www.ubicom.com © 2001-2003 Ubicom, Inc. All rights reserved. 1
1.0 Product Highlights
The Ubicom IP2012™ and IP2022™ Wireless Network
Processors combine support for communication physical
layer, Internet protocol stack, device-specific application,
and device-specific peripheral software modules in a
single chip, and are reconfigurable over the Internet. They
can be programmed, and reprogrammed, using pre-built
software modules and configuration tools to create true
single-chip solutions for a wide range of device-to-device
and device-to-human communication applications. High
speed communication interfaces are available via on-chip
hardware Serializer/Deserializer (SerDes) blocks. These
full-duplex blocks allow the IP2022 or IP2012 to be used
in a variety of communication bridging applications. Each
SerDes block is capable of supporting 10Base-T Ethernet
(MAC and PHY), USB, GPSI, SPI, or UART. The high-
speed operating frequency, combined with most
instructions executing in a single cycle, delivers the
throughput needed for emerging network connectivity
applications. A flash-based program memory allows both
in-system and runtime reprogramming. The IP2022 and
IP2012 implement most peripheral, communications and
control functions via software modules (ipModule™
software), replacing traditional hardware for maximum
system design flexibility. This approach allows rapid,
inexpensive product design and, when needed, quick and
easy reconfiguration to accommodate changes in market
needs or industry standards.
Key Features:
•
Designed to support single-chip networked solutions
•
Fast processor core
•
64kB Flash program memory
•
16kB SRAM data/program memory
•
4kB SRAM data memory
•
Two SerDes communication blocks supporting com-
mon PHYs (Ethernet, USB, UARTs, etc.) and bridging
applications (IP2012 has only one SerDes unit)
•
Advanced RISC processors
•
IP2022 — 120 and 160 MHz versions
•
IP2012 — 120 MHz version
•
High speed packet processing
•
Instruction set optimized for communication functions
•
Supports software implementation of traditional hard-
ware functions
•
In-system reprogrammable for highest flexibility
•
Run time self-programmable
•
Vpp = Vcc supply voltage
Figure 1-1 IP2012 / IP2022 Block Diagram
515-063b.eps
IP2022/IP2012
10Base-T Ethernet
(MAC/PHY on chip)
USB 1.1 (SIE on chip)
GPSI
SPI
UART/Modem
Bluetooth HCI
Customer Application
HTTP/SMTP/TFTP
TCP/UDP
IP/ICMP
Network Access Layer
ipModule Software
PHY Firmware
ipOS Operating System
ISP/ISD
Interface
8-Input
10-Bit
A/DC
PLL
Clock
Multiplier
5
Timers
External
Memory
Interface
4-Kbyte
Data
RAM
16-Kbyte
Inst./Data
RAM
64-Kbyte
Flash
Memory
Internet
Processor
CPU
8/16-Bit
Parallel
Slave Port
ISA (802.11b)
Mini-PCI/Cardbus
(802.11g/802.11a)
I2C
General-Purpose I/O
Choices for
Communication:
10Base-T
Ethernet
(MAC/PHY on chip)
USB 1.1 (SIE on chip)
GPSI
SPI
UART/Modem
Bluetooth HCI
Host Bus
Choices for
Communication:
TM
High-Speed
Serial Unit 1
(SERDES)
High-Speed
Serial Unit 2
(SERDES)
Not available on IP2012
General
Purpose
I/O Ports
IP2012 / IP2022 Wireless Network Processors
Features and Performance Optimized for Network Connectivity
2 www.ubicom.com
IP2012 / IP2022 Data Sheet
1.0 Product Highlights
1
1.1 Additional Features. . . . . . . . . . . . . . . . . . . . . . . . . . .3
1.2 Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.1 CPU
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.2 Serializer/Deserializers
. . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.3 Low-Power Support
. . . . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.4 Memory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.5 Instruction Set
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
1.2.6 Other Supported Functions
. . . . . . . . . . . . . . . . . . . . . . .5
1.2.7 Programming and Debugging
Support
5
2.0 Pin Definitions
6
2.1 PQFP (Plastic Quad Flat Package) for IP2022. . . . . .6
2.2 PQFP (Plastic Quad Flat Package) for IP2012. . . . . .7
2.3 µBGA (Micro Ball Grid Array) IP2022-120 Only . . . . .8
2.4 Signal Descriptions — IP2022 . . . . . . . . . . . . . . . . . .9
2.5 Signal Descriptions — IP2012 . . . . . . . . . . . . . . . . .12
3.0 System Architecture
15
3.1 CPU Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
3.2 Data Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
3.3 Program Memory . . . . . . . . . . . . . . . . . . . . . . . . . . .19
3.3.1 Loading the Program RAM
. . . . . . . . . . . . . . . . . . . . . .19
3.3.2 Program Counter
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
3.4 Low Power Support . . . . . . . . . . . . . . . . . . . . . . . . .20
3.4.1 Clock Stop Mode (SLEEP)
. . . . . . . . . . . . . . . . . . . . . .21
3.4.2 Wakeup
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
3.5 Speed Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
3.6 Instruction Timing . . . . . . . . . . . . . . . . . . . . . . . . . . .21
3.7 Interrupt Support. . . . . . . . . . . . . . . . . . . . . . . . . . . .22
3.7.1 Interrupt Processing
. . . . . . . . . . . . . . . . . . . . . . . . . . .22
3.7.2 Global Interrupt Enable Bit
. . . . . . . . . . . . . . . . . . . . . .25
3.7.3 Interrupt Latency
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
3.7.4 Return From Interrupt
. . . . . . . . . . . . . . . . . . . . . . . . . .25
3.7.5 Disabled Interrupt Resources
. . . . . . . . . . . . . . . . . . . .26
3.8 Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
3.8.1 Brown-Out Detector
. . . . . . . . . . . . . . . . . . . . . . . . . . .28
3.8.2 Reset and Interrupt Vectors
. . . . . . . . . . . . . . . . . . . . . .28
3.8.3 Register States Following Reset
. . . . . . . . . . . . . . . . . .28
3.9 Clock Oscillator. . . . . . . . . . . . . . . . . . . . . . . . . . . . .29
3.9.1 External Connections
. . . . . . . . . . . . . . . . . . . . . . . . . .30
3.10 Configuration Block. . . . . . . . . . . . . . . . . . . . . . . . . .31
3.10.1 FUSE0 Register (not run-time programmable)
. . . . . . . . .32
3.10.2 FUSE1 Register (not run-time programmable)
. . . . . . . . .33
3.10.3 TRIM0 Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
4.0 Instruction Set Architecture
35
4.1 Addressing Modes . . . . . . . . . . . . . . . . . . . . . . . . . .35
4.1.1 Pointer Registers
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
4.1.2 Direct Addressing Mode
. . . . . . . . . . . . . . . . . . . . . . . .36
4.1.3 Indirect Addressing Mode
. . . . . . . . . . . . . . . . . . . . . . .36
4.1.4 Indirect-with-Offset Addressing Mode
. . . . . . . . . . . . . . .37
4.2 Instruction Set. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
4.2.1 Instruction Formats
. . . . . . . . . . . . . . . . . . . . . . . . . . .38
4.2.2 Instruction Types
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
4.3 Instruction Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . .40
4.4 Subroutine Call/Return Stack . . . . . . . . . . . . . . . . . .41
4.5 Key to Abbreviations and Symbols. . . . . . . . . . . . . .42
4.6 Instruction Set Summary Tables. . . . . . . . . . . . . . . .42
4.7 Program Memory Instructions. . . . . . . . . . . . . . . . . .47
4.7.1 Flash Timing Control
. . . . . . . . . . . . . . . . . . . . . . . . . .48
4.7.2 Interrupts During Flash Operations
. . . . . . . . . . . . . . . . .48
5.0 Peripherals
49
5.1 I/O Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
5.1.1 Port B Interrupts
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
5.1.2 Reading and Writing the Ports
. . . . . . . . . . . . . . . . . . . .50
5.1.3 RxIN Registers
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
5.1.4 RxOUT Registers
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
5.1.5 RxDIR Registers
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
5.1.6 INTED Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
5.1.7 INTF Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
5.1.8 INTE Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
5.1.9 Port Configuration Upon Power-Up
. . . . . . . . . . . . . . . .51
5.2 Timer 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
5.3 Real-Time Timer (RTTMR). . . . . . . . . . . . . . . . . . . .52
5.4 Multi-Function Timers (T1 and T2) . . . . . . . . . . . . . .54
5.4.1 Timers T1, T2 Operating Modes
. . . . . . . . . . . . . . . . . .54
5.4.2 T1 and T2 Timer Pin Assignments
. . . . . . . . . . . . . . . . .56
5.4.3 T1 and T2 Timer Registers
. . . . . . . . . . . . . . . . . . . . . .56
5.5 Watchdog Timer (WDT) . . . . . . . . . . . . . . . . . . . . . .57
5.6 Serializer/Deserializer (SERDES). . . . . . . . . . . . . . .58
5.6.1 SERDES TX/RX Buffers
. . . . . . . . . . . . . . . . . . . . . . . .58
5.6.2 SERDES Configuration
. . . . . . . . . . . . . . . . . . . . . . . . .58
5.6.3 SERDES Interrupts
. . . . . . . . . . . . . . . . . . . . . . . . . . .59
5.6.4 Protocol Modes
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
5.6.5 10base-T Ethernet
. . . . . . . . . . . . . . . . . . . . . . . . . . . .61
5.6.6 USB
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64
5.6.7 UART
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
5.6.8 SPI
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68
5.6.9 GPSI
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70
5.7 Analog to Digital Converter (ADC) . . . . . . . . . . . . . .72
5.7.1 ADC Reference Voltage
. . . . . . . . . . . . . . . . . . . . . . . .72
5.7.2 A/D Converter Registers
. . . . . . . . . . . . . . . . . . . . . . . .72
5.7.3 Using the A/D Converter
. . . . . . . . . . . . . . . . . . . . . . . .73
5.7.4 ADC Result Justification
. . . . . . . . . . . . . . . . . . . . . . . .73
5.8 Comparator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73
5.8.1 CMPCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .73
5.9 Linear Feedback Shift Register (LFSR) . . . . . . . . . .74
5.10 Parallel Slave Peripheral (PSP) . . . . . . . . . . . . . . . .79
5.10.1 PSPCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .79
5.11 External Memory Interface (IP2022 only) . . . . . . . . .80
5.11.1 EMCFG Register (IP2022 only)
. . . . . . . . . . . . . . . . . . .80
6.0 In-System Programming
83
7.0 Memory Reference
84
7.0.1 Registers (sorted by address)
. . . . . . . . . . . . . . . . . . . .84
7.0.2 Program Memory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .88
7.1 Register Bit Definitions . . . . . . . . . . . . . . . . . . . . . . .89
7.1.1 ADCCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .89
7.1.2 ADCTMR Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .89
7.1.3 CMPCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .89
7.1.4 EMCFG Register (IP2022 only)
. . . . . . . . . . . . . . . . . . .90
7.1.5 FCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91
7.1.6 INTSPD Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
7.1.7 LFSRA Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93
7.1.8 PSPCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .93
7.1.9 RTCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .94
7.1.10 SxINTE/SxINTF Register
. . . . . . . . . . . . . . . . . . . . . . .95
7.1.11 SxMODE Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .96
7.1.12 SxRCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .96
7.1.13 SxRCNT Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .97
7.1.14 SxRSYNC Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .97
7.1.15 SxSMASK Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .98
7.1.16 SxTCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .98
7.1.17 SxTMRH/SxTMRL Register
. . . . . . . . . . . . . . . . . . . . . .99
7.1.18 SPDREG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .99
7.1.19 STATUS Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .100
7.1.20 T0CFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . .101
7.1.21 TxCFG1H Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .101
7.1.22 TxCFG2H Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .102
7.1.23 TxCFG1L Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .102
7.1.24 TxCFG2L Register
. . . . . . . . . . . . . . . . . . . . . . . . . . .103
7.1.25 TCTRL Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
7.1.26 XCFG Register
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
8.0 Electrical Characteristics
105
8.1 Absolute Maximum Ratings). . . . . . . . . . . . . . . . . .105
8.2 DC Specifications: IP2022-120, IP2012-120 . . . . .106
8.3 DC Specifications: IP2022-160. . . . . . . . . . . . . . . .108
8.4 AC Specifications: IP2022-120, IP2012-120 . . . . .110
8.5 AC Specifications: IP2022-160. . . . . . . . . . . . . . . .111
8.6 Comparator DC and AC Specifications . . . . . . . . .112
8.7 ADC 10-bit Converter DC and AC Specifications. .112
9.0 Package Dimensions
113
9.1 PQFP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113
9.2 µBGA (available for IP2022-120 only) . . . . . . . . . .114
10.0 Part Numbering
115
IP2012 / IP2022 Data Sheet
www.ubicom.com 3
1.1 Additional Features
Internet Processor Capabilities
Foundation for Highly Flexible Connectivity Solution
•Performance: 120 MIPS @ 120 MHz,
160 MIPS @ 160 MHz
• Predictable execution for hard real-time applications
•Fast and deterministic 3-cycle (25ns @120MHz,
18.75ns @ 160 MHz) internal interrupt response
• Hardware save/store of key registers
•Functions implemented via software tightly coupled
with hardware assist peripherals
Multiple Networking Protocols and Physical Layer
Support Hardware
•Two full-duplex serializer/deserializer (SERDES)
channels (IP2022 has two, IP2012 has one)
– Flexible to support 10Base-T, GPSI, SPI, UART,
USB protocols
– Two channels for protocol bridging
– On-chip squelch function for 10Base-T Ethernet
on each SERDES
•Four hardware LFSR (Linear Feedback Shift Regis-
ter) units
– CRC generation/checking
– Data whitening
– Encryption
Memory
• 64-Kbyte (32K
∞
16) on-chip program flash memory
• 16-Kbyte (8K
∞
16) on-chip program/data RAM
• 4-Kbyte on-chip linear-addressed data RAM
•Self-programming with built-in charge pump: instruc-
tions to read, write, and erase flash memory
•Addresses up to 2 Mbytes of external memory
(IP2022 only)
CPU Features
• RISC engine core
• IP2022-120, IP2012-120
– DC to 120 MHz operation
– 8.33 ns instruction cycle at max frequency
• IP2022-160
– DC to 120 MHz and 160 MHz operation only
– 6.25 ns instruction cycle at max frequency
• Compact 16-bit fixed-length instructions
•Single-cycle instruction execution on most instruc-
tions (3 cycles for jumps and calls)
•Sixteen-level hardware stack for high-performance
subroutine linkage
•8 ∞ 8 signed/unsigned single-cycle multiply
•Pointers and stack operation optimized for C compiler
• Uniform, linear address space (no register banks)
General-Purpose Hardware Peripherals
•Two 16-bit timers with 8-bit prescalers supporting:
– Timer mode
– PWM mode
– Capture/Compare mode
•Parallel host interface, 8/16-bit selectable for use as a
communications coprocessor (IP2012 supports 8-bit
only)
• External memory interface (IP2022 only)
• One 8-bit timer with programmable 8-bit prescaler
•One 8-bit real-time clock/counter with programmable
15-bit prescaler and 32 kHz crystal input
•Watchdog timer with prescaler
• 10-bit, 8-channel ADC with 1/2 LSB accuracy
• Analog comparator with hysteresis enable/disable
• Brown-out minimum supply voltage detector
• External interrupt inputs on 8 pins (Port B)
Sophisticated Power and Frequency/Clock
Management Support
• Operating voltage of 2.3V to 2.7V (120 MHz)
•Switching the system clock frequencies between dif-
ferent clock sources
•On-chip PLL clock multiplier with pre- and post-divider
– 120 MHz on-chip clock from 4.8 MHz ext. crystal
– 160 MHz on-chip clock from 3.2 MHz ext. crystal
• Changing the core clock using a selectable divider
• Shutting down the PLL and/or the OSC input
• Dynamic CPU speed control with speed instruction
•Power-On-Reset (POR) logic
Flexible I/O
• 52 I/O Pins (48 on IP2012)
•2.3V to 3.6V symmetric CMOS output drive (120MHz
part)
•5V-tolerant inputs
•Port A pins capable of sourcing/sinking 24 mA
• Optional I/O synchronization to CPU core clock
Re-configurable Over The Internet
• Customer application program updatable
– Run-time self programming
• On-chip in-system programming interface
• On-chip in-system debugging support interface
• Debugging at full IP2022 operating speed
• Programming at device supply voltage level
•Real-time emulation, program debugging, and inte-
grated software development environment offered by
leading third-party tool vendors
4 www.ubicom.com
IP2012 / IP2022 Data Sheet
1.2 Architecture
1.2.1 CPU
The IP2012 and IP2022 implement an enhanced Harvard
architecture (i.e. separate instruction and data memories)
with independent address and data buses. The 16-bit
program memory and 8-bit dual-port data memory allow
instruction fetch and data operations to occur in parallel.
The advantage of this architecture is that instruction fetch
and memory transfers can be overlapped by a multistage
pipeline, so that the next instruction can be fetched from
program memory while the current instruction is executed
with data from the data memory.
Ubicom has developed a revolutionary RISC-based
architecture that is deterministic, jitter free, and
completely reprogrammable.
The architecture implements a four-stage pipeline (fetch,
decode, execute, and write back). At the maximum
operating frequency of 160 MHz, instructions are
executed at the rate of one per 6.25 ns clock cycle.
1.2.2 Serializer/Deserializers
One of the key elements in optimizing the IP2012 and
IP2022 for device-to-device and device-to-human
communication is the inclusion of on-chip
serializer/deserializer units. Each unit supports popular
communication protocols such as GPSI, SPI, UART, USB,
and 10Base-T Ethernet, allowing the IP2000 series
devices to be used in bridge, access point and gateway
applications.
By performing data serialization and deserialization in
hardware, the CPU bandwidth needed to support serial
communications is greatly reduced, especially at high
baud rates. Providing two units (IP2022 only) allows easy
implementation of protocol conversion or bridging
functions between two fast serial devices, such as USB-
to-Ethernet, GPSI to ethernet, or Ethernet to Ethernet. A
single SerDes unit (IP2012) provides the ability to bridge
RS232, SPI, or WLAN (802.11b) to Ethernet.
1.2.3 Low-Power Support
Particular attention has been paid to minimizing power
consumption. For example, an on-chip PLL allows use of
a lower-frequency external source (e.g., an inexpensive
4.8MHz crystal can be used to produce a 120 MHz on-
chip clock; a 3.2 MHz crystal to produce a 160 MHz on-
chip clock), which reduces both power consumption and
EMI. In addition, software can change the execution
speed of the CPU to reduce power consumption, and a
mechanism is provided for automatically changing the
speed on entry and return from an interrupt service
routine. The speed instruction specifies power-saving
modes that include a clock divisor between 1 and 128.
This divisor only affects the clock to the CPU core, not the
timers. The speed instruction also specifies the clock
source (OSC1 clock, RTCLK oscillator, or PLL clock
multiplier), and whether to disable the OSC1 clock
oscillator or the PLL. The speed instruction executes
using the current clock divisor.
1.2.4 Memory
The IP2012 / IP2022 CPU executes from a 32K ∞ 16 flash
program memory and an 8K ∞ 16 RAM program/data
memory. In addition, the ability to write into the program
flash memory allows flexible non-volatile data storage. An
interface is available (IP2022 only) for up to 128K bytes of
linearly addressed external memory, which can be
expanded to 2M bytes with additional software-based I/O
addressing. At 120 MHz operation, the maximum
execution rate is 40 MIPS from flash memory and 120
MIPS from RAM. At 160 MHz operation, the maximum
execution rate is 53.33 MIPS from flash memory and 160
MIPS from RAM. Speed-critical routines can be copied
from the flash memory to the RAM for faster execution.
The IP2000 series devices have a mechanism for in-
system programming of their flash and RAM program
memories through a four-wire SPI interface, and software
has the ability to reprogram the program memories at run
time. This allows the functionality of a device to be
changed in the field over the Internet.
1.2.5 Instruction Set
The IP2000 series instruction set, using 16-bit words,
implements a rich set of arithmetic and logical operations,
including signed and unsigned 8-bit ∞ 8-bit integer
multiply with a 16-bit product.
IP2012 / IP2022 Data Sheet
www.ubicom.com 5
1.2.6 Other Supported Functions
On-chip dedicated hardware also includes a PLL, an 8-
channel 10-bit ADC, general-purpose timers, single-cycle
multiplier, analog comparator, LFSR units, external
memory interface (IP2022 only), parallel slave port,
brown-out power voltage detector, watchdog timer, low-
power support, multi-source wakeup capability, user-
selectable clock modes, high-current outputs, and 52
general-purpose I/O pins (48 on IP2012).
1.2.7 Programming and Debugging
Support
The IP2000 series has advanced in-system programming
and debug support on-chip. This unobtrusive capability is
provided through the ISP/ISD interface. There is no need
for a bond-out chip for software development. This
eliminates concerns about differences in electrical
characteristics between a bond-out chip and the actual
chip used in the target application. Designers can test and
revise code on the same part used in the actual
application.
Ubicom provides the complete Red Hat GNUPro tools,
including C compiler, assembler, linker, utilities and GNU
debugger. In addition, Ubicom offers an integrated
graphical development environment which includes an
editor, project manager, graphical user interface for the
GNU debugger, device programmer, and ipModule™
configuration tool.
6 www.ubicom.com
IP2012 / IP2022 Data Sheet
2.0 Pin Definitions
2.1 PQFP (Plastic Quad Flat Package) for IP2022
Figure 2-1 IP2022 PQFP Pin Definition (Top View)
515-001b.eps
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
RG3
RG2
RG1
RG0
RF7
RF6
RF5
RF4
DVdd
DVss
IOVss
IOVdd
RF3
RF2
RF1
RF0
RE7
RE6
RE5
RE4
RE3
RE2
RE1
RE0
TSS
TSCK
TSI
TSO
RA0
RA1
RA2
RA3
DVdd
DVss
IOVss
IOVdd
RB0
RB1
RB2
RB3
RB4
RB5
RB6
RB7
RC0
RC1
RC2
RC3
IP2022/PQ80-120
or
IP2022/PQ80-160
80
79
78
77
76
75
74
73
72
71
70
69
68
67
RST
OSC2
OSC1
XVss
XVdd
RTCLK2
RTCLK1
DVss
DVdd
AVss
AVdd
RG7
RG6
66 RG4
65
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 GVdd
RG5
RC4
RC5
RC6
RC7
RD0
RD1
DVdd
DVss
IOVss
IOVdd
RD2
RD3
RD4
RD6
RD7
RD5
IP2012 / IP2022 Data Sheet
www.ubicom.com 7
2.2 PQFP (Plastic Quad Flat Package) for IP2012
Figure 2-2 IP2012 PQFP Pin Definition (Top View)
515-099.eps
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
RG3
RG2
RG1
RG0
RF7
RF6
RF5
RF4
DVdd
DVss
IOVss
IOVdd
RF3
RF2
NC
NC
RE7
RE6
RE5
RE4
RE3
RE2
RE1
RE0
TSS
TSCK
TSI
TSO
RA0
RA1
RA2
RA3
DVdd
DVss
IOVss
IOVdd
RB0
RB1
RB2
RB3
RB4
RB5
RB6
RB7
RC0
RC1
RC2
RC3
IP2012/PQ80-120
80
79
78
77
76
75
74
73
72
71
70
69
68
67
RST
OSC2
OSC1
XVss
XVdd
RTCLK2
RTCLK1
DVss
DVdd
AVss
AVdd
RG7
RG6
66 RG4
65
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 GVdd
RG5
RC4
RC5
RC6
RC7
NC
NC
DVdd
DVss
IOVss
IOVdd
RD2
RD3
RD4
RD6
RD7
RD5
IP2012 / IP2022 Data Sheet
www.ubicom.com 9
2.4 Signal Descriptions — IP2022
I = Digital Input, AI = Analog Input, O/DO = Digital Output, HiZ = High Impedance, P = Power, PLP = On-Chip Pullup,
ST = Schmitt Trigger
Table 2-1 Signal Descriptions
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
AVDD 70 B6 P Analog Supply
AVSS 71 A6 P Analog Ground
DVDD 9, 31,
56, 72
D1,D6,
E9,G5 P Logic Supply
DVSS 10, 32,
55, 73
E1,K5,
E10,D5 P Logic Ground
GVDD 65 A8 P I/O Port G supply
IOVDD 12, 34,
53
E5,G6,
E6
P I/O Supply (except Port G)
IOVSS 11, 33,
54
E2,K6,
E7
P I/O Ground (all ports)
XVDD 76 A4 P PLL Supply
XVSS 77 D4 P PLL Ground
OSC1 78 B4 I/ST Clock/Crystal Input
OSC2 79 A3 O/HiZ Crystal Output (tri-state if FUSE0 bit 15 = 1)
RST 80 A2 I/ST/
PLP
Reset Input. There is a weak pull-up on this pin, but
floating this pin does not guarantee Vih.
RTCLK1 74 A5 I Real-Time Clock/Crystal Input
RTCLK2 75 B5 O/HiZ Real-Time Crystal Output (tri-state if FUSE0 bit 14 =
1)
TSS 1A1I/ST/
PLP
Target SPI Slave Select (used only for in-system pro-
gramming and debug)
TSCK 2 C2 I/ST/
PLP
Target SPI Clock (used only for in-system program-
ming and debug)
TSI 3 B1 I/ST
/PLP
Target SPI Serial Data Input (used only for in-system
programming and debug)
TSO 4 B2 O/HiZ Target SPI Serial Data output (used only for in-sys-
tem programming and debug; high Z unless TSS low)
RA0 5 D2 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Capture 1 Input
RA1 6 C1 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Capture 2 Input
RA2 7 B3 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Clock Input
RA3 8 E4 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Output
RB0 13 F5 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Capture 1 Input
RB1 14 F1 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Capture 2 Input
RB2 15 F2 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Clock Input
10 www.ubicom.com
IP2012 / IP2022 Data Sheet
RB3 16 G1 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Output
RB4 17 F4 I/O 8 mA 8 mA I/O Port, External Interrupt, External Memory WR
RB5 18 J3 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
HOLD, External Memory RD
RB6 19 G2 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
R/W, External Memory LE
RB7 20 H1 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
CS, External Memory A0
RC0 21 J2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D8, External
Memory A9
RC1 22 H2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D9, External
Memory A10
RC2 23 J1 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D10, External
Memory A11
RC3 24 K1 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D11, External
Memory A12
RC4 25 K2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D12, External
Memory A13
RC5 26 K3 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D13, External
Memory A14
RC6 27 J4 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D14, External
Memory A15
RC7 28 K4 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D15, External
Memory A16
RD0 29 G4 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D0, External
Memory shared A1/D0
RD1 30 J5 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D1, External
Memory shared A2/D1
RD2 35 J6 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D2, External
Memory shared A3/D2
RD3 36 G7 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D3, External
Memory shared A4/D3
RD4 37 K7 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D4, External
Memory shared A5/D4
RD5 38 J7 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D5, External
Memory shared A6/D5
RD6 39 K8 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D6, External
Memory shared A7/D6
RD7 40 K9 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D7, External
Memory shared A8/D7
RE0 41 K10 I/O 8 mA 8 mA I/O Port, S1CLK - SCLK (SPI), RxCLK (GPSI),
optional SERDES clock input for UART or USB.
Table 2-1 Signal Descriptions (continued)
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
IP2012 / IP2022 Data Sheet
www.ubicom.com 11
RE1 42 H9 I/O 8 mA 8 mA I/O Port, S1RXP - VP (USB), SS (SPI Slave), TxEN
(GPSI Master), RxEN (GPSI Slave)
RE2 43 J10 I/O 8 mA 8 mA I/O Port, S1RXM - VM (USB)
RE3 44 J9 I/O 8 mA 8 mA I/O Port, S1RXD - RCV (USB), RXD (UART), DI
(SPI), TxD (GPSI Master), RxD (GPSI Slave)
RE4 45 H10 I/O 8 mA 8 mA I/O Port, S1TXPE/S1OE - TxD+ (Ethernet), OE
(USB), RxEN (GPSI Master), TxEN (GPSI Slave)
RE5 46 G9 I/O 24 mA 24 mA I/O Port, High Power Output, S1TXP - Tx+ (Ethernet),
VPO (USB), TXD (UART), DO (SPI), RxD (GPSI
Master), TxD (GPSI Slave)
RE6 47 G10 I/O 24 mA 24 mA I/O Port, High Power Output, S1TXM - Tx- (Ethernet),
VMO (USB), TxCLK/RxCLK (GPSI Master), TxCLK
(GPSI Slave)
RE7 48 J8 I/O 8 mA 8 mA I/O Port, S1TXME - TxD- (Ethernet), TxBUSY (GPSI)
RF0 49 F7 I/O 8 mA 8 mA I/O Port, S2TXPE/S2OE - TxD+ (Ethernet), OE
(USB), RxEN (GPSI Master), TxEN (GPSI Slave)
RF1 50 F9 I/O 24 mA 24 mA I/O Port, High Power Output, S2TXP - Tx+ (Ethernet),
VPO (USB), TXD (UART), DO (SPI), RxD (GPSI
Master), TxD (GPSI Slave)
RF2 51 F10 I/O 24 mA 24 mA I/O Port, High Power Output, S2TXM - Tx- (Ethernet),
VMO (USB), TxCLK/RxCLK (GPSI Master), TxCLK
(GPSI Slave)
RF3 52 F6 I/O 8 mA 8 mA I/O Port, S2TXME - TxD- (Ethernet), TxBUSY (GPSI)
RF4 57 B9 I/O 8 mA 8 mA I/O Port, S2CLK - SCLK (SPI), RxCLK (GPSI),
optional SERDES clock input for UART or USB.
RF5 58 A9 I/O 8 mA 8 mA I/O Port, S2RXP - VP (USB), SS (SPI Slave), TxEN
(GPSI Master), RxEN (GPSI Slave)
RF6 59 D10 I/O 8 mA 8 mA I/O Port, S2RXM - VM (USB)
RF7 60 D9 I/O 8 mA 8 mA I/O Port, S2RXD - RCV (USB), RXD (UART), DI
(SPI), TxD (GPSI Master), RxD (GPSI Slave)
RG0 61 C10 AI/DO 4 mA* 4 mA* Output Port, ADC0 Input, Comparator Output
RG1 62 C9 AI/DO 4 mA* 4 mA* Output Port, ADC1 Input, Comparator – Input
RG2 63 B10 AI/DO 4 mA* 4 mA* Output Port, ADC2 Input, Comparator + Input
RG3 64 A10 AI/DO 4 mA* 4 mA* Output Port, ADC3 Input, ADC reference Input
RG4 66 B7 AI/DO 4 mA* 4 mA* Output Port, ADC4 Input, S1RX-
RG5 67 B8 AI/DO 4 mA* 4 mA* Output Port, ADC5 Input, S1RX+
RG6 68 A7 AI/DO 4 mA* 4 mA* Output Port, ADC6 Input, S2RX-
RG7 69 D7 AI/DO 4 mA* 4 mA* Output Port, ADC7 Input, S2RX+
* GVDD = 2.5V
Table 2-1 Signal Descriptions (continued)
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
12 www.ubicom.com
IP2012 / IP2022 Data Sheet
2.5 Signal Descriptions — IP2012
I = Digital Input, AI = Analog Input, O/DO = Digital Output, HiZ = High Impedance, P = Power, PLP = On-Chip Pullup,
ST = Schmitt Trigger
Table 2-2 Signal Descriptions
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
AVDD 70 B6 P Analog Supply
AVSS 71 A6 P Analog Ground
DVDD 9, 31,
56, 72
D1,D6,
E9,G5 P Logic Supply
DVSS 10, 32,
55, 73
E1,K5,
E10,D5 P Logic Ground
GVDD 65 A8 P I/O Port G supply
IOVDD 12, 34,
53
E5,G6,
E6
P I/O Supply (except Port G)
IOVSS 11, 33,
54
E2,K6,
E7
P I/O Ground (all ports)
XVDD 76 A4 P PLL Supply
XVSS 77 D4 P PLL Ground
OSC1 78 B4 I/ST Clock/Crystal Input
OSC2 79 A3 O/HiZ Crystal Output (tri-state if FUSE0 bit 15 = 1)
RST 80 A2 I/ST/
PLP
Reset Input. There is a weak pull-up on this pin, but
floating this pin does not guarantee Vih.
RTCLK1 74 A5 I Real-Time Clock/Crystal Input
RTCLK2 75 B5 O/HiZ Real-Time Crystal Output (tri-state if FUSE0 bit 14 =
1)
TSS 1A1I/ST/
PLP
Target SPI Slave Select (used only for in-system pro-
gramming and debug)
TSCK 2 C2 I/ST/
PLP
Target SPI Clock (used only for in-system program-
ming and debug)
TSI 3 B1 I/ST
/PLP
Target SPI Serial Data Input (used only for in-system
programming and debug)
TSO 4 B2 O/HiZ Target SPI Serial Data output (used only for in-sys-
tem programming and debug; high Z unless TSS low)
RA0 5 D2 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Capture 1 Input
RA1 6 C1 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Capture 2 Input
RA2 7 B3 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Clock Input
RA3 8 E4 I/O 24 mA 24 mA I/O Port, High Power Output, Timer 1 Output
RB0 13 F5 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Capture 1 Input
RB1 14 F1 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Capture 2 Input
RB2 15 F2 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Clock Input
IP2012 / IP2022 Data Sheet
www.ubicom.com 13
RB3 16 G1 I/O 8 mA 8 mA I/O Port, External Interrupt, Timer 2 Output
RB4 17 F4 I/O 8 mA 8 mA I/O Port, External Interrupt
RB5 18 J3 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
HOLD
RB6 19 G2 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
R/W
RB7 20 H1 I/O 8 mA 8 mA I/O Port, External Interrupt, Parallel Slave Peripheral
CS
RC0 21 J2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D8
RC1 22 H2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D9
RC2 23 J1 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D10
RC3 24 K1 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D11
RC4 25 K2 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D12
RC5 26 K3 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D13
RC6 27 J4 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D14
RC7 28 K4 I/O 4 mA 4 mA I/O Port, Parallel Slave Peripheral Data D15
RD2 35 J6 I/O 4 mA 4 mA I/O Port
RD3 36 G7 I/O 4 mA 4 mA I/O Port
RD4 37 K7 I/O 4 mA 4 mA I/O Port
RD5 38 J7 I/O 4 mA 4 mA I/O Port
RD6 39 K8 I/O 4 mA 4 mA I/O Port
RD7 40 K9 I/O 4 mA 4 mA I/O Port
RE0 41 K10 I/O 8 mA 8 mA I/O Port, S1CLK - SCLK (SPI), RxCLK (GPSI),
optional SERDES clock input for UART or USB.
RE1 42 H9 I/O 8 mA 8 mA I/O Port, S1RXP - VP (USB), SS (SPI Slave), TxEN
(GPSI Master), RxEN (GPSI Slave)
RE2 43 J10 I/O 8 mA 8 mA I/O Port, S1RXM - VM (USB)
RE3 44 J9 I/O 8 mA 8 mA I/O Port, S1RXD - RCV (USB), RXD (UART), DI
(SPI), TxD (GPSI Master), RxD (GPSI Slave)
RE4 45 H10 I/O 8 mA 8 mA I/O Port, S1TXPE/S1OE - TxD+ (Ethernet), OE
(USB), RxEN (GPSI Master), TxEN (GPSI Slave)
RE5 46 G9 I/O 24 mA 24 mA I/O Port, High Power Output, S1TXP - Tx+ (Ethernet),
VPO (USB), TXD (UART), DO (SPI), RxD (GPSI
Master), TxD (GPSI Slave)
RE6 47 G10 I/O 24 mA 24 mA I/O Port, High Power Output, S1TXM - Tx- (Ethernet),
VMO (USB), TxCLK/RxCLK (GPSI Master), TxCLK
(GPSI Slave)
RE7 48 J8 I/O 8 mA 8 mA I/O Port, S1TXME - TxD- (Ethernet), TxBUSY (GPSI)
RF2 51 F10 I/O 24 mA 24 mA I/O Port, High Power Output
Table 2-2 Signal Descriptions (continued)
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
14 www.ubicom.com
IP2012 / IP2022 Data Sheet
RF3 52 F6 I/O 8 mA 8 mA I/O Port
RF4 57 B9 I/O 8 mA 8 mA I/O Port
RF5 58 A9 I/O 8 mA 8 mA I/O Port
RF6 59 D10 I/O 8 mA 8 mA I/O Port
RF7 60 D9 I/O 8 mA 8 mA I/O Port
RG0 61 C10 AI/DO 4 mA* 4 mA* Output Port, ADC0 Input, Comparator Output
RG1 62 C9 AI/DO 4 mA* 4 mA* Output Port, ADC1 Input, Comparator – Input
RG2 63 B10 AI/DO 4 mA* 4 mA* Output Port, ADC2 Input, Comparator + Input
RG3 64 A10 AI/DO 4 mA* 4 mA* Output Port, ADC3 Input, ADC reference Input
RG4 66 B7 AI/DO 4 mA* 4 mA* Output Port, ADC4 Input, S1RX-
RG5 67 B8 AI/DO 4 mA* 4 mA* Output Port, ADC5 Input, S1RX+
RG6 68 A7 AI/DO 4 mA* 4 mA* Output Port, ADC6 Input
RG7 69 D7 AI/DO 4 mA* 4 mA* Output Port, ADC7 Input
* GVDD = 2.5V
Table 2-2 Signal Descriptions (continued)
Name
Pin
Type
Sink
@ 3.3V
IOVDD
Source
@ 3.3V
IOVDD
Function
PQFP µBGA
IP2012 / IP2022 Data Sheet
www.ubicom.com 15
3.0 System Architecture
The IP2012 / IP2022 CPUs execute from a 32K ∞ 16 flash
program memory and an 8K ∞ 16 RAM program memory.
Figure 3-1 shows the IP2022 detailed block diagram, and
Figure 3-2 shows the IP2012 detailed block diagram. At
120 MHz operation, the maximum execution rate is 40
MIPS from Flash and 120 MIPS from RAM. At 160 MHz
operation, the maximum execution rate is 53.33 MIPS
from Flash and 160 MIPS from RAM. Speed-critical
routines can be copied from the flash memory to the RAM
for faster execution.
The CPU operates on 8-bit data in 128 special-purpose
registers, 128 global registers, and 3840 bytes of data
memory. The special-purpose registers hold control and
status bits used for CPU control and for interface with
hardware peripherals (timers, I/O ports, A/D converter,
etc.)Although the philosophy followed in the design of
Ubicom products emphasizes the use of fast RISC CPUs
with predictable execution times to emulate peripheral
devices in software (the ipModule™ concept), there are a
few hardware peripherals which are difficult to emulate in
software alone (e.g. an A/D converter) or consume an
excessive number of instruction cycles when operating at
high speed (e.g. data serialization/deserialization). The
design of the IP2012 / IP2022 incorporates only those
hardware peripherals which can greatly accelerate or
extend the reach of the ipModule™ concept. The
hardware peripherals included on-chip are:
• 52 I/O port pins (48 on IP2012)
•Watchdog timer
• Real-time timer
•2 Multifunction 16-bit timers with compare and cap-
ture registers
•2 Real-time 8-bit timers
•2 Serializer/deserializer (SERDES) units (IP2012 has
one unit)
•4 Linear feedback shift register (LFSR) units
• 10-bit, 8-channel A/D converter
• Analog comparator
•Parallel slave peripheral interface
• External SRAM interface (IP2022 only)
There is a single interrupt vector which can be
reprogrammed by software. On-chip peripherals and up to
8 external inputs can raise interrupts.
There are five sources of reset:
• RST external reset input
•Power-On Reset (POR) logic
•Brown-Out Reset (BOR) logic (detects low AVdd con-
dition)
•Watchdog timer reset
• In-system debugging/programming interface reset
An on-chip PLL clock multiplier (x50) enables high-speed
operation (up to 160 MHz) from a slow-speed external
Figure 3-1 IP2022 Detailed Block Diagram
Internal Data Bus
Port A
RTCLK
SxCLK
Real-Time
Clock Driver
OSC
Driver
Timer 1
(T1)
Port E Port F Port G
(2) Serializer/
Deserializers
Analog
Comparator,
2x Ethernet
Squelch
Timer 2
(T2)
Port B
Timer 0
(T0)
Interrupt
Port DPort C
Edge Det.
Parallel
8/16-bit Slave
Peripheral
ADC
Multiplexer
Divider
CPU Core Clock
Multiplexer
Divider
PLL
Divider
Divider
4KB Data
Memory
515-038b.eps
System
Clock
Real-Time
Timer
Reset
W
ALU
Watchdog Timer
with Pre-Scaler
ISD
Writeback
ISP
Brown
Out
POR
RST
Execute
Decode
64KB Flash
Program
Memory
16KB RAM
Program
Memory
Fetch
Internal
RC
Clock
(4) LFSR
Units
Ext. Memory
Interface
16 www.ubicom.com
IP2012 / IP2022 Data Sheet
clock input or crystal. A CPU clock-throttling mechanism
allows fine control over power consumption in modes that
do not require maximum speed, such as waiting for an
interrupt.
The IP200 series has a mechanism for in-system
programming of its flash and RAM program memories
through a four-wire SPI interface. This provides easy
programming and reprogramming of devices on
assembled circuit boards. In addition, the flash memory
can be programmed by software at run time, for example
to store user-specific data such as phone numbers and to
receive software upgrades downloaded over the Internet.
The devices also have an on-chip debugging facility which
makes the internal operation of the chip visible to third-
party debugging tools.
3.1 CPU Registers
Figure 3-3 shows the CPU registers, which consist of
seven 8-bit registers, seven 16-bit registers, and one 24-
bit register. The 16-bit registers are formed from pairs of
8-bit registers, and the 24-bit register is formed from three
8-bit registers. For the register quick reference guide, see
Section 7.0 and Section 7.1.
The W or working register is used as the source or
destination for most arithmetic, movement, and logical
instructions.
The STATUS register holds the condition flags for the
results of arithmetic and logical operations, the page bits
(used for jumps and subroutine calls), and bits which
indicate the skipping state of the core and control of
continuation skip after return from interrupt. Figure 3-4
shows the assignment of the bits in the STATUS register.
Figure 3-2 IP2012 Detailed Block Diagram
Internal Data Bus
Port A
RTCLK
SxCLK
Real-Time
Clock Driver
OSC
Driver
Timer 1
(T1)
Port E Port F Port G
(1) Serializer/
Deserializer
Analog
Comparator,
2x Ethernet
Squelch
Timer 2
(T2)
Port B
Timer 0
(T0)
Interrupt
Port DPort C
Edge Det.
Parallel
8-bit Slave
Peripheral
ADC
Multiplexer
Divider
CPU Core Clock
Multiplexer
Divider
PLL
Divider
Divider
4KB Data
Memory
515-100.eps
System
Clock
Real-Time
Timer
Reset
W
ALU
Watchdog Timer
with Pre-Scaler
ISD
Writeback
ISP
Brown
Out
POR
RST
Execute
Decode
64KB Flash
Program
Memory
16KB RAM
Program
Memory
Fetch
Internal
RC
Clock
(4) LFSR
Units
IP2012 / IP2022 Data Sheet
www.ubicom.com 17
.
•PA2:PA0—Program memory page select bits. Used to
extend the 13-bit address encoded in jump and call in-
structions (selects 8K-word pages). Modified using
the page instruction.
•SAR—Skip After Return bit. This bit should be set if
the core should be in the skipping state, and should
not be set if the core should not be in the skipping
state after the completion of the return instruction
(ret, retnp, or retw instructions, but not reti).
The return instruction will also clear the SAR control
bit to ensure correct behavior after the dynamic jump.
•SSF—Shadowed Skipping/not state Flag. Gives the
ISR the ability to know if the interrupt occurred imme-
diately following a skip instruction. The software can
choose either to clear the SSF flag in the ISR or to
make the first instruction of the context switching code
a nop to flush out the skip state.
•Z—Zero bit. Affected by most logical, arithmetic, and
data movement instructions. Set if the result was zero,
otherwise cleared.
•DC—Digit Carry bit. After addition, set if carry from bit
3 occurred, otherwise cleared. After subtraction,
cleared if borrow from bit 3 occurred, otherwise set.
•C—Carry bit. After addition, set if carry from bit 7 of
the result occurred, otherwise cleared. After subtrac-
tion, cleared if borrow from bit 7 of the result occurred,
otherwise set. After rotate (rr or rl) instructions,
loaded with the LSB or MSB of the operand, respec-
tively.
The MULH register receives the upper 8 bits of the 16-bit
product from signed or unsigned multiplication. The lower
8 bits are loaded into the W register.
The SPDREG register holds bits that control the CPU
speed and clock source settings, and is loaded by using
the speed instruction, as shown in Figure 3-5. The
SPDREG register is read-only, and its contents may only
be changed by executing a speed instruction, taking an
interrupt, or returning from an interrupt. For more
information about the speed instruction and the clock
throttling mechanism, see Section 3.4 and Figure 3-17.
Note: The speed instruction should be followed by a nop
instruction if Port B interrupt is used to wake up from sleep
mode.
• PLL—enable x50 PLL clock multiplier; 0 = enabled; 1
= disabled. Power consumption can be reduced by
disabling it. See Figure 3-17.
• OSC—enable OSC oscillator; 0 = enabled; 1 = dis-
abled (stops OSC oscillator and blocks propagation of
OSC1 external clock input). Power consumption can
be reduced by disabling it.
•CLK1:0—selects the system clock source, as shown
in Table 3-1. See Figure 3-17 for the clock logic. See
Section 7.1.5 (FCFG register, FRDTS1:0 bits) for ex-
ceptions.
Figure 3-3 CPU Registers
76543210
PA2:0 SAR SSF Z DC C
Figure 3-4 STATUS Register
INTVECH/INTVECL Register *
W Register
STATUS Register
MULH Register
SPDREG Register
IPH/IPL Register
DPH/DPL Register
SPH/SPL Register
DATAH/DATAL Register *
ADDRX/ADDRH/ADDRL Register
515-040b.eps
IPCH/IPCL Register *
7
15
0
0
Interrupt Registers
15 0
Pointer Registers
15
0
Program Memory/External Memory Interface Registers
PCH/PCL Register
15 XCFG Register
INTSPD Register
ADDRSEL Register
23
7
* Low byte doesn't carry to high byte
76543 0
PLL OSC CLK1:0 CDIV3:0
Figure 3-5 SPDREG Register
18 www.ubicom.com
IP2012 / IP2022 Data Sheet
•CDIV3:0—selects the clock divisor used to generate
the CPU core clock from the system clock, as shown
in Table 3-2 (also see Figure 3-17).
The INTSPD register holds bits that control the CPU
speed and clock source during interrupt service routines
(it is copied to the SPDREG register when an interrupt
occurs). It has the same format as the SPDREG register.
When the OSC crystal driver is stopped (SPDREG bit 6 =
1) and Port B or Real Time Timer interrupts are enabled,
then INTSPD bits 5 and 4 must not both be 0, because the
crystal startup time plus PLL startup time may be greater
than WUDP2:0 (see Figure 3-17).
The XCFG register holds additional control and status
bits, as shown in Figure 3-6.
•GIE—global interrupt enable bit. When set, interrupts
are enabled. When clear, interrupts are disabled. For
more information about interrupt processing, see
Section 3.7.
•FWP—flash write protect bit. When clear, writes to
flash memory are ignored. For more information
about programming the flash memory, see Section
4.7.
•RTEOS—real-time timer oversampling enable bit.
When set, oversampling is used. For more informa-
tion, see Section 5.3.
•RTOSC_EN—RTCLK oscillator enable bit. When
clear, the RTCLK oscillator is operational. When set,
the RTCLK oscillator is turned off.
•INT_EN—int instruction interrupt enable bit. When
set, int instructions cause interrupts. When clear,
int instructions only increment the PC, like nop.
•FBUSY—read-only flash memory busy bit. Set while
fetching instructions out of flash memory or while
busy processing an iread, ireadi, iwrite,
iwritei, fwrite, fread or ferase instruction
that operates on Flash, otherwise clear. For more in-
formation about programming the flash memory, see
Section 4.7.
The PCH and PCL register pair form a 16-bit program
counter. The PCH register is read-only. The PCL register
can be used to implement a lookup table, by moving a
variable to the w register, then executing an add PCL,w
instruction. If w=01 when the add occurs, the instruction
after the add will be skipped; if w=02, two instructions will
be skipped, etc.
The IPCH and IPCL register pair specifies the return
address when a reti instruction is executed.
The INTVECH and INTVECL register pair specifies the
interrupt vector. It has a default value of 0 following reset.
On a return from interrupt, an option of the reti
instruction allows software to save the incremented value
of the program counter in the INTVECH and INTVECL
registers.
The IPH and IPL register pair is used as a pointer for
indirect addressing. For more information about indirect
addressing, see Section 4.1.3.
The DPH and DPL register pair and the SPH and SPL
register pair are used as pointer registers for indirect-with-
offset addressing. For more information about indirect-
Table 3-1 CLK1:0 Field Encoding
CLK1:0 System Clock Source
00 PLL Clock Multiplier
01 OSC Oscillator/External OSC1 Input
10 RTCLK oscillator/external clock on
RTCLK1 input
11 System Clock Off
Table 3-2 System Clock Divisor
CDIV3:0
System
Clock
Divisor
CPU Core Frequency
120 MHz Sys-
tem Clock
160 MHz Sys-
tem Clock
0000 1 120 MHz 160 MHz
0001 2 60 MHz 80 MHz
0010 3 40 MHz 53.33 MHz
0011 4 30 MHz 40 MHz
0100 5 24 MHz 32 MHz
0101 6 20 MHz 26.66 MHz
0110 8 15 MHz 20 MHz
0111 10 12 MHz 16 MHz
1000 12 10 MHz 13.33 MHz
1001 16 7.5 MHz 10 MHz
1010 24 5 MHz 6.66 MHz
1011 32 3.75 MHz 5 MHz
1100 48 2.5 MHz 3.33 MHz
1101 64 1.875 MHz 2.5 MHz
1110 128 0.9375 MHz 1.25 MHz
1111 Clock Off 0 MHz 0 MHz
76 5 4 3 21 0
GIE FWP RTEOS RTOSC_EN INT_EN Rsvd FBUSY
Figure 3-6 XCFG Register
IP2012 / IP2022 Data Sheet
www.ubicom.com 19
with-offset addressing, see Section 4.1.4. The SPH and
SPL registers are automatically post-decremented when
storing to memory with a push instruction, and they are
automatically pre-incremented when reading from
memory with a pop instruction.
The ADDRSEL register holds an index to one of eight 24-
bit pointers used to address program memory. The current
program memory/external memory 24-bit address
selected by the ADDRSEL register is accessible in the
ADDRX (bits 23:16), ADDRH (bits 15:8), and ADDRL (bits
7:0) registers. The upper 5-bits of the ADDRSEL register
are unused. All 8 banks of 24-bits are initialized to
0x000000 upon reset.
Program memory is always read or written as 16-bit
words. On reads, the data from program memory is
loaded into the DATAH and DATAL register pair. On writes,
the contents of the DATAH and DATAL register pair are
loaded into the program memory.
3.2 Data Memory
Figure 3-7 is a map of the data memory. The special-
purpose registers and the first 128 data memory locations
(between addresses 0x080 and 0x0FF) can be accessed
with a direct addressing mode in which the absolute
address of the operand is encoded within the instruction.
The remaining 3840 bytes of data memory (between
addresses 0x100 and 0xFFF) must be accessed using
indirect or indirect-with-offset addressing modes. There is
one 16-bit register for the indirect address pointer, and two
16-bit registers for indirect-with-offset address pointers.
The offset is a 7-bit value encoded within the instruction.
For more information about the addressing modes, see
Section 4.1.
Figure 3-7 Data Memory Map
3.3 Program Memory
Figure 3-8 is a map of the program memory. A program
memory address in the INTVECH/INVECL, IPCH/IPCL,
or PCH/PCL registers or on the hardware stack is a word
address. However, the GNU software tools require byte
addresses when referring to locations in program
memory. An address loaded in the
ADDRX/ADDRH/ADDL register is a byte address.
The program memory is organized as 8K-word pages
(16K bytes). Single-instruction jumps and subroutine calls
are restricted to be within the same page. Longer jumps
and calls require using a page instruction to load the
upper address bits into the PA2:0 bits of the STATUS
register. The page instruction must immediately precede
the jump or call instruction. The PA2:0 bits should not be
modified by writing directly to the STATUS register,
because this may cause a mismatch between the PA2:0
bits in the STATUS register and the current program
counter (see Section 3.3.2). For more information about
the flash program memory, see Section 4.7 and Section
7.0.2.
Figure 3-8 Program Memory Map
External memory is not shown in Figure 3-8 because the
CPU cannot execute instructions directly out of external
memory. For more information about external memory,
see Section 5.11.
3.3.1 Loading the Program RAM
Software loads the program RAM from program flash
memory using the iread/ireadi and
iwrite/iwritei instructions. The iread instruction
reads the 16-bit word specified by the address held in the
ADDRX/ADDRH/ADDRL register. This word can be in
program flash memory, program RAM, or external
memory. When the iread instruction is executed, bits
515-028a.eps
127
Special-Purpose
Registers
3840 Bytes
Data Memory
0x001
70
0x080
0xFFF
128
Global Registers
0x100
0x07F
0x0FF
515-006b.eps
Program RAM
Reserved
(undefined data)
0x000000
15 0
0x004000
Flash Program Memory
0x014000 Flash Program Memory
0x018000 Flash Program Memory
0x01C000 Flash Program Memory
0x01FFFE
0x010000
0x0000
0x2000
0xA000
0xC000
0xE000
0xFFFF
0x8000
Byte
Address
Word
Address
0x003FFE
0x013FFE
0x017FFE
0x01BFFE
0x00FFFE
0x1FFF
0x9FFF
0xBFFF
0xDFFF
0x7FFF
20 www.ubicom.com
IP2012 / IP2022 Data Sheet
15:8 of the word are loaded into the DATAH register, and
bits 7:0 are loaded into the DATAL register. The address is
a word-aligned byte address (i.e. an address that is zero
in its LSB). The ireadi instruction is identical to the
iread instruction, except that it also increments the
address by 2.
The iwrite instruction writes the 16-bit word held in the
DATAH/DATAL registers to the program RAM location
specified by the address held in the
ADDRX/ADDRH/ADDRL register. The iwritei
instruction is identical, except that it also increments the
address by 2. For more information about the
iread/ireadi and iwrite/iwritei instructions,
see Section 4.7.
3.3.2 Program Counter
The program counter holds the 16-bit address of the
instruction to be executed. The lower eight bits of the
program counter are held in the PCL register, and the
upper eight bits are held in the PCH register. A write to the
PCL register will cause a jump to the 16-bit address
specified by the PCH and PCL registers. If the PCL
register is written as the destination of an add or addc
instruction and carry occurs, the PCH register is
automatically incremented. (This may cause a mismatch
between the PA2:0 bits in the STATUS register and the
current program counter, therefore it is strongly
recommended that direct modification of the PCL register
is only used for jumps within a page.) The PCH register is
read-only.
The PA2:0 bits in the STATUS register are not used for
address generation, except when a jump or subroutine
call instruction is executed. However, when an interrupt is
taken, the PA2:0 bits are automatically updated with the
upper three bits of the interrupt vector (INTVECH/L).
These bits are restored from the STATUS shadow register
when the interrupt service routine returns (i.e. executes a
reti instruction).
3.4 Low Power Support
Software can change the execution speed of the CPU to
reduce power consumption. A mechanism is also
provided for automatically changing the speed on entry
and return from the interrupt service routine. The speed
instruction specifies power-saving modes that include a
clock divisor between 1 and 128. This divisor only affects
the clock to the CPU core, not the timers, SERDES,
external memory or ADC (see Figure 3-17). The speed
instruction also specifies the clock source (OSC clock,
RTCLK oscillator, or PLL clock multiplier) and whether to
disable the OSC clock oscillator or the PLL.
For maximum power savings when running from the OSC
clock, disable the RTCLK oscillator (RTOSC_EN bit in the
XCFG register), disable the watchdog timer (WDTE bit in
the FUSE1 register), disable the A/D converter (ADCGO
bit in the ADCCFG register, disable the analog
comparator (CMPEN bit in the CMPCFG register) and
check that no flash operation is in progress (FBUSY bit in
the XCFG register) before executing a speed #$FF
instruction.
To summarize settings for lowest power:
• XCFG bit 4 = 1
• FUSE1 bit 3 = 0
• CMPCFG bit 7 = 0
• ADCCFG bit 3 = 0
• XCFG bit 0 = 0
Note: Before executing the speed instruction or
executing an interrupt (an interrupt will cause INTSPD to
be copied to SPDREG), insure that the FCFG register has
appropriate settings for the new clock frequency.
The SPDREG register (see Figure 3-5) holds the current
settings for the clock divisor, clock source, and disable
bits. These settings can be explicitly changed by
executing a speed instruction, and they change
automatically on interrupts. The SPDREG register is read-
only, and its contents may only be changed by executing
a speed instruction, taking an interrupt, or returning from
an interrupt. Two consecutive speed instructions are not
allowed. The INTSPD register specifies the settings used
during execution of the interrupt service routine. The
INTSPD register is both readable and writable.
On return from interrupts, the reti instruction includes a
bit that specifies whether the pre-interrupt speed is
restored or the current speed is maintained (see Table 3-
5).
The actual speed of the CPU is indicated by the SPDREG
register unless the specified speed is faster than the flash
access time and the program is executing out of flash.
When program execution moves from program RAM to
program flash memory, the new clock divisor will be the
greater (slower) of the clock divisor indicated by the
SPDREG register and the clock divisor required to avoid
violating the flash memory access time. The SPDREG
register does not indicate if the flash clock divisor is being
used. The speed indicated by the SPDREG will be
overridden only if the speed is too fast for the flash
memory.
The FCFG register holds bits that specify the minimum
number of system clock cycles for each flash memory
cycle (see Section 4.7.1).
IP2012 / IP2022 Data Sheet
www.ubicom.com 21
3.4.1 Clock Stop Mode (SLEEP)
When a speed instruction occurs, it is possible for the
CPU clock source to be disabled. The clock to the CPU
core may be disabled while the system clock is left
running, or the system clock may be disabled which also
disables the CPU core clock. See SPDREG, Section
7.1.18.
3.4.2 Wakeup
Recovery from SLEEP (core clock stop) mode to normal
execution is possible from these sources:
• External interrupts (i.e. Port B interrupts)
• Real-time timer interrupts
•Watchdog timer overflow reset
• Brown-out voltage reset
• RST external reset
The first two sources listed do not reset the chip, so
register and CPU states are maintained. The last three
sources reset the chip, so software must perform all of its
reset initialization tasks to recover. This usually requires
additional time, as compared to recovery through an
interrupt. If a Port B or Real Time Timer interrupt occurs
during core clock stop mode, the INTSPD register will be
copied to the SPDREG register, the ISR will be executed,
then mainline code will resume execution at the
instruction after the speed command that caused the
clock to stop.
Note: If wakeup triggers an ISR that has a reti
instruction which reinstates the pre-interrupt speed (see
Table 3-5), the device goes back to sleep. If a subsequent
wakeup occurs which does not reinstate the pre-interrupt
speed, then a nop must be inserted after the speed
instruction which puts it to sleep.
3.5 Speed Change
The speed instruction executes using the current clock
divisor. The new clock divisor takes effect with the
following instruction, as shown in the following code
example.
The automatic speed changes require a certain amount of
delay to take effect (see Figure 3-5 and Figure 3-17):
•Changing the Core Clock Divisor—there is no delay
when the clock divisor is changed (the instruction af-
ter the speed instruction is executed at the new
speed).
•Changing the System Clock Source—the delay is up
to one cycle of the slower clock. For example, chang-
ing between 4 MHz and 120 MHz could require up to
0.25 microseconds.
•Tu r ning on the OSC Clock Oscillator (clearing the
OSC bit in the SPDREG register)—the system clock
suspend time is specified in the WUDX2:0 bits in the
FUSE0 register (see Section 3.10.1).
•Tu r ning on the PLL Clock Multiplier (clearing the PLL
bit in the SPDREG register)—the system clock sus-
pend time is specified in the WUDP2:0 bits in the
FUSE0 register.
If both the OSC oscillator and PLL are re-enabled
simultaneously, the delay is controlled by only the
WUDP2:0 bits. Bits in the FUSE0 register are flash
memory cells which cannot be changed dynamically
during program execution.
3.6 Instruction Timing
All instructions that perform branches take 3 cycles to
complete, consisting of 1 cycle to execute and 2 cycles to
load the pipeline.
In the case of an automatic speed change, the execution
time will be with respect to the original speed and the
pipeline load time will be with respect to the new speed.
Conditional branching is implemented in the IP2012 /
IP2022 by using conditional skip instructions to branch
over an unconditional jump instruction. To support
conditional branching to other pages, the conditional skip
instructions will skip over two instructions if the first
instruction is a page instruction. The loadh and loadl
instructions also cause an additional instruction to be
skipped. When any of these conditions occur, it is called
an extended skip instruction.
nop ;assume divisor is 4, so this
;instruction takes 4 cycles
speed #0x06 ;change the divisor to 8,
;instruction takes 4 cycles
nop ;instruction takes 8 cycles
speed #0x00 ;change the divisor to 1,
;instruction takes 8 cycles
nop ;instruction takes 1 cycle
Table 3-3 Branch Timing
Instruction Execution Time Pipeline Load
Time
jmp 12
call 12
ret 12
reti 12
22 www.ubicom.com
IP2012 / IP2022 Data Sheet
Skip instructions take 1 cycle if they do not skip, or 2
cycles if they skip over one instruction. An extended skip
instruction may skip over more than one loadh, loadl,
or page instruction, however this operation is
interruptible and does not affect interrupt latency.
The iread and iwrite instructions take 4 cycles. The
multiply instructions take 1 cycle.
3.7 Interrupt Support
There are three types of interrupt sources:
•On-Chip Peripherals–the serializer/deserializer units,
real-time timer, timer 0, timer 1, and timer 2 are capa-
ble of generating interrupts. The Parallel Slave Periph-
eral does not generate interrupts on its own; it
requires programming one of the Port B external inter-
rupt inputs to generate interrupts on its behalf.
•External Interrupts–the eight pins on Port B can be
programmed to generate interrupts on either rising or
falling edges (see Section 5.1.1).
•int Instruction–the int instruction can be executed
by software to generate an interrupt. The INT_EN bit
can be considered as the interrupt flag for the int in-
struction, if the ISR checks for interrupt source. The
INT_EN bit in the XCFG register must be set to enable
the int instruction to trigger an interrupt. Because
the reti instruction returns to the int instruction,
the INT_EN bit must be cleared in the interrupt ser-
vice routine (ISR) before returning.
Figure 3-9 shows the system interrupt logic. Each
interrupt source has an interrupt enable bit. To be capable
of generating an interrupt, the interrupt enable bit and the
global interrupt enable (GIE) bit must be set.
Figure 3-9 System Interrupt Logic
3.7.1 Interrupt Processing
There is one interrupt vector held in the INTVECH and
INTVECL registers, which is reprogrammable by
software. When an interrupt is taken, the current PC is
saved in the IPCH and IPCL registers. On return from
interrupt (i.e. execution of the reti instruction), the PC is
restored from the IPCH and IPCL registers. Optionally, the
reti instruction may also copy the incremented PC to
the INTVECH and INTVECL registers before returning.
This has the effect of loading the INTVECH and INTVECL
registers with the address of the next instruction following
the reti instruction. This option can be used to directly
implement a state machine, such as a simple round-robin
scheduling mechanism for a series of interrupt service
routines (ISRs) in consecutive memory locations.
If multiple sources of interrupts have been enabled, the
ISR must check the interrupt flags of each source to
determine the cause of the interrupt. The ISR must clear
the interrupt flag for the source of the interrupt to prevent
retriggering of the interrupt on completion of the ISR (i.e.
execution of the reti instruction). Because the interrupt
logic adds a 2-cycle delay between clearing an interrupt
flag and deasserting the interrupt request to the CPU, the
flag must be cleared at least 2 cycles before the reti
instruction is taken.
When an interrupt is taken, the registers shown in Figure
3-10 are copied to a shadow register set. Each shadow
register is actually a 2-level push-down stack, so one level
of interrupt nesting is supported in hardware. The interrupt
processing mechanism is completely independent of the
16-level call/return stack used for subroutines.
The contents of the DATAH and DATAL registers are
pushed to their shadow registers 4 cycles after the
interrupt occurs, to protect the result of any pending
iread instruction. Therefore, software should not access
the DATAH or DATAL registers during the first instruction
of an ISR.
515-067c.eps
Interrupt
to CPU
Port B Interrupt
SerDes Interrupt
Timer 0 T0IF Bit
Real-Time Timer RTIF Bit
Timer 1 OFIF Bit
Timer 1 CAP2IF/CMP2IF Bit
Timer 1 CAP1IF Bit
Timer 1 CMP1IF Bit
Timer 2 OFIF Bit
int Instruction
GIE Bit
INT_EN Bit
Timer 2 CAP2IF/CMP2IF Bit
Timer 2 CAP1IF Bit
Timer 2 CMP1IF Bit
IP2012 / IP2022 Data Sheet
www.ubicom.com 23
Figure 3-10 Interrupt Processing (On Entry to the ISR)
Note: On entry to the ISR the W, MULH, IPH/IPL, DPH/DPL, SPH/SPL, ADDRSEL and DATAH/DATAL register values
don’t change from their mainline code values (they are copied to their shadow registers).
INTVECH/INTVECL
Register
W
Register
STATUS
Register
MULH
Register
IPH/IPL
Register
DPH/DPL
Register
SPH/SPL
Register
DATAH/DATAL
Register
515-068d.eps
W
Shadow Register 1
STATUS
Shadow Register 1
MULH
Shadow Register 1
IPH/IPL
Shadow Register 1
DPH/DPL
Shadow Register 1
SPH/SPL
Shadow Register 1
DATAH/DATAL
Shadow Register 1
IPCH/IPCL
Register
PC
W
Shadow Register 2
STATUS
Shadow Register 2
MULH
Shadow Register 2
IPH/IPL
Shadow Register 2
DPH/DPL
Shadow Register 2
SPH/SPL
Shadow Register 2
DATAH/DATAL
Shadow Register 2
INTSPD
Register
SPDREG
Register
SPDREG
Shadow Register 1
SPDREG
Shadow Register 2
IPCH/IPCL
Shadow Register
INTVECH
bits 7:5 copied
GIE
XCFG bit 7 GIE Shadow bit 1 GIE Shadow bit 2
GIE = 0
24 www.ubicom.com
IP2012 / IP2022 Data Sheet
On return from the ISR, these registers are restored from the shadow registers, as shown in Figure 3-11.
Figure 3-11 Interrupt Return Processing (upon execution of reti)
IPCH/IPCL
Register
W
Register
STATUS
Register
MULH
Register
W
Shadow Register 1
STATUS
Shadow Register 1
MULH
Shadow Register 1
IPH/IPL
Register
DPH/DPL
Register
SPH/SPL
Register
DATAH/DATAL
Register
IPH/IPL
Shadow Register 1
DPH/DPL
Shadow Register 1
SPH/SPL
Shadow Register 1
DATAH/DATAL
Shadow Register 1
515-069c.eps
PC
INTVECH/INTVECL
Register
PC + 1
If reti
Instruction
Bit 1 is Set
W
Shadow Register 2
STATUS
Shadow Register 2
MULH
Shadow Register 2
IPH/IPL
Shadow Register 2
DPH/DPL
Shadow Register 2
SPH/SPL
Shadow Register 2
DATAH/DATAL
Shadow Register 2
SPDREG
Register
If reti
Instruction
Bit 2 is Set
SPDREG
Shadow Register 1
SPDREG
Shadow Register 2
T0TMR
Register
T0TMR + W
If reti
Instruction
Bit 0 is Set
GIE
XCFG bit 7
GIE Shadow bit 1
GIE Shadow bit 2
IPCH/IPCL
Shadow
IP2012 / IP2022 Data Sheet
www.ubicom.com 25
3.7.2 Global Interrupt Enable Bit
The GIE bit serves two purposes:
•Preventing an interrupt in a critical section of mainline
code
• Supporting nested interrupts
The GIE bit is automatically cleared when an interrupt
occurs, to disable interrupts while the ISR is executing.
The GIE bit is automatically set by the reti instruction to
re-enable interrupts when the ISR returns.
To re-enable interrupts during ISR execution, the ISR
code must first clear the source of the first interrupt. It may
also be desirable to disable specific interrupts before
setting the GIE bit to provide interrupt prioritization. Even
with GIE deasserted, interrupt triggers are still captured
but an interrupt won’t be triggered until GIE is re-enabled.
Caution must be taken not to exceed the interrupt shadow
register stack depth of 2.
Clearing the GIE bit in the ISR cannot be used to globally
disable interrupts so that they remain disabled when the
ISR returns, because the reti instruction automatically
sets the GIE bit. To disable interrupts in the ISR so that
they remain disabled after the ISR returns, the individual
interrupt enable bits for each source of interrupts must be
cleared.
3.7.3 Interrupt Latency
The interrupt latency is the time from the interrupt event
occurring to first ISR instruction being latched from the
decode to the execute stage (see Section 4.3). If the
interrupt comes from a Port B input and the SYNC bit in
the FUSE1 register is 0, an additional two core clock
cycles of synchronization delay are added to the interrupt
latency.
The ireadi or iwritei instructions are blocking (i.e.
prevent other instructions and interrupts from being
executed) for 4 core clock cycles. The iread or
iwrite instructions are blocking for 4 core clock cycles
while operating on program RAM, and non-blocking
(single cycle) while operating on external memory.
When an interrupt event is triggered, the CPU speed is
changed to the speed specified by the INTSPD register
(the SPDREG register is copied to a shadow register, then
loaded with the value from the INTSPD register).
If INTSPD is set the same as SPDREG when an interrupt
occurs, then the interrupt latency is 3 core clock cycles for
synchronous interrupts. If not, then the interrupt latency is
3 core clock cycles, plus the speed change (delay
described in Section 3.5).
3.7.4 Return From Interrupt
The reti instruction word includes three bits which
control its operation, as shown in Table 3-5. The three bits
are specified from assembly language in a literal (e.g.
reti #0x7 to specify all bits as 1).
Updating the interrupt vector allows the programmer to
implement a sequential state machine. The next interrupt
will resume the code directly after the previous reti
instruction.
The reti instruction takes 1 cycle to execute, and there
is a further delay of 2 cycles at the mainline code speed to
load the pipeline before the mainline code is resumed.
Table 3-4 GIE Bit Handling
Event Effect
Enter ISR (interrupt) GIE bit is cleared
Exit ISR
(reti instruction)
GIE bit is set
setb xcfg,7
instruction (inside ISR)
Enable interrupts for
nested interrupt support
clrb xcfg,7
instruction (inside ISR)
Nothing, the GIE bit is
already clear
setb xcfg,7
instruction (mainline code)
Enable interrupts
clrb xcfg,7
instruction (mainline code)
Disable interrupts
Table 3-5 reti Instruction Options
Bit Function
2 Reinstate the pre-interrupt speed
1 = enable, 0 = disable
1 Store the PC+1 value in the INTVECH and
INTVECL registers
1 = enable, 0 = disable
0 Add W to the T0TMR register
1 = enable, 0 = disable
26 www.ubicom.com
IP2012 / IP2022 Data Sheet
Note: If RETI can return to Flash program memory, insure
that all Flash reads or writes are complete (XCFG bit 0 =
0) before RETI is executed.
3.7.5 Disabled Interrupt Resources
If a peripheral is disabled and its interrupt flag is cleared,
the peripheral does not have the ability to set an interrupt
flag. The interrupt flag, however, is still a valid source of
interrupt (If software sets an interrupt flag, the
corresponding interrupt enable bit is set, and the GIE bit
is set, then the CPU will be interrupted whether or not the
peripheral is enabled or disabled).
If a peripheral is disabled inside the ISR, then its interrupt
flag must be cleared to prevent an undesired interrupt
from being taken when the ISR completes or when GIE is
enabled (enabling nested interrupts - see Section 3.7.2).
3.8 Reset
There are five sources of reset:
•Power-On Reset (POR; reset occurs at power up)
• Brown-Out Reset (BOR)
•Watchdog Reset
• External Reset (from the RST pin)
•Tool Reset (from the debugging interface)
Each of these reset conditions causes the program
counter to branch to the reset vector at the top of the
program memory (word address 0xFFF0 or byte address
0x1FFE0).
The IP2012 / IP2022 incorporates a Power-On Reset
(POR) detector that generates an internal reset as DVdd
rises during power-up. Figure 3-12 is a block diagram of
the reset logic. The startup timer controls the reset time-
out delay. The reset latch controls the internal reset signal.
On power-up, the reset latch is cleared (CPU held in
reset), and the startup timer starts counting once it
detects a valid logic high signal on the RST pin. Once the
startup timer reaches the end of the timeout period, the
reset latch is cleared, releasing the CPU from reset.
Note: CPU operation does not start until the CPU is
released from reset and valid core clocks are received
past the system clock suspend circuit (see WUDX block in
Figure 3-17). So, for a POR with FUSE0 register
WUDX=350us, for example, the core starts operation
~70ms after power up. For a POR with WUDX= 1.1sec,
the core starts operation ~1.1sec after power up.
The PSPCFG (address 0x06E) register contains two bits
to indicate possible sources of the reset, WD and BO. The
WD bit is cleared on reset unless the reset was caused by
the watchdog timer, in which case the WD bit is set. The
BO bit is cleared on reset unless the reset was caused by
the brown-out logic, in which case, the BO bit is set.
Figure 3-13 shows a power-up sequence in which RST is
not tied to the DVDD pin and the DVDD signal is allowed
to rise and stabilize before RST pin is brought high. The
WUDX2:0 bits of the FUSE0 register specify the length of
time from the rising edge of RST until the device leaves
reset.
Figure 3-12 On-Chip Reset Circuit Block Diagram
Start-Up
Timer
Brown-Out
Detection
RST
DVDD
Watchdog Timer Overflow
Watchdog
RC Clock
(~14KHz)
Time-Out
Internal
Reset Signal
(active low)
POR Power-On Reset
515-023d.eps
Clear
For RST, Tool Reset or Watchdog = WUDX2:0 (FUSE0)
For POR or BOR ≈ (1025 x 70us) ≈ 70ms
AVDD
Tool Reset
Core Clock
Core Reset
(active low,
initially low)
FF FF
IP2012 / IP2022 Data Sheet
www.ubicom.com 27
Figure 3-13 Power-Up, Separate RST Signal
Figure 3-14 shows the on-chip Power-On Reset sequence
in which the RST and DVDD pins are tied together. The
DVDD signal is stable before the startup timer expires. In
this case, the CPU receives a reliable reset.
Figure 3-14 Power-On Reset, RST Tied To DVdd
However, Figure 3-15 depicts a situation in which DVDD
rises too slowly. In this scenario, the startup timer will time
out prior to DVDD reaching a valid operating voltage level
(DVDD min). This means the CPU will come out of reset
and start operating with the supply voltage below the level
required for reliable performance. In this situation, an
external RC circuit is recommended for driving RST. The
RC delay should exceed five times the time period
required for DVDD to reach a valid operating voltage.
Figure 3-15 DVdd Rise Time Exceeds Tstartup
Figure 3-16 shows the recommended external reset
circuit. The external reset circuit is required only if the
DVDD rise time has the possibility of being too slow (refer
to SVdd specification in Section 8.4).
Figure 3-16 External Reset Circuit
The diode D discharges the capacitor when DVDD is
powered down.
R1 = 100 Ω to 1K Ω will limit any current flowing into RST
from external capacitor C1. This protects the RST pin from
breakdown due to Electrostatic Discharge (ESD) or
Electrical Overstress (EOS).
R2 < 40K Ω is recommended to make sure that voltage
drop across R2 leaves the RST pin above a Vih level.
C1 should be chosen so that R2 ∞ C1 exceeds five times
the time period required for DVDD to reach a valid
operating voltage.
515-019c.eps
DVDD
RST
POR
WUDX
Startup Timer
(Time-Out)
Internal
Reset Signal
DVDD
RST
POR
Startup Timer
(Time-Out)
Internal
Reset Signal
515-022e.eps
70ms
DVDD
RST
POR
Startup Timer
(Time-Out)
Internal
Reset Signal 515-020e.eps
70ms
DVDD
C1
R2
R1
515-021a.eps
RST
IP2022
28 www.ubicom.com
IP2012 / IP2022 Data Sheet
3.8.1 Brown-Out Detector
The on-chip brown-out detection circuitry resets the CPU
when AVdd dips below the brown-out voltage level
programmed in the BOR2:0 bits of the FUSE1 register
(refer to Section 3.10.2). Bits in the FUSE1 register are
flash memory cells which cannot be changed dynamically
during program execution.
The device is held in reset as long as AVdd stays below
the brown-out voltage. The CPU will come out of reset
when AVdd rises between 100mV and 200mV above the
brown-out voltage (the CPU may never come out of
brownout reset, even after AVdd returns to acceptable
level, if the brownout setting is too high). Therefore, the
2.10V setting is recommended. The brown-out level can
be programmed using the BOR2:0 bits in the FUSE1
register, as shown in Section 3.10.2.
3.8.2 Reset and Interrupt Vectors
After reset, the PC is loaded with 0xFFF0, which is near
the top of the program memory space. Typical activities for
the reset initialization code include:
•Setting up the FCFG register with appropriate values
for flash timing compensation.
•Issuing a speed instruction to initialize the CPU core
clock speed.
•Checking for the cause of reset (brown-out voltage,
watchdog timer overflow, or other cause). In some ap-
plications, a “warm” reset allows some data initializa-
tion procedures to be skipped.
•Copying speed-critical sections of code from flash
memory to program RAM.
•Setting up data memory structures (stacks, tables,
etc.).
• Initializing peripherals for operation (timers, etc.).
•Initializing the dynamic interrupt vector and enabling
interrupts.
Because the default interrupt vector location is 0, which is
in program RAM, interrupts should not be enabled until
the ISR is loaded in shadow RAM or the dynamic interrupt
vector is loaded with the address of an ISR in flash
memory. There is a single dynamic interrupt vector shared
by all interrupts. The interrupt vector can be changed by
loading the INTVECH and INTVECL registers, or by
issuing a reti instruction with an option specifying that
the interrupt vector should be updated with the current PC
value plus 1.
3.8.3 Register States Following Reset
The effect of different reset sources on a register depends
on the register and the type of reset operation. Some
registers are initialized to specific values, some are left
unchanged, and some are undefined.
A register that starts with an unknown value should be
initialized by the software to a known value if it is going to
be used (no need to initialize unused registers nor data
memory). Do not simply test the initial state and rely on it
starting in that state consistently. See Table 7-1 for more
detailed information.
IP2012 / IP2022 Data Sheet
www.ubicom.com 29
3.9 Clock Oscillator
There are two clock oscillators, the OSC oscillator and the
RTCLK oscillator. Using the PLL clock multiplier, the OSC
clock is intended to provide the time base for running the
CPU core at speeds up to 120MHz for the standard
version, and at 160 MHz for the faster version. The RTCLK
oscillator operates at 32.768kHz using an external crystal.
This oscillator is intended for running the real-time timer
when the OSC oscillator and PLL clock multiplier are
turned off. Either clock source can be driven by an
external clock signal up to 120MHz.
Figure 3-17 shows the clock logic for the IP2022, and
Figure 3-18 shows that of the IP2012. The PLL clock
multiplier has a fixed multiplication factor of 50. The PLL is
preceded by a divider capable of any integer divisor
between 1 and 8, as controlled by the PIN2:0 bits of the
FUSE0 register (refer to Section 3.10.1). The PLL is
followed by a second divider capable of any integer divisor
between 1 and 4, as controlled by the POUT1:0 bits of the
FUSE0 register. A third divider which only affects the clock
to the CPU core is controlled by the speed change
mechanism described in Section 3.4. See Section 3.10.1
for a description of the FUSE0 WUDX2:0 and WUDP2:0
bits.
Note: Bits in the FUSE0 register are flash memory cells
which cannot be changed dynamically during program
execution.
Figure 3-17 IP2022-120 and IP2012-120 Clock Logic
515-002h.eps
Timer 1
Timer 0
CPU Core
0–120MHz
Timer 2
RTTMR
SERDES Clock
SPDREG
Divider
Post-
Scaler
Pre-
Scaler
RTCLK
Driver
RTCLK1
RTCLK2
50X PLL
Clock
Multiplier
OSC
Driver
OSC1
OSC2
System
Clock
0–120
MHz
PLL Bypass
External Memory Logic
(IP2022 only)
IP2022-120, IP2012-120:
Crystal 4.75–5 MHz
Ext. 0–120 MHz
Crystal 32.768 kHz
Ext. 0–120 MHz
4.75–5
MHz
FUSE0
Register
(bits 11:9)
FUSE0
Register
(bits 13:12)
237.5–250
MHz
speed
Instruction
(SPDREG
bits 5:4)
speed
Instruction
(SPDREG
bits 3:0)
0–120
MHz
RTCFG Register, RTSS bit
SxMode Register (CLKS1:0)
ADC
WUDX2:0
WUDP2:0
WUDX2:0
SxCLK (RE0 or RF4)
FUSE0
0
1
11
10
01
00
01
00
10
11
30 www.ubicom.com
IP2012 / IP2022 Data Sheet
Figure 3-18 IP2022-160 Clock Logic
3.9.1 External Connections
Figure 3-19 shows the connections for driving the OSC
and/or RTCLK clock sources with an external signal. To
drive the OSC clock source, the external clock signal is
driven on the OSC1 pin and the OSC2 pin is left open. The
external clock signal driven on the OSC1 pin may be any
frequency up to 120 MHz. To drive the RTCLK clock
source, the external clock signal is driven on the RTCLK1
input and the RTCLK2 output is left open. The external
clock signal driven on the RTCLK1 pin may be any
frequency up to 120 MHz.
Figure 3-19 External Clock Inputs
Figure 3-20 shows the connections for attaching a crystal
to the OSC and/or RTCLK oscillator. For the OSC
oscillator, a crystal is connected across the OSC1 and
OSC2 pins. For the RTCLK oscillator, a 32.768kHz crystal
is connected across the RTCLK1 and RTCLK2 pins.
There is about 4pf of capacitance on each of OSC1 and
OSC2 pins to DVss and about 10pf of capacitance on
each of RTCLK1 and RTCLK2 pins to DVss. There is also
an internal feedback resistor (no external feedback
resistor needed). For the OSC crystal, a parallel resonant
crystal is recommended that has a maximum ESR of 100
ohms and a load capacitance rating of 12pF (requires
24pF on each of OSC1 and OSC2 pins). For the optional
RTCLK crystal, a parallel resonant crystal is
recommended that has a maximum ESR of 50K ohms,
and a load capacitance rating of 12.5pF (requires 25pF on
each of RTCLK1 and RTCLK2 pins).
The crystal manufacturer’s load capacitance rating (CL)
should be equal to (C1 x C2)/(C1 + C2), where C1 =
capacitance on OSC1 (4pF + stray board capacitance +
added capacitance), and C2 = capacitance on OSC2 (4pF
+ stray board capacitance + added capacitance). The
trace length between the OSC pins and the crystal should
be as short as possible, to avoid noise coupling.
When RTCLK1 is unused, it should be tied to GND.
515-101.eps
Timer 1
Timer 0
CPU Core
0–120MHz, 160MHz
Timer 2
RTTMR
SERDES Clock
SPDREG
Divider
Post-
Scaler
Pre-
Scaler
RTCLK
Driver
RTCLK1
RTCLK2
50X PLL
Clock
Multiplier
OSC
Driver
OSC1
OSC2
System
Clock
0–120, 160
MHz
PLL Bypass
External Memory Logic
Crystal 32.768 kHz
Ext. 0–120 MHz
3.2
MHz
or
4.75-5
MHz
FUSE0
Register
(bits 11:9)
FUSE0
Register
(bits 13:12)
160
MHz
or
237.5-250
MHz
speed
Instruction
(SPDREG
bits 5:4)
speed
Instruction
(SPDREG
bits 3:0)
0–120
MHz
RTCFG Register, RTSS bit
SxMode Register (CLKS1:0)
ADC
WUDX2:0
WUDP2:0
WUDX2:0
SxCLK (RE0 or RF4)
FUSE0
0
1
11
10
01
00
01
00
10
11
IP2022-160:
Crystal 4.75–5 MHz,
3.2 MHz
Ext. 0–120 MHz
RTCLK1 RTCLK2
Externally
Generated Clock
Open
515-024a.eps
IP2012/IP2022
OSC1 OSC2
Externally
Generated Clock
Open
FUSE0
bit 15=1
FUSE0
bit 14=1
IP2012 / IP2022 Data Sheet
www.ubicom.com 31
Figure 3-20 Crystal Connection
3.10 Configuration Block
The configuration block is a set of flash memory registers
outside of both program memory and data memory. These
registers are not readable or writable at run time.
The FUSE0, FUSE1 registers hold settings that must be
specified by system designers. The other configuration
block registers are used by software tools. The
configuration block is readable but not writable when
Code Protection is enabled. Table 3-6 lists the
configuration block registers.
515-025c.eps
Crystal
Crystal
IP2012/IP2022
OSC1 OSC2
XTAL
(FUSE0 bit 15=0)
RTCLK1 RTCLK2
RTCLK
(FUSE0 bit 14=0)
20pF 20pF 15pF 15pF
Table 3-6 Configuration Block
Word Address Words Name Description
0x00010000 1 FUSE0 FUSE0 register
0x00010001 1 FUSE1 FUSE1 register
0x00010002 to 0x00010003 2 - Reserved
0x00010004 1 TRIM0 TRIM0 register, factory programmed to
FBFE
0x00010005 to 0x0001001D 25 - Reserved
0x0001001E-
0x0001001F
2 FREQ OSC1 input frequency during device pro-
gramming - used by tools only
0x00010020 to 0x00010027 8 VCOMPANY Company name
0x00010028 to 0x0001002F 8 VPRODUCT Product name
0x00010030 to 0x00010031 2 VVERSION Software version
0x00010032 to 0x00010033 2 VSOFTDATE Software date
0x00010034 to 0x00010035 2 VPROGDATE Programming date
0x00010036 to 0x0001003F 10 - Reserved
Total 16-bit words 64
32 www.ubicom.com
IP2012 / IP2022 Data Sheet
3.10.1 FUSE0 Register (not run-time programmable)
15 14 13 12 11 10 9876543210
XTAL RTCLK POUT1:0 PIN2:0 Reserved WUDP2:0 WUDX2:0
Figure 3-21 FUSE0 Register
XTAL OSC2 crystal drive output
0 = Enabled — Use for crystal clock
1 = Disabled — Use for external clock
RTCLK RTCLK2 crystal drive output
0 = Enabled — Use for crystal clock
1 = Disabled — Use for external clock
POUT1:0 Specifies PLL clock multiplier postscaler divisor
00 = 1 (reserved)
01 = 2
10 = 3
11 = 4
PIN2:0 Specifies PLL clock multiplier prescaler divisor
000 = 1 100 = 5
001 = 2 101 = 6
010 = 3 110 = 7
011 = 4 111 = 8
WUDP2:0 Specifies system clock suspend time during PLL startup (after a speed instruction clears the PLL bit in
the SPDREG register).
000 = 140 µs 100 = 1 ms
001 = 210 µs 101 = 2 ms
010 = 350 µs 110 = 5 ms
011 = 630 µs 111 = 9 ms
WUDX2:0 Specifies system clock suspend time during OSC and RTCLK start-up. Used to keep the clock from prop-
agating to the core before the crystal achieves valid signal levels (see Figure 3-17). Also keeps RST
asserted except for POR and BOR.
000 = 450 µs
001 = 1 ms
010 = 5 ms
011 = 9 ms
100 = 18 ms
101 = 72 ms
110 = 574 ms†
1147 ms†
†. Clock suspend time after POR is twice this long if the Watchdog is enabled in FUSE1 to a value less than
WUDX.
111 =
IP2012 / IP2022 Data Sheet
www.ubicom.com 33
3.10.2 FUSE1 Register (not run-time programmable)
15 14 13 7 6 5 4 3 2 1 0
CP SYNC Reserved BOR2:0 WDTE WDPS2:0
Figure 3-22 FUSE1 Register
CP Clear to enable code protection. Once cleared, this bit cannot be set until the entire device is erased.
When code protection is enabled, program memory reads as all 0 to an external device programmer.
This bit does not affect access to program memory made by software, using the iread, ireadi,
iwrite, iwritei, ferase, fwrite and fread instructions. In-system debugging is not
available when code protection is enabled. Code protection does not protect the configuration block
against reading, only against writing. Note: After clearing this bit during programming, Code Protect is
not activated until the part is powered down or reset.
0 = enabled
disabled
1 =
SYNC Set to read directly from the port pins through the RxIN register, clear to read through a CPU core
clock synchronization register. This bit should be clear if any external devices that can be read from
I/O port pins are running asynchronously to the CPU core clock. See Figure 5-1.
0 = enabled
disabled
1 =
BOR2:0 Specifies brown-out voltage level. If AVdd goes below this level, the IP2012 / IP2022 is reset. This set-
ting should be at least 0.2V below the minimum operating AVdd, because there is a maximum of 0.2V
hysteresis to leave brownout reset after power up and after brownout reset occurs.
000 = 2.30V ± 0.1V Do not use unless AVdd ≥ 2.50V
001 = 2.25V ± 0.1V Do not use unless AVdd ≥ 2.45V
010 = 2.20V ± 0.1V
011 = 2.15V ± 0.1V
100 = 2.10V ± 0.1V Recommended
101 = Reserved
110 = Reserved
111 = Disabled, no brown-out reset can occur.
WDTE Enables Watchdog Timer in run mode. Disabled in debug mode regardless of this bit.
0 = disabled
enabled
1 =
WDPS2:0 Specifies the Watchdog Timer prescaler divisor. This controls the time period before the Watchdog
Timer expires. If the Watchdog Timer is enabled, software must clear the Watchdog Timer periodically
within this time period to prevent a reset of the IP2012 / IP2022 from occurring. The cwdt instruction
or any reset clears both the Watchdog Timer and its prescaler. Care must be taken to ensure that this
setting is not greater than the maximum crystal start-up time plus the time required to get to the first
cwdt instruction.
000 = 1 (~20 ms) = 256 x WRC 100 = 16 (~320 ms)
001 = 2 (~40 ms)
4 (~80 ms)
8 (~160 ms)
101 = 32 (~640 ms)
010 = 110 = 64 (~1280 ms)
011 = 111 = 128 (~2560 ms)
34 www.ubicom.com
IP2012 / IP2022 Data Sheet
3.10.3 TRIM0 Register
Factory programmed to $FBFE for 120MHz versions,
$FBFD for 160MHz versions operating at 160MHz. Must
leave at factory default for proper operation.
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
SQUELT3:0 SQUELT5 FPERT CMPT2:0 SQUELT4 VCOT3 SQUELT7:6 VCOT2:0
Figure 3-23 TRIM0 Register
SQUELT7:0 SERDES squelch trim bits
FPERT Controls flash block pulse erase, both for self-programming ferase and for the FERASE command
from the ISD/ISP interface
0 = 20 ms, if OSC1 frequency and FCFG register settings are optimal
1 = Reserved - 10ms block erase (do not use)
CMPT2:0 Comparator offset trim bits
VCOT3:0 PLL VCO frequency trim bits
1110 = 4.75–5.0 MHz into PLL
1101 = 3.2 MHz into PLL
IP2012 / IP2022 Data Sheet
www.ubicom.com 35
4.0 Instruction Set Architecture
The IP2012 / IP2022 implements a powerful load-store
RISC architecture with a rich set of arithmetic and logical
operations, including signed and unsigned 8-bit ∞ 8-bit
integer multiply with a 16-bit product.
The CPU operates on data held in 128 special-purpose
registers, 128 global registers, and 3840 bytes of data
memory. The special-purpose registers are dedicated to
control and status functions for the CPU and peripherals.
The global registers and data memory may be used for
any functions required by software, the only distinction
among them being that the 128 global registers
(addresses 0x080 to 0x0FF) can be accessed using a
direct addressing mode. The remaining 3840 bytes of data
memory (between addresses 0x100 and 0xFFF) must be
accessed using indirect or indirect-with-offset addressing
modes. The IPH/IPL register is the pointer for the indirect
addressing mode, and the DPH/DPL and SPH/SPL
registers are the pointers for the indirect-with-offset
addressing modes.
4.1 Addressing Modes
A 9-bit field within the instruction, called the “fr” field,
specifies the addressing mode and the address (in the
case of direct addressing) or the address offset (in the
case of indirect-with-offset addressing), as shown in Table
4-1. (See Figure 3-7 for data RAM map.)
4.1.1 Pointer Registers
When an addition or increment instruction (i.e. add, inc,
incsz, or incsnz) on the low byte of a pointer register
(i.e. IPL, DPL, SPL, or ADDRL) generates a carry, the
high part of the register is incremented. For example, if the
IP register holds 0x00FF and an inc ipl instruction is
executed, the register will hold 0x0100 after the
instruction. When a subtraction or decrement instruction
(i.e. sub, subc, dec, decsz, or decsnz) generates a
borrow, the high part of the register is decremented.
Note: Because carry and borrow are automatically
handled, the addc and subc instructions are not needed
for arithmetic operations on pointer registers.
Table 4-1 Addressing Mode Summary
“fr” Field Mode Syntax Effective Address (EA) Restrictions
0 0000 0000 Indirect mov w,(ip)
mov (ip),w
IPH || IPL 0x020 < EA < 0xFFF
0 0nnn nnnn Direct, special-
purpose registers
mov w,fr
mov fr,w
nnnnnnn 0x002 < EA < 0x07F
0 1nnn nnnn Direct, global
registers
mov w,fr
mov fr,w
0x080 + nnnnnnn 0x080 < EA < 0x0FF
1 0nnn nnnn Indirect with offset,
data pointer
mov w,offset(dp)
mov offset(dp),w
DPH || DPL + nnnnnnn 0x000 < nnnnnnn < 0x07F
0x020 < EA < 0xFFF
1 1nnn nnnn Indirect with offset,
stack pointer
mov w,offset(sp)
mov offset(sp),w
SPH || SPL + nnnnnnn 0x000 < nnnnnnn < 0x07F
0x020 < EA < 0xFFF
36 www.ubicom.com
IP2012 / IP2022 Data Sheet
4.1.2 Direct Addressing Mode
Figure 4-1 shows the direct addressing mode used to
reference the special-purpose registers. Seven bits from
the “fr” field allow addressing up to 128 special-purpose
registers. (Not all 128 locations in this space are
implemented in the IP2012 / IP2022; several locations are
reserved for future expansion.)
Figure 4-1 Direct Mode, Special-Purpose Registers
The following code example uses direct mode.
Figure 4-2 shows the direct addressing mode used to
reference the global registers. This mode is distinguished
from the mode used to access the special-purpose
registers with bit 7 of the “fr” field. Because these registers
have this additional addressing mode not available for the
other data memory locations, they are especially useful
for holding global variables and frequently accessed data.
Figure 4-2 Direct Mode, Global Registers
Note: Addresses from 0x000 to 0x01F can only be
accessed with Direct mode.
4.1.3 Indirect Addressing Mode
The indirect addressing mode is used when all of the bits
in the “fr” field are clear. The location of the operand is
specified by a 12-bit pointer in the IPH and IPL registers.
The upper four bits of the IPH register are not used. Figure
4-3 shows indirect mode.
Figure 4-3 Indirect Mode
The following code example uses indirect mode.
mov w,0x0012 ;load W with the contents of
;the memory location at 0x0012
;(the DATAL register)
515-007a.eps
127
Special-Purpose
Registers
70
9-Bit "fr" Field
from Instruction
0
8
0nnnnnnn
0
76
515-008a.eps
127
Global Registers
70
9-Bit "fr" Field
from Instruction
0
8
1nnnnnnn
0
76
mov w,#0x03 ;load W with 0x03
mov iph,w ;load the high byte of the
;indirect pointer from W
mov w,#0x85 ;load W with 0x85
mov ipl,w ;load the low byte of the
;indirect pointer from W
mov w,(ip) ;load W with the contents of
;the memory location at
;effective address 0x0385
515-009a.eps
127
Special-Purpose
Registers
3840 Bytes
Data Memory
IPH Register IPL Register
70
n
0
n
7
nnnnnnnX
347
XXXnn n
0
9-Bit "fr" Field
from Instruction
0
8
00000000
0
128
Global Registers
IP2012 / IP2022 Data Sheet
www.ubicom.com 37
4.1.4 Indirect-with-Offset Addressing
Mode
The indirect-with-offset addressing mode is used when bit
8 of the “fr” field is set. The location of the operand is
specified by a 7-bit unsigned immediate from the “fr” field
added to a 12-bit base address in a pointer register.
When bit 7 of the “fr” field is clear, the DPH/DPL register
is selected as the pointer register. This register is
accessed using the loadh and loadl instructions,
which load its high and low bytes, respectively. The upper
four bits of the DPH register are not used. Figure 4-4
shows indirect-with-offset addressing using the DPH/DPL
register as the pointer register.
Figure 4-4 Indirect-with-Offset Mode, Data Pointer
The following code example uses indirect-with-offset
addressing mode.
When bit 7 of the “fr” field is set, the SPH/SPL register is
selected as the pointer register. The upper four bits of the
SPH register are not used. Figure 4-5 shows indirect-with-
offset mode using the SPH/SPL register. In addition to this
indirect-with-offset addressing mode, there are also
push and pop instructions which automatically
increment and decrement the SPH/SPL register while
performing a data transfer between the top of stack and a
data memory location specified by the “fr” field. Stacks
grow down from higher addresses to lower addresses.
This stack addressing mechanism is completely
independent from the hardware stack used for subroutine
call and return.
When a pop instruction is used with the indirect-with-
offset addressing mode, the address calculation for the
“fr” operand is made using the value in the SPH/SPL
register before the automatic increment, even though the
stack operand itself is addressed using the value after the
automatic increment.
Figure 4-5 Indirect-with-Offset Mode, Stack Pointer
MyStuff= 0x038D ;define address MyStuff
loadh MyStuff ;load the high byte of the
;DPH/DPL pointer register
;with 0x03
loadl MyStuff ;load the low byte of the
;DPH/DPL pointer register
;with 0x8D
515-026a.eps
127
Special-Purpose
Registers
DPH Register DPL Register
70
n
0
n
7
nnnnnnnX
347
XXXnn n
0
9-Bit "fr" Field
from Instruction
1
8
0mmmmmmm
0
+
76
3840 Bytes
Data Memory
128
Global Registers
mov w,8(dp) ;load W with the contents of
;the memory location at
;effective address 0x038D
;(i.e. 0x0385 + 0x0008)
515-027a.eps
127
Special-Purpose
Registers
SPH Register SPL Register
70
n
0
n
7
nnnnnnnX
347
XXXnn n
0
9-Bit "fr" Field
from Instruction
1
8
1mmmmmmm
0
+
76
3840 Bytes
Data Memory
128
Global Registers
38 www.ubicom.com
IP2012 / IP2022 Data Sheet
4.2 Instruction Set
The instruction set consists entirely of single-word (16-bit)
instructions, most of which can be executed at a rate of
one instruction per clock cycle, for a throughput of up to
120 MIPS when executing out of program RAM.
Assemblers may implement additional instruction
mnemonics for the convenience of programmers, such as
a long jump instruction which compiles to multiple IP2012
/ IP2022 instructions for handling the page structure of
program memory. Refer to the assembler documentation
for more information about any instruction mnemonics
implemented in the assembler.
4.2.1 Instruction Formats
There are five instruction formats:
•Two-operand arithmetic and logical instructions
•Immediate-operand arithmetic and logical instruc-
tions
•Jumps and subroutine calls
• Bit operations
• Miscellaneous instructions
Figure 4-6 shows the two-operand instruction format. The
two-operand instructions perform an arithmetic or logical
operation between the W register and a data memory
location specified by the “fr” field. The D bit indicates the
destination operand. When the D bit is clear, the
destination operand is the W register. When the D bit is
set, the destination operand is specified by the “fr” field.
There are some exceptions to this behavior. The multiply
instructions always load the 16-bit product into the MULH
and W registers. The MULH register receives the upper 8
bits, and the W register receives the lower 8 bits.
Traditionally single-operand instructions, such as
increment, are available in two forms distinguished by the
D bit. When the D bit is clear, the source operand is
specified by the “fr” field and the destination operand is the
W register. When the D bit is set, the data memory
location specified by the “fr” field is both the source and
destination operand.
Also, there are a few cases of unrelated instructions, such
as clr and cmp, which are distinguished by the D bit.
Figure 4-7 shows the immediate operand instruction
format. In this format, an 8-bit literal value is encoded in
the instruction field. Usually the W register is the
destination operand, however this format also includes
instructions that use the top of the stack or a special-
purpose register as the destination operand.
Figure 4-8 shows the format of the jump and subroutine
call instructions. 13 bits of the entry point address are
encoded in the instruction. The remaining three bits come
from the PA2:0 bits of the STATUS register.
Figure 4-9 shows the format of the instructions that clear,
set, and test individual bits within registers. The register is
specified by the “fr” field, and a 3-bit field in the instruction
selects one of the eight bits in the register.
Figure 4-10 shows the format of the remaining
instructions.
4.2.2 Instruction Types
The instructions are grouped into the following functional
categories:
• Logical instructions
•Arithmetic and shift instructions
• Bit operation instructions
• Data movement instructions
• Program control instructions
• System control instructions
Logical Instructions
Each logic instruction performs a standard logical
operation (AND, OR, exclusive OR, or logical
complement) on the respective bits of the 8-bit operands.
The result of the logic operation is written to W or to the
data memory location specified by the “fr” field.
All of these instructions take one clock cycle for execution.
Arithmetic and Shift Instructions
Each arithmetic or shift instruction performs an operation
such as add, subtract, add with carry, subtract with carry,
15 10 9 8 0
Opcode D “fr” Field
Figure 4-6 Two-Operand Instruction Format
15 8 7 0
Opcode 8-Bit Literal (“#lit8”)
Figure 4-7 Immediate-Operand Instruction Format
15 13 12 0
Opcode Entry Point Address (“addr13”)
Figure 4-8 Jump and Call Instruction Format
15 12 11 9 8 0
Opcode Bit “fr” Field
Figure 4-9 Bit Operation Instruction Format
15 14 13 12 11 10 9 8 0
0000000 Opcode
Figure 4-10 Miscellaneous Instruction Format
IP2012 / IP2022 Data Sheet
www.ubicom.com 39
rotate left or right through carry, increment, decrement,
clear to zero, or swap high/low nibbles. The compare
(cmp) instruction performs the same operation as
subtract, but it only updates the C, DC, and Z flags of the
STATUS register; the result of the subtraction is
discarded.
There are instructions available (incsz, decsz) that
increment or decrement a register and simultaneously
test the result. If the 8-bit result is zero, the next instruction
in the program is skipped. These instructions can be used
to make program loops. There are also compare-and-skip
instructions (cse, csne) which perform the same
operation as subtract, but perform a conditional skip
without affecting either operand or the condition flags in
the STATUS register.
All of the arithmetic and shift instructions take one clock
cycle for execution, except in the case of the test-and-skip
instructions when the tested condition is true and a skip
occurs, in which case the instruction takes at least two
cycles. If a skip instruction is immediately followed by a
loadh, loadl, or page instruction (and the tested
condition is true) then two instructions are skipped and the
operation consumes three cycles. This is useful for
skipping over a conditional branch to another page, in
which a page instruction precedes a jmp instruction. If
several page or loadh/loadl instructions immediately
follow a skip instruction, then they are all skipped plus the
next instruction and a cycle is consumed for each. These
“extended skip instructions” are interruptible, so they do
not affect interrupt latency.
Bit Operation Instructions
There are four bit operation instructions:
•setb—sets a single bit in a data register without af-
fecting other bits
•clrb—clears a single bit in a data register without af-
fecting other bits
•sb—tests a single bit in a data register and skips the
next instruction if the bit is set
•snb—tests a single bit in a data register and skips the
next instruction if the bit is clear
All of the bit operation instructions take one clock cycle for
execution, except for test-and-skip instructions when the
tested condition is true and a skip occurs.
Data Movement Instructions
A data movement instruction moves a byte of data from a
data memory location to either the W register or the top of
stack, or it moves the byte from either the W register or the
top of stack to a data memory location. The location is
specified by the “fr” field. The SPH/SPL register pair points
to the top of stack. This stack is independent of the
hardware stack used for subroutine call and return.
Program Control Instructions
A program control instruction alters the flow of the
program by changing the contents of the program counter.
Included in this category are the jump, call, return-from-
subroutine, and interrupt instructions.
The jmp instruction has a single operand that specifies
the entry point at which to continue execution. The entry
point is typically specified in assembly language with a
label, as in the following code example:
If the carry bit is set to 1, the jmp instruction is executed
and program execution continues where the do_carry
label appears in the program.
The call instruction works in a similar manner, except
that it saves the contents of the program counter to the
CALLH/CALLL registers before jumping to the new
address. It calls a subroutine that is terminated by a ret,
retw, or retnp instruction, as shown in the following
code example:
Returning from a subroutine restores the saved program
counter contents from the CALLH/CALLL registers, which
causes program to resume execution with the instruction
immediately following the call instruction (a nop
instruction, in the above example)
A program memory address contains 16 bits. The jmp
and call instructions specify only the lowest thirteen bits
of the jump/call address. The upper 3 bits come from the
PA2:0 bits of the STATUS register. An indirect relative
jump can be accomplished by adding the contents of the
W register to the PCL register (i.e. an add pcl,w
instruction).
Program control instructions such as jmp, call, and
ret alter the normal program sequence. When one of
snb status,0 ;test the carry bit
jmp do_carry ;jump to do_carry routine
;if C = 1
...
do_carry: ;jump destination label
... ;execution continues here
call add_2bytes ;call subroutine
;add_2bytes
nop ;execution returns to
;here after the
;subroutine is finished
...
add_2bytes: ;subroutine label
... ;subroutine code goes
;here
ret ;return from subroutine
40 www.ubicom.com
IP2012 / IP2022 Data Sheet
these instructions is executed, the execution pipeline is
automatically cleared of pending instructions and refilled
with new instructions, starting at the new program
address. Because the pipeline must be cleared, three
clock cycles are required for execution, one to execute the
instruction and two to reload the pipeline.
System Control Instructions
A system control instruction performs a special-purpose
operation that sets the operating mode of the device or
reads data from the program memory. Included in this
category are the following types of instructions:
•speed—changes the CPU core speed (for saving
power)
•break—enters debug mode
•page—writes to the PA2:0 bits in the STATUS regis-
ter
•loadh/loadl—loads a 16-bit pointer into the DPH
and DPL registers
•iread—reads a word from external memory, pro-
gram flash memory, or program RAM
•ireadi—reads a word (and auto-increments ADDR
by 2) from program flash memory, or program RAM
•iwrite—writes a word to external memory or pro-
gram RAM
•iwritei—writes a word (and auto-increments
ADDR by 2) to program RAM
•fread—reads a word from flash program memory
•fwrite—writes a word to flash program memory
•ferase—erases a block of flash program memory
•cwdt—clears the Watchdog Timer
4.3 Instruction Pipeline
An instruction goes through a four-stage pipeline to be
executed, as shown in Figure 4-11. The first instruction is
fetched from the program memory on the first core clock
cycle. On the second clock cycle, the first instruction is
decoded and a second instruction is fetched. On the third
clock cycle, the first instruction is executed, the second
instruction is decoded, and a third instruction is fetched.
On the fourth clock cycle, the first instruction’s results are
written to its destination, the second instruction is
executed, the third instruction is decoded, and a fourth
instruction is fetched. Once the pipeline is full, instructions
are executed at the rate of one per clock cycle.
Instructions that directly affect the contents of the program
counter (such as jumps and calls) require that the pipeline
be cleared and subsequently refilled. Therefore, these
instructions take two additional clock cycles (the PC will
be changed during the execute cycle of a jump
instruction).
Stage Core
Cycle 1
Core
Cycle 2
Core
Cycle 3
Core
Cycle 4
Fetch Instruction 1 Instruction 2 Instruction 3 Instruction 4
Decode Instruction 1 Instruction 2 Instruction 3
Execute Instruction 1 Instruction 2
Write Instruction 1
Figure 4-11 Pipeline Execution
IP2012 / IP2022 Data Sheet
www.ubicom.com 41
4.4 Subroutine Call/Return Stack
A 16-level hardware call/return stack is provided for
saving the program counter on a subroutine call and
restoring the program counter on subroutine return. The
stack is not mapped into the data memory address space
except for the top level, which is accessible as the CALLH
and CALLL registers. Software can read and write these
registers to implement a deeper stack, in those cases
which require nesting subroutines more than 16 levels
deep. This stack is completely independent of the stack
used with the push and pop instructions and the
SPH/SPL register pair.
Note: The CALLL and CALLH registers require special
attention as modification of these values (the top of the
stack) changes the return vector.
When a subroutine is called, the return address is pushed
onto the subroutine stack, as shown in Figure 4-12.
Specifically, each saved address in the stack is moved to
the next lower level to make room for the new address to
be saved. Stack 1 receives the contents of the program
counter. Stack 16 is overwritten with what was in Stack 15.
The contents of stack 16 are lost.
Figure 4-12 Stack Operation on Subroutine Call
When a return instruction is executed the subroutine stack
is popped, as shown in Figure 4-13. Specifically, the
contents of Stack 1 are copied into the program counter
and the contents of each stack level are moved to the next
higher level. When a value is popped off the stack, the
bottom entry is initialized to 0xFFFF. For example, Stack 1
receives the contents of Stack 2, etc., and Stack 15 is
overwritten with the contents of Stack 16. Stack 16 is
initialized to 0xFFFF.
Figure 4-13 Stack Operation on Subroutine Return
For program bugs involving stack underflow, the
instruction at byte address 0x1FFFE (word address
0xFFFF) can be used to jump to an appropriate handler.
For example, system recovery may be possible by
jumping to the reset vector at byte address 0x1FFE0
(word address 0xFFF0).
The options for returning from a CALL are:
1. RET - The stack will be popped (CALLH/L will be load-
ed into PCH/L) and the page bits (PA2:0 in the STA-
TUS register) will be loaded with the upper 3 bits of
CALLH.
2. RETNP - Same as above, but PA2:0 are not changed.
3. RETW #lit - Same as RET, but also moves literal to W.
515-010a.eps
Stack 1
Stack 2
Program Counter (15:0)
Stack 16 Contents
are Discarded
Stack 3
Stack 4
Stack 5
Stack 6
Stack 7
Stack 8
Stack 9
Stack 10
Stack 11
Stack 12
Stack 13
Stack 14
Stack 15
Stack 16
CALLH/CALLL
!!!!
"
#
$
%
&
"
#
42 www.ubicom.com
IP2012 / IP2022 Data Sheet
4.5 Key to Abbreviations and Symbols
4.6 Instruction Set Summary Tables
Table 4-2 through Table 4-7 list all of the IP2012 / IP2022
instructions, organized by category. For each instruction,
the table shows the instruction mnemonic (as written in
assembly language), a brief description of what the
instruction does, the number of instruction cycles required
for execution, the binary opcode, and the flags in the
STATUS register affected by the instruction.
Although the number of clock cycles for execution is
typically 1, for the skip instructions the exact number of
cycles depends whether the skip is taken or not taken.
Taking the skip adds 1 cycle. The effect of extended skip
instructions (i.e. a skip followed by a loadh, loadl, or
page instruction) is not shown.
For more detailed description, refer to the Programmer’s
Reference Manual.
Symbol Description
addr13 13-bit address in assembly language instruc-
tion
addr16 16-bit address in assembly language instruc-
tion
bit Bit position selector bit in opcode
BO Brown-out bit in the PSPCFG register (bit 0)
C Carry bit in the STATUS register (bit 0)
DC Digit Carry bit in the STATUS register (bit 1)
DPH Upper half of data pointer for indirect-with-off-
set addressing (global file register 0x00C)
DPL Lower half of data pointer for indirect-with-off-
set addressing (global file register 0x00D)
f File register address bit in opcode
fr File register field (a 9-bit file register address
specified in the instruction)
IPH Indirect Pointer High - Upper half of pointer for
indirect addressing (global file register 0x004)
IPL Indirect Pointer Low - Lower half of pointer for
indirect addressing (global file register 0x005)
k Constant value bit in opcode
n Numerical value bit in opcode
PA2:PA0 Page bits in the STATUS register (bits 7:5)
PCL Virtual register for direct PC modification (glo-
bal file register 0x009)
SPH Upper half of stack pointer for indirect-with-off-
set addressing (global file register 0x006)
SPL Lower half of stack pointer for indirect-with-off-
set addressing (global file register 0x007)
STATUS STATUS register (global file register 0x00B)
WWorking register
WD Watchdog Timeout bit in the PSPCFG register
(bit 1)
WDT Watchdog Timer counter and prescaler
Z Zero bit in the STATUS register (bit 2
, File register/bit selector separator
(e.g. clrb status,z)
!= inequality
# Immediate literal designator in assembly lan-
guage instruction (e.g. mov w,#0xff)
#lit8 8-bit literal value in assembly language
instruction
& Logical AND
(address)
Contents of memory referenced by address
^ Logical exclusive OR
| Logical OR
|| Concatenation
Symbol Description
IP2012 / IP2022 Data Sheet
www.ubicom.com 43
Table 4-2 Logical Instructions
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
and fr,w fr = fr & W AND fr,W into fr 1 0001 011f ffff ffff Z
and w,fr W = W & fr AND W,fr into W 1 0001 010f ffff ffff Z
and w,#lit8 W = W & lit8 AND W,literal into W 1 0111 1110 kkkk kkkk Z
not fr fr = fr Complement fr into fr 1 0010 011f ffff ffff Z
not w,fr W = fr Complement fr into W 1 0010 010f ffff ffff Z
or fr,w fr = fr | W OR fr,W into fr 1 0001 001f ffff ffff Z
or w,fr W = W | fr OR W,fr into W 1 0001 000f ffff ffff Z
or w,#lit8 W = W | lit8 OR W,literal into W 1 0111 1101 kkkk kkkk Z
xor fr,w fr = fr ^ W XOR fr,W into fr 1 0001 101f ffff ffff Z
xor w,fr W = W ^ fr XOR W,fr into W 1 0001 100f ffff ffff Z
xor w,#lit8 W = W ^ lit8 XOR W,literal into W 1 0111 1111 kkkk kkkk Z
Table 4-3 Arithmetic and Shift Instructions
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
add fr,w fr = fr + W Add fr,W into fr 1 0001 111f ffff ffff C, DC, Z
add w,fr W = W + fr Add W,fr into W 1 0001 110f ffff ffff C, DC, Z
add w,#lit8 W = W + lit8 Add W,literal into W 1 0111 1011 kkkk kkkk C, DC, Z
addc fr,w fr = C + fr + W Add carry,fr,W into fr 1 0101 111f ffff ffff C, DC, Z
addc w,fr W = C + W + fr Add carry,W,fr into W 1 0101 110f ffff ffff C, DC, Z
clr fr fr = 0 Clear fr 1 0000 011f ffff ffff Z
cmp w,fr fr - W Compare W,fr
then update STATUS
10000 010f ffff ffff C, DC, Z
cmp w,#lit8 lit8 - W Compare W,literal
then update STATUS
10111 1001 kkkk kkkk C, DC, Z
cse w,fr if (fr - W) = 0
then skip
Compare W,fr then skip if equal 1 or
2 (skip)
0100 001f ffff ffff None
cse w,#lit8 if (lit8 - W) = 0
then skip
Compare W,literal then skip if
equal
1 or
2 (skip)
0111 0111 kkkk kkkk None
csne w,fr if (fr - W) != 0
then skip
Compare W,fr then skip if not
equal
1 or
2 (skip)
0100 000f ffff ffff None
csne w,#lit8 if (lit8 - W) != 0
then skip
Compare W,literal then skip if not
equal
1 or
2 (skip)
0111 0110 kkkk kkkk None
cwdt WDT = 0 Clear Watchdog Timer 1 0000 0000 0000 0100 None
dec fr fr = fr - 1 Decrement fr into fr 1 0000 111f ffff ffff Z
dec w,fr W = fr -1 Decrement fr into W 1 0000 110f ffff ffff Z
decsnz fr fr = fr - 1
if fr != 0 then skip
Decrement fr into fr then skip if
not zero (STATUS not updated)
1 or
2 (skip)
0100 111f ffff ffff None
44 www.ubicom.com
IP2012 / IP2022 Data Sheet
decsnz w,fr W = fr - 1
if fr != 0 then skip
Decrement fr into W then skip if
not zero (STATUS not updated)
1 or
2 (skip)
0100 110f ffff ffff None
decsz fr fr = fr - 1
if fr = 0 then skip
Decrement fr into fr then skip if
zero (STATUS not updated)
1 or
2 (skip)
0010 111f ffff ffff None
decsz w,fr W = fr - 1
if fr = 0 then skip
Decrement fr into W then skip if
zero (STATUS not updated)
1 or
2 (skip)
0010 110f ffff ffff None
inc fr fr = fr + 1 Increment fr into fr 1 0010 101f ffff ffff Z
inc w,fr W = fr + 1 Increment fr into W 1 0010 100f ffff ffff Z
incsnz fr fr = fr + 1
if fr != 0 then skip
Increment fr into fr then skip if
not zero (STATUS not updated)
1 or
2 (skip)
0101 101f ffff ffff None
incsnz w,fr W = fr + 1
if fr != 0 then skip
Increment fr into W then skip if
not zero (STATUS not updated)
1 or
2 (skip)
0101 100f ffff ffff None
incsz fr fr = fr + 1
if fr = 0 then skip
Increment fr into fr then skip if
zero (STATUS not updated)
1 or
2 (skip)
0011 111f ffff ffff None
incsz w,fr W = fr + 1
if fr = 0 then skip
Increment fr into W then skip if
zero (STATUS not updated)
1 or
2 (skip)
0011 110f ffff ffff None
muls w,fr MULH || W = W ∞ fr Signed 8 ∞ 8 multiply (bit 7 =
sign); W x fr into MULH || W (bit 7
of MULH is result sign)
10101 010f ffff ffff None
muls w,#lit8
MULH || W = W ∞ lit8
Signed 8 ∞ 8 multiply (bit 7 =
sign); W x literal into MULH || W
(bit 7 of MULH is result sign)
10111 0011 kkkk kkkk None
mulu w,fr MULH || W = W ∞ fr Unsigned 8 ∞ 8 multiply; W x fr
into MULH || W
10101 000f ffff ffff None
mulu w,#lit8
MULH || W = W ∞ lit8
Unsigned 8 ∞ 8 multiply; W x lit-
eral into MULH || W
10111 0010 kkkk kkkk None
rl fr fr || C = C || fr Rotate fr left through carry into fr 1 0011 011f ffff ffff C
rl w,fr W || C = C || fr Rotate fr left through carry into W 1 0011 010f ffff ffff C
rr fr C || fr = fr || C Rotate fr right through carry into
fr
1 0011 001f ffff ffff C
rr w,fr C || W = fr || C Rotate fr right through carry into
W
1 0011 000f ffff ffff C
sub fr,w fr = fr - W Subtract W from fr into fr 1 0000 101f ffff ffff C, DC, Z
sub w,fr W = fr - W Subtract W from fr into W 1 0000 100f ffff ffff C, DC, Z
sub w,#lit8 W = lit8 - W Subtract W from literal into W 1 0111 1010 kkkk kkkk C, DC, Z
subc fr,w fr = fr - C - W Subtract carry,W from fr into fr 1 0100 101f ffff ffff C, DC, Z
subc w,fr W = fr - C - W
Subtract
carry,
W from fr into W
10100 100f ffff ffff C, DC, Z
Table 4-3 Arithmetic and Shift Instructions (continued)
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
IP2012 / IP2022 Data Sheet
www.ubicom.com 45
swap fr fr = fr3:0 || fr7:4 Swap high,low nibbles of fr into fr 1 0011 101f ffff ffff None
swap w,fr W = fr3:0 || fr7:4 Swap high,low nibbles of fr into
W
10011 100f ffff ffff None
test fr if fr = 0 then Z = 1
else Z = 0
Test fr for zero and update Z 1 0010 001f ffff ffff Z
Table 4-3 Arithmetic and Shift Instructions (continued)
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
Table 4-4 Bit Operation Instructions
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
clrb fr,bit fr,bit = 0 Clear bit in fr 1 1000 bbbf ffff ffff None
sb fr,bit if fr,bit = 1 then skip Test bit in fr then skip if set 1 or
2 (skip)
1011 bbbf ffff ffff None
setb fr,bit fr,bit = 1 Set bit in fr 1 1001 bbbf ffff ffff None
snb fr,bit if fr,bit = 0 then skip Test bit in fr then skip if clear 1 or
2 (skip)
1010 bbbf ffff ffff None
Table 4-5 Data Movement Instructions
Assembler
Syntax
Pseudocode
Definition Description Core
Cycles Opcode Flags
Affected
mov fr,w fr = W Move W into fr 1 0000 001f ffff ffff None
mov w,fr W = fr Move fr into W 1 0010 000f ffff ffff Z
mov w,#lit8 W = lit8 Move literal into W 1 0111 1100 kkkk kkkk None
push fr (SP) = fr, then
SP = SP - 1
Move fr onto top of stack 1 0100 010f ffff ffff None
push #lit8 (SP) = lit8, then
SP = SP - 1
Move literal onto top of stack 1 0111 0100 kkkk kkkk None
pop fr fr = (SP + 1), then
SP = SP + 1
Move top of stack + 1 into fr 1 0100 011f ffff ffff None
46 www.ubicom.com
IP2012 / IP2022 Data Sheet
Table 4-6 Program Control Instructions
Assembler
Syntax Description Core
Cycles Opcode Flags
Affected
call addr13 Call subroutine 3 110k kkkk kkkk kkkk None
jmp addr13 Jump 3 111k kkkk kkkk kkkk None
int Software interrupt 3 0000 0000 0000 0110 None
nop No operation 1 0000 0000 0000 0000 None
ret Return from subroutine 3 0000 0000 0000 0111 PA2:0
retnp Return from subroutine, without updating
page bits
30000 0000 0000 0010 None
reti #lit3 Return from interrupt (see Section 3.7.4) 3 0000 0000 0000 1nnn All
retw #lit8 Return from subroutine with literal into W 3 0111 1000 kkkk kkkk PA2:0
Table 4-7 System Control Instructions
Assembler
Syntax Description Core Cycles Opcode Flags
Affected
break Software breakpoint. Keeps PC from
advancing and stops timers, including the
Watchdog Timer
10000 0000 0000 0001 None
breakx Software breakpoint, extending the skip 1 0000 0000 0000 0101 None
ferase Erase a 256 word flash block 1†0000 0000 0000 0011 None
fread Read flash memory 1†0000 0000 0001 1011 None
fwrite Write flash memory 1†0000 0000 0001 1010 None
iread Read external/program memory 4(blocking),
1†(nonblocking)
0000 0000 0001 1001 None
ireadi Read program memory and increment
ADDRL to next even ADDRL
4(blocking),
1†(nonblocking)
0000 0000 0001 1101 None
iwrite Write into external memory/program RAM 4(blocking),
1†(nonblocking)
0000 0000 0001 1000 None
iwritei Write into program RAM and increment
ADDRL to next even ADDRL
4(blocking),
1†(nonblocking)
0000 0000 0001 1100 None
loadh addr8 Load high data address into DPH 1 0111 0000 kkkk kkkk None
loadl addr8 Load low data address into DPL 1 0111 0001 kkkk kkkk None
page addr3 Load page bits from program address into
PA2:0 of the STATUS register
10000 0000 0001 0nnn PA2:0
speed #lit8 Change CPU speed by writing into the
SPDREG register
10000 0001 nnnn nnnn None
†. Only occupies the CPU pipeline for 1 cycle, but the operation is not complete until XCFG:0 = 0. (Refer to Sec-
tion 4.7)
IP2012 / IP2022 Data Sheet
www.ubicom.com 47
4.7 Program Memory Self-Programming and Read Instructions
The IP2012 / IP2022 has several instructions used to read
and write the program RAM and the program flash
memory. These instructions allow the program flash
memory to be read and written through special-purpose
registers in the data memory space, which allows the
flash memory to be used to store both program code and
data.
Because no special programming voltage is required to
write to the flash memory, any application may take
advantage of this feature at run-time. Typical uses include
saving phone numbers and passwords, downloading new
or updated software, and logging infrequent events such
as errors and Watchdog Timer overflow.
The self-programming instructions are not affected by the
code-protection flag (the CP bit of the FUSE1 register), so
the entire program memory is readable and writable by
any software running on the IP2012 / IP2022.
Note: It is highly recommended to enable the brown-out
reset feature if self-programming instructions are being
used in user program code (see Section 3.8.1 for more
information about BOR). This will avoid corruption of flash
memory during power down.
There are seven instructions used for self-programming,
as shown in Table 4-8. Certain uses of the instructions are
not valid. In these cases, the instruction is executed as
though it were a nop instruction (i.e. the program counter
is incremented, but no other registers or bits are affected).
Blocking instructions take 4 cycles to complete, and
prevent other instructions from executing. Non-blocking
instructions occupy the CPU pipeline for only one cycle,
but they launch a multi-cycle operation which is not
complete until indicated by the FBUSY bit in the XCFG
register becoming clear.
The DATAH/DATAL register is a 16-bit data buffer used for
loading or unloading data in program memory. The
ADDRX/ADDRH/ADDRL register holds a 24-bit byte
address used to specify the low-byte of the desired word
location in program memory. Like the other pointer
registers (IPH/IPL, DPH/DPL, and SPH/SPL), addition to
the low byte of the register that results in carry will cause
the high part of the register (ADDRX/ADDRH) to be
incremented. Subtraction from the low byte of the register
that results in borrow will cause the high part of the
register to be decremented.
Note: If ADDRSEL is modified in the ISR, it must first be
shadowed in software, and restored before reti.
Note: ADDRL bit 0 is ignored as the A0 address bit is
handled automatically in hardware.
Software should use the FBUSY bit to check that a
previous flash write or erase operation has completed
before executing another instruction that accesses flash
memory, before jumping to or calling program code in
flash memory, and before changing the CPU core speed.
It is not necessary to check the FBUSY bit if enough
cycles are allowed for the flash operation to complete. See
description of FRDTS1:0, FRDTC1:0 and FWRT3:0 in
Section 7.1.5 for more details. Software must not attempt
to execute out of flash memory while the FBUSY bit is set,
because the flash memory is unreadable during that time.
Therefore, code which reads, writes, or erases flash
memory, using the fread, fwrite or ferase
instructions, must execute from program RAM. Software
must provide at least four cycles between an fread and
Table 4-8 Instructions Used for Self-Programming
Operation Program RAM
(ADDRX = 00)
Flash
(ADDRX =01)
External Memory
(ADDRX = 80 or 81)
Read iread (Blocking)
ireadi (Blocking)
fread (Nonblocking) 1
iread2
ireadi2
iread (Nonblocking)
Write iwrite (Blocking)
iwritei (Blocking)
fwrite (Nonblocking) 3 iwrite (Nonblocking)
Erase N/A ferase (Nonblocking) 3 N/A
1 — Rules 1, 2, 3, 5, 7, and 10 below apply.
2 — Rules 2, 3, 5, 7, and 10 below apply. If executed from program RAM, the instruction is nonblocking; if
executed from flash, it is blocking.
3 — Rules 1, 2, and 4–10 below apply.
48 www.ubicom.com
IP2012 / IP2022 Data Sheet
reading DATAH/DATAL, or ensure that the minimum flash
read time is met.
Unlike RAM, flash memory requires an explicit erase
operation before being written. The ferase instruction is
used to erase a 512-byte (256-word) block of flash
memory (it brings all bits to 1, see Table 4-9). After the
block has been erased, individual words can be written
with the fwrite instruction (fwrite will not change a
0 to 1). For example, an ferase instruction executed on
any byte address from 0x10000 to 0x100FE erases the
whole block spanning those addresses. The self-
programming instructions have no access to the flash
memory bits in the configuration block.
Rules/Troubleshooting for fread/fwrite/ferase
and iread/ireadi of flash:
1. Must be executing out of program RAM, with ADDRX
= 01.
2. FCFG register must be correctly configured (refer to
Section 7.1.5).
3. For an fread or iread/ireadi of flash, there
must be at least 3 core cycles between the read in-
struction and a read of DATAH or DATAL, or the mini-
mum flash read time must be met.
4. No speed commands while fread, fwrite or fe-
rase are busy (while XCFG bit 0 = 1).
5. Do not jump to flash memory while executing
fread/fwrite/ferase. If INTVEC is in flash, en-
sure interrupts are disabled.
6. fwrite will not change a 0 to a 1 (use ferase
first).
7. XCFG bit 0 = 0 before execution (even during ISR) or
sufficient time is allowed to complete previous opera-
tions on flash.
8. XCFG bit 6 = 1, otherwise fwrite and ferase be-
have as nop.
9. Make sure interrupts are disabled or that the INTSPD
value matches the SPDREG value. For iread or
ireadi of flash from flash, a more practical solution
is to jump to a routine in program RAM.
10. Wait 1 cycle after changing ADDRX bit 7, EMCFG bit
7, or ADDRSEL before executing an fread,
fwrite, ferase, iread, ireadi, iwrite, or
iwritei instruction.
11. Do not write to DATAH at the same time an iread of
External Memory is causing a write of DATAL.
4.7.1 Flash Timing Control
The FCFG register controls the timing of flash memory
operations. See Section 7.1.5 for a description of the
FCFG register.
4.7.2 Interrupts During Flash Operations
Before starting a flash write or erase operation, the FCFG
register (see Section 7.1.5) must be set up properly for the
current speed. The CPU core clock is the time base for the
flash write timing compensation, so it is critical that the
CPU core clock speed is not changed during a flash write
or erase operation. Interrupts may be taken during a flash
write or erase operation, if the INTSPD register is set up
so the speed does not change when an interrupt occurs.
If the flash read timing compensation is set up for a clock
divisor of 1 (i.e. fastest speed), interrupts will not cause
fread/iread instructions to fail, so no special
precautions need to be taken to avoid violating the flash
read access time.
Table 4-9 ferase Addresses (ADDRX=01,
ADDRL=xx)
ADDRH Flash Byte
Addresses
0x00 0x10000 - 0x101FE
0x02 0x10200 - 0x102FE
... ...
0xFE 0x1FE00 - 0x1FFFE
IP2012 / IP2022 Data Sheet
www.ubicom.com 49
5.0 Peripherals
The IP2012 / IP2022 provides an array of on-chip
peripherals needed to support a broad range of
embedded Internet applications:
•2 Serializer/Deserializer (SerDes) units (IP2012 has
one unit)
• Real-time timer
• T0 timer
•2 General-purpose timers with compare and capture
Registers
•Watchdog timer
• 10-bit, 8-channel A/D converter
• Analog comparator
•Parallel slave peripheral interface
• External memory interface (IP2022 only)
All of the peripherals except the Watchdog Timer and the
Real-Time Timer use alternate functions of the I/O port
pins to interface with external signals.
5.1 I/O Ports
The IP2022 contains one 4-bit I/O port (Port A) and six 8-
bit I/O ports (Port B through Port G). The IP2012 contains
one 4-bit I/O port (Port A), two 6-bit I/O ports (Ports D and
F) and four 8-bit I/O ports (Ports B, C, E, and G). The four
Port A pins have 24 mA current drive capability. All the
ports have symmetrical drive. Inputs are 5V-tolerant.
Outputs can use the same 2.3–2.7V power supply used
for the CPU core and peripheral logic, or they can use a
higher voltage (up to 3.6V). The IOVdd pins are provided
for the I/O port pin output drivers. Port G has a separate
GVdd pin which can be used to run the Port G output
drivers at a voltage different from that used for the other
ports, since Port G must run from a 2.3–2.7V power
supply.
Each port has separate input (RxIN), output (RxOUT),
and direction (RxDIR) registers, which are memory
mapped. The numbers in the pin names correspond to the
bit positions in these registers. These registers allow each
port bit to be individually configured as a general-purpose
input or output under software control. Unused pins
should be configured as outputs, to prevent them from
floating. Port B has three additional registers for
supporting external interrupts (see Section 5.1.1).
Each port pin has an alternate function used to support
the on-chip hardware peripherals, as listed in Table 2-1.
Port A and Port B support the multi-function timers Timer
1 and Timer 2. On the IP2022, Port B, Port C, and Port D
support the Parallel Slave Peripheral (PSP) and external
memory functions. On the IP2012, Port B and Port C
support the PSP. Port E and Port F support the
serializer/deserializer (SERDES) units (only Port E on the
IP2012). Port G supports the analog to digital converter
(ADC) and the analog comparator. Before enabling a
hardware peripheral, configure the port pins for input or
output as required by the peripheral.
Note: There is positive-feedback circuitry present on the
I/O ports when configured as input. This causes an input
that was previously high, then subsequently tri-stated (i.e.
not driven), to be actively driven by the IP2012 / IP2022 to
a voltage level of approximately 1.7V, or one diode drop
below (IOVdd). See Section 8.2 for details.
Figure 5-1 shows the internal hardware structure and
configuration registers for each pin of a port.
Figure 5-1 Port Pin Block Diagram
5.1.1 Port B Interrupts
Any of the 8 Port B pins can be configured as an external
interrupt input. Logic on these inputs can be programmed
to sense rising or falling edges. When an edge is detected,
the interrupt flag for the port pin is set.
The recommended initialization sequence is:
1. Configure the port pins used for interrupts as inputs
by programming the RBDIR register.
2. Be sure all enabled interrupt pins are driven to valid
logic levels, not floating.
3. Select the desired edge for triggering the interrupt
by programming the INTED register. This may set
interrupt flags.
4. nop, nop.
5. Clear the interrupt flags in the INTF register.
6. nop
515-030a.eps
Data Bus
Port Pin
RxDIR
Register
0 = Output
1 = Hi-Z Input
RxOUT
Register
RxIN
Register
M
SYNC bit set
U
X
1
0
SYNC bit
(FUSE1)
FF FF
Core Clock
SYNC bit clear
50 www.ubicom.com
IP2012 / IP2022 Data Sheet
7. Enable the interrupt input(s) by setting the corre-
sponding bit(s) in the INTE register.
8. Set the GIE bit.
Figure 5-2 shows the Port B interrupt logic. Port B has
three registers for supporting external interrupts, the
INTED (Section 5.1.6), INTF (Section 5.1.7), and INTE
(Section 5.1.8) registers. The INTED register controls the
logic which selects the edge sensitivity (i.e. rising or falling
edge) of the Port B pins. When an edge of the selected
type occurs, the corresponding flag in the INTF register is
set, whether or not the interrupt is enabled. The interrupt
signal passed to the system interrupt logic is the OR
function of the AND of each interrupt flag in the INTF
register with its corresponding enable bit in the INTE
register. See Section 5.1.8.
5.1.2 Reading and Writing the Ports
The port registers are memory-mapped into the data
memory address space between 0x020 and 0x03A. In
addition, Port B has three extra registers located at 0x017
through 0x019 (INTED, INTF, and INTE), which support
external interrupt inputs.
Generally, successive read and write operations on the
same I/O port is not an issue, as there are separate IN an
OUT registers for each I/O port. Care must be given to
ensure that enough time is allowed for data written to the
OUT register to propagate to the IN register on a given
port. If this is an issue, two instructions (or four
instructions if the SYNC bit in the FUSE1 register is clear)
should be inserted between any read-modify-write
instruction sequences (or more nop instructions if the pin
is capacitively loaded).
5.1.3 RxIN Registers
The RxIN registers are virtual registers that provide read-
only access to the physical I/O pins. Reading these
registers returns the states on the pins, which may be
driven either by the IP2012 / IP2022 or an external device.
If the SYNC bit in the FUSE1 register is clear, the states
are read from a synchronization register. If an application
reads data from a device running asynchronously to the
IP2012 / IP2022, the SYNC bit should be cleared to avoid
the occurrence of metastable states (i.e. corrupt data
caused by an input which fails to meet the setup time
before the sampling clock edge, which theoretically could
interfere with the operation of the CPU).
5.1.4 RxOUT Registers
The RxOUT registers are data output buffer registers. The
data in these registers is driven on any I/O pins that are
configured as outputs. On reads, the RxOUT registers
return the data previously written to the data output buffer
registers, which might not correspond to the states
Figure 5-2 Port B Interrupt Logic
IP2012 / IP2022 Data Sheet
www.ubicom.com 51
actually present on pins configured as inputs or pins
forced to another state by an external device.
5.1.5 RxDIR Registers
The RxDIR registers select the direction of the port pins.
For each output port pin, clear the corresponding RxDIR
bit. For each input port pin, set the corresponding RxDIR
bit. Unused pins that are left open-circuit should be
configured as outputs, to keep them from floating.
For example, to configure Port A pins RA3 and RA2 as
outputs and RA1 and RA0 as inputs, the following code
could be used:
The second move instruction in this example writes the
RADIR register, located at address 0x022. Because Port
A has only four I/O pins, only the four least significant bits
of this register are used.
To drive the RA1 pin low and the RA0 pin high, the
following code then could be executed:
The second move instruction shown above writes the
RAOUT register, located at address 0x021. When reading
the Port A pins through the RAIN register (0x020), the
upper four bits always read as zero.
When a write is performed to the RxOUT register of a port
pin that has been configured as an input, the write is
performed but it has no immediate effect on the pin. If that
pin is later configured as an output, the pin will be driven
with the data that had been previously written to the
RxOUT register.
5.1.6 INTED Register
The INTED register consists of 8 edge detection bits that
correspond to the 8 pins of Port B. A set bit in the INTED
register makes the corresponding port pin trigger on
falling edges, while a clear bit makes the pin trigger on
rising edges.
5.1.7 INTF Register
The INTF register consists of 8 interrupt flags that
correspond to the 8 pins of Port B. If the trigger condition
for a Port B pin occurs, the corresponding bit in the INTF
register is set. The bit is set even if the port pin is not
enabled as a source of interrupts.
The interrupt service routine (ISR) can check this register
to determine the source of an external interrupt. If a Port
B pin enabled for generating interrupts has a set bit in the
INTF register, software must clear the bit prior to exiting to
prevent repeated calls to the ISR.
The Port B interrupt logic is asynchronous (e.g. functions
without a clock in clock-stop mode). A side effect is that
there is a 2-cycle delay between the instruction that clears
a INTF bit and the bit being cleared. This means that
software must clear the bit at least 2 cycles before
executing a return from interrupt (reti) instruction.
5.1.8 INTE Register
The INTE register consists of 8 interrupt enable bits that
correspond to the 8 pins of Port B. A Port B pin is enabled
as a source of interrupts by setting the corresponding bit
in the INTE register. The pin is disabled as an interrupt
source by clearing the corresponding INTE bit, but takes
up to 1 core clock cycle for the interrupt to be disabled.
5.1.9 Port Configuration Upon Power-Up
On power-up, all the port control registers (RxDIR) are
initialized to 0xFF. Therefore, each port pin is configured
as a high-impedance input. This prevents any false
signalling to external components which could occur if the
ports were allowed to assume a random configuration at
power-up.
5.2 Timer 0
Timer 0 is an 8-bit timer with an 8-bit prescaler intended
to generate periodic interrupts for ipModule™ instances
that require being called at a constant rate, such as UART
and DTMF functions. When the T0TMR register counts up
to FF and rolls over to 00, the T0IF flag in the T0CFG
register will be set, and an interrupt will occur if the T0IE
and T0EN bit are set (see T0CFG register description in
Section 7.1.20). To clear the interrupt, either the T0IE or
T0EN bit should be cleared, and then the T0IF flag must
be cleared.
Note: If T0IF is not cleared after disabling the Timer0
interrupt (T0IE = 0) or disabling Timer0 (T0EN = 0), it is
assumed that another interrupt has occurred, and the
mov w,#0x03 ;load W with the value 0x03
;(bits 3:2 low, and bits 1:0
;high)
mov 0x022,w ;write 0x03 to RADIR
;register
mov w,#0x01 ;load W with the value 0x01
;(bits 3:1 low, and bit 0
;high)
mov 0x021,w ;write 0x01 to RAOUT
;register
52 www.ubicom.com
IP2012 / IP2022 Data Sheet
interrupt will occur on the next return, or when GIE is set
(enabling nested interrupts - see Section 3.7.2).
The Timer 0 interrupt is also supported in the instruction
set by an option for the reti instruction which adds the
W register to the T0TMR register when returning from an
interrupt. Figure 5-3 shows the Timer 0 logic.
Operation of Timer 0 to generate periodic interrupts:
• T0TMR = 00 when entering ISR from T0 interrupt
•Keeps counting up while in ISR
•Add W to T0TMR with execution of reti (refer to Ta-
ble 3-5). Interrupt frequency is adjusted by adjusting
value loaded in W, and depending on core clock divid-
er, since T0TMR runs on the system clock. If W added
to T0TMR exceeds 0xFF, no interrupt is taken until the
T0TMR rolls over from 0xFF to 0x00 again. If the
T0TMR rolls over during the 3 core cycles in the re-
turn from interrupt, the ISR is executed again (and
never again returns to mainline code as long as the
ISR executes the same).
Note: Do not enable Timer 0 interrupt before enabling the
Timer 0 itself.
Figure 5-3 Timer 0 Block Diagram
The control and status register for Timer 0 is the T0CFG
register, described in detail in Section 7.1.20.
Note: T0IF can only be asserted when T0IE = 1, T0EN =
1 and T0TMR overflow occurs.
5.3 Real-Time Timer (RTTMR)
The Real-Time Timer is an 8-bit timer intended to provide
a periodic system wake-up interrupt. Unlike the other
peripherals (except the Watchdog Timer and Port B
interrupts), the Real-Time Timer continues to function
when the system clock is disabled. For those applications
which spend much of their time with the OSC clock
oscillator turned off to conserve power, there are 5
available mechanisms to exit this mode: external reset
(RST pin), reset from the Watchdog Timer, reset from
Brown-out, interrupt from a Port B input, and interrupt
from the Real-Time Timer. By using an interrupt rather
than reset, more of the CPU state is preserved and some
reset procedures such as initializing the port direction
registers can be skipped. Figure 5-4 shows the Real-Time
Timer logic. (When RTCLK1 is not used, it should be tied
to GND.)
When the RTTMR register counts up to FF and rolls over
to 00, the RTIF flag in the RTCFG register will be set, and
an interrupt will occur if the RTIE and RTEN bit are set
(see RTCFG register description in Section 7.1.9). To
clear the interrupt, either the RTIE or RTEN bit should be
cleared, and then the RTIF flag be cleared.
Note: A nop is required between a speed instruction
and an instruction that enables or writes to RTTMR.
Note: If RTIF is not cleared after disabling the Real-Time
Timer interrupt (RTIE = 0) or disabling the Real-Time
Timer (RTEN = 0), it is assumed that another interrupt has
occurred, and the interrupt will occur on the next return, or
when GIE is set (enabling nested interrupts - see Section
3.7.2).
Note: The system clock must be slower or equal to the
RTCLK clock, for a write to the RTTMR to work correctly.
The real-time timer is readable and writable as the
RTTMR register. The control and status register for the
timer is the RTCFG register, as described in Section
7.1.9.
The RTEOS bit (XCFG bit 5, see Section 7.1.26) selects
the sampling mode for the external input. If the RTEOS bit
is set, the external input is over-sampled with the system
clock. The CPU can always read the value in the RTTMR
register, if the system clock is at least twice the frequency
of the external input. If the system clock source is
changed to RTCLK or turned off, then the RTEOS bit must
be clear for the Real-Time Timer to function.
Note: if the RTEOS bit is cleared, expect a 3 cycle system
clock delay for the overflow interrupt, due to
synchronization circuitry.
If the RTEOS bit is clear then the external input directly
clocks the Real-Time Timer (i.e. RTCLK is not
oversampled). The Real-Time Timer will always function
whether the clock input is synchronous or asynchronous.
However, the CPU cannot reliably read the value in the
RTTMR register unless the RTCLK clock is synchronous
to the system clock (RTEOS=1).
If the value in the RTTMR register does not need to be
used by the CPU (i.e. only the interrupt flag is of interest),
then the RTEOS bit should be clear (i.e. RTCLK not
oversampled), which allows the Real-Time Timer to
function for any configuration of the system clock.
515-091c.eps
8-bit
T0TMR
Register
Data Bus
T0IE
T0IF
8-Bit
Prescaler
8
System
Clock
T0ENT0PS 3:0
IP2012 / IP2022 Data Sheet
www.ubicom.com 53
If the value in the RTTMR register needs to be used by the
CPU, but the Real-Time Timer is not required to function
when the system clock is set to RTCLK or turned off, then
the RTEOS bit should be set to ensure the CPU can
reliably read the RTTMR register.
If the value in the RTTMR register needs to be used by the
CPU and the Real-Time Timer is required to function
when the system clock is set to RTCLK or off, then
software must change the RTEOS bit when changing the
system clock source. To read the RTTMR register when
the system clock is not synchronous to the RTCLK, the
RTEOS bit must be set to ensure reliable operation.
Before the system clock is changed to RTCLK or turned
off, the RTEOS bit must be clear (i.e. RTCLK not
oversampled) for the Real-Time Timer to continue to
function.
Note: When using development tools in single stepping
mode, the RTSS bit must be cleared and RTEOS must be
set, otherwise the counter will behave erratically.
Note: Care must be exercised if Port B interrupts and
RTTMR interrupts are enabled, because the RTTMR may
receive sporadic clocks during crystal startup while the
system clock is waiting for WUDX2:0 (see Figure 3-17).
Figure 5-4 Real-Time Timer Block Diagram
515-015b.eps
OSC1
RTCLK1
RTEOS
RTSS
0
1
0
1
0
1
RTTMR
CLK
Enable
System Clock
"1"
RTEOS
One
Shot
54 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.4 Multi-Function Timers (T1 and T2)
Figure 5-5 Multifunction Timer Block Diagram
The IP2012 / IP2022 contains two independent 16-bit
multi-function timers, called T1 and T2 (notated below as
Tx). These versatile, programmable timers reduce the
software burden on the CPU in real-time control
applications such as PWM generation, motor control, triac
control, variable-brightness display control, sine-wave
generation, and data acquisition.
Each timer consists of a 16-bit counter register supported
by a dedicated 16-bit capture register and two 16-bit
compare registers. The second compare register can also
serve as capture register. Each timer may use up to four
external pins: TxCPI1 (Capture Input), TxCPI2 (Capture
Input), TxCLK (Clock Input), TxOUT (Output). These pins
are multiplexed with general-purpose I/O port pins. The
port direction register has priority over the timer
configuration, so the port direction register must be
programmed appropriately for each of these four signals if
their associated timer functions are used.
Figure 5-5 is a block diagram showing the registers and
I/O pins of one timer. Each timer is based on a 16-bit
counter/timer driven by a 15-bit prescaler. The input of the
prescaler can be either the system clock or an external
clock signal which is internally synchronized to the system
clock. The counter cannot be directly written by software,
but it may be cleared by writing to the TxRST bit in the
TxCTRL register.
5.4.1 Timers T1, T2 Operating Modes
Each timer can be configured to operate in one of the
following modes:
• Pulse-Width Modulation (PWM)
• Timer
• Capture/Compare
PWM Mode
In PWM Mode, the timer can generate a pulse-width
modulated signal on its output pin, TxOUT. The period of
the PWM cycle (high + low), in number of system clocks,
is specified by the value in the TxCAP2H/TxCAP2L
register. The high time of the pulse is specified by the
value in the TxCMP1H/TxCMP1L register.
PWM mode can be used to generate an external clock
signal that is synchronous to the IP2012 / IP2022 system
clock. For example, by loading TxCMP1H/TxCMP1L with
1 and TxCAP2H/TxCAP2L with 2 (the high registers must
be written last for this to work), a symmetric external clock
can be generated at ∫ the frequency of the system clock.
In some applications, this can eliminate crystals or
oscillators required to produce clock signals for other
components in the system. SerDes GPSI mode can also
produce clock outputs.
The 16-bit counter/timer counts upward, starting with the
TxOUT output driven high. After reaching the value stored
in the TxCMP1H/TxCMP1L register minus one, at the next
clock edge the TxOUT pin is driven low. The counter/timer
is unaffected by this event and continues to increment.
After reaching the value stored in the TxCAP2H/TxCAP2L
register minus one, at the next clock edge the timer is
cleared. When the counter is cleared, the TxOUT output is
driven high, unless the TxCMP1H/TxCMP1L register is
clear, in which case the TxOUT pin is driven low.
515-005a.eps
TxCPI1 TxCAP1H/TxCAP1L
Register
TxOUT
TxCMP1H/TxCMP1L
Register
TxCLK
D Q
TxCPI2
TxCAP2H/TxCAP2L
or TxCMP2H/TxCMP2L
Register
TxCNTH/TxCNTL
Register
15-Bit Prescaler
System
Clock
System
Clock
0
1
TxCFG1L bit 6
(OEN)
RAOUT bit 3 for T1
RBOUT bit 3 for T2
RA2 for T1
RB2 for T2
RA0 for T1
RB0 for T2
RA1 for T1
RB1 for T2
RA3 for T1
RB3 for T2
IP2012 / IP2022 Data Sheet
www.ubicom.com 55
There are two special cases. When the
TxCMP1H/TxCMP1L register is clear, the TxOUT pin is
driven with a continuous low, corresponding to a duty-
cycle of 0%. When the value in the TxCMP1H/TxCMP1L
register is equal to the value in the TxCAP2H/TxCAP2L
register, the TxOUT output is driven with a continuous
high, corresponding to a duty-cycle of 100%.
The behavior of the timers is undefined when the value in
the TxCMP1H/TxCMP1L register is greater than the value
in the TxCAP2H/TxCAP2L register.
The timer is glitch-free no matter when the
TxCMP1H/TxCMP1L register or the TxCMP2H/TxCMP2L
register are changed relative to the value of the internal
counter/timer. The new duty cycle or period values do not
take effect until the current PWM cycle is completed (the
counter/timer is reset).
Interrupts, if enabled through the TxCFG1H register, can
be generated whenever the timer output is set or cleared.
If the TxCMP1H/TxCMP1L register is clear, or if the value
in the TxCMP1H/TxCMP1L register is equal to the value
in the TxCAP2H/TxCAP2L register, an interrupt can be
generated each time the counter/timer is reset to zero.
In PWM mode, the Capture 1 input remains active (if
enabled by the CPI1EN bit in the TxCFG1L register) and,
when triggered, captures the current counter/timer value
into the TxCAP1 register.
The multifunction timers can be configured to interrupt on
a Capture 1 event and reset the counter/timer on the
event. For PWM operation without Capture 1, software
must disable the Capture 1 input by clearing the CPI1EN
bit in the TxCFG1L register.
Timer Mode
This is not a separate timer mode (from the hardware
point of view), but is a conceptual mode for programmers.
It is the PWM mode, except that software disables the
timer output by clearing the OEN bit in the TxCFG register.
Capture/Compare Mode
In Capture/Compare mode, one or both of the timer
capture inputs (TxCPI1 and TxCPI2) may be used. Their
pin functions must be enabled in the TxCFG1 register.
Each capture input can be programmed in the TxCFG2
register to trigger on a rising edge, falling edge, or both
rising and falling edges.
When a trigger event occurs on either capture pin, the
current value of the counter/timer is captured into the
TxCAP1H/TxCAP1L register or the TxCAP2H/TxCAP2L
register for that input pin.
The counter/timers can also be configured to reset on a
TxCPI1 input event, in which case the value of the
counter/timer before it was reset is captured in the
TxCAP1H/TxCAP1L register and the counter/timer is
reset to zero. This mode is useful for measuring the
frequency (or width) of external signals. By using both
capture inputs and configuring them for opposite edges,
the duty cycle of a signal can also be measured. To avoid
wasting I/O port pins in this configuration, the CPI2EN bit
in the TxCFG1L register is provided to internally tie the
TxCPI1 and TxCPI2 inputs together, which frees the
TxCPI2 pin to be used as a general-purpose I/O port pin.
An interrupt can be generated for any capture event and
for counter/timer overflows.
This mode also features an output-compare function. The
TxCMP1H/TCMP1L register is constantly compared
against the internal counter/timer. When the counter/timer
reaches the value of the TxCMP1H/TxCMP1L register
minus one, at the next counter clock the TxOUT output is
toggled. The TxOUT output, if enabled via the OEN bit,
can be driven high or low by writing to the TOUTSET and
TOUTCLR bits in the TxCFG2L register. An interrupt can
be enabled for this event.
Interrupts
When a Multi-Function Timer interrupt occurs, the
corresponding interrupt flag (depending on the mode;
OFIF, CAP2IF/CMP2IF, CAP1IF or CMP1IF) in the
TxCFG1H register will be set, and an interrupt will occur if
the TMREN bit (TxCFG1L register), the TxIE bit (TCTRL
register) and an interrupt source is enabled (depending on
the mode; OFIE, CAP2IE/CMP2IE, CAP1IE or CMP1IE)
are set (TxCFG1H register). To clear the interrupt, either
the TMREN bit, TxIE bit or the interrupt source (OFIE,
CAP2IE/CMP2IE, CAP1IE or CMP1IE) should be cleared,
and then the interrupt flag (OFIF, CAP2IF/CMP2IF,
CAP1IF or CMP1IF) should be cleared.
Note: The interrupt flag can only be asserted when the
multi-function timers are enabled, the timer interrupts are
enabled, an interrupt source is enabled, and timer event
occurs.
Note: If the interrupt flag is not cleared after disabling
either the interrupt enable or the Multi-Function Timer
enable (TMREN = 0), it is assumed that another interrupt
has occurred, and the interrupt will occur on the next
return, or when GIE is set (enabling nested interrupts -
see Section 3.7.2).
56 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.4.2 T1 and T2 Timer Pin Assignments
The following table lists the I/O port pins associated with
the Timer T1 and Timer T2 I/O functions.
5.4.3 T1 and T2 Timer Registers
Each timer has six 16-bit register pairs, which are
accessed as 8-bit registers in the special-purpose register
space. There is also one 8-bit register shared by both
timers.
TxCNTH/TxCNTL Register
The TxCNTH/TxCNTL register indicates the value of the
counter/timer and increments synchronously with the
rising edge of the system clock. This register is read-only.
The timer counter may be cleared by writing to the TxRST
bit in the TCTRL register.
Reading the TxCNTL register returns the least-significant
8 bits of the internal TxCNT counter and causes the most-
significant 8 bits of the counter to be latched into the
TxCNTH register. This allows software to read the
TxCNTH register later and still be assured of atomicity.
TxCAP1H/TxCAP1L Register
The TxCAP1H/TxCAP1L register captures the value of
the counter/timer when the TxCPI1 input is triggered. This
register is read-only.
Reading the TxCAP1L register returns the least-
significant 8 bits of an internal capture register and causes
the most-significant 8-bits of the register to be latched into
the TxCAP1H register. This allows software to read the
TxCAP1H register later and still be assured of atomicity.
TxCMP1H/TxCMP1L Register
In Capture/Compare mode, the TxOUT output pin is
toggled (if enabled by the OEN bit in the TxCFG1 register)
when the counter/timer increments to the value in the
TxCMP1 register. In this mode, the value written to the
TxCMP1 register takes effect immediately.
Writing to the TxCMP1L register causes the value to be
stored in the TxCMP1L register with no other effect.
Writing to the TxCMP1H register causes an internal
compare register to be loaded with a 16-bit value in which
the low 8 bits come from the TxCMP1L register and high
8 bits come from the value being written to the TxCMP1H
register. Software should write the TxCMP1L register
before writing the TxCMP1H register, because writing to
the TxCMP1H register is used as an indication that a new
compare value has been written. Writing to the TxCMP1H
register is required for the new compare value to take
effect - this means that TxCMP1H must be written AFTER
TxCMP1L for the value to have any effect. In PWM mode,
the 16-bit number latched into the internal compare
register by writing to the TxCMP1H register does not take
effect until the end of the current PWM cycle.
Reading the TxCMP1H or TxCMP1L registers returns the
previously written value whether or not the value stored in
these registers has been transferred to the internal
compare register by writing to the TxCMP1H register.
TxCAP2H/TxCAP2L or TxCMP2H/TxCMP2L Register
This register may be called the TxCAP2H/TxCAP2L
register or TxCMP2H/TxCMP2L register.
In PWM mode, this register determines the period of the
PWM signal. In this mode, this register is both readable
and writable. However, on writes the value is not applied
until the end of the current PWM cycle.
Writing to the TxCAP2L register causes the value to be
stored in the TxCAP2L register with no other effect.
Writing to the TxCAP2H register causes an internal
compare register to be loaded with a 16-bit value in which
the low 8 bits come from the TxCAP2L register and the
high 8 bits come from the value being written to the
TxCAP2H register. Software should write the TxCAP2L
Table 5-1 Timer T1/T2 Pin Assignments
I/O Pin Timer T1/T2 Function
RA0 Timer T1 Capture 1 Input
RA1 Timer T1 Capture 2 Input
RA2 Timer T1 External Event Clock Source
RA3 Timer T1 Output
RB0 Timer T2 Capture 1 Input
RB1 Timer T2 Capture 2 Input
RB2 Timer T2 External Event Clock Source
RB3 Timer T2 Output
IP2012 / IP2022 Data Sheet
www.ubicom.com 57
register before writing the TxCAP2H register, because
writing to the TxCAP2H register is used as an indication
that a new compare value has been written. Writing to the
TxCAP2H register is required for the new compare value
to take effect. In PWM mode, the 16-bit number latched
into the internal compare register by writing to the
TxCAP2H register does not take effect until the end of the
current PWM cycle.
Reading the TxCAP2H or TxCAP2L registers returns the
previously written value regardless of whether the value
stored in these registers has been transferred to the
internal compare register by writing to the TxCAP2H
register.
In Capture/Compare mode, this register captures the
value of the counter/timer when the TxCPI2 input is
triggered. In this mode, this register is read-only.
Reading the TxCAP2L register returns the least-
significant 8 bits of an internal capture register and causes
the most-significant 8-bits to be latched into the TxCAP2H
register. This allows software to read the TxCAP2H
register later and still be assured of atomicity.
TxCFG1H/TxCFG1L Register
Selects timer operation mode, pin functions, interrupts
and other configuration settings. See Section 7.1.21 for
the description of TxCFG1H and Section 7.1.23 for the
description of TxCFG1L.
TxCFG2H/TxCFG2L Register
Selects capture input trigger edges, prescaler setting, and
other configuration settings. See Section 7.1.22 for the
description of TxCFG2H and Section 7.1.24 for the
description of TxCFG2L.
TCTRL Register
Unlike the other timer control registers, one TCTRL
register is used to synchronize both timers. Setting the
TxRST bit clears the TxCNTH/TxCNTL register pair and
the prescaler counter, which allows global
synchronization of all timers on the device. There are also
individual timer interrupt-enable bits. See Section 7.1.25
for description.
5.5 Watchdog Timer (WDT)
A Watchdog Timer (WDT) is available for recovering from
unexpected system software hang-ups. When the
Watchdog Timer is enabled, software must periodically
clear the timer by executing a cwdt instruction.
Otherwise, the timer will overflow, which resets the IP2012
/ IP2022 but doesn’t clear the WD bit in the PSPCFG
register (this bit should be set before the first cwdt
instruction is executed). Any other source of reset clears
the WD bit, so software can use this bit to identify a reset
caused by the Watchdog Timer. The Watchdog Timer is
shown in Figure 5-6.
Figure 5-6 Watchdog Timer
The Watchdog Timer is enabled by setting the WDTE bit
in the FUSE1 register. The time period between coming
out of reset or clearing the timer and timer overflow is
controlled by the WDPS2:0 bits in the FUSE1 register, as
discussed in Section 3.10.2.
Since the watchdog timer period varies by up to 50% over
temperature and voltage, the minimum timeout period
selected in FUSE1 that works in nominal conditions,
should not be used. For instance, if the 640ms setting
works in nominal conditions, the 1280ms setting should
be used in production.
The Watchdog Timer register is not visible to software.
The only feature of the Watchdog Timer visible to software
is the WD bit in the PSPCFG register (see Section 7.1.8).
Note: When using the development tools, the watchdog
timer is disabled while in debug mode, except when “Run”
command is issued. The break and breakx
instructions suspend the Watchdog Timer, so that debug
mode works correctly. Therefore, if the watchdog feature
is used, the break and breakx instructions should not
be used, and the program RAM should be initialized to
instructions that do not include break and breakx.
515-076a.eps
8-Bit
Timer
Watchdog
Timer Reset
Prescaler
Internal
RC Clock
(14 kHz)
WDTE WDPS 2:0 cwdt Instruction
58 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.6 Serializer/Deserializer (SERDES)
There are two SERDES units in the IP2022 (one unit in
the IP2012), which support a variety of serial
communication protocols, including GPSI, SPI, UART,
USB, and 10Base-T Ethernet. By performing data
serialization/deserialization in hardware, the CPU
bandwidth needed to support serial communication is
greatly reduced, especially at high baud rates. Providing
two units allows easy implementation of protocol
conversion or bridging functions between the two high-
speed serial interfaces, such as a USB to 10Base-T
Ethernet bridge.
Each SERDES unit uses up to 8 external digital signals:
SxCLK, SxRXD, SxRXM, SxRXP, SxTXM, SxTXME,
SxTXP, and SxTXPE/SxOE. The signals for SERDES1
are multiplexed with the Port E pins, and the signals for
SERDES2 are multiplexed with the Port F pins (IP2022
only). The port direction bits must be set appropriately for
each pin that is used. Not all signals are used in all
protocol modes. See Table 5-3 for details on signal port
pin usage in various protocol modes. In addition to the
digital signals, there are also two analog signals only used
in 10Base-T Ethernet mode: SxRX+ and SxRX-.
Note: Proper operation of the SERDES requires that the
core-clock be present — don’t turn off core clock while
SERDES is still transmitting.
SERDES Configuration Registers
The descriptions for the SxMODE, SxRSYNC, SxSMASK,
SxRCFG, SxRCNT, SxTCFG, SxTMRH/SxTMRL and
SxINTE/SxINTF registers can be found in Section 7.1.
Note: A one cycle delay is required between consecutive
read-modify-write instructions to the same SERDES
register (for example, clrb reg and setb reg)
5.6.1 SERDES TX/RX Buffers
SxRBUFH/SxRBUFL Registers
16-bit register pair for unloading received data. The RXBF
bit in the SxINTF register indicates when new data has
been loaded into this register. If the corresponding bit in
the SxINTE register is set, an interrupt is generated.
SxTBUFH/SxTBUFL Registers
16-bit register pair for loading data to be transmitted. The
TXBE bit in the SxINTF register indicates when the data
has been transmitted and the register is ready to be
loaded with new data. If the corresponding bit in the
SxINTE register is set, an interrupt is generated.
Regardless of the number of bits to be transmitted, both
the high and low registers need to be written to initiate
transmission.
5.6.2 SERDES Configuration
Software prepares a SERDES unit to receive data by
programming the receive shift count register (SxRCFG)
and the clock select bits in the SxMODE register
appropriately for the selected protocol. The SxRCFG
register is copied to an internal counter, and when that
number of bits of data has been received, the received
data is loaded into the SxRBUF register.
In 10Base-T, GPSI, or USB mode, when an EOP is
detected, the SxRCNT register is loaded with the number
of bits actually received, the EOP bit of the SxINTF
register is set, and the data bits are loaded into the
SxRBUF register. The RXBF bit in the SxINTE register
can be set to enable an interrupt on this event.
The SxTXP and SxTXM pins correspond to the differential
outputs of the USB or Ethernet bus. Other serial protocols
require only one output pin, which is SxTXP by default.
The SxTXP and SxTXM pins have high current outputs for
driving Ethernet magnetics directly without the use of
transceivers.
When the clock select register is programmed with the
value for 10Base-T, the transmit pre-emphasis
requirement enables the SxTXPE and SxTXME outputs,
which have a 50ns-delayed version of the transmit output
that is resistively combined outside the chip before driving
the magnetics.
For transmitting, software must specify the number of bits
to transmit (specified in the SxTCFG register) and load
the data in the SxTBUF register. This data is then
transferred to an internal register, from which it is serially
shifted out to the transmit logic. The TXBE bit in the
SxINTE register can be set to enable an interrupt when
the data has been transferred from the SxTBUF register.
When there is a transmit buffer underrun event (i.e. all of
the data has been shifted out from the internal register,
but the SxTBUF register has not been reloaded), an EOP
condition is generated on the SxTXP and SxTXM outputs
after an internal counter decrements to zero. The TXEOP
bit in the SxINTE register can be set to enable an interrupt
when an underrun event occurs.
For protocols other than USB and Ethernet, the EOP
generator is bypassed.
IP2012 / IP2022 Data Sheet
www.ubicom.com 59
5.6.3 SERDES Interrupts
Figure 5-7 shows the interrupt logic for the two SERDES
units. For a detailed description of the SxINTE/SxINTF
register bits, refer to Section 7.1.10.
Figure 5-7 SERDES Interrupt Logic
60 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.6.4 Protocol Modes
Table 5-2 shows the features which are enabled for each
protocol, as controlled by the PRS3:0 bits in the SxMODE
register. These features affect which registers and register
fields are used, for example the SxRSYNC register is only
used in the USB and 10Base-T modes. The protocol
mode also affects the signal usage, as shown in Table 5-3.
Pins not used for protocols can be used for general I/O.
•SxCLK - Serial Clock in SPI or GPSI Slave modes,
optional external SERDES clock input for USB or
UART modes.
•SxRXP - Positive-side differential input (USB only),
Slave Select (for SPI Slave), or data valid (GPSI).
•SxRXM - Negative-side differential input (USB only).
•SxRXD - Serial data for USB, UART, SPI and GPSI
modes (10base-T Ethernet only when comparator is
used).
•SxTXPE - Positive-side delayed differential output for
pre-emphasis (10base-T Ethernet), output enable for
external transceiver (USB), or data valid for GPSI
mode.
•SxTXP - Positive-side differential output (10base-T
Ethernet and USB modes), or serial data (UART, SPI
and GPSI modes).
•SxTXM - Negative-side differential output (10base-T
Ethernet and USB modes), transmit clock (GPSI
Slave), or transmit and receive clock (GPSI Master).
•SxTXME - Negative-side delayed differential output
for pre-emphasis (10base-T Ethernet), or TxBUSY in
GPSI mode.
•SxRX+ - Positive-side analog differential input, used
for 10base-T Ethernet squelch function.
•SxRX- - Negative-side analog differential input, used
for 10base-T Ethernet squelch function.
Table 5-2 Protocol Features
PRS3:0 Mode Encoding
Method
Differential or
Single-Ended?
Synchronization
Register
Enabled?
EOP
Generation/
Detection?
Bit
Stuffing/
Unstuffing?
Pre-Emphasis
Outputs
Enabled?
0001 10Base-T Manchester Differential Yes Yes N/A Yes
0010 USB Bus NRZI Differential Yes Yes Yes N/A
0011 UART None Single-Ended No N/A N/A N/A
0101 SPI None Single-Ended No Yes N/A N/A
0110 GPSI None Single-Ended No Yes N/A N/A
Table 5-3 SERDES Protocol Modes And Pin Usage
Signal Names SxCLK SxRXP SxRXM SxRXD SxTXPE SxTXP SxTXM SxTXME SxRX+ SxRX-
SERDES1 Pins RE0 RE1 RE2 RE3 RE4 RE5 RE6 RE7 RG5 RG4
SERDES2 Pins
(IP2022 only) RF4 RF5 RF6 RF7 RF0 RF1 RF2 RF3 RG7 RG6
Mode
10Base-T
Ethernet
---RXD
Note 1
TxD+
(O)
Tx+
(O)
Tx-
(O)
TxD-
(O)
RX+
(I)
RX-
(I)
USB Bus Optional VP (I) VM (I) RCV (I) OE (O) VPO (O) VMO (O) - - -
UART Optional - - RXD - TXD - - - -
SPI Master SCLK - - DI (I) - DO (O) - - - -
Slave SCLK SS (I) - DI (I) - DO (O) - - - -
GPSI
Master - TxEN
(O)
- TxD
(O)
RxEN
(I)
RxD
(I)
TxCLK/
RxCLK (O)
TxBUSY
(I)
--
Slave RxCLK
(I)
TxEN
(O)
- TxD
(O)
RxEN
(I)
RxD
(I)
TxCLK
(I)
TxBUSY
(I)
--
1.
I.
O.
Used in comparator mode only.
Input
Output
IP2012 / IP2022 Data Sheet
www.ubicom.com 61
5.6.5 10base-T Ethernet
Hardware
Each SERDES unit provides 4 transmission digital
signals: SxTXM, SxTXME, SxTXP, and SxTXPE. These
signals for SERDES1 are multiplexed with the Port E pins,
and the signals for SERDES2 (IP2022 only) are
multiplexed with the Port F pins. The differential receive
signals are multiplexed on port G pins for 10Base-T
Ethernet mode: SxRX+ and SxRX-. The port direction bits
must be set appropriately for each pin that is used. Other
unused pins from SERDES and port G remain available
for other functions usage. See Table 5-4 for details on
signal port pin usage in various protocol modes.
Note: Proper operation of the 10Base-T requires that the
core-clock be present as SERDES master clock is derived
from it - don't turn off core clock while SERDES is still
transmitting.
Figure 5-8 shows the clock/data separation and End-of-
Packet (EOP) detection logic of a 10Base-T receiver unit.
The SxRXP and SxRXM pins correspond to the
differential inputs. Providing both inputs allows sensing of
an EOP condition. To set up a SERDES unit for 10Base-T
Ethernet, the input data from a differential line receiver is
connected to the SxRX+ and SxRX- input. The signals
designated Tx+, Tx-, TxD+, and TxD- correspond to the
SxTXP, SxTXM, SxTXPE, and SxTXME pins of the
corresponding SERDES. These pins are connected to an
RJ45 jack through a transformer with terminations.
Figure 5-8 Clock/Data Separation and EOP Detection
Figure 5-9 shows an example circuit. RTXPE, RTXME,
RTXP, RTXM and RL values vary depending on the
Ethernet magnetics used. Please refer to IP2022 Native
Ethernet application notes for more details.
For 10Base-T Ethernet operation, each SERDES is
equipped with a squelch circuit for discriminating between
noise, link pulses, and data. Link pulses are sent
periodically to keep the channel open when no data is
being transmitted. The squelch circuit handles link pulse
detection, link pulse polarity detection, carrier sense, and
EOP detection.
The 10Base-T mode requires only a fixed SFD (start of
frame) pattern, so the SFD pattern for 10Base-T is
hardwired to be 11010101 and the synchronization
pattern register (SxRSYNC) is used to configure features
of 10Base-T other than SFD pattern. Refer to Section
7.1.14 for detailed information. The incoming data stream,
after passing through the polarity inversion logic (which
can be turned on or off under software control) is
compared to the synchronization pattern. Once a match is
found, an internal counter is set to zero and data is shifted
into a shift register. The synchronization matching
operation is then disabled until an EOP condition is
detected, because the synchronization pattern potentially
could be embedded in the data stream as valid data.
515-003e.eps
Clock/Data
Separation
and
Start
Condition
Detection
Receive Polarity
Reversal Bit
Receive
Data
Data
EOP
EOP
Detection
Synchronization
Pattern
Register
(SxRSYNC)
SxRXD
Input
post-PLL Clock
OSC Clock
Ethernet
Squelch
Circuit
SxRx+
SxRx-
SxCLK
SxMODE
62 www.ubicom.com
IP2012 / IP2022 Data Sheet
Figure 5-9 Ethernet Interface Example
Figure 5-10 shows the receive data paths. When an EOP
is detected the SxRCNT register is loaded with the
number of bits actually received, the EOP bit of the
SxINTF register is set, and the data bits are loaded into
the SxRBUF register. The RXBF bit in the SxINTE register
can be set to enable an interrupt on this event.
The data encode block performs 10Base-T Manchester
encoding. The encoded TX signal are sent to TX pins in a
differential mode. The encode block is bypassed for all
other protocols. The SxTXP and SxTXM pins have high
current outputs for driving Ethernet magnetics directly
without the use of transceivers. The pre-emphasis TX
outputs are enabled on SxTXPE and SxTXME outputs,
which have a 50ns-delayed and inverted version of the
transmit outputs. The resistively combined TX outputs
outside the chip are used to drive the magnetics. The
output pins of the serializer are driven low when not
transmitting.
Figure 5-11 shows the transmit data paths. For
transmitting, software must specify the number of bits to
transmit (specified in the SxTCFG register) and load the
data in the SxTBUF register. This data is then transferred
to an internal register, from which it is serially shifted out
to the transmit logic. The TXBE bit in the SxINTE register
can be set to enable an interrupt when the data has been
transferred from the SxTBUF register. When there is a
transmit buffer underrun event (i.e. all of the data has been
shifted out from the internal register, but the SxTBUF
register has not been reloaded), an EOP condition is
automatically generated on the TX output pins after an
internal counter decrements to zero. The TXEOP bit in the
SxINTE register can be set to enable an interrupt when an
underrun event occurs.
Table 5-4 10base-T Ethernet Interface Signal and Port Pin Usage
10base-T
Signal Name
SERDES
Signal Name
SERDES1P
in Name
SERDES2
Pin Name
(IP2022 only)
Direction Description
Tx+ SxTXP RE5 RF1 Output Plus-side differential output
Tx- SxTXM RE6 RF2 Output Minus-side differential output
TxD+ SxTXPE RE4 RF0 Output Plus-side differential output with pre-
emphasis
TxD- SxTXME RE7 RF3 Output Minus-side differential output with pre-
emphasis
Rx+ SxRX+ RG5 RG7 Input Plus-side analog differential input, used
for 10base-T Ethernet squelch function
Rx- SxRX- RG4 RG6 Input Minus-side analog differential input, used
for 10base-T Ethernet squelch function
Clock recovery Rx+
Rx-
Manchester
decoding
coding
Serializer/Deserializer
RJ45
Data
Encoder
Data
Decoder
Manchester
IP2022 or IP2012
Rtxpe
Squelch
SxRX-
SxRX+
SxTXPE
Ethernet
Magnetics
(LPF required)
RL
SxTXM
SxTXME
SxTXP
TxD+
TxD-
Tx-
Tx+
Rtxme
Rtxm
Rtxp
515-064d.eps
IP2012 / IP2022 Data Sheet
www.ubicom.com 63
Figure 5-10 Receive Data Paths
Figure 5-11 Transmit Data Paths
Software
The SERDES 10Base-T mode is designed to run at a
fixed 8x oversampling for line receiver. So a fixed 80MHz
master must be configured through SxMODE. The core
PLL clock multiplier must be programmed to be an integer
multiple of 80MHz for 10Base-T operation. The received
data stream is used, together with the clock recovery
circuit, to recover the original transmit clock and data.
Typical operation with Ubicom's SDK uses a 4.8 MHz
crystal with a PLL post-divide by 2 to yield 120 MHz core
operation, or a 3.2 MHz crystal with no PLL post-divide to
yield 160 MHz operation.
Software must perform the following functions:
•Polarity detection and reversal.
• Carrier sense.
•Jabber detection.
• Link integrity test and link pulse generation.
• Random back off in case of collision.
•When a collision is detected, sending a 32-bit jam se-
quence. Collisions can be detected by positive detec-
tion of carrier sense during active transmission.
•Formation of Ethernet packet by putting the preamble,
SFD, destination address, source address,
length/type, MAC client data into the transmit buffer.
Frame check computation can be done in software or
through the LFSR units (see Section 5.9).
•MAC layer functions.
!""
!#$
!
&!$
'
515-018a.eps
Transmit
Configuration
Register
(SxTCFG)
T
ransmit
Clock
Data
Encoder
Transmit
Buffer
Register
(SxTBUF)
Data Bus
SxTXPE
Pre-Emphasis
EOP
Generator
SxTXME
SxTXP
SxTXM
TXBE, TXEOP
Transmit
Interrupt
64 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.6.6 USB
Each SERDES provides support for USB revision 1.1 host
and device modes of operation.
Hardware
To set up a SERDES unit for USB mode, the received data
output of the USB transceiver should be connected to
SxRXD. The VP and VM pins of the transceiver are
connected to the SxRXP and SxRXM pins to allow
detection of the EOP condition. Figure 5-12 shows the
connections required between an external USB
transceiver and the IP2012 / IP2022. Table 5-6 shows the
mapping of USB signals to the SERDES pins. If desired,
an external clock source can be connected to the SxCLK
pin. For additional hardware configuration information,
please refer to USB reference design available.
Software
The SxMODE register must be programmed with values
for the desired USB mode, Full Speed or Low Speed.
Serdes clock dividers, SxTMRH and SxTMRL, also need
to be programmed to generate the appropriate frequency
according to the USB submode selection. Table 5-7
shows the PLL clock frequencies required for the low and
full speed modes of the USB mode. For example, if the
PLL clock is 240 MHz, it can be programmed at 48 MHz
for full speed with a divisor of 4 (=5). A divisor of 39 will
make it suitable for low-speed operation. If external clock
mode is selected, the clock divisor value is ignored and
the clock is used directly for USB operation.
In USB mode, the SerDes uses two registers, SxRSYNC
and SxSMASK, to detect the sync pattern marking the
beginning of a USB data stream. This sequence is defined
to be “7 zeros and a single 1” by the USB specification,
and only the last 3 bits need to be matched to start
receiving data, also defined in the specification. In order
to achieve this, SxRSYNC needs to be programmed with
0x80 and SxSMASK needs to be programmed with 0xE0.
Receive behavior is controlled by the SxRCFG and
SxRCNT registers. For USB operation, the higher 3 bits of
SxRCFG should be set to zero, and the lower 5 bits should
be set to the desired number of bits received, usually 8 or
16. SxRCNT should be cleared to make sure receive is
performed LSB first. Once the SerDes matches the USB
SYNC pattern, the internal receive count is reset to zero
and the SerDes receives bits from the line until either the
desired count is received or an EOP is encountered, at
which point the received data is transferred to the Serdes
Rx Data registers. If more data is coming in, the procedure
will be repeated. Software is responsible for reading the
data from the data registers before the next write by the
hardware. Notice that this will be a short time, if RxCFG is
configured with a small receive count, or an EOP is
received before the desired count is reached. When the
EOP is received, the SerDes remains idle until the next
match of the SYNC pattern.
Transmit behavior is controlled by the SxTCFG register.
For proper USB operation, the higher 3 bits should be set
to 100, and the lower 5 bits should be set to the desired
number of bits transmitted. Transmit can be initiated by
writing to the SerDes Tx Data registers. Notice that both
registers must be written even if the number of bits to
transmit is less than 8. If the transmit count needs to be
changed, it must be changed before the Tx Data registers
are written. For continued transmission, Tx Data registers
have to be written before the Tx bit count is reached.
Otherwise, the SerDes automatically inserts the EOP
signaling.
While receiving data, the clock/data separation circuit
performs NRZI decoding, after which bit unstuffing is
performed. This means every bit after a series of six
consecutive ones is dropped. On transmit, the SerDes
performs bit stuffing, and the clock/data separation circuit
NRZI encodes the data.
Note: While configured for USB mode, the SerDes cannot
be configured to interrupt on carrier status (RxCRS,
SxRCNT bit 5; see Section 7.1.13).
Table 5-5 shows the function of SerDes status flags in
USB Mode. See also Section 7.1.10.
Table 5-5 Summary of Status Flags in USB Mode
Flag USB Mode Function
RXERROR This bit indicates the presence of 7 ones
in the USB bit stream.
RXEOP This bit indicates the presence of EOP
signaling, which is 2 bit times of SE0 and
1 bit time of J condition on the bus.
SYND This bit indicates that SerDes
successfully matched a USB SYNC
pattern as configured in the SxSYNC
register.
TXBE This bit indicates that new data can be
written to the Tx Data registers to
continue uninterrupted data stream.
TXEOP This bit indicates that an EOP condition
was generated to signal the end of the
USB data stream. Tx Data registers
should be written with the sync pattern to
start a new stream once this condition
occurs.
IP2012 / IP2022 Data Sheet
www.ubicom.com 65
Software must perform the following functions to
implement the USB protocol for a device:
•CRC generation and checking (can be done using the
LFSR, Section 5.9).
•Detecting reset of the device function, which is indi-
cated by 10 milliseconds of a single-ended zero (SE0)
condition on the bus.
•Detecting the suspend state, which is indicated by
more than 3 milliseconds of idle. Software muse make
sure that the suspend current of 500 uA will be drawn
after 10 milliseconds of bus inactivity.
•Formation of the USB packet by putting the sync, pid,
and data into the transmit data registers and setting
the proper count.
•Endpoint and device management and other higher
level protocol tasks.
Timing Considerations
USB relies on certain timing limitations for error detection
and recovery. Response time requirements are
specifically harder to meet. ISR for USB needs to be
carefully structured to satisfy these requirements, and this
is possible because of IP2012 / IP2022's deterministic
ISR execution times. The time from the SE0 on bus to the
RXEOP indication is about 208 ns. The time from writing
to TX data registers and the data put on the bus is about
125 ns. Software tasks like address, error, CRC checking,
and determining the endpoint response have to be
carefully timed and cycle counted to assure the required
timing limitations are satisfied.
Figure 5-12 USB Interface Example
SXLINK-
PULSE
TxIdle in USB mode. This bit indicates
whether the SerDes is actively
transmitting data.
RXBF This bit indicates that Rx data registers
are written with new data and should be
read before the next desired amount of
bits are received.
RXXCRS This bit is a live status indicator in USB
mode. If set, the SerDes is receiving data
into the shift register, but there might not
be any data in the data registers yet.
Table 5-5 Summary of Status Flags in USB Mode
Flag USB Mode Function
PDIUSBP11A
SxRXD
IP2022
or
IP2012
515-034c.eps
SxRXM
SxTXP
SxTXM
D+
D-
USB Bus
SxRXP
+
-
RCV
VM
VPO
VMO
VP
SxTXPE OE
Table 5-6 USB Interface Signal Usage
USB Signal
Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
Direction Description
VP SxRXP RE1 RF5 Input Plus-side differential input
VM SxRXM RE2 RF6 Input Minus-side differential input
VPO SxTXP RE5 RF1 Output Plus-side differential output
VMO SxTXM RE6 RF2 Output Minus-side differential output
66 www.ubicom.com
IP2012 / IP2022 Data Sheet
OE SxTXPE RE4 RF0 Output Output enable
RCV SxRXD RE3 RF7 Input Receive data
Clock SxCLK RE0 RF4 Input External clock input (optional)
Table 5-6 USB Interface Signal Usage (continued)
USB Signal
Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
Direction Description
Table 5-7 Required Clock Frequencies from SerDes
Clock in USB Mode
Protocol Receive
USB 1.1 Full Speed 48 MHz
USB 1.1 Low Speed 6 MHz
IP2012 / IP2022 Data Sheet
www.ubicom.com 67
5.6.7 UART
For UART operation, two internal divide-by-16 circuits are
used. Based on the clock source (either internal or
external), the receive section and the transmit section use
two divided-by-16 clocks that potentially can be out of
phase. This is due to the nature of the UART bus transfers.
The receive logic, based on the 16x bit clock (the clock
source chosen by user), will sample the incoming data for
an falling edge. Once the edge is detected, the receive
logic counts 8 clock cycles and samples the number of
bits specified in the SxRCNT register using the bit clock
(which is obtained by dividing the clock source by 16).
Hardware
Figure 5-13 shows an example circuit to connect the
SERDES in UART mode. Table 5-8 shows the UART
signal to port pin usage.
Software
To set up a SERDES unit for UART mode, select UART
mode in the PRS3:0 bits of the SxMODE register. This
causes the data to be clocked in after a valid start bit is
detected. Make sure that the polarity selected by the
RPOREV bit in the SxRCFG register and the TPOREV bit
in the SxTCFG register match the polarity provided by the
RS-232 transceiver. (Most of them are inverted.) Make
sure the bit order is compatible with the data format (RS-
232 uses LSB-first bit order). The receiver uses 16X
oversampling, so select a SERDES clock divisor (see
Section 7.1.17 for information on the SxTMRH/L
registers) that is 16 times the desired baud rate.
To operate in UART mode, depending on the application,
either transmit or receive can be performed first. In both
cases, the configuration register needs to be programmed
with a bit count that is appropriate for the format. The bit
count depends on the number of data bits, stop bits, and
parity bits. The start bit is included in the bit count. The
receiver does not check for the presence of stop bits. To
detect framing errors caused by missing stop bits,
increase the receiver’s bit count (i.e. the RXSCNT field in
the SxRCFG register) and test the trailing bit(s) in
software.
Figure 5-13 UART Interface Example
RS-232
Transceiver
SxRXD
IP2022
or
IP2012
515-094a.eps
SxTXP
RxIN
TxOUT
RS-232
RxOUT
TxIN
Table 5-8 UART Interface Signal Usage
UART
Signal Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
Direction Description
RXD SxRXD RE3 RF7 Input Receive data
TXD SxTXP RE5 RF1 Output Transmit data
68 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.6.8 SPI
Hardware
Figure 5-14 shows example circuits to connect the
SERDES in SPI mode. Table 5-10 and Table 5-11 show
the SPI signal to port pin usage.
Configuration
The SERDES can be configured for either master or slave
mode:
SxRCFG[7] = 0: slave
SxRCFG[7] = 1: master
The SERDES SCK idle-level (i.e. when SS is de-asserted)
can be configured:
SxMode[4] (CPOL) = 0: idle is low
SxMode[4] (CPOL) = 1: idle is high
Finally, the SERDES can be configured for the phase
relationship of the SDO/SDI pins with respect to the SCK
edge:
SxMode[3] (CPHA) = 0:
SDO is set up by the other device a half clock period
before the first edge following the assertion of SS*.
Therefore SDI will be sampled by this device (the
slave) on the first edge (transition).
SxMode[3] (CPHA) = 1:
SDO is set up by the other device on the first edge
following the assertion of SS*. Therefore SDI will be
sampled by this device (the slave) on the second edge
(transition).
Note: The use of the term "edge" in the above paragraphs
implies any transition, not a specific type (i.e. rising or
falling) of transition. Therefore, "first edge" implies a rising
edge when CPOL=0, and implies a falling edge when
CPOL=1.
In the SPI scheme implemented by Motorola, which the
IP2012 / IP2022 follows, data being output on SDO and
data being sampled on SDI always occur on opposing
edges of the clock, on either master or slave. Transmitting
and sampling on the same-edge of the clock is not
supported by the SERDES.
CPOL, in conjunction with CPHA, determines which clock
edges the SERDES will be using to output and sample
data on, as given by Table 5-9.
When the SERDES is configured as a slave, the state of
the SDO line when SS* is de-asserted will be determined
by the value in the RxOUT GPIO register for that pin,
which the user can configure.
Figure 5-14 SPI Interface Examples
Table 5-9 SERDES Output and Sample Configuration
CPOL CPHA
00output on falling, sample on rising
01output on rising, sample on falling
10output on rising, sample on falling
11output on falling, sample on rising
SPI
Master
SxCLK
IP2022 or IP2012
(SPI Slave)
515-095b.eps
SxTXP(DO)
SCLK
DI
SxRXD(DI) DO
SxRXP SS
OR
SPI
Slave
SxCLK
IP2022 or IP2012
(SPI Master)
SxTXP(DO)
SCLK
DI
SxRXD(DI) DO
GPIO SS
SCLK
DO
DI
SS
SCLK
DO
DI
SS
IP2012 / IP2022 Data Sheet
www.ubicom.com 69
Figure 5-15 SPI Signal Timing
Table 5-10 SPI Master Interface Signal Usage
SPI Device
Signal Name
IP2012 /
IP2022 SPI
Signal Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2 Pin
Name
(IP2022 only)
Direction Description
SCLK SCLK SxCLK RE0 RF4 Output Serial clock output in mas-
ter mode, input in slave
mode
DO DI SxRXD RE3 RF7 Input Receive data
DI DO SxTXP RE5 RF1 Output Transmit data
SS SS GPIO RE1 RF5 Output Slave select pin used in
slave mode only (Master
select handled by software)
Table 5-11 SPI Slave Signal Usage
SPI Device
Signal Name
IP2012 /
IP2022 SPI
SIgnal Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
Direction Description
SCLK SCLK SxCLK RE0 RF4 Input Serial clock output in mas-
ter mode, input in slave
mode
DO DI SxRXD RE3 RF7 Input Receive data
DI DO SxTXP RE5 RF1 Output Transmit data
SS SS SxRXP RE1 RF5 Input Slave select pin used in
slave mode only (Master
select handled by software)
12345678
SCK CYCLE #
SCK (CPOL = 0)
SCK (CPOL = 1)
SAMPLE INPUT
(CPHA = 0) DATA OUT
SAMPLE INPUT
(CPHA = 1) DATA OUT
SS (TO SLAVE)
MSB 6 5 4321LSB
MSB 654321 LSB
515-098.eps
70 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.6.9 GPSI
Hardware
Figure 5-16 shows example circuits to connect the
SERDES in GPSI (General Purpose Serial Interface)
mode. Table 5-12 shows the GPSI signal to port pin
mapping in Master mode, and Table 5-13 shows the GPSI
signal to port pin mapping in Slave mode.
Software
GPSI is a general-purpose, point-to-point, full-duplex
serial bus protocol. Only two devices are allowed to exist
on a bus. The GPSI PHY device is responsible for
maintaining bus timing by driving two continuously
running clocks, TxClk and RxClk. The device that does not
drive the clocks is the MAC device. The TxEn and TxD
signals are synchronized to the TxClk clock. The RxD and
RxEn signals are synchronized to the RxClk clock.
The COLLISION and TxBUSY signals do not participate
in actual data transfer on the GPSI bus. COLLISION and
TxBUSY provide additional flow control capabilities for the
software device driver. The COLLISION signal is used to
indicate that a PHY device has detected a collision
condition. This signal is only useful when the SERDES is
connected to a PHY device or acting as a PHY device.
The TxBUSY signal is used by a GPSI device to indicate
that the device is currently busy, and that another device
should not attempt to start a data transfer.
Refer to SxRCNT register bit 5. Refer to Section 7.1.10
through Section 7.1.17 for detailed configurations.
Example: To get the PLL clock divided by 2 out on RE6:
S1MODE=63
S1RCFG=80
S1TCFG=80
Figure 5-16 GPSI Interface Examples
515-096c.eps
OR
GPSI
Master
SxCLK
IP2022
or
IP2012
(GPSI Slave) SxTXP
TxCLK
RxD
SxRXD TxD
SxRXP TxEN
SxTXPE RxEN
SxTXM RxCLK
GPIO COL
SxTXME TxBUSY
GPSI
Slave
IP2022
or
IP2012
(GPSI Master)
SxTXP
RxCLK
TxD
SxRXD
RxD
SxRXP
RxEN
SxTXPE
TxEN
SxTXM TxCLK
GPIO COL
GPIO TxBUSY
RxCLK
TxD
RxD
RxEN
TxEN
TxCLK
COL
TxBUSY
RxCLK
RxD
TxD
TxEN
RxEN
TxCLK/RxCLK
COL
TxBUSY
IP2012 / IP2022 Data Sheet
www.ubicom.com 71
Note: In GPSI master mode, the SxTXM SERDES pin
should be used by the GPSI slave for both TxCLK and
RxCLK inputs
Table 5-12 IP2012 / IP2022 GPSI Master Interface Signal Usage
GPSI Slave
Signal Name
IP2012 /
IP2022 GPSI
Signal Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
IP2012 /
IP2022’s
Direction
Description
TxCLK and
RxCLK
TxCLK and
RxCLK
SxTXM RE6 RF2 Output Transmit and
Receive clock
TxD RxD SxRXD RE3 RF7 Input Transmit data
TxEN RxEN SxRXP RE1 RF5 Input Transmit data valid
RxD TxD SxTXP RE5 RF1 Output Receive data
RxEN TxEN SxTXPE RE4 RF0 Output Receive data valid
TxBUSY TxBUSY GPIO - - Output Indicates a data
transfer in progress
(handled by soft-
ware)
COLLISION COLLISION GPIO - - Output Indicates a collision
at PHY layer (han-
dled by software)
Table 5-13 IP2012 / IP2022 GPSI Slave Interface Signal Usage
GPSI Master
Signal Name
IP2012 /
IP2022 GPSI
Signal Name
SERDES
Signal Name
SERDES1
Pin Name
SERDES2
Pin Name
(IP2022 only)
IP2012 /
IP2022’s
Direction
Description
TxCLK RxCLK SxTXM RE6 RF2 Input Transmit clock
RxD TxD SxTXP RE5 RF1 Output Transmit data
RxEN TxEN SxTXPE RE4 RF0 Output Transmit data valid
TxCLK RxCLK SxCLK RE0 RF4 Input Receive clock
TxD RxD SxRXD RE3 RF7 Input Receive data
TxEN RxEN SxRXP RE1 RF5 Input Receive data valid
TxBUSY TxBUSY SxTXME RE7 RF3 Input Indicates a data
transfer in progress
(handled by soft-
ware)
COLLISION COLLISION GPIO - - Input Indicates a collision
at PHY layer (han-
dled by software)
72 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.7 Analog to Digital Converter (ADC)
The on-chip A/D converter has the following features:
• 10-bit ADC (when Vref > 2.3V)
•8 input channels
• 48 kHz maximum sampling rate
• One-shot conversion.
• Optional external reference voltage
• Vmax = AVdd (max 2.7V)
• Result returned in the ADCH and ADCL registers
Figure 5-17 shows the A/D converter circuitry. The ADC
input pins use alternate functions of the Port G pins. The
result of an ADC sample is the analog value measured on
the selected pin. To correctly read an external voltage, the
pin being sampled must be configured as an input in the
port direction register (i.e. the RGDIR register). If the pin
is configured as an output, then the result will indicate the
voltage level being driven by the output buffer. The RG1
and RG2 port pins are also used as the analog
comparator input pins. The result of sampling the RG1 or
RG2 pins will be correct whether or not the comparator is
operating. The RG0 pin is also used as the comparator
output pin. If the comparator is enabled, then sampling the
RG0 pin will indicate the voltage level being driven by the
comparator. The RG3 pin is multiplexed with the external
reference voltage.
Figure 5-17 A/D Converter Block Diagram
5.7.1 ADC Reference Voltage
The reference voltage (Vref) can come from either the
RG3 port pin or from the AVdd supply voltage. If AVdd is
used, the RG3 port pin may be used as a channel of
analog input or as a general-purpose port pin.
Vref defines a voltage level which reads as one increment
of resolution below the full-scale voltage. The full-scale
voltage reads as 0x3FF, so the Vref voltage reads as
0x3FE and the A/D converter resolution is 10 bits. Table
5-14 shows the values reported at the upper and lower
limits of the ADC input voltage range.
5.7.2 A/D Converter Registers
ADCTMR Register
The ADCTMR register (see Section 7.1.2) is used to
specify the number of system clock cycles required for a
delay of 1736 ns, which is used to provide the 1.152MHz
(48 kHz ∞ 24) clock period reference clock for the A/D
converter.
For example, at a system clock frequency of 120 MHz, the
timer register should be set to 53 ((120 MHz/1.152
MHz)/2). The minimum value that may be loaded into the
ADCTMR register is 2, so the system clock must be at
least 24 times the ADC sampling frequency for the ADC to
function.
ADCCFG Register
The A/D converter configuration register (ADCCFG)
provides the control and status bits for the A/D converter,
as shown in Section 7.1.1.
!"
#
$$
! %
&'
(
&)%
!*
+, +,
Table 5-14 ADC Values
Vin Voltage ADC Value
0V 0x000
Vref/0x3FE 0x001
Vref 0x3FE
Vref + (Vref/0x3FE) 0x3FF
IP2012 / IP2022 Data Sheet
www.ubicom.com 73
5.7.3 Using the A/D Converter
The following sequence is recommended:
1. Set the ADCTMR register to the correct value for
the system clock speed.
2. Load the ADCCFG register to specify the channel
and set the ADCGO bit. Setting the ADCGO bit en-
ables and resets the ADC timer.
3. After a period of time (24 timer overflows = 20.8 µs)
the conversion will complete, the ADCGO bit will be
cleared, and the ADC timer will be disabled.
4. A timer-based interrupt service routine can detect or
assume the ADCGO bit has been cleared and read
the ADC value.
5. Another load to the ADCCFG register can then be
used to start another conversion.
5.7.4 ADC Result Justification
The 10 bits of the ADC value can be mapped to the 16 bits
of the ADCH/ADCL register pair in three different ways, as
shown in Table 5-15. In this table, the numbers in the cells
represent bit positions in the 10-bit ADC value, Z
represents zero (as opposed to bit position 0), and -9
represents the inversion of bit position 9.
5.8 Comparator
The IP2012 / IP2022 has an on-chip analog comparator
which uses alternate functions of the RG0, RG1, and RG2
port pins. The RG1 and RG2 pins are the comparator
negative and positive inputs, respectively, while the RG0
pin is the comparator output pin. To use the comparator,
software must program the port direction register
(RGDIR) so that RG1 and RG2 are inputs. RG0 may be
set up as a comparator output pin.
Figure 5-18 Analog Comparator
The comparator enable bits are cleared on reset, which
disables the comparator. To avoid drawing additional
current during power-down mode, the comparator should
be disabled before entering power-down mode. A 50 mV
hysteresis is applied between the inputs, when the
CMPHYS bit is set in the CMPCFG register.
5.8.1 CMPCFG Register
The CMPCFG register is used to enable the comparator,
to read the output of the comparator internally, to enable
the output of the comparator to the comparator output pin,
and to enable the hysteresis. Section 7.1.3 shows the bits
in this register.
Table 5-15 Justification of the ADC Value
Mode ADCH Register Bits ADCL Register Bits
7654321076543210
Left Justified 9876543210ZZZZZZ
Right Justified ZZZZZZ9876543210
Signed -9 -9 -9 -9 -9 -9 -9 8 7 6 5 4 3210
515-017.eps
+
-
CMPEN,
CMPHYS CMPRES CMPOE
RG2
RG0
RG1
74 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.9 Linear Feedback Shift Register (LFSR)
Figure 5-19 LFSR Block Diagram
Four linear feedback shift register (LFSR) units provide
hardware support for the computation-intensive inner
loops of algorithms commonly used in data
communications, such as:
• Cyclic Redundancy Check (CRC)
• Data Scrambling
•Data Whitening
• Encryption/Decryption
• Hashing
The LFSR units implement a programmable architecture,
which can be adapted for algorithms used by the
Bluetooth, Ethernet, Homeplug, HomePNA, HomeRF,
IEEE 802.11, and USB communication protocols. Figure
5-19 is a block diagram of the LFSR architecture.
The 40-bit residue register and its surrounding circuits are
the computational core of an LFSR unit. On every clock
cycle, 39 output bits from the register are available at the
input for performing a shift operation or a polynomial
add/subtract-and-shift operation. Four 40-bit multiplexers
at the output of the residue register allow selecting up to
four terms of the register for feedback into the input (D0),
polynomial operation control (POLY_XOR_EN), and
output (DOUT) bit streams. A fifth multiplexer is only used
for generating the output bit stream.
The polynomial and residue registers are mapped as five
8-bit registers. The mapping of the residue register is
controlled by the ML_OUT bit of the LFSRCFG3 register,
as shown in Figure 5-20.
Figure 5-20 Mapping of the Residue Register
Input data is shifted serially out of the 16-bit DATAIN
register, which can be programmed to provide the data
LSB-first or MSB-first. Output data is shifted serially LSB-
first into the 16-bit DATAOUT register.
Polynomial Register (POLYx)
Residue Register (RESx)
1..39
1..39
0..39
0
0
01
0..38
0..39
FB1
40:1
0..39
FB2
40:1
0..39
FB3
40:1
0..39
DOUT
40:1
D0
Source
Gating
POLY
Source
Gating
DOUT
Source
Gating
16-bit DATAOUT Register
DATAIN Register
515-082b.eps
D0
POLY_XOR_EN
DOUT
DIN
0..39
FB4
40:1
40
39
39
111
1
1
1
1
39
IP2012 / IP2022 Data Sheet
www.ubicom.com 75
A 32-bit RESCMP register (not shown) can be used to
compare the result in the residue register against an
expected value. When ML_OUT is set, residue register
bits 0:31 are compared against RESCMP bits 0:31. When
ML_OUT is clear, residue register bits 39:8 are compared
to RESCMP bits 0:31, respectively. If there are bits in the
residue register which do not participate in the
programmed LFSR operation, be sure that the
corresponding bits in the RESCMP register are initialized
to the same values as these non-participating bits.
Each LFSR unit has three configuration registers
(LFSRCFG1, LFSRCFG2, and LFSRCFG3) for various
control and status bits. The HL_TRIGGER bit in the
LFSRCFG3 register controls whether operation of the
LFSR unit is triggered by a write to the high byte or the low
byte of the DATAIN register (i.e. DATAINH or DATAINL).
The operation then proceeds for some number of cycles
programmed in the SHIFT_COUNT3:0 field of the
LFSRCFG1 register. Completion of the operation is
indicated when the DONE bit in the LFSRCFG1 register is
set. (Alternatively, software can wait 1 cycle/bit of DATAIN
before reading the result.)
An autoloading option is available for each LFSR unit to
load the DATAIN register automatically from the SERDES
RX buffers (SxRBUF register of the corresponding
SERDES unit). LFSR0 and LFSR2 are paired with
SERDES1, and LFSR1 and LFSR3 are paired with
SERDES2.
Three registers in data memory are used to access the
LFSR register banks, as shown in Table 5-16.
The LFSRA register is loaded to point to a specific LFSR
unit and a register pair within the unit. The LFSRA register
has the format shown in Figure 5-21
Figure 5-21 LFSRA Register
Only 0, 1, 2, and 3 are valid as the UNIT, LFSRA bits 7 and
6 = don’t care. The valid encodings for the index are
shown in Table 5-17.
The LFSR registers do not support consecutive read-
modify-write operations. For example, the following
instruction sequence loads unpredictable values:
clrb lfsrh,7
clrb lfsrh,4
5.9.1 LFSRCFG1 Register
Figure 5-22 LFSRCFG1 Register
•SET_RES—set to initialize the residue register to all
ones (write-only, reads as zero).
•DONE—clear while the LFSR is busy, set when the
operation is completed (read only).
•CMP_RES—set if last LFSR operation result
matched contents of RESCMP register (read only).
•SHIFT_COUNT3:0—specifies number of bits to shift,
load with N for an operation of N+1 shifts.
Table 5-16 LFSR Registers in Data Memory
Address Name Description
0x23 LFSRH High data byte
0x27 LFSRL Low data byte
0x2B LFSRA Address register
7430
UNIT3:0 INDEX3:0
Table 5-17 LFSRA Register INDEX Encoding
INDEX3:0 High Byte (LFSRH) Low Byte (LFSRL)
0x0 DATAINH DATAINL
0x1 DATAOUTH DATAOUTL
0x2 FB2 FB1
0x3 LFSRCFG2 RES4
0x4 RES3 RES2
0x5 RES1 RES0
0x6 FB4 FB3
0x7 LFSRCFG3 DOUT
0x8 LFSRCFG1 POLY4
0x9 POLY3 POLY2
0xA POLY1 POLY0
0xB RESCMP3 RESCMP2
0xC RESCMP1 RESCMP0
All registers initialized to 0x00 upon reset, except RESx
= 0xFF, RESCMPx = 0xFF, and LFSRCFG1 = 0x10).
76 5 43 0
Rsvd SET_RES DONE CMP_RES SHIFT_COUNT3:0
76 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.9.2 LFSRCFG2 Register
Figure 5-23 LFSRCFG2 Register
•DOUT_DOUT_EN—set to enable DOUT multiplexer
output in source gating for DOUT node.
•DIN_DOUT_EN—set to enable DIN signal in source
gating for DOUT node.
•FB1_DOUT_EN—set to enable FB1 signal in source
gating for DOUT node.
•FB2_DOUT_EN—set to enable FB2 signal in source
gating for DOUT node.
•DIN_D0_EN—set to enable DIN signal in source gat-
ing for D0 node.
•FB1_D0_EN—set to enable FB1 signal in source gat-
ing for D0 node.
•FB2_D0_EN—set to enable FB2 signal in source gat-
ing for D0 node.
•DATA_IN_POLYXOR_EN—set to enable DIN signal
in source gating for POLY_XOR_EN node.
5.9.3 LFSRCFG3 Register
Figure 5-24 LFSRCFG3 Register
•AUTOLOAD_EN—set to enable autoloading DATAIN
register when SxRBUF register of corresponding
SERDES unit is loaded.
•ML_OUT—set to shift data out of residue register
LSB, and into MSB, clear to shift data out of residue
register MSB, and into LSB. See Figure 5-20 for effect
on RESx mapping.
•ML_IN—set to shift data from DATAIN register MSB-
first to DIN node, clear to shift data LSB-first.
•HL_TRIGGER—set to trigger operation start on load-
ing DATAINH register, clear to trigger on DATAINL.
•FB3_D0_EN—set to enable FB3 signal in source gat-
ing for D0 node.
•FB4_D0_EN—set to enable FB4 signal in source gat-
ing for D0 node.
5.9.4 DATAIN Register
The 8-bit DATAINH and DATAINL registers together
comprise the 16-bit DATAIN register. For LFSR0 and
LFSR2, the AUTOLOAD_EN bit in the LFSRCFG3
register can be used to enable automatic loading from
SERDES1. For LFSR1 and LFSR3, the AUTOLOAD_EN
bit in the LFSRCFG3 register can be used to enable
automatic loading from SERDES2. The HL_TRIGGER bit
in the LFSRCFG3 register controls whether loading the
DATAINH or DATAINL register triggers the start of the
LFSR operation. The ML_IN bit in the LFSRCFG3 register
controls whether data is shifted MSB-first or LSB-first from
the DATAIN register to the DIN node.
5.9.5 DATAOUT Register
The 8-bit DATAOUTH and DATAOUTL registers together
comprise the 16-bit DATAOUT register. Data shifted out of
the residue register is shifted LSB-first into the DATAOUT
register.
5.9.6 DOUT Register
The DOUT register controls a 40:1 multiplexer on the
residue register outputs. It selects a term which can be
used in the source gating for the DOUT bit stream.
5.9.7 FBx Registers
The four FBx registers control four 40:1 multiplexers on
the residue register outputs. They select feedback terms
which can be used in the source gating for the D0,
POLY_XOR_EN, and DOUT bit streams.
5.9.8 POLYx Registers
The five POLYx registers hold the 40-bit polynomial used
in the LFSR operation.
5.9.9 RESx Registers
The five RESx registers hold the 40-bit residue used in the
LFSR operation. The ML_OUT bit controls the mapping of
the residue register to the RESx registers, as shown in
76543210
DOUT_DOUT_EN
DIN_DOUT_EN
FB1_DOUT_EN
FB2_DOUT_EN
DIN_D0_EN
FB1_D0_EN
FB2_D0_EN
DATA_IN_POLYXOR_EN
76543210
Reserved
AUTOLOAD_EN
ML_OUT
ML_IN
HL_TRIGGER
FB3_D0_EN
FB4_D0_EN
IP2012 / IP2022 Data Sheet
www.ubicom.com 77
Figure 5-20. The residue register can be initialized to all
ones by setting the SET_RES bit in the LFSRCFG1
register.
5.9.10 RESCMPx Registers
The four RESCMPx registers hold a 32-bit value for
comparison with the contents of the residue register. After
an LFSR operation is completed, the CMP_RES bit in the
LFSRCFG1 register indicates whether the result of the
operation matched the 32-bit value. When the ML_OUT
bit in the LFSRCFG3 register is clear, bits 39:8 of the
residue register are compared against bits 0:31 of the
RESCMP register. When ML_OUT is set, bits 0:31 of the
residue register are compared against bits 0:31 of
RESCMP.
78 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.9.11 LFSR Configuration
The LFSR units were designed with the following
communication protocols in mind: Bluetooth, Ethernet,
Homeplug, HomePNA, HomeRF, IEEE 802.11, and USB.
Table 5-18 shows the LFSR configurations used to
support these protocols.
Table 5-18 LFSR Configurations for Various Protocols
Protocol Subfunction D0 In Feedback D Out
USB CRC16 Din^D15 Din^D15
CRC5 Din^D4 Din^D4
Ethernet CRC32 Din^D31 Din^D31
Scrambler Din^D17^D22 Din^D17^D22
Descrambler Din Din^D17^D22
HomePlug CRC8 Din^D7 Din^D7
CRC16 Din^D15 Din^D15
Scrambler D6^D3 Din^D6^D3
HomePNA CRC8 Din^D7 Din^D7
CRC16 Din^D15 Din^D15
Scrambler D17^D22 Din^D17^D22
802.11 CRC32 (FCS) Din^D31 Din^D31
CRC16 (HEC) Din^D15 Din^D15
Data
Whitening
D3^D6 Din^D3^D6
CRC16 (CRC) Din^D15 Din^D15
Scrambler Din^D3^D6 Din^D3^D6
Descrambler Din Din^D3^D6
Home-RF CRC Din^D31 Din^D31
Scrambler Din Din^D3^D8
Descrambler Din Din^D3^D8
Bluetooth FEC Din^D4 Din^D4
HEC Din^D7 Din^D7
CRC16 Din^D15 Din^D15
Data
Whitening
D6 D6 Din^D6
Encryption Din^D8^D12^D20^D25 D24
Din^D12^D16^D24^D31 D24
Din^D4^D24^D28^D33 D32
Din^D4^D28^D36^D39 D32
IP2012 / IP2022 Data Sheet
www.ubicom.com 79
5.10 Parallel Slave Peripheral (PSP)
The Parallel Slave Peripheral allows the IP2012 / IP2022
to operate as an 8- or 16-bit slave to an external device,
much like a memory chip. The IP2022 supports either 8-
bit or 16-bit wide bus operation, while the IP2012 is 8-bit
wide bus only. Alternate functions of Port C and Port D are
used for transferring data, and alternate functions of Port
B are used for control signals. Figure 5-25 shows the
connections between an external master and the Parallel
Slave Peripheral interface.
Figure 5-25 Parallel Slave Peripheral
To read or write through the Parallel Slave Peripheral
interface, the external master asserts the chip select (CS)
signal low. This signal is an alternate function of port pin
RB7. The direction of transfer is indicated by the R/W
signal, which is an alternate function of port pin RB6.
When the R/W signal is high, the master is reading from
the slave. When the R/W signal is low, the master is
writing to the slave.
Optionally, a HOLD signal may be enabled as an alternate
function of port pin RB5. Assertion of HOLD indicates to
the external master that the Parallel Slave Peripheral
interface is not ready to allow the data transfer to
complete. The HOLD signal is driven like an open-
collector signal, i.e. low when asserted and high-
impedance when not asserted. When the CS signal is not
asserted (i.e. the IP2012 / IP2022 is not selected), the
HOLD signal is in high-impedance mode. The HOLD
signal should have an external pullup resistor (R1 = 10K
Ω is recommended). The CS signal must not be allowed
to float.
When CS is asserted, an interrupt is generated and
HOLD (if enabled) is automatically asserted. If the data
transfer is a write from the external master, software reads
the Port C, Port D, or both. If the data transfer is a read,
software writes the data to the port or ports. Finally, if
HOLD is asserted, software releases assertion of HOLD
by writing to the PSPRDY bit in the PSPCFG register.
The Parallel Slave Peripheral does not generate interrupts
by itself. Software is required to enable port pin RB7 (the
CS input) as a falling-edge interrupt input for the Parallel
Slave Peripheral to function. The CS signal must go high,
then back low, for each data transfer. RB6 (the R/W input)
must also be configured as an input. The setting in the
RBDIR register for RB5 (the HOLD output) is overridden
by the programming of the Parallel Slave Peripheral.
5.10.1 PSPCFG Register
The PSPCFG register is used to enable the Parallel Slave
Peripheral, select which ports are used for data transfer,
enable the HOLD output, and release the HOLD output
when the data transfer is ready to complete.
•PSPEN2—set to enable Port D for data transfer, clear
to disable. (If this bit is set, the Parallel Slave Periph-
eral overrides the RDDIR register.)
•PSPEN1—set to enable Port C for data transfer, clear
to disable. (If this bit is set, the Parallel Slave Periph-
eral will immediately override the RCDIR register.)
•PSPHEN—set to enable HOLD output, clear to dis-
able. (If this bit is set, the Parallel Slave Peripheral will
immediately override bit 5 of the RBDIR register.)
•PSPRDY—set to release HOLD. This bit always reads
as 0.
•WD—Watchdog time-out bit. Set at reset, if reset was
triggered by Watchdog Timer overflow, otherwise
cleared.
•BO—Brown-out reset bit. Set at reset, if reset was
triggered by brown-out voltage level detection, other-
wise cleared.
External
Master
Data
HOLD
IP2012
Slave
515-033a.eps
R/W
CS RB7
RB6
RB5
IOVDD R1
RC7:0
External
Master
Data
Data
HOLD
IP2022
Slave
R/W
CS RB7
RB6
RB5
IOVDD R1
RD7:0
RC7:0
765 43210
PSPEN2 PSPEN1 PSPHEN PSPRDY Res WD BO
Table 5-19 PSPCFG Register
80 www.ubicom.com
IP2012 / IP2022 Data Sheet
5.11 External Memory Interface (IP2022
only)
Port C and Port D can also be used for a parallel interface
for up to 128K bytes of linear-addressed external memory,
(not program memory) as shown in Figure 5-26. With
additional software-based addressing on I/O, up to 2M
bytes is possible. Port C implements the high address bits,
and Port D is multiplexed between data and the low
address bits. A level-triggered 8-bit latch (TI part number
SN74AC573 or equivalent) is required for demultiplexing.
This latch passes the RD7:0 data when LE is high, and
holds the data when LE is low.
Figure 5-26 External Memory Interface
External memory is accessed as 16-bit words at word-
aligned byte addresses 0x800000 to 0x81FFFE, as
shown in Figure 5-27. External 8-bit memory can only be
accessed through the current ADDRX/ADDRH/ADDRL
pointer using the iread and iwrite instructions.
Programs cannot execute directly out of external memory,
and commands on the ISD/ISP interface cannot directly
access external memory.
Note: In order to use the external memory interface
correctly, RB[4:7], RC[0:7] and RD[0:7] must ALL be
configured as outputs through their respective port-
direction configuration registers.
5.11.1 EMCFG Register (IP2022 only)
Refer to Section 7.1.4 for field definitions and other
information about the EMCFG register.
Note: When external memory is enabled (EMEN = 1), the
RDDIR register value is overridden. PSP function will
need to be disabled. Port B bits 4-7 interrupts need to be
disabled.
Note: Wait one cycle after changing ADDRX bit 7, EMCFG
bit 7, or ADDRSEL before executing an fread,
fwrite, or ferase, or an iread or ireadi.
Figure 5-27 External Memory Map
Software is responsible for inserting a one-instruction
delay between changing the address (i.e. the contents of
the ADDRSEL, ADDRX, ADDRH, or ADDRL registers)
and executing the iread or iwrite instruction, if
required by the timing of the external latch.
A read cycle to external memory has the timing shown in
Figure 5-28. Write cycle timing is shown in Figure 5-29.
The timings shown are recommended for 10ns and 12ns
SRAMs. All external memory cycles are 16-bit transfers,
with the low byte (A0 = 0) followed by the high byte (A0 =
1).
The number of system clocks required for one read or
write access cycle is programmable to meet the SRAM
timing. Figure 5-28 shows a typical SRAM read access
cycle, with EMCFG register set to C9h, that is used to
access 10ns SRAM with a 120 MHz system clock. SRAM
grade reflects SRAM access time (Taa) in nanoseconds
and typically is a last digit of an SRAM part number.
SRAM has several important parameters which should be
taken into account when calculating EMCFG register
settings (refer to Table 5-21).
Table 5-20 shows maximum capacitance allowed on any
SRAM controller signal line vs. a given SRAM grade and
EMCFG setting that provides reliable access.
IP2022
WR
External
Memory
515-087.eps
RD
A0
A8:1
A0
D7:0
LE
Latch
A16:9
RD
WR
Address
Data
Addr
RB6
RB7
RD7:0
RC7:0
RB5
RB4
515-090a.eps
Program RAM
Reserved
0x000000
70
0x004000
Flash Program Memory
0x01FFFE
0x010000
Byte
Address
0x003FFE
0x00FFFE
Reserved
0x020000
External Memory
0x800000
0x7FFFFE
0x81FFFE
IP2012 / IP2022 Data Sheet
www.ubicom.com 81
Note: The formulas and tables above assume that the
address latch chip propagation delay is < Tsys_clk.
Software is responsible for allowing a memory cycle to
complete before reading DATAH/DATAL registers by
inserting instructions as follows:
•Instructions between IREAD and read of DATAH =
EMRDT + 1
•Instructions between IREAD and read of DATAL = 2 *
(EMRDT + 1)
•Instructions between consecutive IREADs (not in-
cluding IREAD access cycles) =2 * ( EMRDT + 1)
Figure 5-28 SRAM Read Cycle (EMBRT=1; EMRDT=2 cycles)
Table 5-20 EMCFG Settings
IP2022
System
CLK
Max Load
Capacitance
on LE, RD,
WR, A0, and
RD7:0(pF)
Tipd
(ns)
Taa
(ns)
EMCFG
Register
Setting
120 10 8 12 C9h
120 30 9 12 C9h
120 50 10 10 C9h
160 10 8 10 D2h
160 30 9 12 D2h
*
**
System Clock is the IP2022’s internal clock, and is shown for reference only.
Setting EMBRT adds one extra Z state cycle on the RD7:0 bus between cycle 7
and cycle 0, and increases the IREAD cycle time from 8 to 9 cycles.
RC7:0
RD7:0
LE (RB6)
System Clock*
515-088f.eps
RD (RB5)
A0 (RB7)
ADDR16:9
ADDR8:1
DATAH DATAL
ADDR16:9
ADDR8:1
IREAD
ADDR Stable
01234567801
1st IREAD 2nd IREAD
th
th
toe
**
taa tohz
82 www.ubicom.com
IP2012 / IP2022 Data Sheet
Figure 5-29 SRAM Write Cycle (EMWRT=2 cycles)
Table 5-21 Timing Requirements for SRAM
IP2022's Timing Requirements Signal Name Units Note
Data Hold Time th ≥0ns
Set EMRDT and EMBRT in the EMCFG Register such that:
SRAM Signal Description Signal Name Units Note
Output Enable to Output Data toe ≤(EMRDT+1)*Tsys_clk - tipd ns 1,2
Address to Output Data taa ≤(EMRDT+1.5)*Tsys_clk - tipd ns 1,2
Output Disable to Output in Hi-Z tohz ≤(EMBRT+1)*Tsys_clk - 6ns ns 1,2,3
Add instruction(s) after any ADDR change and before the IREAD or IWRITE in software, such that:
Latch Signal Description Signal Name
Setup Time of Address Before LE tsu ≤(# of instructions added + 1) * Tsys_clk - 3ns
Notes:
1. IP2022's RD is connected to the external SRAM's OE pin.
2. Tsys_clk is the period of the IP2022 system clock, and tipd is the propagation delay internally in the IP2022, of
the SRAM control signals (see Table 5-20).
3. The IP2022 should not output on RD7:0 before the SRAM tri-states this bus (set EMBRT = 1 if bus contention
can exist with EMBRT = 0).
* System Clock is the IP2022’s internal clock, and is
shown for reference only.
RC7:0
RD7:0
LE (RB6)
System Clock*
515-089f.eps
WR (RB4)
A0 (RB7)
ADDR16:9
ADDR8:1
DATAH DATAL
ADDR16:9
ADDR8:1
IWRITE
ADDR Stable
0123
456701
DATAH
IP2012 / IP2022 Data Sheet
www.ubicom.com 83
6.0 In-System Programming
The IP2000 series devices provide a dedicated serial
interface for in-system programming (ISP) of the flash
program memory and configuration block. ISP allows
designers to incorporate a small connector which can be
used to interface to a device programmer for programming
or reprogramming the part after it has been soldered to a
circuit board.
The interface used for in-system programming (ISP) and
in-system debugging (ISD) is compatible with the SPI
serial interface protocol. Whenever possible, a standard
connector should be incorporated in the system design for
in-system debugging and programming. The
recommended connector layout for the ISD/ISP interface
is shown in Figure 6-1. The connector is a male 10-pin
connector with 100-mil pin spacing, whose pin
assignments are listed in Table 6-1. The connector is
keyed to prevent backward insertion.
Figure 6-1 ISD/ISP Connector
Signal levels on the connector are LVTTL-compatible. The
target system provides the TSCK, TSI, TRST
, and TSS
signals with 10K ohm pullup resistors.
For more information about the ISD/ISP interface and the
interaction between the debugger/programmer and the
target system, see the IP2000 Series User’s Manual.
Table 6-1 Connector Pin Assignments
Pin Name Description
1 KEY Key (not a signal)
2 TSS Target Slave Select—Active-low sig-
nal which enables the IP2012 / IP2022
to communicate on the SPI bus. Con-
nect to pin 1 on the IP2012 / IP2022.
3 GND Ground
1
515-053.eps
3
5
7
9
2
4
6
8
10
4 TSCK Target Data Clock—Serial clock. Con-
nect to pin 2 on the IP2012 / IP2022.
5 OSC Target Clock Oscillator—If the debug-
ger/programmer is capable of supply-
ing an OSC clock for the target
system, then this clock must be con-
figurable so that it can be disabled to
prevent it from interfering with the tar-
get system (i.e. the OSC clock output
is placed in a high-impedance state).
6 Reserved Reserved
7 TRST Target Reset—The target system may
use the TRST signal to reset the entire
system, to reset only the IP2012 /
IP2022, or it may ignore the TRST sig-
nal. The debugger/programmer may
provide a 100-ms system reset signal
(TRST) to the target system. If sup-
ported, the TRST output must be an
open-collector driver to accommo-
date other sources of reset in the tar-
get system. The minimum source
requirement for this driver is 6 mA.
The debugger/programmer should not
detect or be reset by the TRST signal
being driven low by the target system.
There is no requirement that the part
be connected to the TRST signal, so
the debugger/programmer cannot
assume that the part has been reset if
the target system pulls the TRST pin
low.
8 TSI Target Serial Input—Sampled on the
rising edge of TSCK. Connect to pin 3
on the IP2012 / IP2022.
9 VDD Power. 2.3 - 3.6V (optional)
10 TSO Target Serial Output—Driven by the
IP2012 / IP2022 after the falling edge
of TSCK. Connect to pin 4 on the part.
The IP2012 / IP2022 drives this pin
only if TSS is held low (TSO is
tristated otherwise). The TSO pin is
driven low if TSS is driven low while
the part is in reset; TSO will be driven
high as soon as the part is out of
reset.
Table 6-1 Connector Pin Assignments (continued)
Pin Name Description
84 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.0 Memory Reference
7.0.1 Registers (sorted by address)
Table 7-1 shows the addresses and reset values of all special-purpose registers in data memory, sorted by their address.
Table 7-1 Register Addresses and Reset State
Address Name Description
Register Status Follow-
ing Reset (Power-On,
RST, Brown-Out RST,
Watchdog RST)
0x001 Reserved Reserved Reserved
0x002 ADDRSEL Selector for current external(IP2022 only)/program
memory ADDRX/ADDRH/ADDRL
0000 0000
0x003 ADDRX External(IP2022 only)/program memory pointer
(bits 23:16)
0000 0000
0x004 IPH Indirect Data RAM Pointer (high byte) 0000 0000
0x005 IPL Indirect Data RAM Pointer (low byte, see Section
4.1)
0000 0000
0x006 SPH Data RAM Stack Pointer (high byte) 0000 0000
0x007 SPL Data RAM Stack Pointer (low byte, see Section
4.1)
0000 0000
0x008 PCH Current PC (program counter) bits 15:8 (read-only) 1111 1111
0x009 PCL Virtual register for direct PC modification 1111 0000
0x00A WREG W (working) register 0000 0000
0x00B STATUS STATUS register 1110 0000
0x00C DPH Data Pointer (high byte) 0000 0000
0x00D DPL Data Pointer (low byte, see Section 4.1) 0000 0000
0x00E SPDREG Current speed (read-only, see Section 3.5) 1001 0011
0x00F MULH Multiply result (high byte) 0000 0000
0x010 ADDRH External(IP2022 only)/program memory address
(bits 15:8)
0000 0000
0x011 ADDRL External(IP2022 only)/program memory address
(bits 7:0, see Section 5.11)
0000 0000
0x012 DATAH External(IP2022 only)/program memory data (high
byte)
0000 0000
0x013 DATAL External(IP2022 only)/program memory data (low
byte)
0000 0000
0x014 INTVECH Interrupt vector (high byte) 0000 0000
0x015 INTVECL Interrupt vector (low byte) 0000 0000
0x016 INTSPD Interrupt speed register 0000 0000
0x017 INTF Port B interrupt flags Undefined
0x018 INTE Port B interrupt enable bits 0000 0000
0x019 INTED Port B interrupt edge select bits 0000 0000
IP2012 / IP2022 Data Sheet
www.ubicom.com 85
0x01A FCFG Flash configuration register 0000 0000
0x01B TCTRL Timer 1/2 common control register 0000 0000
0x01C XCFG Extended configuration (bit 0 is read-only) 0000 000x
(See Section 7.1.26 for
FBUSY))
0x01D EMCFG External memory configuration register (IP2022
only)
0000 0000
0x01E IPCH Interrupt return address (high byte) 0000 0000
0x01F IPCL Interrupt return address (low byte) 0000 0000
0x020 RAIN Data on Port A pins N/A
0x021 RAOUT Port A output latch 0000 0000
0x022 RADIR Port A direction register 1111 1111
0x023 LFSRH LFSR data register (high byte) 0000 0000
0x024 RBIN Data on Port B pins N/A
0x025 RBOUT Port B output latch 0000 0000
0x026 RBDIR Port B direction register 1111 1111
0x027 LFSRL LFSR data register (low byte) 0000 0000
0x028 RCIN Data on Port C pins N/A
0x029 RCOUT Port C output latch 0000 0000
0x02A RCDIR Port C direction register 1111 1111
0x02B LFSRA LFSR address register 0000 0000
0x02C RDIN Data on Port D pins N/A
0x02D RDOUT Port D output latch 0000 0000
0x02E RDDIR Port D direction register 1111 1111
0x02F Reserved Reserved Reserved
0x030 REIN Data on Port E pins N/A
0x031 REOUT Port E output latch 0000 0000
0x032 REDIR Port E direction register 1111 1111
0x033 Reserved Reserved Reserved
0x034 RFIN Data on Port F pins N/A
0x035 RFOUT Port F output latch 0000 0000
0x036 RFDIR Port F direction register 1111 1111
0x037 Reserved Reserved Reserved
0x038 Reserved Reserved Reserved
0x039 RGOUT Port G output latch 0000 0000
0x03A RGDIR Port G direction register 1111 1111
Table 7-1 Register Addresses and Reset State (continued)
Address Name Description
Register Status Follow-
ing Reset (Power-On,
RST, Brown-Out RST,
Watchdog RST)
86 www.ubicom.com
IP2012 / IP2022 Data Sheet
0x03B Reserved Reserved Reserved
0x03C Reserved Reserved Reserved
0x03D Reserved Reserved Reserved
0x03E Reserved Reserved Reserved
0x03F Reserved Reserved Reserved
0x040 RTTMR Real-time timer value 0000 0000
0x041 RTCFG Real-time timer configuration register 0000 0000
0x042 T0TMR Timer 0 value 0000 0000
0x043 T0CFG Timer 0 configuration register 0000 0000
0x044 T1CNTH Timer 1 counter register high (read only) 0000 0000
0x045 T1CNTL Timer 1 counter register low (read only) 0000 0000
0x046 T1CAP1H Timer 1 Capture 1 register high (read only) 0000 0000
0x047 T1CAP1L Timer 1 Capture 1 register low (read only) 0000 0000
0x048 T1CAP2H/T1CMP2H Timer 1 Capture 2/Compare 2 register high 0000 0000
0x049 T1CAP2L/T1CMP2L Timer 1 Capture 2/Compare 2 register low 0000 0000
0x04A T1CMP1H Timer 1 Compare 1 register high 0000 0000
0x04B T1CMP1L Timer 1 Compare 1 register low 0000 0000
0x04C T1CFG1H Timer 1 configuration register 1 high 0000 0000
0x04D T1CFG1L Timer 1 configuration register 1 low 0000 0000
0x04E T1CFG2H Timer 1 configuration register 2 high 0000 0000
0x04F T1CFG2L Timer 1 configuration register 2 low 0000 0000
0x050 ADCH ADC value (high) (read only) 0000 0000
0x051 ADCL ADC value (low) (read only) 0000 0000
0x052 ADCCFG ADC configuration register 0000 0000
0x053 ADCTMR ADC timer register 0000 0000
0x054 T2CNTH Timer 2 counter register high (read only) 0000 0000
0x055 T2CNTL Timer 2 counter register low (read only) 0000 0000
0x056 T2CAP1H Timer 2 Capture 1 register high (read only) 0000 0000
0x057 T2CAP1L Timer 2 Capture 1 register low (read only) 0000 0000
0x058 T2CAP2H/T2CMP2H Timer 2 Capture 2/Compare 2 register high 0000 0000
0x059 T2CAP2L/T2CMP2L Timer 2 Capture 2/Compare 2 register low 0000 0000
0x05A T2CMP1H Timer 2 Compare 1 register high 0000 0000
0x05B T2CMP1L Timer 2 Compare 1 register low 0000 0000
0x05C T2CFG1H Timer 2 configuration register 1 high 0000 0000
0x05D T2CFG1L Timer 2 configuration register 1 low 0000 0000
Table 7-1 Register Addresses and Reset State (continued)
Address Name Description
Register Status Follow-
ing Reset (Power-On,
RST, Brown-Out RST,
Watchdog RST)
IP2012 / IP2022 Data Sheet
www.ubicom.com 87
0x05E T2CFG2H Timer 2 configuration register 2 high 0000 0000
0x05F T2CFG2L Timer 2 configuration register 2 low 0000 0000
0x060 S1TMRH SERDES 1 clock timer register (high bits) 0000 0000
0x061 S1TMRL SERDES 1 clock timer register (low bits) 0000 0000
0x062 S1TBUFH SERDES 1 transmit buffer (high bits) Undefined
0x063 S1TBUFL SERDES 1 transmit buffer (low bits) Undefined
0x064 S1TCFG SERDES 1 transmit configuration 0000 0000
0x065 S1RCNT SERDES 1 received bit count (actual) (read-only) 0000 0000
0x066 S1RBUFH SERDES 1 receive buffer (high bits) (read-only) Undefined
0x067 S1RBUFL SERDES 1 receive buffer (low bits) (read-only) Undefined
0x068 S1RCFG SERDES 1 receive configuration 0000 0000
0x069 S1RSYNC SERDES 1 receive bit sync pattern 0000 0000
0x06A S1INTF SERDES 1 status/Interrupt flags 0000 0000
0x06B S1INTE SERDES 1 Interrupt enable bits 0000 0000
0x06C S1MODE SERDES 1 serial mode/clock select register 0000 0000
0x06D S1SMASK SERDES 1 receive sync mask 0000 0000
0x06E PSPCFG Parallel slave peripheral configuration register 0000 00xx
(See Section 7.1.8 for BO,
WD)
0x06F CMPCFG Comparator configuration register 0000 000X
(See Section 7.1.3)
0x070 S2TMRH SERDES 2 clock timer register (high bits) (IP2022
only)
0000 0000
0x071 S2TMRL SERDES 2 clock timer register (low bits) (IP2022
only)
0000 0000
0x072 S2TBUFH SERDES 2 transmit buffer (high bits) (IP2022 only) Undefined
0x073 S2TBUFL SERDES 2 transmit buffer (low bits) (IP2022 only) Undefined
0x074 S2TCFG SERDES 2 transmit configuration (IP2022 only) 0000 0000
0x075 S2RCNT SERDES 2 received bit count (actual) (read-only)
(IP2022 only)
0000 0000
0x076 S2RBUFH SERDES 2 receive buffer (high bits) (read-only)
(IP2022 only)
Undefined
0x077 S2RBUFL SERDES 2 receive buffer (low bits) (read-only)
(IP2022 only)
Undefined
0x078 S2RCFG SERDES 2 receive configuration (IP2022 only) 0000 0000
0x079 S2RSYNC SERDES 2 receive bit sync pattern (IP2022 only) 0000 0000
0x07A S2INTF SERDES 2 status/Interrupt flags (IP2022 only) 0000 0000
0x07B S2INTE SERDES 2 interrupt enable bits (IP2022 only) 0000 0000
Table 7-1 Register Addresses and Reset State (continued)
Address Name Description
Register Status Follow-
ing Reset (Power-On,
RST, Brown-Out RST,
Watchdog RST)
88 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.0.2 Program Memory
Table 7-2 shows the addresses and reset values of all program memory.
0x07C S2MODE SERDES 2 serial mode/clock select register
(IP2022 only)
0000 0000
0x07D S2SMASK SERDES 2 receive sync mask (IP2022 only) 0000 0000
0x07E CALLH Top of call stack (high 8 bits) 1111 1111
0x07F CALLL Top of call stack (low 8 bits) 1111 1111
0x080 to
0x0FF
Directly addressable general-purpose (global) reg-
isters
Undefined after power-on
or brown-out reset,
unchanged after RST or
Watchdog Timer reset
0x100 to
0xFFF
Data memory RAM Undefined after power-on
or brown-out reset,
unchanged after RST or
Watchdog Timer reset
Table 7-1 Register Addresses and Reset State (continued)
Address Name Description
Register Status Follow-
ing Reset (Power-On,
RST, Brown-Out RST,
Watchdog RST)
Table 7-2 Program Memory Addresses
Address Description
Status Following Reset (Power-On,
RST, Brown-Out RST, Watchdog
RST)
0x0000 to 0x1FFF
(word addresses)
Program Memory RAM Undefined after power-on or brown-
out reset, unchanged after RST or
Watchdog Timer reset
0x8000 to 0xFFFF
(word addresses)
Program Memory Flash. Unchanged after power-on, brown-out
reset, RST or Watchdog Timer reset
(changes only during ISP program-
ming, and during flash self-program-
ming)
0x10000 to 0x1003F
(word addresses)
Flash Configuration Block (see Section 3.10).
Flash is factory programmed to:
Unchanged after power-on, brown-out
reset, RST or Watchdog Timer reset
(changes only during ISP program-
ming, and during flash self-program-
ming)
Word Address
$ 10000
$ 10001
$ 10004
FUSE0
FUSE1
TRIM0
= 1000
= FFF7
= FBFE
IP2012 / IP2022 Data Sheet
www.ubicom.com 89
7.1 Register Bit Definitions
For those registers which have special functions assigned
to bits or fields within the register, the definition of those
bits and fields is described below. The registers are
presented alphabetically.
7.1.1 ADCCFG Register
A/D converter configuration.
7.1.2 ADCTMR Register
The ADCTMR register is used to specify the number of
system clock cycles required for a delay of 1736 ns, which
is used to provide the 1.152MHz (48 kHz ∞ 24) clock
period reference clock for the A/D converter.
7.1.3 CMPCFG Register
Comparator configuration.
765432 0
ADCREF ADCJST Rsrvd. ADCGO ADCS2:0
Name Description
ADCREF A/D converter reference voltage select
0 = AVdd is the reference voltage
1 = RG3 port pin is used to receive
an external reference voltage
ADCJST A/D converter result justification mode
select
00 = Right justified
01 = Signed
10 = Left justified
11 = Reserved
ADCGO A/D converter GO/DONE bit
0 = When the last conversion has
completed, this bit reads as 0.
1 = Write 1 to begin a new conver-
sion. While the conversion is in
progress, this bit reads as 1.
ADCS2:0 A/D converter input channel select
000 = Port pin RG0
001 = Port pin RG1
010 = Port pin RG2
011 = Port pin RG3
100 = Port pin RG4
101 = Port pin RG5
110 = Port pin RG6
111 = Port pin RG7
76543210
CMPEN
CMPOE
CMPHYS
Reserved
CMPRES
Name Description
CMPEN Comparator enable bit
0 = Comparator disabled
1 = Comparator enabled
CMPOE Comparator output enable bit
0 = Comparator output disabled.
1 = Comparator output enabled on
port pin RG0.
CMPHYS Comparator hysteresis enable bit
0 = Hysteresis disabled
1 = Hysteresis enabled
CMPRES Comparator result (read-only)
0 = RG2 voltage > RG1
1 = RG1 voltage > RG2
90 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.4 EMCFG Register (IP2022 only)
External memory interface configuration. (Reserved in
IP2012; value 0x00.)
765 320
EMEN EMBRT EMWRT2:0 EMRDT2:0
Name Description
EMEN Enable external memory interface
0 = Port C, Port D and RB7:4 avail-
able for general-purpose I/O
1 = Port C, Port D and RB7:4 used
for external memory interface
EMBRT Enable bus release wait state
0 = No wait state
One wait state cycle added to
IREAD and IWRITE after read of
DATAL and before ADDR8:1 is
put on RD7:0 bus
1 =
EMWRT2:0 WR pulse width, in system clock cycles
000 = 1 100 = 5
001 = 2 101 = 6
010 = 3 110 = 7
011 = 4 111 = 8
EMRDT2:0 RD pulse width, in system clock cycles
000 = 1 100 = 5
001 = 2 101 = 6
010 = 3 110 = 7
011 = 4 111 = 8
IP2012 / IP2022 Data Sheet
www.ubicom.com 91
7.1.5 FCFG Register
Flash configuration.
76543210
FRDTS1:0 FRDTC1:0 FWRT3:0
Name Description
FRDTS1:0 The core clock frequency is automatically reduced (if necessary) when executing out of flash memory to
prevent the flash memory access time from being too short. The FRDTS1:0 bits specify the minimum
number of system clock cycles required for instruction execution from flash memory. The actual execu-
tion speed from flash memory will be the slower of the speed indicated in the SPDREG register and the
speed specified by the FRDTS1:0 bits. The 11 setting can always be used, but it may cause slower
flash operation than necessary.
Set to this:
If System Clock Frequency (MHz) is System Clock Cycles For
Each Flash Instruction
Cycle120 MHz Part 160 MHz Part
00 0–40 0–53.3 1 cycle
01 40–80 53.3–106.7 2 cycles
10 80–120 106.7–160 3 cycles
11 reserved reserved 4 cycles
Note Flash instruction execution = 25ns minimum (18.75ns for 160MHz part)
FRDTC1:0 To prevent the flash memory access time from being too short, it is necessary to specify the number of
CPU core cycles between reading the flash memory using an fread instruction (or an iread or
ireadi instruction while executing from RAM to read flash) and the time that DATAH and DATAL are
written by the IP2012 / IP2022. Because the CPU core is subject to changes in speed, the value pro-
grammed in these bits should be appropriate for the fastest speed that might be used (typically, the
faster of the main line code and the interrupt service routine). The FRDTC1:0 bits specify the number of
CPU core clock cycles required for flash read access. Even with this configured, enough cycles must be
executed after the flash read instruction and before reading DATAH or DATAL to account for the mini-
mum flash access time.
Set to this: If CPU Core Frequency (MHz) is Core Clock Cycles For
Each Flash Read Cycle
120 MHz Part 160 MHz Part
00 0–40 0–53.3 1 cycle
01 40–80 53.3–106.7 2 cycles
10 80–120 106.7–160 3 cycles
11 reserved reserved 4 cycles
Note fread/iread/ireadi of flash= 25ns minimum (18.75ns for 160MHz part)
FWRT3:0 The flash memory ferase, fwrite and ISP flash write, flash block erase, or flash bulk erase timing
is derived from the CPU core clock through a programmable divider. The FWRT3:0 bits specify the divi-
sor. The time base must be 1 to 2 microseconds. Below 1 microsecond, the flash memory will be under-
programmed, and data retention is not guaranteed. Above 2 microseconds, the flash memory will be
overprogrammed, and reliability is not guaranteed.
Set to this: If CPU Core Frequency is: FWRT Frequency Divisor
0000 1–2 MHz 2
0001 2–3 MHz 3
0010 3–4 MHz 4
0011 4–6 MHz 6
92 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.6 INTSPD Register
Configuration of clock and PLL settings to be used during
an interrupt service routine. INTSPD is copied to
SPDREG when an interrupt occurs. See Table 3-5 for
reti options.
0100 6–8 MHz 8
0101 8–12 MHz 12
0110 12–16 MHz 16
0111 16–24 MHz 24
1000 24–32 MHz 32
1001 32–48 MHz 48
1010 48–64 MHz 64
1011 64–96 MHz 96
1100 96–128 MHz 128
1101 128–160 MHz 192
1110 Reserved 256
1111 Reserved 384
Note: If FCFG & OSC1 are optimal:
fwrite = 42us
ferase = 20ms, because FPERT in TRIM0 should be 0
Name Description
76543 0
PLL OSC CLK1:0 CDIV3:0
Name Description
PLL Run-time control of PLL clock multiplier oper-
ation. If the PLL is not required, power con-
sumption can be reduced by disabling it.
0 = PLL clock multiplier enabled
1 = PLL clock multiplier disabled
OSC Run-time control of OSC oscillator operation.
If the crystal oscillator is not required, power
consumption can be reduced by disabling it
(stops OSC oscillator and blocks propaga-
tion of OSC1 external clock input).
0 = OSC oscillator enabled
1 = OSC oscillator disabled
CLK1:0 Selects the system clock source.
00 = PLL clock multiplier. Do not use if
an interrupt can awaken the part
from sleep (can use a speed
instruction in the ISR instead).
01 = OSC oscillator/external clock on
OSC1 input
10 = RTCLK oscillator/external clock
on RTCLK1 input
11 = System clock disabled (off)
Note: If the OSC crystal driver is stopped
(SPDREG bit 6 = 1) and Port B or Real Time
Timer interrupts are enabled, then INTSPD
bits 5 and 4 must not both be 0, because the
crystal startup time plus PLL startup time
may be greater than WUDP2:0 (see Figure
3-16).
CDIV3:0 Selects the system clock divisor.
0000 = 1 1000 = 12
0001 = 2 1001 = 16
0010 = 3 1010 = 24
0011 = 4 1011 = 32
0100 = 5 1100 = 48
0101 = 6 1101 = 64
0110 = 8 1110 = 128
0111 = 10 1111 = System clock
disabled (off)
Name Description
IP2012 / IP2022 Data Sheet
www.ubicom.com 93
7.1.7 LFSRA Register
Linear Feedback Shift Register configuration.
7.1.8 PSPCFG Register
Parallel Slave Peripheral configuration.
7430
UNIT3:0 INDEX3:0
Name Description
UNIT3:0 LFSR unit number (only 0, 1, 2, and 3 are
valid)
INDEX3:0 Index to the LFSR register being
accessed (see Table 5-17)
765 4310
PSPEN2 PSPEN1 PSPHEN PSPRDY Res WD BO
Name Description
PSPEN2 Port D enable bit (IP2022 only)
0 = Port D is available for general-
purpose I/O
1 = Port D is configured for the
Parallel Slave Peripheral inter-
face
PSPEN1 Port C enable bit
0 = Port C is available for general-
purpose I/O
1 = Port C is configured for the
Parallel Slave Peripheral inter-
face
PSPHEN HOLD output enable bit
0 = HOLD output disabled. Port pin
RB5 available for general-pur-
pose I/O.
1 = HOLD output enabled on port
pin RB5.
PSPRDY Ready bit
0 = This bit always reads as zero.
1 = Write 1 to release HOLD when
the IP2012 / IP2022 is ready to
allow the data transfer to com-
plete.
WD Watchdog time-out bit. If using the Watch-
dog feature, set this bit before the first
cwdt instruction. Then this bit is not
cleared by a Watchdog reset, but is
cleared by all other reset sources. How-
ever, if WUDX in FUSE0 is more than
70ms longer than the watchdog timeout
period in FUSE1, a Power-On Reset or a
Brown-Out Reset may set this bit. (Do not
use this bit if WUDX is more than 70ms
longer than the watchdog timeout period).
BO Brown-out reset bit. Set at reset, if reset
was triggered by brown-out voltage level
detection, otherwise cleared
94 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.9 RTCFG Register
Real-Time Timer configuration.
76 3210
RTEN RTPS3:0 RTSS RTIE RTIF
Name Description
RTEN Real-Time Timer enable bit
0 = Real-Time Timer disabled
0 = Real-Time Timer enabled
RTPS3:0 Real-Time Timer prescaler divisor
0000 = 1 1000 = 256
0001 = 2 1001 = 512
0010 = 4 1010 = 1024
0011 = 8 1011 = 2048
0100 = 16 1100 = 4096
0101 = 32 1101 = 8192
0110 = 64 1110 = 16384
0111 = 128 1111 = 32768
RTSS Real-Time Timer clock source select
0 = external OSC clock
1 = external RTCLK clock
RTIE Real-Time Timer interrupt enable bit
0 = Real-Time Timer interrupt dis-
abled
1 = Real-Time Timer interrupt
enabled
RTIF Real-Time Timer interrupt flag
0 = No timer overflow has occurred
since this bit was last cleared
1 = Timer overflow has occurred.
This bit goes high two cycles
after the actual overflow
occurs.
IP2012 / IP2022 Data Sheet
www.ubicom.com 95
7.1.10 SxINTE/SxINTF Register
Indicates the SERDES conditions that may be enabled as
interrupts.
The SxINTE register has the same format as the SxINTF
register. For each condition indicated by a flag in the
SxINTF register, setting the corresponding bit in the
SxINTE register enables the interrupt for that condition.
76543210
RXERROR
RXEOP
SYND
TXBE
TXEOP
SXLINKPULSE
RXBF
RXXCRS
Name Description
RXERROR Receive error interrupt flag
10Base-T mode: Manchester encod-
ing data phase error
USB mode: bit unstuffing error
(1111111 received)
0 = Receive error has not been
detected since this bit was last
cleared
1 = Receive error has been
detected
RXEOP End-of-Packet detection interrupt flag
10Base-T and USB modes: end-of-
packet detected
GPSI mode: RxEN deasserted
SPI mode: Set on rising edge.
0 = End-of-Packet has not been
detected since this bit was last
cleared
1 = End-of-Packet has been
detected
SYND Synchronization pattern detection
interrupt flag (10Base-T and USB
modes only)
0 = Synchronization pattern has not
been detected since this bit was
last cleared
1 = Synchronization pattern has
been detected
TXBE Transmit buffer empty interrupt flag
0 = Transmit buffer has not been
empty since this bit was last
cleared
1 = Transmit buffer has been empty
TXEOP Transmit underrun. This bit is set
when the previous data in the transmit
buffer register (SxTXBUF) has been
transmitted and no new data has been
loaded in the register. In USB and
10Base-T modes, this causes an EOP
condition to be generated.
0 = Transmit underrun has not
occurred since this bit was last
cleared
1 = Transmit underrun has occurred
SXLINKPULSE Set after a link pulse of 60 to 200 ns
duration is detected in Ethernet mode.
Also known as TxIdle in USB mode.
0 = 10Base-T mode: No link pulse
has been detected since this bit
was last cleared
USB mode: SERDES is trans-
mitting
1 = 10Base-T mode: Link pulse
detected
USB mode: SERDES is not
transmitting
RXBF Receive buffer full interrupt flag
0 = Receive buffer has not been full
since this bit was last cleared
1 = Receive buffer has been full
RXXCRS Set while the carrier is sensed.
10bT mode: clear to 0
0 = USB mode: RxBUSY is
detected - SERDES is receiving
1 = USB mode: RxBUSY is not
detected - SERDES is not
receiving
Name Description
96 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.11 SxMODE Register
SERDES protocol mode configuration.
7.1.12 SxRCFG Register
SERDES RX shift count, USB sync detect and data
polarity configuration.
76543210
REV PRS2:0 SUBM1:0 CLKS1:0
Name Description
REV Read-only as “1” in latest revision. Refer to
IP2022 silicon errata sheet.
PRS2:0 Protocol select (see Table 5-3). All other
encodings are reserved.
000 = Disabled
001 = 10Base-T
010 = USB Bus
011 = UART
101 = SPI
110 = GPSI
SUBM1:0 Submode select
USB mode:
01 = Low-speed USB interface
10 = High-speed USB interface
SPI mode:
00 = Positive clock polarity, receive
on rising edge, transmit on fall-
ing edge
01 = Positive clock polarity, receive
on falling edge, transmit on ris-
ing edge
10 = Negative clock polarity, receive
on falling edge, transmit on ris-
ing edge
11 = Negative clock polarity, receive
on rising edge, transmit on fall-
ing edge
GPSI mode:
00 = Receive on rising edge, trans-
mit on falling edge
01 = Receive on falling edge, trans-
mit on falling edge
10 = Receive on rising edge, trans-
mit on rising edge
11 = Receive on falling edge, trans-
mit on rising edge
CLKS1:0 Clock source select (see Figure 5-8).
00 = Clock disabled
01 = SxCLK input
10 = OSC clock oscillator
11 = post-PLL clock
Note: When switching CLKS1:0 to 10, a
delay is needed before reliable writes to
SERDES registers can be made. The
required delay, in number of core instruc-
tions, is (core clock frequency / new SER-
DES clock frequency) x 2.
7654 0
MASSEL SYNCDETEN RPOREV RXSCNT4:0
Name Description
MASSEL 10Base-T mode:
0 = Normal polarity detected
1 = Reverse polarity detected
GPSI or SPI mode:
0 = Slave mode
1 = Master mode
SYNCDETEN Synchronization byte detection enable
(USB mode only)
0 = Synchronization byte detec-
tion enabled
1 = Synchronization byte detec-
tion disabled
RPOREV Receive data polarity reversal select
0 = Data polarity uninverted
1 = Data polarity inverted
RXSCNT4:0 Receive shift count, specifies number
of bits to receive
Name Description
IP2012 / IP2022 Data Sheet
www.ubicom.com 97
7.1.13 SxRCNT Register
SERDES RX activity configuration.
7.1.14 SxRSYNC Register
SERDES sync pattern configuration.
7654 0
BITORDER RxCRSED RxCRS RXACNT4:0
Name Description
BITORDER Bit order for transmit and receive
0 = LSB first
1 = MSB first
RxCRSED Not used for 10baseT mode. For
GPSI Slave mode:
0 = Disable TxBUSY input
1 = Enable TxBUSY input
RxCRS Carrier Sense Status: Current state
of carrier
RXACNT4:0 Receive shift count, actual number
of bits received (read-only). Excep-
tions occur during the last transfer:
RXACNT = 0 if bit count is
less than 8
RXACNT = 8 if bit count is
greater than or equal to 8, but
less than 16
RXACNT = 16 if bit count is
greater than or equal to 16
and the RXSCNT4:0 field in
the SxRCFG register is 16
7210
SYNCPAT7:2 SQUELCHEN DRIBBITEN
Name Description
SYNCPAT7:2 Synchronization pattern, bits 7:2 (USB
mode only)
SQUELCHEN USB mode: synchronization pattern, bit
1
10Base-T mode:
0 = Squelch disabled
1 = Squelch enabled
All other modes: 0
DRIBBITEN USB mode: synchronization pattern, bit
0
10Base-T mode:
0 = Hardware handles dribble bit
1 = Software is responsible for han-
dling dribble bit
98 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.15 SxSMASK Register
SERDES sync-pattern configuration.
10Base-T mode:
USB mode:
7.1.16 SxTCFG Register
SERDES TX shift count configuration.
76 3210
Resrvd. PREAMCNT3:0 Resrvd. CONTPAIR Resrvd.
70
MASK7:0
Name Description
PREAMCNT3:0 Preamble pair count (10Base-T
mode only). All other encodings are
reserved.
0000 = 24 pairs
0001 = 20 pairs
0010 = 16 pairs
0011 = 12 pairs
0100 = 8 pairs
0101 = 4 pairs
CONTPAIR Configures the detection of consecu-
tive pairs of “10” for sync detection.
0 = Sync detected if 6 "10"
pairs + "11”.
1 = Use PREAMCNT for num-
ber of "10" pairs
MASK7:0 Mask bits for SxRSYNC (USB mode
only)
0 = Ignore corresponding bit in
SxRSYNC
1 = Use corresponding bit in
search pattern for synchro-
nization byte
7654 0
GLOBEN LPBACK TPOREV TXSCNT4:0
Name Description
GLOBEN Global enable bit
0 = Disable SERDES output
1 = Enable SERDES output (must
use for TX). If enabling
SERDES1, the REOUT port data
are overridden by SERDES1 out-
puts. If enabling SERDES2
(IP2022 only), the RFOUT port
data are overridden by SERDES2
outputs.
LPBACK Loopback enable bit
0 = Normal operation
1 = Output is driven into input
TPOREV Transmit data polarity reversal select
(UART mode only)
0 = Data polarity uninverted
1 = Data polarity inverted
TXSCNT4:0 Transmit shift count, specifies number of
bits to transmit
IP2012 / IP2022 Data Sheet
www.ubicom.com 99
7.1.17 SxTMRH/SxTMRL Register
Used to specify the divide value for the OSC clock, post-
PLL clock or SxCLK input (specified in the SxMode
register bits CLKS1:0, Section 7.1.11) to generate the
SERDES clock. The effective divide value = {SxTMRH,
SxTMRL} + 1, except, in the case of SPI and GPSI Master,
the effective divide value = {(SxTMRH/SxTMRL) + 1} x 2.
7.1.18 SPDREG Register
Status of clock and PLL settings during run-time.
Note: This is a read-only register, use speed instruction
to change settings.
76543 0
PLL OSC CLK1:0 CDIV3:0
Name Description
PLL Run-time control of PLL clock multiplier
operation. If the PLL is not required,
power consumption can be reduced by
disabling it, but a WUDP delay is required
to start it again (controlled in FUSE0).
0 = PLL clock multiplier on
1 = PLL clock multiplier off
OSC Run-time control of OSC oscillator opera-
tion. If the OSC clock is not required,
power consumption can be reduced by
disabling it (stops OSC crystal oscillator
and blocks propagation of OSC1 external
clock input).
0 = OSC oscillator enabled
1 = OSC oscillator disabled
CLK1:0 Selects the system clock source.
00 = PLL clock multiplier
01 = OSC oscillator/external clock
on OSC1 input
10 = RTCLK oscillator/external clock
on RTCLK1 input
11 = System clock disabled (off)
CDIV3:0 Selects the divisor which divides the sys-
tem clock to give the core clock.
0000 = 1
0001 = 2
0010 = 3
0011 = 4
0100 = 5
0101 = 6
0110 = 8
0111 = 10
1000 = 12
1001 = 16
1010 = 24
1011 = 32
1100 = 48
1101 = 64
1110 = 128
1111 = System clock disabled (off)
Name Description
100 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.19 STATUS Register
Condition flags for the results of arithmetic and logical
operations, the page bits, and bits which indicate the
skipping state of the core and control of continuation skip
after return from interrupt.
7543210
PA2:0 SAR SSF Z DC C
Name Description
PA2:0 Program memory page select bits (read
only). Used to extend the 13-bit address
encoded in jump and call instructions
(these 3 bits are written to the upper 3 bits
of the program counter when a jump or
call occurs). Modified using the page
instruction.
SAR Skip After Return bit. Indicates if the core
should be in the skipping/not state after
the completion of a return instruction
(ret, retnp, or retw instructions, but
not reti). The return instruction will also
clear the SAR bit to ensure correct behav-
ior after the dynamic jump.
0 = The core should not be in a
skipping state upon comple-
tion of the return.
1 = The core should be in a skip-
ping state upon completion of
the return.
SSF Shadowed Skipping/not state Flag. Gives
the ISR the ability to know if the interrupt
occurred immediately following a skip
instruction. The software can choose
either to clear the SSF flag in the ISR or to
make the first instruction of the mainline
context switching code a nop to flush out
the skip state.
0 = The core was not in a skipping
state when interrupted.
1 = The core was in a skipping
state when interrupted.
Z Zero bit. Affected by most logical, arith-
metic, and data movement instructions
(refer to “Flags Affected” column in Table
4-2 through Table 4-7). Set if the result
was zero, otherwise cleared.
0 = Result of last ALU operation
was non-zero.
1 = Result of last ALU operation
was zero.
DC Digit Carry bit. After addition, set if carry
from bit 3 occurred, otherwise cleared.
After subtraction, cleared if borrow from
bit 3 occurred, otherwise set.
0 = Last addition did not generate
carry out of bit 3, or last sub-
traction generated borrow out
of bit 3.
1 = Last addition generated carry
out of bit 3, or last subtraction
did not generate borrow out of
bit 3.
C Carry bit. After addition, set if carry from
bit 7 of the result occurred, otherwise
cleared. After subtraction, cleared if bor-
row from bit 7 of the result occurred, oth-
erwise set. After rotate (rr or rl)
instructions, loaded with the LSB or MSB
of the operand, respectively.
0 = Last addition did not generate
carry out of bit 7, last subtrac-
tion generated borrow out of bit
7, or last rotate loaded a 0.
1 = Last addition generated carry
out of bit 7, last subtraction did
not generate borrow out of bit
7, or last rotate loaded a 1.
Name Description
IP2012 / IP2022 Data Sheet
www.ubicom.com 101
7.1.20 T0CFG Register
Timer 0 configuration.
7.1.21 TxCFG1H Register
Timer 1 and 2 configuration.
76 3 2 10
T0EN T0PS3:0 Rsrvd. T0IE T0IF
Name Description
T0EN Enables Timer 0
0 = Timer 0 disabled
1 = Timer 0 enabled
T0PS3:0 Specifies Timer 0 prescaler divisor
0000 = 1
0001 = 2
0010 = 4
0011 = 8
0100 = 16
0101 = 32
0110 = 64
0111 = 128
1000 = 256
1001 to 1111 = Reserved
T0IE Timer 0 interrupt enable bit
0 = Timer 0 interrupt disabled
1 = Timer 0 interrupt enabled
T0IF Timer 0 interrupt flag
0 = No timer overflow has occurred
since this bit was last cleared
1 = Timer overflow has occurred
76543210
OFIE
CAP2IE
CMP2IE
CAP1IE
CMP1IE
OFIF
CAP2IF
CMP2IF
CAP1IF
CMP1IF
Name Description
OFIE Timer overflow interrupt enable bit
0 = Overflow interrupt disabled
1 = Overflow interrupt enabled
CAP2IE
or
PWM mode: Compare 2 interrupt enable
bit
CMP2IE Capture/Compare mode: Capture 2 inter-
rupt enable bit
0 = Capture/Compare 2 interrupt
disabled
1 = Capture/Compare 2 interrupt
enabled
CAP1IE Capture 1 interrupt enable bit
0 = Capture 1 interrupt disabled
1 = Capture 1 interrupt enabled
CMP1IE Compare 1 interrupt enable bit
0 = Compare 1 interrupt disabled
1 = Compare 1 interrupt enabled
OFIF Timer overflow interrupt flag
0 = No timer overflow has occurred
since this bit was last cleared
1 = Timer overflow has occurred
CAP2IF
or
PWM mode: Compare 2 interrupt flag (i.e.
timer value matched TxCMP2 value)
CMP2IF Capture/Compare mode: Capture 2 flag
(i.e. TxCPI2 input triggered)
0 = No capture/compare 2 event
has occurred since this bit was
last cleared
1 = Capture/compare 2 event has
occurred
CAP1IF Capture 1 interrupt flag
0 = No capture 1 event has
occurred since this bit was last
cleared
1 = Capture 1 event has occurred
CMP1IF Compare 1 interrupt flag
0 = No compare 1 event has
occurred since this bit was last
cleared
1 = Compare 1 event has occurred
Name Description
102 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.22 TxCFG2H Register
Timer 1 and 2 configuration.
7.1.23 TxCFG1L Register
Timer 1 and 2 configuration.
76543 0
0000 PS3:0
Name Description
PS3:0 Timer prescaler divisor
0000 = 1 1000 = 256
0001 = 2 1001 = 512
0010 = 4 1010 = 1024
0011 = 8 1011 = 2048
0100 = 16 1100 = 4096
0101 = 32 1101 = 8192
0110 = 64 1110 = 16384
0111 = 128 1111 = 32768
76543210
MODE
OEN
ECLKEN
CPI2EN
CPI1EN
ECLKEDG
CAP1RST
TMREN
Name Description
MODE Timer mode select
0 = PWM/timer mode
1 = Capture/compare mode
OEN TxOUT enable bit
0 = TxOUT disabled. Port pin avail-
able for general-purpose I/O.
1 = TxOUT enabled. Port pin must
be configured for output in cor-
responding RxDIR register bit.
Output is on RA3 for T1, RB3
for T2
ECLKEN TxCLK enable bit
0 = TxCLK disabled. Port pin avail-
able for general-purpose I/O.
1 = TxCLK enabled as clock
source for timer. Enabling this
bit does not make any other
restrictions on the use of the
TxCLK port pin for general-pur-
pose I/O.
CPI2EN TxCPI2 enable bit
0 = System clock enabled as clock
source for timer. TxCPI2 port
pin available for general-pur-
pose I/O.
1 = TxCLK enabled as clock
source for timer. Enabling this
bit does not make any other
restrictions on the use of the
port pin for general-purpose
I/O.
CPI1EN TxCPI1 enable bit
0 = Capture 1 input disabled.
TxCPI1 port pin available for
general-purpose I/O.
1 = TxCPI1 enabled as capture 1
input. Enabling this bit does not
make any other restrictions on
the use of the port pin for gen-
eral-purpose I/O.
ECLKEDG TxCLK edge sensitivity select. (This bit is
ignored if the ECLKEN bit is clear.)
0 = TxCLK increments timer on ris-
ing edge
1 = TxCLK increments timer on fall-
ing edge
CAP1RST Reset timer on capture 1 event enable bit
0 = Timer value unchanged by
occurrence of a capture 1
event
1 = Timer value cleared by occur-
rence of a capture 1 event
TMREN Timer enable bit
0 = Timer disabled. Timer clock
source shut off to reduce power
consumption.
1 = Timer enabled
Name Description
IP2012 / IP2022 Data Sheet
www.ubicom.com 103
7.1.24 TxCFG2L Register
Timer 1 and 2 configuration.
76 543210
Reserved
TOUTSET
TOUTCLR
CPI2CPI1
CPI2EDG1:0
CPI1EDG1:0
Name Description
TOUTSET Override bit to set the TxOUT output.
This bit always reads as zero.
0 = Writing 0 to this bit has no
effect
1 = Writing 1 to this bit forces the
TxOUT signal high
TOUTCLR Override bit to clear the TxOUT output.
This bit always reads as zero.
0 = Writing 0 to this bit has no
effect
1 = Writing 1 to this bit forces the
TxOUT signal low
CPI2CPI1 Internally connect the TxCPI2 input to
the TxCPI1 input. This makes the
TxCPI2 port pin available for general-
purpose I/O.
0 = No internal connection
between TxCPI1 and TxCPI2
1 = TxCPI1 and TxCPI2 internally
connected
CPI2EDG1:0 TxCPI2 edge sensitivity select
00 = Falling edge on TXCPI2 rec-
ognized as capture 2 event
01 = Rising edge on TXCPI2 rec-
ognized as capture 2 event
10 = Any falling or rising edge on
TXCPI2 recognized as cap-
ture 2 event
11 = Any falling or rising edge on
TXCPI2 recognized as cap-
ture 2 event
CPI1EDG1:0 TxCPI1 edge sensitivity select
00 = Falling edge on TXCPI1 rec-
ognized as capture 1 event
01 = Rising edge on TXCPI1 rec-
ognized as capture 1 event
10 = Any falling or rising edge on
TXCPI1 recognized as cap-
ture 1 event
11 = Any falling or rising edge on
TXCPI1 recognized as cap-
ture 1 event
Name Description
104 www.ubicom.com
IP2012 / IP2022 Data Sheet
7.1.25 TCTRL Register
Timer 1 and 2 configuration.
7.1.26 XCFG Register
Extra configuration bits for various functions.
76 5 4 32 1 0
00T2IE T1IE 0 0 T2RST T1RST
Name Description
T2IE Timer 2 interrupt enable
0 = Timer 2 interrupt disabled
1 = Timer 2 interrupt enabled
T1IE Timer 1 interrupt enable
0 = Timer 1 interrupt disabled
1 = Timer 1 interrupt enabled
T2RST Timer 2 reset bit. This bit always reads as
zero.
0 = Writing 0 to this bit has no
effect.
1 = Writing 1 to this bit clears Timer
2.
T1RST Timer 1 reset bit. This bit always reads as
zero.
0 = Writing 0 to this bit has no
effect.
1 = Writing 1 to this bit clears Timer
1.
76 5 4 3210
GIE FWP RTEOS RTOSC_EN INT_EN Rsvd FBUSY
Name Description
GIE Global interrupt enable bit
0 = Interrupts disabled
1 = Interrupts enabled
FWP Flash write protect bit. This bit only
affects operation of fwrite and
ferase self-programming instructions
on flash, not programming through the
ISD/ISP interface. Does not affect writes
or erases of RAM.
0 = Writes to flash memory dis-
abled (act as nop instructions)
1 = Writes to flash memory
enabled
RTEOS Real-time timer oversampling bit
0 = Oversampling disabled
1 = Oversampling enabled
RTOSC_EN RTCLK oscillator enable bit
0 = RTCLK oscillator is operational
1 = RTCLK oscillator turned off
INT_EN int instruction interrupt enable bit
0 = int instructions only incre-
ment the PC, like nop
1 = int instructions cause inter-
rupts
FBUSY Flash memory busy bit (read-only).
For more information about programming
the flash memory, see Section 4.7.
0 = Flash memory is idle
1 = Fetching instructions out of
flash memory or busy process-
ing an iread, ireadi,
fwrite, fread or ferase
instruction on Flash
IP2012 / IP2022 Data Sheet
www.ubicom.com 105
8.0 Electrical Characteristics
8.1 Absolute Maximum Ratings (beyond which permanent damage may occur. Correct operation not guar-
anteed outside of DC specifications in Section 8.2)
Parameter Minimum Maximum Units
Ambient temperature under bias -40 85 ˚C
Storage temperature -65 150 ˚C
PQFP Soldering temperature ramp to 160-180˚C 2.5 ˚C/sec
PQFP Soldering hold time at 160-180˚C60110 sec
PQFP Soldering temperature ramp from 160-180˚C to 240˚C maximum 3.0 ˚C/sec
PQFP Soldering hold time at 240˚C maximum 10 40 sec
PQFP Soldering temperature ramp down to 180˚C5˚C/sec
Voltage on DVdd, XVdd, AVdd, and GVdd with respect to Vss -0.5 3.5 V
Voltage on IOVdd with respect to Vss -0.5 4.5 V
Voltage on Port A through Port F, OSC1, RST, RTCLK1, TSCK, TSI, and TSS
inputs with respect to Vss
-0.5 5.7 V
Voltage on Port G inputs with respect to Vss -0.5 3.5 V
Total power dissipation 1W
Maximum current out of all DVss pins 400 mA
Maximum current into all DVdd pins 400 mA
Maximum allowable sink current per I/O pin 160 mA
Maximum allowable source current per I/O pin (excluding port G) 160 mA
Maximum allowable source current per G pin 20 mA
Maximum allowable sink current per group of I/O pins between IOVss pins 160 mA
Maximum allowable source current per group of I/O pins between IOVdd pins
(excluding port G)
160 mA
Latchup 200 mA
θJA, 80-pin PQFP Package 48 ˚C/W
θJA, 80-pin µBGA Package 37 ˚C/W
Flash block erase cycle lifetime (if using 20ms block erases - Section 7.1.5) 20K Cycles
Flash bulk erase cycle lifetime 20K Cycles
ESD Human Body Model - all pins 2000 V
ESD Machine Model - all pins 200 V
106 www.ubicom.com
IP2012 / IP2022 Data Sheet
8.2 DC Specifications: IP2022-120, IP2012-120
Operating Temperature -40˚C < Ta < +85˚C
Symbol Parameter Min Typ Max Units Conditions
DVdd Digital supply voltage 2.3 2.5 2.7 V
AVdd Analog supply voltage 2.3 2.5 2.7 V = DVdd (filters between supplies)
GVdd Port G supply voltage 2.3 2.5 2.7 V = DVdd (filters between supplies)
XVdd PLL supply voltage 2.3 2.5 2.7 V = DVdd (filters between supplies)
IOVdd I/O supply voltage
(except Port G)
> DVdd 2.5/3.3 3.6 V See note 4
Idd Supply current, full operation
DVdd + AVdd + GVdd +
XVdd
150 mA DVdd = 2.3 - 2.7V, 120 MHz CPU core
executing 100% of time (during ISR
and main program code)
70 mA DVdd = 2.5V, 120 MHz CPU core in
ISR only, 40MHz CPU core in main pro-
gram code (Webserver application)
IddIO Supply current, full operation
IOVdd only
mA IOVdd = DVdd - 3.6V
No loads, no floating inputs
Isleep Supply current, sleep
DVdd + AVdd + GVdd +
XVdd
200 µA DVdd = 2.7V, PLL and oscillators off
(XCFG bit 4 = 1, CMPCFG bit 7 = 0,
ADCCFG bit 3 = 0, XCFG bit 0 = 0,
FUSE1 bit 3 = 0)
IsleepIO Supply current, sleep
IOVdd only
µA IOVdd = 3.6V,
PLL and oscillators off,
No loads, no floating inputs
Vih Input high voltage, Port A
through Port F
1.8 5.5 V DVdd = 2.3 - 2.7V,
IOVdd = DVdd - 3.6V
Input high voltage, OSC1
and RTCLK1 inputs
1.8 DVdd V
Input high voltage, RST,
TSCK, TSI, and TSS inputs
2.25 5.5 V
Vil Input low voltage, Port A
through Port F
1.0 V DVdd = 2.3 - 2.7V,
IOVdd = DVdd - 3.6V
Input low voltage, OSC1 and
RTCLK1 inputs
0.4 V
Input low voltage, RST,
TSCK, TSI, and TSS inputs
0.9 V
Vina Analog input voltage (Port
G)
AVdd V See note 1
Iil Input leakage current for
Port A through Port G,
RTCLK1, and TSO pins
-1 ±0.001 1 µA Port G = 0V or AVdd (see note 1),
RTCLK1 = 0V or DVdd
All other inputs = 0V or 5.5V
TSO measured while TSS = 1
RTCLK1 measured in sleep mode and
FUSE0 bit 14 = 1
IP2012 / IP2022 Data Sheet
www.ubicom.com 107
1. If Vref is used for the ADC reference voltage (see Section 5.7.1), then the maximum input voltage on a Port G in-
put is Vref.
2. Data in the Typical (“Typ”) column is at 2.5/3.3V, 25˚C unless otherwise stated.
3. The Port A through Port F pins have a weak latch, even when the pin is configured as an input, which drives
floating I/O pins to 0V or to a diode drop below DVdd. Some current is required to toggle the state of the latch.
4. If IOVdd rises before DVdd, the IP2012 / IP2022 may drive the I/O pins to IOVdd before DVdd has stabilized.
However, there is an internal diode from DVdd to IOVdd, so DVdd should never exceed IOVdd.
5. These pins are guaranteed to pull up above their Vih level, even if IOVdd = DVdd = 2.3V.
Iil Input leakage current for
OSC1 pin
-1 10 µA OSC1 = 0V or DVdd
OSC1 measured in sleep mode and
FUSE0 bit 15 = 1
Iil Input pull-up/down leakage
current for Port A through
Port F pins
-300 - 100 µA See note 3.
Iilt Input leakage current for
RST, TSCK, TSI, TSS inputs
-60 1 µA Vin = 0 to 5.5V. These pins have active
internal pull-ups to a diode drop below
IOVdd. See note 5. (60KOhm min.,
103KOhm typ., 173KOhm max)
Ioh Output high current from
Port A pins and RE5, RE6,
RF1 and RF2 pins
24 60 96 mA Voh = 2.4V
IOVdd = 3.0 to 3.6V
Output high current from
Port B pins and RE4:0, RE7,
RF7:3, RF0, and TSO pins
11 24 39 mA
Output high current from
Port C and Port D pins
81829mA
Output high current from
Port G pins
41224mAVoh = 1.8V
GVdd = 2.3 to 2.7V
Iol Output low current from Port
A pins and RE5, RE6, RF1
and RF2 pins
25 40 50 mA Vol = 0.4V
IOVdd = 3.0 to 3.6V
Output low current from Port
B pins and RE4:0, RE7,
RF7:3, RF0, and TSO pins
91624mA
Output low current from Port
C and Port D pins
61115mA
Output low current from Port
G pins
41324mAVol = 0.4V
GVdd = 2.3 to 2.7V
Symbol Parameter Min Typ Max Units Conditions
108 www.ubicom.com
IP2012 / IP2022 Data Sheet
8.3 DC Specifications: IP2022-160
Operating Temperature 0˚C < Ta < +55˚C
Symbol Parameter Min Typ Max Units Conditions
DVdd Digital supply voltage 2.55 2.625 2.70 V
AVdd Analog supply voltage 2.55 2.625 2.70 V = DVdd (filters between supplies)
GVdd Port G supply voltage 2.55 2.625 2.70 V = DVdd (filters between supplies)
XVdd PLL supply voltage 2.55 2.625 2.70 V = DVdd (filters between supplies)
IOVdd I/O supply voltage
(except Port G)
> DVdd 2.625/
3.3
3.6 V See note 4
Idd Supply current, full operation
DVdd + AVdd + GVdd +
XVdd
mA DVdd = 2.55 - 2.70V, 160 MHz CPU
core executing 100% of time (during
ISR and main program code)
mA DVdd = 2.625V, 160 MHz CPU core in
ISR only, 53.33MHz CPU core in main
program code (Webserver application)
IddIO Supply current, full operation
IOVdd only
mA IOVdd = DVdd - 3.6V
No loads, no floating inputs
Isleep Supply current, sleep
DVdd + AVdd + GVdd +
XVdd
200 µA DVdd = 2.70V, PLL and oscillators off
(XCFG bit 4 = 1, CMPCFG bit 7 = 0,
ADCCFG bit 3 = 0, XCFG bit 0 = 0,
FUSE1 bit 3 = 0)
IsleepIO Supply current, sleep
IOVdd only
µA IOVdd = 3.6V,
PLL and oscillators off,
No loads, no floating inputs
Vih Input high voltage, Port A
through Port F
1.8 5.5 V DVdd = 2.55 - 2.70V,
IOVdd = DVdd - 3.6V
Input high voltage, OSC1
and RTCLK1 inputs
1.8 DVdd V
Input high voltage, RST,
TSCK, TSI, and TSS inputs
2.25 5.5 V
Vil Input low voltage, Port A
through Port F
1.0 V DVdd = 2.55 - 2.70V,
IOVdd = DVdd - 3.6V
Input low voltage, OSC1 and
RTCLK1 inputs
0.4 V
Input low voltage, RST,
TSCK, TSI, and TSS inputs
0.9 V
Vina Analog input voltage (Port
G)
AVdd V See note 1
Iil Input leakage current for
Port A through Port G,
RTCLK1, and TSO pins
-1 ±0.001 1 µA Port G = 0V or AVdd (see note 1),
RTCLK1 = 0V or DVdd
All other inputs = 0V or 5.5V
TSO measured while TSS = 1
RTCLK1 measured in sleep mode and
FUSE0 bit 14 = 1
IP2012 / IP2022 Data Sheet
www.ubicom.com 109
1. If Vref is used for the ADC reference voltage (see Section 5.7.1), then the maximum input voltage on a Port G in-
put is Vref.
2. Data in the Typical (“Typ”) column is at 2.625/3.3V, 25˚C unless otherwise stated.
3. The Port A through Port F pins have a weak latch, even when the pin is configured as an input, which drives
floating I/O pins to 0V or to a diode drop below DVdd. Some current is required to toggle the state of the latch.
4. If IOVdd rises before DVdd, the IP2022 may drive the I/O pins to IOVdd before DVdd has stabilized. However,
there is an internal diode from DVdd to IOVdd, so DVdd should never exceed IOVdd.
5. These pins are guaranteed to pull up above their Vih level, even if IOVdd = DVdd = 2.55V.
Iil Input leakage current for
OSC1 pin
-1 10 µA OSC1 = 0V or DVdd
OSC1 measured in sleep mode and
FUSE0 bit 15 = 1
Iil Input pull-up/down leakage
current for Port A through
Port F pins
-80 - 80 µA See note 3.
Iilt Input leakage current for
RST, TSCK, TSI, TSS inputs
-60 1 µA Vin = 0 to 5.5V. These pins have active
internal pull-ups to a diode drop below
IOVdd. See note 5. (60KOhm min.,
103KOhm typ., 173KOhm max)
Ioh Output high current from
Port A pins and RE5, RE6,
RF1 and RF2 pins
24 60 96 mA Voh = 2.4V
IOVdd = 3.0 to 3.6V
Output high current from
Port B pins and RE4:0, RE7,
RF7:3, RF0, and TSO pins
11 24 39 mA
Output high current from
Port C and Port D pins
81829mA
Output high current from
Port G pins
41224mAVoh = 1.8V
GVdd = 2.55 to 2.70V
Iol Output low current from Port
A pins and RE5, RE6, RF1
and RF2 pins
25 40 50 mA Vol = 0.4V
IOVdd = 3.0 to 3.6V
Output low current from Port
B pins and RE4:0, RE7,
RF7:3, RF0, and TSO pins
91624mA
Output low current from Port
C and Port D pins
61115mA
Output low current from Port
G pins
41324mAVol = 0.4V
GVdd = 2.55 to 2.70V
Symbol Parameter Min Typ Max Units Conditions
110 www.ubicom.com
IP2012 / IP2022 Data Sheet
8.4 AC Specifications: IP2022-120, IP2012-120
Operating Temperature: -40˚C < Ta < +85˚C
6. Vdd must start rising from Vss to ensure proper Power-On-Reset when relying on the internal Power-On-Reset
circuitry. If power supply takes more than 50ms to rise from 0 to 2.5V, use RCs on RST pin (see Figure 3-16).
Symbol Parameter Min Typ Max Units Conditions
Fcore CPU core clock frequency 0 120 MHz Execution from
program RAM
Fflash CPU core clock frequency 0 40 MHz Execution from
program flash
Fsys System clock frequency 0 120 MHz
Fosc External clock frequency on OSC1 0 120 MHz
Foscr External clock frequency on RTCLK1 0 120 MHz
Fxo Crystal frequency, OSC 4.75 4.8 5.0 MHz Ext. crystal, ±100
ppm
Fpll PLL input frequency, after predivider 4.75 4.8 5.0 MHz TRIM0=FBFE
Fxr Crystal frequency, RTCLK 32.765 32.768 32.771 kHz Ext. 32.768 KHz
crystal, ±100 ppm
Tosl, Tosh Clock in (OSC1) low or high time 3 ns
Tr l, Trh Clock in (RTCLK1) low or high time 3.5 ns
SVdd DVdd slew rate to ensure Power-On
reset
0.05 V/ms See note 6.
IP2012 / IP2022 Data Sheet
www.ubicom.com 111
8.5 AC Specifications: IP2022-160
Operating Temperature: 0˚C < Ta < +55˚C
7. Vdd must start rising from Vss to ensure proper Power-On-Reset when relying on the internal Power-On-Reset
circuitry. If power supply takes more than 50ms to rise from 0 to 2.5V, use RCs on RST pin (see Figure 3-16).
8. The 160MHz part is specified for everything the 120MHz part is specified for (except temperature range), and it
will also run at 160MHz. CPU core clock and system clock frequencies between 120 and 160MHz are not guar-
anteed, because the PLL input frequency is only specified for 4.75–5MHz (with TRIM0=FBFE) to get a system
clock up to 120MHz, or at 3.2MHz (with TRIM0=FBFD) to get 160MHz. See Figure 3-18, and configure FUSE0,
TRIM0, and FCFG correctly. Typically customers who buy 160MHz parts will only use a 3.2MHz crystal, giving
3.2MHz into the PLL, and a 160MHz system clock, with TRIM0=FBFD and FUSE0 bits 15, 11, 10, and 9 = 0.
Symbol Parameter Min Typ Max Units Conditions
Fcore CPU core clock frequency 0 120 or 160 MHz Execution from program
RAM. See Note 8.
Fflash CPU core clock frequency 0 40 or 53.33 MHz Execution from program
flash. See Note 8.
Fsys System clock frequency 0 120 or 160 MHz See Note 8.
Fosc External clock frequency on OSC1 0 120 MHz
Foscr External clock frequency on
RTCLK1
0 120 MHz
Fxo Crystal frequency, OSC 3.2 MHz Ext. crystal, ±100 ppm.
See Note 8.
Fpll PLL input frequency, after predivider 3.2 MHz TRIM0=FBFD. See Note
8.
Fxr Crystal frequency, RTCLK 32.765 32.768 32.771 kHz Ext. crystal, ±100 ppm
Tosl, Tosh Ext. clock in (OSC1) low or high time 3 ns
Tr l, Trh Ext. clock in (RTCLK1) low or high
time
3.5 ns
SVdd DVdd slew rate to ensure Power-On
reset
0.05 V/ms See Note 7.
112 www.ubicom.com
IP2012 / IP2022 Data Sheet
8.6 Comparator DC and AC Specifications
Operating Temperature: –40˚C < Ta < +85˚C for IP2022-120 and IP2012-120; 0˚C < Ta < +55˚C for IP2022-160.
8.7 ADC 10-bit Converter DC and AC Specifications
Vref = AVdd, -40˚C < Ta < +85˚C for IP2022-120, IP2012-120; 0˚C < Ta < +55˚C for IP2022-160
Parameter Min Typ Max Units Conditions
Input offset voltage ±10 ±25 mV CMPT2:0 bits in TRIM0 reg-
ister are 111
Hysteresis, rising or falling edge 20 50 80 mV
Bandwidth 15 MHz min. 100 mV peak-to-peak
Response time 100 ns Voverdrive = 50 mV
Does not include compara-
tor mode entry stabilization
time
Time from enabling comparator until
output is valid
2000 ns
Input voltage range 0.1 AVdd - 0.1 V
Parameter Min Typ Max Units Conditions
Sampling Rate 48 kHz
Conversion Time 20.8 µs
Differential nonlinearity error (DNL) ±1.0 LSB
Integral nonlinearity error (INL) ±1.25 LSB
Offset error ±1.0 LSB
Full-scale error ±1.0 LSB
IP2012 / IP2022 Data Sheet
www.ubicom.com 113
9.0 Package Dimensions
9.1 PQFP
80-pin, 14 mm ∞ 20 mm ∞ 2.8 mm body, 0.8 mm pitch, 17.9 mm ∞ 23.9 mm tip-to-tip. All dimensions in mm.
c
c
Gauge Plane
Pin
114 www.ubicom.com
IP2012 / IP2022 Data Sheet
9.2 µBGA (available for IP2022-120 only)
80-pin, 9 mm ∞ 9 mm ∞ 0.95 mm body, 0.8 mm ball spacing
REF.
A
A1
A2
D
D1
DIMENSIONAL REFERENCE
D2
E
E1
E2
b
c
aaa
bbb
ccc
e
f
M
N
0.80 0.90
10
80
0.80 BSC
8.80
0.39
9.00
0.44
0.25
7.20 BSC
0.12
0.10
0.10
1.00
9.20
0.49
9.00
9.00
9.00
9.20
9.20
9.20
8.80
8.80
8.80
0.65 0.70 0.75
0.21 0.26 0.31
1.11 1.21 1.31
MIN. NOM. MAX.
7.20 BSC
D
0.10
A
B
E
D2
TOP VIEW BOTTOM VIEW
(E1) e
e
(D1)
10987654321
AB
CD
EF
GH
JK
SIDE VIEW
DETAIL A
DETAIL B
f
f
b4
NX b
0.15 S
S
C
C
AB
SS
0.075
DETAIL A
SEATING PLANE
A2 A1 A
c
ccc C
C
C
bbb
6
5
aaa C
1. All Dimensins are in millimeters.
2. ’e’ represents the basic solder ball pitch.
3. ’M’ represents the basic solder ball matrix size, and
symbol ’N’ is the number of balls after depopulating.
4. ’b’ is measureable at the maximum solder ball diameter
after reflow parallel to primary datum - C - .
5. Dimension ’aaa’ is measured parallel to primary datum - C - .
6. Primary datum - C - and seating plane are defined by the
spherical crowns of the solder balls.
7. Package surface shall be matte finish charmilles 24 to 27.
8. package centering to substrate shall be 0.0760mm maximum
for both x and y direction respectively.
9. Package warp shall be 0.050mm maximum.
10. Substrate material base is BT resin.
11. The overall package thickness ’A’ already considers collapse balls.
12. Dimensioning and tolerancing per ASME Y14.5-1994.
IP2012 / IP2022 Data Sheet
www.ubicom.com 115
10.0 Part Numbering
Table 10-1 Ordering Information
Device Pins I/O Package Program Flash Program RAM Data RAM Temperature
IP2012/PQ80-120 80 48 PQFP 64KBytes (32K x 16) 16KBytes (8K x 16) 4KBytes –40˚ to 85˚C
IP2022/PQ80-120 80 52 PQFP 64KBytes (32K x 16) 16KBytes (8K x 16) 4KBytes –40˚ to 85˚C
IP2022/BG80-120 80 52 µBGA 64KBytes (32K x 16) 16KBytes (8K x 16) 4KBytes –40˚ to 85˚C
IP2022/PQ80-160 80 52 PQFP 64KBytes (32K x 16) 16KBytes (8K x 16) 4KBytes 0˚ to 55˚C
515-013e.eps
IP 2022 xx80 xxx/-
MIPS Rating
Package Type, PQ80 = 80-pin PQFP, BG80 = 80-pin uBGA
Device Number
Device Family
Ubicom, Inc. develops and markets wireless network processor and
software platforms that enable all electronic devices to be connected to
each other – securely, cost-effectively and transparently. With headquarters
in Mountain View, California, Ubicom also has offices in Southern California
as well as Belgium, Taiwan and Hong Kong. For more information, visit
www.ubicom.com.
Copyright © 2001–2003 Ubicom, Inc. All rights reserved. Ubicom, IP2012
and IP2022 are trademarks of Ubicom, Inc. All other trademarks are the
property of their respective holders.
IP2K-DDS-2000-16 (4/17/2003)
635 Clyde Avenue
Mountain View, CA 94043
Tel: 650.210.1500
Fax: 650.210.8715
Email:sales@ubicom.com
Web: www.ubicom.com
Sales and Technical Support Contact Information
For the latest contact and support information on IP devices, please visit the Ubicom Web site at www.ubicom.com. The
site contains technical literature, local sales contacts, tech support, and many other features.
The Products are not authorized for use in life support systems or under conditions where failure of the Product would
endanger the life or safety of the user, except when prior written approval is obtained from Ubicom, Inc. Ask your sales
representive for details.
