FEATURES Matched pair of programmable filters and VGAs Continuous gain control range: -5 dB to +45 dB 6-pole filter 1 MHz to 30 MHz in 1 MHz steps, 0.5 dB corner frequency SPI programmable 6 dB front-end gain step IMD3: >55 dBc for 1.5 V p-p composite output HD2, HD3: >60 dBc for 1.5 V p-p output Differential input and output Adjustable output common-mode voltage Optional dc output offset correction Power-down feature Single 5 V supply operation APPLICATIONS FUNCTIONAL BLOCK DIAGRAM ENBL INP1 INM1 VPS COM GNSW OFS1 VPS VPSD OPP1 COMD OPM1 COM LE CLK ADRF6510 SPI GAIN VOCM DATA SDO COM COM OPM2 VPS OPP2 COM INP2 INM2 VPS COM OFDS OFS2 VPS 09002-001 Data Sheet 30 MHz Dual Programmable Filters and Variable Gain Amplifiers ADRF6510 Figure 1. Baseband I/Q receivers Diversity receivers ADC drivers GENERAL DESCRIPTION The ADRF6510 is a matched pair of fully differential low noise and low distortion programmable filters and variable gain amplifiers (VGAs). Each channel is capable of rejecting large out-ofband interferers while reliably boosting the wanted signal, thus reducing the bandwidth and resolution requirements on the analog-to-digital converters (ADCs). The excellent matching between channels and their high spurious-free dynamic range over all gain and bandwidth settings makes the ADRF6510 ideal for quadrature-based (IQ) communication systems with dense constellations, multiple carriers, and nearby interferers. The filters provide a six-pole Butterworth response with 0.5 dB corner frequencies programmable through the SPI port from 1 MHz to 30 MHz in 1 MHz steps. The preamplifier that precedes the filters offers a pin-programmable option of either 6 dB or 12 dB of gain. The preamplifier sets a differential input impedance of 400 and has a common-mode voltage that defaults to 2.1 V but can be driven from 1.5 V to 2.5 V. The variable gain amplifiers that follow the filters provide 50 dB of continuous gain control with a slope of 30 mV/dB. The output buffers provide a differential output impedance of 20 that is capable of driving 1.5 V p-p into 1 k loads. The output commonmode voltage defaults to VPS/2, but it can be programmed via the VOCM pin. The built-in dc offset correction loop can be disabled if dc-coupled operation is desired. The high-pass corner frequency is defined by external capacitors on the OFS1 and OFS2 pins. The ADRF6510 operates from a 4.75 V to 5.25 V supply and consumes a maximum supply current of 258 mA when programmed to the highest bandwidth setting. When disabled, it consumes 2 mA. The ADRF6510 is fabricated in an advanced silicon-germanium BiCMOS process and is available in a 32-lead, exposed paddle LFCSP. Performance is specified over the -40C to +85C temperature range. Rev. A Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.461.3113 (c)2010-2011 Analog Devices, Inc. All rights reserved. ADRF6510 Data Sheet TABLE OF CONTENTS Features .............................................................................................. 1 Distortion Characteristics ......................................................... 18 Applications....................................................................................... 1 Maximizing the Dynamic Range.............................................. 18 Functional Block Diagram .............................................................. 1 Key Parameters for Quadrature-Based Receivers .................. 19 General Description ......................................................................... 1 Applications Information .............................................................. 20 Revision History ............................................................................... 2 Basic Connections...................................................................... 20 Specifications..................................................................................... 3 Error Vector Magnitude (EVM) Performance........................... 20 Timing Diagrams.......................................................................... 5 EVM ............................................................................................. 21 Absolute Maximum Ratings............................................................ 6 Effect of Filter Bandwidth on EVM ......................................... 21 ESD Caution.................................................................................. 6 Effect of Output Voltage Levels on EVM ................................ 21 Pin Configuration and Function Descriptions............................. 7 Effect of COFS on EVM ............................................................... 22 Typical Performance Characteristics ............................................. 8 Anti-Aliasing Filter .................................................................... 22 Theory of Operation ...................................................................... 15 Evaluation Board ............................................................................ 24 Input Buffers ............................................................................... 15 Evaluation Board Control Software......................................... 24 Programmable Filters................................................................. 15 Schematics and Artwork ........................................................... 24 Variable Gain Amplifiers (VGAs) ............................................ 16 Evaluation Board Configuration Options............................... 27 Output Buffers/ADC Drivers ................................................... 16 Outline Dimensions ....................................................................... 29 DC Offset Compensation Loop................................................ 16 Ordering Guide .......................................................................... 29 Programming the Filters............................................................ 17 Noise Characteristics ................................................................. 17 REVISION HISTORY 10/11--Rev. 0 to Rev. A Changes to Figure 2 and Figure 3................................................... 5 Changes to Table 3............................................................................ 7 Changes to Figure 48, Changes to Error Vector Magnitude (EVM) Performance Section......................................................... 20 Deleted Low IF Image Rejection Section, and Example Baseband Interface Section, Figure 50, and Figure 51; Renumbered Subsequent Figures ................................................. 20 Changes to Figure 49...................................................................... 21 Added EVM Section, Effect of Filter Bandwidth on EVM Section, Effect of Output Voltage Levels on EVM Section, and Effect of COFS on EVM Section...................................................... 21 Added Anti-Aliasing Filter Section.............................................. 22 Changes to Figure 54...................................................................... 22 Changes to Figure 58...................................................................... 24 Changes to Figure 61...................................................................... 25 Changes to Figure 62 and Figure 63............................................. 26 Changes to Table 5.......................................................................... 27 4/10--Revision 0: Initial Version Rev. A | Page 2 of 32 Data Sheet ADRF6510 SPECIFICATIONS VPS = 5 V, TA = 25C, ZSOURCE = 400 , ZLOAD = 1 k, VOUT = 1.5 V p-p, bandwidth setting = 30 MHz, GNSW = 0 V, unless otherwise noted. Table 1. Parameter FREQUENCY RESPONSE Low-Pass Corner Frequency, fC Step Size Corner Frequency Absolute Accuracy Corner Frequency Matching Pass-Band Ripple Gain Matching Group Delay Variation Corner Frequency = 1 MHz Corner Frequency = 30 MHz Group Delay Matching Corner Frequency = 1 MHz Corner Frequency = 30 MHz Stop-Band Rejection Relative to Pass Band INPUT STAGE Maximum Input Swing Differential Input Impedance Input Common-Mode Range GAIN CONTROL Voltage Gain Range Gain Slope Gain Error Gain Step OUTPUT STAGE Maximum Output Swing Differential Output Impedance Output DC Offset Output Common-Mode Range NOISE/DISTORTION 1 MHz Corner Frequency Output Noise Density Second Harmonic, HD2 Third Harmonic, HD3 Test Conditions/Comments Min Six-pole Butterworth filter, 0.5 dB bandwidth 1 Over operating temperature range Channel A and Channel B at same gain and bandwidth settings Channel A and Channel B at same gain and bandwidth settings From midband to peak Typ Max Unit 30 1 15 MHz MHz % fC 0.5 % fC 0.5 0.1 dB p-p dB 135 11 ns ns 5 0.2 ns ns 30 75 dB dB 1 400 V p-p V V Channel A and Channel B at same gain 2 x fC 5 x fC INP1, INM1, INP2, INM2 At minimum gain, VGAIN = 0 V 1 V p-p input voltage Input pins left floating GAIN, GNSW GNSW = 0 V, VGAIN from 0 V to 2 V GNSW = 5 V 1.5 -5 1 Gain = 0 dB at fC/2 Gain = 20 dB at fC/2 Gain = 40 dB at fC/2 250 kHz fundamental, 1.5 V p-p output voltage Gain = 0 dB Gain = 40 dB 250 kHz fundamental, 1.5 V p-p output voltage Gain = 0 dB Gain = 40 dB Rev. A | Page 3 of 32 +45 51 30 0.2 6 VGAIN from 500 mV to 1.7 V GNSW = 0 V to 5 V OPP1, OPM1, OPP2, OPM2, VOCM At maximum gain, RLOAD = 1 k HD2 > 60 dBc, HD3 > 60 dBc Inputs shorted, offset loop disabled 1.5 V p-p output voltage VOCM left floating VPS/2 VPS/2 2 1.5 20 35 dB dB mV/dB dB dB VPS/2 V p-p V p-p mV V V -129 -127 -111 dBV/Hz dBV/Hz dBV/Hz 46.2 43.2 dBc dBc 52.2 51.2 dBc dBc 1.5 3.0 ADRF6510 Parameter IMD3 IMD3 with Input CW Blocker 30 MHz Corner Frequency Output Noise Density Second Harmonic, HD2 Third Harmonic, HD3 IMD3 IMD3 with Input CW Blocker DIGITAL LOGIC Input High Voltage, VINH Input Low Voltage, VINL Input Current, IINH/IINL Input Capacitance, CIN SPI TIMING fCLK tDH tDS tLH tLS tPW tD POWER AND ENABLE Supply Voltage Range Total Supply Current Disable Current Disable Threshold Enable Response Time Disable Response Time Data Sheet Test Conditions/Comments f1 = 500 kHz, f2 = 550 kHz, 1.5 V p-p composite output voltage Gain = 5 dB Gain = 35 dB f1 = 500 kHz, f2 = 550 kHz, 1.5 V p-p composite output, gain = 5 dB; blocker at 5 MHz, 10 dBc relative to two-tone composite output voltage Min Midband, gain = 0 dB Midband, gain = 20 dB Midband, gain = 40 dB 8 MHz fundamental, 1.5 V p-p output voltage Gain = 0 dB Gain = 40 dB 8 MHz fundamental, 1.5 V p-p output voltage Gain = 0 dB Gain = 40 dB f1 = 15 MHz, f2 = 16 MHz, 1.5 V p-p composite output voltage Gain = 5 dB Gain = 35 dB f1 = 15 MHz, f2 = 16 MHz, 1.5 V p-p composite output, gain = 5 dB; blocker at 150 MHz, 10 dBc relative to two-tone composite output voltage LE, CLK, DATA, SDO, OFDS, GNSW LE, CLK, DATA, SDO 1/tCLK DATA hold time DATA setup time LE hold time LE setup time CLK high pulse width CLK to SDO delay VPS, VPSD, COM, COMD, ENBL 4.75 ENBL = 5 V Maximum bandwidth setting Minimum bandwidth setting ENBL = 0 V Delay following ENBL low-to-high transition Delay following ENBL high-to-low transition Rev. A | Page 4 of 32 Typ Max Unit 61 57 40 dBc dBc dBc -130 -130 -123 dBV/Hz dBV/Hz dBV/Hz 63 84 dBc dBc 54 87 dBc dBc 59 77.5 55 dBc dBc dBc >2 <0.8 <1 2 V V A pF 20 5 5 5 5 5 5 MHz ns ns ns ns ns ns 5.0 258 131 2 2.5 20 300 5.25 V mA mA mA V s ns Data Sheet ADRF6510 TIMING DIAGRAMS tPW tCLK CLK tLH tLS LE DATA tDH WRITE BIT LSB LSB + 1 MSB - 2 MSB - 1 MSB- 2 MSB NOTES 1. THE FIRST DATA BIT DETERMINES WHETHER THE PART IS WRITING TO OR READING FROM THE INTERNAL CORNER FREQUENCY WORD REGISTER. FOR A WRITE OPERATION, THE FIRST BIT SHOULD BE A LOGIC 1. THE CORNER FREQUENCY WORD BIT IS THEN REGISTERED INTO THE DATA PIN ON CONSECUTIVE RISING EDGES OF THE CLOCK. 09002-003 tDS Figure 2. Write Mode Timing Diagram tPW tCLK tD CLK tLS LE DATA SDO tDH READ BIT DC LSB DC DC LSB + 1 MSB - 2 DC MSB - 1 DC MSB NOTES 1. THE FIRST DATA BIT DETERMINES WHETHER THE PART IS WRITING TO OR READING FROM THE INTERNAL CORNER FREQUENCY WORD REGISTER. FOR A READ OPERATION, THE FIRST BIT SHOULD BE A LOGIC 0. THE CORNER FREQUENCY WORD BIT IS THEN UPDATED AT THE SDO PIN ON CONSECUTIVE FALLING EDGES OF THE CLOCK. Figure 3. Read Mode Timing Diagram Rev. A | Page 5 of 32 09002-004 tDS ADRF6510 Data Sheet ABSOLUTE MAXIMUM RATINGS Table 2. Parameter Supply Voltages, VPS, VPSD ENBL, GNSW, OFDS, LE, CLK, DATA, SDO INP1, INM1, INP2, INM2 OPP1, OPM1, OPP2, OPM2 OFS1, OFS2 GAIN Internal Power Dissipation JA (Exposed Pad Soldered to Board) Maximum Junction Temperature Operating Temperature Range Storage Temperature Range Lead Temperature (Soldering 60 sec) Rating 5.25 V VPS + 0.6 V VPS + 0.6 V, GND - 0.6 V VPS + 0.6 V VPS + 0.6 V VPS + 0.6 V 1.4 W 37.4C/W 150C -40C to +85C -65C to +150C 300C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. ESD CAUTION Rev. A | Page 6 of 32 Data Sheet ADRF6510 32 31 30 29 28 27 26 25 ENBL INP1 INM1 VPS COM GNSW OFS1 VPS PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 7 8 PIN 1 INDICATOR ADRF6510 TOP VIEW (Not to Scale) 24 23 22 21 20 19 18 17 OPP1 OPM1 COM GAIN VOCM COM OPM2 OPP2 NOTES 1. CONNECT THE EXPOSED PADDLE TO A LOW IMPEDANCE GROUND PAD. 09002-002 COM INP2 INM2 VPS COM OFDS OFS2 VPS 9 10 11 12 13 14 15 16 VPSD COMD LE CLK DATA SDO COM VPS Figure 4. Pin Configuration Table 3. Pin Function Descriptions Pin No. 1 2 3 4 5 6 7, 9, 13, 19, 22, 28 8, 12, 16, 25, 29 10, 11, 30, 31 14 15, 26 17, 18, 23, 24 20 21 27 32 Mnemonic VPSD COMD LE CLK DATA SDO COM VPS INP2, INM2, INM1, INP1 OFDS OFS2, OFS1 OPP2, OPM2, OPM1, OPP1 VOCM GAIN GNSW ENBL EP Description Digital Positive Supply Voltage: 4.75 V to 5.25 V. Digital Common. Connect to external circuit common using the lowest possible impedance. Latch Enable. SPI programming pin. CMOS levels: VLOW < 0.8 V, VHIGH > 2 V. SPI Port Clock. CMOS levels: VLOW < 0.8 V, VHIGH > 2 V. SPI Data Input. CMOS levels: VLOW < 0.8 V, VHIGH > 2 V. SPI Data Output. CMOS levels: VLOW < 0.8 V, VHIGH > 2 V. Analog Common. Connect to external circuit common. Analog Positive Supply Voltage: 4.75 V to 5.25 V. Differential Inputs. 400 input impedance. Common-mode range is 1.5 V to 2.5 V; default is 2.1 V. Offset Correction Loop Disable. Pull high to disable the offset correction loop. Offset Correction Loop Compensation Capacitors. Connect capacitors to circuit common. Differential Outputs. 20 output impedance. Common-mode range is 1.5 V to 3 V; default is VPS/2. Output Common-Mode Setpoint. Defaults to VPS/2 if left open. Analog Gain Control. 0 V to 2 V, 30 mV/dB gain scaling. Front-End Gain Switch, 6 dB or 12 dB. Pull low for 6 dB; pull high for 12 dB. Chip Enable. Pull high to enable. Exposed Paddle. Connect the exposed paddle to a low impedance ground pad. Rev. A | Page 7 of 32 ADRF6510 Data Sheet TYPICAL PERFORMANCE CHARACTERISTICS VPS = 5 V, TA = 25C, ZSOURCE = 400 , ZLOAD = 1 k, VOUT = 1.5 V p-p, GNSW = 0 V, unless otherwise noted. 2.0 1.5 30 1.0 GAIN ERROR (dB) 35 25 20 -40C VPS = 4.75V, 5V, 5.25V 15 +85C VPS = 4.75V, 5V, 5.25V 10 -1.0 0 -2.0 -5 -2.5 400 600 800 1000 1200 1400 1600 1800 2000 2200 -3.0 09002-005 200 VGAIN (mV) 0 -40C VPS = 4.25V, 5V, 5.25V 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 Figure 8. Gain Conformance vs. VGAIN over Supply and Temperature (Bandwidth Setting = 30 MHz) 9 5 BANDWIDTH = 30MHz BANDWIDTH = 30MHz 8 3 1 10 100 FREQUENCY (MHz) 7 -1 6 -3 5 PREAMP GAIN = 6dB -5 4 -7 3 -9 0 5 10 15 20 25 40 2 Figure 9. 6 dB Gain Step and Gain Error vs. Frequency (Bandwidth Setting = 30 MHz, VGAIN = 0 V) 0.25 15 PREAMP GAIN = 6dB PREAMP GAIN = 12dB BANDWIDTH = 30MHz 10 0.15 5 OP1dB (dBV rms) 0.20 0.10 0.05 0 -5 -10 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 VGAIN (V) Figure 7. Gain Matching vs. VGAIN (Bandwidth Setting = 30 MHz) -15 0 5 10 15 20 GAIN (dB) 25 30 35 40 09002-010 0 09002-007 GAIN MISMATCH (dB) 35 FREQUENCY (MHz) Figure 6. Gain vs. Frequency by VGAIN (Bandwidth Setting = 30 MHz) -0.05 30 GAIN STEP (dB) 1 09002-009 GAIN ERROR (dB) PREAMP GAIN = 12dB 09002-006 50 45 40 35 30 25 20 15 10 5 0 -5 -10 -15 -20 -25 -30 -35 -40 -45 -50 +25C VPS = 4.25V, 5V, 5.25V VGAIN (mV) Figure 5. In-Band Gain vs. VGAIN over Supply and Temperature (Bandwidth Setting = 30 MHz) GAIN (dB) 0 -0.5 -1.5 0 +85C VPS = 4.25V, 5V, 5.25V 0.5 5 -10 BANDWIDTH = 30MHz 2.5 +25C VPS = 4.75V, 5V, 5.25V 40 GAIN (dB) 3.0 BANDWIDTH = 30MHz 45 09002-008 50 Figure 10. Output P1dB vs. GAIN at 15 MHz (Bandwidth Setting = 30 MHz) Rev. A | Page 8 of 32 ADRF6510 40 1000 35 900 30 800 25 700 20 15 10 5 500 400 -5 100 FREQUENCY (Hz) BW = 10MHz BW = 20MHz BW = 30MHz 0 0.5 09002-011 100M BW = 5MHz 300 200 10M BW = 1MHz 600 0 -10 1M GAIN = 20dB 5 50 FREQUENCY (MHz) Figure 11. Frequency Response vs. Bandwidth Setting (Gain = 30 dB), Log Scale 09002-014 GROUP DELAY (ns) GAIN (dB) Data Sheet Figure 14. Group Delay vs. Frequency (Gain = 20 dB) 32 1.0 BANDWIDTH = 30MHz 30 29 1 6 11 16 21 26 31 36 40 FREQUENCY (MHz) GAIN = 20dB 0 GAIN = 40dB -0.5 -1.0 0.5 Figure 12. Frequency Response vs. Bandwidth Setting (Gain = 30 dB), Linear Scale 30 29 5 IQ GROUP DELAY MATCHING (ns) 4 +25C 27 26 25 24 +85C 23 22 21 20 19 18 17 BANDWIDTH = 1MHz 3 GAIN = 20dB 2 1 0 -1 GAIN = 0dB -2 -3 -4 16 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 FREQUENCY (MHz) 09002-013 GAIN (dB) 30 Figure 15. Group Delay Mismatch vs. Frequency (Bandwidth Setting = 30 MHz) -40C 28 15 5 FREQUENCY (MHz) Figure 13. Frequency Response over Temperature (Gain = 26 dB, Bandwidth Setting = 30 MHz) -5 0.2 0.4 0.6 0.8 1.0 1.2 FREQUENCY (MHz) Figure 16. IQ Group Delay Mismatch vs. Frequency (Bandwidth Setting = 1 MHz) Rev. A | Page 9 of 32 1.4 09002-016 27 09002-012 28 0.5 09002-015 GROUP DELAY MATCHING (ns) GAIN (dB) 31 ADRF6510 0 Data Sheet FREQUENCY (MHz) 1.0 1.5 2.0 0.5 100 2.5 3.0 0.75 90 BANDWIDTH = 1MHz 0.25 BANDWIDTH = 30MHz 0 80 70 +25C, VPS = 4.75V +25C, VPS = 5V +25C, VPS = 5.25V +85C, VPS = 4.75V +85C, VPS = 5V +85C, VPS = 5.25V -40C, VPS = 4.75V -40C, VPS = 5V -40C, VPS = 5.25V 60 -0.25 50 -0.50 40 0 5 10 15 20 30 25 FREQUENCY (MHz) 0 20 25 30 35 40 100 1.5V p-p OUTPUT @ 8MHz BANDWIDTH = 30MHz 90 80 HD3 @ 24MHz (dBc) 75 70 65 +25C, VPS = 4.75V +25C, VPS = 5V +25C, VPS = 5.25V +85C, VPS = 4.75V +85C, VPS = 5V +85C, VPS = 5.25V -40C, VPS = 4.75V -40C, VPS = 5V -40C, VPS = 5.25V 55 50 45 0 5 10 15 20 25 30 35 70 60 VOCM = 1.5V VOCM = 1.75V VOCM = 2V VOCM = 2.5V 50 40 GAIN (dB) 40 09002-018 60 80 0 5 10 15 20 25 30 35 40 GAIN (dB) Figure 18. HD2 vs. Gain over Supply and Temperature (Bandwidth Setting = 30 MHz) 09002-021 90 Figure 21. HD3 vs. Gain over Output Common-Mode Voltage (Bandwidth Setting = 30 MHz) 100 30 1.5V p-p OUTPUT @ 8MHz BANDWIDTH = 30MHz 90 BANDWIDTH = 30MHz f1 = 14MHz, f2 = 15MHz 25 PREAMP GAIN = 6dB 80 OIP3 (dBV) 20 70 60 15 PREAMP GAIN = 12dB 10 50 VOCM = 1.5V VOCM = 1.75V VOCM = 2V VOCM = 2.5V 30 0 5 10 15 20 25 30 35 5 40 GAIN (dB) 09002-019 40 Figure 19. HD2 vs. Gain over Output Common-Mode Voltage (Bandwidth Setting = 30 MHz) 0 0 5 10 15 20 25 30 35 40 45 50 GAIN (dB) Figure 22. In-Band OIP3 vs. Gain (Bandwidth Setting = 30 MHz) Rev. A | Page 10 of 32 09002-022 HD2 @ 16MHz (dBc) 15 Figure 20. HD3 vs. Gain over Supply and Temperature (Bandwidth Setting = 30 MHz) 85 HD2 @ 16MHz (dBc) 10 GAIN (dB) Figure 17. IQ Amplitude Mismatch vs. Frequency 40 5 09002-017 -0.75 09002-020 HD3 @ 24MHz (dBc) IQ AMPLITUDE MISMATCH (dB) 0.50 ADRF6510 30 BANDWIDTH = 30MHz f1 = 14MHz, f2 = 15MHz 25 -40C OIP3 (dBV) IMD2 @ 15MHz (dBV) +25C 20 +85C 15 10 0 0 5 10 15 20 25 30 35 40 GAIN (dB) 09002-023 5 65 55 BANDWIDTH = 30MHz 45 35 25 15 5 -5 2:1 SLOPE -15 -25 -35 -45 PREAMP -55 GAIN = 12dB -65 OUT-OF-BAND IIP2 -75 -85 -95 PREAMP -105 GAIN = 6dB -115 -125 -135 -145 -45 -35 -25 -15 -5 55 5 15 25 35 45 Figure 23. In-Band OIP3 vs. Gain over Temperature (Preamp Gain = 6 dB, Bandwidth Setting = 30 MHz) Figure 26. Out-of-Band IIP2: IMD2 Tone at Midband (Bandwidth Setting = 30 MHz) 120 10 GAIN = 0dB GAIN = 10dB GAIN = 20dB GAIN = 30dB GAIN = 40dB BANDWIDTH = 30MHz 0 -10 -20 IMD3 @ 15MHz (dBV) 100 80 IMD3 (dBc) 65 INPUT LEVEL @ 115MHz AND 130MHz (dBV/Tone) 09002-060 Data Sheet 60 40 -30 -40 3:1 SLOPE -50 -60 PREAMP GAIN = 12dB -70 OUT-OF-BAND IIP3 -80 -90 -100 20 PREAMP GAIN = 6dB -110 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 COMPOSITE OUTPUT VOLTAGE (V p-p) -130 -45 09002-024 0 -35 -30 -25 -20 -15 -10 -5 0 5 INPUT LEVEL @ 115MHz AND 215MHz (dBV/Tone) Figure 24. In-Band Third-Order Intermodulation Distortion (Preamp Gain = 6 dB, Bandwidth Setting = 30 MHz) Figure 27. Out-of-Band IIP3: IMD3 Tone at Midband (Bandwidth Setting = 30 MHz) 120 70 GAIN = 6dB GAIN = 16dB GAIN = 26dB GAIN = 36dB GAIN = 46dB 1MHz BW 2MHz BW 4MHz BW 6MHz BW 8MHz BW 30MHz BW 65 60 NOISE FIGURE (dB re 50) 100 80 60 40 55 50 45 40 35 30 20 0 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 COMPOSITE OUTPUT VOLTAGE (V p-p) Figure 25. In-Band Third-Order Intermodulation Distortion (Preamp Gain = 12 dB, Bandwidth Setting = 30 MHz) 20 -5 0 5 10 15 20 25 30 35 40 GAIN (dB) Figure 28. Noise Figure vs. Gain over Bandwidth Setting, Preamp Gain = 6 dB (Noise Figure at 1/2 Bandwidth) Rev. A | Page 11 of 32 45 09002-028 25 09002-025 IMD3 (dBc) -40 09002-027 -120 0 ADRF6510 Data Sheet 60 -90 1MHz BW 2MHz BW 4MHz BW 6MHz BW 8MHz BW 30MHz BW 45 40 35 30 25 20 10 15 20 25 30 35 40 45 50 Figure 29. Noise Figure vs. Gain over Bandwidth Setting, Preamp Gain = 12 dB (Noise Figure at 1/2 Bandwidth) -120 -125 -130 1.0 1.5 2.0 2.5 3.0 FREQUENCY (MHz) Figure 32. Output Noise Density vs. Frequency (Bandwidth Setting = 1 MHz) -110 1MHz BW 2MHz BW 4MHz BW 6MHz BW 8MHz BW 30MHz BW -110 GAIN = 0dB GAIN = 20dB GAIN = 40dB -115 OUTPUT NOISE (dBV rms/Hz) -105 -115 -120 -125 -120 -125 -130 -135 -130 0 5 10 15 20 25 30 35 40 45 GAIN (dB) -140 10 09002-030 -135 -5 Figure 30. Output Noise Density vs. Gain by Bandwidth Setting, Preamp Gain = 6 dB (Noise at 1/2 Bandwidth) 12 14 16 18 20 22 24 26 28 30 FREQUENCY (MHz) 09002-033 OUTPUT NOISE (dBV rms/Hz) -115 -140 0.5 -100 Figure 33. Output Noise Density vs. Frequency (Bandwidth Setting = 20 MHz) -100 -90 1MHz BW 2MHz BW 4MHz BW 6MHz BW 8MHz BW 30MHz BW -110 -95 OUTPUT NOISE @ 15MHz (dBV/Hz) -105 -115 -120 -125 -130 -100 -105 GAIN = 40dB -110 -115 -120 GAIN = 20dB -125 -130 GAIN = 0dB -135 0 5 10 15 20 25 30 35 40 45 50 GAIN (dB) 09002-031 OUTPUT NOISE (dBV rms/Hz) -110 09002-032 5 09002-029 0 GAIN (dB) -135 -105 -135 15 10 -100 Figure 31. Output Noise Density vs. Gain by Bandwidth Setting, Preamp Gain = 12 dB (Noise at 1/2 Bandwidth) -140 -35 -30 -25 -20 -15 BLOCKER LEVEL @ 150MHz (dBV rms) Figure 34. Output Noise Density vs. Blocker Level (Bandwidth Setting = 30 MHz, Blocker at 150 MHz) Rev. A | Page 12 of 32 -10 09002-034 NOISE FIGURE (dB re 50) 50 GAIN = 0dB GAIN = 20dB GAIN = 40dB -95 OUTPUT NOISE (dBV rms/Hz) 55 ADRF6510 450 10 280 440 5 260 430 0 240 420 -5 410 -10 400 -15 390 -20 380 -25 370 -30 360 -35 120 -40 100 10 15 20 25 30 ISUPPLY (mA) CIN (pF) 160 140 FREQUENCY (MHz) 200 266 27 100 264 26 0 25 -100 24 -200 23 -300 22 -400 21 -500 250 -600 248 -50 5 10 15 20 25 30 20 25 30 ISUPPLY (mA) LSERIES_OUT (nH) 260 258 256 254 252 FREQUENCY (MHz) VPS = 4.75V VPS = 5V VPS = 5.25V -30 -10 10 30 50 70 90 TEMPERATURE (C) Figure 36. Output Impedance vs. Frequency (Bandwidth Setting = 30 MHz) Figure 39. Current Consumption vs. Temperature over Supply (Bandwidth Setting = 30 MHz) 90 70 BANDWIDTH = 30MHz BANDWIDTH = 30MHz 80 60 70 GAIN = 40dB GAIN = 40dB 50 GAIN = 20dB 50 CMRR (dB) 60 GAIN = 20dB 40 40 30 GAIN = 0dB 30 20 20 10 10 0 0.1 1 10 100 FREQUENCY (MHz) 09002-037 ISOLATION (dB) 15 262 09002-036 0 10 Figure 38. Current Consumption vs. Bandwidth Setting (Gain = 20 dB) 28 20 5 BANDWIDTH SETTING (MHz) Figure 35. Input Impedance vs. Frequency (Bandwidth Setting = 30 MHz) RSERIES_OUT () 0 09002-038 5 180 09002-039 0 200 Figure 37. Channel Isolation, Output to Output, vs. Frequency (Bandwidth Setting = 30 MHz) 0 0 5 10 15 20 25 FREQUENCY (MHz) Figure 40. Common-Mode Rejection Ratio vs. Frequency (Bandwidth Setting = 30 MHz) Rev. A | Page 13 of 32 30 09002-040 350 220 09002-035 RIN () Data Sheet ADRF6510 Data Sheet 28MHz SIGNAL = 45mV p-p TO 450mV p-p VGAIN = 1400mV TO 800mV INPUT SIGNAL = 45mV p-p OUTPUT SIGNAL = 450mV p-p 20dB OF GAIN VGAIN = 800mV TO 1.40V 400ns/DIV 09002-067 30MHz BANDWIDTH, 28MHz MAINTONE Figure 41. Gain Step Response Rev. A | Page 14 of 32 Data Sheet ADRF6510 THEORY OF OPERATION band rejection, ripple, and group delay. The 0.5 dB bandwidth is programmed from 1 MHz to 30 MHz in 1 MHz steps via the serial programming interface (SPI) as described in the Programming the Filters section. The ADRF6510 consists of a matched pair of buffered, programmable filters followed by variable gain amplifiers and output ADC drivers. The block diagram of a single channel is shown in Figure 42. The programmability of the bandwidth and of the pre- and post-filtering gain offers great flexibility when coping with signals of varying levels in the presence of noise and large, undesired signals nearby. The entire differential signal chain is dc-coupled with flexible interfaces at the input and output. The bandwidth and gain setting controls for the two channels are shared, ensuring close matching of their magnitude and phase responses. The ADRF6510 can be fully disabled through the ENBL pin. 50dB VGA where: is the phase in radians. = 2f is the frequency in radians/second. OUTPUT ADC DRIVER BASEBAND INPUTS BASEBAND OUTPUTS Note that for a frequency scaled filter prototype, the absolute magnitude of the group delay scales inversely with the bandwidth; however, the shape is retained. For example, the peak group delay for a 28 MHz bandwidth setting is 14x less than for a 2 MHz setting. PREAMP GAIN SWITCH FILTER PROGRAMMING SPI BUS ANALOG GAIN CONTROL 30mV/dB OUTPUT COMMON-MODE CONTROL 09002-042 SPI INTERFACE 0 -20 Filtering and amplification are fundamental operations in any signal processing system. Filtering is necessary to select the intended signal while rejecting out-of-band noise and interferers. Amplification increases the level of the desired signal to overcome noise added by the system. When used together, filtering and amplification can extract a low level signal of interest in the presence of noise and out-of-band interferers. Such analog signal processing alleviates the requirements on the analog, mixed signal, and digital components that follow. RELATIVE MAGNITUDE (Hz) Figure 42. Signal Path Block Diagram for a Single Channel of the ADRF6510 -40 -60 -80 -100 -120 -140 -160 -180 1M INPUT BUFFERS 10M 100M 1G FREQUENCY (Hz) (VCM - (VPS/2))/200 The input buffers in both channels can be configured simultaneously to a gain of 6 dB or 12 dB through the GNSW pin. When configured for a 6 dB gain, the buffers support up to a 1 V p-p differential input level with >50 dBc harmonic distortion. For a 12 dB gain setting, the buffers support 0.5 V p-p inputs. Figure 43. Sixth-Order Butterworth Magnitude Response for 0.5 dB Bandwidths; Programmed from 2 MHz to 29 MHz in 1 MHz Steps 500 400 2MHz GROUP DELAY (ns) The input buffers provide a convenient interface to the sensitive filter sections that follow. They set a differential input impedance of 400 and sit at a nominal common-mode voltage of VPS/2. The inputs can be dc-coupled or ac-coupled. If using direct dc-coupling, the common-mode voltage, VCM, can range from 1.5 V to 3 V. A current flows into or out of the input pins to accommodate the difference in common-mode voltages. The current into each pin is given by 09002-043 1MHz TO 30MHz PROG. FILTERS g = -/ PROGRAMMABLE FILTERS 28MHz 300 200 14x 100 0 -100 100k 1M 10M 100M FREQUENCY (Hz) The integrated programmable filter is the key signal processing function in the ADRF6510. The filters follow a six-pole Butterworth prototype response that provides a compromise between Rev. A | Page 15 of 32 Figure 44. Sixth-Order Butterworth Group Delay Response for 0.5 dB Bandwidths; Programmed to 2 MHz and 28 MHz 09002-044 6dB/12dB PREAMP The filters are designed so that the Butterworth prototype filter shape and group delay responses vs. frequency are retained for any bandwidth setting. Figure 43 and Figure 44 illustrate the ideal six-pole Butterworth gain and group delay responses, respectively. The group delay, g, is defined as ADRF6510 Data Sheet The corner frequency of the filters is defined by RC products, which can vary by 30% in a typical process. Therefore, all the parts are factory calibrated for corner frequency, resulting in a residual 10% corner frequency variation over the -40C to +85C temperature range. Although absolute accuracy requires calibration, the matching of RC products between the pair of channels is better than 1% by observing careful design and layout practices. Calibration and excellent matching ensure that the magnitude and group delay responses of both channels track together, a critical requirement for digital IQ-based communication systems. VARIABLE GAIN AMPLIFIERS (VGAs) The VGAs are implemented using the Analog Devices, Inc., patented X-AMP(R) architecture, consisting of a tapped 50 dB attenuator followed by a fixed-gain amplifier. The X-AMP architecture generates a linear-in-dB monotonic gain response with low ripple. The gain is controlled through the high impedance GAIN pin with an accurate slope of 30 mV/dB. The gain response shown in Figure 45 shows the GAIN pin voltage range and the absence of gain foldback at high VGAIN. 50 DC OFFSET COMPENSATION LOOP In many signal processing applications, no information is carried in the dc level. In fact, dc voltages and other low frequency disturbances can often dominate the intended signal and consume precious dynamic range in the analog path and bits in the data converters. These dc voltages can be present with the desired input signal or can be generated inside the signal path by inherent dc offsets or other unintended signaldependent processes such as self-mixing or rectification. Because the ADRF6510 is fully dc-coupled, it may be necessary to remove these offsets to realize the maximum signal-to-noise ratio (SNR). This can be achieved with ac-coupling capacitors at the input and output pins, but that requires large values because the impedances are fairly low, and high-pass corners may need to be <10 Hz in some cases. To address the issue of dc offsets, the ADRF6510 provides an offset correction loop that nulls the output differential dc level as shown in Figure 46. If the correction loop is not required, it can be disabled through the OFDS pin. COFS OFDS 0.3 OFSx 0.2 30 0.1 0 20 -0.1 10 BASEBAND OUTPUT FROM FILTERS 50dB VGA OUTPUT ADC DRIVER GAIN 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 -0.3 4.0 VGAIN (V) 09002-045 -10 Figure 46. Offset Compensation Loop Operation around the VGA and Output Buffer -0.2 0 09002-066 40 GAIN ERROR (dB) GAIN (dB) 30mV/dB Figure 45. Linear-in-dB Gain Control Response of the X-AMP VGA Showing Consistent Slope and Low Error OUTPUT BUFFERS/ADC DRIVERS The low impedance (20 ) output buffers of the ADRF6510 are designed to drive either ADC inputs or subsequent amplifier stages. They are capable of delivering up to 4 V p-p composite two-tone signals into 500 differential loads with >60 dBc IM3. The output common-mode voltage defaults to VPS/2, but it can be adjusted from 1.5 V to 3.0 V without loss of drive capability by presenting the VOCM pin with the desired common-mode voltage. The high input impedance of VOCM allows the ADC reference output to be connected directly. Even though the signal path is fully dc-coupled and the dc offset compensation loop can remove undesired dc offsets (see the DC Offset Compensation Loop section), the output buffers can be accoupled to the next stage by properly selecting the coupling capacitors according to the load impedance. The offset control loop creates a high-pass corner, fHP, that is superimposed on the normal Butterworth filter response. Typically, fHP is many orders of magnitude lower than the lower programmed filter bandwidth so that there is no interaction between them. Setting fHP is accomplished with capacitors, COFS, from the OFS1 and OFS2 pins to ground. Because the correction loop works around the VGA section, fHP is also dependent on the gain of the VGA. In general, the expression for fHP is given by fHP (Hz) = 1.2 x (Gain/COFS) where: Gain is expressed in linear terms, not in decibels (dB). COFS is expressed in microfarads (F). Note that fHP increases in proportion to the gain. For this reason, COFS should be chosen at the highest operating gain to guarantee that fHP is always below the maximum limit required by the system. Rev. A | Page 16 of 32 Data Sheet ADRF6510 PROGRAMMING THE FILTERS NOISE CHARACTERISTICS The 0.5 dB corner frequencies for both filters are programmed simultaneously through the SPI port. A 5-bit register stores the codes for corner frequencies of 1 MHz through 30 MHz (see Table 4). The SPI protocol not only allows frequency codes to be written to the DATA pin but also allows the stored code to be read back from the SDO pin. The output noise behavior of the ADRF6510 depends on the gain and bandwidth settings. Both the filter sections and the VGAs contribute to the total noise at the output. The filter contributes a noise spectral density profile that is flat at low frequencies, peaks near the corner frequency, and then rolls off as the filter poles roll off the gain. The magnitude of the noise spectral density, expressed in nV/Hz, varies inversely with the square root of the bandwidth setting, resulting in a total integrated noise in nV that is nearly constant with bandwidth setting. The latch enable (LE) pin must first go to a Logic 0 for a read or write cycle to begin. On the next rising edge of the clock (CLK), a Logic 1 on the DATA pin initiates a write cycle, whereas a Logic 0 on the DATA pin initiates a read cycle. In a write cycle, the next five CLK rising edges latch the frequency code, LSB first. When LE goes high, the write cycle is completed and the frequency code is presented to the filter. In a read cycle, the next five CLK falling edges present the stored frequency code, LSB first. When LE goes high, the read cycle is completed. Detailed timing diagrams are shown in Figure 2 and Figure 3. Table 4. Frequency Code vs. Corner Frequency Lookup Table 1 -115 BANDWIDTH = 20MHz GAIN = 40dB -120 -125 GAIN = 20dB -130 GAIN = 0dB -135 10 15 20 25 30 35 40 45 50 55 60 FREQUENCY (MHz) Figure 47. Total Output Noise with a 20 MHz Corner Frequency for Three Different Gain Settings MSB first. Rev. A | Page 17 of 32 09002-046 Corner Frequency (MHz) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 30 30 Figure 47 and Figure 48 show the total output noise spectral density vs. frequency for different bandwidth settings. At low values of VGA gain, the noise at the output is the flat spectral density contributed by the VGA because the filter noise is suppressed by the VGA attenuation. As the gain increases, more of the filter noise appears at the output. Because the filter noise increases at lower bandwidth settings, it overwhelms the VGA noise floor. In either case, the noise density asymptotically approaches the -130 dBV/Hz limit set by the VGA at the highest frequencies. For other values of VGA gain and bandwidth setting, the detailed shape of the noise spectral density changes. OUTPUT NOISE (dBV/Hz) 5-Bit Binary Frequency Code1 00000 00001 00010 00011 00100 00101 00110 00111 01000 01001 01010 01011 01100 01101 01110 01111 10000 10001 10010 10011 10100 10101 10110 10111 11000 11001 11010 11011 11100 11101 11110 11111 The X-AMP type VGAs used in the ADRF6510 contribute a fixed noise spectral density to the output, independent of the gain setting, of -130 dBV/Hz, which is equivalent to 316 nV/Hz. Although the VGA noise contribution to the output is fixed, the gain of the VGA controls the relative contribution of the filter noise. ADRF6510 Data Sheet -100 To distinguish and quantify the distortion performance of the input section, two different IP3 specifications are presented. The first is called in-band IP3 and refers to a two-tone test where the signals are inside the filter bandwidth. This is exactly the same figure of merit familiar to communications engineers in which the third-order intermodulation level, IM3, is measured. BANDWIDTH = 1MHz OUTPUT NOISE (dBV/Hz) -105 GAIN = 40dB -110 -115 -120 GAIN = 20dB -125 -130 1.0 1.5 2.0 2.5 3.0 FREQUENCY (MHz) 09002-047 GAIN = 0dB -135 0.5 Figure 48. Total Output Noise with a 1 MHz Corner Frequency for Three Different Gain Settings Note that the noise spectral density outside the filter bandwidth is limited by the fixed VGA output noise. It may be necessary to use an external, fixed-frequency, passive filter prior to an analogto-digital conversion to prevent noise aliasing from degrading the signal-to-noise ratio. The higher the sampling rate relative to the maximum ADRF6510 corner frequency setting to be used, the lower the order of the external filter. DISTORTION CHARACTERISTICS The distortion performance of the ADRF6510 is similar to its noise performance. The filters and the VGAs contribute to the overall distortion and signal handling capabilities. Furthermore, the front end must also cope with out-of-band signals that can be larger than the in-band signals. These out-of-band signals are filtered before reaching the VGA. It is important to understand the signals presented to the ADRF6510 and to match these signals with the input and output characteristics of the part. When the gain is low, the distortion is typically limited by the input section because the output is not driven to its maximum capacity. When the gain is high, the distortion is likely limited by the output section because the input is not driven to its maximum capacity. An exception to this is when the input is driven with a small desired signal in combination with a large out-of-band signal. In this case, the out-of-band signal may drive the input to distort. As long as the input is not overdriven, the out-of-band signal is removed by the filter. A high VGA gain is still needed to raise the small desired signal to a higher level at the output. The overall distortion introduced by the part depends on the input drive level, including the out-of-band signals, and the desired output signal level. As noted in the Input Buffers section, the input section can handle a total signal level of 1 V p-p for a 6 dB preamplifier and 500 mV p-p for a 12 dB preamplifier with >50 dBc harmonic distortion. This includes both in-band and out-of-band signals. To quantify the effect of out-of-band signals, a new out-of-band (OOB) IIP3 figure of merit is introduced. This test also involves a two-tone stimulus; however, the two tones are placed out-ofband so that the lower IM3 product lands in the middle of the filter pass band. At the output, only the IM3 product is visible because the original two tones are filtered out. To calculate the OOB IP3 at the input, the IM3 level is referred to the input by the overall gain. The OOB IIP3 allows the user to predict the impact of out-of-band blockers or interferers at an arbitrary signal level on the in-band performance. The ratio of the desired input signal level to the input-referred IM3 at a given blocker level represents a signal-to-distortion limit imposed by the out-of-band signals. MAXIMIZING THE DYNAMIC RANGE The role of the ADRF6510 is to increase the level of a variable in-band signal while minimizing out-of-band signals. Ideally, this is achieved without degrading the SNR of the incoming signal or introducing distortion to the incoming signal. The first goal is to maximize the output signal swing, which can be defined by the ADC input range or the input signal capacity of the next analog stage. For the complex waveforms often encountered in communication systems, the peak-to-average ratio, or crest factor, must be considered when choosing the peak-to-peak output. From the chosen output signal and the maximum gain of the ADRF6510, the minimum input level can be defined. Lower signal levels do not yield the maximum output and suffer a greater degradation in SNR. As the input signal level increases, the VGA gain is reduced from its maximum gain point to maintain the desired fixed output level. The output noise, initially dominated by the filter, follows the gain reduction, yielding a progressively better SNR. At some point, the VGA gain drops sufficiently that the constant VGA noise becomes dominant, resulting in a constant SNR from that point. From the perspective of SNR alone, the maximum input level is reached when the VGA reaches its minimum gain. Distortion must also be considered when maximizing the dynamic range. At low and moderate signal levels, the output distortion is constant and assumed to be adequate for the selected output level. At some point, the input signal becomes large enough that distortion at the input limits the system. The maximum tolerable input signal depends on whether the input distortion becomes unacceptably large or the minimum gain is reached. Rev. A | Page 18 of 32 Data Sheet ADRF6510 The most challenging scenario in terms of dynamic range is the presence of a large out-of-band blocker accompanying a weaker in-band wanted signal. In this case, the maximum input level is dictated by the blocker and its inclination to cause distortion. After filtering, the weak wanted signal must be amplified to the desired output level, possibly requiring maximum gain. Both the distortion limits associated with the blocker at the input and the SNR limits created by the weaker signal and higher gains are present simultaneously. Furthermore, not only does the blocker scenario degrade the dynamic range but it also reduces the range of input signals that can be handled because a larger part of the gain range is used to simply extract the weak desired signal from the stronger blocker. KEY PARAMETERS FOR QUADRATURE-BASED RECEIVERS The majority of digital communication receivers makes use of quadrature signaling, in which bits of information are encoded onto pairs of baseband signals that then modulate in-phase (I) and quadrature (Q) sinusoidal carriers. Both the baseband and modulated signals appear quite complex in the time domain with dramatic peaks and valleys. In a typical receiver, the goal is to recover the pair of quadrature baseband signals in the presence of noise and interfering signals after quadrature demodulation. In the process of filtering out-of-band noise and unwanted interferers and restoring the levels of the wanted I and Q baseband signals, it is critical to retain their gain and phase integrity over the bandwidth. The ADRF6510 delivers flat in-band gain and group delay, consistent with a six-pole Butterworth prototype filter as described in the Programmable Filters section. Furthermore, careful design ensures excellent matching of these parameters between the I and Q channels. Although absolute gain flatness and group delay can be corrected with digital equalization, mismatch introduces quadrature errors and intersymbol interference that degrade bit error rates in digital communication systems. Rev. A | Page 19 of 32 ADRF6510 Data Sheet APPLICATIONS INFORMATION BASIC CONNECTIONS ERROR VECTOR MAGNITUDE (EVM) PERFORMANCE Figure 49 shows the basic connections for operating the ADRF6510. A voltage from 4.75 V to 5.25 V should be applied to the supply pins. Each supply pin should be decoupled with at least one low inductance, surface-mount ceramic capacitor of 0.1 F placed as close as possible to the device. Error vector magnitude (EVM) is a measure used to quantify the performance of a digital radio transmitter or receiver by measuring the fidelity of the digital signal transmitted or received. Various imperfections in the link, such as magnitude and phase imbalance, noise, and distortion, cause the constellation points to deviate from their ideal locations. The input buffers provide an interface to the sensitive filter sections that follow. They set a differential input impedance of 400 and sit at a nominal common-mode voltage of VPS/2. The inputs can be dc-coupled or ac-coupled. If using direct dc-coupling, the common-mode voltage, VCM, can range from 1.5 V to 3 V. In general, a receiver exhibits three distinct EVM limitations vs. received input signal power. As signal power increases, the distortion components increase. * The output buffers of the ADRF6510 are low impedance (~20 ) designed to drive either ADC inputs or subsequent amplifier stages. The output common-mode voltage defaults to VPS/2 but can be adjusted from 1.5 V to 3.0 V without loss of drive capability by presenting the VOCM pin with the desired common-mode voltage. The high input impedance of VOCM allows the ADC reference output to be connected directly. At large enough signal levels, where the distortion components due to the harmonic nonlinearities in the device dominate, EVM degrades as signal levels increase. At medium signal levels, where the signal chain behaves in a linear manner and the signal is well above any notable noise contributions, EVM has a tendency to reach an optimal level determined dominantly by either the quadrature accuracy and I/Q gain match of the signal chain or the precision of the test equipment. As signal levels decrease, such that noise is a major contributor, EVM performance vs. the signal level exhibits a decibel-for-decibel degradation with decreasing signal levels. At these lower signal levels, where noise is the dominant limitation, decibel EVM is directly proportional to the SNR. * * To enable the ADRF6510, the ENBL pin must be pulled high. Taking ENBL low disables the device, reducing current consumption to approximately 2 mA at ambient temperature. VPS VPS INM1 INP1 VPS VPS OPP1 OPM1 ENBL INM1 OFS1 COM GNSW VPS VPS INP1 OPP1 VPSD VPSD OPM1 COMD LE CLK DATA SDO COM LE CLK VPS ADRF6510 GAIN DATA VOCM SDO COM COM OPM2 VPS VPS OPM2 OPP2 VPS OFDS VPS INP2 OFS2 COM INM2 COM VPS OPP2 VPS INP2 VPS Figure 49. Basic Connections Rev. A | Page 20 of 32 09002-050 VPS INM2 Data Sheet ADRF6510 EVM The basic setup to test EVM for the ADRF6510 consisted of an Agilent E4438C used as a RF signal source with an Agilent InfiniiVision DSO7104B oscilloscope in conjunction with the Agilent 89600 VSA software to sample the signal and compute the EVM. The E4438C RF output drove the RF port of the ADL5380 IQ demodulater, which in turn drove the baseband differential inputs of the ADRF6510. While low-pass filtering with the ADRF6510 to reject out-ofband undesired signals (blockers), more rejection of the undesired signals may be required. If the filter bandwidth is set to approximately the same as the signal bandwidth, the user may trade some degradation of EVM for a gain in rejection of the out-of-band undesired signals, by lowering the low-pass filter bandwidth corner (for example, by 1 MHz). The I and Q outputs of the ADRF6510 were taken differentially into two AD8130 difference amplifiers to convert them into single-ended signals. The single-ended signals were connected to the input channels of the oscilloscope, which captured the modulated waveforms. Lowering the filter bandwidth to gain more rejection works progressively better the lower the signal and filter bandwidths are set to (see Figure 43). A 1 MHz change from 3 MHz filter bandwidth to 2 MHz filter bandwidth yields about 20 dB more rejection. Compare that to a 1 MHz change from 29 MHz filter band-width to 28 MHz filter bandwidth, which will yield about 1 dB more in rejection. An overall baseband EVM performance was measured on the ADRF6510. A modulation setting of 4 QAM and, unless otherwise noted, a 5 MHz symbol rate were used, with a pulse shaping filter alpha of 0.35. The analog gain of the ADRF6510 was adjusted to maintain 1.5 V p-p into a 1 k differential load impedance. Figure 50 shows EVM vs. input power for three different IF frequencies. The input power is the integrated input power over the bandwidth of the modulated signal. In Figure 50, the ADRF6510 shows excellent EVM of better than --35 dB over a 50 dB range at a 0Hz IF. The user can chose to use a complex IF of 5 MHz to achieve even a better EVM of at least -40 dB over a 50 dB range. Figure 51 shows that degradation of EVM as signal bandwidth (positive frequency only) is swept while keeping the filter bandwidth set to 5 MHz. Three different COFS capacitor values were used. 0 -10 COFS = 1nF EVM (dB) -15 0 -20 -5 -25 -10 -30 -15 COFS = 100nF -35 -20 0 2 4 6 8 10 SIGNAL BANDWIDTH (MHz) -30 14 Figure 51. EVM vs. Signal Bandwidth over COFS Values While Maintaining a Filter Bandwidth of 5 MHz 2.5MHz IF -35 12 09002-151 COFS = 1F -40 -25 -40 0Hz IF EFFECT OF OUTPUT VOLTAGE LEVELS ON EVM -45 5MHz IF Output voltage level can affect EVM greatly when the signal is compressed. When changing the output voltage levels of the ADRF6510, take care that the output signal is not in compression, which causes EVM degradation. -50 -80 -70 -60 -50 -40 -30 -20 -10 PIN (dBm) 0 10 09002-058 EVM (dB) FILTER BW CORNER -5 Figure 50. EVM vs. RF Input Power Level; OFDS Pulled Low, COFS = 1 F EFFECT OF FILTER BANDWIDTH ON EVM Care should be taken when selecting the filter bandwidth. In a digital transceiver, the modulated signal is filtered by a pulse shaping filter (such as a root-raised cosine filter) at both the transmit and receive ends to guard against intersymbol interference (ISI). If additional filtering of the modulated signal is done, the signal must be within the pass band of the filter. When the corner frequency of the ADRF6510 filter begins to encroach on the modulated signal, ISI is introduced and degrades EVM, which can lead to loss of signal lock. Rev. A | Page 21 of 32 ADRF6510 Data Sheet 0 -10 -15 -20 -5 -10 -15 EVM (dB) -25 -30 -20 -30 -35 COFS = 100nF -35 -40 -40 -50 -80 -45 -70 -60 -50 -40 -30 -20 -10 0 10 PIN (dBm) 09002-152 -45 EFFECT OF COFS ON EVM When enabled, the dc offset compensation loop effectively nulls any information below the high-pass corner set by the COFS capacitor. However, loss of the low frequency information of the modulated signal can degrade the EVM in some cases. As the signal bandwidth becomes larger, the percentage of information that is corrupted by the high-pass corner becomes smaller. In such cases, it is important to select a COFS capacitor large enough to minimize the high-pass corner frequency, which prevents loss of information and degraded EVM. Figure 53 shows the effect of COFS values at a single signal bandwidth of 6.75 MHz = 1.35 x 5 MHz over input power. Figure 54 shows that EVM can be improved by using a bigger COFS value and/or increasing the signal bandwidth. Increasing signal bandwidth will improve EVM to a point after which the bandwidth limitations of the source, the part, and/or the receiver will start to dominate and degrade EVM. -5 -10 COFS = 1nF COFS = 100nF -30 -35 COFS = 1F -70 -60 -50 -40 -30 -20 15 20 25 30 35 ANTI-ALIASING FILTER The noise spectral density of the ADRF6510 outside the filter bandwidth is limited by the fixed VGA output noise. It may be necessary to use an external, fixed-frequency, passive filter prior to an analog-to-digital conversion to prevent noise aliasing from degrading the signal-to-noise ratio. As shown in Figure 47 and Figure 48, the noise density at higher frequencies tends to be flat, and any higher IF noise aliasing into the Nyquist zone has minimal effects. When designing an antialiasing filter, it is necessary to consider the overall source and load impedance presented by the ADRF6510 and the ADC input to design the filter network. The differential baseband output impedance of the ADRF6510 is 20 and is designed to drive a high impedance ADC input. It may be desirable to terminate the ADC input to a lower impedance by using a terminating resistor, such as 500 . The terminating resistor helps to better define the input impedance at the ADC input at the cost of a slightly reduced gain. The order and type of filter network depend on the desired high frequency rejection required, the pass-band ripple, and the group delay. Filter design tables provide outlines for various filter types and orders, illustrating the normalized inductor and capacitor values for a 1 Hz cutoff frequency and 1 load. -10 0 PIN (dBm) 10 09002-153 -45 -50 -80 10 As an example, a second-order Butterworth, low-pass filter design is shown in Figure 55 where the differential load impedance is 500 and the source impedance is 50 . The normalized series inductor value for the 10-to-1, load-to-source impedance ratio is 0.074 H, and the normalized shunt capacitor is 14.814 F. For a 31 MHz cutoff frequency, the single-ended equivalent circuit consists of a 0.191 H series inductor followed by a 152 pF shunt capacitor. -20 -40 5 After scaling the normalized prototype element values by the actual desired cutoff frequency and load impedance, the series reactance elements are halved to realize the final balanced filter network component values. 0 -25 0 Figure 54. EVM vs. Signal BW over COFS Values Figure 52 shows EVM degradation as the signal level nears compression. At 2.25 V p-p the signal is already degraded a few decibels. When the output level is near the absolute limits of the output stage, the EVM becomes much more erratic over the RF input power level. -15 COFS = 1F SIGNAL BANDWIDTH (MHz) Figure 52. EVM vs. RF Input Power over Output Voltage Levels, IF = 5 MHz, OFDS Pulled High EVM (dB) COFS = 1nF -25 09002-154 EVM (dB) 0 750mV p-p 1.50V p-p 1.75V p-p 2.00V p-p 2.25V p-p 2.90V p-p 2.95V p-p -5 Figure 53. EVM vs. RF Input Power over COFS Values Rev. A | Page 22 of 32 Data Sheet ADRF6510 10 The balanced configuration is realized as the 0.191 H inductor is split in half to achieve the network that is shown in Figure 55. LN = 0.074H -10 CN 14.814F RS = 0.1 RL GAIN (dB) NORMALIZED SINGLE-ENDED CONFIGURATION RL = 1 fC = 1Hz -15 -20 -25 -30 0.191H -35 -40 152pF RL = 500 -45 -50 fC = 31MHz RS = 25 2 0.096H 140 160 180 200 GROUP DELAY (ns) 12 10 8 6 4 2 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 BASEBAND FREQUENCY (MHz) 500 Figure 58. Third-Order Baseband Filter Group Delay Response 09002-064 15pF 120 14 0 20 100 20 1H 1H 80 16 A complete design example is shown in Figure 56. A third-order Chebyshev differential filter with a 31 MHz corner frequency interfaces the output of the ADRF6510 to that of an ADC input. The 20 source impedance reflects the impedance of the output buffer stage. The 500 load resistor defines the input impedance of the ADC. The filter adheres to a 0.1 dB in-band flatness and offers sufficient out-of-band rejection to act as an antialiasing filter. 1.8H 60 18 Figure 55. Second-Order Butterworth, Low-Pass Filter Design Example 1.8H 40 Figure 57. Third-Order Baseband Filter Response RL 2 = 250 RL = 250 2 09002-061 RS = 25 2 152pF 20 BASEBAND FREQUENCY (MHz) 0.096H BALANCED CONFIGURATION VS 0 09002-048 DENORMALIZED SINGLE-ENDED EQUIVALENT Figure 56. Third-Order Chebyshev Differential Filter Design Example Figure 57 and Figure 58 show the measured frequency response and group delay of the third-order Chebyshev differential filter. Rev. A | Page 23 of 32 09002-049 RS = 50 VS 0 -5 RS = 0.1 VS 5 ADRF6510 Data Sheet EVALUATION BOARD The ADRF6510 evaluation board is available with software control to program the filter bandwidth. It is a 4-layer board with split ground plane for analog and digital sections. Special care is taken to place the power decoupling capacitors close to the device pins. The board is designed for easy single-ended (through a Mini-Circuits(R) ADT8-1T+ 8:1 balun) or differential configuration for each channel. EVALUATION BOARD CONTROL SOFTWARE 09002-051 The ADRF6510 evaluation board is configured with a USBfriendly interface to program the filter bandwidth of the ADRF6510. The software GUI (see Figure 59) allows users to select a particular frequency to write to the device and also to read back data from the SDO pin that shows the currently programmed filter setting. The software setup files can be downloaded from the ADRF6510 product page at www.analog.com. Figure 59. Evaluation Control Software SCHEMATICS AND ARTWORK VPOSD R45 R17 OPEN OPEN R31 0 R47 0 T1 1 C6 0.1F 6 5 R56 OPEN 4 3 R55 OPEN R48 0 INP1 DIG_VPOS R57 0 C9 100nF R4 P3 10k R58 0 C10 100nF VPS LE CLK DATA SDO INP2 C27 330pF C27 330pF VPSD C4 0.1F COM ADRF6510 DATA VPS R1 10k C5 0.1F 6 5 3 4 COM OPM2 R50 0 C17 0.1F R5 0 C18 0.1F R6 0 OPP2 VPS VPS OFDS VPS INP2 OFS2 COM INM2 COM R8 300 R19 0 C23 0.1F R11 OPEN R53 0 C7 100nF R54 0 C8 100nF R3 10k C13 1000pF C15 0.1F C22 100nF R10 300 3 6 1 R23 R38 OPEN OPEN VOCM R9 300 T3 4 5 R21 R39 OPEN OPEN VGAIN VOCM C21 100nF OPP1 R41 0 R20 0 VGAIN VPS C11 0.1F R52 OPEN VOCM COM R2 0 R51 OPEN GAIN SDO R46 R15 OPEN OPEN T2 R7 300 OPM1 COMD CLK 1 C19 100nF R12 OPEN ENBL INM1 OFS1 COM GNSW VPS VPS INP1 OPP1 LE R49 0 C3 0.1F VPS VPSD R33 0 R34 0 COM R24 R37 OPEN OPEN C20 100nF R30 100 R32 0 INM2_SE_P P2 COMD C16 0.1F C12 0.1F VPS C2 10F L1 33H C14 1000pF R43 R18 OPEN OPEN R29 100 VPOS C1 10F VPS R14 OPEN R35 0 C24 0.1F R13 OPEN VPS P4 OPM1_SE_P OPM2_SE_P R42 0 T4 4 5 3 6 1 R36 0 R22 R40 OPEN OPEN OPP2 LEGEND NET NAME VPS TEST POINT SMA INPUT/OUTPUT DIGITAL GROUND R44 R16 OPEN OPEN ANALOG GROUND Figure 60. Evaluation Board Schematic Rev. A | Page 24 of 32 09002-065 INM1_SE_P VPOS L2 33H Data Sheet ADRF6510 Y1 24 MHz 3V3_USB 3 1 C51 22pF C54 22pF 2 4 R62 100k 3V3_USB R64 100k C45 0.1F 53 52 51 50 49 48 47 46 VCC CLKOUT GND PD7_FD15 PD6_FD14 PD5_FD13 PD4_FD12 PD3_FD11 PD2_FD10 PD1_FD9 1 RDY0_SLRD 45 44 43 VCC 54 PD0_FD8 55 WAKEUP 56 GND C37 0.1F GND 41 PA7_FLAGD_SCLS_N 40 3 AVCC C49 0.1F 5V_USB RESET_N 42 2 RDY1_SLWR C48 10pF PA6_PKTEND 39 4 XTALOUT 5 XTALIN PA5_FIFOARD1 38 6 AGND PA4_FIFOARD0 37 P5 1 2 7 AVCC 3V3_USB 3 4 5 G1 G2 G3 G4 3V3_USB PA3_WU2 36 CY7C68013A-56 LTXC U4 8 DPLUS LE PA2_SLOE 35 CLK 9 DMINUS PA1_INT1_N 34 DATA 10 AGND PA0_INT0_N 33 SDO 11 VCC VCC 32 12 GND CTL2_FLAGC 31 13 IFCLK CTL1_FLAGB 30 VCC PB0_FD0 PB1_FD1 PB2_FD2 PB3_FD3 PB4_FD4 PB5_FD5 PB6_FD6 PB7_FD7 GND VCC GND 15 SDA SCL 14 RESERVED 24LC64-I_SN U2 3V3_USB 16 17 18 19 20 21 22 23 24 25 26 27 28 R61 2k CTL0_FLAG A 29 CR2 3V3_USB 1 A0 SDA 5 2 A1 SCL 6 WC_N 7 VCC 8 3 A2 4 GND 3V3_USB 3V3_USB C52 1.0F R70 140k C50 1000pF R69 78.7k 3V3_USB C40 0.1F C41 0.1F C42 0.1F C35 0.1F C36 0.1F C44 0.1F 5V_USB ADP3334 U3 3V3_USB 1 OUT1 IN2 8 2 OUT2 IN1 7 3 FB SD 6 4 NC GND 5 C47 1.0F R65 2k CR1 C46 0.1F DGND Figure 61. Schematic for the USB Section of the Evaluation Board Rev. A | Page 25 of 32 09002-026 3V3_USB Data Sheet 09002-052 ADRF6510 09002-053 Figure 62. Top Layer Silkscreen Figure 63. Component Side Layout Rev. A | Page 26 of 32 Data Sheet ADRF6510 EVALUATION BOARD CONFIGURATION OPTIONS Table 5 lists the components of the main section of the ADRF6510 evaluation board. Table 5. Components C1, C2, C4, C5, C11, C12, C15, C16, L1, L2 Function Power supply and ground decoupling. Nominal supply decoupling consists of a 0.1 F capacitor to ground. T1, T2, C3, C6, C7 to C10, R15 to R18, R31, R32, R43 to R58 Input interface. Input SMAs INM1_SE_P and INP2_SE_P are used to drive the baluns in a single-ended fashion. The default configuration of the evaluation board is for single-ended operation. T1 and T2 are 8:1 impedance ratio baluns to transform a 50 , single-ended input into a 400 balanced differential signal. R31, R32, and R47 to R50 are populated for appropriate balun interface. R51 to R58 are provided for generic placement of matching components. C3 and C6 are balun decoupling capacitors. R15 to R18 and R43 to R46 can be populated with 0 , and the balun interfacing resistors can be removed to bypass T1 and T2 for differential interfacing. C7 to C10 can be used for ac coupling with differential interfacing. Output interface. Output SMAs OPP1_SE_P and OPM2_SE_P are used to drive the baluns in a single-ended fashion. The default configuration of the evaluation board is for single-ended operation. T3 and T4 are 8:1 impedance ratio baluns to transform a 50 , single-ended output into a 400 balanced differential load. R19, R20, R35, R36, R41, and R42 are populated for appropriate balun interface. R7 to R14 are provided for generic placement of matching components. R7 to R10 are set to 300 to present a 1 k load (with the balun used) at the DUT output. C19 to C22 are used for ac coupling when differential outputs are used. C23 and C24 are balun decoupling capacitors. R21 to R24 and R37 to R40 can be populated with 0 , and the balun interfacing resistors can be removed to bypass T3 and T4 for differential interfacing. Enable interface. The ADRF6510 is powered up by applying a logic high voltage to the ENBL pin (Jumper P2 is connected to VPS). Serial interface control. The digital interface sets the corner frequency of the device using the serial interface via the LE, CLK, DATA, and SDO pins. T3, T4, C19 to C24, R7 to R14, R19 to R24, R35 to R42 P2 C27, C28, R1, R29, R30, R33, R34 P4, C13, C14, R3 DC offset correction loop compensation. The dc offset correction loop is enabled (low) with Jumper P4. When enabled, the capacitors are connected to circuit common. The high-pass corner frequency is expressed as follows: fHP (Hz) = 1.2 x ((Linear Gain)/COFS (F)). C18, R6 C17, R5 Output common-mode setpoint. The output common-mode voltage can be set externally when applied to the VOCM pin. If the VOCM pin is left open, the output common-mode voltage defaults to VPS/2. Analog gain control. 0 V to 2 V, 30 mV/dB gain scaling. P3, R4 Front-end 6 dB or 12 dB gain switch. Pull low for 6 dB; pull high for 12 dB. Rev. A | Page 27 of 32 Default Conditions C1, C2 = 10 F (Size 1210) C4, C5, C11, C12, C15, C16 = 0.1 F (Size 0603) L1, L2 = 33 H (Size 1812) R2 = 0 (Size 0402) T1, T2 = ADT8-1T+ (Mini-Circuits) C3, C6 = 0.1 F (Size 0402) C7 to C10 = 100 nF (Size 0402) R15 to R18, R43 to R46, R51, R52, R55, R56 = open (Size 0402) R31, R32, R47 to R50, R53, R54, R57, R58 = 0 (Size 0402) T3, T4 = ADT8-1T+ (Mini-Circuits) C19 to C22 = 100 nF (Size 0402) C23, C24 = 0.1 F (Size 0402) R7 to R10 = 300 (Size 0402) R11 to R14 = open R19, R20, R35, R36, R41, R42 = 0 (Size 0402) R21 to R24, R37 to R40 = open (Size 0402) P2 = installed for enable R1 = 10 k (Size 0402) C27, C28 = 330 pF (Size 0402) R29, R30 = 100 (Size 0402) R33, R34 = 0 (Size 0402) P4 = installed C13, C14 = 1000 pF (Size 0402) R3 = 10 k (Size 0402) C18 = 0.1 F (Size 0402) R6 = 0 (Size 0402) C17 = 0.1 F (Size 0402) R5 = 0 (Size 0402) P3 = installed R4 = 10 k (Size 0402) ADRF6510 Data Sheet USB Section Configuration Options Table 6 lists the components of the USB section of the ADRF6510 evaluation board. Table 6. Components XC1, XC2, XC6 XC3 to XC5, XC7, XC8, XC12 to XC19 XC9 to XC11 XD1 XJ1 XR1, XR2 XR3 XR4, XR5 XR6 XU1 XU2 XU3 XY1 Default Conditions 22 pF (Size 0603) 0.1 F (Size 0402) 10 pF (Size 0402) Green LED ( Panasonic LNJ308G8TRA) USB SMT connector (Hirose Electric UX60A-MB-5ST 240-0003-4) 2 k (Size 0603) 1 k (Size 0603) 100 k (Size 0603) 0 (Size 0603) USB microcontroller (Cypress CY7C68013A-56LFXC) 64 kb EEPROM (Microchip 24LC64-I/SN) Low dropout regulator (Analog Devices ADP3303ARZ-3.3) 24 MHz crystal oscillator (AEL Crystals X24M000000S244) Rev. A | Page 28 of 32 Data Sheet ADRF6510 OUTLINE DIMENSIONS 0.60 MAX 5.00 BSC SQ 0.60 MAX PIN 1 INDICATOR 0.50 BSC 4.75 BSC SQ 0.50 0.40 0.30 12 MAX 17 16 0.80 MAX 0.65 TYP 0.30 0.23 0.18 3.25 3.10 SQ 2.95 EXPOSED PAD (BOTTOM VIEW) 9 8 0.25 MIN 3.50 REF 0.05 MAX 0.02 NOM SEATING PLANE 1 0.20 REF COPLANARITY 0.08 FOR PROPER CONNECTION OF THE EXPOSED PAD, REFER TO THE PIN CONFIGURATION AND FUNCTION DESCRIPTIONS SECTION OF THIS DATA SHEET. COMPLIANT TO JEDEC STANDARDS MO-220-VHHD-2 011708-A TOP VIEW 1.00 0.85 0.80 PIN 1 INDICATOR 32 25 24 Figure 64. 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 5 mm x 5 mm Body, Very Thin Quad (CP-32-2) Dimensions shown in millimeters ORDERING GUIDE Model 1 ADRF6510ACPZ-R7 ADRF6510ACPZ-WP ADRF6510-EVALZ 1 Temperature Range -40C to +85C -40C to +85C Package Description 32-Lead LFCSP_VQ, 7" Tape and Reel 32-Lead LFCSP_VQ, Waffle Pack Evaluation Board Z = RoHS Compliant Part. Rev. A | Page 29 of 32 Package Option CP-32-2 CP-32-2 ADRF6510 Data Sheet NOTES Rev. A | Page 30 of 32 Data Sheet ADRF6510 NOTES Rev. A | Page 31 of 32 ADRF6510 Data Sheet NOTES (c)2010-2011 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D09002-0-10/11(A) Rev. A | Page 32 of 32