (UI). Table 5-14. EyeQ Phase Step Encoding Stratix IV Device Handbook Volume 2: Transceivers Desired Phase Step Setting EyeQ Phase Step Encoding 0 6'b111111 1 6'b111110 2 6'b111101 3 6'b111100 4 6'b111011 5 6'b111010 6 6'b111001 7 6'b111000 8 6'b110111 9 6'b110110 10 6'b110101 11 6'b110100 12 6'b110011 13 6'b110010 14 6'b110001 15 6'b110000 16 6'b010000 17 6'b010001 18 6'b010010 19 6'b010011 20 6'b010100 21 6'b010101 22 6'b010110 23 6'b010111 24 6'b011000 25 6'b011001 26 6'b011010 27 6'b011011 28 6'b011100 29 6'b011101 30 6'b011110 31 6'b011111 September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-73 Table 5-15 lists the register memory of the 16-bit EyeQ interface registers. Table 5-15. EyeQ Interface Register Mapping Address Description Control/Status register (EyeQ CSR) 0x0 Bit [0]--Start: Writing a 1 to this bit instructs the ALTGX_RECONFIG instance to program the EyeQ hardware. Writing to this bit automatically clears any error bits. Bit [1]--Read/Write: Writing a 0 to this bit writes the contents of the data register to one of the EyeQ registers depending on the address stored in the EyeQ register address register. Writing a 1 reads the contents of the EyeQ register. Bit [12:2]--11'b00000000000 Bit [13]--Channel address error: This bit is set to 1 if the programmed channel address is invalid. Writing a 1 to this bit clears the error. Bit [14]--EyeQ register address error: this bit is set to 1 if the programmed word address is invalid. Writing a 1 to this bit clears the error. Bit [15]--Busy status: The value of this bit can be polled to determine if the ALTGX_RECONFIG read/write request has completed. When this active-high bit is asserted, all registers become read only until this bit is de-asserted. 0x1 Channel address [15:0]--Specifies the transceiver channel for the desired EyeQ operation. This must match the logical_channel_address input port. 0x2 EyeQ register address [15:0]--Specifies the address EyeQ register to be read from or written to. The values supported are 0x0 or 0x1. Data [15:0]-- 0x3 For a write operation, the data in this register is written to the EyeQ register selected. For a read operation, this register contains the contents of the EyeQ register selected. The data in this register is only valid when the busy status is low. A read operation overwrites the current contents of this register. To control the EyeQ hardware, follow these steps: 1. Read the EyeQ interface register 0x0 (the control and status register) to check the busy status. The clear status bit indicates an idle status. 2. Issue a write to the EyeQ interface register 0x1 (the channel address register) to select the desired channel. 3. Issue a write to the EyeQ interface register 0x2 (the eye monitor register address) to select the desired EyeQ register. 4. Issue a write to the EyeQ interface register 0x3 (the data register) to provide the data to be written to the target EyeQ register. 5. Issue a write to EyeQ interface register 0x0 (the control and status register) to specify read/~write and to issue the start command. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-74 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 6. Poll the EyeQ interface register 0x0 (the control and status register) and wait for the busy status to be de-asserted. After the status is no longer busy, the data is considered successfully written for write transactions. For read transactions, this indicates that the contents of the data register has been updated and can be read out. Note that all writes that occur when the busy status is asserted are ignored; all registers become read only. 7. If the next operation is to the same EyeQ register and same channel, you do not need to repeat steps 2 and 3. Example of Using the EyeQ Feature Consider a design with one regular transceiver channel configured in Basic functional mode. The channel has a data rate of 2.5 Gbps with the EyeQ feature enabled in both the ALTGX and ALTGX_RECONFIG instances. Figure 5-37 shows how the EyeQ mode is first enabled by writing into the EyeQ registers using the EyeQ interface registers. A phase step value of 25 is written to the EyeQ register. Before performing any operation, the following conditions must be met: busy is 0 in the EyeQ CSR ctrl_waitrequest is low Figure 5-37. Enabling EyeQ Mode Set EyeQ Phase Step Value to 25 Enable EyeQ reconfig_mode_sel[3:0] 4'b1011 reconfig_clk (1) ctrl_address[15:0] 0 1 2 (1) 3 0 ctrl_read ... ctrl_write ... ctrl_waitrequest ... ctrl_readdata[15:0] ctrl_writedata[15:0] ... 0 x 4 0 1 2 0 3 0 25 1 busy Note to Figure 5-37: (1) Writing a '1' here instructs the ALTGX_RECONFIG instance to program the EyeQ hardware. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-75 Adaptive Equalization (AEQ) High-speed interface systems require different equalization settings to compensate for changing data rates and backplane losses. Manual tuning of the receiver channel's equalization stages involves finding the optimal settings through trial and error, and then locking in those values at compile time. This manual method is cumbersome under varying system characteristics. The AEQ feature solves this problem by automatically tuning an active receiver channel's equalization filters based on a frequency content comparison between the incoming signal and internally generated reference signals. User logic can dynamically control the AEQ hardware in the receiver through the dynamic reconfiguration controller. This section describes how to enable different options and use them to control the AEQ hardware. Adaptive Equalization Limitations The following are the AEQ feature requirements and limitations: The receive data must be 8B/10B encoded Not available in PCIe functional mode (because the AEQ hardware cannot perform the equalization process when the receive link is under the electrical idle condition) The receiver input signal must have a minimum envelope of 400 mv (differential peak-to-peak). The Quartus II software does not check for this requirement The AEQ hardware is not present in the CMU channels f For more information about speed grade, data rates, receiver input signal level, and other specifications that support the AEQ feature, refer to the DC and Switching Characterization for Stratix IV Devices chapter. Enabling the AEQ Control Logic and AEQ Hardware To use the AEQ feature, enable the AEQ hardware in the ALTGX MegaWizard Plug-In Manager and the AEQ control block in the ALTGX_RECONIG MegaWizard Plug-In Manager. To enable the AEQ hardware and the AEQ control logic: Select the Enable adaptive equalizer control option in the Reconfiguration Settings screen of the ALTGX MegaWizard Plug-In Manager. The AEQ hardware is available for each transceiver channel in the receiver data path. Select the Enable adaptive equalizer control option in the ALTGX_RECONFIG MegaWizard Plug-In Manager. The AEQ control logic is available in the dynamic reconfiguration controller. When you select the above two options, the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers provide the following additional ports: aeq_fromgxb[] aeq_togxb[] The aeq_fromgxb[] and aeq_togxb[] ports provide the interface between the receiver channel and the dynamic reconfiguration controller. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-76 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation The following section describes the connections between the AEQ control block of the ALTGX_RECONFIG instance and the AEQ hardware of the ALTGX instance. Connections Between the ALTGX and ALTGX_RECONFIG Instances Enable the adaptive equalization options in the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers, as explained in the previous section. To use the AEQ control block and AEQ hardware, you must connect the ALTGX receivers to the ALTGX_RECONFIG instance using the reconfig_{to/from}gxb and aeq_{to/from}gxb ports. You must also connect the ALTGX_RECONFIG instance to your design. If you have multiple transceiver instances and a single ALTGX_RECONFIG instance, connect the LSB of the aeq_togxb[] and aeq_fromgxb[] ports of the ALTGX_RECONFIG instance to the transceiver channel with a logical_channel_address value of 0. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Modes Implementation 5-77 Figure 5-38 shows the aeq_fromgxb[] and aeq_togxb[] connections between multiple ALTGX instances and the dynamic reconfiguration controller. Figure 5-38. Connecting the ALTGX and ALTGX_RECONFIG Instances with AEQ Enabled ALTGX Instance 1 Receiver Channel 0 aeq_togxb[23:0] logical_channel_address = 0 aeq_fromgxb[7:0] rx_datain[0] AEQ Hardware ALTGX_RECONFIG busy User Logic AEQ Control Block error ALTGX Instance 2 Receiver Channel 1 logical_channel_address = 4 aeq_fromgxb[15:8] rx_datain[1] AEQ Hardware reconfig_mode_sel[3:0] aeq_togxb[47:24] ALTGX Instance 3 Receiver Channel 2 to 4 aeq_fromgxb[23:16] logical_channel_address = 8 AEQ Hardware rx_datain[2] aeq_togxb[63:48] September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-78 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List One Time Mode for a Single Channel Stratix IV GX and GT devices only support one-time adaptation mode for the AEQ feature. Figure 5-39 shows the AEQ timing diagram in this mode. Figure 5-39. AEQ Timing Diagram in One-Time Adaptation Mode write_all reconfig_mode_sel[3:0] logical_channel_address[] 4'b1001 indicates AEQ convergence and entry into stand-by mode 4 (logical channel 4) busy Offset calibration and AEQ convergence After assertion of the write_all signals, the dynamic reconfiguration controller performs the following steps sequentially: 1. Powers down the receiver buffer and performs offset calibration for the target channel. 2. Powers up the receiver buffer and runs the convergence algorithm to set the appropriate equalization settings. 3. Puts the AEQ circuitry in stand-by mode maintaining the converged equalization setting. In standby mode, no further adaptation occurs. If you observe bit errors over time with the converged equalization settings, you can re-initiate one-time adaptation by following the timing diagram shown in Figure 5-39. Each time you re-initiate one-time adaptation, the receiver buffer is powered down for offset calibration, thereby interrupting the link during this time. Dynamic Reconfiguration Controller Port List Table 5-16 lists the input control ports and output status ports of the dynamic reconfiguration controller. Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 1 of 13) Port Name Input/ Output (3), (4) Description Clock Inputs to the ALTGX_RECONFIG Instance The frequency range of this clock depends on the following transceiver channel configuration modes: reconfig_clk Input Receiver only (37.5 MHz to 50 MHz) Receiver and Transmitter (37.5 MHz to 50 MHz) Transmitter only (2.5 MHz to 50 MHz) By default, the Quartus II software assigns a global clock resource to this port. This clock must be a free-running clock sourced from an I/O clock pin. Do not use dedicated transceiver REFCLK pins or any clocks generated by transceivers. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-79 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 2 of 13) Port Name Input/ Output (3), (4) Description ALTGX and ALTGX_RECONFIG Interface Signals An output port in the ALTGX instance and an input port in the ALTGX_RECONFIG instance. This signal is transceiver-block based. Therefore, the width of this signal increases in steps of 17 bits per transceiver block. In the ALTGX MegaWizard Plug-In Manager, the width of this signal depends on the following: Whether the channels configured in the ALTGX instance are regular transceiver channels or PMA-only channels. The number of channels you select in the What is the number of channels? option in the General screen. For example, if the channels in the ALTGX instance are regular transceiver channels and if you select the number of channels as follows: 1 Channels 4, then the output port reconfig_fromgxb = 17 bits 5 Channels 8, then the output port reconfig_fromgxb = 34 bits reconfig_fromgxb Input 9 Channels 12, then the output port reconfig_fromgxb = 51 bits However, if the channels in the ALTGX instance are PMA-only channels and if you select the number of channels as follows: Number of PMA-only channels = n, then the output port reconfig_fromgxb = n*17 bits For example, reconfig_fromgxb = 6 * 17 bits for 6 PMA-only channels. In the ALTGX_RECONFIG MegaWizard Plug-In Manager, the width of this signal depends on the value you select in the What is the number of channels controlled by the reconfig controller? option in the Reconfiguration settings screen. For example, if you select the total number of channels controlled by ALTGX_RECONFIG instance as follows: 1 Channels 4, then the input port reconfig_fromgxb = 17 bits 5 Channels 8, then the input port reconfig_fromgxb = 34 bits 9 Channels 12, then the input port reconfig_fromgxb = 51 bits September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-80 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 3 of 13) Port Name Input/ Output (3), (4) Description To connect the reconfig_fromgxb port between the ALTGX_RECONFIG instance and multiple ALTGX instances, follow these rules: reconfig_fromgxb (continued) Input Connect the reconfig_fromgxb[16:0] of ALTGX Instance 1 to the reconfig_fromgxb[16:0] of the ALTGX_RECONFIG instance. Connect the reconfig_fromgxb[] port of the next ALTGX instance to the next available bits of the ALTGX_RECONFIG instance, and so on. Connect the reconfig_fromgxb port of the ALTGX instance, which has the highest What is the starting channel number? option, to the MSB of the reconfig_fromgxb port of the ALTGX_RECONFIG instance. The Quartus II Fitter produces an error if the dynamic reconfiguration option is enabled in the ALTGX instance but the reconfig_fromgxb and reconfig_togxb ports are not connected to the ALTGX_RECONFIG instance. For more information, refer to "Connecting the ALTGX and ALTGX_RECONFIG Instances" on page 5-11. reconfig_togxb[3:0] Output An input port of the ALTGX instance and an output port of the ALTGX_RECONFIG instance. You must connect the reconfig_togxb[3:0] input port of every ALTGX instance controlled by the dynamic reconfiguration controller to the reconfig_togxb[3:0] output port of the ALTGX_RECONFIG instance. The width of this port is always fixed to 3 bits. For more information, refer to "Connecting the ALTGX and ALTGX_RECONFIG Instances" on page 5-11. FPGA Fabric and ALTGX_RECONFIG Interface Signals Assert this signal for one reconfig_clk clock cycle to initiate a write transaction from the ALTGX_RECONFIG instance to the ALTGX instance. You can use this signal in two ways for .mif-based modes: write_all Stratix IV Device Handbook Volume 2: Transceivers Continuous write operation--Select the Enable continuous write of all the words needed for reconfiguration option to pulse the write_all signal only once for writing a whole .mif. The What is the read latency of the MIF contents option is available for selection in this case only. Enter the desired latency in terms of the reconfig_clk cycles. Regular write operation--When the Enable continuous write of all the words needed for reconfiguration option is disabled, every word of the .mif requires its own write cycle. Input September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-81 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 4 of 13) Port Name Input/ Output (3), (4) Description Used to indicate the busy status of the dynamic reconfiguration controller during offset cancellation. After the device powers up, this signal remains low for the first reconfig_clk clock cycle. It then is asserted and remains high when the dynamic reconfiguration controller performs offset cancellation on all the receiver channels connected to the ALTGX_RECONFIG instance. Output busy De-assertion of the busy signal indicates the successful completion of the offset cancellation process. For more information, refer to "Operation" on page 5-67. Input read PMA controls reconfiguration mode--This signal is high when the dynamic reconfiguration controller performs a read or write transaction. All other dynamic reconfiguration modes--This signal is high when the dynamic reconfiguration controller writes the .mif into the transceiver channel. Assert this signal for one reconfig_clk clock cycle to initiate a read transaction. The read port is applicable only to the PMA controls reconfiguration mode and data rate division in transmitter mode. The read port is available when you select Analog controls in the Reconfiguration settings screen and select at least one of the PMA control ports in the Analog controls screen. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. Applicable only to PMA controls reconfiguration mode. This port indicates the validity of the data read from the transceiver by the dynamic reconfiguration controller. Output data_valid The current data on the output read ports is the valid data ONLY if data_valid is high. This signal is enabled when you enable at least one PMA control port used in read transactions, for example tx_vodctrl_out. Output error Indicates that an unsupported operation is attempted. You can select this in the Error checks/Data rate switch screen. The dynamic reconfiguration controller de-asserts the busy signal and asserts the error signal for two reconfig_clk cycles when you attempt an unsupported operation. For more information, refer to the "Error Indication During Dynamic Reconfiguration" on page 5-90. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-82 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 5 of 13) Port Name Input/ Output (3), (4) Description Enabled by the ALTGX_RECONFIG MegaWizard Plug-In Manager when you enable the Use 'logical_channel_address' port for Analog controls reconfiguration option in the Analog controls screen. logical_channel_address [8:0] Input The width of the logical_channel_address port depends on the value you set in the What is the number of channels controlled by the reconfig controller? option in the Reconfiguration settings screen. This port can be enabled only when the number of channels controlled by the dynamic reconfiguration controller is more than one. For more information, refer to "Logical Channel Addressing of Regular Transceiver Channels" on page 5-6 and "Logical Channel Addressing of PMA-Only Channels" on page 5-7. A 2 bit wide signal. You can select this in the Error checks/Data rate switch screen. The advantage of using this optional port is that it allows you to reconfigure only the transmitter portion of a channel, even if the channel configuration is duplex. rx_tx_duplex_sel[1:0] Stratix IV Device Handbook Volume 2: Transceivers Input For a setting of: rx_tx_duplex_sel[1:0] = 2'b00--the transmitter and receiver portion of the channel is reconfigured. rx_tx_duplex_sel[1:0] = 2'b01--the receiver portion of the channel is reconfigured. rx_tx_duplex_sel[1:0] = 2'b10--the transmitter portion of the channel is reconfigured. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-83 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 6 of 13) Port Name Input/ Output (3), (4) Description Analog Settings Control/Status Signals An optional transmit buffer VOD control signal. It is 3 bits per transmitter channel. The number of settings varies based on the transmit buffer supply setting and the termination resistor setting on the TX Analog screen of the ALTGX MegaWizard Plug-In Manager. The width of this signal is fixed to 3 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 3 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. tx_vodctrl[2:0] (1) Input The following shows the VOD values corresponding to the tx_vodctrl settings for 100- termination. For more information, refer to the "Programmable Output Differential Voltage" section of the Transceiver Architecture in Stratix IV Devices chapter. tx_vodctrl[2:0] VOD (mV) for 1.4 V VCCH 3'b000 200 3'b001 400 3'b010 600 3'b011 700 3'b100 800 3'b101 900 3'b110 1000 3'b111 1200 An optional transmit buffer VOD control signal for Gen2. The signal is 3 bits per transmitter channel. The following shows the VOD values corresponding to the tx_vodctrla port for 100 termination: tx_vodctrla tx_vodctrla[2:0] tx_vodctrla_out[2:0] September 2012 Altera Corporation Input Output VOD (mV) Value 0 200 1 400 2 600 3 800 4 1000 5 1200 6 700 7 900 An optional transmit VOD read control signal. The tx_vodctrla_out[2:0] signal reads out the Gen2 VOD value written in the control register. Stratix IV Device Handbook Volume 2: Transceivers 5-84 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 7 of 13) Port Name Input/ Output (3), (4) Description An optional pre-emphasis control for pre-tap for the transmit buffer. Depending on what value you set at this input, the controller dynamically writes the value to the pre-emphasis control register of the transmit buffer. This signal controls both pre-emphasis positive and its inversion. The width of this signal is fixed to 5 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 5 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. tx_preemp_0t[4:0] (1) Input The following values are the legal settings allowed for this signal: 0 represents 0 1-15 represents -15 to -1 16 represents 0 17 - 31 represents 1 to 15 In the PCIe configuration, set tx_preemp_0t[4:0] to 5'b00000 when you do a rate switch from Gen 1 mode to Gen 2 mode. This is to ensure that tx_preemp_0t[4:0] does not add to the signal boost when tx_pipemargin and tx_pipedeemph take affect in PCIe Gen 2 mode. For more information, refer to the "Programmable Pre-Emphasis" section of the Transceiver Architecture in Stratix IV Devices chapter. An optional pre-emphasis write control for the first post-tap for the transmit buffer. Depending on what value you set at this input, the controller dynamically writes the value to the first post-tap control register of the transmit buffer. tx_preemp_1t[4:0] (1) Input The width of this signal is fixed to 5 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 5 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13 and the "Programmable Pre-Emphasis" section of the Transceiver Architecture in Stratix IV Devices chapter. tx_preemp_1ta[4:0] tx_preemp_1ta_out[4:0] tx_preemp_1tb[4:0] Stratix IV Device Handbook Volume 2: Transceivers Input An optional pre-emphasis control for the first post-tap for the transmit buffer in Gen2 mode. Depending on what value you set at this input, the controller dynamically writes the value to the pre-emphasis control register of the transmit buffer. Output An optional first post-tap, pre-emphasis read control signal for Gen2. The tx_preemp_1ta_out[4:0] signal reads out the value written by its input control signal. Input An optional de-emphasis control for the first post-tap for the transmit buffer in Gen2 mode. Depending on what value you set at this input, the controller dynamically writes the value to the de-emphasis control register of the transmit buffer. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-85 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 8 of 13) Port Name tx_preemp_1tb_out[4:0] Input/ Output Output (3), (4) Description An optional first post-tap, de-emphasis read control signal for Gen2. The tx_preemp_1tb_out[4:0] signal reads out the value written by its input control signal. An optional pre-emphasis write control for the second post-tap for the transmit buffer. This signal controls both pre-emphasis positive and its inversion. Depending on what value you set at this input, the controller dynamically writes the value to the pre-emphasis control register of the transmit buffer. The width of this signal is fixed to 5 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 5 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. tx_preemp_2t[4:0] (1) Input The following values are the legal settings allowed for this signal: 0 represents 0 1-15 represents -15 to -1 16 represents 0 17-31 represents 1 to 15 In the PCIe configuration, set tx_preemp_2t[4:0] to 5'b00000 when you do a rate switch from Gen 1 mode to Gen 2 mode. This is to ensure that tx_preemp_2t[4:0] does not add to the signal boost when tx_pipemargin and tx_pipedeemph take affect in PCIe Gen 2 mode. For more information, refer to the "Programmable Pre-Emphasis" section of the Transceiver Architecture in Stratix IV Devices chapter. An optional write control to write an equalization control value for the receive side of the PMA. rx_eqctrl[3:0] (1) Input The width of this signal is fixed to 4 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 4 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13 and the "Programmable Equalization and DC Gain" section of the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-86 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 9 of 13) Port Name Input/ Output (3), (4) Description An optional equalizer DC gain write control. The width of this signal is fixed to 3 bits if you enable either the Use 'logical_channel_address' port for Analog controls reconfiguration option or the Use same control signal for all the channels option in the Analog controls screen. Otherwise, the width of this signal is 3 bits per channel. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. The following values are the legal settings allowed for this signal: rx_eqdcgain[2:0] (1), (2) Input 3'b000 => 0 dB 3'b001 => 3 dB 3'b010 => 6 dB 3'b011 => 9 dB 3'b100 => 12 dB All other values => N/A For more information, refer to the "Programmable Equalization and DC Gain" section of the Transceiver Architecture in Stratix IV Devices chapter. tx_vodctrl_out[2:0] tx_preemp_0t_out[4:0] tx_preemp_1t_out[4:0] tx_preemp_2t_out[4:0] rx_eqctrl_out[3:0] rx_eqdcgain_out[2:0] Stratix IV Device Handbook Volume 2: Transceivers Output An optional transmit VOD read control signal. This signal reads out the value written into the VOD control register. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. Output An optional pre-tap, pre-emphasis read control signal. This signal reads out the value written by its input control signal. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. Output An optional first post-tap, pre-emphasis read control signal. This signal reads out the value written by its input control signal. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. Output An optional second post-tap pre-emphasis read control signal. This signal reads out the value written by its input control signal. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. Output An optional read control signal to read the equalization setting of the ALTGX instance. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. Output An optional equalizer DC gain read control signal. This signal reads out the settings of the ALTGX instance DC gain. The width of this output signal depends on the number of channels controlled by the dynamic reconfiguration controller. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-87 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 10 of 13) Port Name Input/ Output (3), (4) Description Transceiver Channel Reconfiguration Control/Status Signals Set the following values at this signal to activate the appropriate dynamic reconfiguration mode: 3'b000 = PMA controls reconfiguration mode. This is the default value. 3'b011 = data rate division in transmitter mode 3'b100 = CMU PLL reconfiguration mode 3'b101 = channel and CMU PLL reconfiguration mode 3'b110 = channel reconfiguration with transmitter PLL select mode reconfig_mode_sel[3:0] Input 3'b111 = central control unit reconfiguration mode The reconfig_mode_sel signal is 4 bits wide when you enable Adaptive Equalization control or EyeQ control: 4'b1000 = AEQ control (continuous mode for a single channel) 4'b1001 = AEQ control (one time mode for a single channel) 4'b1010 = AEQ control (power down for a single channel) 4'b1011 = EyeQ control reconfig_mode_sel[] is available as an input only when you enable more than one dynamic reconfiguration mode. Always available for you to select in the Channel and TX PLL reconfiguration screen. This signal is applicable only in the dynamic reconfiguration modes grouped under the Channel and TX PLL select/reconfig option. reconfig_address_out[5:0] Output This signal represents the current address used by the ALTGX_RECONFIG instance when writing the .mif into the transceiver channel. This signal increments by 1, from 0 to the last address, then starts at 0 again. You can use this signal to indicate the end of all the .mif write transactions (reconfig_address_out[5:0] changes from the last address to 0 at the end of all the .mif write transactions). An optional signal you can select in the Channel and TX PLL reconfiguration screen. This signal is applicable only in dynamic reconfiguration modes grouped under the Channel and TX PLL select/reconfig option. reconfig_address_en reset_reconfig_address Output Input The dynamic reconfiguration controller asserts reconfig_address_en to indicate that reconfig_address_out[5:0] has changed. This signal is asserted only after the dynamic reconfiguration controller completes writing one 16-bit word of the .mif. An optional signal you can select in the Channel and TX PLL reconfiguration screen. This signal is applicable only in dynamic reconfiguration modes grouped under the Channel and TX PLL select/reconfig option. Enable this signal and assert it for one reconfig_clk clock cycle if you want to reset the reconfiguration address used by the ALTGX_RECONFIG instance during reconfiguration. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-88 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 11 of 13) Port Name reconfig_data[15:0] reconfig_address[5:0] (3), (4) Input/ Output Description Input Applicable only in the dynamic reconfiguration modes grouped under the Channel and TX PLL select/reconfig option. This is a 16bit word carrying the reconfiguration information. It is stored in a .mif that you must generate. The ALTGX_RECONFIG instance requires that you provide reconfig_data [15:0]on every .mif write transaction using the write_all signal. Input Available for selection only in .mif-based transceiver channel reconfiguration modes. For more information, refer to "Reduced .mif Reconfiguration" on page 5-24. Available when you select data rate division in transmitter mode. Based on the value you set here, the divide-by setting of the local divider in the transmitter channel gets modified. The legal values for this port are: rate_switch_ctrl[1:0] Input 2'b00 = Divide by 1 2'b01 = Divide by 2 2'b10 = Divide by 4 2'b11 = Not supported Available when you select data rate division in transmitter mode. You can read the existing local divider settings of a transmitter channel at this port. The decoding for this signal is listed below: rate_switch_out[1:0] Input 2'b00 = Division of 1 2'b01 = Division of 2 2'b10 = Division of 4 2'b11= Not supported logical_tx_pll_sel logical_tx_pll_sel_en channel_reconfig_done Stratix IV Device Handbook Volume 2: Transceivers Input Specify the identity of the transmitter PLL you want to reconfigure. You can also specify the identity of the transmitter PLL that you want the transceiver channel to listen to. When you enable this signal, the value set at this signal overwrites the logical_tx_pll value contained in the .mif. The value at this port must be held at a constant logic level until reconfiguration is done. Input If you want to use the logical_tx_pll_sel port only under some conditions and use the logical_tx_pll value contained in the .mif otherwise, enable this optional logical_tx_pll_sel_en port. Only when logical_tx_pll_sel_en is enabled and set to 1 does the dynamic reconfiguration controller use logical_tx_pll_sel to identify the transmitter PLL. The value at this port must be held at a constant logic level until reconfiguration is done. Output The channel_reconfig_done signal goes high to indicate that the dynamic reconfiguration controller has finished writing all the words of the .mif. The channel_reconfig_done signal is automatically de-asserted at the start of a new dynamic reconfiguration write sequence. This signal is applicable only in channel and CMU PLL reconfiguration and channel reconfiguration with transmitter PLL select modes. September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Controller Port List 5-89 Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 12 of 13) Port Name reconfig_reset Input/ Output Input (3), (4) Description An optional signal that you can use to reset the ALTGX_RECONFIG instance. reconfig_reset must be held high for at least one clock cycle to take effect. When feeding into the reconfig_reset port, the reset signal must be synchronized to the reconfig_clk domain. The width of the aeq_fromgxb[7:0] signal depends on the number of channels controlled by the ALTGX_RECONFIG instance. For example, if you select the total number of channels controlled by the ALTGX_RECONFIG instance as follows: 1 Channels 4, then the input port reconfig_fromgxb = 8 bits 5 Channels 8, then the input port reconfig_fromgxb = aeq_fromgxb[7:0] Input 16 bits 9 Channels 12, then the input port reconfig_fromgxb = 24 bits This signal is available only when you enable the AEQ control option. You must connect this signal between the ALTGX_RECONFIG and ALTGX instances when using AEQ control. The width of the aeq_togxb signal depends on the number of channels controlled by the ALTGX_RECONFIG instance. For example, if you select the total number of channels controlled by the ALTGX_RECONFIG instance as follows: 1 Channels 4, then the input port reconfig_fromgxb = 24 bits Output aeq_togxb 5 Channels 8, then the input port reconfig_fromgxb = 48 bits 9 Channels 12, then the input port reconfig_fromgxb = 64 bits This signal is available only when you enable the AEQ control option. You must connect this signal between the ALTGX_RECONFIG and ALTGX instances when using AEQ control. ctrl_address[15:0] Input Used for EyeQ control. This port is used to specify the address of the EyeQ interface register for read and write operations. ctrl_writedata[15:0] Input Used for EyeQ control. Data present on this port is written to the EyeQ interface register selected using the ctrl_address port. ctrl_readdata[15:0] Output Used for EyeQ control. Contents of the EyeQ interface register selected using the ctrl_address port are available on this port after a read operation. ctrl_write Input Used for EyeQ control. Assert this signal high to write the data present on the ctrl_writedata port to the EyeQ interface registers. ctrl_read Input Used for EyeQ control. Assert this signal high to read the contents of the EyeQ interface registers to the ctrl_readdata port. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-90 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Error Indication During Dynamic Reconfiguration Table 5-16. Dynamic Reconfiguration Controller Port List (ALTGX_RECONFIG Instance) (Part 13 of 13) Port Name ctrl_waitrequest (3), (4) Input/ Output Description Output Used for EyeQ control. If asserted, this port indicates that the EyeQ controller is busy with a read or write operation. You must wait until this signal goes low before you perform the next operation. Ensure that the values on the ctrl_read, ctrl_write, ctrl_readdata, and ctrl_writedata ports are constant when ctrl_waitrequest is asserted. Notes to Table 5-16: (1) Not all combinations of the input bits are legal values. (2) In PCIe mode, this input must be tied to 001 to be PCIe-compliant. (3) For the various dynamic reconfiguration controller input and output ports and the software settings, refer to the ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices chapter. (4) For the various transceiver input and output ports and the software settings, refer to the ALTGX Transceiver Setup Guide for Stratix IV Devices chapter. Error Indication During Dynamic Reconfiguration The ALTGX_RECONFIG MegaWizard Plug-In Manager provides an error status signal when you select the Enable illegal mode checking option or the Enable self recovery option in the Error checks/data rate switch screen. The conditions under which the error signal is asserted are: Stratix IV Device Handbook Volume 2: Transceivers Enable illegal mode checking option--When you select this option, the dynamic reconfiguration controller checks whether an attempted operation falls under one of the conditions listed below. The dynamic reconfiguration controller detects these conditions within two reconfig_clk cycles, de-asserts the busy signal, and asserts the error signal for two reconfig_clk cycles. PMA controls, read operation--None of the output ports (rx_eqctrl_out, rx_eqdcgain_out, tx_vodctrl_out, tx_preemp_0t_out, tx_preemp_1t_out, and tx_preemp_2t_out) are selected in the ALTGX_RECONFIG instance and the read signal is asserted. PMA controls, write operation--None of the input ports (rx_eqctrl, rx_eqdcgain, tx_vodctrl, tx_preemp_0t, tx_preemp_1t, and tx_preemp_2t) are selected in the ALTGX_RECONFIG instance and the write_all signal is asserted. TX Data Rate Switch using Local Divider-read operation option--The read transaction is valid only for data rate division in transmitter mode TX Data Rate Switch using Local Divider-write operation with unsupported value option: The rate_switch_ctrl input port is set to 11 The reconfig_mode_sel input port is set to 3 (if other reconfiguration mode options are selected in the Reconfiguration settings screen) The write_all signal is asserted September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Duration 5-91 TX Data Rate Switch using Local Divider-write operation without input port option: The rate_switch_ctrl input port is not used The reconfig_mode_sel port is set to 3 (if other reconfiguration mode options are selected in the Reconfiguration settings screen) The write_all signal is asserted TX Data Rate Switch using Local Divider- read operation without output port option: The rate_switch_out output port is not used The reconfig_mode_sel port is set to 3 (if other reconfiguration mode options are selected in the Reconfiguration settings screen) The read signal is asserted Channel and/or TX PLL reconfig/select-read operation option: The reconfig_mode_sel input port is set to 4, 5, 6, or 7 The read signal is asserted Adaptive Equalization option--read operation: The reconfig_mode_sel input port is set to 7, 8, 9, or 10 The read signal is asserted EyeQ option--read operation: The reconfig_mode_sel input port is set to 11 The read signal is asserted Enable self recovery option--When you select this option, the controller automatically recovers if the operation did not complete within the expected time. The error signal is driven high whenever the controller performs a self recovery. Dynamic Reconfiguration Duration Dynamic reconfiguration duration is the number of cycles the busy signal is asserted when the dynamic reconfiguration controller performs write transactions, read transactions, or offset cancellation of the receiver channels. PMA Controls Reconfiguration Duration The following section contains an estimate of the number of reconfig_clk clock cycles the busy signal is asserted during PMA controls reconfiguration using Method 1, Method 2, or Method 3. For more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-92 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Duration PMA Controls Reconfiguration Duration When Using Method 1 The logical_channel_address port is used in Method 1. The write transaction and read transaction duration is as follows: Write Transaction Duration For writing values to the following PMA controls, the busy signal is asserted for 260 reconfig_clk clock cycles for each of these controls: tx_preemp_1t (pre-emphasis control first post-tap) tx_vodctrl (voltage output differential) rx_eqctrl (equalizer control) rx_eqdcgain (equalizer DC gain) For writing values to the following PMA controls, the busy signal is asserted for 520 reconfig_clk clock cycles for each of these controls: tx_preemp_0t (pre-emphasis control pre-tap) tx_preemp_2t (pre-emphasis control second post-tap) Read Transaction Duration For reading the existing values of the following PMA controls, the busy signal is asserted for 130 reconfig_clk clock cycles for each of these controls. The data_valid signal is then asserted after the busy signal goes low. tx_preemp_1t_out (pre-emphasis control first post-tap) tx_vodctrl_out (voltage output differential) rx_eqctrl_out (equalizer control) rx_eqdcgain_out (equalizer DC gain) For reading the existing values of the following PMA controls, the busy signal is asserted for 260 reconfig_clk clock cycles for each of these controls. The data_valid signal is then asserted after the busy signal goes low. tx_preemp_0t_out (pre-emphasis control pre-tap) tx_preemp_2t_out (pre-emphasis control second post-tap) PMA Controls Reconfiguration Duration When Using Method 2 or Method 3 The logical_channel_address port is not used in Method 2 and Method 3. The write transaction duration and read transaction duration are as follows: Write Transaction Duration For writing values to the following PMA controls, the busy signal is asserted for 260 reconfig_clk clock cycles per channel for each of these controls: Stratix IV Device Handbook Volume 2: Transceivers tx_preemp_1t (pre-emphasis control first post-tap) tx_vodctrl (voltage output differential) rx_eqctrl (equalizer control) rx_eqdcgain (equalizer DC gain) September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Duration 5-93 For writing values to the following PMA controls, the busy signal is asserted for 520 reconfig_clk clock cycles per channel for each of these controls: tx_preemp_0t (pre-emphasis control pre-tap) tx_preemp_2t (pre-emphasis control second post-tap) Read Transaction Duration For reading the existing values of the following PMA controls, the busy signal is asserted for 130 reconfig_clk clock cycles per channel for each of these controls. The data_valid signal is then asserted after the busy signal goes low. tx_preemp_1t_out (pre-emphasis control first post-tap) tx_vodctrl_out (voltage output differential) rx_eqctrl_out (equalizer control) rx_eqdcgain_out (equalizer DC gain) For reading the existing values of the following PMA controls, the busy signal is asserted for 260 reconfig_clk clock cycles per channel for each of these controls. The data_valid signal is then asserted after the busy signal goes low. tx_preemp_0t_out (pre-emphasis control pre-tap) tx_preemp_2t_out (pre-emphasis control second post-tap) Offset Cancellation Duration When the device powers up, the busy signal remains low for the first reconfig_clk clock cycle. Offset cancellation control is only for the receiver channels. The ALTGX_RECONFIG instance takes 18500 reconfig_clk clock cycles per channel for Receiver only and Receiver and Transmitter channels. It takes 900 reconfig_clk clock cycles per channel for Transmitter only channels to determine if the channel under reconfiguration is a receiver channel or not. The ATLGX_RECONFIG requires an additional 130,000 clock cycles for these values to take effect. The ALTGX_RECONFIG instance takes approximately two reconfig_clk clock cycles per channel for the unused logical channels. To demonstrate offset cancellation duration, consider the following example: One ALTGX_RECONFIG instance is connected to two ALTGX instances. ALTGX Instance 1 has one Transmitter only channel (logical_channel_address = 0) ALTGX Instance 2 has one Receiver only channel (logical_channel_address = 4) For this example, the ALTGX_RECONFIG instance consumes the following number of reconfig_clk clock cycles for offset cancellation: 900 cycles for the Transmitter only channel 18,500 cycles for the Receiver only channel 2 cycles each for non-existent channels with logical_channel_addresses = 1, 2, and 3 and 130,000 cycles as a baseline for the values to take affect. The offset cancellation duration for the ALTGX_RECONFIG instance to reconfigure the Transmitter only channel, Receiver only channel, non-existent logical channels 1, 2, and 3 = 149,406 cycles (900 +18,500 +6 + 130,000). September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-94 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration (ALTGX_RECONFIG Instance) Resource Utilization Dynamic Reconfiguration Duration for Channel and Transmitter PLL Select/Reconfig Modes Table 5-17 lists the number of reconfig_clk clock cycles it takes for the dynamic reconfiguration controller to reconfigure various parts of the transceiver channel and CMU PLL. Table 5-17. Dynamic Reconfiguration Duration for Transceiver Channel and CMU PLL Reconfiguration Transceiver Portion Under Reconfiguration Number of reconfig_clk Clock Cycles Transmitter channel reconfiguration 1,518 Receiver channel reconfiguration 5,255 Transmitter and receiver channel reconfiguration 6,762 CMU PLL only reconfiguration 863 Transmitter channel and CMU PLL reconfiguration 2,370 Transceiver channel and CMU PLL reconfiguration 7,614 Central control unit reconfiguration 925 Dynamic Reconfiguration (ALTGX_RECONFIG Instance) Resource Utilization You can observe the resources used during dynamic reconfiguration in the ALTGX_RECONFIG MegaWizard Plug-In Manager. This section contains an estimate of the LE resources used during dynamic reconfiguration. You can obtain resource utilization for all other PMA controls from the ALTGX_RECONFIG MegaWizard Plug-In Manager. For example, the number of LEs used by one dynamic reconfiguration controller is 43 with only tx_vodctrl selected and the number of registers is 130. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Functional Simulation of the Dynamic Reconfiguration Process 5-95 Figure 5-40 shows resource utilization in the ALTGX_RECONFIG MegaWizard Plug-In Manager. Figure 5-40. Resource Utilization in the ALTGX_RECONFIG MegaWizard Plug-In Manager Functional Simulation of the Dynamic Reconfiguration Process This section describes the points to be considered during functional simulation of the dynamic reconfiguration process. September 2012 You must connect the ALTGX_RECONFIG instance to the ALTGX_instance/ALTGX instances in your design for functional simulation. The functional simulation uses a reduced timing model for offset cancellation. Therefore, the duration of the offset cancellation process is 16 reconfig_clk clock cycles for functional simulation only. The gxb_powerdown signal must not be asserted during the offset cancellation sequence (for functional simulation and silicon). Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-96 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples Dynamic Reconfiguration Examples The following examples help to describe the dynamic reconfiguration feature. Example 1 Consider a design with the following configuration: Seven regular transceiver channels in Basic functional mode. You can configure the seven regular transceiver channels from 2.5 Gbps to 5 Gbps and vice versa using a single CMU. Four channels in Basic (PMA Direct) functional mode. You can reconfigure the four PMA-only channels from 3.125 Gbps to 5 Gbps and vice versa. You can reconfigure the PMA controls for any one of these channels. Figure 5-41 shows the arrangement of these channels in the S4GX230 device. Figure 5-41. Dynamic Reconfiguration Configuration for the S4GX230 Device (Example 1) Transceiver Block GXBR2 Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 Transceiver Block GXBR1 Seven regular transceiver channels use a single CMU0 PLL (GXBR2) in the design Channel 3 Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 ATX R0 PLL Block Transceiver Block GXBR0 Channel 3 Four PMA-Only channels use a single CMU0 PLL (GXBR0) in the design Channel 2 CMU1 Channel CMU0 Channel Channel 1 Channel 0 Regular Transceiver Channels PMA-Only Channels Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples 5-97 Because this example does not require the use of the alternate CMU transmitter PLL or additional transmitter PLLs, the logical channel addressing remains the same as explained in "Logical Channel Addressing" on page 5-5. Table 5-18 lists how to set the starting channel number in the ALTGX MegaWizard Plug-In Manager, the total number of channels in the ALTGX_RECONFIG MegaWizard Plug-In Manager, and how to connect the ALTGX instances to the ALTGX_RECONFIG instance. Table 5-18. Logical Channel Addressing Combination of Regular Transceiver Channels and PMA-Only Channels (Example 1) ALTGX Settings and Instances ALTGX Instance 1 (Basic Functional Mode) ALTGX Setting What is the number of channels? option in the General screen What is the starting channel number? option in the Reconfig screen reconfig_ fromgxb1 and reconfig_ fromgxb2 outputs September 2012 ALTGX_RECONFIG Setting and Instance ALTGX Instance 2 (Basic [PMA Direct] Functional Mode) 7 (Regular Transceiver Channels) Set this option to 0. The logical channel addresses of the first to sixth channels are 0, 1, 2, 3, 4, 5, and 6, respectively. 4 (PMA-only Channels) Set this option to 8. This is because the starting channel numbers 0 and 4 have already been used in ALTGX instance 1. The logical channel addresses of the first to fourth channels are 8, 12, 16, and 20, respectively. reconfig_ fromgxb1 is 34 bits wide reconfig_ fromgxb2 is 68 bits wide (2 * 17) (4 * 17) Altera Corporation ALTGX_RECONFIG Setting What is the number of channels controlled by the reconfig controller? option in the Reconfiguration settings screen. -- reconfig_ fromgxb input ALTGX_RECONFIG Instance 1 Determine the highest logical channel address (20). Round it up to the next multiple of 4. Set this option to 24. -- reconfig_ fromgxb is 102 bits wide (24 regular transceiver channels can logically fit into 6 transceiver blocks; 6 * 17 = 102) Stratix IV Device Handbook Volume 2: Transceivers 5-98 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples Figure 5-42 shows how the logical channel addresses of all the channels are set based on what you set as the starting channel number. Figure 5-42. Logical Channel Addresses for Example 1 ALTGX instance 1 Basic Functional Mode Starting channel number = 0 Set the What is the number of channels controlled by the reconfig controller? option = 24 (1) Channel 0 (logical channel address = 0) reconfig_fromgxb1[33:0] Channel 1 (logical channel address = 1) reconfig_togxb[3:0] Channel 2 (logical channel address = 2) reconfig_fromgxb[101:0] (2) Channel 3 (logical channel address = 3) ALTGX_RECONFIG instance 1 Channel 4 (logical channel address = 4) Channel 5 (logical channel address = 5) Channel 6 (logical channel address = 6) ALTGX instance 2 Basic (PMA Direct) Configuration Starting channel number = 8 Channel 0 (logical channel address = 8) Channel 1 (logical channel address = 12) reconfig_fromgxb2[67:0] Channel 2 (logical channel address = 16) Channel 3 (logical channel address = 20) Notes to Figure 5-42: (1) For more information, refer to "Total Number of Channels Option in the ALTGX_RECONFIG Instance" on page 5-10. (2) reconfig_fromgxb[101:0] = { reconfig_fromgxb2[67:0], reconfig_fromgxb1[33:0]} Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples 5-99 Different Dynamic Reconfiguration Modes Involved 1. Channel and CMU PLL reconfiguration mode: is used for reconfiguring the seven regular transceiver channels from one data rate to another using the same CMU0 PLL (in GXBR2) 1 This mode is chosen because both the receiver and transmitter of the regular channels must be re-configured using a single CMU. 2. Channel and CMU PLL select reconfiguration mode: is used for reconfiguring the four PMA-only channels from one data rate to another using the CMU0 PLL (in GXBR0) and CMU1 PLL (GXBR0) 1 This mode is chosen because both the receiver and transmitter of the regular channels must be re-configured and more than one CMU can be used. 3. The rx_tx_duplex_sel[1:0] port allows you to reconfigure the transmitter and receiver channels to operate at the different data rates. 4. PMA controls reconfiguration mode used to configure the PMA settings for all the channels. For more information, refer to "Transceiver Channel Reconfiguration Mode Details" on page 5-19. .mif Generation The following .mifs are required for this example: For the seven regular transceiver channels, you must generate two .mifs. Use one to move from a data rate of 2.5 Gbps to 5 Gbps and the other to revert back to 2.5 Gbps. For the for PMA-only channels, you must generate two .mifs. Use one to move from a data rate of 3.125 Gbps to 5 Gbps and the other to revert back to 3.125 Gbps. For more information, refer to "Memory Initialization File (.mif)" on page 5-20. Various Dynamic Reconfiguration Transactions The following dynamic reconfiguration transactions are required "Example 1" on page 5-96: September 2012 .mif write transaction--for more information, refer to "Channel and CMU PLL Reconfiguration Mode Details" on page 5-24 and "Channel Reconfiguration with Transmitter PLL Select Mode Details" on page 5-48. Reconfiguring PMA controls--for more information, refer to "Dynamically Reconfiguring PMA Controls" on page 5-13. Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-100 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples Example 2 Consider a design with the following configuration: Four regular transceiver channels in XAUI configuration. You can configure these channels from the XAUI configuration (the primary configuration) to the PCIe Gen2 x4 configuration (the secondary configuration) and vice versa. Figure 5-43 shows the arrangement of these channels in the S4GX230 device. Figure 5-43. Dynamic Reconfiguration Configuration for the S4GX230 Device (Example 2) Transceiver Block GXBR0 Channel 3 Channel 2 CMU1 Channel Four regular channels using the CMU0 PLL in the design 1 100 MHz for PIPE Gen2 x4 mode 0 156.25 MHz for XAUI CMU0 Channel Channel 1 Channel 0 Because this example does not require the use of the alternate CMU transmitter PLL or additional transmitter PLLs, the logical channel addressing remains the same as explained in "Logical Channel Addressing" on page 5-5. Table 5-19 lists how to set the starting channel number in the ALTGX MegaWizard Plug-In Manager, the total number of channels in the ALTGX_RECONFIG MegaWizard Plug-In Manager, and how to connect the ALTGX instances to the ALTGX_RECONFIG instance. Table 5-19. Logical Channel Addressing Combination x4 Bonded Channels (Example 2) (Part 1 of 2) ALTGX Settings and Instances ALTGX Instance 1 (XAUI Mode) ALTGX Setting 4 What is the number of channels? option in the General screen What is the starting channel number? option in the Reconfig screen Stratix IV Device Handbook Volume 2: Transceivers (Regular transceiver channels) Set this option to 0. The logical channel addresses of the first to sixth channels are 0, 1, 2, and 3, respectively. ALTGX_RECONFIG Setting and Instance ALTGX_RECONFIG Instance 1 ALTGX_RECONFIG Setting What is the number of channels controlled by the reconfig controller? option in the Reconfiguration settings screen. -- Determine the highest logical channel address (3). Round it up to the next multiple of 4. Set this option to 4. -- September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples 5-101 Table 5-19. Logical Channel Addressing Combination x4 Bonded Channels (Example 2) (Part 2 of 2) ALTGX Settings and Instances ALTGX_RECONFIG Setting and Instance ALTGX Instance 1 (XAUI Mode) ALTGX Setting Settings in Reconfiguration settings page reconfig_fromgxb1 and reconfig_fromgxb2 outputs Enable Channel and Transmitter PLL reconfiguration Enable Channel interface Set 2 for the How many clock inputs are used? option. Set 0 for the XAUI ALTGX instance. Set 1 for the PCIe s x4 ALTGX instance for the What is the selected input clock source for Tx/Rx PLLs? option. ALTGX_RECONFIG Setting ALTGX_RECONFIG Instance 1 -- -- reconfig_fromgxb is 17 bits wide reconfig_fromgxb1 is 17 bits wide reconfig_fromgxb input (4 regular transceiver channels can logically fit into 1 transceiver blocks; 1 * 17 = 17) Figure 5-44 shows how the logical channel addresses of all the channels are set based on what you set as the starting channel number. Figure 5-44. Logical Channel Addresses for Example 2 Set the What is the number of channels controlled by the reconfig controller? option = 4 ALTGX Instance 1 Four regular transceiver channels XAUI/PCI Express (PIPE) Gen2 x4 Functional Mode Starting channel number = 0 Channel 0 (logical channel address = 0) reconfig_fromgxb[16:0] * reconfig_fromgxb[16:0] Channel 1 (logical channel address = 1) reconfig_togxb[3:0] ALTGX_RECONFIG Instance Channel 2 (logical channel address = 2) Channel 3 (logical channel address = 3) September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers 5-102 Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples Different Dynamic Reconfiguration Modes Involved 1. Channel and CMU PLL reconfiguration mode--used for reconfiguring four regular transceiver channels and the CMU0 PLL (in GXBR0) from XAUI mode to PCIe x4 mode and vice versa. 1 Use this mode instead of channel reconfiguration with transmitter PLL select mode because the central clock divider used for bonded modes is only available in CMU0; therefore, you cannot use the CMU1 PLL as an alternate TX PLL. 2. Central control unit reconfiguration mode--used for reconfiguring central control unit logic used in bonded modes from XAUI mode to PCIe x4 mode. For more information, refer to "Transceiver Channel Reconfiguration Mode Details" on page 5-19. .mif Generation The following .mifs are required for this example: One .mif is required to move from XAUI mode to PCIe x4 mode Another .mif is required to revert back to XAUI mode from PCIe x4 mode For more information, refer to "Memory Initialization File (.mif)" on page 5-20. Various Dynamic Reconfiguration Transactions The following dynamic reconfiguration transactions are required for this example: .mif write transaction--for more information, refer to "Channel and CMU PLL Reconfiguration Mode Details" on page 5-24. Alternatively, you may use reduced .mif reconfiguration. Reduced .mifs are generated using the altgx_diffmifgen.exe command. For more information, refer to "Reduced .mif Reconfiguration" on page 5-24. Document Revision History Table 5-20 lists the revision history for this chapter. Table 5-20. Document Revision History (Part 1 of 2) Date Version September 2012 December 2011 Stratix IV Device Handbook Volume 2: Transceivers 3.4 Changes Updated Table 5-13 title. Updated Figure 5-6 to close FB #29348. Updated Figure 5-37. Updated the "Logical Channel Addressing of PMA-Only Channels" and "Transceiver Channel Reconfiguration Mode Details" sections. Updated Table 5-6, Table 5-9, and Table 5-16. 3.3 September 2012 Altera Corporation Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples 5-103 Table 5-20. Document Revision History (Part 2 of 2) Date Version February 2011 Changes Updated Table 5-5, Table 5-6, and Table 5-16. Updated Figure 5-1. Updated the "Transceiver Channel Reconfiguration Mode Details". "PMA Controls Reconfiguration Mode Details", "Connecting the ALTGX and ALTGX_RECONFIG Instances", "One Time Mode for a Single Channel", "Applying a .mif in the User Design", and "Functional Simulation of the Dynamic Reconfiguration Process" sections. Removed the "Continuous Mode for a Single Channel" and "Powerdown for a Single Channel" sections. Updated chapter title. Applied new template. Minor text edits. Updated Table 5-5, Table 5-6, Table 5-15, Table 5-16, and Table 5-17. Updated Figure 5-1, Figure 5-14, Figure 5-16, Figure 5-26, and Figure 5-37. Updated the "Blocks Reconfigured in the Data Rate Division in Transmitter Mode", "Logical Channel Addressing of PMA-Only Channels", "Central Control Unit Reconfiguration Mode Details", "EyeQ", "Error Indication During Dynamic Reconfiguration", and "Functional Simulation of the Dynamic Reconfiguration Process" sections. Added a note to the "Central Control Unit Reconfiguration Mode Details" section. Minor text edits. Completely re-wrote and re-organized chapter. Updated all graphics and tables. Updated Figure 5-4, Figure 5-8, Figure 5-9, Figure 5-10, Figure 5-11, Figure 5-15, Figure 5-22, Table 5-37, Table 5-38, Figure 5-44, Figure 5-47, Figure 5-48, Figure 5-49, Figure 5-50, Figure 5-51, Figure 5-52, Figure 5-53, and Figure 5-54 Updated Table 5-2 and Table 5-31 Changed "logical_tx_pll_sel[1:0]" to "logical_tx_pll_sel" throughout Updated "The reconfig_clk Clock Requirements for the ALTGX Instance and ALTGX_RECONFIG Instance", "The logical_tx_pll_sel and logical_tx_pll_sel_en Ports", "How to Use the logical_tx_pll_sel Port?", and "When Can the logical_tx_pll_sel and logical_tx_pll_sel_en Ports be Used?" Minor text edits 3.2 March 2010 3.1 November 2009 June 2009 3.0 2.1 Complete re-write and re-organization of the chapter. Added or revised: March 2009 2.0 November 2008 September 2012 1.0 Altera Corporation Offset Cancellation Control for Receiver Channels PMA Controls Reconfiguration Channel and CMU PLL Reconfiguration Mode Data Rate Division in Transmitter: Operation Channel Reconfiguration with Transmitter PLL Select Mode CMU PLL Reconfiguration Mode Initial release. Stratix IV Device Handbook Volume 2: Transceivers 5-104 Stratix IV Device Handbook Volume 2: Transceivers Chapter 5: Dynamic Reconfiguration in Stratix IV Devices Dynamic Reconfiguration Examples September 2012 Altera Corporation Additional Information This chapter provides additional information about the document and Altera. How to Contact Altera To locate the most up-to-date information about Altera products, refer to the following table. Contact (1) Technical support Technical training Product literature Contact Method Address Website www.altera.com/support Website www.altera.com/training Email custrain@altera.com Website www.altera.com/literature Nontechnical support (general) Email nacomp@altera.com (software licensing) Email authorization@altera.com Note to Table: (1) You can also contact your local Altera sales office or sales representative. Typographic Conventions The following table shows the typographic conventions this document uses. Visual Cue Meaning Bold Type with Initial Capital Letters Indicate command names, dialog box titles, dialog box options, and other GUI labels. For example, Save As dialog box. For GUI elements, capitalization matches the GUI. bold type Indicates directory names, project names, disk drive names, file names, file name extensions, software utility names, and GUI labels. For example, \qdesigns directory, D: drive, and chiptrip.gdf file. Italic Type with Initial Capital Letters Indicate document titles. For example, Stratix IV Design Guidelines. Indicates variables. For example, n + 1. italic type Variable names are enclosed in angle brackets (< >). For example, <file name> and <project name>.pof file. Initial Capital Letters Indicate keyboard keys and menu names. For example, the Delete key and the Options menu. "Subheading Title" Quotation marks indicate references to sections in a document and titles of Quartus II Help topics. For example, "Typographic Conventions." September 2012 Altera Corporation Stratix IV Device Handbook Volume 2: Transceivers Info-2 Additional Information Typographic Conventions Visual Cue Meaning Indicates signal, port, register, bit, block, and primitive names. For example, data1, tdi, and input. The suffix n denotes an active-low signal. For example, resetn. Courier type Indicates command line commands and anything that must be typed exactly as it appears. For example, c:\qdesigns\tutorial\chiptrip.gdf. Also indicates sections of an actual file, such as a Report File, references to parts of files (for example, the AHDL keyword SUBDESIGN), and logic function names (for example, TRI). r An angled arrow instructs you to press the Enter key. 1., 2., 3., and a., b., c., and so on Numbered steps indicate a list of items when the sequence of the items is important, such as the steps listed in a procedure. Bullets indicate a list of items when the sequence of the items is not important. 1 The hand points to information that requires special attention. h The question mark directs you to a software help system with related information. f The feet direct you to another document or website with related information. m The multimedia icon directs you to a related multimedia presentation. c A caution calls attention to a condition or possible situation that can damage or destroy the product or your work. w A warning calls attention to a condition or possible situation that can cause you injury. The envelope links to the Email Subscription Management Center page of the Altera website, where you can sign up to receive update notifications for Altera documents. Stratix IV Device Handbook Volume 2: Transceivers September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 101 Innovation Drive San Jose, CA 95134 www.altera.com SIV5V3-4.3 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as ISO trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and 9001:2008 services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service Registered described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 Contents Chapter Revision Dates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Section I. Transceiver Configuration Guide Chapter 1. ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 General Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 PLL/Ports Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16 Ports/Calibration Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-19 Loopback Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22 RX Analog Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-24 TX Analog Screen for the Parameter Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-26 Reconfiguration Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-28 Modes Screen for the Reconfiguration Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-29 Transmitter PLL Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-31 Clocking/Interface Screen for the Reconfiguration Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-35 Protocol Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-36 8B10B Screen for the Protocol Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-37 Word Aligner Screen for the Protocol Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-40 Rate Match/Byte Order Screen for the Protocol Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-45 Protocol Settings Screen for GIGE and XAUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-48 Protocol Settings Screen for the (OIF) CEI Phy Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-51 Protocol Settings Screen for PCIe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-52 Protocol Settings Screen for SONET/SDH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-57 EDA Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-60 Summary Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-61 Chapter 2. Transceiver Design Flow Guide for Stratix IV Devices Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 Device Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 Transceiver Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 Dynamic Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4 Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5 Power Supplies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6 Board Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6 Implementation and Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6 Create Transceiver Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 Reset and Status Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 Create Dynamic Reconfiguration Controller Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 Create Reset and Control Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 Create Data Processing and Other User Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 PPM Detector When the Receiver CDR Is Used in Manual Lock Mode . . . . . . . . . . . . . . . . . . . . . 2-8 Synchronization State Machine in Manual Word Alignment Mode . . . . . . . . . . . . . . . . . . . . . . . . 2-8 Gear Boxing Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 Integrate the Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9 Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10 Report Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Fitter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 iv Contents Pin-Out File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Resource Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Functional Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12 Guidelines to Debug Transceiver-Based Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14 Guidelines to Debug the FPGA Logic and the Transceiver Interface . . . . . . . . . . . . . . . . . . . . . . . . . 2-14 Guidelines to Debug System Level Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15 Example 1: Fibre Channel Protocol Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17 Phase 1--Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17 Device Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17 Transceiver Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18 Dynamic Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18 Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19 Phase 2--Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-20 Create the Transceiver Instance for an FC4G Configuration (Channel 0) . . . . . . . . . . . . . . . . . . . 2-21 Create the Transceiver Instance for an FC1G Configuration (Channel 1) . . . . . . . . . . . . . . . . . . . 2-30 Create the Instance for an FC4G Configuration--Transmitter Only Mode (Channel 2) . . . . . . 2-31 Create the Dynamic Reconfiguration Controller (ALTGX_Reconfig) Instance . . . . . . . . . . . . . . 2-33 Create Reset Logic to Control the FPGA Fabric and Transceivers . . . . . . . . . . . . . . . . . . . . . . . . . 2-34 Create Data Processing and Other User Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36 Phase 3--Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36 Phase 4--Simulating the Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36 Chapter 3. ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 Additional Information How to Contact Altera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 Typographic Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Info-1 Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter Revision Dates The chapters in this document, Stratix IV Device Handbook Volume 3, were revised on the following dates. Where chapters or groups of chapters are available separately, part numbers are listed. Chapter 1. ALTGX Transceiver Setup Guide for Stratix IV Devices Revised: September 2012 Part Number: SIV53001-4.3 Chapter 2. Transceiver Design Flow Guide for Stratix IV Devices Revised: February 2011 Part Number: SIV53002-4.1 Chapter 3. ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Revised: February 2011 Part Number: SIV53004-3.1 September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 vi Stratix IV Device Handbook Volume 3 Chapter Revision Dates September 2012 Altera Corporation Section I. Transceiver Configuration Guide This section includes the following chapters: Chapter 1, ALTGX Transceiver Setup Guide for Stratix IV Devices Chapter 2, Transceiver Design Flow Guide for Stratix IV Devices Chapter 3, ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 I-2 Stratix IV Device Handbook Volume 3 Section I: Transceiver Configuration Guide September 2012 Altera Corporation 1. ALTGX Transceiver Setup Guide for Stratix IV Devices September 2012 SIV53001-4.3 SIV53001-4.3 This chapter describes the options you can choose in the ALTGX MegaWizardTM Plug-In Manager in the Quartus(R) II software to configure Stratix(R) IV GX and GT devices in different functional modes. The MegaWizard Plug-In Manager in the Quartus II software creates or modifies design files that contain custom megafunction variations that can then be instantiated in a design file. You can use the MegaWizard Plug-In Manager to set the ALTGX megafunction features in the design. The ALTGX megafunction allows you to configure one or more transceiver channels. You can select the physical coding sublayer (PCS) and physical medium attachment (PMA) functional blocks depending on your transceiver configuration. This chapter contains the following sections: "Parameter Settings" on page 1-4 "Reconfiguration Settings" on page 1-28 "Protocol Settings" on page 1-36 Start the MegaWizard Plug-In Manager using one of the following methods: From the Tools menu, select MegaWizard Plug-In Manager. When working in the Block Editor, click MegaWizard Plug-In Manager in the Symbol dialog box (Edit menu). Start the standalone version of the MegaWizard Plug-In Manager by typing the following command at the command prompt: qmegawiz. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 3 September 2012 Feedback Subscribe 1-2 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Figure 1-1 shows the first page of the MegaWizard Plug-In Manager. To generate an ALTGX custom megafunction variation, select Create a new custom megafunction variation. Figure 1-1. MegaWizard Plug-In Manager (Page 1) Figure 1-2 shows the second page of the MegaWizard Plug-In Manager. To use the MegaWizard Plug-In Manager to configure a Stratix IV device, follow these steps: 1. Select Stratix IV as the device family. 2. Select either VHDL or Verilog HDL depending on the type of output files you want to create. 3. Select the ALTGX megafunction under the I/O section of the available megafunctions. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices 1-3 4. Name the output file, then Browse to the folder you want to save your file in and click Next. The General screen of the ALTGX MegaWizard Plug-In Manager opens (Figure 1-3). Figure 1-2. MegaWizard Plug-In Manager (Page 2) September 2012 1 All reset and control signals are active high unless otherwise mentioned. 1 All output ports are synchronous to the data path unless otherwise specified. 1 Throughout this chapter, the various functional modes and their settings are explained for Stratix IV GX and GT devices. Altera Corporation Stratix IV Device Handbook Volume 3 1-4 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Parameter Settings This section describes the options available on the individual pages of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. The MegaWizard Plug-In Manager provides a warning if any of the settings you choose are illegal. General Screen for the Parameter Settings Figure 1-3 shows the General screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-3. MegaWizard Plug-In Manager--ALTGX (General Screen for the Parameter Settings) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-5 Table 1-1 lists the available functional modes and their options on the General screen of the MegaWizard Plug-In Manager. Depending on your configuration, you will select one of the following functional modes: Basic Basic (PMA Direct) Deterministic Latency GIGE (OIF) CEI Phy Interface PCI Express(R) (PCIe) SDI Serial RapidIO(R) SONET/SDH XAUI If you select Basic (PMA Direct) mode, all the channels are configured with only the PMA blocks. These channels are called PMA-only channels throughout this chapter. The PMA-only channels include: Regular transceiver channels with PMA blocks only CMU channels (clock multiplier unit phase-locked loops [CMU PLLs] configured as additional transceiver channels with PMA blocks only) Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 1 of 11) ALTGX Setting Description Select GX or GT based on the Stratix IV device used in your design. Which device variation will you be using? Select the speed grade of your device. The available speed grades for the Stratix IV GX device are 2, 2x, 3, and 4. The available speed grades for the Stratix IV GT device are 1, 2 and 3. Reference DC and Switching Characteristics for Stratix IV Devices chapter. Determines the specific protocol under which the transceiver operates. For a specific mode, you must select the desired protocol from the following list: Which protocol will you be using? September 2012 Altera Corporation (OIF) CEI PHY Interface SDI SONET/SDH XAUI Basic Basic (PMA Direct) Deterministic Latency GIGE PCIe Serial RapidIO Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-6 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 2 of 11) ALTGX Setting Description Reference Basic In Basic mode, the subprotocols are diagnostic modes. The available options are as follows: None--This is the normal operation of the transceiver. x4--In this mode, all four channels within the transceiver block are clocked from its central clock divider block to minimize transmitter channel-to-channel skew. x8--In this mode, all eight channels in two transceiver blocks are clocked from the central clock divider of the master transceiver block to minimize transmitter channel-to-channel skew. BIST--This subprotocol is applicable only for Receiver and Transmitter operation mode. This mode loops the parallel data from the built-in self test (BIST) (non-PRBS) back to the BIST verifier in the receiver path. Parallel loopback is allowed only in Basic double-width mode. PRBS--This subprotocol is applicable only for Receiver and Transmitter operation mode.This is another Serial Loopback mode but with the pseudo-random binary sequence (PRBS) BIST block active. The PRBS pattern depends on the serializer/deserializer (SERDES) factor. Which subprotocol will you be using? "Basic Functional Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Basic (PMA Direct) None--This is the normal mode of operation in which each channel is treated independently. XN--In this mode, the "N" in XN represents the number of channels in the bonded configuration. All N channels are clocked by the same transmit clock from the central clock divider block to minimize transmitter channel-to-channel skew. "Basic PMA Direct Functional Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Deterministic Latency Stratix IV Device Handbook Volume 3 x1--In this mode, you can have up to two configured channels per transceiver block. Each channel uses one CMU PLL and its feedback path to compensate for the uncertain latency. x4--In this mode, you can have up to four configured channels per transceiver block. All channels use one CMU PLL per block and its feedback path to compensate for the uncertain latency. "Deterministic Latency Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-7 Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 3 of 11) ALTGX Setting Description Reference PCIe In PCIe mode, there are six subprotocols: Which subprotocol will you be using? Gen1 x1--The transceiver is configured as a single-lane PCIe link for a 2.5 Gbps data rate. Gen1 x4--The transceiver is configured as a four-lane PCIe link for a data rate of 2.5 Gbps. Gen1 x8--The transceiver is configured as an eight-lane PCIe link for a data rate of 2.5 Gbps. Gen2 x1--The transceiver is configured as a single-lane PCIe link for a 5.0 Gbps data rate. Gen2 x4--The transceiver is configured as a four-lane PCIe link for a data rate of 5.0 Gbps. Gen2 x8--The transceiver is configured as an eight-lane PCIe link for a data rate of 5.0 Gbps. "PCIe Mode" in the Transceiver Architecture in Stratix IV Devices chapter. SDI In SDI mode, the two available subprotocols are: 3G--third-generation (3 Gbps) SDI at 2967 Mbps or 2970 Mbps. HD--high-definition SDI at 1483.5 Mbps or 1485 Mbps. "SDI Mode" in the Transceiver Architecture in Stratix IV Devices chapter. SONET/SDH In SONET/SDH mode, the three available subprotocols and their data rates are: OC-12--622 Mbps OC-48--2488.32 Mbps OC-96--4976.64 Mbps "SONET/SDH Mode" in the Transceiver Architecture in Stratix IV Devices chapter. Deterministic Latency GIGE (OIF) CEI PHY Interface Enforce default settings for this protocol. PCIe -- SONET/SDH XAUI If you select this option, all mode-specific ports and settings are used. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-8 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 4 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) Deterministic Latency SDI Serial RapidIO SONET/SDH -- The available operation modes are Receiver only, Transmitter only, and Receiver and Transmitter. What is the operation mode? For Basic (PMA Direct) xN mode, the available operation modes are Transmitter only and Receiver and Transmitter. However, if you set the subprotocol to None, the available operation modes are Receiver only, Transmitter only, and Receiver and Transmitter. GIGE The available operation modes are Transmitter only, and Receiver and Transmitter. -- PCIe XAUI -- Only Receiver and Transmitter mode is allowed. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-9 Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 5 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) Deterministic Latency SDI -- Serial RapidIO The number of channels required with the same configuration. This option determines how many identical channels this ALTGX instance contains. GIGE (OIF) CEI PHY Interface SONET/SDH What is the number of channels? -- This option allows you to select how many channels this ALTGX instance contains. In these modes, the number of channels increments by one. PCIe This is the number of channels required with the same configuration. In a x4 subprotocol, the number of channels increments by 4. In a x8 subprotocol, the number of channels increment by 8. -- XAUI This option allows you to select how many identical channels this ALTGX instance contains. In XAUI mode, the number of channels increments by 4. September 2012 Altera Corporation -- Stratix IV Device Handbook Volume 3 1-10 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 6 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) Deterministic Latency This option sets the transceiver data path width. Single-width--This mode operates from 600 Mbps to 3.75 Gbps. Double-width--This mode operates from 1 Gbps to 8.5 Gbps. "Basic Single-Width Mode Configurations" and "Basic Double-Width Mode Configurations" sections in the Transceiver Architecture in Stratix IV Devices chapter. GIGE PCIe SDI Serial RapidIO What is the deserializer block width? -- XAUI These modes only operate in single-width mode. Double-width mode is not allowed. (OIF) CEI PHY Interface The (OIF) CEI PHY Interface mode only operates in double-width mode. Single-width mode is not allowed. -- SONET/SDH This option allows you to set the transceiver data path width. Stratix IV Device Handbook Volume 3 Single-width--Selected automatically in OC-12 and OC-48 configurations. The transceiver data path width is 8 bits. -- Double-width--Selected automatically in OC-96 configurations. The transceiver data path width is 16 bits. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-11 Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 7 of 11) ALTGX Setting Description Reference Basic Deterministic Latency This option determines the FPGA fabric-Transceiver interface width. Single-width mode--Selecting 8 or 10 bits bypasses the byte serializer/deserializer. Selecting 16 or 20 bits uses the byte serializer/deserializer. Double-width mode--Selecting 16 or 20 bits bypasses the byte serializer/deserializer. Selecting 32 or 40 bits uses the byte serializer/deserializer. Basic (PMA Direct) This option determines the FPGA fabric-Transceiver interface width. Single-width mode--You can select 8 or 10 bits. Double-width mode-- You can select 16 or 20 bits. GIGE This option determines the FPGA fabric-Transceiver interface width. In GIGE mode, only 8 bits are allowed. What is the channel width? (OIF) CEI PHY Interface This option selects the FPGA fabric-Transceiver width. In (OIF) CEI PHY Interface mode, only 32 bits are allowed. "Byte Serializer" and "Byte Deserializer" sections in the Transceiver Architecture in Stratix IV Devices chapter. PCIe This option determines the FPGA fabric-Transceiver interface width. In PCIe Gen1 (2.5 Gbps) mode, 8 and 16 bits are allowed. In PCIe Gen2 (5 Gbps) mode, only 16 bits are allowed. SDI This option determines the FPGA fabric-Transceiver interface width: HD mode--10-bit and 20-bit channel widths are allowed. 3G mode--only 20-bit channel width is allowed. 10-bit configuration--the byte serializer is not used. 20-bit configuration--the byte serializer is used. Serial RapidIO The channel width is fixed to 16 in Serial RapidIO mode. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-12 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 8 of 11) ALTGX Setting Description Reference SONET/SDH This option selects the FPGA fabric-Transceiver interface width. Depending on your subprotocol selection, choose one of the following: What is the channel width? 8 bits for OC-12 mode 16 bits for OC-48 mode 32 bits for OC-96 mode "Byte Serializer" and "Byte Deserializer" sections in the Transceiver Architecture in Stratix IV Devices chapter. XAUI XAUI mode only operates in single-width mode. Basic Basic (PMA Direct) You can select one of the following options: What would you like to base the setting on? Stratix IV Device Handbook Volume 3 Data rate--Selecting this option allows you to enter the transceiver channel serial data rate. Based on the value you enter, the ALTGX MegaWizard Plug-In Manager populates the input reference clock frequency options in the What is the input clock frequency? field. The ALTGX MegaWizard Plug-In Manager determines these input reference clock frequencies depending on the available multiplier settings. -- Input clock frequency--Selecting this option allows you to enter your input clock frequency. Based on the value you enter, the ALTGX MegaWizard Plug-In Manager populates the data rate options in the What is the effective data rate? field. The ALTGX MegaWizard Plug-In Manager determines these data rate options depending on the available multipler settings. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-13 Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 9 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) Deterministic Latency If you select the Data Rate option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager allows you to specify the effective serial data rate value in this field. -- If you select the Input Clock Frequency option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager displays the list of effective serial data rates in this field. GIGE This option is not available in GIGE mode. The transceiver channel serial data rate is fixed to 1250 Mbps in this mode. -- (OIF) CEI PHY Interface The allowed effective data rate is between 3125 Mbps and 6500 Mbps. Enter the transceiver channel's serial data rate in this field. -- PCIe This option is not available in PCIe mode. The defaults What is the effective data rate? are: 2500 Mbps for PCIe Gen1 mode. -- 5000 Mbps for PCIe Gen 2 mode. SDI The effective data rate is fixed at: 2967 Mbps or 2970 Mbps in 3G mode. 1483.5 Mbps or 1485 Mbps in HD mode. -- Serial RapidIO Enter one of these three data rates in this option: 1250 Mbps. 2500 Mbps. 3125 Mbps. -- SONET/SDH The effective data rate is fixed at: 622 Mbps in OC-12 mode. 2488.32 Mbps in OC-48 mode. -- 4976 Mbps in OC-96 mode. XAUI The effective data rate can be from 3125 Mbps to 3750 Mbps. September 2012 Altera Corporation -- Stratix IV Device Handbook Volume 3 1-14 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 10 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) If you select the Input Clock Frequency option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager allows you to specify the input reference clock frequency in this field. If you select the Data Rate option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager displays the list of input reference clock frequencies in this field. Deterministic Latency GIGE (OIF) CEI PHY Interface SDI SONET/SDH What is the input clock frequency? Based on the effective data rate value in the What is the effective data rate? field, the ALTGX MegaWizard Plug-In Manager determines the input reference clock frequencies depending on the available multiplier settings. "Input Reference Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. PCIe This option is not available in PCIe mode. The input reference clock frequency is fixed to 100 MHz in PCIe mode. Serial RapidIO This option provides the available input reference clock frequencies depending on whether your effective serial data rate is 1250 Mbps, 2500 Mbps, or 3125 Mbps and the available multiplier settings. XAUI This option provides the available input reference clock frequencies depending on whether your effective serial data rate is 3125 Mbps or 3750 Mbps and the available multiplier settings. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-15 Table 1-1. MegaWizard Plug-In Manager Options (General Screen for Basic Mode) (Part 11 of 11) ALTGX Setting Description Reference Basic Basic (PMA Direct) The ALTGX MegaWizard Plug-In Manager provides you the base data rate options for the CMU/advanced technology extended (ATX) PLL and receiver clock data recovery (CDR). -- If you select a value in this field that is greater than the value in the What is the effective data rate? field, the ALTGX MegaWizard Plug-In Manager enables the appropriate local clock divider values. The local divider is present in the receiver channels. GIGE This option is not available in this mode because the data rate is fixed. The ALTGX MegaWizard Plug-In Manager provides you the base data rate options for the CMU PLL and receiver CDR. -- (OIF) CEI PHY Interface Specify base data rate. Serial RapidIO XAUI -- This option is not available in these modes. The ALTGX MegaWizard Plug-In Manager provides you the base data rate options for the CMU PLL and receiver CDR. PCIe For Gen1 x1, an optional base data rate of either 2500 or 5000 Mbps is available. -- SDI This option is not available this mode as the data rate is fixed in 3G and HD modes. The ALTGX MegaWizard Plug-In Manager provides you the base data rate options for the CMU PLL and receiver CDR. -- SONET/SDH This option is not available in this mode as the data rates are fixed in OC-12, OC-48, and OC-96 modes. The ALTGX MegaWizard Plug-In Manager provides you the base data rate options for the CMU PLL and receiver CDR in this option. September 2012 Altera Corporation -- Stratix IV Device Handbook Volume 3 1-16 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings PLL/Ports Screen for the Parameter Settings Figure 1-4 shows the PLL/Ports screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-4. MegaWizard Plug-In Manager--ALTGX (PLL/Ports Screen) Table 1-2 lists the available options on the PLL/ports screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-2. MegaWizard Plug-In Manager Options (PLL/Ports Screen) (Part 1 of 3) ALTGX Setting Description Train receiver clock and data recovery (CDR) from pll_inclk. If you select this option, the input reference clock to the CMU PLL trains the receiver CDR. Use ATX Transmitter PLL Stratix IV Device Handbook Volume 3 This option is only available for certain data rates. Refer to the DC and Switching Characteristics for Stratix IV Devices chapter for the supported data rates. This option enables the auxiliary transmitter PLL. This is a low-jitter PLL that resides between the transceiver blocks and can be used as a transmitter PLL. Reference Table 1-77 in the Transceiver Architecture in Stratix IV Devices chapter. "Auxiliary Transmit (ATX) PLL Block" section in the Transceiver Architecture in Stratix IV Devices, the Transceiver Clocking in Stratix IV Devices chapter, and the DC and Switching Characteristics for Stratix IV Devices section. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-17 Table 1-2. MegaWizard Plug-In Manager Options (PLL/Ports Screen) (Part 2 of 3) ALTGX Setting Description Reference Enable PLL phase frequency detector (PFD) feedback to compensate latency uncertainty in tx_dataout and tx_clkout paths relative to the reference clock. This option applies only when you select Deterministic Latency functional mode. "CMU PLL Feedback" section in the Transceiver Architecture in Stratix IV Devices chapter. What is the TX PLL bandwidth mode? The available options are Auto, Low, Medium, and High. Select the appropriate option based on your system requirements. "PLL Bandwidth Setting" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. What is the receiver CDR bandwidth mode? The available options are Auto, Low, Medium, and High. Select the appropriate option based on your system requirements. "Clock and Data Recovery Unit" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. What is the acceptable PPM threshold between the receiver CDR VCO and the receiver input reference clock? In Automatic Lock mode, the CDR remains in Lock-to-Data (LTD) mode as long as the parts per million (PPM) difference between the CDR VCO output clock and the input reference clock is less than the PPM value that you set in this option. If the PPM difference is greater than the PPM value that you set in this option, the CDR switches to Lock-to-Reference (LTR) mode. "Automatic Lock Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. The range of values available in this option is 62.5 ppm to 1000 ppm. (1) Optional Ports Create a gxb_powerdown port to power down the transceiver block. When asserted, this signal powers down the entire transceiver block. If none of the channels are instantiated in a transceiver block, the Quartus II software automatically powers down the entire transceiver block. "User Reset and Power Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. Create a pll_powerdown port to power down the TX PLL. Each transceiver block has two CMU PLLs. Each CMU/ATX PLL has a dedicated power down signal called pll_powerdown. This signal powers down the CMU/ATX PLL. "User Reset and Power Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. Create a rx_analogreset port for the analog portion of the receiver. September 2012 Altera Corporation The receiver analog reset port is available in Receiver only and Receiver and Transmitter operation modes. This resets part of the analog portion of the receiver CDR in the receiver channel. Altera recommends using this port to implement the recommended reset sequence. The minimum pulse width is two parallel clock cycles. "User Reset and Power Down Signals" in the Reset Control and Power Down in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-18 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-2. MegaWizard Plug-In Manager Options (PLL/Ports Screen) (Part 3 of 3) ALTGX Setting Create a rx_digitalreset port for the digital portion of the receiver. Create a tx_digitalreset port for the digital portion of the transmitter. Description Reference The receiver digital reset port is available in Receiver only and Receiver and Transmitter operation modes. This resets the PCS portion of the receiver channel. Altera recommends using this port to implement the recommended reset sequence. The minimum pulse width is two parallel clock cycles. The transmitter digital reset port is available in Transmitter only and Receiver and Transmitter operation modes. This resets the PCS portion of the transmitter channel. Altera recommends using this port to implement the recommended reset sequence. The minimum pulse width is two parallel clock cycles. "User Reset and Power Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. "User Reset and Power Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. Create a pll_locked port to indicate PLL is in lock with the reference input clock. Each CMU/ATX PLL has a dedicated pll_locked signal that is fed to the FPGA fabric to indicate when the PLL is locked to the input reference clock. "Transceiver Reset Sequences" section in the Reset Control and Power Down in Stratix IV Devices chapter. Create an rx_locktorefclk port to lock the RX CDR to the reference clock. When this signal is asserted high, the LTR/LTD controller forces the receiver CDR to lock to the phase and frequency of the input reference clock. (1), (2) "LTR/LTD Controller" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_locktodata port to lock the RX CDR to the received data. When this signal is asserted high, the LTR/LTD controller forces the receiver CDR to lock to the received data. (1), (2) "LTR/LTD Controller" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_pll_locked port to indicate RX CDR is locked to the input reference clock. Create an rx_freqlocked port to indicate RX CDR is locked to the received data. In LTR mode, this signal is asserted high to indicate that the receiver CDR has locked to the phase and frequency of the input reference clock. In LTD mode, this signal has no significance. "Lock-to-Reference (LTR) Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. (1) This signal is asserted high to indicate that the receiver CDR has switched from LTR to LTD mode. This signal has relevance only in Automatic Lock mode and may be required to control the transceiver resets, as described in the User Reset and Power Down Signals section in the Reset Control and Power Down in Stratix IV Devices chapter. (1) "LTR/LTD Controller" section in the Transceiver Architecture in Stratix IV Devices chapter. Notes to Table 1-2: (1) LTR mode is lock-to-reference mode and LTD mode is lock-to-data mode. (2) When rx_locktorefclk and rx_locktodata are both asserted high, rx_locktodata takes precedence over rx_locktorefclk, forcing the CDR to lock to the received data. When both these signals are de-asserted, the LTR/LTD controller is configured in Automatic Lock mode. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-19 Ports/Calibration Screen for the Parameter Settings Figure 1-5 shows the Ports/Calibration screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-5. MegaWizard Plug-In Manager--ALTGX (Ports/Calibration Screen) Table 1-3 lists the available options on the Ports/Calibration screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Unless indicated otherwise, the options apply to all functional modes. Table 1-3. MegaWizard Plug-In Manager Options (Ports/Calibration Screen) (Part 1 of 3) ALTGX Setting Description Reference Optional Ports/Controls Create an rx_signaldetect port to indicate data input signal detection. This port is only available in Basic and PCIe mode. "Signal Threshold Detection Circuitry" section in the Transceiver Architecture in Stratix IV Devices chapter. Enable TX Phase Comp FIFO in register mode. This option is only available in Deterministic Latency mode. "Deterministic Latency" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_phase_comp_fifo_error output port. This output port indicates a Receiver Phase Compensation FIFO overflow or under-run condition. "Receiver Phase Compensation FIFO Error Flag" section in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-20 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Table 1-3. MegaWizard Plug-In Manager Options (Ports/Calibration Screen) (Part 2 of 3) ALTGX Setting Description Reference Create a tx_phase_comp_fifo_error output port. This output port indicates a Transmitter Phase Compensation FIFO overflow or under-run condition. "TX Phase Compensation FIFO Status Signal" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_coreclk port to connect to the read clock of the RX phase compensation FIFO. You can clock the parallel output data from the receiver using this optional input port. This port allows you to clock the read side of the Receiver Phase Compensation FIFO with a user-provided clock (FPGA fabric clock, FPGA fabric-Transceiver interface clock, or input reference clock). "FPGA Fabric-Transceiver Interface Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. Create a tx_coreclk port to connect to the write clock of the TX phase compensation FIFO. You can clock the parallel transmitter data generated in the FPGA fabric using this optional input port. This port allows you to clock the write side of the Transmitter Phase Compensation FIFO with a user-provided clock (FPGA fabric clock, FPGA fabric-Transceiver interface clock, or input reference clock). "FPGA Fabric-Transceiver Interface Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. Create a tx_forceelecidle input port In Basic and PCIe modes, this optional input signal places the transmitter buffer in the electrical idle state. "Transceiver Channel Architecture" section in the Transceiver Architecture in Stratix IV Devices chapter. Use calibration block. The calibration block is always enabled. "Calibration Blocks" section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-21 Table 1-3. MegaWizard Plug-In Manager Options (Ports/Calibration Screen) (Part 3 of 3) ALTGX Setting Description Create an active high cal_blk_powerdown to power down the calibration block. Asserting this signal high powers down the calibration block. A high-to-low transition on this signal restarts calibration. Reference "Input Signals to the Calibration Block" section in the Transceiver Architecture in Stratix IV Devices chapter. The options available for selection are based on what you specify in the Specify base data rate option: What is the Analog Power (VCCA_L/R)? 3.3 V--Available up to 11.3 Gbps for Stratix IV GT devices only. 3.0 V--Available up to 8.5 Gbps. 2.5 V--Available up to 4.25 Gbps. AUTO--The ALTGX MegaWizard Plug-In Manager automatically sets VCCA_L/R to 2.5 V for the VCO data rates less than 4.25 Gbps. "General Requirements to Combine Channels" section in the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. or VCCA_L/R to 3.0 V for the VCO data rates greater than 4.25 Gbps. It is up to you to connect the correct voltage supply to the VCCA_L/R pins on the board. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-22 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings Loopback Screen for the Parameter Settings Figure 1-6 shows the Loopback screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-6. MegaWizard Plug-In Manager--ALTGX (Loopback Screen) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-23 Table 1-4 lists the available options on the Loopback screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-4. MegaWizard Plug-In Manager Options (Lpbk Screen) ALTGX Setting Description Reference There are two options available: No loopback--This is the default mode. Serial loopback--If you select serial loopback, the rx_seriallpbken port is available to control the serial loopback feature dynamically. Which loopback option would you like? 1'b1--enables serial loopback 1'b0--disables serial loopback "Serial Loopback" section in the Transceiver Architecture in Stratix IV Devices chapter. This signal is asynchronous to the receiver datapath. There are three options available: No reverse loopback--This is the default mode. Reverse Serial loopback (pre-CDR)--This is the loopback before the receiver's CDR block to the transmitter buffer. The receiver path in PCS is active but the transmitter side is not. Reverse Serial loopback--This is a loopback after the receiver's CDR block to the transmitter buffer. The receiver path in PCS is active but the transmitter side is not. Which reverse loopback option would you like? September 2012 Altera Corporation "Loopback Modes" section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-24 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings RX Analog Screen for the Parameter Settings Figure 1-7 shows the RX Analog screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-7. MegaWizard Plug-In Manager--ALTGX (RX Analog Screen) Table 1-5 lists the available options on the RX Analog screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-5. MegaWizard Plug-In Manager Options (RX Analog Screen) (Part 1 of 2) ALTGX Setting Enable static equalizer control. Description This option enables the static equalizer settings. Reference "Programmable Equalization and DC Gain" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. This DC gain option has five settings: What is the DC gain? Stratix IV Device Handbook Volume 3 0 - 0 dB 1 - 3 dB 2 - 6 dB 3 - 9 dB 4 - 12 dB "Programmable Equalization and DC Gain" section in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-25 Table 1-5. MegaWizard Plug-In Manager Options (RX Analog Screen) (Part 2 of 2) ALTGX Setting Description Reference What is the receiver common mode voltage (RX VCM)? The receiver common mode voltage is programmable to 0.82 V or 1.1 V. "Receiver Channel Datapath" section in the Transceiver Architecture in Stratix IV Devices chapter. Force signal detection. In PCIe mode, this option disables the signal threshold detect circuit for the receiver CDR. The receiver CDR no longer depends on the signal detect criterion to switch from LTR to LTD mode. "Signal Threshold Detection Circuitry" section in the Transceiver Architecture in Stratix IV Devices chapter. What is the signal detect threshold? Use external receiver termination. Use this option in PCIe or Basic mode with the 8B/10B block enabled and the rx_signaldetect port selected to determine the threshold level for the signal detect circuit. PIPE mode--The levels are fixed. Basic mode--A range of values depending on the data rate are available. The levels will be determined after characterization. Select this option if you want to use an external termination resistor instead of differential on-chip termination (OCT). If checked, this option turns off the receiver OCT. This option allows you to select the receiver differential termination value. The settings allowed are: What is the receiver termination resistance? September 2012 Altera Corporation 85 100 120 150. "Signal Threshold Detection Circuitry" section in the Transceiver Architecture in Stratix IV Devices chapter. "Programmable Differential On-Chip Termination" section in the Transceiver Architecture in Stratix IV Devices chapter. "Programmable Differential On-Chip Termination" section in the Transceiver Architecture in Stratix IV Devices chapter, and the DC and Switching Characteristics for Stratix IV Devices section. Stratix IV Device Handbook Volume 3 1-26 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings TX Analog Screen for the Parameter Settings Figure 1-8 shows the TX Analog screen of the ALTGX MegaWizard Plug-In Manager for the Parameter Settings. Figure 1-8. MegaWizard Plug-In Manager--ALTGX (TX Analog Screen) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Parameter Settings 1-27 Table 1-6 lists the available options on the TX Analog screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-6. MegaWizard Plug-In Manager Options (TX Analog Screen) (Part 1 of 2) ALTGX Setting Description Reference The options available for selection are based on what you enter in the What is the effective data rate? option. What is the transmitter buffer power (VCCH)? 1.4 V--Available up to 8.5 Gbps. 1.5 V is available up to 6.5 Gbps (not available for Stratix IV GT). AUTO--The ALTGX MegaWizard Plug-In Manager automatically sets VCCH to 1.5 V for the effective data rates less than 6.5 Gbps "Programmable Transmit Output Buffer Power (VCCH)" section in the Transceiver Architecture in Stratix IV Devices chapter. or VCCH to 1.4 V for effective data rates greater than 6.5 Gbps. It is up to you to connect the correct voltage supply to the VCCH pins on the board. The transmitter common mode voltage is fixed to What is the transmitter common mode voltage (VCM)? 0.65 V. "Transmitter Output Buffer" in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. This option is available if you want to use an external termination resistor instead of the differential OCT. Checking this option turns off the transmitter differential OCT. "Programmable Transmitter Termination" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. This option selects the transmitter differential termination value. The settings allowed are 85 100 , 120 , and 150 . "Programmable Transmitter Termination" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. What is the voltage output differential (VOD) control setting? This option selects the VOD of the transmitter buffer. The available VOD settings change based on the transmitter termination resistance value. "Programmable Output Differential Voltage" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. What is the pre-emphasis first post-tap setting (% of VOD)? This option sets the amount of pre-emphasis on the transmitter buffer using first post-tap. "Programmable Pre-Emphasis" section in the Transceiver Architecture in Stratix IV Devices chapter. Use external transmitter termination. Select the transmitter termination resistance. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-28 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings Table 1-6. MegaWizard Plug-In Manager Options (TX Analog Screen) (Part 2 of 2) ALTGX Setting Description Reference What is the pre-emphasis pre-tap setting (% of VOD)? This option sets the amount of pre-emphasis on the transmitter buffer using pre-tap. "Programmable Pre-Emphasis" section in the Transceiver Architecture in Stratix IV Devices chapter. What is the pre-emphasis second post-tap setting (% of VOD)? This option sets the amount of pre-emphasis on the transmitter buffer using second post-tap. "Programmable Pre-Emphasis" section in the Transceiver Architecture in Stratix IV Devices chapter. Reconfiguration Settings This section describes the various dynamic reconfiguration modes and settings for Stratix IV GX and GT transceivers. In Reconfiguration Settings, when you enable the Enable Channel and Transmitter PLL reconfiguration option, the following screens become available: Modes Transmitter PLLs Clocking/Interface The following sections describe these screens and their corresponding settings. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings 1-29 Modes Screen for the Reconfiguration Settings Figure 1-9 shows the Modes screen, listing the various dynamic reconfiguration modes available. Figure 1-9. MegaWizard Plug-In Manager--Reconfiguration Settings September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-30 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings Table 1-7 lists the different options available in the Modes screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-7. MegaWizard Plug-In Manager Options (Modes Screen) (Part 1 of 2) ALTGX Setting Description Reference Dynamic Reconfiguration Settings The different dynamic reconfiguration modes available are listed in the Reconfiguration Settings screen. Based on which portion of the transceiver you want to reconfigure, select the corresponding options and connect the ALTGX_RECONFIG instance to the ALTGX instance. What do you want to be able to dynamically reconfigure in the transceiver? Analog controls (VOD, Pre-emphasis, and Manual Equalization and EyeQ)--Enable this option to dynamically reconfigure the PMA control settings similar to VOD, pre-emphasis, manual equalization, DC gain, and EyeQ. Enable adaptive equalizer control--Selecting this option enables the Adaptive Equalization (AEQ) hardware and provides the following additional ports: aeq_togxb[] aeq_fromgxb[] These ports provide the interface between the receiver channel and the dynamic reconfiguration controller. "Dynamic Reconfiguration Modes Implementation" section, "PMA Controls Reconfiguration Mode Details" section, "Enabling the AEQ Control Logic and AEQ Hardware" section, and the "Offset Cancellation Feature" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Offset cancellation for receiver channels--This option is enabled by default for Receiver only and Receiver and Transmitter configurations. It is not available for Transmitter only configurations. Ensure that you connect a dynamic reconfiguration controller to all the transceiver channels in the design. You must enable this option to reconfigure one of the following: Transmitter local divider block, CMU PLL, Transceiver channel, or Both the CMU PLL and transceiver channel. Enable Channel and Transmitter PLL Reconfiguration Channel Interface--This option allows channel interface reconfiguration. Use alternate CMU Transmitter PLL--This option sets up the alternate PLL so that the transceiver channel can optionally select between the output of the main and alternate transmitter PLL. Use additional CMU/ATX Transmitter PLLs from outside the Transceiver Block--This option allows you to select a maximum of four transmitter PLLs. For example, you can select the ATX PLL as the main PLL and three additional PLLs. Stratix IV Device Handbook Volume 3 "Transceiver Channel Reconfiguration Modes Details" section, "FPGA Fabric-Transceiver Channel Interface Selection" section, "Transceiver Channel Reconfiguration Modes Details" section. and the "Multi-PLL Settings" section in the Transceiver Architecture in Stratix IV Devices chapter. How many additional PLLs are used?--You can have a maximum of two PLLs outside the transceiver block. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings 1-31 Table 1-7. MegaWizard Plug-In Manager Options (Modes Screen) (Part 2 of 2) ALTGX Setting Description Reference How many input clocks are used? Enter the number of input clocks available for selection for the transmitter PLLs and receiver PLL. You have a choice of up to 10 input clock sources (clock 1, clock 2, and so on). "Guidelines for Specifying the Input Reference Clocks" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. What is the starting channel number? You must set the starting channel number of the first ALTGX instance controlled by the dynamic reconfiguration controller to 0. Set the starting channel number of the consecutive ALTGX instances controlled by the same dynamic reconfiguration controller, if any, in the next available multiples of 4. "Logical Channel Addressing while Reconfiguring the PMA Controls" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Transmitter PLL Settings Depending on the number of additional PLLs you select in the How many additional PLLs are used? option in Reconfiguration Settings, the corresponding PLL screens become available. Each of these PLL screens have the same settings available for selection. Table 1-8 lists each of these settings. 1 September 2012 The Main PLL is the PLL you configure in the General screen. Therefore, some of the options are already enabled or disabled for this PLL. Some of the options differ when compared with the additional transmitter PLLs. Altera Corporation Stratix IV Device Handbook Volume 3 1-32 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings Figure 1-10 shows the options available on the Main PLL screen of the ALTGX MegaWizard Plug-In Manager. Figure 1-10. MegaWizard Plug-In Manager Options--Main PLL Screen Table 1-8 lists the available options on the Main PLL screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-8. MegaWizard Plug-In Manager Options (Main PLL Screen) (Part 1 of 3) ALTGX Setting Description Reference Main Tx PLL/Rx PLL Settings "Selecting the PLL Logical Reference Index for Additional PLLs" and the "Multi-PLL Settings" sections in the Dynamic Reconfiguration in Stratix IV Devices chapter. Use central clock divider to drive the transmitter channels using x4/xN lines If this option is enabled, the transmitter PLL is outside the transceiver block. If this option is disabled, the transmitter PLL is one of the CMU PLLs within the same transceiver block. What is the PLL logical reference index (used in reconfiguration)? "Selecting the PLL Logical Reference Index for The PLL logical reference index is selected based on the Additional PLLs" and location of the alternate PLL. If the Use central clock divider "Selecting the Logical to drive the transmitter channels using x4/xN lines option is Reference Index of the CMU unchecked this must be 0 or 1, otherwise this must be 2 or 3. PLL" sections in the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings 1-33 Table 1-8. MegaWizard Plug-In Manager Options (Main PLL Screen) (Part 2 of 3) ALTGX Setting Description Reference What is the selected input clock source for the Rx/Tx PLLs? Assign identification numbers to all input reference clocks that are used by the transmitter PLLs in their corresponding PLL screens. You can set up a maximum of 10 input reference clocks and assign identification numbers from 1 to 10. "Guidelines for Specifying the Input Reference Clocks" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. What is the protocol to be reconfigured to? Select the desired functional mode here, if you intend to dynamically reconfigure the transceiver channel to a different functional mode using the alternate transmitter PLL. "Channel Reconfiguration with Transmitter PLL Select Mode Details" in the Dynamic Reconfiguration in Stratix IV Devices chapter. This option is not available for Basic, (OIF) CEI PHY Interface, Serial RapidIO, GIGE, and XAUI functional modes. What is the subprotocol to be reconfigured to? This option is available for the following protocols and subprotocols: Protocol = PCIe; Subprotocols = Gen 1 and Gen 2 Protocol = SDI; Subprotocols = 3G and HD Protocol = SONET/SDH; Subprotocols = OC12, OC48, and OC96 -- This option is available only for Basic mode.You can select one of the following options for the alternate transmitter PLL: Input clock frequency--Selecting this option allows you to enter your input clock frequency. Based on the value you enter, the ALTGX MegaWizard Plug-In Manager populates the data rate options in the What is the effective data rate? field. The ALTGX MegaWizard Plug-In Manager determines these data rate options depending on the available multiplier settings. Data rate--Selecting this option allows you to enter the transceiver channel serial data rate. Based on the value you enter, the ALTGX MegaWizard Plug-In Manager populates the input reference clock frequency options in the What is the input clock frequency? field. The ALTGX MegaWizard Plug-In Manager determines these input reference clock frequencies depending on the available multiplier settings. What would you like to base the setting on? -- These settings are to dynamically reconfigure the transceiver channel to listen to the alternate transmitter PLL. What is the data rate? September 2012 Altera Corporation If you select the data rate option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager allows you to specify the effective serial data rate value in this field. -- If you select the input clock frequency option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager displays the list of effective serial data rates in this field. Stratix IV Device Handbook Volume 3 1-34 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings Table 1-8. MegaWizard Plug-In Manager Options (Main PLL Screen) (Part 3 of 3) ALTGX Setting Description Reference These settings are to dynamically reconfigure the transceiver channel to listen to the alternate transmitter PLL. What is the input clock frequency? If you select the input clock frequency option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager displays the list of effective serial data rates in this field. If you select the data rate option in the What would you like to base the setting on? field, the ALTGX MegaWizard Plug-In Manager allows you to specify the effective serial data rate value in this field. "CMU PLL Reconfiguration Mode Details" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. What is the PLL bandwith mode? The available options are Auto, Low, Medium, and High. Select the appropriate option based on your system requirements. "PLL Bandwidth Setting" section in the Transceiver Architecture in Stratix IV Devices chapter. Create powerdown port to power down the PLL. Each transceiver block has two CMU PLLs. Each CMU/ATX PLL has a dedicated power down signal called pll_powerdown. This signal powers down the CMU PLL. "User Reset and Power-Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. Create locked port to indicate that the PLL is in lock with the reference clock. Each CMU/ATX PLL has a dedicated pll_locked signal that is fed to the FPGA fabric to indicate when the PLL is locked to the input reference clock. "User Reset and Power-Down Signals" section in the Reset Control and Power Down in Stratix IV Devices chapter. Use Auxiliary Transmitter (ATX) PLL (available only if central clock divider is used) Stratix IV Device Handbook Volume 3 This option is only available for certain data rates. Refer to the DC and Switching Characteristics for Stratix IV Devices chapter for the supported data rates. This option enables the auxiliary transmitter PLL. This is a low-jitter PLL that resides between the transceiver blocks and can be used as a transmitter PLL. "Auxiliary Transmit (ATX) PLL Block" section in the Transceiver Architecture in Stratix IV Devices chapter and the DC and Switching Characteristics for Stratix IV Devices section. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Reconfiguration Settings 1-35 Clocking/Interface Screen for the Reconfiguration Settings Figure 1-11 shows the Clocking/Interface screen of the ALTGX MegaWizard Plug-In Manager for the Reconfiguration settings. Figure 1-11. MegaWizard Plug-In Manager Options (Clocking/Interface Screen) Table 1-9 lists the available options on the Clocking/Interface screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. 1 This screen is not available for Basic (PMA Direct) x1 and xN configurations. Table 1-9. MegaWizard Plug-In Manager Options (Clocking/Interface Screen) (Part 1 of 2) ALTGX Setting Description Reference Dynamic Reconfiguration Channel Internal and Interface Settings Select one of the following available options: How should the receivers be clocked? September 2012 Altera Corporation Share a single transmitter core clock between receivers Use the respective channel transmitter core clocks Use the respective channel receiver core clocks "Clocking/Interface Options" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-36 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-9. MegaWizard Plug-In Manager Options (Clocking/Interface Screen) (Part 2 of 2) ALTGX Setting Description Reference Select one of the following available options: How should the transmitters be clocked? Share a single transmitter core clock between transmitters Use the respective channel transmitter core clocks "Clocking/Interface Options" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Create an 'rx_revbitorderwa' input port to use receiver enable bit reversal This optional input port allows you to dynamically reverse the bit order at the output of the receiver word aligner. "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. Check a control box to use the corresponding control port. You can select various control and status signals depending on what protocol(s) you intend to dynamically reconfigure the transceiver channel to. "FPGA Fabric-Transceiver Channel Interface Selection" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Protocol Settings This section describes the various screens available to set up the PCS blocks of the Stratix IV transceiver. 1 Protocol Settings are not available for Basic (PMA Direct) functional mode. Based on the protocol you select in the General screen of Parameter Settings, the screens listed in Table 1-10 become available. Table 1-10. Protocol Settings Protocol Settings Screens Protocols Basic Deterministic Latency SDI Serial RapidIO 8B/10B Word Aligner Rate match/Byte order Y(Basic/8B10B) Y Y Y Det. Latency/8B10B) Y -- Y SDI/8B10B) Y -- Y Serial RapidIO/8B10B) Y Y The following sections describe these screens and the available settings for each of them. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-37 8B10B Screen for the Protocol Settings Figure 1-12 shows the 8B10B screen of the MegaWizard Plug-In Manager for the Protocol Settings. Figure 1-12. MegaWizard Plug-In Manager--ALTGX (8B10B Screen) Table 1-11 lists the available options on the 8B10B screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-11. MegaWizard Plug-In Manager Options (8B10B Screen) (Part 1 of 3) ALTGX Setting Description Reference Enable low latency PCS mode. This option disables all the PCS blocks except the Transmitter/Receiver Phase Comp FIFO and optional byte serializer/de-serializer. "Low Latency PCS Datapath" section in the Transceiver Architecture in Stratix IV Devices chapter. Enable 8B/10B decoder/encoder. This option is available if the channel width is 8-bits, 16-bits, or 32-bits. "8B/10B Decoder" section in theTransceiver Architecture in Stratix IV Devices chapter. 8B/10B encoder force disparity control: Create a tx_forcedisp to enable Force disparity and use tx_dispval to code up the incoming word using positive or negative disparity. September 2012 Altera Corporation When asserted high--forces the 8B/10B encoder to encode the data on the tx_datain port with a positive or negative disparity depending on the tx_dispval signal level. When de-asserted low--the 8B/10B encoder encodes the data on the tx_datain port according to the 8B/10B running disparity rules. "8B/10B Encoder" and "Transceiver Port Lists" sections in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-38 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-11. MegaWizard Plug-In Manager Options (8B10B Screen) (Part 2 of 3) ALTGX Setting Description Reference This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. This signal indicates whether the decoded 8-bit code group is a data or control code group on this port. Create an rx_ctrldetect port to indicate 8B/10B decoder has detected a control code. If the received 10-bit code group is one of the 12 control code groups (/Kx.y/) specified in the IEEE802.3 specification, this signal is driven high. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. If the received 10-bit code group is a data code group (/Dx.y/), this signal is driven low. The signal width is 1, 2, and 4 bits for a channel width of 8 bits, 16 bits, and 32 bits, respectively. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric and indicates an 8B/10B code group violation. Create an rx_errdetect port to indicate 8B/10B decoder has detected an error code. This signal is asserted high if the received 10-bit code group has a code violation or disparity error. It is used along with the rx_disperr signal to differentiate between a code violation error and/or a disparity error. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. The signal width is 1, 2 and 4 bits for a channel width of 8 bits, 16 bits, and 32 bits, respectively. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. Create an rx_disperr port to indicate 8B/10B decoder has detected a disparity error. This signal is asserted high if the received 10-bit code or data group has a disparity error. When this signal goes high, rx_errdetect is also asserted high. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. The signal width is 1, 2, and 4 bits for a channel width of 8 bits, 16 bits, and 32 bits, respectively. Create an rx_runningdisp port to indicate the current running disparity of the 8B10B decoded byte. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric to indicate the current running disparity of the 8B/10B decoded byte. "8B/10B Decoder" section of Table 1-77 in the Transceiver Architecture in Stratix IV Devices chapter. Flip receiver output data bits. This option reverses the bit order of the parallel receiver data at a byte level at the output of the receiver phase compensation FIFO. For example, if the 16-bit parallel receiver data at the output of the receiver phase compensation FIFO is '10111100 10101101' (16'hBCAD), enabling this option reverses the data on rx_dataout port to '00111101 10110101' (16'h3DB5). -- This option reverses the bit order of the parallel transmitter data at a byte level at the input of the transmitter phase compensation FIFO. For example, if the 16-bit parallel transmitter data at the tx_datain Flip transmitter input data bits. port is '10111100 10101101' (16'hBCAD), enabling this option reverses the input data to the transmitter phase compensation FIFO to '00111101 10110101' (16'h3DB5). -- Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-39 Table 1-11. MegaWizard Plug-In Manager Options (8B10B Screen) (Part 3 of 3) ALTGX Setting Description Reference Enabling this option in: Enable transmitter bit reversal. Single-width mode--the 8-bit D[7:0] or 10-bit D[9:0] data at the input of the serializer gets rewired to D[0:7] or D[0:9], respectively. Double-width mode--the 16-bit D[15:0] or 20-bit D[19:0] data at the input of the serializer gets rewired to D[0:15] or D[0:19], respectively. "Transmitter Bit Reversal" section in the Transceiver Architecture in Stratix IV Devices chapter. For example, if the 8-bit parallel data at the input of the serializer is '00111101', enabling this option reverses this serializer input data to '10111100.' Create a tx_invpolarity port to allow Transmitter polarity inversion. This optional port allows you to dynamically reverse the polarity of every bit of the data word fed to the serializer in the transmitter data path. Use this option when the positive and negative signals of the differential output from the transmitter (tx_dataout) are erroneously swapped on the board. You can only select this option when you use the Create tx_bitslipboundary Transmitter only or Receiver and Transmitter select port to control the operation mode. This option enables the number of words slipped in tx_bitslipboundaryselect input to control the the TX bitslipper. number of bits slipped in the TX bitslipper. September 2012 Altera Corporation "Transmitter Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. -- Stratix IV Device Handbook Volume 3 1-40 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Word Aligner Screen for the Protocol Settings Figure 1-13 shows the Word Aligner screen of the MegaWizard Plug-In Manager for the Protocol Settings. Figure 1-13. MegaWizard Plug-In Manager--ALTGX (Word Aligner Screen) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-41 Table 1-12 lists the available options on the Word Aligner screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. 1 The word aligner and rate matcher operations and patterns are pre-configured for PCIe, GIGE, and XAUI modes, and cannot be altered. Table 1-12. MegaWizard Plug-In Manager Options (Word Aligner Screen) (Part 1 of 4) ALTGX Setting Use manual word alignment mode. Description Reference Enabling this option sets the word aligner in Manual Alignment mode. In Manual Alignment mode, the word aligner operation is controlled by the input signal rx_enapatternalign. "Manual Alignment Mode Word Aligner with 8-bit PMA-PCS Interface Modes" and "Manual Alignment Mode Word Aligner with 10-bit PMA-PCS Interface Modes" sections in the Transceiver Architecture in Stratix IV Devices chapter. Two options are available in manual mode: When should the word aligner realign? Realign continuously while the rx_enapatternalign signal is high. Realign at the rising edge of the rx_enapatternalign signal. This option sets the word aligner in Bit-Slip mode. Enabling this option creates an input signal rx_bitslip to control the word aligner. At every rising edge of the rx_bitslip signal, the bit slip circuitry slips one bit into the received data stream, effectively shifting the word boundary by one bit. Use manual bitslipping mode. SDI Because word alignment and framing occur after de-scrambling, the word aligner in the receiver data path is not useful in SDI systems. Altera recommends driving the ALTGX rx_bitslip signal low to prevent the word aligner from inserting bits in the received data stream. Use the Automatic synchronization state machine mode. This option sets the word aligner in Automatic Synchronization State Machine mode. This mode is available only in Single-width mode for 8B/10B encoded data: 10-bit PCS-PMA Interface where the 8B/10B encoder is enabled or September 2012 Altera Corporation "Manual Alignment Mode Word Aligner with 8-bit PMA-PCS Interface Modes" and "Manual Alignment Mode Word Aligner with 10-bit PMA-PCS Interface Modes" sections in the Transceiver Architecture in Stratix IV Devices chapter. "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. 10-bit PCS-PMA Interface where the 8B/10B is disabled but the data is already 8B/10B encoded Stratix IV Device Handbook Volume 3 1-42 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-12. MegaWizard Plug-In Manager Options (Word Aligner Screen) (Part 2 of 4) ALTGX Setting Description Reference Use this option in Automatic Synchronization State Machine mode to indicate the number of consecutive valid words that it must receive between erroneous words to reduce the error count by one. The rx_syncstatus stays high as long as the error count is less than the programmed error count. "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Number of bad data words before loss of synch state. Use this option in Automatic Synchronization State Machine mode to indicate the number of bad data words (error count) that it must receive to lose synchronization. The loss-of-synch is indicated by the rx_syncstatus signal going low. "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Number of valid patterns before synch state is reached. Use this option in Automatic Synchronization State Machine mode to indicate the number of word alignment patterns that it must receive without intermediate erroneous code groups to achieve synchronization. The rx_syncstatus signal is driven high to indicate that synchronization has been achieved. "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Number of consecutive valid words before synch state is reached. This option sets the word alignment pattern length. The available choices depend on the following conditions: Whether the data is 8B/10B encoded or not Which mode is used in Single-width mode: What is the word alignment pattern length? What is the word alignment pattern? Stratix IV Device Handbook Volume 3 for 8-bit PCS-PMA Interface (8B/10B encoder disabled), only 16 bits are allowed. for 10-bit PCS-PMA, 7 and 10 bits are allowed. Which mode is used in Double-width mode: for 16-bit PCS-PMA Interface (8B/10B encoder disabled), 8, 16, and 32 bits are allowed. for 20-bit PCS-PMA Interface, 7, 10, and 20 bits are allowed. Enter the word alignment pattern in MSB to LSB order with MSB at the left most bit position. The length of the alignment pattern is based on the What is the word alignment pattern length? option. The word aligner restores the word boundary by looking for the pattern that you enter here. For example, if you want to set the word alignment pattern to /K28.5/: You must enter the word alignment pattern length: 10. You must enter the word alignment pattern: 0101111100 (17C). "Word Aligner in Single-Width Mode" and "Word Aligner in Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. "Word Aligner in Single-Width Mode" and "Word Aligner in Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-43 Table 1-12. MegaWizard Plug-In Manager Options (Word Aligner Screen) (Part 3 of 4) ALTGX Setting Flip word alignment pattern bits. Description Reference When this option is enabled, the ALTGX MegaWizard Plug-In Manager flips the bit order of the pattern that you enter in the What is the word alignment pattern? option and uses the flipped version as the word alignment pattern. For example, if you enter '0101111100' (17C) as the word alignment pattern and enable this option, the word aligner uses '0011111010' as the word alignment pattern. -- This option creates the output signal rx_rlv. Enabling this option also activates the run-length violation circuit. If the number of continuous 1s and 0s exceeds the number that you set in this option, the run-length violation circuit asserts the rx_rlv signal. The rx_rlv signal is asynchronous to the receiver data path and is asserted for a minimum of two recovered clock cycles in Single-width mode. Similarly, it is asserted for a minimum of three recovered clock cycles in Double-width mode. Enable run-length violation checking with a run length of: The run length limits are as follows: Single-width mode: 8-bit and 16-bit channel width: 4 to 128 in increments of four 10-bit and 20-bit channel width: 5 to 160 in increments of five "Programmable Run Length Violation Detection" section in the Transceiver Architecture in Stratix IV Devices chapter. Double-width mode: 16-bit and 32-bit channel width: 8 to 512 in increments of eight 20-bit and 40-bit channel width: 10 to 640 in increments of 10 Enable word aligner output reverse bit ordering. In manual bit-slip mode, this option creates an input port rx_revbitorderwa to dynamically reverse the bit order at the output of the receiver word aligner. "Receiver Bit Reversal" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_syncstatus output port for pattern detector and word aligner. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that synchronization has been achieved. This signal is synchronous with the parallel receiver data on the rx_dataout port. This signal is not available in bit-slip mode. Signal width is 1, 2, and 4 bits for a channel width of 8-bits/10-bits, 16-bits/20-bits, and 32-bits/40-bits, respectively. Table 1-77, "Word Aligner in Single-Width Mode" and "Word Aligner in Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_patterndetect port to indicate pattern detected. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that the word alignment pattern programmed has been detected in the current word boundary. Signal width is 1, 2, and 4 bits for a channel width of 8-bits/10-bits, 16-bits/20-bits, and 32-bits/40-bits, respectively. Table 1-77 and "Word Aligner in Single-Width Mode" and "Word Aligner in Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-44 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-12. MegaWizard Plug-In Manager Options (Word Aligner Screen) (Part 4 of 4) ALTGX Setting Create an rx_invpolarity port to enable word aligner polarity inversion. Description Reference This optional port allows you to dynamically reverse the polarity of every bit of the received data at the input of the word aligner. Use this option when the positive and negative signals of the differential input to the receiver (rx_datain) are erroneously swapped on the board. "Receiver Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. This is an optional input port that is available only in the double-width mode. It creates an rx_revbyteorderwa port to dynamically swap the MSByte and LSByte of the data at the output of the word aligner in the receiver data path. Enabling this Create an option compensates for the erroneous swapping of rx_revbyteorderwa to enable Receiver symbol swap. bytes at the upstream transmitter and corrects the data received by the downstream systems. "Receiver Byte Reversal in Basic Double-Width Modes" section in the Transceiver Architecture in Stratix IV Devices chapter. For example, if the 16-bit output of the word aligner is 0B0A, asserting the rx_revbyteorderwa signal swaps the two bytes so the output becomes 0A0B. Create rx_bitslipboundaryselec tout port to indicate the number of bits slipped in the word aligner. Stratix IV Device Handbook Volume 3 This option is available for selection only when you are in Receiver only or Receiver and Transmitter operation mode. This option enables the rx_bitslipboundaryselectout output to indicate the number of bits slipped in the word aligner. -- September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-45 Rate Match/Byte Order Screen for the Protocol Settings Figure 1-14 shows the Rate Match/Byte Order screen of the MegaWizard Plug-In Manager for the Protocol Settings. Figure 1-14. MegaWizard Plug-In Manager--ALTGX (Rate Match/Byte Order Screen) September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-46 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-13 lists the available options on the Rate Match/Byte Order screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-13. MegaWizard Plug-In Manager Options (Rate Match/Byte Order Screen) (Part 1 of 3) ALTGX Setting Description This option enables the rate match (clock rate compensation) FIFO. The rate match block consists of a 20-word deep FIFO. Depending on the PPM difference, the rate match FIFO controls insertion and deletion of skip characters based on the 20-bit rate match pattern you enter in the What is the 20-bit rate match pattern1? and What is the 20-bit rate match pattern2? options. To enable this block: Enable rate match FIFO. The transceiver channel must have both the transmitter and the receiver channels instantiated. You must select the Receiver and Transmitter option in the What is the operation mode? field in the General screen. You must also enable the 8B/10B encoder/decoder in the 8B10B screen. Reference "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. The rate match block is capable of compensating up to 300 PPM difference between the upstream transmitter clock and the local receiver's input reference clock. What is the 20-bit rate match pattern1? (usually used for +ve disparity pattern) Enter a 10-bit skip pattern and a 10-bit control pattern. In the skip pattern field, you must choose a 10-bit code group that has neutral disparity. When the rate matcher receives the 10-bit control pattern followed by the 10-bit skip pattern, it inserts or deletes the 10-bit skip pattern as necessary to avoid rate match FIFO overflow or underflow conditions. (1) "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. What is the 20-bit rate match pattern2? (usually used for -ve disparity pattern) Enter a 10-bit skip pattern and a 10-bit control pattern. In the skip pattern field, you must choose a 10-bit code group that has neutral disparity. When the rate matcher receives the 10-bit control pattern followed by the 10-bit skip pattern, it inserts or deletes the 10-bit skip pattern as necessary to avoid rate match FIFO overflow or underflow conditions. (1) "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Create the rx_rmfifofull port to indicate when the rate match FIFO is full. This option creates the output port rx_rmfifofull when you enable the Enable Rate Match FIFO option. It is a status flag that the rate match block forwards to the FPGA fabric. It indicates when the rate match FIFO block is full (20 words). This signal remains high as long as the FIFO is full. It is asynchronous to the receiver data path. "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-47 Table 1-13. MegaWizard Plug-In Manager Options (Rate Match/Byte Order Screen) (Part 2 of 3) ALTGX Setting Description Reference Create the rx_rmfifoempty port to indicate when the rate match FIFO is empty. This option creates the output port rx_rmfifoempty when you enable the Enable Rate Match FIFO option. It is a status flag that the rate match block forwards to the FPGA fabric. It indicates when the rate match FIFO block is empty (5 words full). This signal remains high as long as the FIFO is empty. It is asynchronous to the receiver data path. "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Create the rx_rmfifodatainserted port to indicate when data is inserted in the rate match FIFO. This option creates the output port rx_rmfifodatainserted flag when you enable the Enable Rate Match FIFO option. It is a status flag that the rate match block forwards to the FPGA fabric. This indicates the insertion of skip patterns. For every deletion, this signal is high for one parallel clock cycle. "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Create the rx_rmfifodatadeleted port to indicate when data is deleted in the rate match FIFO. This option creates the output port rx_rmfifodatadeleted flag when you enable the Enable Rate Match FIFO option. It is a status flag that the rate match block forwards to the FPGA fabric. This indicates the deletion of skip patterns. For every insertion, this signal is high for one parallel clock cycle. "Rate Match FIFO in Basic Single-Width Mode" and "Rate Match FIFO in Basic Double-Width Mode" sections in the Transceiver Architecture in Stratix IV Devices chapter. Enable insertion or deletion of consecutive characters or ordered sets This option enables the back-to-back insertion or deletion of skip characters in the rate match FIFO. This option is available for selection in Single-width mode. It is enabled by default in Double-width mode. -- This option enables the byte ordering block. It is available in both Single-width and Double-width modes. It is available only when the channel width is: Enable byte ordering block. 16-bits/20-bits in Single-width mode 32-bits/40-bits in Double-width mode As soon as the byte ordering block sees the rising edge of the appropriate signal, it compares the LSByte coming out of the byte deserializer with the byte ordering pattern. If they do not match, the byte ordering block inserts the pad character that you enter in the What is the byte ordering pad pattern? option such that the byte ordering pattern is seen in the LSByte position. Inserting this pad character enables the byte ordering block to restore the correct byte order. "Byte Ordering Block" section in the Transceiver Architecture in Stratix IV Devices chapter. What do you want the byte ordering to be based on? This option is available only when the byte ordering block is enabled. This option allows you to trigger the byte ordering block on the rising edge of either the rx_syncstatus signal or the user-controlled rx_enabyteord signal from the FPGA fabric. "Byte Ordering Block" section in the Transceiver Architecture in Stratix IV Devices chapter. What is the byte ordering pattern? This option is available only when the byte ordering block is enabled. Enter the 10-bit pattern that the byte ordering block must place in the LSByte position of the receiver parallel data on the rx_dataout port. "Byte Ordering Block" section in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-48 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-13. MegaWizard Plug-In Manager Options (Rate Match/Byte Order Screen) (Part 3 of 3) ALTGX Setting What is the byte ordering pad pattern? Description When the byte ordering block does not find the byte ordering pattern in the LSByte position of the data coming out of the byte deseriazlier, it inserts this byte ordering pad pattern such that the byte ordering pattern is seen in the LSByte position of the receiver parallel data on the rx_dataout port. Inserting this pad character enables the byte ordering block to restore the correct byte order. Reference "Byte Ordering Block" section in the Transceiver Architecture in Stratix IV Devices chapter. Note to Table 1-13: (1) If you want the rate matcher to insert or delete both the positive and negative disparities of the 20-bit rate matching pattern, enter the positive disparity as pattern1 and negative disparity as pattern2. Protocol Settings Screen for GIGE and XAUI Figure 1-15 shows the Protocol Settings screen for the GIGE and XAUI modes of the MegaWizard Plug-In Manager. Figure 1-15. MegaWizard Plug-In Manager--ALTGX (Protocol Settings Screen--GIGE and XAUI) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-49 Table 1-14 lists the available options for the GIGE and XAUI modes in the Protocol Settings screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-14. MegaWizard Plug-In Manager Options (Protocol Settings --GIGE and XAUI) (Part 1 of 3) ALTGX Setting Enable run-length violation checking with a run length of __. Description This option creates the output signal rx_rlv. Enabling this option also activates the run-length violation circuit. If the number of continuous 1s and 0s exceeds the number that you set in this option, the run-length violation circuit asserts the rx_rlv signal. The rx_rlv signal is asynchronous to the receiver data path and is asserted for a minimum of two recovered clock cycles. Reference "Programmable Run Length Violation Detection" section in the Transceiver Architecture in Stratix IV Devices chapter. The run length limits are five to 160 in increments of five. Create an rx_syncstatus output port for pattern detector and word aligner. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that synchronization has been achieved. This signal is synchronous with the parallel receiver data on the rx_dataout port. Receiver synchronization is indicated on the rx_syncstatus port of each channel. Table 1-33 and the "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_patterndetect port to indicate pattern detected. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that the word alignment pattern programmed has been detected in the current word boundary. Table 1-33 and the "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_invpolarity port to enable word aligner polarity inversion. This optional port allows you to dynamically reverse the polarity of every bit of the received data at the input of the word aligner. Use this option when the positive and negative signals of the differential input to the receiver (rx_datain) are erroneously swapped on the board. "Receiver Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_ctrldetect port to indicate 8B/10B decoder has detected a control code. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. This signal indicates whether the decoded 8-bit code group is a data or control code group on this port. If the received 10-bit code group is one of the 12 control code groups (/Kx.y/) specified in IEEE802.3 specification, this signal is driven high. If the received 10-bit code group is a data code group (/Dx.y/), this signal is driven low. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_errdetect port to indicate 8B/10B decoder has detected an error code. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. This signal indicates an 8B/10B code group violation. It is asserted high if the received 10-bit code group has a code violation or disparity error. It is used along with the rx_disperr signal to differentiate between a code violation error and/or a disparity error. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-50 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-14. MegaWizard Plug-In Manager Options (Protocol Settings --GIGE and XAUI) (Part 2 of 3) ALTGX Setting Create an rx_disperr port to indicate 8B/10B decoder has detected a disparity error. Description Reference This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric.This signal is asserted high if the received 10-bit code or data group has a disparity error. When this signal goes high, rx_errdetect also is asserted high. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. This optional port allows you to dynamically reverse the polarity of every bit of the data word fed to the serializer in the transmitter data path. Create a tx_invpolarity port to Use this option when the positive and negative allow Transmitter polarity inversion. signals of the differential output from the transmitter (tx_dataout) are erroneously swapped on the board. "Transmitter Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_runningdisp port to indicate the current running disparity of the 8B/10B decoded byte. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. This signal is asserted high when the current running disparity of the 8B/10B decoded byte is negative. This signal is low when the current running disparity of the 8B/10B decoded byte is positive. Create an rx_rmfifofull port to indicate when the rate match FIFO is full. This option creates the output port rx_rmfifofull. It is a status flag that the rate match block forwards to the FPGA fabric. This indicates when the rate match FIFO block is full (20 words). This signal remains high as long as the FIFO is full and is asynchronous to the receiver data path. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_rmfifoempty port to indicate when the rate match FIFO is empty. This option creates the output port rx_rmfifoempty. It is a status flag that the rate match block forwards to the FPGA fabric. This indicates when the rate match FIFO block is empty (five words). This signal remains high as long as the FIFO is empty and is asynchronous to the receiver data path. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_rmfifodatainserted port to indicate when data is inserted in the rate match FIFO. This option creates the output port rx_rmfifodatainserted flag. It is a status flag that the rate match block forwards to the FPGA fabric. The rx_rmfifodatainserted flag is asserted when a rate match pattern byte is inserted to compensate for the PPM difference in reference clock frequencies between the upstream transmitter and the local receiver. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. This option creates the output port rx_rmfifodatadeleted. It is a status flag that Create an rx_rmfifodatadeleted the rate match block forwards to the FPGA fabric. port to indicate when data is deleted The rx_rmfifodatadeleted flag is asserted in the rate match FIFO. when a rate match pattern byte is deleted to compensate for the PPM difference in reference clock frequencies between the upstream transmitter and the local receiver. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 -- September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-51 Table 1-14. MegaWizard Plug-In Manager Options (Protocol Settings --GIGE and XAUI) (Part 3 of 3) ALTGX Setting Description Reference Enable transmitter bit reversal. Enabling this option reverses every bit of the 10-bit parallel data at the input of the serializer. The 10-bit input to the serializer D[9:0] is reversed to D[0:9]. "8B/10B Encoder" section in the Transceiver Architecture in Stratix IV Devices chapter. What is the word alignment pattern length? This option sets the word alignment pattern length. The available choices are 7 and 10 for the GIGE and XAUI modes. The default setting for this option is 10. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. Protocol Settings Screen for the (OIF) CEI Phy Interface Table 1-15 lists the available options for the (OIF) CEI Phy Interface mode in the Protocol Settings screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-15. MegaWizard Plug-In Manager Options (Protocol Settings - [OIF] CEI PHY Interface) ALTGX Setting Enable run-length violation checking with a run length of __. Description This option creates the output signal rx_rlv. Enabling this option also activates the run-length violation circuit. If the number of continuous 1s and 0s exceeds the number that you set in this option, the run-length violation circuit asserts the rx_rlv signal. The rx_rlv signal is asynchronous to the receiver data path and is asserted for a minimum of two recovered clock cycles. Reference "Programmable Run Length Violation Detection" section in the Transceiver Architecture in Stratix IV Devices chapter. For a 32-bit channel width, the run length limits are 8 to 512 in increments of eight. September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 1-52 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Protocol Settings Screen for PCIe Figure 1-16 shows the PCIe 1 screen for Protocol Settings of the MegaWizard Plug-In Manager. Figure 1-16. MegaWizard Plug-In Manager--ALTGX (PCIe 1 Screen) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-53 Table 1-16 lists the available options on the PCIe 1 screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-16. MegaWizard Plug-In Manager Options (PCIe 1) (Part 1 of 2) ALTGX Setting Enable low latency synchronous PCIe. Enable run-length violation checking with a run length of __. Description Reference This option puts the rate match FIFO into low latency mode, which forces the system into a 0 ppm mode. Ensure that there is a 0 ppm difference between the upstream transmitter's and the local receiver's input reference clocks. "Rate Match (Clock Rate Compensation) FIFO" section in the Transceiver Architecture in Stratix IV Devices chapter. This option creates the output signal rx_rlv. Enabling this option also activates the run-length violation circuit. If the number of continuous 1s and 0s exceeds the number that you set in this option, the run-length violation circuit asserts the rx_rlv signal. The rx_rlv signal is asynchronous to the receiver data path. "Programmable Run Length Violation Detection" section in the Transceiver Architecture in Stratix IV Devices chapter. For both 8-bit and 16-bit channel widths, the run length limits are 5 to 160 in increments of five. Enable fast recovery mode. This option enables the CDR control block. When this block is enabled, the rx_locktodata and rx_locktorefclk signals are disabled. "Fast Recovery Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Enable the electrical idle inference module by selecting this option. In PCIe mode, the PCS has an optional electrical idle inference module designed to implement the electrical idle inference conditions specified in PCIe base specification 2.0. Enable electrical idle inference functionality. Create an rx_syncstatus output port for pattern detector and word aligner. September 2012 Altera Corporation Enabling this option creates the rx_elecidleinfersel[2:0] input signal. The electrical idle Inference module infers electrical idle depending on the logic level driven on the rx_elecidleinfersel[2:0] input signal. For the electrical idle Inference module to correctly infer an electrical idle condition in each LTSSM sub-state, you must drive the rx_elecidleinfersel[2:0] signal appropriately. The ALTGX MegaWizard Plug-In Manager automatically configures the word aligner in Automatic Synchronization State Machine mode for PCIe mode. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that synchronization has been achieved. This signal is synchronous with the parallel receiver data on the rx_dataout port. The signal width is 1 and 2 bits for a channel width of 8 bits and 16 bits, respectively. "Electrical Idle Inference" section in the Transceiver Architecture in Stratix IV Devices chapter. Table 1-29 and "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-54 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-16. MegaWizard Plug-In Manager Options (PCIe 1) (Part 2 of 2) ALTGX Setting Create an rx_patterndetect output port to indicate pattern detected. Description Reference This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that the word alignment pattern programmed has been detected in the current word boundary. The signal width is 1 and 2 bits for a channel width of 8 bits and 16 bits, respectively. "Automatic Synchronization State Machine Mode Word Aligner with 10-bit PMA-PCS Interface Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. This is an output status signal that the 8B/10B decoder forwards to the FPGA fabric. This signal indicates whether the decoded 8-bit code group is a data or control code group on this port. Create an rx_ctrldetect port to indicate 8B/10B decoder has detected a control code. If the received 10-bit code group is one of the 12 control code groups (/Kx.y/) specified in the IEEE802.3 specification, this signal is driven high. If the received 10-bit code group is a data code group (/Dx.y/), this signal is driven low. The signal width is 1 and 2 bits for a channel width of 8 bits and 16 bits, respectively. "8B/10B Decoder" section in the Transceiver Architecture in Stratix IV Devices chapter. Create a tx_detectrxloop input port as Receiver detect or loopback enable, depending on the power state. Depending on the power-down mode, asserting this signal enables either the receiver detect operation or Loopback mode. (1) "Receiver Detection" and "PCIe Reverse Parallel Loopback" section in the Transceiver Architecture in Stratix IV Devices chapter. Create a tx_forceelecidle input port to force the Transmitter to send Electrical Idle signals. Enabling this port sets the transmitter buffer in electrical idle mode. This port is available in all PCIe power-down modes and has a specific use in each mode. (1) "Transmitter Buffer Electrical Idle" section in the Transceiver Architecture in Stratix IV Devices chapter. A high level on this port forces the associated parallel transmitter data on the tx_datain port to be transmitted with negative current running disparity. For 8-bit transceiver channel width configurations, you must drive tx_forcedispcompliance[1:0] high in the same parallel clock cycle as the first /K28.5/ of the compliance pattern on the tx_datain port. For 16-bit transceiver channel width configurations, you must drive only the LSB of tx_forcedispcompliance[1:0]high in the same parallel clock cycle as /K28.5/D21.5/ of the compliance pattern on the tx_datain port. Create a tx_forcedispcompliance input port to force negative running disparity. Create a tx_invpolarity port to allow Transmitter polarity inversion. This optional port allows you to dynamically reverse the polarity of every bit of the data word fed to the serializer in the transmitter data path. Use this option when the positive and negative signals of the differential output from the transmitter (tx_dataout) are erroneously swapped on the board. "Compliance Pattern Transmission Support" section in the Transceiver Architecture in Stratix IV Devices chapter. "Transmitter Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. Note to Table 1-16: (1) Refer to the table 'Power States and Functions Allowed in Each Power State' in the PIPE Interface section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-55 Figure 1-17 shows the PCIe 2 screen of Protocol Settings for the MegaWizard Plug-In Manager. Figure 1-17. MegaWizard Plug-In Manager--ALTGX (PCIe 2 Screen) Table 1-17 lists the available options on the PCIe 2 screen of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-17. MegaWizard Plug-In Manager Options (PCIe 2 Screen) (Part 1 of 2) ALTGX Setting Description Reference Create a pipestatus output port for PIPE interface status signal. The PCIe interface block receives status signals from the transceiver channel PCS and PMA blocks and encodes the status on a 3-bit output signal (pipestatus[2:0]) that is forwarded to the FPGA fabric. "Receiver Status" section and Table 1-53 in the Transceiver Architecture in Stratix IV Devices chapter. Create a pipedatavalid output port to indicate valid data from the receiver. This is an output status port that indicates the receiver parallel data on the rx_dataout port is valid. September 2012 Altera Corporation -- Stratix IV Device Handbook Volume 3 1-56 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-17. MegaWizard Plug-In Manager Options (PCIe 2 Screen) (Part 2 of 2) ALTGX Setting Description Reference Enabling this option creates the pipeelecidle output status port that is forwarded to the FPGA fabric. Create a pipeelecidle output port for Electrical Idle detect status signal. If you select Enable Electrical Idle Inference Module, the pipeelecidle signal is driven high when the electrical idle inference module infers an electrical idle condition depending on the logic driven on the rx_elecidleinfersel[2:0] port. Otherwise, it is driven low. If you do not select Enable Electrical Idle Inference Module, the rx_signaldetect output signal from the signal threshold detection circuitry is inverted and driven on the pipeelecidle port. "Electrical Idle Inference" section in the Transceiver Architecture in Stratix IV Devices chapter. The pipeelecidle signal is asynchronous to the receiver data path. Create a pipephydonestatus output port to indicate PIPE completed power state transitions. This is an output status signal forwarded to the FPGA fabric. The completion of various PHY functions; for example, receiver detection, power state transition, clock switch, and rate switch, are indicated on this pipephydonestatus signal by driving this signal high for one parallel clock cycle. "PCIe Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Create a pipe8b10binvpolarity port to enable polarity inversion in PIPE. This optional port allows you to dynamically reverse every bit of the received data at the input of the 8B/10B decoder. "PCIe Mode" section in the Transceiver Architecture in Stratix IV Devices chapter. Create a powerdn input port for PIPE powerdown directive. Enabling this option creates an input control port powerdn[1:0] for each transceiver channel. "Power State Management" section and Table 1-51 in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-57 Protocol Settings Screen for SONET/SDH Figure 1-18 shows the SONET/SDH screen for Protocol Settings of the MegaWizard Plug-In Manager. Figure 1-18. MegaWizard Plug-In Manager--ALTGX (Protocol Settings--SONET/SDH) Table 1-18 lists the available options on the SONET/SDH screen for Protocol Settings of the MegaWizard Plug-In Manager for your ALTGX custom megafunction variation. Table 1-18. MegaWizard Plug-In Manager Options (SONET/SDH Screen) (Part 1 of 3) ALTGX Setting When should the word aligner realign? Description This option is not available in SONET/SDH mode. In SONET/SDH mode, the word aligner operates in Manual Alignment mode. By default, the ALTGX MegaWizard Plug-In Manager sets the behavior of the word aligner such that re-alignment occurs when there is a rising edge of the rx_enapatternalign input signal in this mode. This option sets the length of the word alignment pattern. The following options are available: What is the word alignment pattern length? September 2012 Altera Corporation OC-12--only 16-bit pattern is allowed. OC-48--only 16-bit pattern is allowed. OC-96--16-bit and 32-bit patterns are allowed. Reference "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. "SONET/SDH Mode" (OC-12, OC-48, and OC-96) section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-58 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-18. MegaWizard Plug-In Manager Options (SONET/SDH Screen) (Part 2 of 3) ALTGX Setting Description What is the word alignment pattern? Enter the word alignment pattern. By default, the pattern that appears in the MegaWizard Plug-In Manager is '0001010001101111' (16'h146F). Flip word alignment pattern bits. This option is enabled in the MegaWizard Plug-In Manager by default. This option reverses the order of the alignment pattern at a bit level to support MSB-to-LSB transmission in SONET/SDH mode. The ALTGX MegaWizard Plug-In Manager flips the bit order of the default word alignment pattern '0001010001101111 '(16'h146F) and uses the flipped version '1111011000101000' (16'hF628) as the word alignment pattern. What do you want the byte ordering to be based on? This option allows you to trigger the byte ordering block either on the rising edge of the rx_syncstatus signal or the user-controlled rx_enabyteord signal from the FPGA fabric. The byte ordering block is enabled only in OC-48 mode. Enable run-length violation checking with a run length of. This option creates the output signal rx_rlv. Enabling this option also activates the run-length violation circuit. If the number of continuous 1s and 0s exceeds the number that you set in this option, the run-length violation circuit asserts the rx_rlv signal. The rx_rlv signal is asynchronous to the receiver data path and is asserted for a minimum of two recovered clock cycles in OC-12 and OC-48 modes. Similarly, it is asserted for a minimum of three recovered clock cycles in the OC-96 mode. Reference "SONET/SDH Mode" (OC-12, OC-48, and OC-96) section in the Transceiver Architecture in Stratix IV Devices chapter. -- "Byte Ordering Block" section in the Transceiver Architecture in Stratix IV Devices chapter. "Programmable Run Length Violation Detection" section in the Transceiver Architecture in Stratix IV Devices chapter. For the OC-12 and OC-48 modes, the run length limits are 4 to 128 in increments of four. For the OC-96 mode, the run length limits are 5 to 160 in increments of five. Create an rx_syncstatus output port for pattern detector and word aligner. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that synchronization has been achieved. This signal is synchronous with the parallel receiver data on the rx_dataout port. The signal width is 1 bit, 2 bits, and 4 bits for a channel width of 8 bits, 16 bits, and 32 bits, respectively. Table 1-77 and "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. Create an rx_patterndetect port to indicate pattern detected. This is an output status signal that the word aligner forwards to the FPGA fabric to indicate that the word alignment pattern programmed has been detected in the current word boundary. The signal width is 1 bit, 2 bits, and 4 bits for a channel width of 8 bits, 16 bits, and 32 bits, respectively. Table 1-33 and "Word Aligner" section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-59 Table 1-18. MegaWizard Plug-In Manager Options (SONET/SDH Screen) (Part 3 of 3) ALTGX Setting Description Create a rx_invpolarity port to enable word aligner polarity inversion. This optional port allows you to dynamically reverse the polarity of every bit of the received data at the input of the word aligner. Use this option when the positive and negative signals of the differential input to the receiver (rx_datain) are erroneously swapped on the board. "Receiver Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. Create a tx_invpolarity port to allow Transmitter polarity inversion. This optional port allows you to dynamically reverse the polarity of every bit of the data word fed to the serializer in the transmitter data path. Use this option when the positive and negative signals of the differential output from the transmitter (tx_dataout) are erroneously swapped on the board. "Transmitter Polarity Inversion" section in the Transceiver Architecture in Stratix IV Devices chapter. This option reverses the bit order of the parallel receiver data at a byte level at the output of the receiver phase compensation FIFO to support MSB-to-LSB transmission in SONET/SDH mode. Flip receiver output data bits. Flip transmitter input data bits. September 2012 Altera Corporation For example, if the 16-bit parallel receiver data at the output of the receiver phase compensation FIFO is '10111100 10101101' (16'hBCAD), enabling this option reverses the data on the rx_dataout port to '00111101 10110101' (16'h3DB5). This option reverses the bit order of the parallel transmitter data at a byte level at the input of the transmitter phase compensation FIFO to support MSB-to-LSB transmission protocols in SONET/SDH mode. For example, if the 16-bit parallel transmitter data at the tx_datain port is '10111100 10101101' (16'hBCAD), enabling this option reverses the input data to the transmitter phase compensation FIFO to '00111101 10110101' (16'h3DB5). Reference "SONET/SDH Mode" (OC-12, OC-48, and OC-96) section in the Transceiver Architecture in Stratix IV Devices chapter. "SONET/SDH Mode" (OC-12, OC-48, and OC-96) section in the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 1-60 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings EDA Screen Figure 1-19 shows the EDA screen of the MegaWizard Plug-In Manager. The Generate Netlist option generates a netlist for the third party EDA synthesis tool to estimate timing and resource utilization for the ALTGX instance. Figure 1-19. MegaWizard Plug-In Manager--ALTGX (EDA Screen) Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings 1-61 Summary Screen Figure 1-20 shows the Summary screen of the MegaWizard Plug-In Manager. You can select optional files on this page. After you make your selections, click Finish to generate the files. Figure 1-20. MegaWizard Plug-In Manager--ALTGX (Summary Screen) Document Revision History Table 1-19 lists the revision history for this chapter. Table 1-19. Document Revision History (Part 1 of 2) Date Version September 2012 4.3 December 2011 4.2 September 2012 Altera Corporation Changes Updated Table 1-1 to close FB #65275. Updated Table 1-12 to close FB #37243. Updated Table 1-1. Stratix IV Device Handbook Volume 3 1-62 Chapter 1: ALTGX Transceiver Setup Guide for Stratix IV Devices Protocol Settings Table 1-19. Document Revision History (Part 2 of 2) Date Version February 2011 November 2009 June 2009 March 2009 November 2008 4.1 4.0 3.1 3.0 2.0 Changes Updated Table 1-1, Table 1-3, Table 1-7, and Table 1-17. Updated chapter title. Minor text edits. Applied new template. Added Deterministic Latency protocol information. Added AEQ information. Updated PLL setting information. Consolidated Parameter Settings information (Table 1-1 to Table 1-6). Consolidated Reconfiguration Settings information (Table 1-7 to Table 1-9). Consolidated Protocol Settings information (Table 1-10 to Table 1-18). Minor text edits. Updated Table 1-9, Table 1-29 and Table 1-35. Updated Figure 1-10. Added introductory sentences to improve search ability. Minor text edits. Updated the figures to match the software changes. Removed the 'Deterministic Latency' subprotocol from Basic functional mode. Removed the various clock frequencies from the Reconfig Clks screen for all the applicable functional modes. Updated Table 1-1, Table 1-6, and Table 1-11. Updated Figure 1-8. Added Reconfig Clks and Reconfig 2 sections. Added the "Use ATX Transmitter PLL" setting. Changed the "Which device speed grade will you be using?" setting to the "Which device variation will you be using" setting. June 2008 1.1 Minor text edit. May 2008 1.0 Initial release. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation 2. Transceiver Design Flow Guide for Stratix IV Devices February 2011 SIV53002-4.1 SIV53002-4.1 This chapter describes the Altera-recommended basic design flow that simplifies Stratix(R) IV GX transceiver-based designs. Use the following design flow techniques to simplify transceiver implementation. The "Guidelines to Debug Transceiver-Based Designs" on page 2-14 provides guidelines to trouble-shoot transceiver-based designs. An example of a fibre channel protocol application is also described in this chapter. The transceiver-based design is divided into phases and are detailed in the following sections: "Architecture" on page 2-3 "Implementation and Integration" on page 2-6 "Compilation" on page 2-10 "Verification" on page 2-11 "Functional Simulation" on page 2-12 "Example 1: Fibre Channel Protocol Application" on page 2-17 Figure 2-1 shows the design flow chart of the different stages of the design flow. The design flow stages include architecture, functional simulation, compilation, and verification. Each stage of the design flow is explained in the sections that follow. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 3 February 2011 Feedback Subscribe 2-2 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Figure 2-1. Flow Chart of the Different Stages in a Transceiver-Based Design Architecture Device Specification Transceiver Configuration Select Options in the Dynamic Reconfiguration Controller (if required) Clocking Power Supplies Implementation Create Transceiver Instances Is Dynamic Reconfiguration Required ? No Yes Create a Dynamic Reconfiguration Controller using the ALTGX_Reconfig MegaWizard Create Reset and Control Logic Create Data Processing Logic Integrate the Design Yes Functional Simulation If used, Include the Stratix IV GX ALTGX megafunction-generated wrapper file (.v or .vhd) and ALTGX_Reconfig megafunction-generated wrapper file Is Simulation Required ? No Compilation Synthesize the Design Require SignalTap Yes for Verification? Add Altera Simulation Library Files Verification No Simulate the Design Create Pin and OCT Assignments Create Timing Constraints Add signals to SignalTap II Logic Analyzer Include SignalTap file (.stp) in the Compilation Create Clock Grouping Constraints if Required Compile the Design Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Architecture 2-3 Architecture The first step in creating a transceiver-based design is to map your system requirements with the Stratix IV GX device supported features. The Stratix IV GX device contains multiple transceiver channels that you can configure in multiple data rates and protocols. It also provides multiple transceiver clocking options. For your design, identify the transceiver capabilities and clocking options to ensure that the transceiver meets your system requirements. This section describes the critical parameters that you need to identify as part of this architecture phase. Device Specification The following device specifications must meet your requirements: Refer to the device data sheet to ensure that the transceivers meet the data rate and electrical requirements for your target high-speed interface application; for example, the jitter specification and voltage output differential (VOD) range. Check whether the device family that you select supports your design requirements; for example, the number of transceiver channels, FPGA logic density, memory elements, and DSP blocks. If you intend to migrate to a higher logic density or higher transceiver count device in the future, ensure that the migration device is available. f For information about device characteristics, refer to the "Transceiver Performance Specifications" section in the DC and Switching Characteristics for Stratix IV Devices chapter. For information about transceiver resources, refer to the Overview for the Stratix IV Device Family chapter. Transceiver Configuration Use the ALTGX MegaWizardTM Plug-In Manager interface to configure the Stratix IV transceiver channel's features and options. When selecting a transceiver configuration, check for the following parameters: Check whether the transceiver physical coding sublayer (PCS) and physical medium attachment (PMA) functional blocks comply with your system requirements. For example, check whether the rate match (clock rate compensation) FIFO in the receiver channel PCS meets the parts per million (PPM) specifications required for your application. f For more information about transceiver specifications, refer to the "Transceiver Performance Specifications" section of the DC and Switching Characteristics for Stratix IV Devices chapter. February 2011 Select a configuration that meets your latency requirements. If your system has maximum latency requirements through the transceiver data path, consider the appropriate functional configuration. The Stratix IV GX transceiver supports various configurations that differ in latency (for example, low latency PCS mode and Basic [PMA direct] mode). Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Architecture 2-4 In some configurations, specific functional blocks in the transceiver are disabled or bypassed. Before you select a transceiver configuration, understand the functional blocks that must be implemented in the FPGA fabric. For example, Basic (PMA direct) mode provides reduced latency but does not have PCS functional blocks enabled (for example, word aligner and 8B/10B encoder). Therefore, implement these functional blocks in the FPGA fabric if you need them in your application. Some examples of functional blocks that you may need to implement in the FPGA fabric are shown in "Create Data Processing and Other User Logic" on page 2-8. f For more information about the ALTGX MegaWizard Plug-In Manager, refer to the ALTGX Transceiver Setup Guide for Stratix IV Devices chapter. Check whether the loopback features are available for your selected functional mode. The Stratix IV GX transceiver provides diagnostic loopback features between the transmitter channel and the receiver channel at the transceiver PCS and PMA interfaces. These loopback features help in debugging your design. If your design uses multiple transceiver channels within the same transceiver block, based on the transceiver channel configurations, the Quartus(R) II software might impose restrictions on combining these channels. f For more information about these restrictions, refer to the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Dynamic Reconfiguration Use the Stratix IV transceivers in multiple-link interconnect environments by dynamically reconfiguring the PMA controls (for example, VOD, Pre-emphasis, Equalization, DC gain, and the transceiver channel configuration). You can also reconfigure the PMA controls without affecting any other transceiver channel or the logic in the FPGA fabric. Use the transceiver channel reconfiguration to dynamically switch a transceiver channel to multiple protocols and data rates. The Quartus II software allows you to generate a memory initialization file (.mif) that stores unique transceiver settings and provides a dynamic reconfiguration controller, which is soft logic that controls the transceiver reconfiguration with minimal user interface logic. You can generate this soft logic using the ALTGX_RECONFIG MegaWizard interface. f For more information about the ALTGX_RECONFIG interface, refer to the ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices chapter. 1 All receiver channels in the Stratix IV GX device require offset cancellation to counter offset variations in process, voltage, and temperature (PVT) on the receiver. The dynamic reconfiguration controller initiates the sequence to perform offset cancellation on the receiver channels. Therefore, if you configure the Stratix IV GX transceiver channel in Receiver only or Transmitter and Receiver configuration, you must instantiate a dynamic reconfiguration controller. f For more information about offset cancellation or dynamic reconfiguration of PMA controls or channel configuration, refer to the "Offset Cancellation Feature" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Architecture 2-5 Clocking The Stratix IV GX transceiver is clocked by various input reference clocks, for example: Dedicated transceiver reference clock (refclk) pins. Altera recommends using refclk pins whenever possible because the refclk pins yield reduced jitter on the transmitted data. Clock sources connected to global clock lines. Clock outputs from the phase-locked loops (PLLs) in the FPGA fabric. Identify the transceiver channels input reference clock sources, for example: Ensure that your selected device has the required number of input reference clock resources to implement your design. Ensure that the transceiver clock input supports the required I/O standards. Ensure that the clocking restrictions work with your selected device: Check whether the allowed frequencies for the transceiver input reference clocks meet your system requirements. If you use the PLL cascade clock, understand its restrictions. If you are using the auxiliary transmit (ATX) PLL, understand the recommendations for the input reference clock sources and the restrictions on data rate ranges supported by the ATX PLL. For transceiver-FPGA interface clocking: Ensure that the transceiver-FPGA interface clock frequency limits meet your system requirements. f For information about transceiver specifications, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Identify the clocking scheme to clock the transceiver data to the logic in the FPGA fabric. For example, if your design has multiple transceiver channels that run at the same data rate and are connected to the one upstream link, you might be able to use a single transceiver-FPGA clock to provide clocks to the transceiver data path, which can conserve clock routing resources. If you are using Basic (PMA direct) mode, determine whether you require a left/right PLL to provide phase shifted clocks to the FPGA fabric. The left/right PLL clocks the data received and transmitted between the transceiver and the FPGA fabric interface and may be required to meet the timing requirements of the data transfer. f For information about transceiver clocking, refer to the Transceiver Clocking in Stratix IV Devices chapter. After you identify the required transceiver parameters, start the implementation and integration phase. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Implementation and Integration 2-6 Power Supplies The Stratix IV GX device requires multiple power supplies. The pin connection guidelines provide specific recommendations about the type of power supply regulator (linear or switching) and the voltage supply options and restrictions. For example, the transmitter buffer supply VCCHTx has two options--1.5 V and 1.4 V. There are specific data rate restrictions when using 1.5 V. You must understand these restrictions when you select a power supply value. f For more information, refer to the Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. Estimate the power required to run your design. This estimation allows you to select the appropriate power supply modules and to design the power distribution network on your board. Use the Early Power Estimator tool to estimate the transient current requirements. f For more information about the Early Power Estimation tool, refer to the Stratix III, Stratix IV, Stratix V, HardCopy III, and HardCopy IV PowerPlay Early Power Estimator. If your design is already complete, use the power optimization features available in the Stratix IV Devices. f For more information about optimizing power in Stratix IV FPGA devices, refer to AN 514: Power Optimization in Stratix IV FPGAs. Board Design Requirements For improved signal integrity on the high-speed serial interface, follow the best design practices for your power distribution network, PCB design, and stack up. f For detailed guidelines and recommendations about your power distribution network, PCB design, and stack up, refer to the Board Design Resource Center web site. f For more information about the Stratix IV GX design process, refer to AN 519: Stratix IV Design Guidelines. Implementation and Integration There are three steps to the implementation and integration phase: February 2011 "Create Transceiver Instances" on page 2-7 "Create Reset Logic to Control the FPGA Fabric and Transceivers" on page 2-34 "Create Data Processing and Other User Logic" on page 2-36 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Implementation and Integration 2-7 Create Transceiver Instances The ALTGX MegaWizard Plug-In Manager to creates the transceiver instance. In the architecture phase, you identified the transceiver configuration for your design. Using the ALTGX MegaWizard Plug-In Manager, select the appropriate parameters that apply to your architecture requirements. Reset and Status Signals The ALTGX MegaWizard Plug-In Manger provides various reset and status signals: Reset signals--tx_digitalreset, rx_digitalreset, rx_analogreset, and pll_powerdown are required to reset the transceiver PCS and PMA functional blocks. Status signals--rx_freqlocked and pll_locked indicate the state of the receiver CDR and transmitter PLL, respectively. Use these reset and status signals to implement the transceiver reset control logic in the FPGA fabric. For more information, refer to "Create Reset and Control Logic" on page 2-8. If you determine that your application requires dynamic reconfiguration, select the options in the Reconfig screen of the ALTGX MegaWizard interface. If you intend to dynamically reconfigure the channel into other protocol modes or data rates, the Reconfig screen provides multiple options (for example, the channel interface and Use alternate PLL options) to enable this feature. f To understand the logical channel addressing, logical PLL index, and type of reconfiguration to select options in the Reconfig screen, refer to the "Channel and CMU PLL Reconfiguration Mode Details" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Depending on your system, when you use multiple transceiver channels, you might be able to share the transmitter and receiver parallel clocks of one channel with the other channels. If your design requires sharing a clock resource, select the tx_coreclk and rx_coreclk ports. f Transceiver-FPGA fabric interface clock sharing conditions are provided in the Transceiver Clocking in Stratix IV Devices chapter. f For more information about using the ALTGX MegaWizard Plug-In Manager and the functionality of the different options and signals available, refer to the ALTGX Transceiver Setup Guide for Stratix IV Devices chapter. Create Dynamic Reconfiguration Controller Instances Use the ALTGX_RECONFIG MegaWizard interface to create the dynamic reconfiguration controller instance. If you intend to use the channel and CMU PLL reconfiguration feature, select the relevant options in the ALTGX_RECONFIG Megawizard Plug-In Manager. f For descriptions of the options in the ALTGX_RECONFIG megafunction, refer to the ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Implementation and Integration 2-8 f For more information about using the signals, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. Create Reset and Control Logic The reset sequence is important for initializing the transceiver functional blocks to proper operating condition. Altera recommends a reset sequence for different transceiver configurations and protocol functional modes. The ALTGX MegaWizard Plug-In Manager provides the tx_digitalreset, rx_analogreset, rx_digitalreset, and pll_powerdown signals to reset the different functional blocks of the transceiver. You can reset the CMU PLL or the ATX PLL (based on your selection) using the pll_powerdown signal. For transceiver instances that share the same CMU PLL or ATX PLL, the pll_powerdown port of these instances must be driven by the same logic. f For more information about reset sequences, refer to the Reset Control and Power Down in Stratix IV Devices chapter. Create Data Processing and Other User Logic A typical transceiver-based design consists of custom data processing and other user logic that must be implemented in the FPGA fabric based on your application requirements. In addition to application-specific logic, for specific transceiver configurations, you may need additional logic to interface with the transceivers. This section provides examples of such logic. PPM Detector When the Receiver CDR Is Used in Manual Lock Mode Each receiver channel contains a clock data recovery (CDR) that you can use in automatic or manual lock mode. If you use the receiver CDR in manual lock mode, you can control the timing of the CDR to lock to the input reference clock using the rx_locktorefclk port or lock to the recovered data using the rx_locktodata port. When you use the receiver CDR in manual lock mode, you may need to implement the PPM detector in the FPGA fabric to determine the PPM difference between the upstream transmitter and the Stratix IV GX receiver. Synchronization State Machine in Manual Word Alignment Mode Each receiver channel contains a synchronization state machine in the PCS that you can enable in certain functional modes. The synchronization state machine triggers the loss of synchronization status to the FPGA fabric based on invalid 8B/10B code groups. However, the synchronization state machine in the PCS is not available in some functional modes. You may need to implement custom logic in the FPGA fabric to indicate the loss-of-synchronization status of the received data. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Implementation and Integration 2-9 Gear Boxing Logic Some protocols require a wider data path than provided by the transceiver interface; for example, the Interlaken Protocol requires 64/67-bit encoding and decoding, but the maximum data path interface in the Stratix IV GX transceiver is 40 bits. Therefore, you must implement gear box logic to interface the 64/67-bit encoder-decoder with the transceiver interface. Functional Blocks to Interface with the Transceiver Configured in Basic (PMA Direct) Mode In Basic (PMA direct) mode, all the PCS functional blocks in the transceiver channel are disabled. Therefore, you may need to implement the following blocks in the FPGA fabric: Word Alignment--To align the byte boundary on the received data. Byte Deserializer--To increase the data path width to the rest of the user logic and to reduce the clock frequency of the data path by two. Phase Compensation FIFO (for bonded channel applications)--In bonded channel applications in which multiple transceiver channels are connected to the same upstream system (for example, one Interlaken Protocol link using 24 transceiver channels). To minimize the global clock routing resources you use, implement a phase compensation FIFO to interface the receiver side of the transceiver interface with the logic in the FPGA Fabric. Use the recovered clock from each channel to clock the write side of the phase compensation FIFO. Use the recovered clock from any of the channels to clock the read side of the phase compensation FIFO. With this method, you only use one clock resource and the subsequent receive-side logic in the FPGA fabric can operate in this single clock domain. Deskew Logic (for bonded channel applications)--In bonded channel applications in which multiple transceiver channels are connected to the same upstream system, the data received between multiple channels are not aligned due to potential skew in the interconnect and the upstream transmitter system. To compensate for the skew, use deskew logic in the FPGA fabric. Encoding/Decoding or Scrambling/Descrambling--Many protocols require the transmitter data to be encoded or scrambled to maintain signal integrity. This logic may be required in the FPGA fabric based on your application requirements. Integrate the Design After you implement all of the required logic, integrate the transceiver instances with the remaining logic and provide the appropriate transceiver-FPGA fabric interface clocking. Synthesize the design using third-party synthesis tools, such as Synopsys Synplicity or the Quartus II software synthesis tool. This allows you to detect syntax errors in your design. 1 February 2011 If you are using the transceiver in Basic (PMA direct) mode, you must develop all the PCS functionality in the FPGA fabric. Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Compilation 2-10 Compilation When you compile your design, the Quartus II software generates an SRAM Object File (.sof) or programmer object file (.pof) that you can download to the Stratix IV GX hardware. Typically, the first step in compiling the design is assigning pin locations for the I/Os and clocks. Use the pin planner tool in the Quartus II software to assign pins. 1 For a basic tutorial about the Quartus II software, open the Quartus II software, click the Help menu and select Tutorial. Stratix IV GX transceivers support a variety of I/O standards for the input reference clocks and serial data pins. Assign pins and the logic level standard (for example, 1.5-V PCML and LVDS) for the input and output pins. f For more information, refer to the I/O Features in Stratix IV Devices chapter. If you share the same transceiver-FPGA fabric interface clocks for multiple transceiver channels (tx_coreclk and rx_coreclk) in your design, set the 0 ppm constraints. These constraints enable the Quartus II software to relax the legality check restrictions on clocking. f For more information, refer to the "Common Clock Driver Selection Rules" section of the Transceiver Clocking in Stratix IV Devices chapter. For transceiver serial pins and refclk pins, set the on-chip termination (OCT) resistor settings. f For more information about supported OCT settings, refer to "Transmitter Output Buffer" section of the Transceiver Architecture in Stratix IV Devices chapter. Create timing constraints for the clocks and data paths. Use the TimeQuest Timing Analyzer to set timing constraints. f For more information about the TimeQuest Timing Analyzer, refer to the Quartus II Development Software Handbook. Compile the design. This generates a .sof that can be downloaded in the FPGA. The Quartus II software generates multiple report files that contain information such as transceiver configuration and clock resource utilization. The following section describes the report files relevant to using transceivers and clock resource. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Verification 2-11 Report Files The Quartus II software provides a report file in the synthesis, fitter, map, placement, and assembler stages. The report file provides useful information on the device and transceiver configuration generated by the Quartus II software. This section only describes the reports provided in the fitter stage. To access the report, click on the Processing menu, select the Compilation Report option and expand the Fitter tab. Fitter Summary The fitter summary provides high-level information on the FPGA fabric resources and transceiver channels used by your design. For example, to ensure that the Quartus II software has created the number of transceiver channels as specified in your design, refer to the GXB Receiver channels and GXB Transmitter channels field at the bottom of the report. For detailed information on resource utilization, expand the Fitter tab. Pin-Out File Select the Pin-Out file option under the Fitter tab. The Quartus II software displays the I/O standards and bank numbers of all the pins (used and unused) needed to connect to the board. The Quartus II software also generates a PIN file (.pin) with the above information. Altera recommends using the .pin as a guideline. Use the pin connection guidelines for board layout. f For more information about pin connection guidelines for board layout, refer to Stratix IV GX and Stratix IV E Device Family Pin Connection Guidelines. Resource Section Expand the Resource Section option under the Fitter tab to view the following tabs: The GXB Transmitter channel tab--Provides generated settings for all the transmitter channels instantiated in your design. The GXB Transmitter PLL tab--Provides generated settings for all the transmitter PLLs instantiated in your design. The GXB Receiver channel tab--Provides generated settings for all the receiver channels instantiated in your design. The Global and other fast signals tab--Displays the list of clock and other signals in your design that are assigned to the global and regional clock resources. You can use the report file to verify whether the transceiver settings (for example, data rate), are generated per your settings in the ALTGX MegaWizard Plug-In Manager. Verification The SignalTap(R) Logic Analyzer allows you to verify design functionality using the on-chip logic analyzer. SignalTap provides options to create multiple sets of signals that can be sampled using different trigger clocks. You can add the signals to the SignalTap Logic Analyzer and save the file as an STP file (.stp). When you include this .stp along with the design files and compile the design, the Quartus II software creates an .sof that allows you to verify the functionality of the signals that you added in the SignalTap Logic Analyzer file. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Functional Simulation 2-12 You can run the .stp that connects to the device through the JTAG port and displays the signal transitions using the Quartus II software. Because the JTAG port is required to run SignalTap, consider designing the board with the JTAG interface for debugging your system. f For more information about using SignalTap, refer to the Design Debugging Using the SignalTap II Embedded Logic Analyzer section in volume 3 of the Quartus II Development Software Handbook. To verify the functionality of the PCS and PMA blocks, the Stratix IV GX transceiver provides diagnostic loopback features between the transmitter and the receiver channels. f For more information, refer to the "Loopback Modes" section in the Transceiver Clocking in Stratix IV Devices chapter. Functional Simulation Use the ALTGX MegaWizard Plug-In Manager-generated wrapper file to simulate the instantiated transceiver configuration in third-party simulation software such as ModelSim. For simulation, specific Altera(R) simulation library files are required (listed in Table 2-1). The following library files are available in VHDL and Verilog versions: 220pack 220model altera_mf_components altera_mf sgate_pack sgate stratixiv_hssi_component stratixiv_hssi_atoms These simulation files are available under the following folder in the Quartus II installation directory: <Quartus II installation folder>/eda/sim_lib 1 The stratixiv_hssi_component library file is only applicable if the transceiver instance is created using VHDL. For VHDL simulation using ModelSim, create the following libraries in your ModelSim project: lpm sgate altera_mf stratixiv_hssi These simulation files are available under <Quartus II installation folder\quartus\eda\sim_lib>. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Functional Simulation 2-13 Compile the simulation files into the libraries specified in Table 2-1. Table 2-1. Library to Compile Simulation Files Altera Simulation Files Library 220pack lpm 220model lpm sgate pack sgate sgate sgate altera_mf_components altera_mf altera_mf altera_mf stratixiv_hssi_component stratixiv_hssi stratixiv_hssi_atoms stratixiv_hssi user design files work For example, to compile a file into a specific library using ModelSim, right click on the file, select Properties, then click the General tab. In the Compile to library option, select the corresponding library for the file selected. Figure 2-2 shows the ModelSim window compilation of files in a specific library for the Stratix II GX device. Figure 2-2. ModelSim Option to Compile Files in a Specific Library February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-14 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Guidelines to Debug Transceiver-Based Designs Include all the libraries in the search path. Add the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Manager-generated wrapper files (.v or .vhd) and all of the design files to the library. Compile all the library files first, then the design files, and lastly run the simulation. For Verilog simulation, add the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Manager-generated Verilog wrapper files (.v), the Altera library files, and all of the design files. Compile all the library files first, then the simulation model file, followed by the design files. Lastly, run the simulation. These guidelines are further described in "Example 1: Fibre Channel Protocol Application" below. f For more information about functional register transfer level (RTL) simulation or post-fit simulation, refer to the Simulation chapter in volume 3 of the Quartus II Handbook. Guidelines to Debug Transceiver-Based Designs This section provides guidelines to debug transceiver-based designs. If a system failure occurs, the first step is to ensure the functionality of the logic within the FPGA. Use the following information when you observe a system failure. Guidelines to Debug the FPGA Logic and the Transceiver Interface Before checking the functionality in silicon, perform functional simulation to ensure the basic functionality of the RTL and the transceiver-FPGA fabric interface. Understand the limitations of functional simulation. If you intend to simulate timing parameters, consider post-fit simulation. The functional simulation model for transceivers does not model timing-related parameters or uncertainties in the transceiver data path. For example, the PPM difference in the rate matcher clocks (clock rate compensation) or the phase differences between the read and write side of the phase compensation FIFO are not modeled. f For information about functional RTL simulation or post-fit simulation, refer to the Simulation chapter in volume 3 of the Quartus II Handbook. Check whether the compiled design has timing violations in the TimeQuest Timing Analyzer report. Set the appropriate timing constraints on the failing paths. f For information about using the TimeQuest Timing Analyzer, refer to the The Quartus II TimeQuest Timing Analyzer chapter in volume 3 of the Quartus II Handbook. Stratix IV Device Handbook Volume 3 Verify the functionality of the transmitter and receiver data path with serial loopback. Dynamically control the serial loopback through the rx_seriallpbken port. When this signal is asserted, data from the transmitter serializer is looped back to the receiver CDR of the channel. February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Guidelines to Debug Transceiver-Based Designs 2-15 Use SignalTap to verify the behavior of the user logic and the transceiver interface signals. If you have FPGA I/O pins available for debug, you can also use the external logic analyzer to debug the functionality of the device. f For more information, refer to the In-System Debugging Using External Logic Analyzers chapter in volume 3 of the Quartus II Handbook. 1 To use these features, you must connect the JTAG configuration pins in the FPGA. Verify the interconnect on the receive side by configuring the transceiver in reverse serial loopback mode. In this case, the recovered data from the receiver channel is sent to the transmitter buffer. To configure a transceiver channel operating in a different configuration to reverse serial loopback mode, use the dynamic reconfiguration controller. Check whether the transceiver FPGA fabric interface clocking schemes follow the recommendations provided in the "FPGA Fabric-Transceiver Interface Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. Ensure that you have used the recommended transceiver reset sequence. Guidelines to Debug System Level Issues If you have determined that the logic in the FPGA fabric is functionally correct, check for system level issues: Check the voltage ripple across the 2 k resistor that is connected to the RREF pin. The voltage ripple must be less than 60 mv. Measure the eye on the near-end and far-end of the transmitter to understand the jitter added by the transmitter and interconnect. Ensure that the high-speed scopes you use for measurement have sufficient bandwidth (the bandwidth rating on the scope and cables must be at least three times the serial data rate). Check whether the eye meets the eye-mask requirements if specified by the protocol application. Use scopes that provide information on the different jitter components to understand the possible source of the increased jitter. For example, increased intersymbol interface (ISI) indicates potential bandwidth limitations on the interconnect. 1 February 2011 Some scopes, such as Agilent 86100C DCA, require pre-defined patterns (for example, PRBS7 or PRBS23) to provide jitter components. Measure signals on the traces (no connector) using a high-impedance differential probe with short leads. Altera Corporation Stratix IV Device Handbook Volume 3 2-16 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Guidelines to Debug Transceiver-Based Designs Ensure that characteristic impedance on the interconnect matches the source and load systems. Check for impedance discontinuities on the trace by Time Domain Reflectometry (TDR). Revisit the board design, layout, and routing for any inconsistencies that can cause impedance discontinuities. Check whether the termination schemes on the Stratix IV GX device and on the upstream system are matched. Altera recommends using OCT in the Stratix IV GX device instead of external termination to improve signal integrity. Change the transmit output differential voltage to improve eye amplitude. Compensate for high frequency losses in the interconnect by changing the equalization settings of the Stratix IV GX device and check for improvement of the bit error rate. If the upstream system does not have an equalization feature, increase the pre-emphasis (1st post tap) of the Stratix IV GX transmitter. In cases where there are multiple interconnects between the Stratix IV GX device and the upstream system, use the pre-tap and 2nd post tap. Altera provides tools to select the pre-emphasis. Measure the increase in jitter at the near end and far end with one channel turned on at a time if you have multiple transceiver channels connected to the upstream system. This helps to observe the effect of cross talk from adjacent channels on the victim channel. Check the board layout and routing to ensure that you have implemented the design practices to mitigate cross talk. Ensure that the input voltage and duty cycle of the input reference clock source provided to the transmitter PLLs meet the input reference clock requirements. Check whether the voltage drop on the power supplies is within the specified tolerance range. Measure the voltage at the via beneath the power supply pin using a high-impedance probe. Check whether the voltage regulator specifications meet the Stratix IV GX power supply requirements. Revisit the power distribution scheme for the supply voltage to ensure that it is designed to handle the transient current requirements of the transceiver. f For the tolerance values of the different power supplies, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. Check for periodic modulation of other frequency components on the transmit data. Send a high-frequency pattern (1010) from the transmitter side and connect the transmitter serial output to a spectrum analyzer. f For more information about debugging Stratix IV GX transceivers, refer to AN 553: Debugging Transceivers. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-17 Example 1: Fibre Channel Protocol Application Assume that you want to implement a fibre channel protocol application using three transceiver channels. Consider the following system requirements: You need three transceiver channels All the channels need to be placed in the same transceiver block All the channels need to have independent control to reset their PCS and PMA functional blocks Table 2-2 lists the transceiver channel configuration for Example 1. Table 2-2. Transceiver Channel Configuration for Example 1 Channels Mode of Operation Data Rate Input Reference Clock Frequency (MHz) 0 Receiver and Transmitter FC4G (4.25 Gbps) 106.25 1 Receiver and Transmitter FC1G (1.0625 Gbps) 53.125 2 Transmitter Only FC4G (4.25 Gbps) 106.25 Phase 1--Architecture In this phase, check whether the Stratix IV GX device supports or meets your design requirements. Device Specification Consider the questions listed in Table 2-3 before setting device-specific parameters. Table 2-3. Device Specific Parameters Questions Answer Yes Do the parameters meet the fibre channel protocol electrical requirements? For more information, refer to the "Transceiver Performance Characteristics" section in the DC and Switching Characteristics for Stratix IV Devices chapter Are three transceiver channels available? Yes Yes Is there support for 4.25 Gbps and 1.0625 Gbps data rates? Two CMU PLLs are available within each transceiver block to support two different transmitter data rates. Each receiver channel contains a dedicated receiver CDR that supports 4.25 Gbps and 1.0625 Gbps data rates. f For the maximum data rates supported, refer to the "Transceiver Performance Specifications" section in the DC and Switching Characteristics for Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-18 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Transceiver Configuration The fibre channel protocol uses an 8B/10B encoder and requires clock rate compensation. Functional Blocks Consider the questions listed in Table 2-4 before configuring the transceiver. Table 2-4. Configuring the Transceiver Questions Answer No Is the 8B/10B encoder in the PCS block fibre channel compliant? The fibre channel protocol consists of two different End-of-Frame (EOFt) ordered sets. The correct EOFt ordered set sent by the user logic depends on the ending disparity of the word preceeding the EOFt. The Stratix IV GX transceiver does not provide running disparity flags to the user logic. Therefore, the user logic might not be able to select the correct EOFt ordered set. Yes Is there a workaround? Implement the 8B/10B encoder in the FPGA fabric. Is the clock rate compensation block in the PCS available without an 8B/10B encoder? No You can implement this in the FPGA fabric. The design requires a Transmitter and Receiver configuration for two channels and a Transmitter Only configuration for one channel (Table 2-5). Table 2-5. Multiple Channels Questions Answer Yes Does the Stratix IV GX transceiver support these two configurations and allow you to combine them within the same transceiver block The available FPGA fabric interface width is 20 or 40 bits to support 4.25 Gbps and 1.0625 Gbps data rates, respectively. This FPGA fabric interface facilitates 8B/10B encoding and decoding in the FPGA fabric without additional re-arrangement of the received parallel data to a 10-bit boundary. Dynamic Reconfiguration If your application requires you to dynamically reconfigure the transceiver PMA controls, ensure that you understand the settings, options, and user logic required to enable this feature. f For more information, refer to the "Interfacing ALTGX and ALTGX_RECONFIG Instances" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. f For more information about initiating read and write transactions, refer to the "Dynamically Reconfiguring PMA Controls" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-19 If you are using the channel reconfiguration feature, enable the appropriate options in the ALTGX and ALTGX_RECONFIG MegaWizard Plug-In Managers. f You can dynamically use the reconfiguration modes to reconfigure different functional blocks in a transceiver channel using .mifs. For information about generating .mifs, refer to the "Channel and CMU PLL Reconfiguration Mode Details" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Clocking Consider the questions listed in Table 2-6 before configuring clocking. Table 2-6. Configuring Clocking Questions Answer Yes Is there support for two different input reference clocks? Do the refclk pins support the required frequency range? The Stratix IV GX transceiver has two refclk pins for each transceiver block. Yes The minimum frequency range of refclk is 50 MHz; the maximum frequency range is 622.08 MHz. No The design requires independent control on all channels, so you must not share the transceiver-FPGA fabric interface Can transceiver-FPGA fabric interface clocking be shared? clock of one channel with another channel. Each of the channels must use its own tx_clkout and rx_clkout signals to clock the data between the transceiver channels and the FPGA fabric. Does the Stratix IV GX transceiver support this feature? Yes f For more information about clocking the transmitter and receiver channel data path for this type of configuration, refer to the "Transmitter Channel Datapath Clocking" section of the Transceiver Clocking in Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-20 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Figure 2-3 shows the transmitter side of the transceiver setup for Example 1. 1 The transmitter side receives its clocks from the clock multiplier unit (CMU) PLLs. The receiver side contains its dedicated CDR that provides the high-speed serial and low-speed parallel clocks to its PMA and PCS blocks, respectively. Figure 2-3. Top-Level Transceiver Setup--Transmitter-Side Only Transceiver Block Channel 0 (4.25 Gbps) TX RX refclk0 (106.25 MHz) One CMU PLL Configured for 4.25 Gbps Data Rate Channel 1 (1.0625 Gbps) refclk1 (53.125 MHz) Second CMU PLL Configured for 1.0625 Gbps Data Rate TX RX Channel 2 (4.25 Gbps) TX Phase 2--Implementation Create the transceiver instance using the ALTGX MegaWizard Plug-In Manager. f For a description of the individual options, refer to the ALTGX Transceiver Setup Guide for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-21 Create the Transceiver Instance for an FC4G Configuration (Channel 0) Figure 2-4 through Figure 2-14 show the different options available in the ALTGX MegaWizard Plug-In Manager to create the transceiver channel instance for the FC4G data rate. Use this instance for channel 0, with the following settings: General screen--You can configure the Stratix IV GX transceiver for fibre channel protocol using Basic mode. Set the options with the values shown in Figure 2-4. Figure 2-4. FC4G Instance Settings (General Screen) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-22 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application PLL/Ports screen--Check the Train Receiver CDR from PLL inclk option, as shown in Figure 2-5. When you select this option, the same input reference clock used for the CMU PLL is provided as a training clock to the receiver CDR. Figure 2-5. FC4G Instance Settings (PLL/Ports Screen) Check the pll_powerdown signal. This signal allows you to power down the CMU PLL. Use this signal as part of your reset sequence. Check the pll_locked signal. This signal indicates whether the CMU PLL is locked to the input reference clock. The user logic waits until the pll_locked signal goes high before transmitting data. Check the rx_freqlocked signal. This signal indicates whether the receiver CDR is locked to data. When the receiver CDR is configured in automatic lock mode, assert the rx_digitalreset signal if the rx_freqlocked signal goes low to keep the receiver PCS under reset. Altera recommends specific transceiver reset sequences to ensure proper device operation. f For more information about receiver CDR and lock modes, refer to the "Receiver Channel Datapath" section of the Transceiver Architecture in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-23 Ports /Cal Blk screen--The calibration block is required so it is always enabled. Select the options shown in Figure 2-6. Figure 2-6. FC4G Instance Settings (Ports/Cal Blk Screen) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-24 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application RX Analog screen--Select the options shown in Figure 2-7. Figure 2-7. FC4G Instance Settings (RxAnalog Screen) f For a description of the individual options, refer to the ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-25 TX Analog screen--Select the output differential voltage and common mode voltage values that meet the fibre channel protocol specification. If you intend to transmit data through faulty interconnects, select the pre-emphasis settings shown in Figure 2-8. Figure 2-8. FC4G Instance Settings (TX Analog Screen) f For more information about pre-emphasis settings, refer to the DC and Switching Characteristics for Stratix IV Devices chapter. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-26 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Reconfig screen--Set the starting channel number to 0. Because offset cancellation is required for receiver channels, the Offset Cancellation for Receiver Channels option is automatically enabled. Ensure that you connect the reconfig_fromgxb and reconfig_togxb ports with the dynamic reconfiguration controller (Figure 2-9). Figure 2-9. FC4G Instance Settings (Reconfig Screen) f For more information about the starting channel numbers, refer to the "Logical Channel Addressing" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-27 Lpbk screen--The serial loopback option is enabled, as shown in Figure 2-10. Figure 2-10. FC4G Instance Settings (Lpbk Screen) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-28 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Basic/8B10B screen--The Basic/8B10B screen is shown in Figure 2-11. The 8B/10B encoder is not compatible with the fibre channel protocol application; therefore, this option is unchecked. Figure 2-11. FC4G Instance Settings (Basic 8B/10B) Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-29 Word Aligner screen--The fibre channel protocol requires that you use K28.5 to align the byte boundary. In the What is the word alignment pattern? option, set one of the 10-bit disparity values to K28.5. The word aligner automatically detects when the other disparity value is received. Figure 2-12. FC4G Instance Settings (Word Aligner Screen) February 2011 Select the rx_patterndetect and rx_syncstatus signals. The rx_patterndetect signal indicates whenever the word alignment pattern is detected in the word boundary. Click Finish to exit the ALTGX MegaWizard Plug-In Manager. Altera Corporation Stratix IV Device Handbook Volume 3 2-30 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Create the Transceiver Instance for an FC1G Configuration (Channel 1) Creating the instance for FC1G is very similar to that of the FC4G configuration, with the following changes: General screen--Set the values shown in Figure 2-13. Figure 2-13. FC1G Instance (Channel 1) Settings (General Screen) Stratix IV Device Handbook Volume 3 Reconfig screen--Set the starting channel number to 4. February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-31 Create the Instance for an FC4G Configuration--Transmitter Only Mode (Channel 2) This configuration is similar to the channel 0 configuration, with the following changes: Set the operation mode to Transmitter Only, as shown in Figure 2-14. Because this is a Transmitter Only instance, all the options relevant to the receiver are not available in the ALTGX MegaWizard Plug-In Manager. Figure 2-14. FC4G_TXONLY Instance (Channel 1) Settings (General Screen) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 2-32 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Reconfig screen--Set the starting channel number to 8. Select the Analog controls option even if you do not intend to dynamically reconfigure the PMA controls, as shown in Figure 2-15. Selecting this option is required for this example scenario because: For a Transmitter Only instance, offset cancellation is not available; therefore, the reconfig_fromgxb and reconfig_togxb ports are not available. The other two instances (containing a receiver channel) have these ports available because offset cancellation is automatically enabled. If one transceiver instance has the reconfig_fromgxb and reconfig_togxb ports enabled, the Quartus II software requires the other transceiver instances to have these ports enabled to combine them in the same transceiver block. Therefore, for this Transmitter Only instance, the Analog options... must be selected. Figure 2-15. FC4G_TXONLY Instance (Reconfig) Screen f For more information about the requirements to combine multiple transceiver instances, refer to the "Combining Transceiver Instances in Multiple Transceiver Blocks" section in the Configuring Multiple Protocols and Data Rates in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-33 Create the Dynamic Reconfiguration Controller (ALTGX_Reconfig) Instance This section only describes the relevant options that must be set to implement the application. Figure 2-16. ALTGX_Reconfig Settings (Reconfiguration Settings Screen) f For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. Figure 2-16 shows the options that you must set (assuming that you do not require dynamic reconfiguration of the PMA controls in the transceiver channels). f For more information about selecting the Number of Channels option, refer to the "Total Number of Channels Option in the ALTGX_RECONFIG Instance" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Connect the following: February 2011 reconfig_fromgxb[16:0] of the ALTGX_RECONFIG instance to the FC4G instance (channel0) reconfig_fromgxb[33:17] to the FC1G instance (channel1) reconfig_fromgxb[50:34] to the FC4G Transmitter Only instance (channel2) reconfig_togxb[3:0] of the ALTGX_RECONFIG instance to all three transceiver instances Altera Corporation Stratix IV Device Handbook Volume 3 2-34 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Create Reset Logic to Control the FPGA Fabric and Transceivers The design requires independent control on each channel. Altera recommends creating independent reset control logic for each channel. In this design, channel 0 and channel 2 share the same CMU PLL (because they are configured at the same data rate) and channel 1 uses the second CMU PLL. When you create a Transmitter Only or Receiver and Transmitter instance, the ALTGX MegaWizard Plug-In Manager provides a pll_powerdown signal to reset the CMU PLL that provides clocks to the transmitter channel. In this design example, because channels 0 and 2 share the same CMU PLL, drive the pll_powerdown port of channel 0 and channel 2 in the ALTGX instance from the same logic. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-35 Channels 0, 1, and 2 have separate rx_digitalreset, rx_analogreset, and tx_digitalreset signals. Figure 2-17 shows the interface between the three transceiver instances and the FPGA fabric. Figure 2-17. Transceiver--FPGA Fabric Interface User Logic rx_freqlocked Reset Control Logic tx_digitalreset rx_digitalreset rx_analogreset Transmitter Side Logic data 8B/10B processing encoder logic tx_datain ALTGX Instance Channel0 Starting channel number = 0 reconfig_fromgxb[16:0] reconfig_togxb[3:0] Receiver Side Logic data processing logic ALTGX_RECONFIG Instance rx_patterndetect 8B/10B decoder rx_syncstatus rx_dataout reconfig_fromgxb[16:0] reconfig_fromgxb[50:34] pll_locked Reset Control for CMU PLL Reset Control Logic reconfig_fromgxb[33:17] pll_powerdown ALTGX Instance Channel 2 tx_digitalreset reconfig_fromgxb[16:0] reconfig_togxb[3:0] Transmitter Side Logic data processing logic tx_datain 8B/10B encoder Reset Control Logic pll_powerdown pll_locked rx_freqlocked tx_digitalreset rx_digitalreset rx_analogreset Starting channel number = 8 ALTGX Instance Channel 1 reconfig_togxb[3:0] reconfig_fromgxb[16:0] Transmitter Side Logic data processing logic tx_datain 8B/10B encoder Receiver Side Logic data processing logic February 2011 Altera Corporation 8B/10B decoder rx_patterndetect rx_syncstatus Starting channel number = 4 rx_dataout Stratix IV Device Handbook Volume 3 2-36 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application Create Data Processing and Other User Logic For this example, you must implement the 8B/10B encoder and decoder in the FPGA fabric. Figure 2-17 on page 2-35 shows the logic on the transmitter and receiver side and the system logic controls for all channels in the FPGA fabric. This block diagram is a representation of a typical system and may not exactly show the different blocks in a practical application. Interface all the logic blocks with the transceiver. If you would like to add SignalTap for verification, first complete synthesis, then add the transceiver-FPGA fabric or other user logic signals in SignalTap. Lastly, compile the design to generate the .sof. Phase 3--Compilation Assign pins for the input and output signals in your design. The Quartus II software versions 8.1 and earlier do not allow pin assignments for the Stratix IV GX device. Set the OCT values for the transceiver serial pins, add timing constraints for the clocks and data paths in your logic, then compile the design. Phase 4--Simulating the Design To simulate the design, follow the steps outlined in "Functional Simulation" on page 2-12. Document Revision History Table 2-7 lists the revision history for this chapter. Table 2-7. Document Revision History (Part 1 of 2) Date Version February 2011 November 2009 June 2009 March 2009 Stratix IV Device Handbook Volume 3 4.1 4.0 3.1 3.0 Changes Applied new template. Updated chapter title. Minor text edits Added Table 2-3, Table 2-4, Table 2-5, and Table 2-6. Minor text edits. Updated the "Introduction", "Power Supplies", "Transceiver Configuration", "Clocking", "Create Transceiver Instances", "Create Dynamic Reconfiguration Controller Instances", "Create Data Processing and Other User Logic", "Functional Simulation" sections. Added the "Board Design Requirements", "Gear Boxing Logic", "Guidelines to Debug the FPGA Logic and the Transceiver Interface", and "Guidelines to Debug System Level Issues" sections. Added introductory sentences to improve search ability. Add "Power Supplies" on page 2-6 Updated "Dynamic Reconfiguration" on page 2-4 Text edits February 2011 Altera Corporation Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application 2-37 Table 2-7. Document Revision History (Part 2 of 2) Date Version November 2008 May 2008 February 2011 Changes Added "Transceiver Configuration" on page 2-3 Added "Create Dynamic Reconfiguration Controller Instances" on page 2-8 "Dynamic Reconfiguration" on page 2-15 Updated "Create the Instance for an FC4G Configuration--Transmitter Only Mode (Channel 2)" on page 2-28 Added "Create the Dynamic Reconfiguration Controller (ALTGX_Reconfig) Instance" on page 2-30 Updated Figure 2-1, Figure 2-4, Figure 2-5, Figure 2-6, Figure 2-7, Figure 2-8, Figure 2-10, Figure 2-11, Figure 2-12, Figure 2-13, and Figure 2-14 Added Figure 2-9, Figure 2-15, and Figure 2-16 2.0 1.0 Altera Corporation Initial release. Stratix IV Device Handbook Volume 3 2-38 Stratix IV Device Handbook Volume 3 Chapter 2: Transceiver Design Flow Guide for Stratix IV Devices Example 1: Fibre Channel Protocol Application February 2011 Altera Corporation 3. ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices February 2011 SIV53004-3.1 SIV53004-3.1 You can use the ALTGX_RECONFIG MegaWizardTM Plug-In Manager in the Quartus(R) II software to create and modify design files for the Stratix(R) IV device family. This chapter describes the different Quartus II settings for dynamic reconfiguration in the ALTGX_RECONFIG MegaWizard Plug-In Manager. The MegaWizard Plug-In Manager helps you create or modify design files that contain custom megafunction variations. These auto-generated MegaWizard files can then be instantiated in a design file. The MegaWizard Plug-In Manager allows you to specify options for the ALTGX_RECONFIG megafunction. Start the MegaWizard Plug-In Manager using one of the following methods: Choose the MegaWizard Plug-In Manager command (Tools menu). When working in the Block Editor (schematic symbol), open the Edit menu and choose Insert Symbol. The Symbol dialog box appears. In the Symbol dialog box, click MegaWizard Plug-In Manager. Start the stand-alone version of the MegaWizard Plug-In Manager by typing the following command at the command prompt: qmegawiz. Dynamic Reconfiguration This section describes the options available on the individual pages of the ALTGX_RECONFIG MegaWizard Plug-In Manager. 1 The MegaWizard Plug-In Manager provides a warning if any of the settings you choose are illegal. (c) 2011 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 3 February 2011 Feedback Subscribe 3-2 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Figure 3-1 shows the first page of the MegaWizard Plug-In Manager. To generate an ALTGX_RECONFIG custom megafunction variation, select Create a new custom megafunction variation. Click Next. Figure 3-1. MegaWizard Plug-In Manager (Page 1) Figure 3-2 shows the second page of the MegaWizard Plug-In Manager. Select the following options (click Next when you are done): 1. In the list of megafunctions on the left, click the "+" icon beside the I/O item. From the options presented, choose ALTGX_RECONFIG megafunction. 2. From the drop-down menu beside Which device family will you be using?, select Stratix IV. 3. From the radio buttons under Which type of output file do you want to create?, choose your output file format (AHDL, VHDL, or Verilog HDL). 4. In the box beneath What name do you want for the output file?, enter the file name or click the Browse button to search for it. 1 Stratix IV Device Handbook Volume 3 For the design to compile successfully, always enable the dynamic reconfiguration controller for all the ALTGX instances in the design. February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-3 Figure 3-2. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Page 2) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 3-4 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Figure 3-3 shows page 3 of the ALTGX_RECONFIG MegaWizard Plug-In Manager. From the drop-down menu, select the number of channels controlled by the dynamic reconfiguration controller. Figure 3-3. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Reconfiguration Settings) Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-5 Table 3-1 lists the available options on page 3 of the MegaWizard Plug-In Manager for your ALTGX_RECONFIG custom megafunction variation. Select the Match project/default option if you want to change the device Currently selected device family options. Make your selections on page 3, then click Next. Table 3-1. MegaWizard Plug-In Manager Options (Page 3) (Part 1 of 2) ALTGX_RECONFIG Setting Description Reference Determine the highest logical channel address among all the ALTGX instances connected to the ALTGX_RECONFIG instance. Round it up to the next multiple of four and set that number in this option. What is the number of channels controlled by the reconfig controller? Depending on this setting, the ALTGX_RECONFIG MegaWizard Plug-in Manager generates the appropriate signal width for the interface signal (reconfig_fromgxb) between the ALTGX_RECONFIG and the ALTGX instances. It also gives the necessary bus width for all the selected physical media attachment (PMA) signals. "Total Number of Channels Controlled by the ALTGX_RECONFIG Instance" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. Depending on the number of channels set, the resource estimate changes because this is a soft implementation that uses fabric logic resources. The resource estimate is shown in the bottom left of Page 3 of the MegaWizard Plug-in Manager. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 3-6 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Table 3-1. MegaWizard Plug-In Manager Options (Page 3) (Part 2 of 2) ALTGX_RECONFIG Setting Description Reference This feature is always enabled by default: Offset Cancellation for Receiver Channels--After the device powers up, the dynamic reconfiguration controller performs offset cancellation on the receiver portion of all the transceiver channels controlled by it. The "Offset Cancellation" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. These features are available for selection: Analog Controls--Allows dynamic reconfiguration of PMA controls such as Equalization, Pre-emphasis, DC Gain, and voltage offset differential (VOD). Data rate division in TX--Allows dynamic reconfiguration of the transmitter local divider settings to 1, 2, or 4. The transmitter channel data rate is reconfigured based on the local divider settings. Channel and TX PLL select/reconfig--The following features are available under this option: CMU PLL Reconfiguration--Allows you to dynamically reconfigure the clock multiplier unit (CMU) phase-locked loop (PLL) to a different data rate. Channel and CMU PLL reconfiguration--Allows the dynamic reconfiguration of the transceiver channel from one functional mode to another and also the CMU PLL reconfiguration. Channel reconfiguration with TX PLL select-- Allows you to select additional transmitter PLLs for the transceiver channel and reconfigure the functional mode of the channel. Central Control Unit reconfiguration--Allows you to reconfigure bonded mode configurations from one to another. What are the features to be reconfigured by the reconfig controller? What are the features to be reconfigured by the reconfig controller? Stratix IV Device Handbook Volume 3 Adaptive Equalization Control --Allows you to reconfigure the adaptive equalization hardware (AEQ) in the receiver portion of the transceivers. Enable one time mode for a single channel mode is a single stable equalization value is set up and locked for the specified channel by the AEQ hardware. EyeQ control--Allows you to reconfigure the EyeQ hardware in the receiver portion of the transceivers. The "PMA Controls Reconfiguration Mode Details" section, "Data Rate Division in Transmitter Mode Details" section, "CMU PLL Reconfiguration Mode Details'" section, "Channel and CMU PLL Reconfiguration Mode Details" section, "Channel reconfiguration with TX PLL Select Mode Details" section, and the "Adaptive Equalization (AEQ)" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. "EyeQ" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-7 Figure 3-4 shows page 4 of the ALTGX_RECONFIG MegaWizard Plug-In Manager. Figure 3-4. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Analog Controls) February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 3-8 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Table 3-2 lists the available options on page 4 of the MegaWizard Plug-In Manager for your ALTGX_RECONFIG custom megafunction variation. Make your selections on page 4, then click Next. Table 3-2. MegaWizard Plug-In Manager Options (Page 4) (Part 1 of 2) ALTGX_RECONFIG Setting Description This option is applicable only for Analog controls reconfiguration and is available for selection when the number of channels controlled by the ALTGX_RECONFIG instance is more than one. The dynamic reconfiguration controller reconfigures only the channel whose logical Use `logical_channel_add channel address is specified at the logical_channel_address port. ress' port for Analog controls reconfiguration The width of this port is selected by the ALTGX_RECONFIG MegaWizard Plug-In Manager depending on the number of channels controlled by the dynamic reconfiguration controller. The maximum width of the logical_channel_address port is 9 bits. Use the same control signal for all channels This option is available for selection when the number of channels controlled by the ALTGX_RECONFIG instance is more than one. When you enable this option, the dynamic reconfiguration controller writes the same control signals to all the channels connected to it. You cannot select this option if you enable the Use 'logical_channel_address' port for Analog controls reconfiguration option. Stratix IV Device Handbook Volume 3 Reference "Dynamic Reconfiguration Controller Port List" and "Method 1--Using the logical_channel_address Port" sections of the Dynamic Reconfiguration in Stratix IV Devices chapter. Method 2 and Method 3 of the "PMA Controls Reconfiguration Mode Details" section of Dynamic Reconfiguration in Stratix IV Devices chapter. February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-9 Table 3-2. MegaWizard Plug-In Manager Options (Page 4) (Part 2 of 2) ALTGX_RECONFIG Setting Description Reference The PMA control ports available to write various analog settings to the transceiver channels controlled by the dynamic reconfiguration controller are as follows: Write Control tx_vodctrl--VOD; 3 bits per channel tx_preemp_0t--Pre-emphasis control pre-tap; 5 bits per channel tx_preemp_1t--Pre-emphasis control 1st post-tap; 5 bits per channel tx_preemp_2t--Pre-emphasis control 2nd post-tap; 5 bits per channel rx_eqdcgain--Equalizer DC gain; 3 bits per channel rx_eqctrl--Equalizer control; 4 bits per channel These are optional signals. The signal widths are based on the setting you entered for the What is the number of channels controlled by the reconfig controller? option and whether you enabled the Use 'logical_channel_address' port for Analog controls reconfiguration option. The port width is also determined by the Use the same control signal for all channels option. At least one of these PMA control ports must be enabled to configure and use the dynamic reconfiguration controller. The PMA control ports available to read the existing values from the transceiver channels controlled by the dynamic reconfiguration controller are as follows: Read Control tx_vodctrl_out--VOD; 3 bits per channel tx_preemp_0t_out--Pre-emphasis control pre-tap; 5 bits per channel tx_preemp1t_out--Pre-emphasis control 1st post-tap; 5 bits per channel tx_preemp_2t_out--Pre-emphasis control 2nd post-tap; 5 bits per channel rx_eqdcgain_out--Equalizer DC gain; 3 bits per channel rx_eqctrl_out--Equalizer control; 4 bits per channel "Dynamically Reconfiguring PMA Controls" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. These are optional signals. The signal widths are based on the setting you entered for the What is the number of channels controlled by the reconfig controller? option and whether you enabled the Use 'logical_channel_address' port for Analog controls reconfiguration option. The PMA controls are available for selection only if you select the corresponding write control. Read and write transactions cannot be performed simultaneously. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 3-10 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Figure 3-5 shows page 5 of the ALTGX_RECONFIG MegaWizard Plug-In Manager. Figure 3-5. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Channel and TX/PLL Reconfiguration) Table 3-4 lists the available options on page 5 of the MegaWizard Plug-In Manager for your ALTGX_RECONFIG custom megafunction variation. Table 3-3. MegaWizard Plug-In Manager Options (Page 5) (Part 1 of 2) ALTGX_RECONFIG Setting Description Reference Enable continuous write of all the words needed for reconfiguration. For a continuous write operation, select the Enable continuous write of all the words needed for reconfiguration option to pulse the write_all signal once to write an entire memory initialization file (.mif). "Dynamic Reconfiguration Controller Port List" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. What is the read latency of the MIF contents? This option is available only if you have selected the Enable continuous write of all the words needed for reconfiguration option. Enter the desired latency in terms of the reconfig_clk cycles it takes for each .mif word to be present at the reconfig_data port. For more information, refer to Figure 3-6. "Dynamic Reconfiguration Controller Port List" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-11 Table 3-3. MegaWizard Plug-In Manager Options (Page 5) (Part 2 of 2) ALTGX_RECONFIG Setting Description Reference Use `reconfig_address_out' This option is enabled by default when you select the Channel and TX PLL select/reconfig option. The value on reconfig_address_out[5:0] indicates the address associated with the words in the .mif, which contains the dynamic reconfiguration instructions. The dynamic reconfiguration controller automatically increments the address at the end of each .mif write transaction. "Dynamic Reconfiguration Controller Port List" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Use `reconfig_address_en' When high, this optional output status signal indicates that the address used in the .mif write transaction cycle has changed. This signal is asserted when the .mif write transaction is completed (when the busy signal is de-asserted). "Dynamic Reconfiguration Controller Port List" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Use `reset_reconfig_address' When asserted, this optional control signal resets reconfig_address_out (the current reconfiguration address) to 0. "Dynamic Reconfiguration Controller Port List" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. This is an optional control signal. The logical_tx_pll_sel[1:0] signal refers to the logical reference index of the CMU PLL. The functionality of the signal depends on the feature activated, as shown below: Use `logical_tx_pll_sel' Use `logical_tx_pll_sel_en' February 2011 Altera Corporation CMU PLL reconfiguration--The corresponding CMU PLL is reconfigured based on the value at logical_tx_pll_sel[1:0]. Channel and CMU PLL reconfiguration--The corresponding CMU PLL is reconfigured based on the value at this signal. The transceiver channel listens to the CMU PLL selected by logical_tx_pll_sel[1:0]. Channel reconfiguration with TX PLL select-- The transceiver channel listens to the TX PLL selected by logical_tx_pll_sel[1:0]. This is an optional control signal. When you enable this signal, the value set on the logical_tx_pll_sel[1:0] signal is valid only if the logical_tx_pll_sel_en is set to 1. "Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. "Guidelines for logical_tx_pll_sel and logical_tx_pll_sel_en Ports" section in the Dynamic Reconfiguration in Stratix IV Devices chapter. Stratix IV Device Handbook Volume 3 3-12 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Figure 3-6 shows that the read latency of the .mif contents is 2, as it takes two reconfig_clk cycles for the .mif data to become available on the reconfig_data port after providing address on the reconfig_address_out port. Figure 3-6. Read Latency Latency = 2 reconfig_clock reconfig_address_out reconfig_data Address 0 Address 1 Address 2 Invalid Data Data 0 Data 1 Figure 3-7 shows page 6 of the ALTGX_RECONFIG MegaWizard Plug-In Manager. Figure 3-7. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Error Checks/Data Rate Switch) Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-13 Table 3-4 lists the available options on page 6 of the MegaWizard Plug-In Manager for your ALTGX_RECONFIG custom megafunction variation. Make your selections on page 6, then click Next. Table 3-4. MegaWizard Plug-In Manager Options (Page 6) ALTGX_RECONFIG Setting Description Reference Enable illegal mode checking When you select this option, the ALTGX_RECONFIG MegaWizard Plug-In Manager provides the error output port. The dynamic reconfiguration controller detects the error conditions within two reconfig_clk cycles, de-asserts the busy signal, and asserts the error signal for two reconfig_clk cycles. The "Error Indication During Dynamic Reconfiguration" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. Enable self recovery When you select this option, the controller automatically recovers if the operation did not complete within the expected time. The error signal is driven high whenever the controller performs a self recovery. The "Error Indication During Dynamic Reconfiguration" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. Use rate_switch_out port to read out the current data rate division The rate_switch_out[1:0] signal is available when you select Data Rate Division in TX mode. You can read the existing local divider settings of a transmitter channel at this port. The decoding for this signal is listed below: 2'b00--Division of 1 2'b01--Division of 2 The "Data Rate Division in Transmitter Mode Details" mode section in the Dynamic Reconfiguration in Stratix IV Devices chapter. 2'b10--Division of 4 2'b11--Not supported You can read or write the receiver and transmitter settings, only the receiver settings, or only the transmitter settings, based on the value you set at the rx_tx_duplex_sel[1:0] port; Use the rx_tx_duplex_sel port to enable RX only, TX only or duplex configuration 2'b00--Duplex mode 2'b01--RX only mode 2'b10--TX only mode 2'b11--unsupported value (do not use this value) "Dynamically Reconfiguring PMA Controls" section of the Dynamic Reconfiguration in Stratix IV Devices chapter. If you disable the rx_tx_duplex_sel[1:0] port, the dynamic reconfiguration controller reads or writes both the receiver and transmitter settings. February 2011 Altera Corporation Stratix IV Device Handbook Volume 3 3-14 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Figure 3-8 shows page 7 (the Simulation Libraries page) of the MegaWizard Plug-In Manager, which is used for dynamic reconfiguration selection. Make your selections, then click Next. Figure 3-8. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Simulation Libraries) Table 3-5 lists the available option on page 7 of the MegaWizard Plug-In Manager for your ALTGX_RECONFIG custom Megafunction variation. Make your selections on page 7, then click Next. Table 3-5. MegaWizard Plug-In Manager Options (Page 7) ALTGX_RECONFIG Setting Description Reference Generate a netlist for synthesis area and timing estimation Selecting this option generates a netlist file that third-party synthesis tools can use to estimate the timing and resource usage. -- Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration 3-15 Figure 3-9 shows page 8 (the last page) of the MegaWizard Plug-In Manager for the dynamic reconfiguration protocol set up. You can select optional files on this page. After you make your selections, click Finish to generate the files. Figure 3-9. MegaWizard Plug-In Manager--ALTGX_RECONFIG (Summary) Document Revision History Table 3-6 lists the revision history for this chapter. Table 3-6. Document Revision History (Part 1 of 2) Date Version February 2011 November 2009 February 2011 3.1 3.0 Altera Corporation Changes Updated Table 3-1. Applied new template. Updated chapter title. Minor text edits. Updated Table 3-1. Updated Table 3-3. Added Figure 3-6. Made minor text edits. Stratix IV Device Handbook Volume 3 3-16 Chapter 3: ALTGX_RECONFIG Megafunction User Guide for Stratix IV Devices Dynamic Reconfiguration Table 3-6. Document Revision History (Part 2 of 2) Date Version Changes Updated Table 3-3. June 2009 2.1 Added introductory sentences to improve search ability. Minor text edits. March 2009 2.0 Updated screen shots. November 2008 1.0 Added chapter to the Stratix IV Device Handbook Stratix IV Device Handbook Volume 3 February 2011 Altera Corporation Additional Information This chapter provides additional information about the document and Altera. How to Contact Altera To locate the most up-to-date information about Altera products, refer to the following table. Contact Technical support Technical training Product literature Contact Method Address Website www.altera.com/support Website www.altera.com/training Email custrain@altera.com Website www.altera.com/literature Nontechnical support (general) Email nacomp@altera.com (software licensing) Email authorization@altera.com Note to Table: (1) You can also contact your local Altera sales office or sales representative. Typographic Conventions The following table shows the typographic conventions this document uses. Visual Cue Meaning Bold Type with Initial Capital Letters Indicate command names, dialog box titles, dialog box options, and other GUI labels. For example, Save As dialog box. For GUI elements, capitalization matches the GUI. bold type Indicates directory names, project names, disk drive names, file names, file name extensions, software utility names, and GUI labels. For example, \qdesigns directory, D: drive, and chiptrip.gdf file. Italic Type with Initial Capital Letters Indicate document titles. For example, Stratix IV Design Guidelines. Indicates variables. For example, n + 1. italic type Variable names are enclosed in angle brackets (< >). For example, <file name> and <project name>.pof file. Initial Capital Letters Indicate keyboard keys and menu names. For example, the Delete key and the Options menu. "Subheading Title" Quotation marks indicate references to sections in a document and titles of Quartus II Help topics. For example, "Typographic Conventions." September 2012 Altera Corporation Stratix IV Device Handbook Volume 3 Info-2 Additional Information Typographic Conventions Visual Cue Meaning Indicates signal, port, register, bit, block, and primitive names. For example, data1, tdi, and input. The suffix n denotes an active-low signal. For example, resetn. Courier type Indicates command line commands and anything that must be typed exactly as it appears. For example, c:\qdesigns\tutorial\chiptrip.gdf. Also indicates sections of an actual file, such as a Report File, references to parts of files (for example, the AHDL keyword SUBDESIGN), and logic function names (for example, TRI). r An angled arrow instructs you to press the Enter key. 1., 2., 3., and a., b., c., and so on Numbered steps indicate a list of items when the sequence of the items is important, such as the steps listed in a procedure. Bullets indicate a list of items when the sequence of the items is not important. 1 The hand points to information that requires special attention. h The question mark directs you to a software help system with related information. f The feet direct you to another document or website with related information. m The multimedia icon directs you to a related multimedia presentation. c A caution calls attention to a condition or possible situation that can damage or destroy the product or your work. w A warning calls attention to a condition or possible situation that can cause you injury. The envelope links to the Email Subscription Management Center page of the Altera website, where you can sign up to receive update notifications for Altera documents. Stratix IV Device Handbook Volume 3 September 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 101 Innovation Drive San Jose, CA 95134 www.altera.com SIV5V4-5.3 (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as ISO trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and 9001:2008 services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service Registered described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Contents Chapter Revision Dates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Section I. Device Datasheet and Addendum for Stratix IV Devices Chapter 1. DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 Operating Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 Absolute Maximum Ratings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 Recommended Operating Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5 DC Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7 Internal Weak Pull-Up Resistor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11 I/O Standard Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-12 Power Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-15 Switching Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-15 Transceiver Performance Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16 Transceiver Datapath PCS Latency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-47 Core Performance Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-47 Clock Tree Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-47 PLL Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-48 DSP Block Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-50 TriMatrix Memory Block Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-51 Configuration and JTAG Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-52 Temperature Sensing Diode Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-53 Chip-Wide Reset (Dev_CLRn) Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-54 Periphery Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-54 High-Speed I/O Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-54 OCT Calibration Block Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-61 Duty Cycle Distortion (DCD) Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-62 I/O Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-62 Programmable IOE Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-63 Programmable Output Buffer Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-63 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-64 Document Revision History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-67 Chapter 2. Addendum to the Stratix IV Device Handbook Additional Information How to Contact Altera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 Typographic Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum iv Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Contents July 2012 Altera Corporation Chapter Revision Dates The chapters in this document, Stratix IV Device Handbook, were revised on the following dates. Where chapters or groups of chapters are available separately, part numbers are listed. Chapter 1. DC and Switching Characteristics for Stratix IV Devices Revised: July 2012 Part Number: SIV54001-5.3 Chapter 2. Addendum to the Stratix IV Device Handbook Revised: February 2011 Part Number: SIV54002-1.5 July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum vi Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter Revision Dates July 2012 Altera Corporation Section I. Device Datasheet and Addendum for Stratix IV Devices This section includes the following chapters: Chapter 1, DC and Switching Characteristics for Stratix IV Devices Chapter 2, Addendum to the Stratix IV Device Handbook Revision History Refer to each chapter for its own specific revision history. For information on when each chapter was updated, refer to the Chapter Revision Dates section, which appears in the full handbook. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum I-2 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Section I: Device Datasheet and Addendum for Stratix IV Devices July 2012 Altera Corporation 1. DC and Switching Characteristics for Stratix IV Devices July 2012 SIV54001-5.3 SIV54001-5.3 This chapter contains the following sections: "Electrical Characteristics" "Switching Characteristics" "I/O Timing" "Glossary" Electrical Characteristics This chapter covers the electrical and switching characteristics for Stratix(R) IV devices. Electrical characteristics include operating conditions and power consumption. Switching characteristics include transceiver specifications, core, and periphery performance. This chapter also describes I/O timing, including programmable I/O element (IOE) delay and programmable output buffer delay. f For information regarding the densities and packages of devices in the Stratix IV family, refer to the Stratix IV Device Family Overview chapter. Operating Conditions When you use Stratix IV devices, they are rated according to a set of defined parameters. To maintain the highest possible performance and reliability of the Stratix IV devices, you must consider the operating requirements described in this chapter. Stratix IV devices are offered in commercial, industrial, and military grades. Commercial devices are offered in -2 (fastest), -2x, -3, and -4 speed grades. Industrial devices are offered in -1, -2, -3, and -4 speed grades. Military devices are offered in -3 speed grade. For the Stratix IV GT -1 and -2 speed grade specifications, refer to the -2/-2x speed grade column. For the Stratix IV GT -3 speed grade specification, refer to the -3 speed grade column, unless otherwise specified. Absolute Maximum Ratings Absolute maximum ratings define the maximum operating conditions for Stratix IV devices. The values are based on experiments conducted with the devices and theoretical modeling of breakdown and damage mechanisms. The functional operation of the device is not implied for these conditions. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Feedback Subscribe 1-2 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics c Conditions other than those listed in Table 1-1, Table 1-2, and Table 1-3 may cause permanent damage to the device. Additionally, device operation at the absolute maximum ratings for extended periods of time may have adverse effects on the device. Table 1-1. Absolute Maximum Ratings for Stratix IV Devices Symbol Description Minimum Maximum Unit VCC Core voltage and periphery circuitry power supply -0.5 1.35 V VCCPT Power supply for programmable power technology -0.5 1.8 V VCCPGM Configuration pins power supply -0.5 3.75 V VCCAUX Auxiliary supply for the programmable power technology -0.5 3.75 V VCCBAT Battery back-up power supply for design security volatile key register -0.5 3.75 V VCCPD I/O pre-driver power supply -0.5 3.75 V VCCIO I/O power supply -0.5 3.9 V VCC_CLKIN Differential clock input power supply -0.5 3.75 V VCCD_PLL PLL digital power supply -0.5 1.35 V VCCA_PLL PLL analog power supply -0.5 3.75 V VI DC input voltage -0.5 4.0 V IOUT DC output current per pin -25 40 mA TJ Operating junction temperature -55 125 C TSTG Storage temperature (No bias) -65 150 C Minimum Maximum Unit Table 1-2. Transceiver Power Supply Absolute Maximum Ratings for Stratix IV GX Devices Symbol Description VCCA_L Transceiver high voltage power (left side) -0.5 3.75 V VCCA_R Transceiver high voltage power (right side) -0.5 3.75 V VCCHIP_L Transceiver HIP digital power (left side) -0.5 1.35 V VCCHIP_R Transceiver HIP digital power (right side) -0.5 1.35 V VCCR_L Receiver power (left side) -0.5 1.35 V VCCR_R Receiver power (right side) -0.5 1.35 V VCCT_L Transmitter power (left side) -0.5 1.35 V VCCT_R Transmitter power (right side) -0.5 1.35 V VCCL_GXBLn (1) Transceiver clock power (left side) -0.5 1.35 V VCCL_GXBRn (1) Transceiver clock power (right side) -0.5 1.35 V VCCH_GXBLn (1) Transmitter output buffer power (left side) -0.5 1.8 V VCCH_GXBRn (1) Transmitter output buffer power (right side) -0.5 1.8 V Note to Table 1-2: (1) n = 0, 1, 2, or 3. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-3 Table 1-3. Transceiver Power Supply Absolute Maximum Ratings for Stratix IV GT Devices (1) Symbol Description Minimum Maximum Unit VCCA_L Transceiver high voltage power (left side) -0.5 3.75 V VCCA_R Transceiver high voltage power (right side) -0.5 3.75 V VCCHIP_L Transceiver HIP digital power (left side) -0.5 1.35 V VCCHIP_R Transceiver HIP digital power (right side) -0.5 1.35 V VCCR_L Receiver power (left side) -0.5 1.35 V VCCR_R Receiver power (right side) -0.5 1.35 V VCCT_L Transmitter power (left side) -0.5 1.35 V VCCT_R Transmitter power (right side) -0.5 1.35 V VCCL_GXBLn (2) Transceiver clock power (left side) -0.5 1.35 V VCCL_GXBRn (2) Transceiver clock power (right side) -0.5 1.35 V VCCH_GXBLn (2) Transmitter output buffer power (left side) -0.5 1.8 V VCCH_GXBRn (2) Transmitter output buffer power (right side) -0.5 1.8 V Notes to Table 1-3: (1) For the absolute maximum ratings for Stratix IV GT engineering sample (ES1) devices, contact your local Altera sales representative. (2) n = 0, 1, 2, or 3. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-4 Maximum Allowed Overshoot and Undershoot Voltage During transitions, input signals may overshoot to the voltage shown in Table 1-4 and undershoot to -2.0 V for input currents less than 100 mA and periods shorter than 20 ns. Table 1-4 lists the maximum allowed input overshoot voltage and the duration of the overshoot voltage as a percentage of device lifetime. The maximum allowed overshoot duration is specified as a percentage of high time over the lifetime of the device. A DC signal is equivalent to 100% duty cycle. For example, a signal that overshoots to 4.3 V can only be at 4.3 V for ~5% over the lifetime of the device; for a device lifetime of 10 years, this amounts to half of a year. Table 1-4. Maximum Allowed Overshoot During Transitions Symbol Vi (AC) Description Condition (V) Overshoot Duration as % of High Time Unit 4.0 100.000 % 4.05 79.330 % 4.1 46.270 % 4.15 27.030 % AC input voltage 4.2 15.800 % 4.25 9.240 % 4.3 5.410 % 4.35 3.160 % 4.4 1.850 % 4.45 1.080 % 4.5 0.630 % 4.55 0.370 % 4.6 0.220 % Temperature Overshoot Above Maximum Allowed Temperature The maximum allowed operating temperature for Stratix IV industrial grade devices is 100 C. It is recommended that the operating temperature of the device is maintained below 100 C at all times. The temperature excursions over 100 C due to internal heating of the device should not exceed the number of cycles as specified in the Table 1-5. Exceeding the recommended number of cycles may cause solder interconnect failures. Altera(R) recommends using the Stratix IV military grade devices if the application requires operating temperatures over 100 C. Table 1-5. Temperature Overshoot Above Maximum Allowed Temperature Description Device operating temperature (C) July 2012 Altera Corporation Operating Temperature (C) Number of Cycles Over 100 C 100 3200 105 768 110 640 115 480 120 320 125 160 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-5 Recommended Operating Conditions This section lists the functional operation limits for AC and DC parameters for Stratix IV devices. Table 1-6 lists the steady-state voltage and current values expected from Stratix IV devices. Power supply ramps must all be strictly monotonic, without plateaus. f For power supply ripple requirements, refer to the Device-Specific Power Delivery Network (PDN) Tool User Guide. Table 1-6. Recommended Operating Conditions for Stratix IV Devices (Part 1 of 2) Symbol Description Condition Minimum Typical Maximum Unit VCC (Stratix IV GX and Stratix IV E) Core voltage and periphery circuitry power supply -- 0.87 0.90 0.93 V VCC (Stratix IV GT) Core voltage and periphery circuitry power supply -- 0.92 0.95 0.98 V VCCPT Power supply for programmable power technology -- 1.45 1.5 1.55 V VCCAUX Auxiliary supply for the programmable power technology -- 2.375 2.5 2.625 V I/O pre-driver (3.0 V) power supply -- 2.85 3.0 3.15 V I/O pre-driver (2.5 V) power supply -- 2.375 2.5 2.625 V I/O buffers (3.0 V) power supply -- 2.85 3.0 3.15 V I/O buffers (2.5 V) power supply -- 2.375 2.5 2.625 V VCCPD (2) I/O buffers (1.8 V) power supply -- 1.71 1.8 1.89 V I/O buffers (1.5 V) power supply -- 1.425 1.5 1.575 V I/O buffers (1.2 V) power supply -- 1.14 1.2 1.26 V Configuration pins (3.0 V) power supply -- 2.85 3.0 3.15 V Configuration pins (2.5 V) power supply -- 2.375 2.5 2.625 V Configuration pins (1.8 V) power supply -- 1.71 1.8 1.89 V VCCA_PLL PLL analog voltage regulator power supply -- 2.375 2.5 2.625 V VCCD_PLL (Stratix IV GX and Stratix IV E) PLL digital voltage regulator power supply -- 0.87 0.90 0.93 V VCCD_PLL (Stratix IV GT) PLL digital voltage regulator power supply -- 0.92 0.95 0.98 V VCC_CLKIN Differential clock input power supply -- 2.375 2.5 2.625 V Battery back-up power supply (For design security volatile key register) -- 1.2 -- 3.3 V VI DC input voltage -- -0.5 -- 3.6 V VO Output voltage VCCIO VCCPGM VCCBAT (1) TJ (Stratix IV GX and Stratix IV E) TJ (Stratix IV GT) July 2012 Operating junction temperature Operating junction temperature Altera Corporation -- 0 -- VCCIO V Commercial 0 -- 85 C Industrial -40 -- 100 C Military -55 -- 125 C Industrial 0 -- 100 C Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-6 Table 1-6. Recommended Operating Conditions for Stratix IV Devices (Part 2 of 2) Symbol tRAMP Description Power supply ramp time Condition Minimum Typical Maximum Unit Normal POR (PORSEL=0) 0.05 -- 100 ms Fast POR (PORSEL=1) 0.05 -- 4 ms Notes to Table 1-6: (1) If you do not use the volatile security key, you may connect the VCCBAT to either GND or a 3.0-V power supply. (2) VCCPD must be 2.5 V when VCCIO is 2.5, 1.8, 1.5, or 1.2 V. VCCPD must be 3.0 V when VCCIO is 3.0 V. Table 1-7 lists the transceiver power supply recommended operating conditions for Stratix IV GX devices. Table 1-7. Transceiver Power Supply Operating Conditions for Stratix IV GX Devices Symbol Description (1) Minimum Typical Maximum Unit 2.85/2.375 3.0/2.5 (2) 3.15/2.625 V VCCA_L Transceiver high voltage power (left side) VCCA_R Transceiver high voltage power (right side) VCCHIP_L Transceiver HIP digital power (left side) 0.87 0.9 0.93 V VCCHIP_R Transceiver HIP digital power (right side) 0.87 0.9 0.93 V VCCR_L Receiver power (left side) 1.045 1.1 1.155 V VCCR_R Receiver power (right side) 1.045 1.1 1.155 V VCCT_L Transmitter power (left side) 1.045 1.1 1.155 V VCCT_R Transmitter power (right side) 1.045 1.1 1.155 V VCCL_GXBLn (3) Transceiver clock power (left side) 1.05 1.1 1.15 V VCCL_GXBRn (3) Transceiver clock power (right side) 1.05 1.1 1.15 V VCCH_GXBLn (3) Transmitter output buffer power (left side) VCCH_GXBRn (3) Transmitter output buffer power (right side) 1.33/1.425 1.4/1.5 (4) 1.47/1.575 V Notes to Table 1-7: (1) Transceiver power supplies do not have power-on-reset (POR) circuitry. After initial power-up, violating the transceiver power supply operating conditions could lead to unpredictable link behavior. (2) VCCA_L/R must be connected to a 3.0-V supply if the clock multiplier unit (CMU) phase-locked loop (PLL), receiver clock data recovery (CDR), or both, are configured at a base data rate > 4.25 Gbps. For data rates up to 4.25 Gbps, you can connect VCCA_L/R to either 3.0 V or 2.5 V. (3) n = 0, 1, 2, or 3. (4) VCCH_GXBL/R must be connected to a 1.4-V supply if the transmitter channel data rate is > 6.5 Gbps. For data rates up to 6.5 Gbps, you can connect VCCH_GXBL/R to either 1.4 V or 1.5 V. Table 1-8 lists the recommended operating conditions for the Stratix IV GT transceiver power supply. Table 1-8. Transceiver Power Supply Operating Conditions for Stratix IV GT Devices (Part 1 of 2) (1), Symbol Description (2) Minimum Typical Maximum Unit VCCA_L Transceiver high voltage power (left side) 3.17 3.3 3.43 V VCCA_R Transceiver high voltage power (right side) 3.17 3.3 3.43 V VCCHIP_L Transceiver HIP digital power (left side) 0.92 0.95 0.98 V VCCHIP_R Transceiver HIP digital power (right side) 0.92 0.95 0.98 V VCCR_L Receiver power (left side) 1.15 1.2 1.25 V July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-7 Table 1-8. Transceiver Power Supply Operating Conditions for Stratix IV GT Devices (Part 2 of 2) (1), Symbol Description (2) Minimum Typical Maximum Unit VCCR_R Receiver power (right side) 1.15 1.2 1.25 V VCCT_L Transmitter power (left side) 1.15 1.2 1.25 V VCCT_R Transmitter power (right side) 1.15 1.2 1.25 V VCCL_GXBLn (3) Transceiver clock power (left side) 1.15 1.2 1.25 V VCCL_GXBRn (3) Transceiver clock power (right side) 1.15 1.2 1.25 V VCCH_GXBLn (3) Transmitter output buffer power (left side) 1.33 1.4 1.47 V VCCH_GXBRn (3) Transmitter output buffer power (right side) 1.33 1.4 1.47 V Notes to Table 1-8: (1) For the recommended operating conditions for Stratix IV GT engineering sample (ES1) devices, contact your local Altera sales representative. (2) Transceiver power supplies do not have power-on-reset circuitry. After initial power-up, violating the transceiver power supply operating conditions could lead to unpredictable link behavior. (3) n = 0, 1, 2, or 3. DC Characteristics This section lists the supply current, I/O pin leakage current, bus hold, on-chip termination (OCT) tolerance, input pin capacitance, and hot socketing specifications. Supply Current Standby current is the current drawn from the respective power rails used for power budgeting. Use the Excel-based Early Power Estimator (EPE) to get supply current estimates for your design because these currents vary greatly with the resources you use. f For more information about power estimation tools, refer to the PowerPlay Early Power Estimator User Guide and the PowerPlay Power Analysis chapter in the Quartus II Handbook. I/O Pin Leakage Current Table 1-9 lists the Stratix IV I/O pin leakage current specifications. Table 1-9. I/O Pin Leakage Current for Stratix IV Devices Symbol Description Conditions (1) Min Typ Max Unit II Input pin VI = 0V to VCCIOMAX -20 -- 20 A IOZ Tri-stated I/O pin VO = 0V to VCCIOMAX -20 -- 20 A Note to Table 1-9: (1) VREF current refers to the input pin leakage current. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-8 Bus Hold Specifications Table 1-10 lists the Stratix IV device family bus hold specifications. Table 1-10. Bus Hold Parameters VCCIO Parameter Symbol 1.2 V Conditions 1.5 V 1.8 V 2.5 V 3.0 V Unit Min Max Min Max Min Max Min Max Min Max 22.5 -- 25.0 -- 30.0 -- 50.0 -- 70.0 -- A -22.5 -- -25.0 -- -30.0 -- -50.0 -- -70.0 -- A Low sustaining current ISUSL High sustaining current ISUSH Low overdrive current IODL 0V < VIN < VCCIO -- 120 -- 160 -- 200 -- 300 -- 500 A High overdrive current IODH 0V < VIN < VCCIO -- -120 -- -160 -- -200 -- -300 -- -500 A Bus-hold trip point VTRIP -- 0.45 0.95 0.50 1.00 0.68 1.07 0.70 1.70 0.80 2.00 V VIN > VIL (maximum) VIN < VIH (minimum) On-Chip Termination (OCT) Specifications If you enable OCT calibration, calibration is automatically performed at power-up for I/Os connected to the calibration block. Table 1-11 lists the Stratix IV OCT termination calibration accuracy specifications. Table 1-11. OCT Calibration Accuracy Specifications for Stratix IV Devices (Part 1 of 2) (1) Calibration Accuracy Symbol 25- RS (2) 3.0, 2.5, 1.8, 1.5, 1.2 50- RS 3.0, 2.5, 1.8, 1.5, 1.2 50- RT 2.5, 1.8, 1.5, 1.2 20- , 40- , and 60- RS (3) 3.0, 2.5, 1.8, 1.5, 1.2 July 2012 Description Conditions Unit C2 C3,I3, M3 C4,I4 Internal series termination with calibration (25- setting) VCCIO = 3.0, 2.5, 1.8, 1.5, 1.2 V 8 8 8 % Internal series termination with calibration (50- setting) VCCIO = 3.0, 2.5, 1.8, 1.5, 1.2 V 8 8 8 % Internal parallel termination with calibration (50- setting) VCCIO = 2.5, 1.8, 1.5, 1.2 V 10 10 10 % Expanded range for internal series termination with calibration (20- , 40- and 60- RS setting) VCCIO = 3.0, 2.5, 1.8, 1.5, 1.2 V 10 10 10 % Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-9 Table 1-11. OCT Calibration Accuracy Specifications for Stratix IV Devices (Part 2 of 2) (1) Calibration Accuracy Symbol 25- RS_left_shift 3.0, 2.5, 1.8, 1.5, 1.2 Description Internal left shift series termination with calibration (25- RS_left_shift setting) Conditions VCCIO = 3.0, 2.5, 1.8, 1.5, 1.2 V Unit C2 C3,I3, M3 C4,I4 10 10 10 % Notes to Table 1-11: (1) OCT calibration accuracy is valid at the time of calibration only. (2) 25- RS is not supported for 1.5 V and 1.2 V in Row I/O. (3) 20- RS is not supported for 1.5 V and 1.2 V in Row I/O. The calibration accuracy for calibrated series and parallel OCTs are applicable at the moment of calibration. When process, voltage, and temperature (PVT) conditions change after calibration, the tolerance may change. Table 1-12 lists the Stratix IV OCT without calibration resistance tolerance to PVT changes. Table 1-12. OCT Without Calibration Resistance Tolerance Specifications for Stratix IV Devices Resistance Tolerance Symbol 25- RS 3.0 and 2.5 25- RS 1.8 and 1.5 25- RS 1.2 50- RS 3.0 and 2.5 50- RS 1.8 and 1.5 50- RS 1.2 100- RD 2.5 July 2012 Description Conditions Unit C2 C3,I3, M3 C4,I4 Internal series termination without calibration (25- setting) VCCIO = 3.0 and 2.5 V 30 40 40 % Internal series termination without calibration (25- setting) VCCIO = 1.8 and 1.5 V 30 40 40 % Internal series termination without calibration (25- setting) VCCIO = 1.2 V 35 50 50 % Internal series termination without calibration (50- setting) VCCIO = 3.0 and 2.5 V 30 40 40 % Internal series termination without calibration (50- setting) VCCIO = 1.8 and 1.5 V 30 40 40 % Internal series termination without calibration (50- setting) VCCIO = 1.2 V 35 50 50 % Internal differential termination (100- setting) VCCIO = 2.5 V 25 25 25 % Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-10 OCT calibration is automatically performed at power-up for OCT-enabled I/Os. Table 1-13 lists OCT variation with temperature and voltage after power-up calibration. Use Table 1-13 to determine the OCT variation after power-up calibration and Equation 1-1 to determine the OCT variation without re-calibration. Equation 1-1. OCT Variation Without Re-Calibration (1), (2), (3), (4), (5), (6) dR dR R OCT = R SCAL 1 + ------- T ------- V dT dV Notes to Equation 1-1: (1) The ROCT value calculated from Equation 1-1 shows the range of OCT resistance with the variation of temperature and VCCIO. (2) RSCAL is the OCT resistance value at power-up. (3) T is the variation of temperature with respect to the temperature at power-up. (4) V is the variation of voltage with respect to the VCCIO at power-up. (5) dR/dT is the percentage change of RSCAL with temperature. (6) dR/dV is the percentage change of RSCAL with voltage. Table 1-13 lists the OCT variation after the power-up calibration. Table 1-13. OCT Variation after Power-Up Calibration Symbol dR/dV dR/dT Description (1) VCCIO (V) Typical 3.0 0.0297 2.5 0.0344 1.8 0.0499 1.5 0.0744 1.2 0.1241 3.0 0.189 2.5 0.208 1.8 0.266 1.5 0.273 1.2 0.317 OCT variation with voltage without re-calibration OCT variation with temperature without re-calibration Unit %/mV %/C Note to Table 1-13: (1) Valid for VCCIO range of 5% and temperature range of 0 to 85C. Pin Capacitance Table 1-14 lists the Stratix IV device family pin capacitance. Table 1-14. Pin Capacitance for Stratix IV Devices (Part 1 of 2) Symbol Description Value Unit CIOTB Input capacitance on the top and bottom I/O pins 4 pF CIOLR Input capacitance on the left and right I/O pins 4 pF CCLKTB Input capacitance on the top and bottom non-dedicated clock input pins 4 pF CCLKLR Input capacitance on the left and right non-dedicated clock input pins 4 pF July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-11 Table 1-14. Pin Capacitance for Stratix IV Devices (Part 2 of 2) Symbol Description Value Unit COUTFB Input capacitance on the dual-purpose clock output and feedback pins 5 pF CCLK1, CCLK3, CCLK8, and CCLK10 Input capacitance for dedicated clock input pins 2 pF Hot Socketing Table 1-15 lists the hot socketing specifications for Stratix IV devices. Table 1-15. Hot Socketing Specifications for Stratix IV Devices Symbol Description Maximum IIOPIN (DC) DC current per I/O pin 300 A IIOPIN (AC) AC current per I/O pin 8 mA IXCVR-TX (DC) DC current per transceiver TX pin 100 mA IXCVR-RX (DC) DC current per transceiver RX pin 50 mA (1) Note to Table 1-15: (1) The I/O ramp rate is 10 ns or more. For ramp rates faster than 10 ns, |IIOPIN| = C dv/dt, in which C is the I/O pin capacitance and dv/dt is the slew rate. Internal Weak Pull-Up Resistor Table 1-16 lists the weak pull-up resistor values for Stratix IV devices. Table 1-16. Internal Weak Pull-Up Resistor for Stratix IV Devices Symbol RPU Description Value of the I/O pin pull-up resistor before and during configuration, as well as user mode if the programmable pull-up resistor option is enabled. (1), (3) Conditions (V) Value (4) Unit VCCIO = 3.0 5% (2) 25 k VCCIO = 2.5 5% (2) 25 k VCCIO = 1.8 5% (2) 25 k VCCIO = 1.5 5% (2) 25 k VCCIO = 1.2 5% (2) 25 k Notes to Table 1-16: (1) All I/O pins have an option to enable weak pull-up except configuration, test, and JTAG pins. (2) Pin pull-up resistance values may be lower if an external source drives the pin higher than VCCIO. (3) The internal weak pull-down feature is only available for the JTAG TCK pin. The typical value for this internal weak pull-down resistor is approximately 25 k (4) These specifications are valid with 10% tolerances to cover changes over PVT. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-12 I/O Standard Specifications Table 1-17 through Table 1-22 list the input voltage (VIH and VIL), output voltage (VOH and VOL), and current drive characteristics (IOH and IOL) for various I/O standards supported by Stratix IV devices. These tables also show the Stratix IV device family I/O standard specifications. VOL and VOH values are valid at the corresponding IOH and IOL, respectively. For an explanation of terms used in Table 1-17 through Table 1-22, refer to "Glossary" on page 1-64. Table 1-17. Single-Ended I/O Standards I/O Standard VCCIO (V) Min VIL (V) Typ Max Min VIH (V) Max Min VOL (V) VOH (V) Max Min Max IOL (mA) IOH (mA) LVTTL 2.85 3 3.15 -0.3 0.8 1.7 3.6 0.4 2.4 2 -2 LVCMOS 2.85 3 3.15 -0.3 0.8 1.7 3.6 0.2 VCCIO - 0.2 0.1 -0.1 2.5 V 2.375 2.5 2.625 -0.3 0.7 1.7 3.6 0.4 2 1 -1 1.8 V 1.71 1.8 1.89 -0.3 0.35 * VCCIO 0.65 * VCCIO VCCIO + 0.3 0.45 VCCIO 0.45 2 -2 1.5 V 1.425 1.5 1.575 -0.3 0.35 * VCCIO 0.65 * VCCIO VCCIO + 0.3 0.25 * VCCIO 0.75 * VCCIO 2 -2 1.2 V 1.14 1.2 1.26 -0.3 0.35 * VCCIO 0.65 * VCCIO VCCIO + 0.3 0.25 * VCCIO 0.75 * VCCIO 2 -2 3.0-V PCI 2.85 3 3.15 -- 0.3 * VCCIO 0.5 * VCCIO 3.6 0.1 * VCCIO 0.9 * VCCIO 1.5 -0.5 3.0-V PCI-X 2.85 3 3.15 -- 0.35 * VCCIO 0.5 * VCCIO -- 0.1 * VCCIO 0.9 * VCCIO 1.5 -0.5 Table 1-18. Single-Ended SSTL and HSTL I/O Reference Voltage Specifications VCCIO (V) I/O Standard VREF (V) VTT (V) Min Typ Max Min Typ Max Min Typ Max SSTL-2 Class I, II 2.375 2.5 2.625 0.49 * VCCIO 0.5 * VCCIO 0.51 * VCCIO VREF 0.04 VREF VREF + 0.04 SSTL-18 Class I, II 1.71 1.8 1.89 0.833 0.9 0.969 VREF 0.04 VREF VREF + 0.04 SSTL-15 Class I, II 1.425 1.5 1.575 0.47 * VCCIO 0.5 * VCCIO 0.53 * VCCIO 0.47 * VCCIO VREF 0.53 * VCCIO HSTL-18 Class I, II 1.71 1.8 1.89 0.85 0.9 0.95 -- VCCIO/2 -- HSTL-15 Class I, II 1.425 1.5 1.575 0.68 0.75 0.9 -- VCCIO/2 -- HSTL-12 Class I, II 1.14 1.2 1.26 0.47 * VCCIO 0.5 * VCCIO 0.53 * VCCIO -- VCCIO/2 -- July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-13 Table 1-19. Single-Ended SSTL and HSTL I/O Standards Signal Specifications VIL(DC) (V) I/O Standard VIH(DC) (V) VIL(AC) (V) VIH(AC) (V) VOL (V) VOH (V) Iol (mA) Ioh (mA) Min Max Min Max Max Min Max Min SSTL-2 Class I -0.3 VREF 0.15 VREF + 0.15 VCCIO + 0.3 VREF 0.31 VREF + 0.31 VTT 0.57 VTT + 0.57 8.1 -8.1 SSTL-2 Class II -0.3 VREF 0.15 VREF + 0.15 VCCIO + 0.3 VREF 0.31 VREF + 0.31 VTT 0.76 VTT + 0.76 16.2 -16.2 SSTL-18 Class I -0.3 VREF 0.125 VREF + 0.125 VCCIO + 0.3 VREF 0.25 VREF + 0.25 VTT 0.475 VTT + 0.475 6.7 -6.7 SSTL-18 Class II -0.3 VREF 0.125 VREF + 0.125 VCCIO + 0.3 VREF 0.25 VREF + 0.25 0.28 VCCIO 0.28 13.4 -13.4 SSTL-15 Class I -- VREF 0.1 VREF + 0.1 -- VREF 0.175 VREF + 0.175 0.2 * VCCIO 0.8 * VCCIO 8 -8 SSTL-15 Class II -- VREF 0.1 VREF + 0.1 -- VREF 0.175 VREF + 0.175 0.2 * VCCIO 0.8 * VCCIO 16 -16 HSTL-18 Class I -- VREF -0.1 VREF + 0.1 -- VREF 0.2 VREF + 0.2 0.4 VCCIO 0.4 8 -8 HSTL-18 Class II -- VREF 0.1 VREF + 0.1 -- VREF 0.2 VREF + 0.2 0.4 VCCIO 0.4 16 -16 HSTL-15 Class I -- VREF 0.1 VREF + 0.1 -- VREF 0.2 VREF + 0.2 0.4 VCCIO 0.4 8 -8 HSTL-15 Class II -- VREF 0.1 VREF + 0.1 -- VREF 0.2 VREF + 0.2 0.4 VCCIO 0.4 16 -16 HSTL-12 Class I -0.15 VREF 0.08 VREF + 0.08 VCCIO + 0.15 VREF 0.15 VREF + 0.15 0.25* VCCIO 0.75* VCCIO 8 -8 HSTL-12 Class II -0.15 VREF 0.08 VREF + 0.08 VCCIO + 0.15 VREF 0.15 VREF + 0.15 0.25* VCCIO 0.75* VCCIO 16 -16 Table 1-20. Differential SSTL I/O Standards I/O Standard VCCIO (V) VSWING(DC) (V) Min Typ Max Min SSTL-2 Class I, II 2.375 2.5 2.625 0.3 SSTL-18 Class I, II 1.71 1.8 1.89 0.25 SSTL-15 Class I, II 1.425 1.5 1.575 0.2 July 2012 Altera Corporation Max VX(AC) (V) Min VSWING(AC) (V) VOX(AC) (V) Typ Max Min Max Min Typ Max VCCIO + VCCIO/2 0.6 - 0.2 -- VCCIO/2 + 0.2 0.62 VCCIO + 0.6 VCCIO/2 - 0.15 -- VCCIO/2 + 0.15 VCCIO + 0.6 VCCIO/2 0.175 -- VCCIO/2 + 0.175 0.5 VCCIO + 0.6 VCCIO/2 0.125 -- VCCIO/2 + 0.125 -- -- VCCIO/2 -- 0.35 -- -- VCCIO/2 -- Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Electrical Characteristics 1-14 Table 1-21. Differential HSTL I/O Standards I/O Standard VCCIO (V) VDIF(DC) (V) VX(AC) (V) VCM(DC) (V) VDIF(AC) (V) Min Typ Max Min Max Min Typ Max Min Typ Max Min Max HSTL-18 Class I 1.71 1.8 1.89 0.2 -- 0.78 -- 1.12 0.78 -- 1.12 0.4 -- HSTL-15 Class I, II 1.425 1.5 1.575 0.2 -- 0.68 -- 0.9 0.68 -- 0.9 0.4 -- HSTL-12 Class I, II 1.14 1.2 1.26 0.16 VCCIO + 0.3 -- 0.5* VCCIO -- 0.4* VCCIO 0.5* VCCIO 0.6* VCCIO 0.3 VCCIO + 0.48 Table 1-22. Differential I/O Standard Specifications I/O Standard VCCIO (V) (3) Min Typ Max (1), (2) (Part 1 of 2) VID (mV) VICM(DC) (V) Min Condition Max Min Condition VOD (V) Max Min (4) VOCM (V) Typ Max Min Typ (4) Max Transmitter, receiver, and input reference clock pins of high-speed transceivers use PCML I/O standard. For transmitter, receiver, and reference clock I/O pin specifications, refer to Table 1-23 on page 1-16 and Table 1-24 on page 1-25. PCML 2.5 V LVDS (HIO) 2.375 2.5 2.625 100 VCM = 1.25 V -- -- 0.05 (5) 1.05 (5) DMAX 700 Mbps 1.8 0.247 -- 0.6 1.125 1.25 1.375 DMAX > 1.55 ( 0.247 5) 700 Mbps -- 0.6 1.125 1.25 1.375 (5) -- 0.05 DMAX 700 Mbps 1.8 0.247 -- 0.6 1 1.25 1.5 -- 1.05 DMAX> 700 Mbps 1.55 0.247 -- 0.6 1 1.25 1.5 VCM = 1.25 V -- 0.3 -- 1.4 0.1 0.2 0.6 0.5 1.2 1.4 2.375 2.5 2.625 100 VCM = 1.25 V -- 0.3 -- 1.4 0.1 0.2 0.6 0.5 1.2 1.5 Mini-LVDS (HIO) 2.375 2.5 2.625 200 -- 600 0.4 -- 1.325 0.25 -- 0.6 1 1.2 1.4 Mini-LVDS (VIO) 2.375 2.5 2.625 200 -- 600 0.4 -- 1.325 0.25 -- 0.6 1 1.2 1.5 2.375 2.5 2.625 300 -- -- 0.6 DMAX 700 Mbps 1.8 -- -- -- -- -- -- 2.375 2.5 2.625 300 -- -- DMAX > 700 Mbps 1.6 -- -- -- -- -- -- 2.5 V LVDS (VIO) 2.375 2.5 2.625 100 RSDS (HIO) 2.375 2.5 2.625 100 RSDS (VIO) VCM = 1.25 V LVPECL (7) July 2012 Altera Corporation (6) 1 (6) (6) (6) Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-22. Differential I/O Standard Specifications I/O Standard BLVDS (8) VCCIO (V) (3) Min Typ Max (1), (2) 1-15 (Part 2 of 2) VID (mV) VICM(DC) (V) Min Condition Max Min 2.375 2.5 2.625 100 -- -- -- VOD (V) (4) VOCM (V) (4) Condition Max Min Typ Max Min Typ Max -- -- -- -- -- -- -- -- Notes to Table 1-22: (1) Vertical I/O (VIO) is top and bottom I/Os; horizontal I/O (HIO) is left and right I/Os. (2) 1.4-V/1.5-V PCML transceiver I/O standard specifications are described in "Transceiver Performance Specifications" on page 1-16. (3) Differential clock inputs in column I/O are powered by VCC_CLKIN which requires 2.5 V. Differential inputs that are not on clock pins in column I/O are powered by VCCPD which requires 2.5 V. All differential inputs in row I/O banks are powered by VCCPD which requires 2.5V. (4) RL range: 90 RL 110 . (5) The receiver voltage input range for the data rate when DMAX > 700 Mbps is 1.0 V VIN 1.6 V. The receiver voltage input range for the data rate when DMAX 700 Mbps is zero V VIN 1.85 V. (6) The receiver voltage input range for the data rate when DMAX > 700 Mbps is 0.85 V VIN 1.75 V. The receiver voltage input range for the data rate when DMAX 700 Mbps is 0.45 V VIN 1.95 V. (7) Column and row I/O banks support LVPECL I/O standards for input operation only on dedicated clock input pins. (8) For more information about BLVDS interface support in Altera devices, refer to AN522: Implementing Bus LVDS Interfaces in Supported Altera Device Families. Power Consumption Altera offers two ways to estimate power consumption for a design the Excel-based Early Power Estimator and the Quartus(R) II PowerPlay Power Analyzer feature. 1 You typically use the interactive Excel-based Early Power Estimator before designing the FPGA to get a magnitude estimate of the device power. The Quartus II PowerPlay Power Analyzer provides better quality estimates based on the specifics of the design after you complete place-and-route. The PowerPlay Power Analyzer can apply a combination of user-entered, simulation-derived, and estimated signal activities that, when combined with detailed circuit models, yields very accurate power estimates. f For more information about power estimation tools, refer to the PowerPlay Early Power Estimator User Guide and the PowerPlay Power Analysis chapter in the Quartus II Handbook. Switching Characteristics This section provides performance characteristics of Stratix IV core and periphery blocks for commercial, industrial, and military grade devices. The final numbers are based on actual silicon characterization and testing. The numbers reflect the actual performance of the device under worst-case silicon process, voltage, and junction temperature conditions. There are no designations on finalized tables. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-16 Transceiver Performance Specifications This section describes transceiver performance specifications. Table 1-23 lists the Stratix IV GX transceiver specifications. Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 1 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ Max -3 Military (2) and -4 Commercial/Industrial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Min Typ Max Min Typ Unit Max Reference Clock Supported I/O Standards 1.2 V PCML, 1.4 V PCML 1.5 V PCML, 2.5 V PCML, Differential LVPECL Input frequency from REFCLK input pins -- 50 -- 697 50 -- 697 50 -- 637.5 MHz Phase frequency detector (CMU PLL and receiver CDR) -- 50 -- 425 50 -- 325 50 -- 325 MHz Absolute VMAX for a REFCLK pin -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V Operational VMAX for a REFCLK pin -- -- -- 1.5 -- -- 1.5 -- -- 1.5 V Absolute VMIN for a REFCLK pin -- -0.4 -- -- -0.4 -- -- -0.4 -- -- V -- -- -- 0.2 -- -- 0.2 -- -- 0.2 UI Duty cycle -- 45 -- 55 45 -- 55 45 -- 55 % Peak-to-peak differential input voltage -- 200 -- 1600 200 -- 1600 200 -- 1600 mV Spread-spectrum modulating clock frequency PCIe 30 -- 33 30 -- 33 30 -- 33 kHz Spread-spectrum downspread PCIe -- -- -- -- -- -- -- On-chip termination resistors -- -- -- -- -- -- -- VICM (AC coupled) -- VICM (DC coupled) HCSL I/O standard for PCIe reference clock Rise/fall time July 2012 (21) Altera Corporation 0 to -0.5% 100 1100 10% 250 -- 550 0 to -0.5% 100 1100 10% 250 -- (4), LVDS, HCSL 0 to -0.5% 100 1100 10% 550 250 -- mV 550 mV Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-17 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 2 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade -3 Military (2) and -4 Commercial/Industrial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Min Typ Max Min Typ Max Min Typ Max Unit 10 Hz -- -- -50 -- -- -50 -- -- -50 dBc/Hz 100 Hz -- -- -80 -- -- -80 -- -- -80 dBc/Hz 1 KHz -- -- -110 -- -- -110 -- -- -110 dBc/Hz 10 KHz -- -- -120 -- -- -120 -- -- -120 dBc/Hz 100 KHz -- -- -120 -- -- -120 -- -- -120 dBc/Hz 1 MHz -- -- -130 -- -- -130 -- -- -130 dBc/Hz 10 KHz to 20 MHz -- -- 3 -- -- 3 -- -- 3 ps -- -- 2000 1% -- -- 2000 1% -- -- 2000 1% -- Calibration block clock frequency -- 10 -- 125 10 -- 125 10 -- 125 MHz fixedclk clock frequency PCIe Receiver Detect -- 125 -- -- 125 -- -- 125 -- MHz reconfig_clk clock frequency Dynamic reconfiguration clock frequency 2.5/ 37.5 -- 50 2.5/ 37.5 -- 50 2.5/ 37.5 -- 50 -- -- -- -- 2 -- -- 2 -- -- 2 ms -- 1 -- -- 1 -- -- 1 -- -- s 3750 Mbps Transmitter REFCLK Phase Noise Transmitter REFCLK Phase Jitter (rms) for 100 MHz REFCLK (3) RREF Transceiver Clocks Delta time between reconfig_clks (5) (5) (5) (19) Transceiver block minimum power-down (gxb_powerdown) pulse width Receiver Supported I/O Standards 1.4 V PCML, 1.5 V PCML, 2.5 V PCML, LVPECL, LVDS Data rate (Single width, non-PMA Direct) -- 600 -- 3750 600 -- 3750 600 -- Data rate (Double width, non-PMA Direct) -- 1000 -- 8500 1000 -- 6500 1000 -- Data rate (Single width, PMA Direct) -- 600 -- 3250 600 -- 3250 600 Data rate (Double width, PMA Direct) -- 1000 -- 6500 1000 -- 6500 1000 July 2012 Altera Corporation 6375 (22) Mbps -- 3250 Mbps -- 6375 Mbps Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-18 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 3 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) -3 Military (2) and -4 Commercial/Industrial Speed Grade Min Typ Max Min Typ Max Min Typ Max Unit Absolute VMAX for a receiver pin (6) -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V Operational VMAX for a receiver pin -- -- -- 1.5 -- -- 1.5 -- -- 1.5 V Absolute VMIN for a receiver pin -- -0.4 -- -- -0.4 -- -- -0.4 -- -- V Maximum peak-to-peak differential input voltage VID (diff p-p) before device configuration -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V VICM = 0.82 V setting -- -- 2.7 -- -- 2.7 -- -- 2.7 V VICM =1.1 V setting (7) -- -- 1.6 -- -- 1.6 -- -- 1.6 V 100 -- -- 100 -- -- 165 -- -- mV 165 -- -- 165 -- -- 165 -- -- mV Maximum peak-topeak differential input voltage VID (diff p-p) after device configuration Data Rate = 600 Mbps to 5 Gbps Minimum differential eye opening at receiver serial input pins (20) Equalization = 0 DC gain = 0 dB Data Rate > 5 Gbps Equalization = 0 DC gain = 0 dB VICM VICM = 0.82 V setting 820 10% 820 10% 820 10% mV VICM = 1.1 V setting (7) 1100 10% 1100 10% 1100 10% mV Receiver DC Coupling Support Differential on-chip termination resistors July 2012 -- For more information about receiver DC coupling support, refer to the "DCCoupled Links" section in the Transceiver Architecture in Stratix IV Devices chapter. 85 setting 85 20% 85 20% 85 20% 100 setting 100 20% 100 20% 100 20% 120 setting 120 20% 120 20% 120 20% 150- setting 150 20% 150 20% 150 20% Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-19 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 4 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade Min Differential and common mode return loss Typ Max -3 Military (2) and -4 Commercial/Industrial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Min PCIe (Gen 1 and Gen 2), XAUI, HiGig+, CEI SR/LR, Serial RapidIO SR/LR, Typ Max Min Typ Unit Max Compliant -- CPRI LV/HV, OBSAI, SATA 62.5, 100, 125, 200, Programmable PPM detector (8) -- Run length -- -- -- 200 -- -- 200 -- -- 200 UI Programmable equalization (18) -- -- -- 16 -- -- 16 -- -- 16 dB -- -- -- 75 -- -- 75 -- -- 75 s -- 15 -- -- 15 -- -- 15 -- -- s -- -- -- 4000 -- -- 4000 -- -- 4000 ns -- -- -- 4000 -- -- 4000 -- -- 4000 ns tLTR (9) tLTR_LTD_Manual tLTD_Manual tLTD_Auto (10) (11) (12) Receiver CDR 3 dB Bandwidth in lock-to-data (LTD) mode PCIe Gen1 20 - 35 MHz PCIe Gen2 40 - 65 MHz (OIF) CEI PHY at 6.375 Gbps 20 - 35 MHz XAUI 10 - 18 MHz Serial RapidIO 1.25 Gbps 10 - 18 MHz Serial RapidIO 2.5 Gbps 10 - 18 MHz Serial RapidIO 3.125 Gbps 6 - 10 MHz GIGE 6 - 10 MHz SONET OC12 3-6 MHz SONET OC48 14 - 19 Receiver buffer and CDR offset cancellation time (per channel) July 2012 ppm 250, 300, 500, 1000 Altera Corporation -- -- -- 1850 0 -- -- MHz 1850 0 -- -- 18500 recon fig_ clk cycles Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-20 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 5 of 9) Symbol/ Description Conditions -3 Military (2) and -4 Commercial/Industrial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) -2 Commercial Speed Grade Unit Min Typ Max Min Typ Max Min Typ Max DC Gain Setting =0 -- 0 -- -- 0 -- -- 0 -- dB DC Gain Setting =1 -- 3 -- -- 3 -- -- 3 -- dB DC Gain Setting =2 -- 6 -- -- 6 -- -- 6 -- dB DC Gain Setting =3 -- 9 -- -- 9 -- -- 9 -- dB DC Gain Setting =4 -- 12 -- -- 12 -- -- 12 -- dB EyeQ Data Rate -- 600 -- 3250 600 -- 3250 600 -- 3250 Mbps AEQ Data Rate min VID (diff p-p) outer envelope = 600 mV 8B/10B encoded data 2500 -- 6500 2500 -- 6500 -- -- -- Mbps Decision Feedback Equalizer (DFE) Data Rate min VID (diff p-p) outer envelope = 500 mV 3125 -- 6500 3125 -- 6500 -- -- -- Mbps 3750 Mbps Programmable DC gain Transmitter Supported I/O Standards 1.4 V PCML, 1.5 V PCML Data rate (Single width, non-PMA Direct) -- Data rate (Double width, non-PMA Direct) -- Data rate (Single width, PMA Direct) -- Data rate (Double width, PMA Direct) -- (13) VOCM 0.65 V setting Differential on-chip termination resistors July 2012 600 -- 1000 600 -- 1000 -- -- -- 3750 8500 3250 6500 650 -- 600 -- 3750 600 -- 1000 -- 6500 1000 -- 600 -- 3250 600 1000 -- 6500 -- 650 -- 6375 (22) Mbps -- 3250 Mbps 1000 -- 6375 Mbps -- 650 -- mV 85 setting 85 15% 85 15% 85 15% 100 setting 100 15% 100 15% 100 15% 120 setting 120 15% 120 15% 120 15% 150- setting 150 15% 150 15% 150 15% Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-21 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 6 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade Min Differential and common mode return loss Rise time Typ Max -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Min PCIe Gen1 and Gen2 (TX VOD=4), XAUI (TX VOD=6), HiGig+ (TX VOD=6), CEI SR/LR (TX VOD=8), Serial RapidIO SR (VOD=6), Serial RapidIO LR (VOD=8), CPRI LV (VOD=6), CPRI HV (VOD=2), OBSAI (VOD=6), SATA (VOD=4), (14) Typ Max -3 Military (2) and -4 Commercial/Industrial Speed Grade Min Typ Unit Max Compliant -- -- 50 -- 200 50 -- 200 50 -- 200 ps -- 50 -- 200 50 -- 200 50 -- 200 ps XAUI rise time -- 60 -- 130 60 -- 130 60 -- 130 ps XAUI fall time -- 60 -- 130 60 -- 130 60 -- 130 ps Intra-differential pair skew -- -- -- 15 -- -- 15 -- -- 15 ps Intra-transceiver block transmitter channel-to-channel skew x4 PMA and PCS bonded mode Example: XAUI, PCIe x4, Basic x4 -- -- 120 -- -- 120 -- -- 120 ps Inter-transceiver block transmitter channel-to-channel skew x8 PMA and PCS bonded mode Example: PCIe x8, Basic x8 -- -- 500 -- -- 500 -- -- 500 ps Fall time July 2012 (14) Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-22 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 7 of 9) Symbol/ Description Inter-transceiver block skew in Basic (PMA Direct) xN mode (15) Conditions -2 Commercial Speed Grade -3 Military (2) and -4 Commercial/Industrial Speed Grade -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Unit Min Typ Max Min Typ Max Min Typ Max N < 18 channels located across three transceiver blocks with the source CMU PLL located in the center transceiver block -- -- 400 -- -- 400 -- -- 400 ps N 18 channels located across four transceiver blocks with the source CMU PLL located in one of the two center transceiver blocks -- -- 650 -- -- 650 -- -- 650 ps 600 -- 8500 600 -- 6500 600 -- 6375 Mbps CMU0 PLL and CMU1 PLL Supported Data Range -- pll_powerdown minimum pulse width (tpll_powerdown) -- CMU PLL lock time from pll_powerdown de-assertion -- July 2012 Altera Corporation s 1 -- -- 100 -- -- 100 -- -- 100 s Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-23 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 8 of 9) Symbol/ Description Conditions -2 Commercial Speed Grade Min -3 dB Bandwidth Typ Max -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) Min Typ Max -3 Military (2) and -4 Commercial/Industrial Speed Grade Min Typ Unit Max PCIe Gen1 2.5 - 3.5 MHz PCIe Gen2 6-8 MHz (OIF) CEI PHY at 4.976 Gbps 7 - 11 MHz (OIF) CEI PHY at 6.375 Gbps 5 - 10 MHz XAUI 2-4 MHz Serial RapidIO 1.25 Gbps 3 - 5.5 MHz Serial RapidIO 2.5 Gbps 3 - 5.5 MHz Serial RapidIO 3.125 Gbps 2-4 MHz GIGE 2.5 - 4.5 MHz SONET OC12 1.5 - 2.5 MHz SONET OC48 3.5 - 6 MHz ATX PLL (6G) /L = 1 4800-5400 and 6000-6500 4800-5400 and 6000-6500 4800-5400 and 6000-6375 Mbps /L = 2 2400-2700 and 3000-3250 2400-2700 and 3000-3250 2400-2700 and 3000-3187.5 Mbps /L = 4 1200-1350 and 1500-1625 1200-1350 and 1500-1625 1200-1350 and 1500-1593.75 Mbps PCIe Gen 2 1.5 1.5 -- MHz (OIF) CEI PHY at 6.375 Gbps 3 - 4.5 3 - 4.5 -- MHz Supported Data Range (16) -3 dB Bandwidth Transceiver-FPGA Fabric Interface Interface speed (non-PMA Direct) Interface speed (PMA Direct) July 2012 Altera Corporation -- 25 -- 325 25 -- 325 25 -- 250 MHz -- 50 -- 325 50 -- 325 50 -- 325 MHz Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-24 Table 1-23. Transceiver Specifications for Stratix IV GX Devices (Part 9 of 9) Symbol/ Description Conditions Min Digital reset pulse width -3 Commercial/ Industrial and -2x Commercial Speed Grade (1) -2 Commercial Speed Grade Typ -- Max Min Typ Max -3 Military (2) and -4 Commercial/Industrial Speed Grade Min Typ Unit Max Minimum is two parallel clock cycles -- Notes to Table 1-23: (1) The -2x speed grade is the fastest speed grade offered in the following Stratix IV GX devices: EP4SGX70DF29, EP4SGX110DF29, EP4SGX110FF35, EP4SGX230DF29, EP4SGX110FF35, EP4SGX180DF29, EP4SGX230FF35, EP4SGX290FF35, EP4SGX180FF35, EP4SGX290FH29, EP4SGX360FF35, and EPSGX360FH29. (2) Stratix IV GX devices in military speed grade only support selected transceiver configuration up to 3125 Mbps. For more information, contact Altera sales representative. (3) To calculate the REFCLK rms phase jitter requirement at reference clock frequencies other than 100 MHz, use the following formula: REFCLK rms phase jitter at f (MHz) = REFCLK rms phase jitter at 100 MHz * 100/f. (4) Differential LVPECL signal levels must comply to the minimum and maximum peak-to-peak differential input voltage specified in this table. (5) The minimum reconfig_clk frequency is 2.5 MHz if the transceiver channel is configured in Transmitter only mode. The minimum reconfig_clk frequency is 37.5 MHz if the transceiver channel is configured in Receiver only or Receiver and Transmitter mode. For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. (6) The device cannot tolerate prolonged operation at this absolute maximum. (7) You must use the 1.1-V RX VICM setting if the input serial data standard is LVDS. (8) The rate matcher supports only up to 300 parts per million (ppm). (9) Time taken to rx_pll_locked goes high from rx_analogreset de-assertion. Refer to Figure 1-2 on page 1-33. (10) Time for which the CDR must be kept in lock-to-reference (LTR) mode after rx_pll_locked goes high and before rx_locktodata is asserted in manual mode. Refer to Figure 1-2 on page 1-33. (11) Time taken to recover valid data after the rx_locktodata signal is asserted in manual mode. Refer to Figure 1-2 on page 1-33. (12) Time taken to recover valid data after the rx_freqlocked signal goes high in automatic mode. Refer to Figure 1-3 on page 1-33. (13) A GPLL may be required to meet the PMA-FPGA fabric interface timing above certain data rates. For more information, refer to the "Left/Right PLL Requirements in Basic (PMA Direct) Mode" section in the Transceiver Clocking in Stratix IV Devices chapter. (14) The Quartus II software automatically selects the appropriate slew rate depending on the configured data rate or functional mode. (15) For applications that require low transmit lane-to-lane skew, use Basic (PMA Direct) xN to achieve PMA-Only bonding across all channels in the link. You can bond all channels on one side of the device by configuring them in Basic (PMA Direct) xN mode. For more information about clocking requirements in this mode, refer to the "Basic (PMA Direct) Mode Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. (16) The Quartus II software automatically selects the appropriate /L divider depending on the configured data. (17) The maximum transceiver-FPGA fabric interface speed of 265.625 MHz is allowed only in Basic low-latency PCS mode with a 32-bit interface width. For more information, refer to the "Basic Double-Width Mode Configurations" section in the Transceiver Architecture in Stratix IV Devices chapter. (18) Figure 1-1 shows the AC gain curves for each of the 16 available equalization settings. (19) If your design uses more than one dynamic reconfiguration controller (altgx_reconfig) instances to control the transceiver (altgx) channels physically located on the same side of the device AND if you use different reconfig_clk sources for these altgx_reconfig instances, the delta time between any two of these reconfig_clk sources becoming stable must not exceed the maximum specification listed. (20) The differential eye opening specification at the receiver input pins assumes that Receiver Equalization is disabled. If you enable Receiver Equalization, the receiver circuitry can tolerate a lower minimum eye opening, depending on the equalization level. Use H-Spice simulation to derive the minimum eye opening requirement with Receiver Equalization enabled. (21) The rise and fall time transition is specified from 20% to 80%. (22) Stratix IV GX devices in -4 speed grade support Basic mode and deterministic latency mode transceiver configurations up to 6375 Mbps. These configurations are shown in the figures 1-90, 1-92, 1-94, 1-96, and 1-101 in the Transceiver Architecture in Stratix IV Devices chapter. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-25 Figure 1-1 shows the top-to-bottom AC gain curve for equalization settings 0 to 15. Figure 1-1. AC Gain Curves for Equalization Settings 0 to 15 (Bottom to Top) Table 1-24 lists the Stratix IV GT transceiver specifications. Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 1 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade Min Typ Max -2 Industrial Speed Grade Min Typ Max -3 Industrial Speed Grade Min Typ Unit Max Reference Clock Supported I/O Standards 1.2 V PCML, 1.4 V PCML, 1.5 V PCML, 2.5 V PCML, Differential LVPECL (3), LVDS Input frequency from REFCLK input pins -- 50 -- 706.25 50 -- 706.25 50 -- 706.25 MHz Phase frequency detector (CMU PLL and receiver CDR) -- 50 -- 425 50 -- 425 50 -- 425 MHz Absolute VMAX for a REFCLK pin -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V Operational VMAX for a REFCLK pin -- -- -- 1.5 -- -- 1.5 -- -- 1.5 V Absolute VMIN for a REFCLK pin -- -0.3 -- -- -0.3 -- -- -0.3 -- -- V July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-26 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 2 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade -2 Industrial Speed Grade -3 Industrial Speed Grade Min Typ Max Min Typ Max Min Typ Max Unit Rise/fall time -- -- -- 0.2 -- -- 0.2 -- -- 0.2 UI Duty cycle -- 45 -- 55 45 -- 55 45 -- 55 % Peak-to-peak differential input voltage -- 200 -- 1200 200 -- 1200 200 -- 1200 mV On-chip termination resistors -- -- 100 -- -- 100 -- -- 100 -- VICM -- 1200 10% 1200 10% 1200 10% mV 10 Hz -- -- -50 -- -- -50 -- -- -50 dBc/Hz 100 Hz -- -- -80 -- -- -80 -- -- -80 dBc/Hz 1 KHz -- -- -110 -- -- -110 -- -- -110 dBc/Hz 10 KHz -- -- -120 -- -- -120 -- -- -120 dBc/Hz 100 KHz -- -- -120 -- -- -120 -- -- -120 dBc/Hz 1 MHz -- -- -130 -- -- -130 -- -- -130 dBc/Hz 10 KHz to 20 MHz -- -- 3 -- -- 3 -- -- 3 ps -- -- -- 2000 1% -- 2000 1% -- -- 2000 1% -- Calibration block clock frequency -- 10 -- 125 10 -- 125 10 -- 125 MHz reconfig_clk clock frequency Dynamic reconfiguration clock frequency 2.5/ 37.5 -- -- 2.5/ 37.5 -- 50 2.5/ 37.5 -- 50 MHz fixedclk clock frequency PCIe Receiver Detect -- 125 -- -- 125 -- -- 125 -- MHz -- -- -- 2 -- -- 2 -- -- 2 ms -- -- 1 -- -- 1 -- -- 1 -- s 3750 Mbps Transmitter REFCLK Phase Noise Transmitter REFCLK Phase Jitter (rms) for 100 MHz REFCLK (2) RREF Transceiver Clocks Delta time between reconfig_clks (1) (1) (1) (15) Transceiver block minimum (gxb_powerdown) power-down pulse width Receiver Supported I/O Standards Data rate (Single width, non-PMA Direct) July 2012 Altera Corporation 1.4 V PCML, 1.5 V PCML, 2.5 V PCML, LVPECL, LVDS -- 600 -- 3750 600 -- 3750 600 -- Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-27 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 3 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade -2 Industrial Speed Grade Min Typ Max Min Typ Max -3 Industrial Speed Grade Unit Min Typ Max 10312.5 1000 -- 8500 Mbps Data rate (Double width, non-PMA Direct) -- 1000 -- 11300 1000 - Data rate (Single width, PMA Direct) -- 600 - 3250 600 - 3250 600 -- 3250 Mbps Data rate (Double width, PMA Direct) -- 1000 - 6500 1000 - 6500 1000 -- 6500 Mbps Absolute VMAX for a receiver pin (4) -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V Operational VMAX for a receiver pin -- -- -- 1.5 -- -- 1.5 -- -- 1.5 V Absolute VMIN for a receiver pin -- -- -0.4 -- -0.4 -- -- -0.4 -- -- V Maximum peak-to-peak differential input voltage VID (diff p-p) before device configuration -- -- -- 1.6 -- -- 1.6 -- -- 1.6 V VICM = 0.82 V setting -- -- 2.7 -- -- 2.7 -- -- 2.7 V VICM = 1.2 V setting (5) -- -- 1.2 -- -- 1.2 -- -- 1.2 V 85 -- -- 85 -- -- 85 -- -- mV 165 -- -- -- -- -- -- -- -- mV Maximum peak-to-peak differential input voltage VID (diff p-p) after device configuration Minimum differential eye opening at the receiver serial input pins for data rates 10.3125 Gbps. Minimum differential eye opening at the receiver serial input pins for data rates > 10.3125 Gbps. VICM July 2012 Equalization = 0 (6) DC gain = 0 dB Equalization = 0 (6) DC gain = 0 dB VICM = 0.82 V setting 820 10% 820 10% 820 10% mV VICM = 1.2 V setting (5) 1200 10% 1200 10% 1200 10% mV Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-28 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 4 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade Min Differential on-chip termination resistors Differential and common mode return loss Typ Max -2 Industrial Speed Grade Min Typ Max -3 Industrial Speed Grade Min Typ Unit Max 85 setting 85 20% 85 20% 85 20% 100 setting 100 20% 100 20% 100 20% 120 setting 120 20% 120 20% 120 20% 150- setting 150 20% 150 20% 150 20% PCIe (Gen 1 and Gen 2), XAUI, HiGig+, CEI SR/LR, Serial RapidIO SR/LR, Compliant -- CPRI LV/HV, OBSAI, SATA 62.5, 100, 125, 200, Programmable PPM detector (7) -- -- Run length -- -- -- 200 -- -- 200 -- -- 200 UI Programmable equalization -- -- -- 16 -- -- 16 -- -- 16 dB -- -- -- 75 -- -- 75 -- -- 75 s -- 15 -- -- 15 -- -- 15 -- -- s tLTR (8) tLTR_LTD_Manual tLTD_Manual (9) (10) ppm 250, 300, 500, 1000 -- -- -- 4000 -- -- 4000 -- -- 4000 ns (11) -- -- -- 4000 -- -- 4000 -- -- 4000 ns Receiver buffer and CDR offset cancellation time (per channel) -- -- -- 17000 -- -- 17000 -- -- 17000 reconfig_clk DC Gain Setting =0 -- 0 -- -- 0 -- -- 0 -- dB DC Gain Setting =1 -- 3 -- -- 3 -- -- 3 -- dB DC Gain Setting =2 -- 6 -- -- 6 -- -- 6 -- dB DC Gain Setting =3 -- 9 -- -- 9 -- -- 9 -- dB DC Gain Setting =4 -- 12 -- -- 12 -- -- 12 -- dB -- -- -- 4.0 -- -- 4.0 -- -- 4.0 Gbps tLTD_Auto Programmable DC gain EyeQ Max Data Rate July 2012 Altera Corporation cycles Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-29 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 5 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade -2 Industrial Speed Grade -3 Industrial Speed Grade Unit Min Typ Max Min Typ Max Min Typ Max AEQ Data Rate min VID (diff p-p) outer envelope = 600 mV 8B/10B encoded data 2500 -- 6500 2500 -- 6500 -- -- -- Mbps Decision Feedback Equalizer (DFE) Data Rate min VID (diff p-p) outer envelope = 600 mV 3125 -- 6500 3125 -- 6500 -- -- -- Mbps 3750 600 -- 3750 Mbps 10312.5 1000 -- 8500 Mbps Transmitter Supported I/O Standards 1.4 V PCML Data rate (Single width, non-PMA Direct) -- 600 -- 3750 600 -- Data rate (Double width, non-PMA Direct) -- 1000 -- 11300 1000 -- Data rate (Single width, PMA Direct) -- 600 -- 3250 600 -- 3250 600 -- 3250 Mbps Data rate (Double width, PMA Direct) (12) -- 1000 -- 6500 1000 -- 6500 1000 -- 6500 Mbps 0.65 V setting -- 650 -- -- 650 -- -- 650 -- mV VOCM Differential on-chip termination resistors July 2012 85 setting 85 15% 85 15% 85 15% 100 setting 100 15% 100 15% 100 15% 120 setting 120 15% 120 15% 120 15% 150- setting 150 15% 150 15% 150 15% Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-30 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 6 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade Min Differential and common mode return loss Rise time Typ Max -2 Industrial Speed Grade Min PCIe Gen1 and Gen2 (TX VOD=4), XAUI (TX VOD=6), HiGig+ (TX VOD=6), CEI SR/LR (TX VOD=8), Serial RapidIO SR (VOD=6), Serial RapidIO LR (VOD=8), CPRI LV (VOD=6), CPRI HV (VOD=2), OBSAI (VOD=6), SATA (VOD=4), (13) Typ Max -3 Industrial Speed Grade Min Typ Unit Max Compliant -- -- 50 -- 200 50 -- 200 50 -- 200 ps -- 50 -- 200 50 -- 200 50 -- 200 ps XAUI rise time -- 60 -- 130 60 -- 130 60 -- 130 ps XAUI fall time -- 60 -- 130 60 -- 130 60 -- 130 ps Intra-differential pair skew -- -- -- 15 -- -- 15 -- -- 15 ps Intra-transceiver block transmitter channel-to-channel skew x4 PMA and PCS bonded mode Example: XAUI, PCIe, x4, Basic x4 -- -- 120 -- -- 120 -- -- 120 ps Inter-transceiver block transmitter channel-to-channel skew x8 PMA and PCS bonded mode Example: PCIe x8, Basic x8 -- -- 500 -- -- 500 -- -- 500 ps Fall time July 2012 (13) Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-31 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 7 of 8) Symbol/ Description Inter-transceiver block skew in Basic (PMA Direct) xN mode (14) Conditions -1 Industrial Speed Grade -2 Industrial Speed Grade -3 Industrial Speed Grade Unit Min Typ Max Min Typ Max Min Typ Max N < 18 channels located across three transceiver blocks with the source CMU PLL located in the center transceiver block -- -- 400 -- -- 400 -- -- 400 ps N 18 channels located across four transceiver blocks with the source CMU PLL located in one of the two center transceiver blocks -- -- 650 -- -- 650 -- -- 650 ps CMU PLL0 and CMU PLL1 Supported data range -- 600 -- 11300 600 -- 10312.5 600 -- 8500 Mbps CMU PLL lock time from pll_powerdown de-assertion -- -- -- 100 -- -- 100 -- -- 100 s ATX PLL (6G) Supported Data Range /L = 1 4800-5400 and 6000-6500 4800-5400 and 6000-6500 4800-5400 and 6000-6500 Mbps /L = 2 2400-2700 and 3000-3250 2400-2700 and 3000-3250 2400-2700 and 3000-3250 Mbps /L = 4 1200-1350 and 1500-1625 1200-1350 and 1500-1625 1200-1350 and 1500-1625 Mbps -- Mbps ATX PLL (10G) Supported Data Range -- 9900 -- 11300 9900 -- 10312.5 -- 25 -- 325 25 -- 325 25 -- 265.625 MHz -- 50 -- 325 50 -- 325 50 -- 325 MHz Transceiver-FPGA Fabric Interface Interface speed (non-PMA Direct) Interface speed (PMA Direct) July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-32 Table 1-24. Transceiver Specifications for Stratix IV GT Devices (Part 8 of 8) Symbol/ Description Conditions -1 Industrial Speed Grade Min Digital reset pulse width -- Typ Max -2 Industrial Speed Grade Min Typ Max -3 Industrial Speed Grade Min Typ Unit Max Minimum is two parallel clock cycles -- Notes to Table 1-24: (1) The minimum reconfig_clk frequency is 2.5 MHz if the transceiver channel is configured in Transmitter Only mode. The minimum reconfig_clk frequency is 37.5 MHz if the transceiver channel is configured in Receiver only or Receiver and Transmitter mode. For more information, refer to the Dynamic Reconfiguration in Stratix IV Devices chapter. (2) To calculate the REFCLK rms phase jitter requirement at reference clock frequencies other than 100 MHz, use the following formula: REFCLK rms phase jitter at f (MHz) = REFCLK rms phase jitter at 100 MHz * 100/f. (3) Differential LVPECL signal levels must comply to the minimum and maximum peak-to-peak differential input voltage specified in this table. (4) The device cannot tolerate prolonged operation at this absolute maximum. (5) You must use the 1.2-V RXVICM setting if the input serial data standard is LVDS. (6) The differential eye opening specification at the receiver input pins assumes that Receiver Equalization is disabled. If you enable Receiver Equalization, the receiver circuitry can tolerate a lower minimum eye opening, depending on the equalization level. Use H-Spice simulation to derive the minimum eye opening requirement with Receiver Equalization enabled. (7) The rate matcher supports only up to 300 ppm. (8) Time taken to rx_pll_locked goes high from rx_analogreset de-assertion. Refer to Figure 1-2 on page 1-33. (9) Time for which the CDR must be kept in lock-to-reference mode after rx_pll_locked goes high and before rx_locktodata is asserted in manual mode. Refer to Figure 1-2 on page 1-33. (10) Time taken to recover valid data after the rx_locktodata signal is asserted in manual mode. Refer to Figure 1-2 on page 1-33. (11) Time taken to recover valid data after the rx_freqlocked signal goes high in automatic mode. Refer to Figure 1-3 on page 1-33. (12) A GPLL may be required to meet the PMA-FPGA fabric interface timing above certain data rates. For more information, refer to the "Left/Right PLL Requirements in Basic (PMA Direct) Mode" section in the Transceiver Clocking in Stratix IV Devices chapter. (13) The Quartus II software automatically selects the appropriate slew rate depending on the configured data rate or functional mode. (14) For applications that require low transmit lane-to-lane skew, use Basic (PMA Direct) xN to achieve PMA-Only bonding across all channels in the link. You can bond all channels on one side of the device by configuring them in Basic (PMA Direct) xN mode. For more information about clocking requirements in this mode, refer to the "Basic (PMA Direct) Mode Clocking" section in the Transceiver Clocking in Stratix IV Devices chapter. (15) If your design uses more than one dynamic reconfiguration controller (altgx_reconfig) instances to control the transceiver (altgx) channels physically located on the same side of the device AND if you use different reconfig_clk sources for these altgx_reconfig instances, the delta time between any two of these reconfig_clk sources becoming stable must not exceed the maximum specification listed. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-33 Figure 1-2 shows the lock time parameters in manual mode. 1 LTD = Lock-To-Data; LTR = Lock-To-Reference Figure 1-2. Lock Time Parameters for Manual Mode r x_analogreset CDR status LTR LTD r x_pll_locked r x_locktodata Invalid Data Valid data r x_dataout t t LTR LTD_Manual t LTR_LTD_Manual Figure 1-3 shows the lock time parameters in automatic mode. Figure 1-3. Lock Time Parameters for Automatic Mode CDR status LTR LTD r x_freqlocked r x_dataout Invalid Valid data t July 2012 Altera Corporation data LTD_Auto Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-34 Table 1-25 through Table 1-28 lists the typical differential VOD termination settings for Stratix IV GX and GT devices. Table 1-25. Typical VOD Setting, TX Term = 85 VOD Setting (mV) Symbol VOD differential peak-to-peak Typical (mV) 0 1 2 3 4 5 6 7 170 20% 340 20% 510 20% 595 20% 680 20% 765 20% 850 20% 1020 20% 6 7 Table 1-26. Typical VOD Setting, TX Term = 100 VOD Setting (mV) Symbol VOD differential peak-to-peak Typical (mV) 0 1 2 3 4 5 200 20% 400 20% 600 20% 700 20% 800 20% 900 20% 1000 1200 20% 20% Table 1-27. Typical VOD Setting, TX Term = 120 VOD Setting (mV) Symbol VOD differential peak-to-peak Typical (mV) 0 1 2 3 4 5 6 240 20% 480 20% 720 20% 840 20% 960 20% 1080 20% 1200 20% Table 1-28. Typical VOD Setting, TX Term = 150 VOD Setting (mV) Symbol VOD differential peak-to-peak Typical (mV) 0 1 2 3 4 5 300 20% 600 20% 900 20% 1050 20% 1200 20% 1350 20% Table 1-29 lists typical transmitter pre-emphasis levels in dB for the first post tap under the following conditions (low-frequency data pattern [five 1s and five 0s] at 6.25 Gbps). The levels listed in Table 1-29 are a representation of possible pre-emphasis levels under the specified conditions only and that the pre-emphasis levels may change with data pattern and data rate. f To predict the pre-emphasis level for your specific data rate and pattern, run simulations using the Stratix IV HSSI HSPICE models. Table 1-29. Transmitter Pre-Emphasis Levels for Stratix IV Devices (Part 1 of 2) VOD Setting Pre-Emphasis 1st Post-Tap Setting 0 1 2 3 4 5 6 7 0 0 0 0 0 0 0 0 0 1 N/A 0.7 0 0 0 0 0 0 2 N/A 1 0.3 0 0 0 0 0 3 N/A 1.5 0.6 0 0 0 0 0 July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-35 Table 1-29. Transmitter Pre-Emphasis Levels for Stratix IV Devices (Part 2 of 2) VOD Setting Pre-Emphasis 1st Post-Tap Setting 0 1 2 3 4 5 6 7 4 N/A 2 0.7 0.3 0 0 0 0 5 N/A 2.7 1.2 0.5 0.3 0 0 0 6 N/A 3.1 1.3 0.8 0.5 0.2 0 0 7 N/A 3.7 1.8 1.1 0.7 0.4 0.2 0 8 N/A 4.2 2.1 1.3 0.9 0.6 0.3 0 9 N/A 4.9 2.4 1.6 1.2 0.8 0.5 0.2 10 N/A 5.4 2.8 1.9 1.4 1 0.7 0.3 11 N/A 6 3.2 2.2 1.7 1.2 0.9 0.4 12 N/A 6.8 3.5 2.6 1.9 1.4 1.1 0.6 13 N/A 7.5 3.8 2.8 2.1 1.6 1.2 0.6 14 N/A 8.1 4.2 3.1 2.3 1.7 1.3 0.7 15 N/A 8.8 4.5 3.4 2.6 1.9 1.5 0.8 16 N/A N/A 4.9 3.7 2.9 2.2 1.7 0.9 17 N/A N/A 5.3 4 3.1 2.4 1.8 1.1 18 N/A N/A 5.7 4.4 3.4 2.6 2 1.2 19 N/A N/A 6.1 4.7 3.6 2.8 2.2 1.4 20 N/A N/A 6.6 5.1 4 3.1 2.4 1.5 July 2012 21 N/A N/A 7 5.4 4.3 3.3 2.7 1.7 22 N/A N/A 8 6.1 4.8 3.8 3 2 23 N/A N/A 9 6.8 5.4 4.3 3.4 2.3 24 N/A N/A 10 7.6 6 4.8 3.9 2.6 25 N/A N/A 11.4 8.4 6.8 5.4 4.4 3 26 N/A N/A 12.6 9.4 7.4 5.9 4.9 3.3 27 N/A N/A N/A 10.3 8.1 6.4 5.3 3.6 28 N/A N/A N/A 11.3 8.8 7.1 5.8 4 29 N/A N/A N/A 12.5 9.6 7.7 6.3 4.3 30 N/A N/A N/A N/A 11.4 9 7.4 N/A 31 N/A N/A N/A N/A 12.9 10 8.2 N/A Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-36 Table 1-30 lists the Stratix IV GX transceiver jitter specifications for all supported protocols. For protocols supported by Stratix IV GT industrial speed grade devices, refer to the Stratix IV GX -2 commercial speed grade column in Table 1-30. Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions SONET/SDH Transmit Jitter Generation -2 Commercial Speed Grade (2) (Part 1 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Typ Max Min Typ Max Min Typ Max Unit (4) Peak-to-peak jitter at 622.08 Mbps Pattern = PRBS15 -- -- 0.1 -- -- 0.1 -- -- 0.1 UI RMS jitter at 622.08 Mbps Pattern = PRBS15 -- -- 0.01 -- -- 0.01 -- -- 0.01 UI Peak-to-peak jitter at 2488.32 Mbps Pattern = PRBS15 -- -- 0.1 -- -- 0.1 -- -- 0.1 UI RMS jitter at 2488.32 Mbps Pattern = PRBS15 -- -- 0.01 -- -- 0.01 -- -- 0.01 UI SONET/SDH Receiver Jitter Tolerance (4) Jitter frequency = 0.03 KHz > 15 > 15 > 15 UI > 1.5 > 1.5 > 1.5 UI > 0.15 > 0.15 > 0.15 UI > 15 > 15 > 15 UI > 1.5 > 1.5 > 1.5 UI > 0.15 > 0.15 > 0.15 UI > 0.15 > 0.15 > 0.15 UI Pattern = PRBS15 Jitter tolerance at 622.08 Mbps Jitter frequency = 25 KHZ Pattern = PRBS15 Jitter frequency = 250 KHz Pattern = PRBS15 Jitter frequency = 0.06 KHz Pattern = PRBS15 Jitter frequency = 100 KHZ Jitter tolerance at 2488.32 Mbps Pattern = PRBS15 Jitter frequency = 1 MHz Pattern = PRBS15 Jitter frequency = 10 MHz Pattern = PRBS15 Fibre Channel Transmit Jitter Generation (5), (13) Total jitter FC-1 Pattern = CRPAT -- -- 0.23 -- -- 0.23 -- -- 0.23 UI Deterministic jitter FC-1 Pattern = CRPAT -- -- 0.11 -- -- 0.11 -- -- 0.11 UI Total jitter FC-2 Pattern = CRPAT -- -- 0.33 -- -- 0.33 -- -- 0.33 UI Deterministic jitter FC-2 Pattern = CRPAT -- -- 0.2 -- -- 0.2 -- -- 0.2 UI July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-37 Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ Max (2) (Part 2 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Typ Max Unit Max Total jitter FC-4 Pattern = CRPAT -- -- 0.52 -- -- 0.52 -- -- 0.52 UI Deterministic jitter FC-4 Pattern = CRPAT -- -- 0.33 -- -- 0.33 -- -- 0.33 UI Fibre Channel Receiver Jitter Tolerance (5), (14) Deterministic jitter FC-1 Pattern = CJTPAT > 0.37 > 0.37 > 0.37 UI Random jitter FC-1 Pattern = CJTPAT > 0.31 > 0.31 > 0.31 UI Fc/25000 > 1.5 > 1.5 > 1.5 UI Fc/1667 > 0.1 > 0.1 > 0.1 UI Pattern = CJTPAT > 0.33 > 0.33 > 0.33 UI Sinusoidal jitter FC-1 Deterministic jitter FC-2 Random jitter FC-2 Pattern = CJTPAT > 0.29 > 0.29 > 0.29 UI Fc/25000 > 1.5 > 1.5 > 1.5 UI Fc/1667 > 0.1 > 0.1 > 0.1 UI Deterministic jitter FC-4 Pattern = CJTPAT > 0.33 > 0.33 > 0.33 UI Random jitter FC-4 Pattern = CJTPAT > 0.29 > 0.29 > 0.29 UI Fc/25000 > 1.5 > 1.5 > 1.5 UI Fc/1667 > 0.1 > 0.1 > 0.1 UI Sinusoidal jitter FC-2 Sinusoidal jitter FC-4 XAUI Transmit Jitter Generation (6) Total jitter at 3.125 Gbps Pattern = CJPAT -- -- 0.3 -- -- 0.3 -- -- 0.3 UI Deterministic jitter at 3.125 Gbps -- -- 0.17 -- -- 0.17 -- -- 0.17 UI Pattern = CJPAT XAUI Receiver Jitter Tolerance (6) Total jitter -- > 0.65 > 0.65 > 0.65 UI Deterministic jitter -- > 0.37 > 0.37 > 0.37 UI Peak-to-peak jitter Jitter frequency = 22.1 KHz > 8.5 > 8.5 > 8.5 UI Peak-to-peak jitter Jitter frequency = 1.875 MHz > 0.1 > 0.1 > 0.1 UI Peak-to-peak jitter Jitter frequency = 20 MHz > 0.1 > 0.1 > 0.1 UI PCIe Transmit Jitter Generation (7) Total jitter at 2.5 Gbps (Gen1) Compliance pattern -- -- 0.25 -- -- 0.25 -- -- 0.25 UI Total jitter at 5 Gbps (Gen2) (15) Compliance pattern -- -- 0.25 -- -- 0.25 -- -- -- UI PCIe Receiver Jitter Tolerance Total jitter at 2.5 Gbps (Gen1) July 2012 Altera Corporation (7) Compliance pattern > 0.6 > 0.6 > 0.6 UI Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-38 Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ Total jitter at 5 Gbps (Gen2) Compliance pattern Max (2) (Part 3 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Compliant Typ Max Compliant Unit Max -- UI PCIe (Gen 1) Electrical Idle Detect Threshold VRX-IDLE-DETDIFFp-p (16) Compliance pattern (peak-to-peak) Total jitter (peak-to-peak) Data Rate = 1.25, 2.5, 3.125 Gbps Combined deterministic and random jitter tolerance (peak-to-peak) 175 65 -- 175 65 -- 175 UI -- -- 0.17 -- -- 0.17 -- -- 0.17 UI -- -- 0.35 -- -- 0.35 -- -- 0.35 UI Pattern = CJPAT Data Rate = 1.25, 2.5, 3.125 Gbps Pattern = CJPAT Serial RapidIO Receiver Jitter Tolerance Deterministic jitter tolerance (peak-to-peak) -- (8) Serial RapidIO Transmit Jitter Generation Deterministic jitter 65 (8) Data Rate = 1.25, 2.5, 3.125 Gbps > 0.37 > 0.37 > 0.37 UI > 0.55 > 0.55 > 0.55 UI > 8.5 > 8.5 > 8.5 UI > 0.1 > 0.1 > 0.1 UI > 0.1 > 0.1 > 0.1 UI Pattern = CJPAT Data Rate = 1.25, 2.5, 3.125 Gbps Pattern = CJPAT Jitter Frequency = 22.1 KHz Data Rate = 1.25, 2.5, 3.125 Gbps Pattern = CJPAT Jitter Frequency = 1.875 MHz Sinusoidal jitter tolerance (peak-to-peak) Data Rate = 1.25, 2.5, 3.125 Gbps Pattern = CJPAT Jitter Frequency = 20 MHz Data Rate = 1.25, 2.5, 3.125 Gbps Pattern = CJPAT GIGE Transmit Jitter Generation (9) Deterministic jitter (peak-to-peak) Pattern = CRPAT -- -- 0.14 -- -- 0.14 -- -- 0.14 UI Total jitter (peak-to-peak) Pattern = CRPAT -- -- 0.279 -- -- 0.279 -- -- 0.279 UI July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-39 Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ GIGE Receiver Jitter Tolerance Max (2) (Part 4 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Typ Max Unit Max (9) Deterministic jitter tolerance (peak-to-peak) Pattern = CJPAT > 0.4 > 0.4 > 0.4 UI Combined deterministic and random jitter tolerance (peak-to-peak) Pattern = CJPAT > 0.66 > 0.66 > 0.66 UI HiGig Transmit Jitter Generation (10) Deterministic jitter (peak-to-peak) Data Rate = 3.75 Gbps Total jitter (peak-to-peak) Data Rate = 3.75 Gbps Pattern = CJPAT Pattern = CJPAT HiGig Receiver Jitter Tolerance -- -- 0.17 -- -- -- -- -- -- UI -- -- 0.35 -- -- -- -- -- -- UI > 0.37 -- -- -- -- -- -- UI > 0.65 -- -- -- -- -- -- UI > 8.5 -- -- -- -- -- -- UI > 0.1 -- -- -- -- -- -- UI > 0.1 -- -- -- -- -- -- UI -- -- 0.3 -- -- 0.3 UI -- -- >0.675 UI (10) Deterministic jitter tolerance (peak-to-peak) Data Rate = 3.75 Gbps Combined deterministic and random jitter tolerance (peak-to-peak) Data Rate = 3.75 Gbps Pattern = CJPAT Pattern = CJPAT Jitter Frequency = 22.1 KHz Data Rate = 3.75 Gbps Pattern = CJPAT Sinusoidal jitter tolerance (peak-to-peak) Jitter Frequency = 1.875MHz Data Rate = 3.75 Gbps Pattern = CJPAT Jitter Frequency = 20 MHz Data Rate = 3.75 Gbps Pattern = CJPAT (OIF) CEI Transmitter Jitter Generation Total jitter (peak-to-peak) Data Rate = 6.375 Gbps Pattern = PRBS15 BER = 10-12 (OIF) CEI Receiver Jitter Tolerance Deterministic jitter tolerance (peak-to-peak) July 2012 Altera Corporation (11) -- -- 0.3 (11) Data Rate = 6.375 Gbps Pattern = PRBS31 BER = 10-12 > 0.675 > 0.675 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-40 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ Combined deterministic and random jitter tolerance (peak-to-peak) Data Rate = 6.375 Gbps Pattern=PRBS31 Max (2) (Part 5 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Typ Max Unit Max > 0.988 > 0.988 -- -- >0.988 UI >5 >5 -- -- >5 UI > 0.05 > 0.05 -- -- > 0.05 UI > 0.05 > 0.05 -- -- > 0.05 UI BER = 10-12 Jitter Frequency = 38.2 KHz Data Rate = 6.375 Gbps Pattern = PRBS31 BER = 10-12 Jitter Frequency = 3.82 MHz Sinusoidal jitter tolerance (peak-to-peak) Data Rate = 6.375 Gbps Pattern = PRBS31 BER = 10-12 Jitter Frequency = 20 MHz Data Rate= 6.375 Gbps Pattern = PRBS31 BER = 10-12 SDI Transmitter Jitter Generation Alignment jitter (peak-to-peak) (12) Data Rate = 1.485 Gbps (HD) Pattern = Color Bar Low-Frequency Roll-Off = 100 KHz 0.2 -- -- 0.2 -- -- 0.2 -- -- UI Data Rate = 2.97 Gbps (3G) Pattern = Color Bar Low-Frequency Roll-Off = 100 KHz 0.3 -- -- 0.3 -- -- 0.3 -- -- UI Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-41 Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ SDI Receiver Jitter Tolerance Max (2) (Part 6 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Typ Max Unit Max (12) Jitter Frequency = 15 KHz Data Rate = 2.97 Gbps (3G) Pattern = Single Line Scramble Color Bar Sinusoidal jitter tolerance (peak-to-peak) >2 >2 >2 UI > 0.3 > 0.3 > 0.3 UI > 0.3 > 0.3 > 0.3 UI >1 >1 >1 UI > 0.2 > 0.2 > 0.2 UI > 0.2 > 0.2 > 0.2 UI Jitter Frequency = 100 KHz Data Rate = 2.97 Gbps (3G) Pattern = Single Line Scramble Color Bar Jitter Frequency = 148.5 MHz Data Rate = 2.97 Gbps (3G) Pattern = Single Line Scramble Color Bar Jitter Frequency = 20 KHz Data Rate = 1.485 Gbps (HD) Pattern = 75% Color Bar Sinusoidal jitter tolerance (peak-to-peak) Jitter Frequency = 100 KHz Data Rate = 1.485 Gbps (HD) Pattern = 75% Color Bar Jitter Frequency = 148.5 MHz Data Rate = 1.485 Gbps (HD) Pattern = 75% Color Bar SAS Transmit Jitter Generation (17) Total jitter at 1.5 Gbps (G1) Pattern = CJPAT -- -- 0.55 -- -- 0.55 -- -- 0.55 UI Deterministic jitter at 1.5 Gbps (G1) Pattern = CJPAT -- -- 0.35 -- -- 0.35 -- -- 0.35 UI Total jitter at 3.0 Gbps (G2) Pattern = CJPAT -- -- 0.55 -- -- 0.55 -- -- 0.55 UI Deterministic jitter at 3.0 Gbps (G2) Pattern = CJPAT -- -- 0.35 -- -- 0.35 -- -- 0.35 UI Total jitter at 6.0 Gbps (G3) Pattern = CJPAT -- -- 0.25 -- -- 0.25 -- -- 0.25 UI July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-42 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Random jitter at 6.0 Gbps (G3) Conditions Pattern = CJPAT SAS Receiver Jitter Tolerance -2 Commercial Speed Grade (2) (Part 7 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Typ Max Min Typ Max Min Typ Max -- 0.15 -- -- 0.15 -- 0.15 -- -- Unit UI (17) Total Jitter tolerance at 1.5 Gbps (G1) Pattern = CJPAT > 0.65 > 0.65 > 0.65 UI Deterministic Jitter tolerance at 1.5 Gbps (G1) Pattern = CJPAT > 0.35 > 0.35 > 0.35 UI > 0.1 > 0.1 > 0.1 UI Sinusoidal Jitter tolerance at 1.5 Gbps (G1) Jitter Frequency = 900 KHz to 5 MHz Pattern = CJTPAT BER = 1E-12 CPRI Transmit Jitter Generation (18) E.6.HV, E.12.HV Pattern = CJPAT Total Jitter E.6.LV, E.12.LV, E.24.LV, E.30.LV -- -- 0.279 -- -- 0.279 -- -- 0.279 UI -- -- 0.35 -- -- 0.35 -- -- 0.35 UI -- -- 0.14 -- -- 0.14 -- -- 0.14 UI -- -- 0.17 -- -- 0.17 -- -- 0.17 UI Pattern = CJTPAT E.6.HV, E.12.HV Pattern = CJPAT Deterministic Jitter E.6.LV, E.12.LV, E.24.LV, E.30.LV Pattern = CJTPAT CPRI Receiver Jitter Tolerance Total jitter tolerance Deterministic jitter tolerance Total jitter tolerance (18) E.6.HV, E.12.HV Pattern = CJPAT E.6.HV, E.12.HV Pattern = CJPAT E.6.LV, E.12.LV, E.24.LV, E.30.LV > 0.66 > 0.66 > 0.66 UI > 0.4 > 0.4 > 0.4 UI > 0.65 > 0.65 > 0.65 UI > 0.37 > 0.37 > 0.37 UI > 0.55 > 0.55 > 0.55 UI Pattern = CJTPAT Deterministic jitter tolerance Combined deterministic and random jitter tolerance E.6.LV, E.12.LV, E.24.LV, E.30.LV Pattern = CJTPAT E.6.LV, E.12.LV, E.24.LV, E.30.LV Pattern = CJTPAT Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-43 Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions OBSAI Transmit Jitter Generation (Part 8 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Unit Min Typ Max Min Typ Max Min Typ Max -- -- 0.35 -- -- 0.35 -- -- 0.35 UI -- -- 0.17 -- -- 0.17 -- -- 0.17 UI (19) Total jitter at 768 Mbps, 1536 Mbps, and 3072 Mbps REFCLK = 153.6MHz Deterministic jitter at 768 Mbps, 1536 Mbps, and 3072 Mbps REFCLK = 153.6MHz Pattern = CJPAT Pattern = CJPAT OBSAI Receiver Jitter Tolerance -2 Commercial Speed Grade (2) (19) Deterministic jitter tolerance at 768 Mbps, 1536 Mbps, and 3072 Mbps Pattern = CJPAT > 0.37 > 0.37 > 0.37 UI Combined deterministic and random jitter tolerance at 768 Mbps, 1536 Mbps, and 3072 Mbps Pattern = CJPAT > 0.55 > 0.55 > 0.55 UI Jitter Frequency = 5.4 KHz > 8.5 > 8.5 > 8.5 UI > 0.1 > 0.1 > 0.1 UI > 8.5 > 8.5 > 8.5 UI > 0.1 > 0.1 > 0.1 UI Sinusoidal Jitter tolerance at 768 Mbps Pattern = CJPAT Jitter Frequency = 460 MHz to 20 MHz Pattern = CJPAT Jitter Frequency = 10.9 KHz Sinusoidal Jitter tolerance at 1536 Mbps Pattern = CJPAT Jitter Frequency = 921.6 MHz to 20 MHz Pattern = CJPAT July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-44 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-30. Transceiver Block Jitter Specifications for Stratix IV GX Devices (1), Symbol/ Description Conditions -2 Commercial Speed Grade Min Typ Jitter Frequency = 21.8 KHz Sinusoidal Jitter tolerance at 3072 Mbps Max (2) (Part 9 of 9) -3 Commercial/ Industrial and -2x Commercial Speed Grade -3 Military (3) and -4 Commercial/ Industrial Speed Grade Min Min Typ Typ Max Unit Max > 8.5 > 8.5 > 8.5 UI > 0.1 > 0.1 > 0.1 UI Pattern = CJPAT Jitter Frequency = 1843.2 MHz to 20 MHz Pattern = CJPAT Notes to Table 1-30: (1) Dedicated refclk pins were used to drive the input reference clocks. (2) The Jitter numbers are valid for the stated conditions only. (3) Stratix IV GX devices in military speed grade only support selected transceiver configuration up to 3125 Mbps. For more information, contact Altera sales representative. (4) The jitter numbers for SONET/SDH are compliant to the GR-253-CORE Issue 3 Specification. (5) The jitter numbers for Fibre Channel are compliant to the FC-PI-4 Specification revision 6.10. (6) The jitter numbers for XAUI are compliant to the IEEE802.3ae-2002 Specification. (7) The jitter numbers for PCI Express (PIPE) (PCIe) are compliant to the PCIe Base Specification 2.0. (8) The jitter numbers for Serial RapidIO are compliant to the RapidIO Specification 1.3. (9) The jitter numbers for GIGE are compliant to the IEEE802.3-2002 Specification. (10) The jitter numbers for HiGig are compliant to the IEEE802.3ae-2002 Specification. (11) The jitter numbers for (OIF) CEI are compliant to the OIF-CEI-02.0 Specification. (12) The HD-SDI and 3G-SDI jitter numbers are compliant to the SMPTE292M and SMPTE424M Specifications. (13) The fibre channel transmitter jitter generation numbers are compliant to the specification at T interoperability point. (14) The fibre channel receiver jitter tolerance numbers are compliant to the specification at R interoperability point. (15) You must use the ATX PLL adjacent to the transceiver channels to meet the transmitter jitter generation compliance in PCIe Gen2 x8 modes. (16) Stratix IV PCIe receivers are compliant to this specification provided the VTX-CM-DC-ACTIVEIDLE-DELTA of the upstream transmitter is less than 50mV. (17) The jitter numbers for Serial Attached SCSI (SAS) are compliant to the SAS-2.1 Specification. (18) The jitter numbers for CPRI are compliant to the CPRI Specification V3.0. (19) The jitter numbers for OBSAI are compliant to the OBSAI RP3 Specification V4.1. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-45 Table 1-31 lists the transceiver jitter specifications for protocols supported by Stratix IV GT devices. Table 1-31. Transceiver Jitter Specifications for Protocols by Stratix IV GT Devices (Part 1 of 2) Symbol/ Description Conditions XLAUI/CAUI Transmit Jitter Generation Total Jitter -1 Industrial Speed Grade -2 Industrial Speed Grade -3 Industrial Speed Grade Unit Min Typ Max Min Typ Max Min Typ Max -- -- 0.30 -- -- 0.30 -- -- 0.30 UI -- -- 0.17 -- -- 0.17 -- -- 0.17 UI (1), (3) Pattern = PRBS-31 VOD = 800 mV Deterministic Jitter REFCLK = 644.53 MHz 4 (XLAUI)/ 10 (CAUI) channels in Basic x1 mode XLAUI/CAUI Receiver Jitter Tolerance Total Jitter tolerance (1) Pattern = PRBS-31 > 0.62 > 0.62 -- UI >5 >5 -- UI > 0.05 > 0.05 -- UI Jitter Frequency = 40 KHz Pattern = PRBS-31 Equalization = Disabled Sinusoidal Jitter tolerance BER = 1E-12 Jitter Frequency 4 MHz Pattern = PRBS-31 Equalization = Disabled BER = 1E-12 XFI Transmitter Jitter Generation (2), (3) Pattern = PRBS-31 Total jitter at 10.3125 Gbps Vod = 800 mV REFCLK = 644.53 MHz -- -- 0.3 -- -- 0.3 -- -- -- UI -- -- 0.30 -- -- 0.30 -- -- 0.30 UI -- -- 0.17 -- -- 0.17 -- -- 0.17 UI 10 channels in Basic x1 mode OTL 4.10 (1), (3) Total Jitter at 11.18 Gbps Pattern = PRBS-31 Deterministic Jitter REFCLK = 698.75 MHz July 2012 VOD = 800 mV Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-46 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-31. Transceiver Jitter Specifications for Protocols by Stratix IV GT Devices (Part 2 of 2) Symbol/ Description Conditions -1 Industrial Speed Grade Min Typ Max -2 Industrial Speed Grade Min Typ Max -3 Industrial Speed Grade Min Typ Unit Max Jitter Frequency = 40 KHz Pattern = PRBS-31 Equalization = Disabled Sinusoidal Jitter tolerance >5 >5 -- UI > 0.05 > 0.05 -- UI BER = 1E-12 Jitter Frequency 4 MHz Pattern = PRBS-31 Equalization = Disabled BER = 1E-12 Notes to Table 1-31: (1) The jitter numbers for XLAUI/CAUI are compliant to the IEEE P802.3ba specification. (2) Stratix IV GT transceivers are compliant to the XFI datacom transmitter jitter specifications in Table 9 of XFP Revision 4.1. (3) Contact Altera for board and link best practices at BER = 1E-15. Table 1-32 lists the SFI-S transmitter jitter specifications for Stratix IV GT devices. Table 1-32. SFI-S Transmitter Jitter Specifications for Stratix IV GT Devices Symbol/Description Conditions (1), (2) -1 Industrial Speed Grade -2 Industrial Speed Grade -3 Industrial Speed Grade Mean Mean Mean -- -- Unit Pattern = PRBS-31 Total Transmitter jitter at 11.3 Gbps (4) Vod = 800 mV REFCLK = 706.25 MHz 0.23 UI (3) UI 12 channels in Basic x1 mode Notes to Table 1-32: (1) Dedicated refclk pins were used to drive the input reference clocks. (2) The jitter numbers are valid for stated conditions only. (3) Two hundred channels were characterized to derive the mean transmitter jitter specification of 0.23 UI. The maximum jitter across the 200 units characterized was 0.30 UI. (4) Contact Altera for board and link best practices at BER = 1E-15. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-47 Transceiver Datapath PCS Latency f For more information about: Basic mode PCS latency, refer to Figure 1-90 through Figure 1-97 in the Transceiver Architecture in Stratix IV Devices chapter. PCIe mode PCS latency, refer to Figure 1-102 in the Transceiver Architecture in Stratix IV Devices chapter. XAUI mode PCS latency, refer to Figure 1-119 in the Transceiver Architecture in Stratix IV Devices chapter. GIGE mode PCS latency, refer to Figure 1-128 in the Transceiver Architecture in Stratix IV Devices chapter. SONET/SDH mode PCS latency, refer to Figure 1-136 in the Transceiver Architecture in Stratix IV Devices chapter. SDI mode PCS latency, refer to Figure 1-141 in the Transceiver Architecture in Stratix IV Devices chapter. (OIF) CEI PHY mode PCS latency, refer to Figure 1-143 in the Transceiver Architecture in Stratix IV Devices chapter. Core Performance Specifications This section describes the clock tree, phase-locked loop (PLL), digital signal processing (DSP), TriMatrix, configuration, JTAG, and chip-wide reset (Dev_CLRn) specifications. For the Stratix IV GT -1 and -2 speed grade specifications, refer to the -2/-2x speed grade column. For the Stratix IV GT -3 speed grade specification, refer to the -3 speed grade column, unless otherwise specified. Clock Tree Specifications Table 1-33 lists the clock tree specifications for Stratix IV devices. Table 1-33. Clock Tree Performance for Stratix IV Devices Performance Unit Symbol July 2012 -2/-2x Speed Grade -3 Speed Grade -4 Speed Grade Global clock and Regional clock 800 700 500 MHz Periphery clock 550 500 500 MHz Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-48 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics PLL Specifications Table 1-34 lists the Stratix IV PLL specifications when operating in the commercial (0 to 85C), industrial (-40 to 100C), and military (-55C to 125C) junction temperature ranges. Table 1-34. PLL Specifications for Stratix IV Devices (Part 1 of 2) Symbol Parameter Input clock frequency (-2/-2x speed grade) fIN fINPFD fVCO (2) tEINDUTY fOUT fOUT_EXT Min 5 Typ -- Max Unit 800 (1) MHz MHz MHz Input clock frequency (-3 speed grade) 5 -- 717 (1) Input clock frequency (-4 speed grade) 5 -- 717 (1) Input frequency to the PFD 5 -- 325 MHz PLL VCO operating range (-2 speed grade) 600 -- 1600 MHz PLL VCO operating range (-3 speed grade) 600 -- 1300 MHz PLL VCO operating range (-4 speed grade) 600 -- 1300 MHz Input clock or external feedback clock input duty cycle 40 -- 60 % Output frequency for internal global or regional clock (-2/-2x speed grade) -- -- 800 (3) MHz Output frequency for internal global or regional clock (-3 speed grade) -- -- 717 (3) MHz Output frequency for internal global or regional clock (-4 speed grade) -- -- 717 (3) MHz Output frequency for external clock output (-2 speed grade) -- -- 800 (3) MHz 717 (3) MHz 717 (3) MHz Output frequency for external clock output (-3 speed grade) Output frequency for external clock output (-4 speed grade) -- -- -- -- tOUTDUTY Duty cycle for external clock output (when set to 50%) 45 50 55 % tFCOMP External feedback clock compensation time -- -- 10 ns tCONFIGPLL Time required to reconfigure scan chain -- 3.5 -- scanclk cycles tCONFIGPHASE Time required to reconfigure phase shift -- 1 -- scanclk cycles fSCANCLK scanclk frequency -- -- 100 MHz tLOCK Time required to lock from end-of-device configuration or de-assertion of areset -- -- 1 ms tDLOCK Time required to lock dynamically (after switchover or reconfiguring any non-post-scale counters/delays) -- -- 1 ms PLL closed-loop low bandwidth -- 0.3 -- MHz -- 1.5 -- MHz -- 4 -- MHz fCLBW PLL closed-loop medium bandwidth PLL closed-loop high bandwidth (8) tPLL_PSERR Accuracy of PLL phase shift -- -- 50 ps tARESET Minimum pulse width on the areset signal 10 -- -- ns tINCCJ (4), (5) tOUTPJ_DC (6) Input clock cycle to cycle jitter (FREF 100 MHz) -- -- 0.15 UI (p-p) Input clock cycle to cycle jitter (FREF < 100 MHz) -- -- 750 ps (p-p) Period Jitter for dedicated clock output (FOUT 100 MHz) -- -- 175 ps (p-p) Period Jitter for dedicated clock output (FOUT < 100 MHz) -- -- 17.5 mUI (p-p) Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-49 Table 1-34. PLL Specifications for Stratix IV Devices (Part 2 of 2) Symbol tOUTCCJ_DC tOUTPJ_IO (6) (6), (9) tOUTCCJ_IO (6), (9) tCASC_OUTPJ_DC (6), (7) fDRIFT Parameter Min Typ Max Unit Cycle to Cycle Jitter for dedicated clock output (FOUT 100 MHz) -- -- 175 ps (p-p) Cycle to Cycle Jitter for dedicated clock output (FOUT < 100 MHz) -- -- 17.5 mUI (p-p) Period Jitter for clock output on regular I/O (FOUT 100 MHz) -- -- 600 ps (p-p) Period Jitter for clock output on regular I/O (FOUT < 100 MHz) -- -- 60 mUI (p-p) Cycle to Cycle Jitter for clock output on regular I/O (FOUT 100 MHz) -- -- 600 ps (p-p) Cycle to Cycle Jitter for clock output on regular I/O (FOUT < 100 MHz) -- -- 60 mUI (p-p) Period Jitter for dedicated clock output in cascaded PLLs (FOUT 100MHz) -- -- 250 ps (p-p) Period Jitter for dedicated clock output in cascaded PLLs (FOUT < 100MHz) -- -- 25 mUI (p-p) Frequency drift after PFDENA is disabled for duration of 100 us -- -- 10 % Notes to Table 1-34: (1) This specification is limited in the Quartus II software by the I/O maximum frequency. The maximum I/O frequency is different for each I/O standard. (2) The VCO frequency reported by the Quartus II software in the PLL summary section of the compilation report takes into consideration the VCO post-scale counter K value. Therefore, if the counter K has a value of 2, the frequency reported can be lower than the fVCO specification. (3) This specification is limited by the lower of the two: I/O FMAX or FOUT of the PLL. (4) A high input jitter directly affects the PLL output jitter. To have low PLL output clock jitter, you must provide a clean clock source that is less than 120 ps. (5) FREF is fIN/N when N = 1. (6) Peak-to-peak jitter with a probability level of 10-12 (14 sigma, 99.99999999974404% confidence level). The output jitter specification applies to the intrinsic jitter of the PLL, when an input jitter of 30 ps is applied. The external memory interface clock output jitter specifications use a different measurement method and are available in Table 1-51 on page 1-62. (7) The cascaded PLL specification is only applicable with the following condition: A. Upstream PLL: 0.59Mhz Upstream PLL BW < 1 MHz B. Downstream PLL: Downstream PLL BW > 2 MHz (8) High bandwidth PLL settings are not supported in external feedback mode. (9) External memory interface clock output jitter specifications use a different measurement method, which is available in Table 1-49 on page 1-61. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-50 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics DSP Block Specifications Table 1-35 lists the Stratix IV DSP block performance specifications. Table 1-35. Block Performance Specifications for Stratix IV DSP Devices (1) Resources Used Performance Number of Multipliers -1 Industrial and-2/-2x Commercial/ Industrial Speed Grade 9x9-bit multiplier (A, C, E, G) (2) 1 520 460 460 400 400 MHz 9x9-bit multiplier (B, D, F, H) (2) 1 520 460 460 400 400 MHz 1 540 500 500 440 440 MHz 1 540 500 500 440 430 MHz 18x18-bit multiplier 1 600 550 550 480 480 MHz 36x36-bit multiplier 1 480 440 440 380 380 MHz 18x18-bit multiply accumulator 4 490 440 440 380 380 MHz 18x18-bit multiply adder 4 510 470 470 410 400 MHz 18x18-bit multiply adder-signed full precision 2 490 450 440 390 390 MHz 18x18-bit multiply adder with loopback (4) 2 390 350 350 310 300 MHz 36-bit shift (32-bit data) 1 490 440 440 380 380 MHz Double mode 1 480 440 440 380 370 MHz Mode 12x12-bit multiplier (A, E) (3) 12x12-bit multiplier (B, D, F, H) (3) -3 -3 -4 -4 Commercial Industrial Commercial Industrial Speed Speed Speed Speed Grade Grade Grade Grade Unit Notes to Table 1-35: (1) Maximum is for fully pipelined block with Round and Saturation disabled. (2) The DSP block implements eight independent 9b9b multiplies using A, B, C, D for the top DSP half block and E, F, G, H for the bottom DSP half block multipliers. (3) The DSP block implements six independent 12b12b multiplies using A, B, D for the top DSP half block and E, F, H for the bottom DSP half block multipliers. (4) Maximum for loopback input registers disabled, Round and Saturation disabled, and pipeline and output registers enabled. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-51 TriMatrix Memory Block Specifications Table 1-36 lists the Stratix IV TriMatrix memory block specifications. Table 1-36. TriMatrix Memory Block Performance Specifications for Stratix IV Devices (1) (Part 1 of 2) Resources Used Memory MLAB (3) M9K Block (3) July 2012 Mode Performance -3 -1 Industrial -3 -4 -4 Industrial/ and -2 /-2x Commercial/ Industrial Commercial/ Military TriMatrix ALUTs Commercial/ Industrial/ Speed Industrial Speed Memory Industrial Military Grade Grade Speed Grade (2) Speed Grade Speed Grade (2) Unit Single port 64x10 0 1 600 500 450 500 450 MHz Simple dual-port 32x20 0 1 600 500 450 500 450 MHz Simple dual-port 64x10 0 1 600 500 450 500 450 MHz ROM 64x10 0 1 600 500 450 500 450 MHz ROM 32x20 0 1 600 500 450 500 450 MHz Single-port 256x36 0 1 600 540 475 540 475 MHz Simple dual-port 256x36 0 1 550 490 420 490 420 MHz Simple dual-port 256x36, with the read-during-write option set to Old Data 0 1 375 340 300 340 300 MHz True dual port 512x18 0 1 490 430 370 430 370 MHz True dual-port 512x18, with the read-during-write option set to Old Data 0 1 375 335 290 335 290 MHz ROM 1 Port 0 1 600 540 475 540 475 MHz ROM 2 Port 0 1 600 540 475 540 475 MHz Min Pulse Width (clock high time) -- -- 750 800 850 800 850 ps Min Pulse Width (clock low time) -- -- 500 625 690 625 690 ps Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-52 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-36. TriMatrix Memory Block Performance Specifications for Stratix IV Devices (1) (Part 2 of 2) Resources Used Memory M144K Block (3) Performance -3 -4 -1 Industrial -3 -4 Industrial/ Industrial and -2 /-2x Commercial/ TriMatrix Commercial/ Military Speed ALUTs Commercial/ Industrial/ Memory Industrial Speed Grade Industrial Military Speed Grade Grade (2) Speed Grade Speed Grade (2) Mode Unit Single-port 4Kx36 0 1 475 440 380 400 350 MHz Simple dual-port 2Kx72 0 1 465 435 385 375 325 MHz Simple dual-port 2Kx72, with the read-during-write option set to Old Data 0 1 260 240 205 225 200 MHz Simple dual-port 2Kx64 (with ECC) 0 1 335 300 255 295 250 MHz True dual-port 4Kx36 0 1 400 375 330 350 310 MHz True dual-port 4Kx36, with the read-during-write option set to Old Data 0 1 245 230 205 225 200 MHz ROM 1 Port 0 1 540 500 435 450 420 MHz ROM 2 Port 0 1 500 465 400 425 400 MHz Min Pulse Width (clock high time) -- -- 700 755 860 860 950 ps Min Pulse Width (clock low time) -- -- 500 625 690 690 690 ps Notes to Table 1-36: (1) To achieve the maximum memory block performance, use a memory block clock that comes through global clock routing from an on-chip PLL set to 50% output duty cycle. Use the Quartus II software to report timing for this and other memory block clocking schemes. (2) This is only applicable to the Stratix IV E and GX devices. (3) When you use the error detection CRC feature, there is no degradation in FMAX. Configuration and JTAG Specifications Table 1-37 lists the Stratix IV configuration mode specifications. Table 1-37. Configuration Mode Specifications for Stratix IV Devices DCLK FMAX Programming Mode Passive serial Fast passive parallel Fast active serial Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum (1) Unit Min Typ Max -- -- 125 MHz -- -- 125 MHz 17 26 40 MHz July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-53 Table 1-37. Configuration Mode Specifications for Stratix IV Devices DCLK FMAX Programming Mode Remote update only in fast AS mode Unit Min Typ Max 4.3 5.3 10 MHz Note to Table 1-37: (1) This denotes the maximum frequency supported in the FPP configuration scheme. The frequency supported for each device may vary depending on device density. For more information, refer to the Configuration, Design Security, and Remote System Upgrades in Stratix IV Devices chapter. Table 1-38 lists the JTAG timing parameters and values for Stratix IV devices. Table 1-38. JTAG Timing Parameters and Values for Stratix IV Devices Symbol Description Min Max Unit tJCP TCK clock period 30 -- ns tJCH TCK clock high time 14 -- ns tJCL TCK clock low time 14 -- ns tJPSU (TDI) TDI JTAG port setup time 1 -- ns tJPSU (TMS) TMS JTAG port setup time 3 -- ns tJPH JTAG port hold time 5 -- ns tJPCO JTAG port clock to output -- tJPZX JTAG port high impedance to valid output tJPXZ -- JTAG port valid output to high impedance -- 11 (1) ns 14 (1) ns 14 (1) ns Note to Table 1-38: (1) A 1 ns adder is required for each VCCIO voltage step down from 3.0 V. For example, tJPCO = 12 ns if VCCIO of the TDO I/O bank = 2.5 V, or 13 ns if it equals 1.8 V. Temperature Sensing Diode Specifications Table 1-39 lists the specifications for the Stratix IV temperature sensing diode. Table 1-39. External Temperature Sensing Diode Specifications for Stratix IV Devices Description Min Typ Max Unit Ibias, diode source current 8 -- 500 A Vbias, voltage across diode 0.3 -- 0.9 V Series resistance -- -- <5 1.026 1.028 1.030 -- Diode ideality factor Table 1-40 lists the specifications for the Stratix IV internal temperature sensing diode. Table 1-40. Internal Temperature Sensing Diode Specifications for Stratix IV Devices Temperature Range Accuracy Offset Calibrated Option Sampling Rate Conversion Time Resolution Minimum Resolution with No Missing Codes -40 to 100 C 8 C No Frequency: 500 kHz, 1 MHz < 100 ms 8 bits 8 bits July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-54 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Chip-Wide Reset (Dev_CLRn) Specifications Table 1-41 lists the specifications for the Stratix IV chip-wide reset (Dev_CLRn). This specifications denote the minimum pulse width of the Dev_CLRn signal required to clear all the device registers. Table 1-41. Chip-Wide Reset (DEV_CLRn) Specifications Description Dev_CLRn Min Typ Max Unit 500 -- -- s Periphery Performance This section describes periphery performance, including high-speed I/O and external memory interface. I/O performance supports several system interfaces, such as the LVDS high-speed I/O interface, external memory interface, and the PCI/PCI-X bus interface. General-purpose I/O standards such as 3.3-, 2.5-, 1.8-, and 1.5-LVTTL/LVCMOS are capable of typical 167 MHz and 1.2 LVCMOS at 100 MHz interfacing frequency with 10 pF load. For the Stratix IV GT -1 and -2 speed grade specifications, refer to the -2/-2x speed grade column. For the Stratix IV GT -3 speed grade specification, refer to the -3 speed grade column, unless otherwise specified. 1 Actual achievable frequency depends on design- and system-specific factors. You must perform HSPICE/IBIS simulations based on your specific design and system setup to determine the maximum achievable frequency in your system. High-Speed I/O Specification Table 1-42 lists the high-speed I/O timing for Stratix IV devices. Table 1-42. High-Speed I/O Specifications (1), (2) (Part 1 of 3) -2/-2x Speed Grade Symbol fHSCLK_in (input clock frequency) True Differential I/O Standards fHSCLK_in (input clock frequency) Single Ended I/O Standards (10) fHSCLK_in (input clock frequency) Single Ended I/O Standards (11) fHSCLK_OUT (output clock frequency) -4 Speed Grade -3 Speed Grade Conditions Unit Min Typ Max Min Typ Max Min Typ Max 5 -- 800 5 -- 717 5 -- 717 MHz 5 -- 800 5 -- 717 5 -- 717 MHz (3) 5 -- 520 5 -- 420 5 -- 420 MHz -- 5 -- 5 -- 5 -- Clock boost factor W = 1 to 40 (3) Clock boost factor W = 1 to 40 (3) Clock boost factor W = 1 to 40 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 800 (8) 717 (8) 717 (8) July 2012 Altera Corporation MHz Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-42. High-Speed I/O Specifications (1), (2) 1-55 (Part 2 of 3) -2/-2x Speed Grade Symbol -3 Speed Grade -4 Speed Grade Conditions Unit Min Typ Max Min Typ Max Min Typ Max (4) -- 1600 (4) -- 1250 (4) -- 1250 Mbps SERDES factor J = 2, Uses DDR Registers (4) -- (5) (4) -- (5) (4) -- (5) Mbps SERDES factor J = 1, Uses an SDR Register (4) -- (5) (4) -- (5) (4) -- (5) Mbps (4) -- 1250 (4) -- 1152 (4) -- 800 Mbps (4) -- 311 (4) -- 200 (4) -- 200 Mbps Total Jitter for Data Rate, 600 Mbps to 1.6 Gbps -- -- 160 -- -- 160 -- -- 160 ps Total Jitter for Data Rate, < 600 Mbps -- -- 0.1 -- -- 0.1 -- -- 0.1 UI Total Jitter for Data Rate, 600 Mbps to 1.25 Gbps -- -- 300 -- -- 300 -- -- 325 ps Total Jitter for Data Rate < 600 Mbps -- -- 0.2 -- -- 0.2 -- -- 0.25 UI -- -- -- 0.125 -- -- 0.15 -- -- 0.15 UI Tx output clock duty cycle for both True and Emulated Differential I/O Standards 45 50 55 45 50 55 45 50 55 % True Differential I/O Standards -- -- 160 -- -- 200 -- -- 200 ps Emulated Differential I/O Standards with Three External Output Resistor Networks -- -- 250 -- -- 250 -- -- 300 ps Emulated Differential I/O Standards with One External Output Resistor -- -- 460 -- -- 500 -- -- 500 ps Transmitter SERDES factor J = 3 to 10 True Differential I/O Standards - fHSDR (data rate) Emulated Differential I/O Standards with Three External Output Resistor Networks - fHSDR (data rate) (6) (9) SERDES factor J = 4 to 10 Emulated Differential I/O Standards with One External Output Resistor - fHSDR (data rate) tx Jitter - True Differential I/O Standards tx Jitter - Emulated Differential I/O Standards with Three External Output Resistor Network tx Jitter - Emulated Differential I/O Standards with One External Output Resistor Network tDUTY tRISE & tFALL July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-56 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-42. High-Speed I/O Specifications (1), (2) (Part 3 of 3) -2/-2x Speed Grade Symbol TCCS -4 Speed Grade -3 Speed Grade Conditions Unit Min Typ Max Min Typ Max Min Typ Max True Differential I/O Standards -- -- 100 -- -- 100 -- -- 100 ps Emulated Differential I/o Standards -- -- 250 -- -- 250 -- -- 250 ps SERDES factor J = 3 to 10 150 -- 1600 150 -- 1250 150 -- 1250 Mbps SERDES factor J = 3 to 10 (4) -- (7) (4) -- (7) (4) -- (7) Mbps SERDES factor J = 2, Uses DDR Registers (4) -- (5) (4) -- (5) (4) -- (5) Mbps SERDES factor J = 1, Uses an SDR Register (4) -- (5) (4) -- (5) (4) -- (5) Mbps -- -- -- 10000 -- -- 10000 -- -- 10000 UI -- -- -- 300 -- -- 300 -- -- 300 PPM -- -- -- 300 -- -- 300 -- -- 300 ps Receiver True Differential I/O Standards fHSDRDPA (data rate) fHSDR (data rate) DPA Mode DPA run length Soft CDR mode Soft-CDR PPM tolerance Non DPA Mode Sampling Window Notes to Table 1-42: (1) When J = 3 to 10, use the serializer/deserializer (SERDES) block. (2) When J = 1 or 2, bypass the SERDES block. (3) Clock Boost Factor (W) is the ratio between input data rate to the input clock rate. (4) The minimum specification depends on the clock source (for example, the PLL and clock pin) and the clock routing resource (global, regional, or local) that you use. The I/O differential buffer and input register do not have a minimum toggle rate. (5) The maximum ideal frequency is the SERDES factor (J) x the PLL maximum output frequency (fOUT) provided you can close the design timing and the signal integrity simulation is clean. (6) You must calculate the leftover timing margin in the receiver by performing link timing closure analysis. You must consider the board skew margin, transmitter channel-to-channel skew, and receiver sampling margin to determine leftover timing margin. (7) You can estimate the achievable maximum data rate for non-DPA mode by performing link timing closure analysis. You must consider the board skew margin, transmitter delay margin, and the receiver sampling margin to determine the maximum data rate supported. (8) This is achieved by using the LVDS and DPA clock network. (9) If the receiver with DPA enabled and transmitter are using shared PLLs, the minimum data rate is 150 Mbps. (10) This only applies to DPA and soft-CDR modes. (11) This only applies to LVDS source synchronous mode. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-57 Table 1-43 lists the DPA lock time specifications for Stratix IV ES devices. Table 1-43. DPA Lock Time Specifications--Stratix IV ES Devices Only Standard SPI-4 Training Pattern Number of Data Transitions in one repetition of training pattern 00000000001111111111 00001111 2 Number of repetitions per 256 data transitions 128 Parallel Rapid I/O 10010000 10101010 4 64 8 32 Miscellaneous 01010101 8 Condition Maximum without DPA PLL calibration 256 data transitions with DPA PLL calibration 3x256 data transitions + 2x96 slow clock cycles (5) without DPA PLL calibration 256 data transitions with DPA PLL calibration 3x256 data transitions + 2x96 slow clock cycles (5) without DPA PLL calibration 256 data transitions with DPA PLL calibration 3x256 data transitions + 2x96 slow clock cycles (5) without DPA PLL calibration 256 data transitions with DPA PLL calibration 3x256 data transitions + 2x96 slow clock cycles (5) without DPA PLL calibration 256 data transitions with DPA PLL calibration 3x256 data transitions + 2x96 slow clock cycles (5) (4) 128 2 (1), (2), (3) 32 Notes to Table 1-43: (1) The DPA lock time is for one channel. (2) One data transition is defined as a 0-to-1 or 1-to-0 transition. (3) The DPA lock time applies to commercial, industrial, and military speed grades. (4) This is the number of repetition for the stated training pattern to achieve 256 data transitions. (5) Slow clock = Data rate (Mbps)/Deserialization factor. Figure 1-4 shows the DPA lock time specifications with DPA PLL calibration enabled. Figure 1-4. DPA Lock Time Specification with DPA PLL Calibration Enabled rx_reset DPA Lock Time rx_dpa_locked 256 data transitions July 2012 Altera Corporation 96 slow clock cycles 256 data transitions 96 slow clock cycles 256 data transitions Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-58 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-44 lists the DPA lock time specifications for Stratix IV GX and GT devices. Table 1-44. DPA Lock Time Specifications--Stratix IV GX and GT Devices Only Standard Training Pattern SPI-4 Number of Data Transitions in One Repetition of the Training Pattern (1), (2), (3) Number of Repetitions per 256 Data Transitions Maximum (4) 00000000001111111111 2 128 640 data transitions 00001111 2 128 640 data transitions 10010000 4 64 640 data transitions 10101010 8 32 640 data transitions 01010101 8 32 640 data transitions Parallel Rapid I/O Miscellaneous Notes to Table 1-44: (1) The DPA lock time is for one channel. (2) One data transition is defined as a 0-to-1 or 1-to-0 transition. (3) The DPA lock time stated in the table applies to commercial, industrial, and military speed grades. (4) This is the number of repetitions for the stated training pattern to achieve the 256 data transitions. Figure 1-5 shows the LVDS soft-CDR/DPA sinusoidal jitter tolerance specification for a data rate equal to or higher than 1.25 Gbps. Table 1-45 lists this information in table form. Figure 1-5. LVDS Soft-CDR/DPA Sinusoidal Jitter Tolerance Specification for a Data Rate Equal to or Higher Than 1.25 Gbps LVDS Soft-CDR/DPA Sinusoidal Jitter Tolerance Specification Jitter Amphlitude (UI) 25 8.5 0.35 0.1 F1 F2 F3 F4 Jitter Frequency (Hz) Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-59 Table 1-45 lists the LVDS soft-CDR/DPA sinusoidal jitter tolerance specification for a data rate equal to or higher than 1.25 Gbps. Table 1-45. LVDS Soft-CDR/DPA Sinusoidal Jitter Mask Values for a Data Rate Equal to or Higher than 1.25 Gbps Jitter Frequency (Hz) F1 Sinusoidal Jitter (UI) 10,000 25.000 F2 17,565 25.000 F3 1,493,000 0.350 F4 50,000,000 0.350 Figure 1-6 shows the LVDS soft-CDR/DPA sinusoidal jitter tolerance specification for a data rate less than 1.25 Gbps. Figure 1-6. LVDS Soft-CDR/DPA Sinusoidal Jitter Tolerance Specification for a Data Rate Less than 1.25 Gbps Sinusoidal Jitter Amplitude 20db/dec 0.1 UI P-P Frequency 20 MHz baud/1667 When the data rate is equals to 800 Mbps, the LVDS soft-CDR/DPA sinusoidal jitter tolerance allows up to 0.1 UI (125 ps) for jitter frequencies between 479.9 kHz and 20 MHz. DLL and DQS Logic Block Specifications Table 1-46 lists the DLL frequency range specifications for Stratix IV devices. Table 1-46. DLL Frequency Range Specifications for Stratix IV Devices (Part 1 of 2) Frequency Range (MHz) -2/-2x Speed Grade -3 Speed Grade -4 Speed Grade Available Phase Shift DQS Delay Buffer Mode (1) Number of Delay Chains 0 90-140 90-130 90-120 22.5, 45, 67.5, 90 Low 16 1 120-180 120-170 120-160 30, 60, 90, 120 Low 12 2 150-220 150-210 150-200 36, 72, 108, 144 Low 10 3 180-280 180-260 180-240 45, 90,135, 180 Low 8 Frequency Mode July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-60 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics Table 1-46. DLL Frequency Range Specifications for Stratix IV Devices (Part 2 of 2) Frequency Range (MHz) Frequency Mode DQS Delay Buffer Mode (1) Number of Delay Chains -2/-2x Speed Grade -3 Speed Grade -4 Speed Grade Available Phase Shift 4 240-350 240-320 240-290 30, 60, 90, 120 High 12 5 290-430 290-380 290-360 36, 72, 108, 144 High 10 6 360-540 360-450 360-450 45, 90, 135, 180 High 8 7 470-700 470-630 470-590 60, 120, 180, 240 High 6 Note to Table 1-46: (1) Low indicates a 6-bit DQS delay setting; high indicates a 5-bit DQS delay setting. Table 1-47 lists the DQS phase offset delay per stage for Stratix IV devices. Table 1-47. DQS Phase Offset Delay Per Setting for Stratix IV Devices (1), (2), (3) Speed Grade Min Max Unit -2/-2x 7 13 ps -3 7 15 ps -4 7 16 ps Notes to Table 1-47: (1) The valid settings for phase offset are -64 to +63 for frequency modes 0 to 3 and -32 to +31 for frequency modes 4 to 6. (2) The typical value equals the average of the minimum and maximum values. (3) The delay settings are linear, with a cumulative delay variation of 40 ps for all speed grades. For example, when using a -2 speed grade and applying a 10 phase offset settings to a 90 phase shift at 400 MHz, the expected average cumulative delay is [625 ps + (10 x 10.5 ps) 20 ps] = 730 ps 20 ps. Table 1-48 lists the DQS phase shift error for Stratix IV devices. Table 1-48. DQS Phase Shift Error Specification for DLL-Delayed Clock (tDQS_PSERR) for Stratix IV Devices (1) Number of DQS Delay Buffer -2/-2X Speed Grade -3 Speed Grade -4 Speed Grade Unit 1 26 28 30 ps 2 52 56 60 ps 3 78 84 90 ps 4 104 112 120 ps Note to Table 1-48: (1) This error specification is the absolute maximum and minimum error. For example, skew on three DQS delay buffers in a -2/-2x speed grade is 78 ps or 39 ps. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Switching Characteristics 1-61 Table 1-49 lists the memory output clock jitter specifications for Stratix IV devices. Table 1-49. Memory Output Clock Jitter Specification for Stratix IV Devices (1), Clock Network Parameter Symbol (2), (3) -2/-2X Speed Grade -3 Speed Grade -4 Speed Grade Min Max Min Max Min Max Unit Clock period jitter Regional tJIT(per) -50 50 -55 55 -55 55 ps Cycle-to-cycle period jitter Regional tJIT(cc) -100 100 -110 110 -110 110 ps Duty cycle jitter Regional tJIT(duty) -50 50 -82.5 82.5 -82.5 82.5 ps Clock period jitter Global tJIT(per) -75 75 -82.5 82.5 -82.5 82.5 ps Cycle-to-cycle period jitter Global tJIT(cc) -150 150 -165 165 -165 165 ps Duty cycle jitter Global tJIT(duty) -75 75 -90 90 -90 90 ps Notes to Table 1-49: (1) The memory output clock jitter measurements are for 200 consecutive clock cycles, as specified in the JEDEC DDR2/DDR3 SDRAM standard. (2) The clock jitter specification applies to memory output clock pins generated using differential signal-splitter and DDIO circuits clocked by a PLL output routed on a regional or global clock network as specified. Altera recommends using regional clock networks whenever possible. (3) The memory output clock jitter stated in Table 1-49 is applicable when an input jitter of 30 ps is applied. OCT Calibration Block Specifications Table 1-50 lists the OCT calibration block specifications for Stratix IV devices. Table 1-50. OCT Calibration Block Specifications for Stratix IV Devices Symbol Description Min Typ Max Unit OCTUSRCLK Clock required by OCT calibration blocks -- -- 20 MHz TOCTCAL Number of OCTUSRCLK clock cycles required for OCT RS/RT calibration -- 1000 -- Cycles TOCTSHIFT Number of OCTUSRCLK clock cycles required for OCT code to shift out -- 28 -- Cycles TRS_RT Time required between the dyn_term_ctrl and oe signal transitions in a bidirectional I/O buffer to dynamically switch between OCT RS and RT -- 2.5 -- ns July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-62 Chapter 1: DC and Switching Characteristics for Stratix IV Devices I/O Timing Figure 1-7 shows the timing diagram for the oe and dyn_term_ctrl signals. Figure 1-7. Timing Diagram for the oe and dyn_term_ctrl Signals Tristate Tristate TX RX RX oe dyn_term_ctrl TRS_RT TRS_RT Duty Cycle Distortion (DCD) Specifications Table 1-51 lists the worst-case DCD for Stratix IV devices. Table 1-51. Worst-Case DCD on Stratix IV I/O Pins (1) Symbol Output Duty Cycle -2/-2x Speed Grade -3 Speed Grade -4 Speed Grade Min Max Min Max Min Max 45 55 45 55 45 55 Unit % Note to Table 1-51: (1) The listed specification is only applicable to the output buffer across different I/O standards. I/O Timing Altera offers two ways to determine I/O timing--the Excel-based I/O Timing and the Quartus II Timing Analyzer. Excel-based I/O Timing provides pin timing performance for each device density and speed grade. The data is typically used prior to designing the FPGA to get an estimate of the timing budget as part of the link timing analysis. The Quartus II Timing Analyzer provides a more accurate and precise I/O timing data based on the specifics of the design after you complete place-and-route. f The Excel-based I/O Timing spreadsheet is downloadable from the Literature: Stratix IV Devices webpage. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices I/O Timing 1-63 Programmable IOE Delay Table 1-52 lists the Stratix IV IOE programmable delay settings. Table 1-52. IOE Programmable Delay for Stratix IV Devices Fast Model Parameter (1) Available Settings Slow Model Min Offset (2) Industrial/ Military Commercial (3) C2 (3) C3 C4 I3/M3 I4 Unit D1 16 0 0.462 0.505 0.732 0.795 0.857 0.801 0.864 ns D2 8 0 0.234 0.232 0.337 0.372 0.407 0.371 0.405 ns D3 8 0 1.700 1.769 2.695 2.927 3.157 2.948 3.178 ns D4 16 0 0.508 0.554 0.813 0.882 0.952 0.889 0.959 ns D5 16 0 0.472 0.500 0.747 0.799 0.875 0.817 0.882 ns D6 8 0 0.186 0.195 0.294 0.319 0.345 0.321 0.347 ns Notes to Table 1-52: (1) You can set this value in the Quartus II software by selecting D1, D2, D3, D4, D5, and D6 in the Assignment Name column. (2) Minimum offset does not include the intrinsic delay. (3) For the EP4SGX530 device density, the IOE programmable delays have an additional 5% maximum offset. Programmable Output Buffer Delay Table 1-53 lists the delay chain settings that control the rising and falling edge delays of the output buffer. The default delay is 0 ps. Table 1-53. Programmable Output Buffer Delay Symbol DOUTBUF (1) Parameter Rising and/or falling edge delay Typical Unit 0 (default) ps 50 ps 100 ps 150 ps Note to Table 1-53: (1) You can set the programmable output buffer delay in the Quartus II software by setting the Output Buffer Delay Control assignment to either positive, negative, or both edges, with the specific values stated here (in ps) for the Output Buffer Delay assignment. July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-64 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Glossary Glossary Table 1-54 lists the glossary for this chapter. Table 1-54. Glossary Table (Part 1 of 4) Letter Subject Definitions A, B, C -- -- Receiver Input Waveforms Single-Ended Waveform Positive Channel (p) = VIH VID Negative Channel (n) = VIL VCM Ground Differential Waveform VID p-n=0V VID D Differential I/O Standards Transmitter Output Waveforms Single-Ended Waveform Positive Channel (p) = VOH VOD Negative Channel (n) = VOL VCM Ground Differential Waveform VOD p-n=0V VOD E F G, H, I -- -- fHSCLK Left/right PLL input clock frequency. fHSDR High-speed I/O block: Maximum/minimum LVDS data transfer rate (fHSDR = 1/TUI), non-DPA. fHSDRDPA High-speed I/O block: Maximum/minimum LVDS data transfer rate (fHSDRDPA = 1/TUI), DPA. -- Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum -- July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Glossary 1-65 Table 1-54. Glossary Table (Part 2 of 4) Letter Subject J Definitions High-speed I/O block: Deserialization factor (width of parallel data bus). JTAG Timing Specifications: TMS TDI J t JCP JTAG Timing Specifications t JCH t JCL t JPH t JPSU TCK tJPZX t JPXZ t JPCO TDO K, L, M, N, O -- -- Diagram of PLL Specifications (1) Switchover CLKOUT Pins fOUT_EXT CLK fIN N fINPFD PFD CP LF Core Clock P VCO fVCO Counters C0..C9 fOUT PLL Specifications GCLK RCLK M Key Reconfigurable in User Mode External Feedback Note: (1) Core Clock can only be fed by dedicated clock input pins or PLL outputs. Q R July 2012 -- RL Altera Corporation -- Receiver differential input discrete resistor (external to Stratix IV device). Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-66 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Glossary Table 1-54. Glossary Table (Part 3 of 4) Letter Subject SW (sampling window) Definitions Timing Diagram--the period of time during which the data must be valid in order to capture it correctly. The setup and hold times determine the ideal strobe position within the sampling window, as shown: Bit Time 0.5 x TCCS RSKM Sampling Window (SW) RSKM 0.5 x TCCS The JEDEC standard for SSTl and HSTL I/O defines both the AC and DC input signal values. The AC values indicate the voltage levels at which the receiver must meet its timing specifications. The DC values indicate the voltage levels at which the final logic state of the receiver is unambiguously defined. After the receiver input has crossed the AC value, the receiver changes to the new logic state. The new logic state is then maintained as long as the input stays beyond the AC threshold. This approach is intended to provide predictable receiver timing in the presence of input waveform ringing, as shown: S Single-Ended Voltage Referenced I/O Standard Single-ended voltage referenced I/O standard VCCIO VOH VIH (AC ) VIH(DC) VREF VIL(DC) VIL(AC ) VOL VSS High-speed receiver/transmitter input and output clock period. tC The timing difference between the fastest and slowest output edges, including tCO variation TCCS (channeland clock skew, across channels driven by the same PLL. The clock is included in the TCCS to-channel-skew) measurement (refer to the Timing Diagram figure under SW in this table). High-speed I/O block: Duty cycle on high-speed transmitter output clock. Timing Unit Interval (TUI) tDUTY The timing budget allowed for skew, propagation delays, and data sampling window. (TUI = 1/(Receiver Input Clock Frequency Multiplication Factor) = tC/w) T U tFALL Signal high-to-low transition time (80-20%) tINCCJ Cycle-to-cycle jitter tolerance on the PLL clock input tOUTPJ_IO Period jitter on the general purpose I/O driven by a PLL tOUTPJ_DC Period jitter on the dedicated clock output driven by a PLL tRISE Signal low-to-high transition time (20-80%) -- Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum -- July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Document Revision History 1-67 Table 1-54. Glossary Table (Part 4 of 4) Letter Subject V W Definitions VCM(DC) DC Common mode input voltage. VICM Input Common mode voltage--The common mode of the differential signal at the receiver. VID Input differential voltage swing--The difference in voltage between the positive and complementary conductors of a differential transmission at the receiver. VDIF(AC) AC differential input voltage--Minimum AC input differential voltage required for switching. VDIF(DC) DC differential input voltage-- Minimum DC input differential voltage required for switching. VIH Voltage input high--The minimum positive voltage applied to the input which is accepted by the device as a logic high. VIH(AC) High-level AC input voltage VIH(DC) High-level DC input voltage VIL Voltage input low--The maximum positive voltage applied to the input which is accepted by the device as a logic low. VIL(AC) Low-level AC input voltage VIL(DC) Low-level DC input voltage VOCM Output Common mode voltage--The common mode of the differential signal at the transmitter. VOD Output differential voltage swing--The difference in voltage between the positive and complementary conductors of a differential transmission at the transmitter. VSWING Differential input voltage VX Input differential cross point voltage VOX Output differential cross point voltage W High-speed I/O block: Clock Boost Factor X, Y, Z -- -- Document Revision History Table 1-55 lists the revision history for this chapter. Table 1-55. Document Revision History (Part 1 of 3) Date July 2012 December 2011 June 2011 July 2012 Version 5.3 5.2 5.1 Altera Corporation Changes Added Table 1-5 and Table 1-40. Updated Table 1-15, Table 1-22, Table 1-23, Table 1-30, Table 1-33, Table 1-35, Table 1-36, Table 1-39, Table 1-42 and Table 1-51. Removed "Schmitt Trigger Input" section. Added Figure 1-7. Updated Table 1-22 and Table 1-41. Added military speed grade information. Updated Table 1-1 and Table 1-30. Updated (Note 3) in Table 1-42 and (Note 3) in Table 1-43. Added military speed grade to Table 1-5, Table 1-10, Table 1-11, Table 1-23, Table 1-30, Table 1-36, and Table 1-51. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-68 Chapter 1: DC and Switching Characteristics for Stratix IV Devices Document Revision History Table 1-55. Document Revision History (Part 2 of 3) Date Version Updated Table 1-1, Table 1-5, Table 1-6, Table 1-13, Table 1-16, Table 1-23, and Table 1-24. 4.9 Updated Table 1-24. 4.8 Removed (Note 17) in Table 1-24. 4.7 Added (Note 17) to Table 1-24. Updated Table 1-1, Table 1-5, Table 1-23, Table 1-24, Table 1-30, Table 1-31, Table 1-32, Table 1-34, Table 1-37, Table 1-41, and Table 1-51. Updated the "Recommended Operating Conditions" section. Added the "Schmitt Trigger Input" section.July Minor text edits. Updated Table 1-29. Updated chapter title. Minor text edits. Applied new template. Updated Table 1-1 and Table 1-5. Updated Table 1-7, Table 1-22, Table 1-23, Table 1-33, Table 1-35, Table 1-36, and Table 1-40. Added Table 1-39. Changed "PCI Express" to "PCIe" throughout. Minor text edits Updated Table 1-22, Table 1-23, Table 1-30, Table 1-46, and Table 1-49. Added Table 1-31. Minor text edits. Updated Table 1-11, Table 1-22, Table 1-23, Table 1-24, Table 1-25, Table 1-26, Table 1-27, Table 1-29, Table 1-32, Table 1-33, Table 1-34, Table 1-35, Table 1-39, Table 1-40, Table 1-43, Table 1-46, and Table 1-49. Added Stratix IV GT speed grade note to Table 1-32, Table 1-35, Table 1-39, Table 1-43, Table 1-44, Table 1-45, and Table 1-46. Added Table 1-28 and Table 1-30. Minor text edits. Added Table 1-9, Table 1-15, Table 1-38, and Table 1-39. Added Figure 1-5 and Figure 1-6. Added the "Transceiver Datapath PCS Latency" section. Updated the "Electrical Characteristics", "Operating Conditions", and "I/O Timing" sections. All tables updated except Table 1-16, Table 1-24, Table 1-25, Table 1-26, Table 1-27, Table 1-33, Table 1-34, and Table 1-45. Updated Figure 1-2 and Figure 1-3. Updated Equation 1-1. Deleted Table 1-28, Table 1-29, Table 1-30, Table 1-42, Table 1-43, and Table 1-44. Minor text edits. April 2011 5.0 March 2011 March 2011 February 2011 February 2011 November 2010 September 2010 July 2010 March 2010 February 2010 November 2009 Changes 4.6 4.5 4.4 4.3 4.2 4.1 4.0 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation Chapter 1: DC and Switching Characteristics for Stratix IV Devices Document Revision History 1-69 Table 1-55. Document Revision History (Part 3 of 3) Date June 2009 March 2009 December 2008 November 2008 August 2008 May 2008 July 2012 Version Changes Added "Preliminary Specifications" to the footer of each page. Updated Table 1-1, Table 1-2, Table 1-7, Table 1-10, Table 1-11, Table 1-12, Table 1-21, Table 1-22, Table 1-23, Table 1-25, Table 1-37, Table 1-38, Table 1-39, Table 1-40, and Table 1-44. Minor text edits. Replaced Table 1-31 and Table 1-37. Updated Table 1-1, Table 1-2, Table 1-5, Table 1-19, Table 1-41, Table 1-44, Table 1-45, Table 1-49, and Table 1-51. Added Table 1-21, Table 1-46, and Table 1-47 Added Figure 1-3. Removed "Timing Model", "Preliminary and Final Timing", "I/O Timing Measurement Methodology", "I/O Default Capacitive Loading", and "Referenced Documents" sections. 3.1 3.0 2.1 2.0 1.1 1.0 Altera Corporation Minor changes. Minor text edits. Updated Table 1-19, Table 1-32, Table 1-34 - Table 1-39. Minor text edits. Updated Table 1-1, Table 1-2, Table 1-4, Table 1-5, and Table 1-26. Removed figures from "Transceiver Performance Specifications" on page 1-10 that are repeated in the glossary. Minor text edits and an additional note to Table 1-26. Initial release. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 1-70 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum Chapter 1: DC and Switching Characteristics for Stratix IV Devices Document Revision History July 2012 Altera Corporation 2. Addendum to the Stratix IV Device Handbook February 2011 SIV54002-1.5 SIV54002-1.5 This chapter describes changes to the published version of the Stratix IV Device Handbook. All changes from Revision 1.4 of this chapter are now incorporated in the main handbook chapters or in AN 612: Decision Feedback Equalization in Stratix IV Devices. Adaptive Equalization (AEQ) f This AEQ information is now located in the Dynamic Reconfiguration in Stratix IV Devices chapter. Decision Feedback Equalization (DFE) f For more information about the DFE feature, refer to AN 612: Decision Feedback Equalization in Stratix IV Devices. Power-On Reset Circuitry The Power-On Reset Circuitry information is now located in the Hot Socketing and Power-On Reset in Stratix IV Devices chapter. Power-On Reset Specifications f The Power-On Reset Specification information is now located in the Hot Socketing and Power-On Reset in Stratix IV Devices chapter. (c) 2012 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, HARDCOPY, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as trademarks or service marks are the property of their respective holders as described at www.altera.com/common/legal.html. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. ISO 9001:2008 Registered Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum February 2011 Feedback Subscribe 2-2 Chapter 2: Addendum to the Stratix IV Device Handbook Document Revision History Table 2-5 lists the revision history for this chapter. Table 2-1. Document Revision History Date February 2011 September 2010 April 2010 March 2010 February 2010 November 2009 Version Changes Removed the "Decision Feedback Equalization (DFE)" section now that AN 612: Decision Feedback Equalization in Stratix IV Devices is published. Moved the "Adaptive Equalization (AEQ)" sections to the Dynamic Reconfiguration in Stratix IV Devices chapter. Moved the "Power-On Reset Circuitry" and "Power-On Reset Specifications" sections to the Hot Socketing and Power-On Reset in Stratix IV Devices chapter. Minor text edits. Added corrections for the Adaptive Equalization (AEQ) section of the Stratix IV Dynamic Reconfiguration chapter. Added new information for the Decision Feedback Equalization (DFE) feature. Added corrections for the "Power-On Reset Circuitry" and "Power-On Reset Specifications" sections to of the Hot Socketing and Power-On Reset in Stratix IV Devices chapter. Moved the "Power-On Reset Circuitry", "Power-On Reset Specifications", "Correct Power-Up Sequence for Production Devices", and "Correct Power-Up Sequence for Production Devices" sections to the Hot Socketing and Power-On Reset in Stratix IV Devices chapter. Moved the "Power-On Reset Circuit" and "JTAG TMS and TDI Pin Pull-Up Resistor Value Specification" sections to the Configuration, Design Security, Remote System Upgrades with Stratix IV Devices chapter. Moved the "Summary of OCT Assignments" section to the I/O Features in Stratix IV Devices chapter. Added the "Power-On Reset Circuitry", "Power-On Reset Specifications", "Correction to POR Signal Pulse Width Delay Times", "Correct Power-Up Sequence for Production Devices", "Power-On Reset Circuit", "Summary of OCT Assignments", and "JTAG TMS and TDI Pin Pull-Up Resistor Value Specification" sections. Minor text edits. Stratix IV GX enhanced transceiver data rate specifications in -4 commercial speed grade. Initial release. 1.5 1.4 1.3 1.2 1.1 1.0 Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum February 2011 Altera Corporation Additional Information This chapter provides additional information about the document and Altera. About this Handbook This handbook provides comprehensive information about the Altera(R) Stratix(R) IV family of devices. How to Contact Altera To locate the most up-to-date information about Altera products, refer to the following table. Contact (1) Technical support Technical training Product literature Contact Method Address Website www.altera.com/support Website www.altera.com/training Email custrain@altera.com Website www.altera.com/literature Nontechnical support (general) Email nacomp@altera.com (software licensing) Email authorization@altera.com Note to Table: (1) You can also contact your local Altera sales office or sales representative. Typographic Conventions The following table shows the typographic conventions this document uses. Visual Cue Meaning Bold Type with Initial Capital Letters Indicate command names, dialog box titles, dialog box options, and other GUI labels. For example, Save As dialog box. For GUI elements, capitalization matches the GUI. bold type Indicates directory names, project names, disk drive names, file names, file name extensions, software utility names, and GUI labels. For example, \qdesigns directory, D: drive, and chiptrip.gdf file. Italic Type with Initial Capital Letters Indicate document titles. For example, Stratix IV Design Guidelines. Indicates variables. For example, n + 1. italic type Variable names are enclosed in angle brackets (< >). For example, <file name> and <project name>.pof file. Initial Capital Letters Indicate keyboard keys and menu names. For example, the Delete key and the Options menu. "Subheading Title" Quotation marks indicate references to sections in a document and titles of Quartus II Help topics. For example, "Typographic Conventions." July 2012 Altera Corporation Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum 2-2 Additional Information Typographic Conventions Visual Cue Meaning Indicates signal, port, register, bit, block, and primitive names. For example, data1, tdi, and input. The suffix n denotes an active-low signal. For example, resetn. Courier type Indicates command line commands and anything that must be typed exactly as it appears. For example, c:\qdesigns\tutorial\chiptrip.gdf. Also indicates sections of an actual file, such as a Report File, references to parts of files (for example, the AHDL keyword SUBDESIGN), and logic function names (for example, TRI). r An angled arrow instructs you to press the Enter key. 1., 2., 3., and a., b., c., and so on Numbered steps indicate a list of items when the sequence of the items is important, such as the steps listed in a procedure. Bullets indicate a list of items when the sequence of the items is not important. 1 The hand points to information that requires special attention. h The question mark directs you to a software help system with related information. f The feet direct you to another document or website with related information. m The multimedia icon directs you to a related multimedia presentation. c A caution calls attention to a condition or possible situation that can damage or destroy the product or your work. w A warning calls attention to a condition or possible situation that can cause you injury. The envelope links to the Email Subscription Management Center page of the Altera website, where you can sign up to receive update notifications for Altera documents. Stratix IV Device Handbook Volume 4: Device Datasheet and Addendum July 2012 Altera Corporation