FEATURES TYPICAL APPLICATION CIRCUIT Support input data rate >2 GSPS Proprietary low spurious and distortion design SFDR = 82 dBc at dc IF, −9 dBFS Flexible 8-lane JESD204B interface Multiple chip synchronization Fixed latency Data generator latency compensation Selectable 1×, 2×, 4×, or 8× interpolation filter Low power architecture Transmit enable function allows extra power saving and instant control of the output status High performance, low noise phase-locked loop (PLL) clock multiplier Digital inverse sinc filter Low power: 1.42 W at 1.6 GSPS full operating conditions 88-lead LFCSP with exposed pad QUAD MOD LPF ADRF6720 SYSREF± DAC RF OUTPUT 0°/90° PHASE SHIFTER JESD204B DAC AD9135/ AD9136 LO_IN SYNCOUT0± SYNCOUT1± MOD_SPI CLK± DAC SPI 12578-001 Data Sheet Dual, 11-/16-Bit, 2.8 GSPS, TxDAC+® Digital-to-Analog Converters AD9135/AD9136 Figure 1. APPLICATIONS Wireless communications 3G/4G W-CDMA base stations Wideband repeaters Software defined radios Wideband communications Point to point Local multipoint distribution service (LMDS) and multichannel multipoint distribution service (MMDS) Transmit diversity, multiple input/multiple output (MIMO) Instrumentation Automated test equipment GENERAL DESCRIPTION PRODUCT HIGHLIGHTS The AD9135/AD9136 are dual, 11-/16-bit, high dynamic range digital-to-analog converters (DACs) that provide a maximum sample rate of 2800 MSPS, permitting a multicarrier generation over a very wide bandwidth. The DAC outputs are optimized to interface seamlessly with the ADRF6720, as well as other analog quadrature modulators (AQMs) from Analog Devices, Inc. An optional 3-wire or 4-wire serial port interface (SPI) provides for programming/readback of many internal parameters. The fullscale output current can be programmed over a typical range of 13.9 mA to 27.0 mA. The AD9135/AD9136 are available in an 88-lead LFCSP. 1. 2. 3. 4. 5. 6. Rev. A Greater than 2 GHz, ultrawide complex signal bandwidth enables emerging wideband and multiband wireless applications. Advanced low spurious and distortion design techniques provide high quality synthesis of wideband signals from baseband to high intermediate frequencies. JESD204B Subclass 1 support simplifies multichip synchronization in software and hardware design. Fewer pins for data interface width with a serializer/ deserializer (SERDES) JESD204B eight-lane interface. Programmable transmit enable function allows easy design balance between power consumption and wake-up time. Small package size with 12 mm × 12 mm footprint. Document Feedback 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 ©2014 Analog Devices, Inc. All rights reserved. Technical Support www.analog.com AD9135/AD9136 Data Sheet TABLE OF CONTENTS Features .............................................................................................. 1 JESD204B Setup ......................................................................... 30 Applications ....................................................................................... 1 SERDES Clocks Setup ................................................................ 32 Typical Signal Chain......................................................................... 1 Equalization Mode Setup .......................................................... 32 General Description ......................................................................... 1 Link Latency Setup ..................................................................... 32 Product Highlights ........................................................................... 1 Crossbar Setup ............................................................................ 34 Revision History ............................................................................... 3 JESD204B Serial Data Interface .................................................... 35 Functional Block Diagram .............................................................. 4 JESD204B Overview .................................................................. 35 Specifications..................................................................................... 5 Physical Layer ............................................................................. 36 DC Specifications ......................................................................... 5 Data Link Layer .......................................................................... 39 Digital Specifications ................................................................... 6 Transport Layer .......................................................................... 48 Maximum DAC Update Rate Speed Specifications by Supply ..... 7 JESD204B Test Modes ............................................................... 58 JESD204B Serial Interface Speed Specifications ...................... 7 JESD204B Error Monitoring..................................................... 59 SYSREF Signal to DAC Clock Timing Specifications.............. 8 Hardware Considerations ......................................................... 61 Digital Input Data Timing Specifications ................................. 8 Digital Datapath ............................................................................. 65 Latency Variation Specifications ................................................ 9 DAC Paging ................................................................................. 65 JESD204B Interface Electrical Specifications ........................... 9 Data Format ................................................................................ 65 AC Specifications........................................................................ 10 Interpolation Filters ................................................................... 65 Absolute Maximum Ratings.......................................................... 11 Inverse Sinc ................................................................................. 66 Thermal Resistance .................................................................... 11 Digital Gain, DC Offset, and Group Delay ............................. 66 ESD Caution ................................................................................ 11 Downstream Protection ............................................................ 68 Pin Configuration and Function Descriptions ........................... 12 Datapath PRBS ........................................................................... 69 Terminology .................................................................................... 15 DC Test Mode ............................................................................. 70 Typical Performance Characteristics ........................................... 16 Interrupt Request Operation ........................................................ 71 Theory of Operation ...................................................................... 22 Interrupt Service Routine .......................................................... 71 Serial Port Operation ..................................................................... 23 DAC Input Clock Configurations ................................................ 72 Data Format ................................................................................ 23 Driving the CLK± Inputs .......................................................... 72 Serial Port Pin Descriptions ...................................................... 23 DAC PLL Fixed Register Writes ............................................... 72 Serial Port Options ..................................................................... 23 Clock Multiplication .................................................................. 72 Chip Information ............................................................................ 25 Starting the PLL .......................................................................... 74 Device Setup Guide ........................................................................ 26 Analog Outputs............................................................................... 75 Overview...................................................................................... 26 Transmit DAC Operation.......................................................... 75 Step 1: Start Up the DAC ........................................................... 26 Device Power Dissipation.............................................................. 78 Step 2: Digital Datapath ............................................................. 27 Temperature Sensor ................................................................... 78 Step 3: Transport Layer .............................................................. 27 Start-Up Sequence .......................................................................... 79 Step 4: Physical Layer ................................................................. 28 Step 1: Start Up the DAC ........................................................... 79 Step 5: Data Link Layer .............................................................. 28 Step 2: Digital Datapath............................................................. 79 Step 6: Optional Error Monitoring .......................................... 29 Step 3: Transport Layer .............................................................. 80 Step 7: Optional Features ........................................................... 29 Step 4: Physical Layer ................................................................. 80 DAC PLL Setup ........................................................................... 30 Step 5: Data Link Layer.............................................................. 81 Interpolation ............................................................................... 30 Step 6: Error Monitoring ........................................................... 81 Rev. A | Page 2 of 117 Data Sheet AD9135/AD9136 Register Maps and Descriptions ....................................................82 Outline Dimensions ......................................................................116 Device Configuration Register Map .........................................82 Ordering Guide .........................................................................117 Device Configuration Register Descriptions ..........................88 REVISION HISTORY 7/15—Rev. 0 to Rev. A Changed Functional Block Diagram Section to Typical Application Circuit Section.............................................................. 1 Changes to General Description Section ....................................... 1 Changed Detailed Functional Block Diagram Section to Functional Block Diagram Section ................................................. 4 Changes to Offset Drift Parameter, Table 1 ................................... 5 Deleted Reference Voltage Parameter, Table 1 .............................. 5 Changed 1× Interpolation Mode Parameter to 1× Interpolation Mode, JESD Mode 8, 8 SERDES Lanes Parameter, Table 1 ......... 5 Changes to Output Voltage (VOUT) Logic High Parameter, Output Voltage (VOUT) Logic Low Parameter, JESD204B Serial Interface Speed Minimum Parameter, and SYSREF± Frequency Parameter, Table 2 ............................................................................. 6 Changes to Table 4 ............................................................................ 7 Changes to Interpolation Parameter, Table 6 ................................ 8 Changed Junction Temperature Parameter to Operating Junction Temperature, Table 10 ....................................................11 Changes to Terminology Section ..................................................17 Changes to Figure 34 Caption and Figure 37 Caption ...............21 Changes to Device Revision Parameter, Table 14 .......................25 Changes to Overview Section, Table 15, Table 16, and Table 17 .............................................................................................26 Changes to Step 3: Transport Layer Section and Table 19 .........27 Changes to Table 20 and Table 21 .................................................28 Changes to Step 7: Optional Features Section .............................29 Added Table 25; Renumbered Sequentially .................................30 Changes to DAC PLL Setup Section, Table 26, and Table 27 ......30 Changes to Table 28 and CurrentLink Section............................31 Added DAC Power-Down Setup Section ....................................31 Changes to Table 30 ........................................................................32 Changes to Table 32 ........................................................................33 Changes to Table 36 and Figure 44 ...............................................36 Changes to Table 37 ........................................................................37 Added SERDES PLL Fixed Register Writes Section and Table 38 .............................................................................................37 Changes to Table 39 ........................................................................38 Changes to Figure 47 and Data Link Layer Section ...................39 Added Figure 50; Renumbered Sequentially ............................... 40 Changes to Table 45 and Table 46 ................................................. 49 Changes to Figure 61 ...................................................................... 51 Changes to Figure 62 ...................................................................... 52 Changes to Figure 63 ...................................................................... 53 Changes to Figure 64 ...................................................................... 54 Changes to Figure 65 ...................................................................... 56 Changes to Figure 66 ...................................................................... 57 Changes to Power Supply Recommendations Section ............... 61 Added Figure 68 .............................................................................. 62 Changes to Figure 72 ...................................................................... 64 Changes to Data Format Section and Table 61 ........................... 65 Changes to Figure 76 ...................................................................... 66 Changed 0x13D[7:0] to 0x13D[3:0], Table 62............................. 67 Changes to Group Delay Section .................................................. 67 Changes to DC Test Mode Section ............................................... 95 Deleted Table 70; Renumbered Sequentially ............................... 71 Moved Figure 78 and Table 68 ...................................................... 71 Added DAC PLL Fixed Register Writes Section and Table 69 ............................................................................................. 72 Changes to Clock Multiplication Section .................................... 73 Added Loop Filter Section and Charge Pump Filter Section ...... 73 Added Temperature Tracking Section and Table 73 .................. 74 Changes to Starting the PLL Section and Figure 82 ................... 74 Changes to Transmit DAC Operation Section ............................ 75 Changes to Start-Up Sequence Section, Table 77, and Table 78 ............................................................................................. 79 Changes to Table 80 and Table 81 ................................................. 80 Changes to Table 82 and Table 83 ................................................. 81 Changes to Table 84 ........................................................................ 82 Changes to Table 85 ........................................................................ 88 Deleted Lookup Tables for Three Different DAC PLL Reference Frequencies Section and Table 83 to 85 ..................................... 112 Added Figure 92 ............................................................................ 116 Updated Outline Dimensions...................................................... 116 Changes to Ordering Guide ......................................................... 117 9/14—Revision 0: Initial Version Rev. A | Page 3 of 117 AD9135/AD9136 Data Sheet FUNCTIONAL BLOCK DIAGRAM DACCLK HB2 HB3 OUT1+ FSC Q-GAIN DACCLK MODE CONTROL I-GAIN HB1 HB2 I-OFFSET HB3 OUT0+ FSC SERDIN0± CLOCK DISTRIBUTION AND CONTROL LOGIC PLL_CTRL SERIAL I/O PORT POWER-ON RESET DAC ALIGN DETECT CLK_SEL CONFIG REGISTERS OUT0– REF AND BIAS I120 SYSREF+ SYSREF– CLK+ CLK– DAC PLL 12578-002 TXEN1 IRQ TXEN0 SYSREF Rx CLK Rx DACCLK PLL_LOCK RESET SYNCOUT1+ SYNCOUT1– SYNCHRONIZATION LOGIC SDO SDIO SCLK CS SYNCOUT0+ SYNCOUT0– OUT1– Q-OFFSET INV SINC SERDIN7± HB1 CLOCK DATA RECOVERY AND CLOCK FORMATTER VTT INV SINC SERDES PLL Figure 2. Rev. A | Page 4 of 117 Data Sheet AD9135/AD9136 SPECIFICATIONS DC SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted. Table 1. AD9135 Parameter RESOLUTION ACCURACY Differential Nonlinearity (DNL) Integral Nonlinearity (INL) MAIN DAC OUTPUTS Gain Error I/Q Gain Mismatch Full-Scale Output Current (IOUTFS) Maximum Setting Minimum Setting Output Compliance Range Output Resistance Output Capacitance Gain DAC Monotonicity Settling Time MAIN DAC TEMPERATURE DRIFT Offset Gain REFERENCE Internal Reference Voltage ANALOG SUPPLY VOLTAGES AVDD33 PVDD12 CVDD12 DIGITAL SUPPLY VOLTAGES SIOVDD33 VTT DVDD12 SVDD12 IOVDD POWER CONSUMPTION 1× Interpolation Mode, JESD Mode 8, 8 SERDES Lanes AVDD33 PVDD12 CVDD12 SVDD12 DVDD12 SIOVDD33 IOVDD Test Conditions/Comments Min Typ 11 AD9136 Max Min Typ 16 Max Unit Bits With calibration ±0.175 ±0.35 With internal reference ±1.0 ±2.0 LSB LSB −2.5 −0.6 +2 +5.5 +0.6 −2.5 −0.6 +2 +5.5 +0.6 % FSR % FSR 25.5 13.1 −250 27.0 13.9 28.6 14.8 +750 25.5 13.1 −250 27.0 13.9 28.6 14.8 +750 mA mA mV MΩ pF Based on a 4 kΩ external resistor between I120 and GND To within ±0.5 LSB 1.2 V nominal supply voltage 1.3 V nominal supply voltage 1.2 V nominal supply voltage 1.3 V nominal supply voltage 0.2 3.0 Guaranteed 20 0.2 3.0 Guaranteed 20 0.04 32 0.04 32 ppm ppm/°C 1.2 1.2 V ns 3.13 1.14 1.14 3.3 1.2 1.2 3.47 1.26 1.26 3.13 1.14 1.14 3.3 1.2 1.2 3.47 1.26 1.26 V V V 3.13 1.1 1.14 1.274 1.14 1.274 1.71 3.3 1.2 1.2 1.3 1.2 1.3 1.8 3.47 1.37 1.26 1.326 1.26 1.326 3.47 3.13 1.1 1.14 1.274 1.14 1.274 1.71 3.3 1.2 1.2 1.3 1.2 1.3 1.8 3.47 1.37 1.26 1.326 1.26 1.326 3.47 V V V V V V V 1.42 1.74 1.42 1.74 W 68 100 101 554 196 11 36 73 113.4 112 665 224 12 50 68 100 101 554 196 11 36 73 113.4 112 665 224 12 50 mA mA mA mA mA mA μA fDAC = 1.6 GSPS, IF = 40 MHz, PLL on, digital gain on, inverse sinc on, DAC full-scale current (IOUTFS) = 20 mA Includes VTT Rev. A | Page 5 of 117 AD9135/AD9136 Data Sheet DIGITAL SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted. Table 2. Parameter CMOS INPUT LOGIC LEVEL Input Voltage (VIN) Logic High Low Symbol CMOS OUTPUT LOGIC LEVEL Output Voltage (VOUT) Logic High Low Test Conditions/Comments Min Typ Max Unit 1.8 V IOVDD 3.3 V 1.8 V IOVDD 3.3 V 0.7 × IOVDD 0.3 × IOVDD V V 1.8 V IOVDD 3.3 V 1.8 V IOVDD 3.3 V 0.75 × IOVDD 0.25 × IOVDD V V 1× interpolation2 (see Table 4) 2× interpolation2 4× interpolation3 8× interpolation3 2120 2120 2800 2800 MSPS MSPS MSPS MSPS 1× interpolation 2× interpolation 4× interpolation 8× interpolation 2120 1060 700 350 MSPS MSPS MSPS MSPS MAXIMUM DAC UPDATE RATE1 ADJUSTED DAC UPDATE RATE INTERFACE4 Number of JESD204B Lanes JESD204B Serial Interface Speed Minimum Maximum DAC CLOCK INPUT (CLK+, CLK−) Differential Peak-to-Peak Voltage Common-Mode Voltage Maximum Clock Rate REFCLK5 Frequency (PLL Mode) SYSTEM REFERENCE INPUT (SYSREF+, SYSREF−) Differential Peak-to-Peak Voltage Common-Mode Voltage SYSREF± Frequency6 SYSREF SIGNAL TO DAC CLOCK7 Setup Time Hold Time Keep Out Window SPI Maximum Clock Rate Minimum SCLK Pulse Width High Low SDIO to SCLK Setup Time Hold Time 8 Per lane Per lane, SVDD12 = 1.3 V ± 2% 1.44 Gbps Gbps 2000 1000 mV mV MHz MHz 2000 2000 fDATA/(K × S) mV mV Hz 10.64 400 Self biased input, ac-coupled 6.0 GHz ≤ fVCO ≤ 12.0 GHz Lanes 1000 600 2800 35 400 0 1000 SYSREF differential swing = 0.4 V, slew rate = 1.3 V/ns, common modes tested: ac-coupled, 0 V, 0.6 V, 1.25 V, 2.0 V tSSD tHSD KOW SCLK 131 119 ps ps ps 20 IOVDD = 1.8 V 10 tPWH tPWL MHz 8 12 tDS tDH 5 2 Rev. A | Page 6 of 117 ns ns ns ns Data Sheet Parameter SDO to SCLK Data Valid Window CS to SCLK Setup Time Hold Time AD9135/AD9136 Symbol Test Conditions/Comments Min Typ Max Unit tDV 25 ns tSCS 5 ns tHCS 2 ns 1 See Table 3 for detailed specifications for DAC update rate conditions. The maximum speed for 1× and 2× interpolation is limited by the JESD204B interface with increased supply levels. See Table 4 for details. The maximum speed for 4× and 8× interpolation is limited by the DAC core. See Table 4 for details. 4 See Table 4 for detailed specifications for JESD204B speed conditions. 5 REFCLK is the reference clock. 6 K, F, and S are JESD204B transport layer parameters. See Table 43 for the full definitions. 7 See Table 5 for detailed specifications for SYSREF signal to DAC clock timing conditions. 2 3 MAXIMUM DAC UPDATE RATE SPEED SPECIFICATIONS BY SUPPLY AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted. Table 3. Parameter MAXIMUM DAC UPDATE RATE 2×, 4×, and 8× Interpolation Test Conditions/Comments Min Typ Max Unit DVDD12, CVDD12 = 1.2 V ± 5% DVDD12, CVDD12 = 1.2 V ± 2% DVDD12, CVDD12 = 1.3 V ± 2% 2.23 2.41 2.80 GSPS GSPS GSPS DVDD12, CVDD12 = 1.2 V ± 5% DVDD12, CVDD12 = 1.2 V ± 2% DVDD12, CVDD12 = 1.3 V ± 2% 1.81 1.93 2.21 GSPS GSPS GSPS 1× Interpolation JESD204B SERIAL INTERFACE SPEED SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted. Table 4. Parameter HALF RATE FULL RATE OVERSAMPLING Test Conditions/Comments SVDD12 = 1.2 V ± 5% SVDD12 = 1.2 V ± 2% SVDD12 = 1.3 V ± 2% SVDD12 = 1.2 V ± 5% SVDD12 = 1.2 V ± 2% SVDD12 = 1.3 V ± 2% SVDD12 = 1.2 V ± 5% SVDD12 = 1.2 V ± 2% SVDD12 = 1.3 V ± 2% Rev. A | Page 7 of 117 Min 5.75 5.75 5.75 2.88 2.88 2.88 1.44 1.44 1.44 Typ Max 8.92 9.42 10.64 4.63 4.93 5.52 2.31 2.46 2.76 Unit Gbps Gbps Gbps Gbps Gbps Gbps Gbps Gbps Gbps AD9135/AD9136 Data Sheet SYSREF SIGNAL TO DAC CLOCK TIMING SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, SYSREF± common-mode voltages = 0.0 V, 0.6 V, 1.25 V, and 2.0 V, unless otherwise noted. Table 5. Parameter SYSREF DIFFERENTIAL SWING = 0.4 V, SLEW RATE = 1.3 V/ns Setup Time Hold Time SYSREF DIFFERENTIAL SWING = 0.7 V, SLEW RATE = 2.28 V/ns Setup Time Hold Time SYSREF SWING = 1.0 V, SLEW RATE = 3.26 V/ns Setup Time Hold Time Test Conditions/Comments Min Typ Max Unit AC-coupled DC-coupled AC-coupled DC-coupled 126 131 92 119 ps ps ps ps AC-coupled DC-coupled AC-coupled DC-coupled 96 104 77 95 ps ps ps ps AC-coupled DC-coupled AC-coupled DC-coupled 83 90 68 84 ps ps ps ps DIGITAL INPUT DATA TIMING SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = 25°C, IOUTFS = 20 mA, unless otherwise noted. Table 6. Parameter LATENCY Interface Interpolation 1× 2× 4× 8× Inverse Sinc Digital Gain Adjust POWER-UP TIME 1 Min Typ Max Unit 17 PClock1 cycles 66 137 251 484 17 12 60 DAC clock cycles DAC clock cycles DAC clock cycles DAC clock cycles DAC clock cycles DAC clock cycles μs PClock is the AD9135/AD9136 internal processing clock and equals the lane rate ÷ 40. Rev. A | Page 8 of 117 Data Sheet AD9135/AD9136 LATENCY VARIATION SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = 25°C, IOUTFS = 20 mA, unless otherwise noted. Table 7. Parameter DAC LATENCY VARIATION SYNC On PLL Off PLL On Min Typ Max Unit 0 1 +1 DAC clock cycles DAC clock cycles −1 JESD204B INTERFACE ELECTRICAL SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V, TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted. Table 8. Parameter JESD204B DATA INPUTS Input Leakage Current Logic High Logic Low Unit Interval Common-Mode Voltage Differential Voltage VTT Source Impedance Differential Impedance Differential Return Loss Common-Mode Return Loss DIFFERENTIAL OUTPUTS (SYNCOUTx±)2 Output Differential Voltage Normal Swing Mode High Swing Mode Output Offset Voltage DETERMINISTIC LATENCY Fixed Variable SYSREF± to LOCAL MULTIFRAME COUNTER (LMFC) DELAY Symbol Test Conditions/Comments Min TA = 25°C Input level = 1.2 V ± 0.25 V, VTT = 1.2 V Input level = 0 V UI VRCM R_VDIFF ZTT ZRDIFF RLRDIF RLRCM AC-coupled, VTT = SVDD121 At dc At dc Typ Max Unit 714 +1.85 1050 30 120 μA μA ps V mV Ω Ω dB dB 235 394 1.27 mV mV V 17 2 PClock3 cycles PClock3 cycles DAC clock cycles 10 −4 94 −0.05 110 80 100 8 6 VOD Register 0x2A5[0] = 0 Register 0x2A5[0] = 1 VOS 192 341 1.19 4 1 As measured on the input side of the ac coupling capacitor. IEEE Standard 1596.3 LVDS compatible. 3 PClock is an AD9135/AD9136 internal processing clock and equals the lane rate ÷ 40. 2 Rev. A | Page 9 of 117 AD9135/AD9136 Data Sheet AC SPECIFICATIONS AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V,1 VTT = 1.2 V, TA = 25°C, IOUTFS = 20 mA, unless otherwise noted. Table 9. Parameter SPURIOUS-FREE DYNAMIC RANGE (SFDR) fDAC = 983.04 MSPS fDAC = 983.04 MSPS fDAC = 1966.08 MSPS fDAC = 1966.08 MSPS TWO-TONE INTERMODULATION DISTORTION (IMD) fDAC =983.04 MSPS fDAC = 983.04 MSPS fDAC = 1966.08 MSPS fDAC = 1966.08 MSPS NOISE SPECTRAL DENSITY (NSD), SINGLE-TONE fDAC = 983.04 MSPS fDAC = 1966.08 MSPS W-CDMA FIRST ADJACENT CHANNEL LEAKAGE RATIO (ACLR), SINGLE-CARRIER fDAC = 983.04 MSPS fDAC = 983.04 MSPS fDAC = 1966.08 MSPS W-CDMA SECOND ACLR, SINGLE-CARRIER fDAC = 983.04 MSPS fDAC = 983.04 MSPS fDAC = 1966.08 MSPS 1 Test Conditions/Comments −9 dBFS single-tone fOUT = 20 MHz fOUT = 150 MHz fOUT = 20 MHz fOUT = 170 MHz −9 dBFS fOUT = 20 MHz fOUT = 150 MHz fOUT = 20 MHz fOUT = 170 MHz 0 dBFS fOUT = 150 MHz fOUT = 150 MHz 0 dBFS fOUT = 30 MHz fOUT = 150 MHz fOUT = 150 MHz 0 dBFS fOUT = 30 MHz fOUT = 150 MHz fOUT = 150 MHz SVDD12 = 1.3 V for all fDAC = 1966.08 MSPS conditions in Table 9. Rev. A | Page 10 of 117 Min Typ Max Unit 82 76 81 69 dBc dBc dBc dBc 90 82 90 81 dBc dBc dBc dBc −162 −163 dBm/Hz dBm/Hz 82 80 80 dBc dBc dBc 84 85 85 dBc dBc dBc Data Sheet AD9135/AD9136 ABSOLUTE MAXIMUM RATINGS THERMAL RESISTANCE Table 10. Parameter I120 to Ground SERDINx±, VTT, SYNCOUT1±/ SYNCOUT0±, TXENx OUTx± SYSREF± CLK± to Ground RESET, IRQ, CS, SCLK, SDIO, SDO to Ground LDO_BYP1 LDO_BYP2 LDO24 Ambient Operating Temperature (TA) Operating Junction Temperature Storage Temperature Range Rating −0.3 V to AVDD33 + 0.3 V −0.3 V to SIOVDD33 + 0.3 V −0.3 V to AVDD33 + 0.3 V GND − 0.5 V to +2.5 V −0.3 V to PVDD12 + 0.3 V −0.3 V to IOVDD + 0.3 V −0.3 V to SVDD12 + 0.3 V −0.3 V to PVDD12 + 0.3 V −0.3 V to AVDD33 + 0.3 V −40°C to +85°C 125°C −65°C to +150°C The exposed pad (EPAD) must be soldered to the ground plane for the 88-lead LFCSP. The EPAD provides an electrical, thermal, and mechanical connection to the board. Typical θJA, θJB, and θJC values are specified for a 4-layer JESD51-7 high effective thermal conductivity test board for leaded surface-mount packages. θJA is obtained in still air conditions (JESD51-2). Airflow increases heat dissipation, effectively reducing θJA. θJB is obtained following double-ring cold plate test conditions (JESD51-8). θJC is obtained with the test case temperature monitored at the bottom of the exposed pad. ΨJT and ΨJB are thermal characteristic parameters obtained with θJA in still air test conditions. Junction temperature (TJ) can be estimated using the following equations: TJ = TT + (ΨJT × P) Stresses at or above those listed under Absolute Maximum Ratings may cause permanent damage to the product. This is a stress rating only; functional operation of the product at these or any other conditions above those indicated in the operational section of this specification is not implied. Operation beyond the maximum operating conditions for extended periods may affect product reliability. or TJ = TB + (ΨJB × P) where: TT is the temperature measured at the top of the package. P is the total device power dissipation. TB is the temperature measured at the board. Table 11. Thermal Resistance Package 88-Lead LFCSP1 1 θJA 22.6 θJB 5.59 θJC 1.17 ΨJT 0.1 ΨJB 5.22 The exposed pad must be securely connected to the ground plane. ESD CAUTION Rev. A | Page 11 of 117 Unit °C/W AD9135/AD9136 Data Sheet 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 LDO_BYP2 CVDD12 I120 AVDD33 OUT0+ OUT0– LDO24 CVDD12 DNC DNC DNC AVDD33 CVDD12 AVDD33 OUT1+ OUT1– LDO24 CVDD12 DNC DNC DNC AVDD33 PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 AD9135/AD9136 TOP VIEW (Not to Scale) 66 65 64 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 IOVDD CS SCLK SDIO SDO RESET IRQ PROTECT_OUT0 PROTECT_OUT1 PVDD12 PVDD12 GND GND DVDD12 SERDIN7+ SERDIN7– SVDD12 SERDIN6+ SERDIN6– SVDD12 VTT SVDD12 NOTES 1. DNC = DO NOT CONNECT. DO NOT CONNECT TO THIS PIN. 2. THE EXPOSED PAD MUST BE SECURELY CONNECTED TO THE GROUND PLANE. 12578-003 SYNCOUT0+ SYNCOUT0– VTT SERDIN2+ SERDIN2– SVDD12 SERDIN3+ SERDIN3– SVDD12 SVDD12 SVDD12 LDO_BYP1 SIOVDD33 SVDD12 SERDIN4– SERDIN4+ SVDD12 SERDIN5– SERDIN5+ VTT SYNCOUT1– SYNCOUT1+ 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 PVDD12 CLK+ CLK– PVDD12 SYSREF+ SYSREF– PVDD12 PVDD12 PVDD12 PVDD12 TXEN0 TXEN1 DVDD12 DVDD12 SERDIN0+ SERDIN0– SVDD12 SERDIN1+ SERDIN1– SVDD12 VTT SVDD12 Figure 3. Pin Configuration Table 12. Pin Function Descriptions Pin No. 1 2 Mnemonic PVDD12 CLK+ 3 CLK− 4 5 PVDD12 SYSREF+ 6 SYSREF− 7 8 9 10 11 12 13 14 15 PVDD12 PVDD12 PVDD12 PVDD12 TXEN0 TXEN1 DVDD12 DVDD12 SERDIN0+ 16 SERDIN0− 17 18 SVDD12 SERDIN1+ Description 1.2 V Supply. PVDD12 provides a clean supply. PLL Reference/Clock Input, Positive. When the PLL is used, this pin is the positive reference clock input. When the PLL is not used, this pin is the positive device clock input. This pin is self biased and must be ac-coupled. PLL Reference/Clock Input, Negative. When the PLL is used, this pin is the negative reference clock input. When the PLL is not used, this pin is the negative device clock input. This pin is self biased and must be ac-coupled. 1.2 V Supply. PVDD12 provides a clean supply. Positive Reference Clock for Deterministic Latency. This pin is self biased for ac coupling. It can be ac-coupled or dc-coupled. Negative Reference Clock for Deterministic Latency. This pin is self biased for ac coupling. It can be ac-coupled or dc-coupled. 1.2 V Supply. PVDD12 provides a clean supply. 1.2 V Supply. PVDD12 provides a clean supply. 1.2 V Supply. PVDD12 provides a clean supply. 1.2 V Supply. PVDD12 provides a clean supply. Transmit Enable for DAC0. CMOS levels are determined with respect to IOVDD. Transmit Enable for DAC1. CMOS levels are determined with respect to IOVDD. 1.2 V Digital Supply. 1.2 V Digital Supply. Serial Channel Input 0, Positive. CML compliant. SERDIN0+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 0, Negative. CML compliant. SERDIN0− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. Serial Channel Input 1, Positive. CML compliant. SERDIN1+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Rev. A | Page 12 of 117 Data Sheet Pin No. 19 Mnemonic SERDIN1− 20 21 22 23 24 25 26 SVDD12 VTT SVDD12 SYNCOUT0+ SYNCOUT0− VTT SERDIN2+ 27 SERDIN2− 28 29 SVDD12 SERDIN3+ 30 SERDIN3− 31 32 33 34 35 36 37 SVDD12 SVDD12 SVDD12 LDO_BYP1 SIOVDD33 SVDD12 SERDIN4− 38 SERDIN4+ 39 40 SVDD12 SERDIN5− 41 SERDIN5+ 42 43 44 45 46 47 48 VTT SYNCOUT1− SYNCOUT1+ SVDD12 VTT SVDD12 SERDIN6− 49 SERDIN6+ 50 51 SVDD12 SERDIN7− 52 SERDIN7+ 53 54 55 56 57 58 59 60 DVDD12 GND GND PVDD12 PVDD12 PROTECT_OUT1 PROTECT_OUT0 IRQ AD9135/AD9136 Description Serial Channel Input 1, Negative. CML compliant. SERDIN1− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. 1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins. 1.2 V JESD204B Receiver Supply. Positive LVDS Sync (Active Low) Output Signal Channel Link 0. Negative LVDS Sync (Active Low) Output Signal Channel Link 0. 1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins. Serial Channel Input 2, Positive. CML compliant. SERDIN2+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 2, Negative. CML compliant. SERDIN2− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. Serial Channel Input 3, Positive. CML compliant. SERDIN3+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 3, Negative. CML compliant. SERDIN3− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. 1.2 V JESD204B Receiver Supply. 1.2 V JESD204B Receiver Supply. LDO SERDES Bypass. This pin requires a 1 Ω resistor in series with a 1 μF capacitor to ground. 3.3 V Supply for SERDES. 1.2 V JESD204B Receiver Supply. Serial Channel Input 4, Negative. CML compliant. SERDIN4− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 4, Positive. CML compliant. SERDIN4+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. Serial Channel Input 5, Negative. CML compliant. SERDIN5− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 5, Positive. CML compliant. SERDIN5+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins. Negative LVDS Sync (Active Low) Output Signal Channel Link 1. Positive LVDS Sync (Active Low) Output Signal Channel Link 1. 1.2 V JESD204B Receiver Supply. 1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins. 1.2 V JESD204B Receiver Supply. Serial Channel Input 6, Negative. CML compliant. SERDIN6− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 6, Positive. CML compliant. SERDIN6+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V JESD204B Receiver Supply. Serial Channel Input 7, Negative. CML compliant. SERDIN7− is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. Serial Channel Input 7, Positive. CML compliant. SERDIN7+ is internally terminated to the VTT pin voltage using a calibrated 50 Ω resistor. This pin is ac-coupled only. 1.2 V Digital Supply. Ground. Connect GND to the ground plane. Ground. Connect GND to the ground plane. 1.2 V Supply. PVDD12 provides a clean supply. 1.2 V Supply. PVDD12 provides a clean supply. Power Detection and Protection Pin Output for DAC1. Pin 58 is high when power protection is in process. Power Detection and Protection Pin Output for DAC0. Pin 59 is high when power protection is in process. Interrupt Request (Active Low, Open Drain). Rev. A | Page 13 of 117 AD9135/AD9136 Pin No. 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 Mnemonic RESET SDO SDIO SCLK CS IOVDD AVDD33 DNC DNC DNC CVDD12 LDO24 OUT1− OUT1+ AVDD33 CVDD12 AVDD33 DNC DNC DNC CVDD12 LDO24 OUT0− OUT0+ AVDD33 I120 CVDD12 LDO_BYP2 EPAD Data Sheet Description Reset. This pin is active low. CMOS levels are determined with respect to IOVDD. Serial Port Data Output. CMOS levels are determined with respect to IOVDD. Serial Port Data Input/Output. CMOS levels are determined with respect to IOVDD. Serial Port Clock Input. CMOS levels are determined with respect to IOVDD. Serial Port Chip Select. This pin is active low. CMOS levels are determined with respect to IOVDD. IOVDD Supply for CMOS Input/Output and SPI. Operational for 1.8 V IOVDD 3.3 V. 3.3 V Analog Supply for DAC Cores. Do not connect to this pin. Do not connect to this pin. Do not connect to this pin. 1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 71. 2.4 V LDO. Requires a 1 μF capacitor to ground. DAC1 Negative Current Output. DAC1 Positive Current Output. 3.3 V Analog Supply for DAC Cores. 1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 76. 3.3 V Analog Supply for DAC Cores. Do not connect to this pin. Do not connect to this pin. Do not connect to this pin. 1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 81. 2.4 V LDO. Requires a 1 μF capacitor to ground. DAC0 Negative Current Output. DAC0 Positive Current Output. 3.3 V Analog Supply for DAC Cores. Output Current Generation Pin for DAC Full-Scale Current. Tie a 4 kΩ resistor from the I120 pin to ground. 1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 87. LDO Clock Bypass for DAC PLL. This pin requires a 1 Ω resistor in series with a 1 μF capacitor to ground. Exposed Pad. The exposed pad must be securely connected to the ground plane. Rev. A | Page 14 of 117 Data Sheet AD9135/AD9136 TERMINOLOGY Integral Nonlinearity (INL) INL is the maximum deviation of the actual analog output from the ideal output, determined by a straight line drawn from zero scale to full scale. Differential Nonlinearity (DNL) DNL is the measure of the variation in analog value, normalized to full scale, associated with a 1 LSB change in digital input code. Offset Error Offset error is the deviation of the output current from the ideal of 0 mA. For OUTx+, 0 mA output is expected when all inputs are set to 0. For OUTx−, 0 mA output is expected when all inputs are set to 1. Gain Error Gain error is the difference between the actual and ideal output span. The actual span is determined by the difference between the output when the input is at its minimum code and the output when the input is at its maximum code. Output Compliance Range The output compliance range is the range of allowable voltages at the output of a current output DAC. Operation beyond the maximum compliance limits can cause either output stage saturation or breakdown, resulting in nonlinear performance. Temperature Drift Offset drift is a measure of how far from full-scale range (FSR) the DAC output current is at 25°C (in ppm). Gain drift is a measure of the slope of the DAC output current across its full ambient operating temperature range, TA (in ppm/°C). Power Supply Rejection (PSR) PSR is the maximum change in the full-scale output as the supplies are varied from minimum to maximum specified voltages. Settling Time Settling time is the time required for the output to reach and remain within a specified error band around its final value, measured from the start of the output transition. Spurious-Free Dynamic Range (SFDR) SFDR is the difference, in decibels, between the peak amplitude of the output signal and the peak spurious signal within the dc to Nyquist frequency of the DAC. Typically, energy in this band is rejected by the interpolation filters. This specification, therefore, defines how well the interpolation filters work and the effect of other parasitic coupling paths on the DAC output. Signal-to-Noise Ratio (SNR) SNR is the ratio of the rms value of the measured output signal to the rms sum of all other spectral components below the Nyquist frequency, excluding the first six harmonics and dc. The value for SNR is expressed in decibels. Interpolation Filter If the digital inputs to the DAC are sampled at a multiple rate of fDATA (interpolation rate), a digital filter can be constructed that has a sharp transition band near fDATA/2. Images that typically appear around fDAC (output data rate) can be greatly suppressed. Adjacent Channel Leakage Ratio (ACLR) ACLR is the ratio in decibels relative to the carrier (dBc) between the measured power within a channel relative to its adjacent channel. Complex Image Rejection In a traditional two-part upconversion, two images are created around the second IF frequency. These images have the effect of wasting transmitter power and system bandwidth. By placing the real part of a second complex modulator in series with the first complex modulator, either the upper or lower frequency image near the second IF can be rejected. Adjusted DAC Update Rate The adjusted DAC update rate is defined as the DAC update rate divided by the smallest interpolating factor. For clarity on DACs with multiple interpolating factors, the adjusted DAC update rate for each interpolating factor may be given. Physical Lane Physical Lane x refers to SERDINx±. Logical Lane Logical Lane x refers to physical lanes after optionally being remapped by the crossbar block (Register 0x308 to Register 0x30B). Link Lane Link Lane x refers to logical lanes considered per link. When paging Link 0 (Register 0x300[2] = 0), Link Lane x = Logical Lane x. When paging Link 1 (Register 0x300[2] = 1, dual link only), Link Lane x = Logical Lane x + 4. Rev. A | Page 15 of 117 AD9135/AD9136 Data Sheet TYPICAL PERFORMANCE CHARACTERISTICS 0 –20 –40 –60 –80 –40 –60 100 200 300 400 500 fOUT (MHz) 12578-104 0 –100 0 MEDIAN 300 400 500 0dBFS –6dBFS –9dBFS –12dBFS –20 SFDR (dBc) –40 –60 –80 –40 –60 100 200 300 400 500 fOUT (MHz) 12578-305 0 –100 0 100 200 300 400 500 fOUT (MHz) 12578-108 –80 –100 Figure 8. Single-Tone SFDR vs. fOUT in the First Nyquist Zone over Digital Back Off, fDAC = 1966 MHz Figure 5. Single-Tone SFDR vs. fOUT in the First Nyquist Zone, fDAC = 1966 MHz and 2456 MHz 0 0 fDAC = 983MHz fDAC = 1228MHz fDAC = 1474MHz IN-BAND SECOND HARMONIC IN-BAND THIRD HARMONIC MAX DIGITAL SPUR –20 –20 –40 –40 IMD3 (dBc) –60 –60 –80 –80 –100 100 200 300 fOUT (MHz) 400 500 12578-106 –100 0 Figure 6. Single-Tone Second and Third Harmonics and Maximum Digital Spur in the First Nyquist Zone, fDAC = 1966 MHz, 0 dB Back Off Rev. A | Page 16 of 117 0 100 200 300 400 fOUT (MHz) Figure 9. Two-Tone Third IMD (IMD3) vs. fOUT, fDAC = 983 MHz, 1228 MHz, and 1474 MHz 500 12578-109 SFDR (dBc) 200 Figure 7. Single-Tone SFDR vs. fOUT in the First Nyquist Zone over Digital Back Off, fDAC = 983 MHz fDAC = 1966MHz fDAC = 2456MHz –20 100 fOUT (MHz) Figure 4. Single-Tone SFDR vs. fOUT in the First Nyquist Zone, fDAC = 983 MHz, 1228 MHz, and 1474 MHz 0 0 12578-107 –80 –100 SFDR (dBc) 0dBFS –6dBFS –9dBFS –12dBFS –20 SFDR (dBc) SFDR (dBc) 0 fDAC = 983MHz fDAC = 1228MHz fDAC = 1474MHz Data Sheet 0 fDAC = 1966MHz fDAC = 2456MHz –20 –20 –40 –40 IMD3 (dBc) –60 –80 1MHz TONE SPACING 16MHz TONE SPACING 35MHz TONE SPACING –60 –80 0 100 200 300 400 500 fOUT (MHz) 100 200 300 400 500 fOUT (MHz) –130 0dBFS –6dBFS –9dBFS –12dBFS –20 0 Figure 13. Two-Tone Third IMD (IMD3) vs. fOUT over Tone Spacing at 0 dB Back Off, fDAC = 983 MHz and 1966 MHz Figure 10. Two-Tone Third IMD (IMD3) vs. fOUT, fDAC = 1966 MHz and 2456 MHz 0 –100 12578-110 –100 fDAC = 983MHz fDAC = 1966MHz 12578-113 IMD3 (dBc) 0 AD9135/AD9136 fDAC = 983MHz fDAC = 1228MHz fDAC = 1474MHz –135 NSD (dBm/Hz) IMD3 (dBc) –140 –40 –60 –145 –150 –155 –160 –80 100 200 300 400 500 fOUT (MHz) –170 0 300 400 500 500 Figure 14. AD9136 Single-Tone (0 dBFS) NSD vs. fOUT, fDAC = 983 MHz, 1228 MHz, and 1474 MHz –130 0dBFS –6dBFS –9dBFS –12dBFS –20 200 fOUT (MHz) Figure 11. Two-Tone Third IMD (IMD3) vs. fOUT over Digital Back Off, fDAC = 983 MHz, Each Tone Is at −6 dBFS 0 100 12578-114 0 12578-111 –100 12578-421 –165 fDAC = 983MHz fDAC = 1228MHz fDAC = 1474MHz –135 NSD (dBm/Hz) IMD3 (dBc) –140 –40 –60 –145 –150 –155 –160 –80 –100 0 100 200 300 fOUT (MHz) 400 500 12578-112 –165 Figure 12. Two-Tone Third IMD (IMD3) vs. fOUT over Digital Back Off, fDAC = 1966 MHz, Each Tone Is at −6 dBFS Rev. A | Page 17 of 117 –170 0 100 200 300 400 fOUT (MHz) Figure 15. AD9135 Single-Tone (0 dBFS) NSD vs. fOUT, fDAC = 983 MHz, 1228 MHz, and 1474 MHz AD9135/AD9136 Data Sheet –135 –140 –145 –145 NSD (dBm/Hz) –140 –150 –155 –165 –165 200 300 400 500 –170 –130 fDAC = 1966MHz fDAC = 2456MHz –145 NSD (dBm/Hz) –145 –150 –155 –165 –165 400 500 fOUT (MHz) –170 12578-422 300 –130 –145 –145 NSD (dBm/Hz) –140 –150 –155 –165 300 400 500 fOUT (MHz) 12578-116 –165 200 400 500 –155 –160 100 300 –150 –160 0 200 0dBFS –6dBFS –9dBFS –12dBFS –135 –140 –170 100 Figure 20. AD9136 Single-Tone NSD vs. fOUT over Digital Back Off, fDAC = 1966 MHz 0dBFS –6dBFS –9dBFS –12dBFS –135 0 fOUT (MHz) Figure 17. AD9135 Single-Tone (0 dBFS) NSD vs. fOUT, fDAC = 1966 MHz and 2456 MHz –130 500 –155 –160 200 400 –150 –160 100 300 0dBFS –6dBFS –9dBFS –12dBFS –135 –140 0 200 Figure 19. AD9135 Single-Tone NSD vs. fOUT over Digital Back Off, fDAC = 983 MHz –140 –170 100 fOUT (MHz) –130 –135 0 12578-117 100 12578-115 0 Figure 16. AD9136 Single-Tone (0 dBFS) NSD vs. fOUT, fDAC = 1966 MHz and 2456 MHz NSD (dBm/Hz) –155 –160 fOUT (MHz) NSD (dBm/Hz) –150 –160 –170 0dBFS –6dBFS –9dBFS –12dBFS Figure 18. AD9136 Single-Tone NSD vs. fOUT over Digital Back Off, fDAC = 983 MHz –170 0 100 200 300 400 500 fOUT (MHz) Figure 21. AD9135 Single-Tone NSD vs. fOUT over Digital Back Off, fDAC = 1966 MHz Rev. A | Page 18 of 117 12578-424 –135 NSD (dBm/Hz) –130 fDAC = 1966MHz fDAC = 2456MHz 12578-423 –130 Data Sheet –130 PLL OFF PLL ON –135 fDAC = 983MHz fDAC = 1966MHz –140 NSD (dBm/Hz) AD9135/AD9136 –145 –150 –155 –160 –170 0 100 200 300 400 500 fOUT (MHz) 12578-118 12578-317 –165 Figure 22. AD9136 Single-Tone NSD (0 dBFS) vs. fOUT, fDAC = 983 MHz and 1966 MHz, PLL On and Off –130 PLL OFF PLL ON –135 fDAC = 983MHz fDAC = 1966MHz –140 NSD (dBm/Hz) Figure 25. AD9136 Four-Carrier W-CDMA ACLR, fOUT = 30 MHz, fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz –145 –150 –155 12578-417 –160 –170 0 100 200 300 400 500 fOUT (MHz) 12578-425 –165 Figure 23. AD9135 Single-Tone NSD (0 dBFS) vs. fOUT, fDAC = 983 MHz and 1966 MHz, PLL On and Off –60 fOUT = 30MHz fOUT = 200MHz fOUT = 400MHz –80 PHASE NOISE (dBc/Hz) Figure 26. AD9135 Four-Carrier W-CDMA ACLR, fOUT = 30 MHz, fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz PLL OFF PLL ON –100 –120 –140 10 100 1k 10k 100k OFFSET FREQUENCY (Hz) 1M 10M 12578-119 –180 12578-318 –160 Figure 24. AD9136 Single-Tone Phase Noise vs. Offset Frequency over fOUT, fDAC = 2.0 GHz, PLL On and Off Rev. A | Page 19 of 117 Figure 27. AD9136 Four-Carrier W-CDMA ACLR, fOUT = 122 MHz, fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz Data Sheet 12578-418 12578-320 AD9135/AD9136 Figure 31. AD9136 Four-Carrier W-CDMA ACLR, fOUT = 122 MHz, fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 122 MHz 11675-420 12578-319 Figure 28. AD9135 Four-Carrier W-CDMA ACLR, fOUT = 122 MHz, fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz Figure 32. AD9135 Four-Carrier W-CDMA ACLR, fOUT = 122 MHz, fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 122 MHz Figure 30. AD9135 Four-Carrier W-CDMA ACLR, fOUT = 30 MHz, fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 245 MHz 12578-433 12578-419 Figure 29. AD9136 Four-Carrier W-CDMA ACLR, fOUT = 30 MHz, fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 245 MHz Figure 33. AD9136 Output Performance of an Ultra Wideband (900 MHz) QAM Signal, fDAC = 2 GHz, 1× Interpolation, Inverse sinc On, JESD204B Mode 11 Rev. A | Page 20 of 117 Data Sheet 1300 1200 1100 1000 1500 2000 2500 Figure 34. Total Power Consumption vs. fDAC over Interpolation, 8 SERDES Lanes Enabled, 2 DACs Enabled, Digital Gain, Inverse Sinc and DAC PLL Disabled 300 1 SUPPLY CURRENT (mA) 200 40 3 4 5 6 7 8 Figure 36. SVDD12 Current vs. Lane Rate over Number of SERDES Lanes and Supply Voltage Setting 250 60 2 LANE RATE (Gbps) DVDD12 CVDD12 PVDD12 AVDD33 1.2V SUPPLY 1.3V SUPPLY 3.3V SUPPLY 150 100 400 600 800 1000 1200 1400 1600 fDAC (MHz) Figure 35. Power Consumption vs. fDAC over Digital Functions 0 400 600 800 100 fDAC (MHz) 1200 1400 1600 12578-329 50 20 12578-327 POWER CONSUMPTION (mW) 400 100 PLL (fDAC /fREF RATIO:4) DIGITAL GAIN INVERSE SINC 80 0 200 500 12578-328 1000 fDAC (MHz) 100 1.2V SVDD12 SUPPLY 1.3V SVDD12 SUPPLY 200 900 500 120 2 LANES 4 LANES 8 LANES 600 SVDD12 CURRENT (mA) 1400 700 1× 2× 4× 12578-326 TOTAL POWER CONSUMPTION (mW) 1500 AD9135/AD9136 Figure 37. DVDD12, CVDD12, PVDD12, and AVDD33 Supply Currents vs. fDAC over Supply Voltage Setting, 2 DACs Enabled Rev. A | Page 21 of 117 AD9135/AD9136 Data Sheet THEORY OF OPERATION The AD9135/AD9136 are 11-/16-bit, dual DACs with a SERDES interface. Figure 2 shows a detailed functional block diagram of the AD9135/AD9136. Eight high speed serial lanes carry data at a maximum speed of 10.64 Gbps, and a 2120 MSPS input data rate to each DAC. Compared to either LVDS or CMOS interfaces, the SERDES interface simplifies pin count, board layout, and input clock requirements to the device. The clock for the input data is derived from the device clock (required by the JESD204B specification). This device clock can be sourced with a PLL reference clock used by the on-chip PLL to generate a DAC clock or a high fidelity direct external DAC sampling clock. The device can be configured to operate in one-, two-, four-, or eight-lane modes, depending on the required input data rate. The digital datapath of the AD9135/AD9136 offers four interpolation modes (1×, 2×, 4×, and 8×) through three half-band filters with a maximum DAC sample rate of 2.8 GSPS. An inverse sinc filter is provided to compensate for sinc related roll-off. The AD9135/AD9136 DAC cores provide a fully differential current output with a nominal full-scale current of 20 mA. The full-scale current, IOUTFS, is user adjustable to between 13.9 mA and 27.0 mA, typically. The differential current outputs are complementary and are optimized for easy integration with the Analog Devices the ADRF6720 AQM. The AD9135/AD9136 are capable of multichip synchronization that can both synchronize multiple DACs and establish a constant and deterministic latency (latency locking) path for the DACs. The latency for each of the DACs remains constant from link establishment to link establishment. An external alignment (SYSREF±) signal makes the AD9135/AD9136 Subclass 1 compliant. Several modes of SYSREF± signal handling are available for use in the system. An SPI configures the various functional blocks and monitors their statuses. The various functional blocks and the data interface must be set up in a specific sequence for proper operation (see the Device Setup Guide section). Simple SPI initialization routines set up the JESD204B link and are included in the evaluation board package. The following sections describe the various blocks of the AD9135/AD9136 in greater detail. Descriptions of the JESD204B interface, control parameters, and various registers to set up and monitor the device are provided. The recommended start-up routine reliably sets up the data link. Rev. A | Page 22 of 117 Data Sheet AD9135/AD9136 SERIAL PORT OPERATION The serial port is a flexible, synchronous serial communications port that allows easy interfacing with many industry-standard microcontrollers and microprocessors. The serial input/output (I/O) is compatible with most synchronous transfer formats, including both the Motorola SPI and Intel® SSR protocols. The interface allows read/write access to all registers that configure the AD9135/AD9136. MSB first or LSB first transfer formats are supported. The serial port interface can be configured as a 4-wire interface or a 3-wire interface in which the input and output share a single-pin I/O (SDIO). CS 65 The serial clock pin synchronizes data to and from the device and runs the internal state machines. The maximum frequency of SCLK is 10 MHz. All data input is registered on the rising edge of SCLK. All data is driven out on the falling edge of SCLK. SPI PORT 12578-044 SCLK 64 Figure 38. Serial Port Interface Pins There are two phases to a communication cycle with the AD9135/AD9136. Phase 1 is the instruction cycle (the writing of an instruction byte into the device), coincident with the first 16 SCLK rising edges. The instruction word provides the serial port controller with information regarding the data transfer cycle, Phase 2 of the communication cycle. The Phase 1 instruction word defines whether the upcoming data transfer is a read or write, along with the starting register address for the following data transfer. A logic high on the CS pin followed by a logic low resets the serial port timing to the initial state of the instruction cycle. From this state, the next 16 rising SCLK edges represent the instruction bits of the current I/O operation. The remaining SCLK edges are for Phase 2 of the communication cycle. Phase 2 is the actual data transfer between the device and the system controller. Phase 2 of the communication cycle is a transfer of one or more data bytes. Eight × N SCLK cycles are needed to transfer N bytes during the transfer cycle. Registers change immediately upon writing to the last bit of each transfer byte. DATA FORMAT The instruction byte contains the information shown in Table 13. Table 13. Serial Port Instruction Word I[15] (MSB) R/W SERIAL PORT PIN DESCRIPTIONS Serial Clock (SCLK) SDO 62 SDIO 63 A14 to A0, Bit 14 to Bit 0 of the instruction word, determine the register that is accessed during the data transfer portion of the communication cycle. For multibyte transfers, A[14:0] is the starting address. The remaining register addresses are generated by the device based on the address increment bits. If the address increment bits are set high (Register 0x000, Bit 5 and Bit 2), multibyte SPI writes start at A[14:0] and increment by 1 every 8 bits sent/received. If the address increment bits are set to 0, the address decrements by 1 every 8 bits. I[14:0] A[14:0] Chip Select (CS) An active low input starts and gates a communication cycle. CS allows more than one device to be used on the same serial communications lines. The SDIO pin goes to a high impedance state when this input is high. During the communication cycle, chip select must stay low. Serial Data I/O (SDIO) This pin is a bidirectional data line. In 4-wire mode, this pin acts as the data input, and SDO acts as the data output. SERIAL PORT OPTIONS The serial port can support both MSB first and LSB first data formats. This functionality is controlled by the LSB first bits (Register 0x000, Bit 6 and Bit 1). The default is MSB first (LSBFIRST/LSBFIRST_M = 0). When the LSB first bits = 0 (MSB first), the instruction and data bits must be written from MSB to LSB. R/W is followed by A[14:0] as the instruction word, and D[7:0] is the data-word. When the LSB first bits = 1 (LSB first), the opposite is true. A[0:14] is followed by R/W, which is subsequently followed by D[0:7]. The serial port supports a 3-wire or 4-wire interface. When the SDO active bits = 1 (Register 0x000, Bit 4 and Bit 3), a 4-wire interface with a separate input pin (SDIO) and output pin (SDO) is used. When the SDO active bits = 0, the SDO pin is unused and the SDIO pin is used for both input and output. R/W, Bit 15 of the instruction word, determines whether a read or a write data transfer occurs after the instruction word write. Logic 1 indicates a read operation, and Logic 0 indicates a write operation. Rev. A | Page 23 of 117 AD9135/AD9136 Data Sheet INSTRUCTION CYCLE Multibyte data transfers can be performed as well. This is done by holding the CS pin low for multiple data transfer cycles (eight SCLKs) after the first data transfer word following the instruction cycle. The first eight SCLKs following the instruction cycle read from or write to the register provided in the instruction cycle. For each additional eight SCLK cycles, the address is either incremented or decremented and the read/write occurs on the new register. The direction of the address can be set using the address increment bits (Register 0x000, Bit 5 and Bit 2). When the address increment bits is 1, the multicycle addresses are incremented. When the address increment bits is 0, the addresses are decremented. A new write cycle can always be initiated by bringing CS high and then low again. DATA TRANSFER CYCLE CS SDIO R/W A14 A13 A3 A2 A1 A0 D7N D6N D5N D30 D20 D10 D00 Figure 39. Serial Register Interface Timing, MSB First, ADDRINC = 0 INSTRUCTION CYCLE DATA TRANSFER CYCLE CS SDIO A0 A1 A2 A12 A13 A14 R/W D0 0 D10 D20 D4N D5N D6N D7N Figure 40. Serial Register Interface Timing, LSB First, ADDRINC = 1 CS SDIO DATA BIT n DATA BIT n – 1 Figure 41. Timing Diagram for Serial Port Register Read tSCLK CS tPWH tPWL SDIO tDH INSTRUCTION BIT 15 INSTRUCTION BIT 14 Figure 42. Timing Diagram for Serial Port Register Write Rev. A | Page 24 of 117 12578-047 SCLK 12578-048 SCLK tDV tDS 12578-046 SCLK To prevent confusion and to ensure consistency between devices, the chip tests the first nibble following the address phase, ignoring the second nibble. This test is completed independently from the LSB first bit and ensures that there are extra clock cycles following the soft reset bits (Register 0x000, Bit 0 and Bit 7). This only applies when writing to Register 0x000. tDCS 12578-045 SCLK Data Sheet AD9135/AD9136 CHIP INFORMATION Register 0x003 to Register 0x006 contain chip information, as shown in Table 14. Table 14. Chip Information Information Chip Type Product ID Product Grade Device Revision Description The product type is high speed DAC, which is represented by a code of 0x04 in Register 0x003. Eight MSBs in Register 0x005 and eight LSBs in Register 0x004. The product ID is 0x9144. Register 0x006[7:4]. The product grade is 0x6 for the AD9136 and 0x4 for the AD9135. Register 0x006[3:0]. The device revision is 0x6. Rev. A | Page 25 of 117 AD9135/AD9136 Data Sheet DEVICE SETUP GUIDE OVERVIEW The sequence of steps to properly set up the AD9135/AD9136 is as follows: 1. 2. 3. 4. 5. 6. 7. Set up the SPI interface, power up necessary circuit blocks, make the required writes to the configuration registers, and set up the DAC clocks (see the Step 1: Start Up the DAC section). Set the digital features of the AD9135/AD9136 (see the Step 2: Digital Datapath section). Set up the JESD204B links (see the Step 3: Transport Layer section). Set up the physical layer of the SERDES interface (see the Step 4: Physical Layer section). Set up the data link layer of the SERDES interface (see the Step 5: Data Link Layer section). Check for errors (see the Step 6: Optional Error Monitoring section). Optionally, enable any needed features as described in the Step 7: Optional Features section. The register writes listed in Table 15 to Table 21 give the register writes necessary to set up the AD9135/AD9136. Consider printing this setup guide and filling in the Value column with the appropriate variable values for the conditions of the desired application. The notation 0x, shaded in gray, indicates register settings that must be filled in by the user. To fill in the unknown register values, select the correct settings for each variable listed in the Variable column of Table 15 to Table 21. The Description column describes how to set variables or provides a link to a section where this is described. A variable is noted by concatenating multiple terms. For example, PdDACs is a variable corresponding to the value that is determined for Register 0x011[6:3] in the Device Setup Guide section. STEP 1: START UP THE DAC This section describes how to set up the SPI interface, power up necessary circuit blocks, write to the required configuration registers, and set up the DAC clocks, listed in Table 15. Table 15. Power-Up and DAC Initialization Settings Addr. 0x000 0x000 0x011 Bit No. 7 [6:3] 2 Value1 0xBD 0x3C 0x 0 Variable PdDACs 0x080 0 0x PdClocks 0x081 0x PdSysref 1 Description Soft reset. Deassert reset, set 4-wire SPI. Power up band gap. PdDACs = 0x05 to power up DAC0/DAC1. PdDACs = 0x07 if only using DAC0. Power up master DAC. PdClocks = 0 if DAC0/DAC1 are being used. PdClocks = 0x40 if only using DAC0. PdSysref = 0x00 for Subclass 1. PdSysref = 0x10 for Subclass 0. See the Subclass Setup section for details on subclass. The registers in Table 16 must be written from their default values to be the values listed in the table for the device to work correctly. These registers must be written after any soft reset, hard reset, or power-up occurs. Table 16. Required Device Configurations Addr. 0x12D 0x146 0x2A4 0x232 0x333 Value 0x8B 0x01 0xFF 0xFF 0x01 Description Digital datapath configuration Digital datapath configuration Clock configuration SERDES interface configuration SERDES interface configuration If using the optional DAC PLL, also set the registers in Table 17. Table 17. Optional DAC PLL Configuration Procedure Addr. 0x087 Value1 0x62 0x088 0xC9 0x089 0x0E 0x08A 0x12 0x08D 0x7B 0x1B0 0x00 0x1B9 0x24 0x1BC 0x0D 0x1BE 0x02 0x1BF 0x8E 0x1C0 0x2A 0x1C1 0x2A 0x1C4 0x08B 0x08C 0x085 Various 0x7E 0x 0x 0x 0x 0x083 0x10 1 Variable LODivMode RefDivMode BCount LookUpVals Description Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL charge pump settings Optimal DAC LDO settings for DAC PLL Power DAC PLL blocks when power machine disabled Optimal DAC PLL charge pump settings Optimal DAC PLL VCO control settings Optimal DAC PLL VCO power control settings Optimal DAC PLL VCO calibration settings Optimal DAC PLL lock counter length setting Optimal DAC PLL charge pump setting Optimal DAC PLL varactor settings See the DAC PLL Setup section See the DAC PLL Setup section See the DAC PLL Setup section See Table 25 in the DAC PLL Setup section for the list of register addresses and values for each. Enable the DAC PLL2 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the appropriate register value. 2 Verify that Register 0x084[1] reads back 1 after enabling the DAC PLL to indicate that the DAC PLL has locked. 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the appropriate register value. Rev. A | Page 26 of 117 Data Sheet AD9135/AD9136 STEP 2: DIGITAL DATAPATH Table 19. Transport Layer Settings This section describes which interpolation filters to use and how to set the data format being used. Additional digital features are available, including digital gain scaling and an inverse sinc filter used to improve pass-band flatness. Table 22 provides further details on the feature blocks available. Addr. 0x200 0x201 Bit No. 0x300 Value1 0x00 0x Variable UnusedLanes 0x 6 CheckSumMode 3 DualLink 2 CurrentLink Table 18. Digital Datapath Settings Addr. 0x112 Bit No. Variable InterpMode Description Select interpolation mode; see the Interpolation section. 0x450 0x DID DataFmt = 0 if twos complement; DataFmt = 1 if unsigned binary. 0x451 0x BID 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the appropriate register value. 0x452 0x LID 0x110 0x 7 1 Value1 0x DataFmt STEP 3: TRANSPORT LAYER 0x453 This section describes how to set up the JESD204B links. The parameters are determined by the desired JESD204B operating mode. See the JESD204B Setup section for details. Table 19 shows the register settings for the transport layer. If using dual-link mode, perform writes from Register 0x300 to Register 0x47D with CurrentLink = 0 and then repeat the same set of register writes with CurrentLink = 1 (Register 0x200 and Register 0x201 need only be written once). See the JESD204B Setup section. See the JESD204B Setup section. See the JESD204B Setup section. Set DID to match the device ID sent by the transmitter. Set BID to match the bank ID sent by the transmitter. Set LID to match the lane ID sent by the transmitter. 0x 7 Scrambling [4:0] L − 12 0x454 0x F − 12 0x455 0x K − 12 0x456 0x M − 12 0x457 0x458 0x 0x N − 12 [7:5] Subclass NP − 12 [4:0] 0x459 See the JESD204B Setup section. See the JESD204B Setup section. See the JESD204B Setup section. See the JESD204B Setup section. See the JESD204B Setup section. N = 16. See the JESD204B Setup section. NP = 16. 0x [7:5] JESDVer [4:0] S − 12 0x45A JESDVer = 1 for JESD204B, JESDVer = 0 for JESD204A. See the JESD204B Setup section. 0x 7 [4:0] HD 0x45D 0 0x CF Lane0Checksum 0x46C 0x Lanes 0x476 0x F 0x47D 0x Lanes 1 Description Power up the interface. See the JESD204B Setup section. See the JESD204B Setup section. CF must equal 0. See the JESD204B Setup section. Deskew lanes. See the JESD204B Setup section. See the JESD204B Setup section. Enable lanes. See the JESD204B Setup section. 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the correct register value. 2 This JESD204B link parameter is programmed in n − 1 notation as noted. For example, if the setup requires L = 8 (8 lanes per link), program L − 1 or 7 into Register 0x453[4:0]. Rev. A | Page 27 of 117 AD9135/AD9136 Data Sheet STEP 4: PHYSICAL LAYER STEP 5: DATA LINK LAYER This section describes how to set up the physical layer of the SERDES interface. In this section, the input termination settings are configured along with the CDR sampling and SERDES PLL. This section describes how to set up the data link layer of the SERDES interface. This section deals with SYSREF signal processing, setting deterministic latency, and establishing the link. Table 20. Device Configurations and Physical Layer Settings Addr. 0x2AA 0x2AB 0x2B1 0x2B2 0x2A7 0x2AE 0x314 0x230 Bit No. Table 21. Data Link Layer Settings Value1 0xB7 0x87 0xB7 0x87 0x01 0x01 0x01 0x 5 [4:2] 1 0x206 0x206 0x289 2 [1:0] Halfrate 0x2 OvSmp 0x00 0x01 0x 1 PLLDiv 0x284 0x285 0x286 0x287 0x62 0xC9 0x0E 0x12 0x28A 0x28B 0x7B 0x00 0x290 0x89 0x294 0x24 0x296 0x297 0x299 0x03 0x0D 0x02 0x29A 0x8E 0x29C 0x2A 0x29F 0x78 0x2A0 0x06 0x280 0x268 0x01 0x [7:6] [5:0] 1 Variable EqMode 0x22 Description SERDES interface termination setting SERDES interface termination setting Addr. 0x301 Autotune PHY setting Autotune PHY setting SERDES SPI configuration Set up the CDR; see the SERDES Clocks Setup section SERDES PLL default configuration Set up the CDR; see the SERDES Clocks Setup section Reset the CDR Release the CDR reset SERDES PLL configuration Set the CDR oversampling for PLL; see the SERDES Clocks Setup section Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL charge pump Optimal SERDES PLL VCO LDO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO Optimal SERDES PLL VCO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO varactor Optimal SERDES PLL VCO varactor Enable the SERDES PLL2 See the Equalization Mode Setup section Required value (default) Value1 0x Variable Subclass 0x304 0x LMFCDel 0x305 0x LMFCDel 0x306 0x LMFCVar 0x307 0x LMFCVar 0x03A 0x01 0x03A 0x81 0x03A 0xC1 SYSREF± Signal 0x308 to 0x30B 0x XBarVals 0x334 0x InvLanes 0x300 1 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the correct register value. 2 Verify that Register 0x281[0] reads back 1 after enabling the SERDES PLL to indicate that the SERDES PLL has locked. Bit No. 0x 6 3 2 CheckSumMode DualLink CurrentLink [1:0] EnLinks Description See the JESD204B Setup section. See the Link Latency Setup section. See the Link Latency section. See the Link Latency Setup section. See the Link Latency Setup section. Set sync mode = oneshot sync; see the Syncing LMFC Signals section for other sync options. Enable the sync machine. Arm the sync machine. If Subclass = 1, ensure that at least one SYSREF± edge is sent to the device.2 If remapping lanes, set up crossbar; see the Crossbar Setup section. Invert the polarity of the desired logical lanes. Bit x of InvLanes must be a 1 for each Logical Lane x to invert. Enable the links. See the JESD204B Setup section. Set to 0 to access Link 0 status or 1 for Link 1 status readbacks. See the JESD204B Setup section. EnLinks = 3 if DualLink = 1 (enables Link 0 and Link 1); EnLinks = 1 if DualLink = 0 (enables Link 0 only). 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the correct register value. 2 Verify that Register 0x03B[3] reads back 1 after sending at least one SYSREF± edge to the device to indicate that the LMFC sync machine has properly locked. Rev. A | Page 28 of 117 Data Sheet AD9135/AD9136 STEP 6: OPTIONAL ERROR MONITORING Table 22. Optional Features For JESD204B error monitoring, see the JESD204B Error Monitoring section. For other error checks, see the Interrupt Request Operation section. Feature Inverse Sinc Default On Digital Gain 2.7 dB DC Offset Off Group Delay 0 Downstream Protection Off Self Calibration Off STEP 7: OPTIONAL FEATURES There are a number of optional features that can be enabled. Table 22 provides links to the sections describing each feature. These features can be enabled during the Digital Datapath configuration step, or after the link is set up, because it is not required to configure them for the link to be established, unlike interpolation. Unless otherwise noted, these features are paged as described in the DAC Paging section. Paging is particularly important for DAC specific settings like digital gain and dc offset. Rev. A | Page 29 of 117 Description Improves pass-band flatness. See the Inverse Sinc section. Multiplies data by a factor. Can compensate inverse sinc usage or balance I/Q amplitude. See the Digital Gain section. Used to cancel LO leakage. See the DC Offset section. Used to control overall latency. See the Group Delay section. Used to protect downstream components. See the Downstream Protection section. Used to improve DAC linearity. Not paged by the dual paging register. See the Self Calibration section. AD9135/AD9136 Data Sheet DAC PLL SETUP INTERPOLATION This section explains how to select appropriate values for LODivMode, RefDivMode, and BCount in the Step 1: Start Up the DAC section. These parameters depend on the desired DAC clock frequency (fDAC) and DAC reference clock frequency (fREF). When using the DAC PLL, the reference clock signal is applied to the CLK± differential pins (Pin 2 and Pin 3). The transmit path can use zero to three cascaded interpolation filters, which each provides a 2× increase in output data rate and a low-pass function. Table 26 shows the different interpolation modes and the respective usable bandwidth along with the maximum fDATA rate attainable. Table 23. DAC PLL LODivMode Settings LO_DIV_MODE, Register 0x08B[1:0] 1 2 3 DAC Frequency Range (MHz) 1500 to 2800 750 to 1500 420 to 750 Table 24. DAC PLL RefDivMode Settings DAC PLL Reference Frequency (fREF) (MHz) 35 to 80 80 to 160 160 to 320 320 to 640 640 to 1000 Divide by (RefDivFactor) 1 2 4 8 16 REF_DIV_MODE, Register 0x08C[2:0] 0 1 2 3 4 The VCO frequency (fVCO) is related to the DAC clock frequency according to the following equation: fVCO = fDAC × 2LODivMode + 1 where 6 GHz fVCO 12 GHz. BCount must be between 6 and 127 and is calculated based on fDAC and fREF as follows: BCount = floor((fDAC)/(2 × fREF/RefDivFactor)) where RefDivFactor = 2RefDivMode (see Table 24). Finally, to finish configuring the DAC PLL, set the VCO control registers up as described in Table 25 based on the VCO frequency (fVCO). Write the registers listed in the table with the corresponding LookUpVals. Table 25. VCO Control Lookup Table Reference VCO Frequency Range (GHz) fVCO < 6.3 6.3 ≤ fVCO < 7.25 fVCO ≥ 7.25 Register 0x1B4 Setting 0x08 0x09 0x09 Register 0x1B6 Setting 0x03 0x03 0x13 Register 0x1BB Setting 0x07 0x06 0x06 For more information on the DAC PLL, see the DAC Input Clock Configurations section. Table 26. Interpolation Modes and Their Usable Bandwidth Interpolation Mode 1× (bypass) InterpMode 0x00 Usable Bandwidth 0.5 × fDATA 2× 0x01 0.4 × fDATA 4× 8× 0x03 0x04 0.4 × fDATA 0.4 × fDATA Maximum fDATA (MSPS) 2120 (SERDES limited) 1060 (SERDES limited) 700 350 The usable bandwidth is defined for 1×, 2×, 4×, and 8× modes as the frequency band over which the filters have a pass-band ripple of less than ±0.001 dB and an image rejection of greater than 85 dB. For more information, see the Interpolation Filters section. JESD204B SETUP This section explains how to select a JESD204B operating mode for a desired application. This section in turn defines appropriate values for CheckSumMode, UnusedLanes, DualLink, CurrentLink, Scrambling, L, F, K, M, N, NP, Subclass, S, HD, Lane0Checksum, and Lanes needed for the Step 3: Transport Layer section. Note that DualLink, Scrambling, F, K, N, NP, S, HD, and Subclass must be set the same on the transmit side. For Mode 8, Mode 9, and Mode 10, the number of converters (M) and the lane count (L) on the transmit side must also match the receive side. For Mode 11, Mode 12, and Mode 13, M and L on the transmit side do not match the receive side. See Table 28 for details. For a summary of how a JESD204B system works and what each parameter means, see the JESD204B Serial Data Interface section. Available Operating Modes Table 27. JESD204B Operating Modes (Single- or Dual-Link) (Applies to Both JESD204B Tx and Rx) Parameter M (Converter Count) L (Lane Count) S ((Samples per Converter) per Frame) F ((Octets per Frame) per Lane) 1 81 1 4 2 1 Mode 9 1 2 1 1 10 1 1 1 2 Mode 8 can only be used with 1× interpolation. Other interpolation options are not available in this mode. Rev. A | Page 30 of 117 Data Sheet AD9135/AD9136 Table 28. JESD204B Operating Modes (Single-Link Only) Parameter M (Converter Count) (Tx Setting) AD9135 and AD9136 M Setting1 (Rx Setting) L (Lane Count) (Tx Setting) AD9135 and AD9136 L Setting1 (Rx Setting) S ((Samples per Converter) per Frame) F ((Octets per Frame) per Lane) 112 2 1 Mode 12 2 1 13 2 1 8 4 4 2 2 1 2 1 1 1 1 2 1 Note that for Mode 11 to Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. 2 Mode 11 can only be used with 1× interpolation. Other interpolation options are not available in this mode. For a particular application, the number of converters to use per link (M) and the fDATA (DataRate) are known. The LaneRate and number of lanes (L) can be traded off as follows: DataRate = (DACRate)/(InterpolationFactor) LaneRate = (20 × DataRate × M)/L where LaneRate is between 1.44 Gbps and 10.64 Gbps. Octets per frame per lane (F) and samples per convertor per frame (S) define how the data is packed. If F = 1, the high density setting must be set to one (HD = 1). Otherwise, set HD = 0. Converter resolution and bits per sample (N and NP) must both be set to 16. Frames per multiframe (K) must be set to 32 for Mode 8, Mode 9, Mode 11, and Mode 12. Other modes can use either K = 16 or K = 32. DualLink CurrentLink Set CurrentLink to either 0 or 1 depending on whether Link 0 or Link 1, respectively, needs to be configured. Lanes Lanes is used to enable and deskew particular lanes in two thermometer coded registers. The lanes setting for each of the modes is given in Table 29. Table 29. Lanes Setting per JESD Operating Mode JESD Mode ID Lanes 8 0x0F 9 0x03 10 0x01 11 0xFF 12 0x33 13 0x11 UnusedLanes UnusedLanes is used to turn off unused circuit blocks to save power. Each physical lane that is not being used (SERDINx±) must be powered off by writing a 1 to the corresponding bit of Register 0x201. For example, if using Mode 9 in dual-link mode and sending data on SERDIN0±, SERDIN1±, SERDIN4±, and SERDIN5±, set UnusedLanes = 0xCC to power off Physical Lane 2, Lane 3, Lane 6, and Lane 7. CheckSumMode CheckSumMode must match the checksum mode used on the transmit side. If the checksum used is the sum of fields in the link configuration table, CheckSumMode = 0. If summing the registers containing the packed link configuration fields, CheckSumMode = 1. For more information on the how to calculate the two checksum modes, see the Lane0Checksum section. Lane0Checksum DualLink sets up two independent JESD204B links, which allows each link to be reset independently. If this functionality is desired, set DualLink to 1; if a single link is desired, set DualLink to 0. Note that Link 0 and Link 1 must have identical parameters. The operating modes available when using dual- or single-link mode are shown in Table 27. Additional single-link modes that are available are shown in Table 28. Scrambling Scrambling is a feature that makes the spectrum of the link data independent. This avoids spectral peaking and provides some protection against data dependent errors caused by frequency selective effects in the electrical interface. Set this variable to 1 if scrambling is being used, or to 0 if it is not. Subclass Subclass determines whether the latency of the device is deterministic, meaning it requires an external synchronization signal. See the Subclass Setup section for more information. Lane0Checksum can be used for error checking purposes to ensure that the transmitter is set up as expected. If CheckSumMode = 0, the checksum is the lower eight bits of the sum of the L − 1, M − 1, K − 1, N − 1, NP − 1, S − 1, Scrambling, HD, Subclass, and JESDVer variables. If CheckSumMode = 1, Lane0Checksum is the lower eight bits of the sum of Register 0x450 to Register 0x45A. Select whether to sum by fields or by registers, matching the setting on the transmitter. DAC Power-Down Setup As described in the Step 1: Start Up the DAC section, PdDACs must be set to 5 if both converters are being used either in a single- or dual-link mode. If only one DAC is being used (M = 1 and in single-link mode), PdDACs must be set to 7. Rev. A | Page 31 of 117 AD9135/AD9136 Data Sheet SERDES CLOCKS SETUP Link Delay Setup This section describes how to select the appropriate Halfrate, OvSmp, and PLLDiv settings in the Step 4: Physical Layer section. These parameters depend solely on the lane rate (the lane rate is established in the JESD204B Setup section). LMFCVar and LMFCDel are used to impose delays such that all lanes in a system arrive in the same LMFC cycle. Table 30. SERDES Lane Rate Configuration Settings Lane Rate (Gbps) 1.44 to 2.76 2.88 to 5.52 5.75 to 10.64 Halfrate 0 0 1 OvSmp 1 0 0 PLLDiv 2 1 0 The unit used internally for delays is the period of the internal processing clock (PClock), whose rate is 1/40th the lane rate. Delays that are not in PClock cycles must be converted before they are used. Some useful internal relationships are defined by PClockPeriod = 40/LaneRate Halfrate and OvSmp set how the clock detect and recover (CDR) circuit samples. See the SERDES PLL section for an explanation of how that circuit blocks works and the role of PLLDiv in the block. EQUALIZATION MODE SETUP Set EqMode = 1 for a low power setting. Select this mode if the insertion loss in the printed circuit board (PCB) is less than 12 dB. For insertion losses greater than 12 dB but less than 17.5 dB, set EqMode = 0. More details can be found in the Equalization section. LINK LATENCY SETUP This section describes the steps necessary to guarantee multichip deterministic latency in Subclass 1 and to guarantee synchronization of links within a device in Subclass 0. Use this section to fill in LMFCDel, LMFCVar, and Subclass in the Step 5: Data Link Layer section. For more information, see the Syncing LMFC Signals section. PClockPeriod can be used to convert from time to PClock cycles when needed. PClockFactor = 4/F (frames per PClock) PClockFactor is used to convert from units of PClock cycles to FrameClock cycles, which is needed to set LMFCDel in Subclass 1. PClocksPerMF= K/PClockFactor (PClocks per LMFC cycle) where PClocksPerMF is the number or PClock cycles in a multiframe cycle. The values for PClockFactor and PClockPerMF are given per JESD mode in Table 31. Table 31. PClockFactor and PClockPerMF JESD Mode ID PClockFactor PClockPerMF (K = 32) PClockPerMF (K = 16) 1 8 4 8 N/A1 9 4 8 N/A1 10 2 16 8 11 4 8 N/A1 12 4 8 N/A1 13 2 16 8 N/A means not applicable. Subclass Setup With Known Delays The AD9135/AD9136 support JESD204B Subclass 0 and Subclass 1 operation. With information about all the system delays, LMFCVar and LMFCDel can be calculated directly. Subclass 1 This mode gives deterministic latency and allows links to be synced to within ½ DAC clock periods. It requires an external SYSREF± signal that is accurately phase aligned to the DAC clock. Subclass 0 This mode does not require any signal on the SYSREF± pins, which can be left disconnected. Subclass 0 still requires that all lanes arrive within the same LMFC cycle and that the two DACs must be synchronized to each other; they are synchronized to an internal clock instead of to the SYSREF± signal. RxFixed (the fixed receiver delay in PClock cycles) and RxVar (the variable receiver delay in PClock cycles) can be found in Table 8. TxFixed (the fixed transmitter delay in PClock cycles) and TxVar (the variable receiver delay in PClock cycles) can be found in the data sheet of the transmitter used. PCBFixed (the fixed PCB trace delay in PClock cycles) can be extracted from software; because this is generally much smaller than a PClock cycle, it can also be omitted. For both the PCB and transmitter delays, convert the delays into PClock cycles. For each lane, Set Subclass to 0 or 1 as desired. MinDelayLane = floor(RxFixed + TxFixed + PCBFixed) MaxDelayLane = ceiling(RxFixed + RxVar + TxFixed + TxVar + PCBFixed)) Rev. A | Page 32 of 117 Data Sheet AD9135/AD9136 For safety, add a guard band of 1 PClock cycle to each end of the link delay as in the following equations: LMFCVar = (MaxDelay + 1) − (MinDelay − 1) where: MinDelay is the minimum of all MinDelayLane values across lanes, links, and devices. MaxDelay is the maximum of all MaxDelayLane values across lanes, links, and devices. Note that if LMFCVar must be more than 10, the AD9135/ AD9136 cannot tolerate the variable delay in the system. Table 32. Register Configuration and Procedure for OneShot Sync Addr. 0x301 0x03A Bit. No. Value1 0x Variable Subclass 0x01 0x03A 0x81 0x03A 0xC1 SYSREF± Signal For Subclass 1 LMFCDel = ((MinDelay − 1) × PClockFactor) % K For Subclass 0 0x 0x300 LMFCDel = (MinDelay − 1) % PClockPerMF Program the same LMFCDel and LMFCVar across all links and devices. See the Link Delay Setup Example, with Known Delays section for an example calculation. 6 CheckSumMode 3 DualLink 2 CurrentLink Set to 0 to access Link 0 status or 1 for Link 1 status readbacks. See the JESD204B Setup section. [1:0] EnLinks EnLinks = 3 if in DualLink mode to enable Link 0 and Link 1; EnLinks = 1 if not in DualLink mode to enable Link 0 Without Known Delays If comprehensive delay information is not available or known, the AD9135/AD9136 can read back the link latency between the local LMFC for each link (LMFCRX) and the last arriving LMFC boundary in PClock cycles. This information is then used to calculate LMFCVar and LMFCDel. For each link (on each device), 1. 2. 3. 4. 5. Power up the board. Follow the steps in Table 15 through Table 21 of the Device Setup Guide. Set the subclass and perform a sync. For one-shot sync, perform the writes in Table 32. See the Syncing LMFC Signals section for alternate sync modes. Record DYN_LINK_LATENCY_0 (Register 0x302) as a value of Delay for that link and power cycle. Record DYN_LINK_LATENCY_1 (Register 0x303) as a value of Delay for that link and power cycle the system. Repeat Step 1 to Step 5 twenty times for each device in the system. Keep a single list of the Delay values across all runs and devices. 1 Description Set subclass Set sync mode to one-shot sync Enable the sync machine Arm the sync machine If Subclass = 1, ensure that at least one SYSREF± edge is sent to the device Enable the links See the JESD204B Setup section See the JESD204B Setup section 0x denotes a register value that the user must fill in. See the Variable and Description columns for information on selecting the appropriate register value. The list of Delay values is used to calculate LMFCDel and LMFCVar, however, first some of the Delay values may need to be remapped. The maximum possible value for DYN_LINK_LATENCY_x is one less than the number of PClocks in a multiframe (PClocksPerMF). It is possible that a rollover condition may be encountered; that is, the set of recorded Delay values may roll over the edge of a multiframe. If so, Delay values may be near both 0 and PClocksPerMF. If this occurs, add PClocksPerMF to the set of values near 0. For example, for Delay value readbacks of 6, 7, 0, and 1, the 0 and 1 Delay values must be remapped to 8 and 9, making the new set of Delay values 6, 7, 8, and 9. Rev. A | Page 33 of 117 AD9135/AD9136 Data Sheet Across power cycles, links, and devices, CROSSBAR SETUP Register 0x308 to Register 0x30B allow arbitrary mapping of physical lanes (SERDINx±) to logical lanes used by the SERDES deframers. MinDelay is the minimum of all Delay measurements MaxDelay is the maximum of all Delay measurements For safety, a guard band of 1 PClock cycle is added to each end of the link delay and calculate LMFCVar and LMFCDel with the following equation: LMFCVar = (MaxDelay + 1) − (MinDelay − 1) Note that if LMFCVar must be more than 10, the AD9135/ AD9136 cannot tolerate the variable delay in the system. For Subclass 1 LMFCDel = ((MinDelay − 1) × PClockFactor) For Subclass 0 LMFCDel = (MinDelay − 1) % PClockPerMF Program the same LMFCDel and LMFCVar across all links and devices. See the Link Delay Setup Example, Without Known Delay section for an example calculation. Table 33. Crossbar Registers Address 0x308 0x308 0x309 0x309 0x30A 0x30A 0x30B 0x30B Bits [2:0] [5:3] [2:0] [5:3] [2:0] [5:3] [2:0] [5:3] Logical Lane LOGICAL_LANE0_SRC LOGICAL_LANE1_SRC LOGICAL_LANE2_SRC LOGICAL_LANE3_SRC LOGICAL_LANE4_SRC LOGICAL_LANE5_SRC LOGICAL_LANE6_SRC LOGICAL_LANE7_SRC Write each LOGICAL_LANEy_SRC with the number (x) of the desired physical lane (SERDINx±) from which to receive data. By default, all logical lanes use the corresponding physical lane as their data source. For example, by default LOGICAL_LANE0_ SRC = 0, meaning that Logical Lane 0 receives data from Physical Lane 0 (SERDIN0±). To use SERDIN4± as the source for Logical Lane 0, write LOGICAL_LANE0_SRC = 4. Rev. A | Page 34 of 117 Data Sheet AD9135/AD9136 JESD204B SERIAL DATA INTERFACE JESD204B OVERVIEW The AD9135/AD9136 have eight JESD204B data ports that receive data. The eight JESD204B ports can be configured as part of a single JESD204B link or as part of two separate JESD204B links (dual-link mode) that share a single system reference (SYSREF±) and device clock (CLK±). The JESD204B serial interface hardware consists of three layers: the physical layer, the data link layer, and the transport layer. These sections of the hardware are described in subsequent sections, including information for configuring every aspect of the interface. Figure 43 shows the communication layers implemented in the AD9135/AD9136 serial data interface to recover the clock and deserialize, descramble, and deframe the data before it is sent to the digital signal processing section of the device. The physical layer establishes a reliable channel between the transmitter and the receiver, the data link layer unpacks the data into octets and descrambles the data, and the transport layer receives the descrambled JESD204B frames and converts them to DAC samples. A number of JESD204B parameters (L, F, K, M, N, NP, S, HD, and Scrambling) defines how the data is packed and instruct the device how to turn the serial data into samples. These parameters are defined in detail in the Transport Layer section. Only certain combinations of parameters are supported. Each supported combination is called a mode. In total, six modes are supported by the AD9135/AD9136. There are three supported single-link modes, as described in Table 35, and three modes that can operate in either single- or dual-link mode, as described in Table 34. These tables show the associated clock rates when the lane rate is 10 Gbps. For a particular application, the number of converters to use (M) and DataRate are known. Calculate LaneRate and number of lanes (L) as follows: DataRate = (DACRate)/(InterpolationFactor) LaneRate = (20 × DataRate × M)/L where LaneRate must be between 1.44 Gbps and 10.64 Gbps. Achieving and recovering synchronization of the lanes is very important. To simplify the interface to the transmitter, the AD9135/AD9136 designate a master synchronization signal for each JESD204B link. In single-link mode, SYNCOUT0± is used as the master signal for all lanes; in dual-link mode, SYNCOUT0± is used as the master signal for Link 0, and SYNCOUT1± is used as the master signal for Link 1. If any lane in a link loses synchronization, a resynchronization request is sent to the transmitter via the synchronization signal of the link. The transmitter stops sending data and instead sends synchronization characters to all lanes in that link until resynchronization is achieved. SYNCOUT0± SYNCOUT1± PHYSICAL LAYER SERDIN0± DATA LINK LAYER DESERIALIZER I DATA[15:0] OR [11:0] QBD/ DESCRAMBLER FRAME TO SAMPLES TO DAC Q DATA[15:0] OR [11:0] DESERIALIZER 12578-004 SERDIN7± TRANSPORT LAYER SYSREF± Figure 43. Functional Block Diagram of Serial Link Receiver Rev. A | Page 35 of 117 AD9135/AD9136 Data Sheet Table 34. Single-Link and Dual-Link JESD204B Operating Modes Parameter M (Converter Counts) L (Lane Counts) S ((Samples per Converter) per Frame) F ((Octets per Frame) per Lane) Example Clocks for 10 Gbps Lane Rate PClock Rate (MHz) Frame Rate (MHz) Data Rate (MHz) 8 1 4 2 1 250 1000 2000 Mode 9 1 2 1 1 10 1 1 1 2 250 1000 1000 250 500 500 The AD9135/AD9136 autocalibrate the input termination to 50 Ω. Before running the termination calibration, write to Register 0x2AA, Register 0x2AB, Register 0x2B1, and Register 0x2B2 as described in Table 36 to guarantee proper calibration. The termination calibration begins when Register 0x2A7[0] and Register 0x2AE[0] transition from low to high. Register 0x2A7 controls autocalibration for PHY 0, PHY 1, PHY 6, and PHY 7. Register 0x2AE controls autocalibration for PHY 2, PHY 3, PHY 4, and PHY 5. The PHY termination autocalibration routine is shown in Table 36. Table 36. PHY Termination Autocalibration Routine 8 4 4 2 13 2 1 2 1 2 1 1 1 1 2 250 1000 2000 250 1000 1000 250 500 500 Note that for Mode 11 to Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. PHYSICAL LAYER The physical layer of the JESD204B interface, hereafter referred to as the deserializer, has eight identical channels. Each channel consists of the terminators, an equalizer, a clock and data recovery (CDR) circuit, and the 1:40 demux function (see Figure 45). Value 0xB7 0x87 0xB7 0x87 0x01 0x01 Description SERDES interface termination configuration SERDES interface termination configuration SERDES interface termination configuration SERDES interface termination configuration Autotune PHY terminations Autotune PHY terminations The input termination voltage of the DAC is sourced externally via the VTT pins (Pin 21, Pin 25, Pin 42, and Pin 46). Set VTT by connecting it to SVDD12. It is recommended that the JESD204B inputs be ac-coupled to the JESD204B transmit device using 100 nF capacitors. Receiver Eye Mask The AD9135/AD9136 comply with the JESD204B specification regarding the receiver eye mask and are capable of capturing data that complies with this mask. Figure 44 shows the receiver eye mask normalized to the data rate interval with a VTT swing of 600 mV. See the JESD204B specification for more information regarding the eye mask and permitted receiver eye opening. LV-OIF-11G-SR RECEIVER EYE MASK (3.125Mbps ≥ UI ≤ 12.5Gbps) 525 JESD204B data is input to the AD9135/AD9136 via the SERDINx± 1.2 V differential input pins as per the JESD204B specification. Interface Power-Up and Input Termination Before using the JESD204B interface, it must be powered up by setting Register 0x200[0] = 0. In addition, each physical lane that is not being used (SERDINx±) must be powered down. To do so, set the corresponding Bit x for Physical Lane x in Register 0x201 to 0 if the physical lane is being used, and to 1 if it is not being used. 55 0 –55 –525 0 0.35 0.5 0.65 TIME (UI) Figure 44. Receiver Eye Mask DESERIALIZER SERDINx± TERMINATION EQUALIZER CDR 1:40 SPI CONTROL FROM PLL Figure 45. Deserializer Block Diagram Rev. A | Page 36 of 117 12578-006 1 11 2 1 Address 0x2AA 0x2AB 0x2B1 0x2B2 0x2A7 0x2AE 1.00 12578-007 Parameter M (Converter Count) (Tx setting) AD9135 and AD9136 M Setting1 (Rx Setting) L (Lane Count) (Tx setting) AD9135 and AD9136 L Setting1 (Rx Setting) S ((Samples per Converter) per Frame) F ((Octets per Frame) per Lane) Example Clocks for 10 Gbps Lane Rate PClock Rate (MHz) Frame Rate (MHz) Data Rate (MHz) Mode 12 2 1 AMPLITUDE (mV) Table 35. Single-Link JESD204B Operating Modes Data Sheet AD9135/AD9136 Clock Relationships Register 0x280 controls the synthesizer enable and recalibration. The following clocks rates are used throughout the rest of the JESD204B section. The relationship between any of the clocks can be derived from the following equations: To enable the SERDES PLL, first set the PLL divider register according to Table 37, and then enable the SERDES PLL by writing 1 to Register 0x280[0]. DataRate = (DACRate)/(InterpolationFactor) Confirm that the SERDES PLL is working by reading Register 0x281. If Register 0x281[0] = 1, the SERDES PLL has locked. If Register 0x281[3] = 1, the SERDES PLL was successfully calibrated. If Register 0x281[4] or Register 0x281[5] are high, the PLL has reached the upper or lower end of its calibration band and must be recalibrated by writing 0 and then 1 to Register 0x280[2]. LaneRate = (20 × DataRate × M)/L ByteRate = LaneRate/10 where: M is the JESD204B parameter for converters per link. L is the JESD204B parameter for lanes per link. SERDES PLL Fixed Register Writes This relationship comes from 8-bit/10-bit encoding, where each byte is represented by 10 bits. To optimize the SERDES PLL across all operating conditions, the register writes in Table 38 are recommended. PClockRate = ByteRate/4 Table 38. SERDES PLL Fixed Register Writes The processing clock is used for a quad-byte decoder. FrameRate = ByteRate/F where F is defined as bytes per frame per lane. PClockFactor = FrameRate/PClockRate = 4/F where F is the JESD204B parameter for octets per frame per lane. SERDES PLL Functional Overview of the SERDES PLL The independent SERDES PLL uses integer-N techniques to achieve clock synthesis. The entire SERDES PLL is integrated on-chip, including the VCO and the loop filter. The SERDES PLL VCO operates over the range of 5.65 GHz to 11.04 GHz. In the SERDES PLL, a VCO divider block divides the VCO clock by 2 to generate a 2.825 GHz to 5.52 GHz quadrature clock for the deserializer cores. This clock is the input to the clock and data recovery block that is described in the Clock and Data Recovery section. The reference clock to the SERDES PLL is always running at a frequency, fREF, that is equal to 1/40 of the lane rate (PClockRate). This clock is divided by the DivFactor value to deliver a clock to the PFD block that is between 35 MHz and 80 MHz. Table 37 includes the respective SERDES_PLL_DIV_MODE register settings for each of the desired DivFactor options available. Address 0x284 0x285 0x286 0x287 0x28A 0x28B 0x290 0x294 0x296 0x297 0x299 0x29A 0x29C 0x29F 0x2A0 Divide by (DivFactor) 1 2 4 Description Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL charge pump Optimal SERDES PLL VCO LDO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO Optimal SERDES PLL VCO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO varactor Optimal SERDES PLL VCO varactor SERDES PLL IRQ SERDES PLL lock and lost signals are available as IRQ events. Use Register 0x01F[3:2] to enable these signals, and then use Register 0x023[3:2] to read back their statuses and reset the IRQ signals. See the Interrupt Request Operation section for more information. Table 37. SERDES PLL Divider Settings LaneRate (Gbps) 1.44 to 2.76 2.88 to 5.52 5.75 to 10.64 Value 0x62 0xC9 0x0E 0x12 0x7B 0x00 0x89 0x24 0x03 0x0D 0x02 0x8E 0x2A 0x78 0x06 SERDES_PLL_DIV_MODE, Register 0x289[1:0] 2 1 0 Rev. A | Page 37 of 117 AD9135/AD9136 Data Sheet 2.825GHz TO 5.52GHz OUTPUT VCO LDO CHARGE PUMP PFD 80MHz REF DIVIDER MAX N = 1, 2, 4 fREF BIT RATE ÷ 40 DIVIDER (1, 2, 4) C1 R1 UP C2 C3 I Q LC VCO 5.65GHz TO 11.04GHz ÷2 DOWN ÷80 R3 ALC CAL CAL CONTROL BITS 12578-011 FO CAL 3.2mA Figure 46. SERDES PLL Synthesizer Block Diagram Including VCO Divider Block Clock and Data Recovery Equalization The deserializer is equipped with a CDR circuit. Instead of recovering the clock from the JESD204B serial lanes, the CDR recovers the clocks from the SERDES PLL. The 2.825 GHz to 5.52 GHz output from the SERDES PLL, shown in Figure 46, is the input to the CDR. To compensate for signal integrity distortions for each PHY channel due to PCB trace length and impedance, the AD9135/ AD9136 employ an easy to use, low power equalizer on each JESD204B channel. The AD9135/AD9136 equalizers can compensate for insertion losses far greater than required by the JESD204B specification. The equalizers have two modes of operation that are determined by the EQ_POWER_MODE register setting in Register 0x268[7:6]. In low power mode (Register 0x268[7:6] = 2b’01) and operating at the maximum lane rate of 10 Gbps, the equalizer can compensate for up to 12 dB of insertion loss. In normal mode (Register 0x268[7:6] = 2b’00), the equalizer can compensate for up to 17.5 dB of insertion loss. This performance is shown in Figure 47 as an overlay to the JESD204B specification for insertion loss. Figure 47 shows the equalization performance at 10.0 Gbps, near the maximum baud rate for the AD9135/AD9136. A CDR sampling mode must be selected to generate the lane rate clock inside the device. If the desired lane rate is greater than 5.65 GHz, half rate CDR operation must be used. If the desired lane rate is less than 5.65 GHz, disable half rate operation. If the lane rate is less than 2.825 GHz, disable half rate operation and enable 2× oversampling to recover the appropriate lane rate clock. Table 39 gives a breakdown of CDR sampling settings that must be set dependent on the LaneRate. Table 39. CDR Operating Modes LaneRate (Gbps) 1.44 to 2.76 2.88 to 5.52 5.75 to 10.64 ENHALFRATE, Register 0x230[5] 0 0 1 CDR_OVERSAMP, Register 0x230[1] 1 0 0 The CDR circuit synchronizes the phase used to sample the data on each serial lane independently. This independent phase adjustment per serial interface ensures accurate data sampling and eases the implementation of multiple serial interfaces on a PCB. After configuring the CDR circuit, reset it and then release the reset by writing 1 and then 0 to Register 0x206[0]. Power-Down Unused PHYs Note that any unused and enabled lanes consume extra power unnecessarily. Each lane that is not being used (SERDINx±) must be powered off by writing a 1 to the corresponding bit of PHY_PD (Register 0x201). Figure 48 and Figure 49 are provided as points of reference for hardware designers and show the insertion loss for various lengths of well laid out stripline and microstrip transmission lines. See the Hardware Considerations section for specific layout recommendations for the JESD204B channel. Low power mode is recommended if the insertion loss of the JESD204B PCB channels is less than that of the most lossy supported channel for low power mode (shown in Figure 47). If the insertion loss is greater than that, but still less than that of the most lossy supported channel for normal mode (shown in Figure 47), use normal mode. At 10 Gbps operation, the equalizer in normal mode consumes about 4 mW more power per lane used than in low power equalizer mode. Note that either mode can be used in conjunction with transmitter preemphasis to ensure functionality and/or to optimize for power. Rev. A | Page 38 of 117 Data Sheet AD9135/AD9136 0 JESD204B SPEC ALLOWED CHANNEL LOSS 2 INSERTION LOSS (dB) 4 AD9135/AD9136 ALLOWED CHANNEL LOSS (LOW POWER MODE) 6 8 EXAMPLE OF AD9135/AD9136 COMPATIBLE CHANNEL (LOW POWER MODE) 10 12 EXAMPLE OF AD9135/AD9136 COMPATIBLE CHANNEL (NORMAL MODE) AD9135/AD9136 ALLOWED CHANNEL LOSS (NORMAL MODE) 14 16 DATA LINK LAYER EXAMPLE OF JESD204B COMPLIANT CHANNE L 18 20 22 24 5.0 7.5 12578-339 2.5 FREQUENCY (GHz) Figure 47. Insertion Loss Allowed 0 The AD9135/AD9136 can operate as a single-link or dual-link high speed JESD204B serial data interface. When operating in dual-link mode, configure both links with the same JESD204B parameters because they share a common device clock and system reference. All eight lanes of the JESD204B interface handle link layer communications such as code group synchronization, frame alignment, and frame synchronization. –5 ATTENUATION (dB) –10 –15 –20 STRIPLINE = 6” STRIPLINE = 10” STRIPLINE = 15” STRIPLINE = 20” STRIPLINE = 25” STRIPLINE = 30” –25 –30 –40 1 2 3 4 5 6 7 8 9 10 FREQUENCY (GHz) 12578-009 –35 0 Figure 48. Insertion Loss of 50 Ω Striplines on FR4 0 –5 –15 –20 6” MICROSTRIP 10” MICROSTRIP 15” MICROSTRIP 20” MICROSTRIP 25” MICROSTRIP 30” MICROSTRIP –30 –35 –40 0 1 2 3 4 5 6 7 8 9 FREQUENCY (GHz) 10 The AD9135/AD9136 decode 8-bit/10-bit control characters, allowing marking of the start and end of the frame and alignment between serial lanes. Each AD9135/AD9136 serial interface link can issue a synchronization request by setting its SYNCOUT0±/ SYNCOUT1± signal low. The synchronization protocol follows Section 4.9 of the JESD204B standard. When a stream of four consecutive /K/ symbols is received, the AD9135/AD9136 deactivate the synchronization request by setting the SYNCOUT0±/ SYNCOUT1± signal high at the next internal LMFC rising edge. Then, the AD9135/AD9136 wait for the transmitter to issue an ILAS. During the ILAS sequence, all lanes are aligned using the /A/ to /R/ character transition as described in the JESD204B Serial Link Establishment section. Elastic buffers hold early arriving lane data until the alignment character of the latest lane arrives. At this point, the buffers for all lanes are released and all lanes are aligned (see Figure 52). 12578-010 ATTENUATION (dB) –10 –25 The data link layer of the AD9135/AD9136 JESD204B interface accepts the deserialized data from the PHYs and deframes and descrambles them so that data octets are presented to the transport layer to be put into DAC samples. Figure 50 shows the link mode block diagrams for single-link and dual-link configurations and the interaction between the physical layer and logical layer. The DACs can only be configured in sequential order; for example, in Mode 10, when in single-link mode, the AD9135/AD9136 only uses Logical Lane 0 and DAC0. Logical lanes must be set according to Table 29 for the desired mode. See the Mode Configuration Maps section for further details on each of the mode configurations supported. The architecture of the data link layer is shown in Figure 51. The data link layer consists of a synchronization FIFO for each lane, a crossbar switch, a deframer, and descrambler. Figure 49. Insertion Loss of 50 Ω Microstrips on FR4 Rev. A | Page 39 of 117 AD9135/AD9136 Data Sheet LOGICAL LANE 0/LINK 0 LANE 0 SERDIN1±/PHYSICAL LANE 1 LOGICAL LANE 1/LINK 0 LANE 1 SERDIN2±/PHYSICAL LANE 2 SERDIN3±/PHYSICAL LANE 3 SERDIN4±/PHYSICAL LANE 4 SERDIN5±/PHYSICAL LANE 5 LOGICAL LANE 2/LINK 0 LANE 2 CROSSBAR REGISTER 0x308 TO REGISTER 0x30B DAC0 LOGICAL LANE 3/LINK 0 LANE 3 LOGICAL LANE 4/LINK 0 LANE 4 LOGICAL LANE 6/LINK 0 LANE 6 SERDIN7±/PHYSICAL LANE 7 LOGICAL LANE 7/LINK 0 LANE 7 LOGICAL LANE 0/LINK 0 LANE 0 SERDIN1±/PHYSICAL LANE 1 LOGICAL LANE 1/LINK 0 LANE 1 SERDIN2±/PHYSICAL LANE 2 LOGICAL LANE 2/LINK 0 LANE 2 SERDIN5±/PHYSICAL LANE 5 DAC1 LOGICAL LANES SERDIN0±/PHYSICAL LANE 0 SERDIN4±/PHYSICAL LANE 4 QUAD-BYTE DEFRAMER (QBD0) LOGICAL LANE 5/LINK 0 LANE 5 SERDIN6±/PHYSICAL LANE 6 SERDIN3±/PHYSICAL LANE 3 DAC CORE QBD SERDIN0±/PHYSICAL LANE 0 PHYSICAL LANES CROSSBAR REGISTER 0x308 TO REGISTER 0x30B DAC0 QUAD-BYTE DEFRAMER 1 (QBD1) DAC1 LOGICAL LANE 4/LINK 1 LANE 0 LOGICAL LANE 5/LINK 1 LANE 1 SERDIN6±/PHYSICAL LANE 6 LOGICAL LANE 6/LINK 1 LANE 2 SERDIN7±/PHYSICAL LANE 7 LOGICAL LANE 7/LINK 1 LANE 3 LOGICAL LANES 12578-550 PHYSICAL LANES QUAD-BYTE DEFRAMER 0 (QBD0) LOGICAL LANE 3/LINK 0 LANE 3 Figure 50. Link Mode Functional Diagram DATA LINK LAYER SYNCOUTx± QUAD-BYTE DEFRAMER QBD DES_DATA0 SERDIN0_CLK SERDIN0 FIFO CROSS BAR SWITCH SYSREF SERDIN7 FIFO LANE 0 OCTETS LANE 7 OCTETS SYSTEM CLOCK PHASE DETECT 12578-012 DES_DATA7 SERDIN7_CLK DESCRAMBLE DUAL-LINK MODE LOGICAL LAYER CROSSBAR 10-BIT/8-BIT DECODE SINGLE-LINK MODE PHYSICAL LAYER (PHY) PCLK SPI CONTROL Figure 51. Data Link Layer Block Diagram Rev. A | Page 40 of 117 Data Sheet AD9135/AD9136 L RECEIVE LANES (EARLIEST ARRIVAL) K K K R D D D D A R Q C L RECEIVE LANES (LATEST ARRIVAL) K K K K K K K R D D C D D A R Q C D D A R D D C D D A R D D 0 CHARACTER ELASTIC BUFFER DELAY OF LATEST ARRIVAL 4 CHARACTER ELASTIC BUFFER DELAY OF EARLIEST ARRIVAL L ALIGNED RECEIVE LANES K K K K K K K R D D D D A R Q C C 12578-013 K = K28.5 CODE GROUP SYNCHRONIZATION COMMA CHARACTER A = K28.3 LANE ALIGNMENT SYMBOL F = K28.7 FRAME ALIGNMENT SYMBOL R = K28.0 START OF MULTIFRAME Q = K28.4 START OF LINK CONFIGURATION DATA C = JESD204B LINK CONFIGURATION PARAMETERS D = Dx.y DATA SYMBOL D D A R D D Figure 52. Lane Alignment During ILAS JESD204B Serial Link Establishment A brief summary of the high speed serial link establishment process for Subclass 1 is provided. See Section 5.3.3 of the JESD204B specifications document for complete details. After the last /A/ character of the last ILAS, multiframe data begins streaming. The receiver adjusts the position of the /A/ character such that it aligns with the internal LMFC of the receiver at this point. Step 1: Code Group Synchronization Step 3: Data Streaming Each receiver must locate K (K28.5) characters in its input data stream. After four consecutive K characters are detected on all link lanes, the receiver block deasserts the SYNCOUTx± signal to the transmitter block at the receiver LMFC edge. In this phase, data is streamed from the transmitter block to the receiver block. The transmitter captures the change in the SYNCOUTx± signal, and at a future transmitter LMFC rising edge, starts the initial lane alignment sequence (ILAS). The receiver block processes and monitors the data it receives for errors, including Step 2: Initial Lane Alignment Sequence The main purposes of this phase are to align all the lanes of the link and to verify the parameters of the link. Before the link is established, write each of the link parameters to the receiver device to designate how data is sent to the receiver block. The ILAS consists of four or more multiframes. The last character of each multiframe is a multiframe alignment character, /A/. The first, third, and fourth multiframes are populated with predetermined data values. Note that Section 8.2 of the JESD204B specifications document describes the data ramp that is expected during ILAS. By default, the AD9135/AD9136 do not require this ramp. Register 0x47E[0] can be set high to require the data ramp. The deframer uses the final /A/ of each lane to align the ends of the multiframes within the receiver. The second multiframe contains an R (K28.0), Q (K28.4), and then data corresponding to the link parameters. Additional multiframes can be added to the ILAS if needed by the receiver. By default, the AD9135/AD9136 use four multiframes in the ILAS (this can be changed in Register 0x478). If using Subclass 1, exactly four multiframes must be used. Optionally, data can be scrambled. Scrambling does not start until the very first octet following the ILAS. Bad running disparity (8-bit/10-bit error) Not in table (8-bit/10-bit error) Unexpected control character Bad ILAS Interlane skew error (through character replacement) If any of these errors exist, they are reported back to the transmitter in one of a few ways (see the JESD204B Error Monitoring section for details). SYNCOUTx± signal assertion: resynchronization (SYNCOUTx± signal pulled low) is requested at each error for the last two errors. For the first three errors, an optional resynchronization request can be asserted when the error counter reaches a set error threshold. For the first three errors, each multiframe with an error in it causes a small pulse on SYNCOUTx±. Errors can optionally trigger an IRQ event, which can be sent to the transmitter. Various test modes for verifying the link integrity can be found in the JESD204B Test Modes section. Rev. A | Page 41 of 117 AD9135/AD9136 Data Sheet Lane FIFO The FIFOs in front of the crossbar switch and deframer synchronize the samples sent on the high speed serial data interface with the deframer clock by adjusting the phase of the incoming data. The FIFO absorbs timing variations between the data source and the deframer; this allows up to two PClock cycles of drift from the transmitter. The FIFO_STATUS_REG_0 register and FIFO_STATUS_REG_1 register (Register 0x30C and Register 0x30D, respectively) can be monitored to identify whether the FIFOs are full or empty. Lane FIFO IRQ An aggregate lane FIFO error bit is also available as an IRQ event. Use Register 0x01F[1] to enable the FIFO error bit, and then use Register 0x023[1] to read back its status and reset the IRQ signal. See the Interrupt Request Operation section for more information. Crossbar Switch Register 0x308 to Register 0x30B allow arbitrary mapping of physical lanes (SERDINx±) to logical lanes used by the SERDES deframers. Table 40. Crossbar Registers Address 0x308 0x308 0x309 0x309 0x30A 0x30A 0x30B 0x30B Bits [2:0] [5:3] [2:0] [5:3] [2:0] [5:3] [2:0] [5:3] Logical Lane LOGICAL_LANE0_SRC LOGICAL_LANE1_SRC LOGICAL_LANE2_SRC LOGICAL_LANE3_SRC LOGICAL_LANE4_SRC LOGICAL_LANE5_SRC LOGICAL_LANE6_SRC LOGICAL_LANE7_SRC In single-link mode, Deframer 0 is used exclusively and Deframer 1 remains inactive. In dual-link mode, both QBDs are active and must be configured separately using the LINK_PAGE bit (Register 0x300[2]) to select which link to configure. The LINK_MODE bit (Register 0x300[3]) is 1 for dual-link, or 0 for single-link. Each deframer uses the JESD204B parameters that the user has programmed into the register map to identify how the data has been packed and how to unpack it. The JESD204B parameters are described in detail in the Transport Layer section; many of the parameters are also needed in the transport layer to convert JESD204B frames into samples. Descrambler The AD9135/AD9136 provide an optional descrambler block using a self synchronous descrambler with a polynomial: 1 + x14 + x15. Enabling data scrambling reduces spectral peaks that are produced when the same data octets repeat from frame to frame. It also makes the spectrum data independent so that possible frequency selective effects on the electrical interface do not cause data dependent errors. Descrambling of the data is enabled by setting the SCR bit (Register 0x453[7]) to 1. Syncing LMFC Signals The first step in guaranteeing synchronization across links and devices begins with syncing the LMFC signals. Each DAC has its own LMFC signal. In Subclass 0, the LMFC signals for each of the two DACs are synchronized to an internal processing clock. In Subclass 1, all LMFC signals (for all DACs and devices) are synchronized to an external SYSREF signal. All LMFC sync registers are paged as described in the DAC Paging section. SYSREF Signal Write each LOGICAL_LANEy_SRC with the number (x) of the desired physical lane (SERDINx±) from which to receive data. By default, all logical lanes use the corresponding physical lane as their data source. For example, by default LOGICAL_LANE0_ SRC = 0; thus, Logical Lane 0 receives data from Physical Lane 0 (SERDIN0±). If instead the user wants to use SERDIN4± as the source for Logical Lane 0, the user must write LOGICAL_LANE0_ SRC = 4. Lane Inversion Register 0x334 allows inversion of desired logical lanes, which can be used to ease routing of the SERDINx± signals. For each Logical Lane x, set Bit x of Register 0x334 to 1 to invert it. Deframers The AD9135/AD9136 consist of two quad-byte deframers (QBDs). Each deframer receives the 8-bit/10-bit encoded data from the deserializer (via the crossbar switch), decodes it, and descrambles it into JESD204B frames before passing it to the transport layer to be converted to DAC samples. The deframer processes four symbols (or octets) per processing clock (PClock) cycle. The SYSREF signal is a differential source synchronous input that synchronizes the LMFC signals in both the transmitter and receiver in a JESD204B Subclass 1 system to achieve deterministic latency. The SYSREF signal is an active high signal that is sampled by the device clock rising edge. It is best practice that the device clock and SYSREF signals be generated by the same source, such as the AD9516-1 clock generator, so that the phase alignment between the signals is fixed. When designing for optimum deterministic latency operation, consider the timing distribution skew of the SYSREF signal in a multipoint link system (multichip). The AD9135/AD9136 support a single pulse or step, or a periodic SYSREF± signal. The periodicity can be continuous, strobed, or gapped periodic. The SYSREF± signal can always be dc-coupled (with a common-mode voltage of 0 V to 2 V). When dc-coupled, a small amount of common-mode current (<500 μA) is drawn from the SYSREF± pins. See Figure 53 for the SYSREF± internal circuit. Rev. A | Page 42 of 117 Data Sheet AD9135/AD9136 To avoid this common-mode current draw, a 50% duty-cycle periodic SYSREF± signal can be used with ac coupling capacitors. If ac-coupled, the ac coupling capacitors combine with the resistors shown in Figure 53 to make a high-pass filter with a RC time constant, τ = RC. Select C such that τ > 4/SYSREF frequency. In addition, the edge rate must be sufficiently fast— at least 1.3 V/ns is recommended per Table 5—to meet the SYSREF± vs. DAC clock keepout window (KOW) requirements. Continuous mode differs from one-shot mode in two ways. First, no SPI cycle is required to arm the device; the alignment edge seen after continuous mode is enabled results in a phase check. Second, a phase check (and when necessary, clock rotation) occurs on every alignment edge in continuous mode. The one caveat to the previous statement is that when a phase rotation cycle is underway, subsequent alignment edges are ignored until the logic lane is ready again. It is possible to use ac-coupled mode without meeting the frequency to time-constant constraint by using SYSREF± hysteresis (Register 0x081 and Register 0x082). However, this increases the DAC clock KOW (Table 5 does not apply) by an amount depending on SYSREF± frequency, level of hysteresis, capacitor choice, and edge rate. The maximum acceptable phase error (in DAC clock cycles) between the alignment edge and the LMFC edge is set in the error window tolerance register. If continuous sync mode is used with a nonzero error window tolerance, a phase check occurs on every SYSREF± pulse, but an alignment occurs only if the phase error is greater than the specified error window tolerance. If the jitter of the SYSREF signal violates the KOW specification given in Table 5 and therefore causes phase error uncertainty, the error tolerance can be increased to avoid constant clock rotations. Note that this means the latency is less deterministic by the size of the window. 1.2V 2kΩ SYSREF– 2kΩ ~800mV 12578-015 SYSREF+ Figure 53. SYSREF± Input Circuit Sync Processing Modes Overview The AD9135/AD9136 support various LMFC sync processing modes. These modes are one-shot, continuous, windowed continuous, and monitor modes. All sync processing modes perform a phase check to see that the LMFC is phase aligned to an alignment edge. In Subclass 1, the SYSREF pulse acts as the alignment edge; in Subclass 0, an internal processing clock acts as the alignment edge. If the signals are not in phase, a clock rotation occurs to align the signals. The sync modes are described in the following sections. See the Sync Procedure section for details on the procedure for syncing the LMFC signals. One-Shot Sync Mode (SYNCMODE = 0x1) In one-shot sync mode, a phase check occurs on only the first alignment edge that is received after the sync machine is armed. If the phase error is larger than a specified window error tolerance, a phase adjustment occurs. Though an LMFC synchronization occurs only once, the SYSREF signal can still be continuous. Continuous Sync Mode (SYNCMODE = 0x2) Continuous mode can only be used in Subclass 1 with a periodic SYSREF signal. In continuous mode, a phase check/alignment occurs on every alignment edge. For debug purposes, SYNCARM (Register 0x03A[6]) can be used to inform the user that alignment edges are being received in continuous mode. Because the SYNCARM bit is self cleared after an alignment edge is received, the user can arm the sync (SYNCARM (Register 0x03A[6]) = 1), and then read back SYNCARM. If SYNCARM = 0, the alignment edges are being received and phase checks are occurring. Arming the sync machine in this mode does not affect the operation of the device. One-Shot Then Monitor Sync Mode (SYNCMODE = 0x9) In one-shot then monitor mode, the user can monitor the phase error in real time. Use this sync mode with a periodic SYSREF signal. A phase check and alignment occurs on the first alignment edge received after the sync machine is armed. On all subsequent alignment edges, the phase is monitored and reported, but no clock phase adjustment occurs. The phase error can be monitored on the SYNC_CURRERR_L register (Register 0x03C[3:0]). Immediately after an alignment occurs, CURRERROR = 0 indicates that there is no difference between the alignment edge and the LMFC edge. On every subsequent alignment edge, the phase is checked. If the alignment is lost, the phase error is reported in the SYNC_ CURRERR_L register in DAC clock cycles. If the phase error is beyond the selected window tolerance (Register 0x034[2:0]), one bit of Register 0x03D[7:6] is set high depending on whether the phase error is on the low or high side. When an alignment occurs, snapshots of the last phase error (Register 0x03C[3:0]) and the corresponding error flags (Register 0x03D[7:6]) are placed into readable registers for reference (Register 0x038 and Register 0x039, respectively). Rev. A | Page 43 of 117 AD9135/AD9136 Data Sheet Sync Procedure LMFC Sync IRQ The procedure for enabling the sync is as follows: The sync status bits (SYNC_LOCK, SYNC_ROTATE, SYNC_TRIP, and SYNC_WLIM) are available as IRQ events. 1. 2. 3. 4. 5. 6. 7. Set Register 0x008 to 0x03 to sync the LMFC for both DAC0 and DAC1. Set the desired sync processing mode. The sync processing mode settings are listed in Table 41. For Subclass 1, set the error window according to the uncertainty of the SYSREF signal relative to the DAC clock and the tolerance of the application for deterministic latency uncertainty. Sync window tolerance settings are given in Table 42. Enable sync by writing 1 to SYNCENABLE (Register 0x03A[7]). If in one-shot mode, arm the sync machine by writing 1 to SYNCARM (Register 0x03A[6]). If in Subclass 1, ensure that at least one SYSREF± pulse is sent to the device. Check the status by reading the following bit fields: a) SYNC_BUSY (Register 0x03B[7]) = 0 to indicate that the sync logic is no longer busy. b) SYNC_LOCK (Register 0x03B[3]) = 1 to indicate that the signals are aligned. This bit updates on every phase check. c) SYNC_WLIM (Register 0x03B[1]) = 0 to indicate that the phase error is not beyond the specified error window. This bit updates on every phase check. d) SYNC_ROTATE (Register 0x03B[2]) = 1. If the phases were not aligned before the sync and an alignment occurred, this bit indicates that a clock alignment occurred. This bit is sticky and can be cleared only by writing to the SYNCCLRSTKY control bit (Register 0x03A[5]). e) SYNC_TRIP (Register 0x03B[0]) = 1 to indicate that the alignment edge was received and a phase check occurred. This bit is sticky and can be cleared only by writing to the SYNCCLRSTKY control bit (Register 0x03A[5]). Table 41. Sync Processing Modes Sync Processing Mode One-shot Continuous One-shot then monitor SYNCMODE (Register 0x03A[3:0]) 0x01 0x02 0x09 Use Register 0x021[3:0] to enable the sync status bits for DAC0 and then use Register 0x025[3:0] to read back their statuses and to reset the IRQ signals. Use Register 0x022[3:0] to enable the sync status bits for DAC1 and then use Register 0x026[3:0] to read back their statuses and to reset the IRQ signals. See the Interrupt Request Operation section for more information. Deterministic Latency JESD204B systems contain various clock domains distributed throughout each system. Data traversing from one clock domain to a different clock domain can lead to ambiguous delays in the JESD204B link. These ambiguities lead to nonrepeatable latencies across the link from power cycle to power cycle with each new link establishment. Section 6 of the JESD204B specification addresses the issue of deterministic latency with mechanisms defined as Subclass 1 and Subclass 2. The AD9135/AD9136 support JESD204B Subclass 0 and Subclass 1 operation, but not Subclass 2. Write the subclass to Register 0x301[2:0] and once per link to Register 0x458[7:5]. Subclass 0 This mode does not require any signal on the SYSREF± pins, which can be left disconnected. Subclass 0 still requires that all lanes arrive within the same LMFC cycle and that the two DACs be synchronized to each other. Minor Subclass 0 Caveats Because the AD9135/AD9136 require an ILAS, the nonmultiple converter device alignment single lane (NMCDA-SL) case from the JESD204A specification is supported only when using the optional ILAS. Error reporting using SYNCOUTx± is not supported when using Subclass 0 with F = 1. Subclass 1 This mode gives deterministic latency and allows links to be synced to within ½ of a DAC clock period. It requires an external SYSREF± signal that is accurately phase aligned to the DAC clock. Table 42. Sync Window Tolerance Sync Error Window Tolerance ±½ DAC clock cycles ±1 DAC clock cycles ±2 DAC clock cycles ±3 DAC clock cycles ERRWINDOW (Register 0x034[2:0]) 0x00 0x01 0x02 0x03 Rev. A | Page 44 of 117 Data Sheet AD9135/AD9136 account for any amount of fixed delay. As a result, the LMFC period must only be larger than the variation in the link delays, and the AD9135/AD9136 can achieve proper performance with a smaller total latency. Figure 54 and Figure 55 show a case where the link delay is larger than an LMFC period. Note that it can be accommodated by delaying LMFCRx. DETERMINISTIC LATENCY REQUIREMENTS Several key factors are required for achieving deterministic latency in a JESD204B Subclass 1 system. LMFC ILAS ALIGNED DATA DATA LATE ARRIVING LMFC REFERENCE EARLY ARRIVING LMFC REFERENCE Link Delay Figure 54. Link Delay > LMFC Period Example POWER CYCLE VARIANCE The link delay of a JESD204B system is the sum of fixed and variable delays from the transmitter, channel, and receiver as shown in Figure 56. LMFC ILAS ALIGNED DATA For proper functioning, all lanes on a link must be read during the same LMFC period. Section 6.1 of the JESD204B specification states that the LMFC period must be larger than the maximum link delay. For the AD9135/AD9136, this is not necessarily the case; instead, the AD9135/AD9136 use a local LMFC for each link (LMFCRx) that can be delayed from the SYSREF aligned LMFC. Because the LMFC is periodic, this can DATA LMFCRX LMFC_DELAY_x LMFC REFERENCE FOR ALL POWER CYCLES FRAME CLOCK Figure 55. LMFC_DELAY_x, to Compensate for Link Delay > LMFC LINK DELAY = DELAYFIXED + DELAYVARIABLE LOGIC DEVICE (JESD204B Tx) CHANNEL JESD204B Rx DSP DAC POWER CYCLE VARIANCE LMFC ALIGNED DATA CGS ILAS DATA FIXED DELAY VARIABLE DELAY Figure 56. JESD204B Link Delay = Fixed Delay + Variable Delay Rev. A | Page 45 of 117 12578-019 POWER CYCLE VARIANCE 12578-018 SYSREF± signal distribution skew within the system must be less than the desired uncertainty. SYSREF± setup and hold time requirements must be met for each device in the system. The total latency variation across all lanes, links, and devices must be ≤10 PClock periods. This includes both variable delays and the variation in fixed delays from lane to lane, link to link, and device to device in the system. 12578-017 AD9135/AD9136 Data Sheet 1. The method to set the LMFCDel and LMFCVar values is described in the Link Delay Setup section. Setting LMFCDel appropriately ensures that all the corresponding data samples arrive in the same LMFC period. Then LMFCVar is written into the receive buffer delay (RBD) to absorb all link delay variation. This ensures that all data samples have arrived before reading. By setting these to fixed values across runs and devices, deterministic latency is achieved. 2. The RBD described in the JESD204B specification takes values from 1 frame clock cycle to K frame clock cycles, whereas the RBD of the AD9135/AD9136 take values from 0 PClock cycles to 10 PClock cycles. As a result, up to 10 PClock cycles of total delay variation can be absorbed. Because LMFCVar is in PClock cycles, and LMFCDel is in frame clock cycles, a conversion between these two units is needed. The PClockFactor, or number of frame clock cycles per PClock cycle, is equal to 4/F. For more information on this relationship, see the Clock Relationships section. 3. 4. Two examples follow that show how to determine LMFCVar and LMFCDel. After they are calculated, write LMFCDel into both Register 0x304 and Register 0x305 for all devices in the system, and write LMFCVar to both Register 0x306 and Register 0x307 for all devices in the system. 5. Link Delay Setup Example, with Known Delays All the known system delays can be used to calculate LMFCVar and LMFCDel as described in the Link Delay Setup section. 6. The example shown in Figure 57 is demonstrated in the following steps according to the procedure outlined in the Link Delay Setup section. Note that this example is in Subclass 1 to achieve deterministic latency, which has a PClockFactor (4/F) of two frame clock cycles per PClock Cycle, and uses K = 32 (frames/multiframe). Because PCBFixed << PClockPeriod, PCBFixed is negligible in this example and not included in the calculations. 7. Find the receiver delays using Table 8. RxFixed = 17 PClock cycles RxVar = 2 PClock cycles Find the transmitter delays. The equivalent table in the example JESD204B core (implemented on a GTH or GTX transceiver on a Virtex-6 FPGA) states that the delay is 56 ± 2 byte clock cycles. Because the PClockRate = ByteRate/4 as described in the Clock Relationships section, the transmitter delays in PClock cycles are as follows: TxFixed = 54/4 = 13.5 PClock cycles TxVar = 4/4 = 1 PClock cycle Calculate MinDelayLane as follows: MinDelayLane = floor(RxFixed + TxFixed + PCBFixed) = floor(17 + 13.5 + 0) = floor(30.5) MinDelayLane = 30 Calculate MaxDelayLane as follows: MaxDelayLane = ceiling(RxFixed + RxVar + TxFixed + TxVar + PCBFixed)) = ceiling(17 + 2 + 13.5 + 1 + 0) = ceiling(33.5) MaxDelayLane = 34 Calculate LMFCVar as follows: LMFCVar = (MaxDelay + 1) − (MinDelay − 1) = (34 + 1) − (30 − 1) = 35 − 29 LMFCVar = 6 PClock cycles Calculate LMFCDel as follows: LMFCDel = ((MinDelay − 1) × PClockFactor) % K = ((30 − 1) × 2) % 32 = (29 × 2) % 32 = 58 % 32 LMFCDel = 26 frame clock cycles Write LMFCDel to both Register 0x304 and Register 0x305 for all devices in the system. Write LMFCVar to both Register 0x306 and Register 0x307 for all devices in the system. LMFC PCLK FCLK DYN_LINK_LATENCY_CNT 0 1 2 ALIGNED LANE DATA (A) ALIGNED LANE DATA (B) ALIGNED LANE DATA (C) 3 4 5 6 7 0 1 2 3 4 5 6 7 DATA ILAS ILAS DATA ILAS DATA LMFCRX ILAS LMFC_DELAY_x = 12 (FCLK CYCLES) DATA LMFC_VAR = 6 (PCLK CYCLES) Figure 57. LMFC_DELAY_x Calculation Example Rev. A | Page 46 of 117 12578-024 DETERMINISTICALLY DELAYED DATA Data Sheet AD9135/AD9136 Link Delay Setup Example, Without Known Delay If the system delays are not known, the AD9135/AD9136 can read back the link latency between LMFCRX for each link and the SYSREF± aligned LMFC. This information is then used to calculate LMFCVar and LMFCDel, as shown in the Without Known Delays section. Figure 59 shows how DYN_LINK_LATENCY_x (Register 0x302 and Register 0x303) provides a readback showing the delay (in PClock cycles) between LMFCRX and the transition from ILAS to the first data sample. By repeatedly power cycling and taking this measurement, the minimum and maximum delays across power cycles can be determined and used to calculate LMFCVar and LMFCDel. 2. 3. 4. The example shown in Figure 59 is demonstrated in the following steps according to the procedure outlined in the Without Known Delays section. Note that this example is in Subclass 1 to achieve deterministic latency, which has a PClockFactor (frame clock rate/ PClockRate) of 2 and uses K = 16; therefore, PClocksPerMF = 8. In Figure 59, for Link A, Link B, and Link C, the system containing the AD9135/AD9136 (including the transmitter) is power cycled and configured 20 times. The AD9135/AD9136 are configured as described in the Device Setup Guide. Because the point of this exercise is to determine LMFCDel and LMFCVar, the LMFCDel is programmed to 0 and DYN_LINK_LATENCY_x is read from Register 0x302 and Register 0x303 for Link 0 and 6. SYSREF LMFCRX ILAS ALIGNED DATA DATA 12578-022 DYN_LINK_LATENCY Figure 58. DYN_LINK_LATENCY Illustration LMFC PCLK FRAME CLOCK DYN_LINK_LATENCY_CNT 0 1 2 ALIGNED LANE DATA (A) ALIGNED LANE DATA (B) ALIGNED LANE DATA (C) 3 4 5 6 7 8 9 10 11 ILAS 12 13 14 15 DATA ILAS DATA ILAS DATA LMFCRX DETERMINISTICALLY DELAYED DATA ILAS LMFC_DELAY = 6 (FCLK CYCLES) DATA LMFC_VAR = 7 (PCLK CYCLES) Figure 59. Multilink Synchronization Settings, Derived Method Example Rev. A | Page 47 of 117 12578-025 1. 5. Link 1, respectively. The variation in the link latency over the 20 runs is shown in Figure 59 in gray. Link A gives readbacks of 6, 7, 0, and 1. Note that the set of recorded delay values rolls over the edge of a multiframe at the boundary K/PClockFactor = 8. Add PClocksPerMF = 8 to low set. Delay values range from 6 to 9. Link B gives Delay values from 5 to 7. Link C gives Delay values from 4 to 7. Calculate the minimum of all Delay measurements across all power cycles, links, and devices: MinDelay = min(all Delay values) = 4 Calculate the maximum of all Delay measurements across all power cycles, links, and devices: MaxDelay = max(all Delay values) = 9 Calculate the total Delay variation (with guard band) across all power cycles, links, and devices: LMFCVar = (MaxDelay + 1) − (MinDelay − 1) = (9 + 1) − (4 − 1) = 10 − 3 = 7 PClock cycles Calculate the minimum delay in frame clock cycles (with guard band) across all power cycles, links, and devices: LMFCDel = ((MinDelay − 1) × PClockFactor) % K = ((4 − 1) × 2) % 16 = (3 × 2) % 16 = 6 % 16 = 6 frame clock cycles Write LMFCDel to both Register 0x304 and Register 0x305 for all devices in the system. Write LMFCVar to both Register 0x306 and Register 0x307 for all devices in the system. AD9135/AD9136 Data Sheet TRANSPORT LAYER TRANSPORT LAYER (QBD) LANE 0 OCTETS DELAY BUFFER 0 F2S_0 DAC0 [15:0] OR [11:0] DELAY BUFFER 1 F2S_1 DAC1 [15:0] OR [11:0] LANE 3 OCTETS PCLK_0 SPI CONTROL LANE 4 OCTETS LANE 7 OCTETS 12578-026 PCLK_1 SPI CONTROL Figure 60. Transport Layer Block Diagram The transport layer receives the descrambled JESD204B frames and converts them to DAC samples based on the programmed JESD204B parameters shown in Table 43. A number of device parameters are defined in Table 44. Table 44. JESD204B Device Parameters Parameter CF Table 43. JESD204B Transport Layer Parameters CS Parameter F K HD L M S Description Number of octets per frame per lane: 1, 2, or 4. Number of frames per multiframe. K = 32 if F = 1, K = 16 or 32 otherwise. Number of lanes per converter device (per link), as follows: 1, 2, 4, or 8 (single-link mode). 1, 2, or 4 (dual-link mode). Number of converters per device (per link), as follows: 1 or 2 (single-link mode). 1 (dual-link mode). Number of samples per converter, per frame: 1 or 2. N Nʹ (or NP) Description Number of control words per device clock per link. Not supported, must be 0. Number of control bits per conversion sample. Not supported, must be 0. High density user data format. Used when samples must be split across lanes. Set to 1 when F = 1, otherwise 0. Converter resolution = 16. Total number of bits per sample = 16. Certain combinations of these parameters, called JESD204B operating modes, are supported by the AD9135/AD9136. See Table 45 and Table 46 for a list of supported modes, along with their associated clock relationships. Rev. A | Page 48 of 117 Data Sheet AD9135/AD9136 Table 45. Single-Link and Dual-Link JESD204B Operating Modes Mode Parameter M (Converter Count) L (Lane Count) S (Samples per Converter per Frame) F (Octets per Frame per Lane) K2 (Frames per Multiframe) HD (High Density) N (Converter Resolution) NP (Bits per Sample) Example Clocks for 10 Gbps Lane Rate PClock Rate (MHz) Frame Clock Rate (MHz) Data Rate (MHz) 1 2 81 1 4 2 1 32 1 16 16 9 1 2 1 1 32 1 16 16 10 1 1 1 2 16 or 32 0 16 16 250 250 250 250 1000 1000 250 500 500 Mode 8 can only be used with 1× interpolation. Other interpolation options are not available in this mode. K must be 32 in Mode 8 and Mode 9. It can be 16 or 32 in Mode 10. Table 46. Single-Link JESD204B Operating Modes Mode Parameter M (Converter Count) AD9135/AD9136 M Setting2 L (Lane Count) AD9135/AD9136 L Setting2 S (Samples per Converter per Frame) F (Octets per Frame, per Lane) K3 (Frames per Multiframe) HD (High Density) N (Converter Resolution) NP (Bits per Sample) Example Clocks for 10 Gbps Lane Rate PClock Rate (MHz) Frame Clock Rate (MHz) Data Rate (MHz) 111 2 1 8 4 2 1 32 1 16 16 12 2 1 4 2 1 1 32 1 16 16 13 2 1 2 1 1 2 16 or 32 0 16 16 250 250 250 250 250 250 250 1000 1000 1 Mode 11 can only be used with 1× interpolation. Other interpolation options are not available in this mode. Note that for Mode 11 through Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. 3 K must be 32 in Mode 11 and Mode 12. It can be 16 or 32 in Mode 13. 2 Rev. A | Page 49 of 117 AD9135/AD9136 Data Sheet Configuration Parameters Data Flow Through the JESD204B Receiver The AD9135/AD9136 modes refer to the link configuration parameters for L, K, M, N, NP, S, and F. Table 47 provides the description and addresses for these settings. The link configuration parameters determine how the serial bits on the JESD204B receiver interface are deframed and passed on to the DACs as data samples. Figure 61 shows a detailed flow of the data through the various hardware blocks for Mode 11 (L = 8, M = 2, S = 2, F = 1). Simplified flow diagrams for all other modes are shown in Figure 62 through Figure 66. Table 47. Configuration Parameters JESD204B Setting L−1 F − 11 K−1 M−1 N−1 NP − 1 S−1 HD F1 DID BID LID0 JESDV 1 Description Number of lanes − 1. Number of ((octets per frame) per lane) − 1. Number of frames per multiframe − 1. Number of converters − 1. Converter bit resolution − 1. Bit packing per sample − 1. Number of ((samples per converter) per frame) − 1. High density format. Set to 1 if F = 1. Leave at 0 if F ≠ 1. F parameter, in ((octets per frame) per lane). Device ID. Match the device ID sent by the transmitter. Bank ID. Match the bank ID sent by the transmitter. Lane ID for Lane 0. Match the lane ID sent by the transmitter on Logical Lane 0. JESD204x version. Match the version sent by the transmitter (0x0 = JESD204A, 0x1 = JESD204B). Address 0x453[4:0] 0x454[7:0] 0x455[4:0] 0x456[7:0] 0x457[4:0] 0x458[4:0] 0x459[4:0] 0x45A[7] 0x476[7:0] 0x450[7:0] 0x451[3:0] 0x452[4:0] 0x459[7:5] Single- and Dual-Link Configuration The AD9135/AD9136 use the settings contained in Table 45 and Table 46. Mode 8 to Mode 13 can be used for single-link operation. Mode 8 to Mode 10 can also be used for dual-link operation. To use dual-link mode, set LINK_MODE (Register 0x300[3]) to 1. In dual-link mode, Link 1 must be programmed with identical parameters to Link 0. To write to Link 1, set LINK_PAGE (Register 0x300[2]) to 1. If single-link mode is being used, a small amount of power can be saved by powering down the output buffer for SYNCOUT1±, which can be done by setting Register 0x203[0] = 1. Checking Proper Configuration As a convenience, the AD9135/AD9136 provide some quick configuration checks. Register 0x030[5] is high if an illegal LMFC_DELAY value is used. Register 0x030[3] is high if an unsupported combination of L, M, F, or S is used. Register 0x030[2] is high if an illegal K is used. Register 0x030[1] is high if an illegal SUBCLASSV is used. Deskewing and Enabling Logical Lanes The values that need to be written in Register 0x454 and Register 0x476 are different, F − 1 and F, respectively. After proper configuration, the logical lanes must be deskewed and enabled to capture data. Set Bit x in Register 0x46C to 1 to deskew Logical Lane x and to 0 if that logical lane is not being used. Then, set Bit x in Register 0x47D to 1 to enable Logical Lane x and to 0 if that logical lane is not being used. Rev. A | Page 50 of 117 Data Sheet AD9135/AD9136 S0 S19 J19 J8 J1 J10 DESERIALIZER S10 S9 SERDINx± J9 J8 J1 J0 DESERIALIZER S0 SERDINx± J19 J18 S19 J11 J10 S10 S9 J1 J0 DESERIALIZER S0 S19 SERDINx± J19 J8 J1 J10 DESERIALIZER S10 S9 SERDINx± J9 J8 J1 J0 DESERIALIZER S0 SERIAL JESD204B DATA (L = 8) SAMPLES SPLIT ACROSS LANES (HD = 1) D0 DAC0 D15 D0 DESCRAMBLER DESERIALIZER SERDINx± J9 J8 10-BIT/8-BIT DECODE D15 D15 D0 DAC1 D15 D0 2 CONVERTERS (M = 2) 2 SAMPLES PER CONVERTER PER FRAME (S = 2) 40 BITS PARALLEL DATA (ENCODED AND SCRAMBLED) 16-BIT NIBBLE GROUP (N = 16) 1 OCTET PER LANE (F = 1) 12578-350 SERDINx± CONV0, SMPL0 DESERIALIZER CONV1, SMPL0 J0 NIBBLE GROUP 0 J1 NIBBLE GROUP 0 S10 S9 CONV0, SMPL1 DESERIALIZER SERDINx± J9 J8 LANE 3 LANE 2 LANE 1 LANE 0 OCTET 0 OCTET 0 OCTET 0 OCTET 0 S19 J11 J10 TRANSPORT LAYER LANE 7 LANE 6 LANE 5 LANE 4 OCTET 0 OCTET 0 OCTET 0 OCTET 0 J19 J18 DATA LINK LAYER CONV1, SMPL1 PHYSICAL LAYER SERDINx± Figure 61. JESD204B Mode 11 Data Deframing Mode Configuration Maps Table 48 to Table 53 contain the SPI configuration map for each mode shown in Figure 61 through Figure 66. Figure 61 through Figure 66 show the associated data flow through the deframing process of the JESD204B receiver for each of the modes. Mode 8 to Mode 13 apply to single-link operation. Mode 8 to Mode 10 also apply to dual-link operation. Register 0x300 must be set accordingly for single- or dual-link operation. For additional details regarding all the SPI registers, see the Register Maps and Descriptions section. Rev. A | Page 51 of 117 AD9135/AD9136 Data Sheet Table 48. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 8 LANE 0, OCTET 0 J0 J1 SERDINx± J7 LANE 2, OCTET 0 J6 J8 J9 SERDINx± J1 SERDINx± LANE 1, OCTET 0 J6 J15 J14 1 OCTET PER LANE (F = 1) 16-BIT NIBBLE GROUP (N = 16) 2 SAMPLES PER CONVERTER PER FRAME (S = 2) J7 SERIAL JESD204B DATA (L = 4) SAMPLES SPLIT ACROSS LANES (HD = 1) J9 J8 J0 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x3: L = 4 lanes per link Register 0x454[7:0] = 0x00: F = 1 octet per frame Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x1: S = 2 (samples per converter) per frame Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0 to Link Lane 3 Register 0x476[7:0] = 0x01: F = 1 octet per frame Register 0x47D[7:0] = 0x0F: enable Link Lane 0 to Link Lane 3 SERDINx± LANE 3, OCTET 0 NIBBLE GROUP 0 CONVERTER 0, SAMPLE 0 CONVERTER 0, SAMPLE 1 D15 ... D0 (0) D15 ... D0 (1) 1 CONVERTER (M = 1) DAC0 12578-352 Setting 0x03 or 0x83 0x00 0x1F 0x00 0x0F 0x0F or 0x2F 0x21 0x80 0x0F 0x01 0x0F J15 J14 Address 0x453 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D Figure 62. JESD204B Mode 8 Data Deframing Rev. A | Page 52 of 117 Data Sheet AD9135/AD9136 Table 49. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 9 J0 J1 J8 SERDINx± LANE 0, OCTET 0 LANE 1, OCTET 0 NIBBLE GROUP 0 CONVERTER 0, SAMPLE 0 1 CONVERTER (M = 1) D15 ... D0 DAC0 Figure 63. JESD204B Mode 9 Data Deframing Rev. A | Page 53 of 117 12578-035 1 OCTET PER LANE (F = 1) 16-BIT NIBBLE GROUP (N = 16) 1 SAMPLE PER CONVERTER PER FRAME (S = 1) J9 SERIAL JESD204B DATA (L = 2) SAMPLES SPLIT ACROSS LANES (HD = 1) J11 J10 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x1: L = 2 lanes per link Register 0x454[7:0] = 0x00: F = 1 octet per frame Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample per converter) per frame Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0 and Link Lane 1 Register 0x476[7:0] = 0x01: F = 1 octet per frame Register 0x47D[7:0] = 0x03: enable Link Lane 0 and Link Lane 1 J19 J18 Setting 0x01 or 0x81 0x00 0x1F 0x00 0x0F 0x0F or 0x2F 0x20 0x80 0x03 0x01 0x03 SERDINx± Address 0x453 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D AD9135/AD9136 Data Sheet Table 50. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 10 2 OCTETS PER LANE (F = 2) 16-BIT NIBBLE GROUP (N = 16) 1 SAMPLE PER CONVERTER PER FRAME (S = 1) LANE 0, OCTET 0 LANE 1, OCTET 0 NIBBLE GROUP 0 CONVERTER 0, SAMPLE 0 D15 ... D0 1 CONVERTER (M = 1) DAC0 Figure 64. JESD204B Mode 10 Data Deframing Rev. A | Page 54 of 117 12578-036 SERIAL JESD204B DATA (L = 1) SAMPLES SPLIT ACROSS LANES (HD = 0) J1 J0 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x0: L = 1 lane per link Register 0x454[7:0] = 0x01: F = 2 octets per frame Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample per converter) per frame Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0 Register 0x476[7:0] = 0x02: F = 2 octets per frame Register 0x47D[7:0] = 0x01: enable Link Lane 0 J19 J18 Setting 0x00 or 0x80 0x01 0x0F or 0x1F 0x00 0x0F 0x0F or 0x2F 0x20 0x00 0x01 0x02 0x01 SERDINx± Address 0x453 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D Data Sheet AD9135/AD9136 Table 51. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 11 Address 0x453 Setting 0x03 or 0x83 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D 0x00 0x1F 0x00 0x0F 0x0F or 0x2F 0x21 0x80 0xFF 0x01 0xFF 1 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x3: L = 4 lanes per link (L = 8 on transmit side)1 Register 0x454[7:0] = 0x00: F = 1 octet per frame Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link (M = 2 on transmit side)1 Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x1: S = 2 (samples per converter) per frame Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0 to Link Lane 7 Register 0x476[7:0] = 0x01: F = 1 octet per frame Register 0x47D[7:0] = 0x0F: enable Link Lane 0 to Link Lane 7 Note that for Mode 11 through Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. See Figure 61 for an illustration of the AD9135/AD9136 JESD204B Mode 11 data deframing process. Rev. A | Page 55 of 117 AD9135/AD9136 Data Sheet Table 52. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 12 Address 0x453 Setting 0x01 or 0x81 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D 0x00 0x1F 0x00 0x0F 0x0F or 0x2F 0x20 0x80 0x33 0x01 0x33 2 CONVERTERS (M = 2) LANE 1, OCTET 0 J1 J9 J8 SERDINx± J0 J11 J10 SERDINx± J19 J18 J1 J8 SERDINx± J0 J11 J10 LANE 0, OCTET 0 LANE 2, OCTET 0 LANE 3, OCTET 0 NIBBLE GROUP 0 NIBBLE GROUP 1 CONVERTER 0, SAMPLE 0 CONVERTER 1, SAMPLE 0 D15 ... D0 D15 ... D0 DAC0 DAC1 12578-354 1 OCTET PER LANE (F = 1) 16-BIT NIBBLE GROUP (N = 16) 1 SAMPLE PER CONVERTER PER FRAME (S = 1) J9 SERIAL JESD204B DATA (L = 4) SAMPLES SPLIT ACROSS LANES (HD = 1) SERDINx± Note that for Mode 11 through Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. J19 J18 1 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x1: L = 2 lanes per link (L = 4 on transmit side)1 Register 0x454[7:0] = 0x00: F = 1 octet per frame Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link (M = 2 on transmit side)1 Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample per converter) per frame Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0, Link Lane 1, Link Lane 4, and Link Lane 5 Register 0x476[7:0] = 0x01: F = 1 octet per frame Register 0x47D[7:0] = 0x03: enable Link Lane 0, Link Lane 1, Link Lane 4, and Link Lane 5 Figure 65. JESD204B Mode 12 Data Deframing Rev. A | Page 56 of 117 Data Sheet AD9135/AD9136 Table 53. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 13 Address 0x453 Setting 0x00 or 0x80 0x454 0x455 0x456 0x457 0x458 0x459 0x45A 0x46C 0x476 0x47D 0x01 0x0F or 0x1F 0x00 0x0F 0x0F or 0x2F 0x20 0x00 0x11 0x02 0x11 J0 J1 J19 J18 DAC0 D0 D1 D4 D5 D6 D7 D8 D9 D10 D11 D12 D13 D14 D15 D0 D1 D2 D3 D4 D5 D6 D7 D8 D9 CONVERTER 1, SAMPLE 0 D10 CONVERTER 0, SAMPLE 0 D2 LANE 1, OCTET 1 D3 J1 LANE 1, OCTET 0 NIBBLE GROUP 1 D11 D12 D13 LANE 0, OCTET 1 NIBBLE GROUP 0 DAC1 12578-033 2 CONVERTERS (M = 2) LANE 0, OCTET 0 D14 2 OCTETS PER LANE (F = 2) 16-BIT NIBBLE GROUP (N = 16) 1 SAMPLE PER CONVERTER PER FRAME (S = 1) SERDINx± SERIAL JESD204B DATA (L = 2) SAMPLES NOT SPLIT ACROSS LANES (HD = 0) J19 J18 SERDINx± J0 Note that for Mode 11 through Mode 13, the M and L parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. D15 1 Description Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x0: L = 1 lane per link (L = 2 on transmit side)1 Register 0x454[7:0] = 0x01: F = 2 octets per frame Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe Register 0x456[7:0] = 0x00: M = 1 converter per link (M = 2 on transmit side)1 Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample per converter) per frame Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0 Register 0x46C[7:0] = 0x0F: deskew Link Lane 0 and Link Lane 4 Register 0x476[7:0] = 0x02: F = 2 octets per frame Register 0x47D[7:0] = 0x01: enable Link Lane 0 and Link Lane 4 Figure 66. JESD204B Mode 13 Data Deframing Rev. A | Page 57 of 117 AD9135/AD9136 Data Sheet JESD204B TEST MODES Transport Layer Testing PHY PRBS Testing The JESD204B receiver in the AD9135/AD9136 supports the short transport layer (STPL) test as described in the JESD204B standard. This test can be used to verify the data mapping between the JESD204B transmitter and receiver. To perform this test, this function must be implemented in the logic device and enabled there. Before running the test on the receiver side, the link must be established and running without errors (see the Device Setup Guide). The JESD204B receiver on the AD9135/AD9136 includes a PRBS pattern checker on the back end of its physical layer. This functionality enables bit error rate (BER) testing of each physical lane of the JESD204B link. The PHY PRBS pattern checker does not require that the JESD204B link be established. The pattern checker can synchronize with a PRBS7, PRBS15, or PRBS31 data pattern. PRBS pattern verification can be performed on multiple lanes at once. The error counts for failing lanes are reported for one JESD204B lane at a time. The process for performing PRBS testing on the AD9135/AD9136 is as follows: 1. 2. 3. 4. 5. 6. 7. 8. 9. Start sending a PRBS7, PRBS15, or PRBS31 pattern from the JESD204B transmitter. Select and write the appropriate PRBS pattern to Register 0x316[3:2], as shown in Table 54. Enable the PHY test for all lanes being tested by writing to PHY_TEST_EN (Register 0x315). Each bit of Register 0x315 enables the PRBS test for the corresponding lane. For example, writing a 1 to Bit 0 enables the PRBS test for Physical Lane 0. Toggle PHY_TEST_RESET (Register 0x316[0]) from 0 to 1 then back to 0. Set PHY_PRBS_ERROR_THRESHOLD (Register 0x317 to Register 0x319) as desired. Write a 0 and then a 1 to PHY_TEST_START (Register 0x316[1]). The rising edge of PHY_TEST_START starts the test. Wait 500 ms. Stop the test by writing PHY_TEST_START (Register 0x316[1]) = 0. Read the PRBS test results. a. Each bit of PHY_PRBS_PASS (Register 0x31D) corresponds to one SERDES lane: 0 is fail, 1 is pass. b. The number of PRBS errors seen on each failing lane can be read by writing the lane number to check (0 to 7) in the PHY_SRC_ERR_CNT (Register 0x316[6:4]) and reading the PHY_PRBS_ERR_CNT (Register 0x31A to Register 0x31C). The maximum error count is 224 − 1. If all bits of Register 0x31A to Register 0x31C are high, the maximum error count on the selected lane has been exceeded. Table 54. PHY PRBS Pattern Selection PHY_PRBS_PAT_SEL Setting, (Register 0x316[3:2]) 0b00 (default) 0b01 0b10 PRBS Pattern PRBS7 PRBS15 PRBS31 The STPL test ensures that each sample from each converter is mapped appropriately according to the number of converters (M) and the number of samples per converter (S). As specified in the JESD204B standard, the converter manufacturer specifies what test samples are transmitted. Each sample must have a unique value. For example, if M = 2 and S = 2, there are 4 unique samples transmitted repeatedly until the test is stopped. The expected sample must be programmed into the device and the expected sample is compared to the received sample one sample at a time until all have been tested. The process to perform this test on the AD9135/AD9136 is described as follows: 1. 2. 3. 4. 5. 6. 7. 8. Synchronize JESD204B link. Enable the STPL test at the JESD204B transmitter. Select Converter 0 Sample 0 for testing. Write SHORT_TPL_DAC_SEL (Register 0x32C[3:2]) = 0 and SHORT_TPL_SP_SEL (Register 0x32C[5:4]) = 0. Set the expected test sample for Converter 0, Sample 0. Program the expected 11-/16-bit test sample into the SHORT_TPL_REF_SP_x registers (Register 0x32E and Register 0x32D). Enable the STPL test. Write SHORT_TPL_TEST_EN (Register 0x32C[0]) = 1. Toggle the STPL reset. SHORT_TPL_TEST_RESET (Register 0x32C[1]) from 0 to 1 then back to 0. Check for failures. Read SHORT_TPL_FAIL (Register 0x32F[0]): 0 is pass, 1 is fail. Repeat Step 3 to Step 7 for each sample of each converter, Conv0Sample0 through ConvM − 1SampleS − 1. Repeated CGS and ILAS Test As per Section 5.3.3.8.2 of the JESD204B specification, the AD9135/AD9136 can check that a constant stream of /K28.5/ characters is being received, or that CGS followed by a constant stream of ILAS is being received. To run a repeated CGS test, send a constant stream of /K28.5/ characters to the AD9135/AD9136 SERDES inputs. Next, set up the device and enable the links as described in the Device Setup Guide section. Ensure that the /K28.5/ characters are being received by verifying that the SYNCOUTx± has been deasserted and that CGS has passed for all enabled link lanes by reading Register 0x470. Program Register 0x300[2] = 0 to monitor the status of lanes on Link 0, and Register 0x300[2] = 1 to monitor the status of lanes on Link 1 for dual-link mode. Rev. A | Page 58 of 117 Data Sheet AD9135/AD9136 To run the CGS followed by a repeated ILAS sequence test, follow the Device Setup Guide section; however, before performing the last write (enabling the links), enable the ILAS test mode by writing a 1 to Register 0x477[7]. Then, enable the links. When the device recognizes four CGS characters on each lane, it deasserts SYNCOUTx±. At this point, the transmitter starts sending a repeated ILAS sequence. Read Register 0x473 to verify that initial lane synchronization has passed for all enabled link lanes. Program Register 0x300[2] = 0 to monitor the status of lanes on Link 0, and Register 0x300[2] = 1 to monitor the status of lanes on Link 1 for dual-link mode. JESD204B ERROR MONITORING Check for Error Count Over Threshold In addition to reading the error count per lane and error type as described in the Checking Error Counts section, the user can check a register to see if the error count for a given error type has reached a programmable threshold. The same error threshold is used for the three error types (disparity, not in table, and unexpected control character). The error counters are on a per error type basis. To use this feature, complete the following steps: 1. 2. Disparity, Not in Table, and Unexpected Control Character Errors As per Section 7.6 of the JESD204B specification, the AD9135/AD9136 can detect disparity errors, not in table errors, and unexpected control character errors, and can optionally issue a sync request and reinitialize the link when errors occur. Note that the disparity error counter counts all characters with invalid disparity, regardless of whether they are in the 8-bit/10-bit decoding table. This is a minor deviation from the JESD204B specification, which only counts disparity errors when they are in the 8-bit/10-bit decoding table. Checking Error Counts The error count can be checked for disparity errors, not in table errors, and unexpected control character errors. The error counts are on a per lane and per error type basis. Note that the lane select and counter select are programmed into Register 0x46B and the error count is read back from the same address. To check the error count, complete the following steps: 1. 2. Select the desired link lane and error type of the counter to view. Write these to Register 0x46B according to Table 55. To select a link lane, first select a link (Register 0x300[2] = 0 to select Link 0 or Register 0x300[2] = 1 to select Link 1 (dual link only)). Note that when using Link 1, Link Lane x refers to Logical Lane x + 4. Read the error count from Register 0x46B. Note the maximum error count is equal to the error threshold set in Register 0x47C. Table 55. Error Counters Addr. 0x46B Bits [6:4] Variable LaneSel [1:0] CntrSel Description LaneSel = x to monitor the error count of Link Lane x. See the notes on link lane in Step 1 of the Checking Error Counts section. CntrSel = 0b00 for bad running disparity counter. CntrSel = 0b01 for not in table error counter. CntrSel = 0b10 for unexpected control character counter. Program the desired error count threshold into ERRORTHRES (Register 0x47C). Read back the error status for each error type to see if the error count has reached the error threshold. Disparity errors are reported in Register 0x46D. Not in table errors are reported in Register 0x46E. Unexpected control character errors are reported in Register 0x46F. Error Counter and IRQ Control The user can write to Register 0x46D and Register 0x46F to reset or disable the error counts and to reset the IRQ for a given lane. Note that these are the same registers that are used to report error count over threshold (see the Check for Error Count Over Threshold section); therefore, the readback is not the value that was written. For each error type, 1. 2. 3. Select the link lane to access. To select a link lane, first select a link (Register 0x300[2] = 0 to select Link 0, Register 0x300[2] = 1 to select Link 1 (dual link only)). Note that when using Link 1, Link Lane x refers to Logical Lane x + 4. Decide whether to reset the IRQ, disable the error count, and/or reset the error count for the given lane and error type. Write the link lane and desired reset or disable action to Register 0x46D to Register 0x46F according to Table 56. Table 56. Error Counter and IRQ Control: Disparity (Register 0x46D), Not In Table (Register 0x46E), Unexpected Control Character (Register 0x46F) Bits 7 Variable RstIRQ 6 Disable_ErrCnt 5 RstErrCntr [2:0] LaneAddr Rev. A | Page 59 of 117 Description RstIRQ = 1 to reset IRQ for the lane selected in Bits[2:0]. Disable_ErrCnt = 1 to disable the error count for the lane selected in Bits[2:0]. RsteErrCntr = 1 to reset the error count for the lane selected in Bits[2:0]. LaneAddr = x to monitor the error count of Link Lane x. See the notes on link lane in Step 1 of the Checking Error Counts section. AD9135/AD9136 Data Sheet Monitoring Errors via SYNCOUTx± Table 58. Sync Assertion Mask When one or more disparity, not in table, or unexpected control character error occurs, the error is reported on the SYNCOUTx± pins as per Section 7.6 of the JESD204B specification. The JESD204B specification states that the SYNCOUTx± signal is asserted for exactly two frame periods when an error occurs. For the AD9135/AD9136, the width of the SYNCOUTx± pulse can be programmed to ½, 1, or 2 PClock cycles. The settings to achieve a SYNCOUTx± pulse of 2 frame clock cycles are given in Table 57. Addr. 0x47B Table 57. Setting SYNCOUTx± Error Pulse Duration JESD Mode IDs 8, 9, 11, 12 10, 13 1 PClockFactor (Frames/PClock) 4 2 5 UCC_S CGS, Frame Sync, Checksum, and ILAS Monitoring Register 0x470 to Register 0x473 can be monitored to verify that each stage of the JESD204B link establishment has occurred. Program Register 0x300[2] = 0 to monitor the status of the lanes on Link 0, and Register 0x300[2] = 1 to monitor the status of the lanes on Link 1. See the Error Counter and IRQ Control section for information on resetting the IRQ. See the Interrupt Request Operation section for more information on IRQs. Errors Requiring Reinitializing A link reinitialization automatically occurs when four invalid disparity characters are received as per Section 7.1 of the JESD204B specification. When a link reinitialization occurs, the resync request is five frames and nine octets long. The user can optionally reinitialize the link when the error count for disparity errors, not in table errors, or unexpected control characters reaches a programmable error threshold. The process to enable the reinitialization feature for certain error types is as follows: 4. NIT_S 0 (default) 1 For disparity, not in table, and unexpected control character errors, error count over the threshold events are available as IRQ events. Enable these events by writing to Register 0x47A[7:5]. The IRQ event status can be read at the same address (Register 0x47A[7:5]) after the IRQs are enabled. 3. 6 Description Set to 1 to assert SYNCOUTx± if the disparity error count reaches the threshold Set to 1 to assert SYNCOUTx± if the not in table error count reaches the threshold Set to 1 to assert SYNCOUTx± if the unexpected control character count reaches the threshold SYNCB_ERR_DUR Disparity, Not in Table, and Unexpected Control Character IRQs 2. Bit Name BADDIS_S (Register 0x312[5:4]) Setting1 These register settings assert the SYNCOUTx± signal for 2 frame clock cycle pulse widths. 1. Bit No. 7 Set THRESHOLD_MASK_EN (Register 0x477[3]) = 1. Note that when this bit is set, unmasked errors do not saturate at either threshold or maximum value. Enable the sync assertion mask for each type of error by writing to the SYNCASSERTIONMASK register (Register 0x47B[7:5]) according to Table 58. Program the desired error counter threshold into ERRORTHRES (Register 0x47C). For each error type enabled in the SYNCASSERTIONMASK register, if the error counter on any lane reaches the programmed threshold, SYNCOUTx± falls, issuing a sync request. Note that all error counts are reset when a link reinitialization occurs. The IRQ does not reset and must be reset manually. Bit x of CODEGRPSYNCFLAG (Register 0x470) is high if Link Lane x received at least four K28.5 characters and passed code group synchronization. Bit x of FRAMESYNCFLAG (Register 0x471) is high if Link Lane x completed initial frame synchronization. Bit x of GOODCHKSUMFLG (Register 0x472) is high if the checksum sent over the lane matches the sum of the JESD204B parameters sent over the lane during ILAS for Link Lane x. The parameters can be added either by summing the individual fields in registers or summing the packed register. If Register 0x300[6] = 0 (default), the calculated checksums are the lower eight bits of the sum of the following fields: DID, BID, LID, SCR, L − 1, F − 1, K − 1, M − 1, N − 1, SUBCLASSV, NP − 1, JESDV, S − 1, and HD. If Register 0x300[6] = 1, the calculated checksums are the lower eight bits of the sum of Register 0x400 to Register 0x40C and LID. Bit x of INITLANESYNCFLG (Register 0x473) is high if Link Lane x passed the initial lane alignment sequence. CGS, Frame Sync, Checksum, and ILAS IRQs Fail signals for CGS, frame sync, checksum, and ILAS are available as IRQ events. Enable them by writing to Register 0x47A[3:0]. The IRQ event status can be read at the same address (Register 0x47A[3:0]) after the IRQs are enabled. Write a 1 to Register 0x470[7] to reset the CGS IRQ. Write a 1 to Register 0x471 to reset the frame sync IRQ. Write a 1 to Register 0x472 to reset the checksum IRQ. Write a 1 to Register 0x473 to reset the ILAS IRQ. See the Interrupt Request Operation section for more information. Rev. A | Page 60 of 117 Data Sheet AD9135/AD9136 Configuration Mismatch IRQ Power and Ground Planes The AD9135/AD9136 have a configuration mismatch flag that is available as an IRQ event. Use Register 0x47B[3] to enable the mismatch flag (it is enabled by default), and then use Register 0x47B[4] to read back its status and reset the IRQ signal. See the Interrupt Request Operation section for more information. Solid ground planes are recommended to avoid ground loops and to provide a solid, uninterrupted ground reference for the high speed transmission lines that require controlled impedances. Do not use segmented power planes as a reference for controlled impedances unless the entire length of the controlled impedance trace traverses across only a single segmented plane. These and additional guidelines for the topology of high speed transmission lines are described in the JESD204B Serial Interface Inputs (SERDIN0± to SERDIN7±) section. The configuration mismatch event flag is high when the link configuration settings (in Register 0x450 to Register 0x45D) do not match the JESD204B transmitted settings (Register 0x400 to Register 0x40D). All these registers are paged per link (in Register 0x300). For Mode 11 through Mode 13, the configuration mismatch flag is high because the values for the M and L parameters sent over the link do not match the parameters programmed to Register 0x453 and Register 0x456. Table 59. Power Supplies Note that this function is different from the good checksum flags in Register 0x472. The good checksum flags ensure that the transmitted checksum matches a calculated checksum based on the transmitted settings. The configuration mismatch event ensures that the transmitted settings match the configured settings. Supply Domain DVDD121 PVDD122 SVDD123 CVDD121 IOVDD VTT4 SIOVDD33 AVDD33 HARDWARE CONSIDERATIONS 1 Power Supply Recommendations The power supply domains are described in Table 59. The power supplies can be grouped into separate PCB domains as show in Figure 67. All the AD9135/AD9136 supply domains must remain as noise free as possible for the best operation. Power supply noise has a frequency component that affects performance, and is specified in terms of V rms. Figure 68 shows the recommended power supply components. Voltage (V) 1.2 1.2 1.2 1.2 1.8 1.2 3.3 3.3 Circuitry Digital core DAC PLL JESD204B receiver interface DAC clocking SPI interface VTT Sync LVDS transmit DAC This supply requires a 1.3 V supply when operating at maximum DAC sample rates. See Table 3 for details. 2 This supply can be combined with CVDD12 on the same regulator with a separate supply filter network and sufficient bypass capacitors near the pins. 3 This supply requires a 1.3 V supply when operating at maximum interface rates. See Table 4 for details. 4 This supply can be connected to SVDD12 and does not need separate circuitry. An LC filter on the output of the power supply is recommended to attenuate the noise, and must be placed as close to the AD9135/AD9136 as possible. An effective filter is shown in Figure 67. This filter scheme reduces high frequency noise components. Each of the power supply pins of the AD9135/ AD9136 must also have a 0.1 μF capacitor connected to the ground plane, as shown in Figure 67. Place the capacitor as close to the supply pin as possible. Adjacent power pins can share a bypass capacitor. Connect the ground pins of the AD9135/ AD9136 to the ground plane using vias. Rev. A | Page 61 of 117 AD9135/AD9136 Data Sheet 1.8V FPGA VCCIO OR OTHER SYSTEM SUPPLY 10µH IOVDD 66 10µF 1.2V LINEAR REGULATOR DVDD12 13 10µF 14 53 3.3V LINEAR REGULATOR SIOVDD33 35 10µF AVDD33 67 75 1.2V LINEAR REGULATOR SVDD12 17 10µF 77 20 85 22 28 1.2V LINEAR REGULATOR 10µH 31 CVDD12 71 32 10µF 76 33 81 36 87 39 1 45 10µH PVDD12 10µF 47 4 50 7 VTT 8 21 10µF 9 25 10 42 56 46 57 12578-039 NOTES 1. UNLABELED CAPACITORS ARE 0.1µF, CLOSE TO DEVICE PIN(S), WITH MINIMUM DISTANCE AND VIAS BETWEEN CAPACITORS AND PIN(S). Figure 67. JESD204B Interface PCB Power Domain Recommendation AD9135/AD9136 1.8V POWER INPUT +12V ADP1741 ADP2119 ADP1741 ADP1753 BUCK 1.2MHz/600kHz 800mA 3.8V ADM7154-3.3 1.2V 1.2V SVDD12 + VTT DVDD12 CVDD12 + PVDD12 3.3V AVDD33 3.3V IOVDD + SIOVDD33 ADP2370 ADM7160-3.3 Figure 68. Power Supply Connections Rev. A | Page 62 of 117 12578-568 +3.3V STEP-DOWN DDC 1.2MHz, 2A 1.2V Data Sheet AD9135/AD9136 JESD204B Serial Interface Inputs (SERDIN0± to SERDIN7±) Return Loss When considering the layout of the JESD204B serial interface transmission lines, there are many factors to consider to maintain optimal link performance. Among these factors are insertion loss, return loss, signal skew, and the topology of the differential traces. The JESD204B specification limits the amount of return loss allowed in a converter device and a logic device, but does not specify return loss for the channel. However, every effort must be made to maintain a continuous impedance on the transmission line between the transmitting logic device and the AD9135/AD9136. As mentioned in the Insertion Loss section, minimizing the use of vias, or eliminating them altogether, reduces one of the primary sources for impedance mismatches on a transmission line. Maintain a solid reference beneath (for microstrip) or above and below (for stripline) the differential traces to ensure continuity in the impedance of the transmission line. If the stripline technique is used, follow the guidelines listed in the Insertion Loss section to minimize impedance mismatches and stub effects. Insertion Loss The JESD204B specification limits the amount of insertion loss allowed in the transmission channel (see Figure 47). The AD9135/AD9136 equalization circuitry allows significantly more loss in the channel than is required by the JESD204B specification. It is still important that the designer of the PCB minimize the amount of insertion loss by adhering to the following guidelines: Keep the differential traces short by placing the AD9135/AD9136 as near to the transmitting logic device as possible and routing the trace as directly as possible between the devices. Route the differential pairs on a single plane using a solid ground plane as a reference. Use a PCB material with a low dielectric constant (<4) to minimize loss, if possible. When choosing between the stripline and microstrip techniques, keep in mind the following considerations: stripline has less loss (see Figure 48 and Figure 49) and emits less EMI, but requires the use of vias that can add complexity to the task of controlling the impedance; whereas microstrip is easier to implement if the component placement and density allow routing on the top layer and eases the task of controlling the impedance. If using the top layer of the PCB is problematic or the advantages of stripline are desirable, follow these recommendations: Minimize the number of vias. If possible, use blind vias to eliminate via stub effects and use micro vias to minimize via inductance. If using standard vias, use the maximum via length to minimize the stub size. For example, on an 8-layer board, use Layer 7 for the stripline pair (see Figure 69). For each via pair, place a pair of ground vias adjacent to them to minimize the impedance discontinuity (see Figure 69). LAYER 1 LAYER 2 ADD GROUND VIAS y STANDARD VIA LAYER 6 MINIMIZE STUB EFFECT There are many sources for signal skew, but the two sources to consider when laying out a PCB are interconnect skew within a single JESD204B link and skew between multiple JESD204B links. In each case, keeping the channel lengths matched to within 15 mm is adequate for operating the JESD204B link at speeds of up to 10.64 Gbps. Managing the interconnect skew within a single link is fairly straightforward. Managing multiple links across multiple devices is more complex. However, follow the 15 mm guideline for length matching. Topology Structure the differential SERDINx± pairs to achieve 50 Ω to ground for each half of the pair. Stripline vs. microstrip tradeoffs are described in the Insertion Loss section. In either case, it is important to keep these transmission lines separated from potential noise sources such as high speed digital signals and noisy supplies. If using stripline differential traces, route them using a coplanar method, with both traces on the same layer. Although this does not offer more noise immunity than the broadside routing method (traces routed on adjacent layers), it is easier to route and manufacture so that the impedance continuity is maintained. An illustration of broadside vs. coplanar differential routing techniques is shown in Figure 70. Tx DIFF A GND LAYER 7 Signal Skew y DIFF+ LAYER 5 LAYER 8 y DIFF– LAYER 3 LAYER 4 GND 12578-040 Another primary source for impedance mismatch is at either end of the transmission line, where care must be taken to match the impedance of the termination to that of the transmission line. The AD9135/AD9136 handle this internally with a calibrated termination scheme for the receiving end of the line. See the Interface Power-Up and Input Termination section for details on this circuit and the calibration routine. Figure 69. Minimizing Stub Effect and Adding Ground Vias for Differential Stripline Traces Tx DIFF B Tx ACTIVE BROADSIDE DIFFERENTIAL Tx LINES Tx DIFF A Tx DIFF B Tx ACTIVE COPLANAR DIFFERENTIAL Tx LINES Figure 70. Broadside vs. Coplanar Differential Stripline Routing Techniques Rev. A | Page 63 of 117 12578-041 AD9135/AD9136 Data Sheet SYNCOUTx±, SYSREF±, and CLK± Signals When considering the trace width vs. copper weight and thickness, the speed of the interface must be considered. At multigigabit speeds, the skin effect of the conducting material confines the current flow to the surface. Maximize the surface area of the conductor by making the trace width made wider to reduce the losses. Additionally, loosely couple differential traces to accommodate the wider trace widths. This helps reduce the crosstalk and minimize the impedance mismatch when the traces must separate to accommodate components, vias, connectors, or other routing obstacles. Tightly coupled vs. loosely coupled differential traces are shown in Figure 71. TIGHTLY COUPLED DIFFERENTIAL Tx LINES Tx DIFF A Tx DIFF B LOOSELY COUPLED DIFFERENTIAL Tx LINES It is important to keep similar trace lengths for the CLK± and SYSREF± signals from the clock source to each of the devices on either end of the JESD204B links (see Figure 72). If using a clock chip that can tightly control the phase of CLK± and SYSREF±, the trace length matching requirements are greatly reduced. Figure 71. Tightly Coupled vs. Loosely Coupled Differential Traces AC Coupling Capacitors The AD9135/AD9136 require that the JESD204B input signals be ac-coupled to the source. These capacitors must be 100 nF and placed as close as possible to the transmitting logic device. To minimize the impedance mismatch at the pads, select the package size of the capacitor so that the pad size on the PCB matches the trace width as closely as possible. LANE 0 LANE 1 Tx DEVICE Rx DEVICE LANE N – 1 LANE N SYSREF DEVICE CLOCK SYSREF CLOCK SOURCE (AD9516-1, ADCLK925) SYSREF TRACE LENGTH DEVICE CLOCK TRACE LENGTH DEVICE CLOCK SYSREF TRACE LENGTH DEVICE CLOCK TRACE LENGTH Figure 72. SYSREF Signal and Device Clock Trace Length Rev. A | Page 64 of 117 12578-043 Tx DIFF B Separate the SYNCOUTx± signal from other noisy signals, because noise on the SYNCOUTx± may be interpreted as a request for K characters. The SYNCOUTx± signal has two modes of operation available for use. Register 0x2A5[0] defaults to 0, which sets the SYNCOUTx± swing to normal swing mode. When this bit is set to 1, the SYNCOUTx± swing is configured for high swing mode. For more details, see Table 8. 12578-042 Tx DIFF A The SYNCOUTx± and SYSREF± signals on the AD9135/ AD9136 are low speed LVDS differential signals. Use controlled impedance traces routed with 100 Ω differential impedance and 50 Ω to ground when routing these signals. As with the SERDIN0± to SERDIN7± data pairs, it is important to keep these signals separated from potential noise sources such as high speed digital signals and noisy supplies. Data Sheet AD9135/AD9136 DIGITAL DATAPATH INV SINC DIGITAL GAIN, DC OFFSET, AND GROUP DELAY ADJUSTMENT INTERPOLATION FILTERS 12578-049 INTERPOLATION MODES 1×, 2×, 4×, 8× Figure 73. Block Diagram of Digital Datapath The block diagram in Figure 73 shows the functionality of the digital datapath (all blocks can be bypassed). The digital processing includes three half-band interpolation filters, an inverse sinc filter, and gain, offset, and group delay adjustment blocks. Note that the pipeline delay changes when digital datapath functions are enabled/disabled. If fixed DAC pipeline latency is desired, do not reconfigure these functions after initial configuration. DAC PAGING Digital datapath registers are paged to allow configuration of either DAC independently or both simultaneously. Table 60 shows how to use the DAC paging bits. The transmit path contains three half-band interpolation filters, which each provide a 2× increase in output data rate and a lowpass function. The filters can be cascaded to provide a 4× or 8× interpolation ratio. Table 61 shows how to select each available interpolation mode, their usable bandwidths, and their maximum data rates. Note that fDATA = fDAC/Interpolation Factor. Interpolation mode is paged as described in the DAC Paging section. Register 0x030[0] is high if an unsupported interpolation mode is selected. Table 61. Interpolation Modes and Usable Bandwidth Interpolation Mode 1× (bypass)1 2× 4× 8× INTERP_MODE, Reg. 0x112[2:0] 0x00 0x01 0x03 0x04 Usable Bandwidth 0.5 × fDATA 0.4 × fDATA 0.4 × fDATA 0.4 × fDATA Maximum fDATA (MHz) 21202 10602 700 350 1 Mode 8 and Mode 11 can only use 1× interpolation. 2×, 4×, and 8× interpolation are only available in Mode 9, Mode 10, Mode 12, and Mode 13. 2 The maximum speed for 1× and 2× interpolation is limited by the JESD204B interface. See Table 4 for the appropriate supply levels. Table 60. Paging Modes DACs Paged DAC0 DAC1 DAC0 and DAC1 Filter Performance DATA FORMAT BINARY_FORMAT (Register 0x110[7]), paged as described in the DAC Paging section) controls the expected input data format. By default it is 0, which means the input data must be in twos complement. It can also be set to 1, which means the input data is in offset binary. For the AD9136, 0x0000 is negative full scale and 0xFFFF is positive full scale. For the AD9135, 0x0000 is negative full scale and 0xFFE0 is positive full scale. Though the AD9135 is an 11-bit resolution DAC at the output, the input to the part must still be 16-bits wide for proper 8-bit/10-bit decoding. The AD9135 uses a 16-bit datapath that is truncated to 11 bits before going into the DAC core. Either 11-bit zeropadded data or full 16-bit data can be sent into the device, with the latter having slightly better spectral performance by minimizing quantization error along the datapath. The usable bandwidth (as shown in Table 61) is defined as the frequency band over which the filters have a pass-band ripple of less than ±0.001 dB and an image rejection of greater than 85 dB. 0 2× 4× 8× –20 –40 –60 –80 –100 0 0.2 0.4 0.6 0.8 FREQUENCY (×fDAC ) 1.0 12578-368 Several functions are paged by DAC, such as input data format, downstream protection, interpolation, inverse sinc, digital gain, dc offset, group delay, datapath PRBS, and LMFC sync. The interpolation filters interpolate between existing data in such a way that they minimize changes in the incoming data while suppressing the creation of interpolation images. This is shown for each filter in Figure 74. MAGNITUDE (dB) DAC_PAGE, Register 0x008[1:0] 1 2 3 (default) Figure 74. All Band Responses of Interpolation Filters Filter Performance Beyond Specified Bandwidth The interpolation filters are specified to 0.4 × fDATA (with pass band). The filters can be used slightly beyond this ratio at the expense of increased pass-band ripple and decreased interpolation image rejection. Rev. A | Page 65 of 117 AD9135/AD9136 Data Sheet 0 –0.1 70 –0.2 60 –0.3 50 –0.4 40 –0.5 30 20 40 IMAGE REJECTION PASS-BAND RIPPLE 41 42 43 44 45 MAXIMUM PASS-BAND RIPPLE (dB) 80 DIGITAL GAIN, DC OFFSET, AND GROUP DELAY –0.6 12578-369 MINIMUM INTERPOLATION IMAGE REJECTION (dB) 90 BANDWIDTH (% fDATA ) Digital gain and dc offset (as described in the Digital Gain section and DC Offset section) allow compensation of imbalances in the I and Q paths due to analog mismatches between DAC I/Q outputs, quadrature modulator I/Q baseband inputs, and DAC/modulator interface I/Q paths. These imbalances can cause the two following issues: Figure 75. Interpolation Filter Performance Beyond Specified Bandwidth Figure 75 shows the performance of the interpolation filters beyond 0.4 × fDATA. Note that the ripple increases much slower than the image rejection decreases. This means that if the application can tolerate degraded image rejection from the interpolation filters, more bandwidth can be used. INVERSE SINC The AD9135/AD9136 provide a digital inverse sinc filter to compensate the DAC roll-off over frequency. The filter is enabled by setting the INVSINC_ENABLE bit (Register 0x111[7]; paged as described in the DAC Paging section) and is enabled by default. The inverse sinc (sinc−1) filter is a seven-tap FIR filter. Figure 76 shows the frequency response of sin(x)/x roll-off, the inverse sinc filter, and the composite response. The composite response has less than ±0.05 dB pass-band ripple up to a frequency of 0.4 × fDAC. To provide the necessary peaking at the upper end of the pass band, the inverse sinc filter shown has an intrinsic insertion loss of about 3.8 dB; in many cases, this can be partially compensated as described in the Digital Gain section. Group delay allows adjustment of the delay through the DAC, which can be used to adjust digital predistortion (DPD) loop delay. Digital Gain Digital gain can be used to independently adjust the digital signal magnitude being fed into each DAC. This is useful to balance the gain between I and Q channels of a DAC or to cancel out the insertion loss of the inverse sinc filter. Digital gain must be enabled when using the blanking state machine (see the Downstream Protection section). If digital gain is disabled, TXENx must be tied high. Digital gain is enabled by setting the DIG_GAIN_ENABLE bit (Register 0x111[5], paged as described in the DAC Paging section). In addition to enabling the function the amount of digital gain (GainCode) desired must be programmed. By default, digital gain is enabled and GainCode is 0xAEA. 0 ≤ Gain ≤ 4095/2048 −∞ dB ≤ dBGain ≤ 6.018 dB Gain = GainCode × (1/2048) 1 dBGain = 20 × log10(Gain) SIN(x)/x ROLL-OFF SINC–1 FILTER RESPONSE COMPOSITE RESPONSE 0 An unwanted sideband signal to appear at the quadrature modulator output with significant energy. Tuning the quadrature gain adjust values can optimize image rejection in single sideband radios or can optimize the error vector magnitude (EVM) in zero IF (ZIF) architectures. The I/Q mismatch can cause LO leakage through a modulator, which can be tuned out using dc offset. GainCode = 2048 × Gain = 2048 × 10dBGain/20 MAGNITUDE (dB) where GainCode is a 12-bit unsigned binary number. –1 The I/Q digital gain is set as shown in Table 62 and paged as described in the DAC Paging section. –2 –3 –5 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 FREQUENCY (× fIN) (Hz) 0.40 0.45 0.50 12578-058 –4 The default GainCode value (0xAEA = 2.7 dB) is appropriate to counteract the insertion loss of the inverse sinc filter without causing digital clipping when using 2× interpolation. This value can be read from of Figure 76 at 0.25 × fDAC, because that is the Nyquist rate when using a 2× interpolation. The recommended GainCode values for 4× and 8× interpolation are 0xBB3 (3.3 dB) and 0xBF8 (3.5 dB), respectively. Figure 76. Responses of sin(x)/x Roll-Off, the Sinc−1 Filter, and the Composite of the Two-Input Signal and Protection Rev. A | Page 66 of 117 Data Sheet AD9135/AD9136 Table 63. DC Offset Registers Table 62. Digital Gain Registers Register 0x111[5] 0x13C[7:0] 0x13D[3:0] Bit Name DIG_GAIN_ENABLE DAC_DIG_GAIN[7:0] DAC_DIG_GAIN[11:8] Description Set to 1 to enable digital gain LSB gain code MSB gain code Register 0x135[0] 0x136[7:0] 0x137[7:0] 0x13A[4:0] Bit Name DC_OFFSET_ON LSB_OFFSET[7:0] LSB_OFFSET[15:8] SIXTEENTH_OFFSET Description Set to 1 to enable dc offset LSB dc offset code MSB dc offset code Sub-LSB dc offset code DC Offset The dc offset feature is used to individually offset the data into the I or Q DAC. This feature can be used to cancel LO leakage. The offset is programmed as a 16-bit twos complement number in LSBs, plus a 5-bit twos complement number in 16ths of an LSB, as shown in Table 63. DC offset is paged as described in the DAC Paging section. −215 ≤ LSBs Offset < 215 −16 ≤ Sixteenths Offset 15 where LSBs Offset is the value of Register 0x136 and Register 0x137. Sixteenths Offset is the value of Register 0x13A. Group Delay Group delay can be used to delay the I or Q channels. This can be useful, for example, for DPD loop delay adjustment. −4 ≤ DAC Clock Cycles ≤ 3.5 Group Delay = (DAC Clock Cycles × 2) + 8 where Group Delay is a 4-bit twos complement number. Write the GroupDelay to the GROUP_DLY register (Register 0x014[3:0]). This feature is paged as described in the DAC Paging section. Rev. A | Page 67 of 117 AD9135/AD9136 Data Sheet DOWNSTREAM PROTECTION FILTER AND MODULATION DATA FROM LMFC SYNC LOGIC TXENx DIGITAL GAIN BSM Tx ENSM DATA TO DACs BSM_PROTECT Tx_PROTECT 1 0 1 0 PROTECT_OUTx Tx_PROTECT_OUT SPI_PROTECT PROTECT_OUT_INVERT 1 SPI_PROTECT_OUT PROTECT OUTx GENERATION 12578-372 0 Figure 77. Downstream Protection Block Diagram The AD9135/AD9136 have several blocks designed to protect the power amplifier (PA) of the system, as well as other downstream blocks. The AD9135/AD9136 consist of a blanking state machine (BSM) and a transmit enable state machine (Tx ENSM). after the BSM settles (see the Blanking State Machine (BSM) section). If the masks are enabled, an additional delay is imposed; the output is not valid until the BSM settles and the DACs fully power on (nominally an additional ~35 μs). The Tx ENSM is a block that controls delay between TXENx and the Tx_PROTECT signal. The Tx_PROTECT signal is used as an input to the BSM and its inverse can optionally be routed externally. Optionally, the Tx ENSM can also power down its associated DAC. The Tx ENSM is configured as shown in Table 64 and is paged as described in the DAC Paging section. The BSM gently ramps data entering the DAC and flushes the datapath. The BSM is activated by the Tx_PROTECT signal or automatically by the LMFC sync logic during a rotation. For proper function, digital gain must be enabled; tie TXENx high if disabling digital gain. Finally, some simple logic takes the outputs from each of those blocks and uses them to generate a desired PROTECT_OUTx signal on an external pin. This signal can be used to enable/disable downstream components, such as a PA. Transmit Enable State Machine The Tx ENSM is a simple block that controls the delay between the TXENx signal and the TX_PROTECT signal. This signal is used as an input to the BSM and its inverse can be routed to an external pin (PROTECT_OUTx) to turn downstream components on or off as desired. Table 64. Tx ENSM Registers Addr. 0x11F Bit No. [7:6] Bit Name FALL_COUNTERS [5:4] RISE_COUNTERS 0x121 [7:0] RISE_COUNT_0 0x122 [7:0] RISE_COUNT_1 0x123 [7:0] FALL_COUNT_0 0x124 [7:0] FALL_COUNT_1 The TXENx signal can be used to power down the associated DAC. If DAC0_MASK (Register 0x012[0]) = 1, a falling edge of TXENx causes DAC0 to power down. If DAC1_MASK (Register 0x012[1]) = 1, a falling edge of TXENx causes DAC1 to power down. On a rising edge of TXENx, without DAC0_MASK and DAC1_MASK enabled, the output is valid Rev. A | Page 68 of 117 Description Number of fall counters to use (1 to 2). Number of rise counters to use (0 to 2). Delay TX_PROTECT rise from TXENx rising edge by 32 × RISE_COUNT_0 DAC clock cycles. Delay TX_PROTECT rise from TXENx rising edge by 32 × RISE_COUNT_1 DAC clock cycles. Delay TX_PROTECT rise from TXENx rising edge by 32 × FALL_COUNT_0 DAC clock cycles. Must be at least 0x12. Delay TX_PROTECT rise from TXENx rising edge by 32 × FALL_COUNT_1 DAC clock cycles. Data Sheet AD9135/AD9136 Blanking State Machine (BSM) Table 66. PROTECT_OUTx Registers The BSM gently ramps data entering the DAC and flushes the datapath. On a falling edge of TX_PROTECT (the TXENx signal delayed by the Tx ENSM), the datapath holds the latest data value and the digital gain gently ramps from its set value to 0. At the same time, the datapath is flushed with zeros. On a rising edge of TX_PROTECT, the TXENx signal is delayed by the Tx ENSM; data is allowed to flow through the datapath again and the digital gain gently ramps the data from 0 up to the set digital gain. Both of these functions are also triggered automatically by the LMFC sync logic during a rotation to prevent glitching on the output. Ramping For proper ramping, digital gain must be enabled; tie TXENx high if disabling digital gain. The step size to use when ramping gain to 0 or its assigned value can be controlled via the GAIN_RAMP_DOWN_STEP[11:0] registers (Register 0x142 and Register 0x143) and the GAIN_RAMP_UP_STEP[11:0] registers (Register 0x140 and Register 0x141). These registers are paged as described in the DAC Paging section. The current BSM state can be read back as shown in Table 65. Table 65. Blanking State Machine Ramping Readbacks Register 0x147[7:6] Value 0b00 0b01 0b10 0b11 Description Data is being held at midscale. Ramping gain to 0. Data ramping to midscale. Ramping gain to assigned value. Data ramping to normal amplitude. Data at normal amplitude. Blanking State Machine IRQ Blanking completion is available as an IRQ event. Use Register 0x021[5] to enable blanking completion for DAC0 and then use Register 0x025[5] to read back its status and reset the IRQ signal. Use Register 0x022[5] to enable blanking completion for DAC1 and then use Register 0x026[5] to read back its status and reset the IRQ signal. See the Interrupt Request Operation section for more information. PROTECT_OUTx Generation Register 0x013 controls which signals are OR’ed into the external PROTECT_OUTx signal. Register 0x11F[2] can be used to invert the PROTECT_OUTx signal. By default, PROTECT_OUTx is high when the output is valid. Register 0x013 and Register 0x11F are paged as described in the DAC Paging section. Reg. 0x013 0x11F Bit No. 5 Bit Name TX_PROTECT_OUT 3 SPI_PROTECT_OUT 2 2 SPI_PROTECT PROTECT_OUT_INVERT Description 1: Tx ENSM triggers PROTECT_OUT 1: SPI_PROTECT triggers PROTECT_OUT Sets SPI_PROTECT Inverts PROTECT_OUTx DATAPATH PRBS The datapath PRBS can be used to verify that the AD9135/ AD9136 datapath is receiving and correctly decoding data. The datapath PRBS verifies that the JESD204B parameters of the transmitter and receiver match, that the lanes of the receiver are mapped appropriately, that the lanes have been appropriately inverted, if necessary, and in general that the start-up routine has been implemented correctly. The datapath PRBS test is designed to support input data rates of up to 1060 MHz. The datapath PRBS is paged as described in the DAC Paging section. To run the datapath PRBS test, complete the following steps: 1. Set up the device in the desired operating mode. See the Device Setup Guide section for details on setting up the device. 2. Send the PRBS7 or PRBS15 data. 3. Write Register 0x14B[2] = 0 for PRBS7 or 1 for PRBS15. 4. Write Register 0x14B[1:0] = 0b11 to enable and reset the PRBS test. 5. Write Register 0x14B[1:0] = 0b01 to enable the PRBS test and release reset. 6. Wait 500 ms. 7. Check the status by checking the IRQ for DAC0 and DAC1 PRBS as described in the Datapath PRBS IRQ section. 8. If there are failures, set Register 0x008 = 0x01 to view the status of DAC0. Set Register 0x008 = 0x02 to view the status of DAC1. 9. Read Register 0x14B[6]. Bit 6 is 0 if the selected DAC has any errors. This must match the IRQ. 10. Read Register 0x14C to read the error count of the selected DAC. Note that the PRBS processes 32 bits at a time, and compares the 32 new bits to the previous set of 32 bits. It detects (and reports) only 1 error in every group of 32 bits; therefore, the error count partly depends on when the errors are seen. For example, Rev. A | Page 69 of 117 Bits: 32 good, 31 good, 1 bad; 32 good (2 errors) Bits: 32 good, 22 good, 10 bad; 32 good (2 errors) Bits: 32 good, 31 good, 1 bad; 31 good, 1 bad; 32 good (3 errors) AD9135/AD9136 Data Sheet Datapath PRBS IRQ DC TEST MODE The PRBS fail signals for each DAC are available as IRQ events. Use Register 0x020, Bit 2 and Bit 0, to enable the fail signals, and then use Register 0x024, Bit 2 and Bit 0, to read back their statuses and reset the IRQ signals. See the Interrupt Request Operation section for more information. As a convenience, the AD9135/AD9136 provide a dc test mode, which is enabled by setting Register 0x520[1] to 1 and clearing Register 0x146[0] to 0. When this mode is enabled, the datapath is given 0 (midscale) for its data. Register 0x146[0] must be set to 1 for all other modes of operation. In conjunction with dc offset, this test mode can provide the desired dc data to the DACs. Rev. A | Page 70 of 117 Data Sheet AD9135/AD9136 INTERRUPT REQUEST OPERATION The AD9135/AD9136 provide an interrupt request output signal on Pin 60 (IRQ) that can be used to notify an external host processor of significant device events. On assertion of the interrupt, query the device to determine the precise event that occurred. The IRQ pin is an open-drain, active low output. Pull the IRQ pin high external to the device. This pin can be tied to the interrupt pins of other devices with open-drain outputs to wire; OR these pins together. Table 67. IRQ Register Block Details Figure 78 shows a simplified block diagram of how the IRQ blocks works. If IRQ_EN is low, the INTERRUPT_SOURCE signal is set to 0. If IRQ_EN is high, any rising edge of an event causes the INTERRUPT_SOURCE signal to be set high. If any INTERRUPT_SOURCE signal is high, the IRQ pin is pulled low. INTERRUPT_SOURCE can be reset to 0 by either an IRQ_RESET signal or a DEVICE_RESET. Register Block 0x01F to 0x026 Event Reported Per chip 0x46D to 0x46F; 0x470 to 0x473; 0x47A 0x47B[4] Per link and lane Per link EVENT_STATUS INTERRUPT_SOURCE if IRQ is enabled, if not, it is EVENT INTERRUPT_SOURCE if IRQ is enabled, if not, 0 INTERRUPT_SOURCE if IRQ is enabled, if not, 0 INTERRUPT SERVICE ROUTINE Interrupt request management begins by selecting the set of event flags that require host intervention or monitoring. Enable the events that require host action so that the host is notified when they occur. For events requiring host intervention upon IRQ activation, run the following routine to clear an interrupt request: Depending on the STATUS_MODE signal, the EVENT_STATUS bit reads back event or INTERRUPT_SOURCE. The AD9135/ AD9136 have several IRQ register blocks, which can monitor up to 75 events (depending on device configuration). Certain details vary by IRQ register block as described in Table 67. Table 68 shows which registers the IRQ_EN, IRQ_RESET, and STATUS_MODE signals in Figure 78 originate from, as well as the address where EVENT_STATUS is read back. 1. 2. 3. 4. 5. 6. 7. Read the status of the event flag bits that are being monitored. Disable the interrupt by writing 0 to IRQ_EN. Read the event source. For Register 0x01F to Register 0x026, EVENT_STATUS has a live readback. For other events, see their registers. Perform any actions required to clear the cause of the EVENT. In many cases, no specific actions may be required. Verify that the event source is functioning as expected. Clear the interrupt by writing 1 to IRQ_RESET. Enable the interrupt by writing 1 to IRQ_EN. 0 1 EVENT_FLAG IRQ INTERRUPT_ SOURCE INTERRUPT_ENABLE OTHER INTERRUPT SOURCES EVENT_FLAG_SOURCE 12578-060 WRITE_1_TO_EVENT_FLAG DEVICE_RESET Figure 78. Simplified Schematic of IRQ Circuitry Table 68. IRQ Register Block Address of IRQ Signal Details Register Block 0x01F to 0x026 0x46D to 0x46F IRQ_EN 0x01F to 0x022; R/W per chip 0x47A; W per link 0x470 to 0x473 0x47B[4] 0x47A; W per link 0x47B[3]; R/W per link; 1 by default 1 Address of IRQ Signals1 IRQ_RESET STATUS_MODE 0x023 to 0x026; W per chip STATUS_MODE = IRQ_EN N/A, STATUS_MODE = 1 0x46D to 0x46F; W per link and lane 0x470 to 0x473; W per link N/A, STATUS_MODE = 1 0x47B[4]; W per link N/A, STATUS_MODE = 1 N/A means not applicable. Rev. A | Page 71 of 117 EVENT_STATUS 0x023 to 0x026; R per chip 0x47A; R per link 0x47A; R per link 0x47B[4]; R per link AD9135/AD9136 Data Sheet DAC INPUT CLOCK CONFIGURATIONS The AD9135/AD9136 DAC sample clock (DACCLK) can be sourced directly through CLK± (Pin 2 and Pin 3) or by clock multiplication through the CLK± differential input. Clock multiplication employs the on-chip PLL that accepts a reference clock operating at a submultiple of the desired DACCLK rate. The PLL then multiplies the reference clock up to the desired DACCLK frequency, which is used to generate all the internal clocks required by the DAC. The clock multiplier provides a high quality clock that meets the performance requirements of most applications. Using the on-chip clock multiplier removes the burden of generating and distributing the high speed DACCLK. The second mode bypasses the clock multiplier circuitry and allows DACCLK to be sourced directly to the DAC core. This mode allows the user to source a very high quality clock directly to the DAC core. DRIVING THE CLK± INPUTS The CLK± differential input circuitry is shown in Figure 79 as a simplified circuit diagram of the input. The on-chip clock receiver has a differential input impedance of 10 kΩ. It is self biased to a common-mode voltage of about 600 mV. The inputs can be driven by differential PECL or LVDS drivers with accoupling between the clock source and the receiver. CLK+ To optimize the PLL across all operating conditions, the register writes in Table 69 are recommended. These writes properly set up the DAC PLL, including the loop filter and the charge pump. Table 69. DAC PLL Fixed Register Writes Register Address 0x087 0x088 0x089 0x08A 0x08D 0x1B0 Register Value 0x62 0xC9 0x0E 0x12 0x7B 0x00 0x1B9 0x1BC 0x1BE 0x24 0x0D 0x02 0x1BF 0x8E 0x1C0 0x2A 0x1C1 0x1C4 0x2A 0x7E Description Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL charge pump settings Optimal DAC LDO settings for DAC PLL Power DAC PLL blocks when power machine disabled Optimal DAC PLL charge pump settings Optimal DAC PLL VCO control settings Optimal DAC PLL VCO power control settings Optimal DAC PLL VCO calibration settings Optimal DAC PLL lock counter length setting Optimal DAC PLL charge pump setting Optimal DAC PLL varactor settings CLOCK MULTIPLICATION 5kΩ The on-chip PLL clock multiplier circuit can be used to generate the DAC sample rate clock from a lower frequency reference clock. The PLL is integrated on-chip, including the VCO and the loop filter. The VCO operates over the frequency range of 6 GHz to 12 GHz. 600mV 12578-061 5kΩ CLK– DAC PLL FIXED REGISTER WRITES Figure 79. Clock Receiver Input Simplified Equivalent Circuit The minimum input drive level to the differential clock input is 400 mV p-p differential. The optimal performance is achieved when the clock input signal is between 800 mV p-p differential and 1000 mV p-p differential. Whether using the on-chip clock multiplier or sourcing the DACCLK directly (the CLK± pins are used in both cases), the input clock signal to the device must have low jitter and fast edge rates to optimize the DAC noise performance. Direct clocking with a low noise clock produces the lowest noise spectral density at the DAC outputs. The clocks and clock receiver are powered down by default. The clocks must be enabled by writing to Register 0x080. To enable all clocks on the device, write Register 0x080 = 0x00. Register 0x080, Bit 7 powers up the clocks for DAC0. Bit 6 powers up the clocks for DAC1. Bit 5 powers up the digital clocks; Bit 4 powers up the SERDES clocks; and Bit 3 powers up the clock receiver. The PLL configuration parameters must be programmed before the PLL is enabled. Step by step instructions on how to program the PLL can be found in the Starting the PLL section. The functional block diagram of the clock multiplier is shown in Figure 82. The clock multiplication circuit generates the DAC sampling clock from the REFCLK input, which is fed in on the CLK± differential pins (Pin 2 and Pin 3). The frequency of the REFCLK input is referred to as fREF. The REFCLK input is divided by the variable RefDivFactor. Select the RefDivFactor variable to ensure that the frequency into the phase frequency detector (PFD) block is between 35 MHz and 80 MHz. The valid values for RefDivFactor are 1, 2, 4, 8, 16, or 32. Each RefDivFactor value maps to the appropriate REF_DIV_MODE register control according to Table 70. The REF_DIV_MODE register is programmed through Register 0x08C[2:0]. Rev. A | Page 72 of 117 Data Sheet AD9135/AD9136 Table 70. Mapping of RefDivFactor to REF_DIV_MODE Divide by (RefDivFactor) 1 2 4 8 16 REF_DIV_MODE, Reg. 0x08C[2:0] 0 1 2 3 4 The range of fREF is 35 MHz to 1 GHz, and the output frequency of the PLL is 420 MHz to 2.8 GHz. Use the following equations to determine the RefDivFactor: 35 MHz f REF 80 MHz RefDivFactor (1) where: fREF is the reference frequency on the CLK± input pins. RefDivFactor is the reference divider division ratio. Table 72. Common Frequency Examples Frequency (MHz) 368.64 184.32 307.2 122.88 61.44 491.52 245.76 fDAC (MHz) 1474.56 1474.56 1228.88 983.04 983.04 1966.08 1966,08 f DACCLK f REF RefDivFactor 2 BCount (2) TO VCO R1 C2 C3 C1 TO VCO LDO Figure 80. Loop Filter The PFD compares fREF/RefDivRate to fDAC/(2 × BCount) and pulses the charge pump up or down to control the frequency of the VCO. A low noise VCO is tunable over an octave with an oscillation range of 6 GHz to 12 GHz. The clock multiplication circuit operates such that the VCO outputs a frequency, fVCO. (3) and from Equation 2, the DAC sample clock frequency, fDAC, is equal to Charge Pump The charge pump current is 6-bit programmable and varies from 0.1 mA to 6.4 mA in 0.1 mA steps. The charge pump current is programmed into Register 0x08A for the DAC PLL as shown in the DAC PLL Fixed Register Writes section. The charge pump calibration must be run one time during chip initialization to reduce reference spurs. This calibration is on by default. (4) UP TO LOOP FILTER DOWN Table 71. DAC VCO Divider Selection Divide by (LODivFactor) 4 8 16 LO_DIV_MODE, Register 0x08B[1:0] 1 2 3 Rev. A | Page 73 of 117 CHARGE PUMP CURRENT = 0.1mA TO 6.4mA Figure 81. Charge Pump 12578-063 The LODivFactor is chosen to keep fVCO in the operating range between 6 GHz and 12 GHz. The valid values for LODivFactor are 4, 8, and 16. Each LODivFactor maps to a LO_DIV_MODE value. The LO_DIV_MODE (Register 0x08B[1:0]) is programmed as described in Table 71. DAC Frequency Range (MHz) >1500 750 to 1500 420 to 750 BCount 16 16 16 8 8 16 16 R3 FROM CHARGE PUMP The BCount value is programmed with Bits[7:0] of Register 0x085. It is programmable from 6 to 127. f REF RefDivFactor LODivFactor 8 8 8 8 8 4 4 The RF PLL filter is fully integrated on-chip and is a standard passive third-order filter with five 4-bit programmable components (see Figure 80). The C1, C2, C3, R1, and R3 filter components are programmed with Register 0x087 through Register 0x089, as described in the DAC PLL Fixed Register Writes section. where: fDAC is the DAC sample clock. BCount is the feedback loop divider ratio. f DACCLK 2 BCount RefDivFactor 8 4 8 2 1 8 4 Loop Filter The BCount value is the divide ratio of the loop divider. It is set to divide the fDAC to frequency match the fREF/RefDivFactor. Select BCount so that the following equation is true: fVCO = fDAC × LODivFactor fVCO (MHz) 11796.48 11796.48 9831.04 7864.35 7864.35 7864.35 7864.35 12578-062 DAC Reference Frequency Range (MHz) 35 to 80 80 to 160 160 to 320 320 to 640 640 to 1000 Table 72 lists some common frequency examples for the RefDivFactor, LODivFactor, and BCount values that are needed to configure the PLL properly. AD9135/AD9136 Data Sheet Charge pump calibration is run during the first power-up of the PLL, and the coefficient of the calibration is held for all subsequent starts. The PLL is enabled by writing 0x10 into Register 0x083, but the configuration registers must be programmed before the PLL is enabled. The calibration tries to match the up and down current, which minimizes the spurs at the reference frequency that appears at the DAC output. The charge pump calibration takes 64 reference clock cycles. Bit 5 in Register 0x084 notifies the user that the charge pump calibration is completed and is valid. STARTING THE PLL The programming sequence for the DAC PLL is as follows: 1. 2. 3. 4. Temperature Tracking When properly configured, the device automatically selects one of the 512 VCO bands. The PLL settings selected by the device ensure that the PLL remains locked over the full −40°C to +85°C operating temperature range of the device without further adjustment. The PLL remains locked over the full temperature range, even if the temperature during initialization is at one of the temperature extremes. Confirm the PLL lock bit to ensure that the calibration completed properly. The PLL lock bit is Bit 1 of Register 0x084. 5. 6. 7. Register 0x084[5] notifies the user that the DAC PLL calibration is completed and is valid. To properly configure temperature tracking, follow the settings in the DAC PLL Fixed Register Writes section and the fVCO dependent SPI writes shown in Table 73. Register 0x084[1] notifies the user that the PLL has locked. Register 0x084[7] and Register 0x084[6] notify the user that the DAC PLL has reached the upper or lower edge of its operating band, respectively. If either of these bits are high, recalibrate the DAC PLL by setting Register 0x083[7] to 0 and then 1. Table 73. VCO Control Lookup Table Reference 4-BIT PROGRAMMABLE, INTEGRATED LOOP FILTER PFD 80MHz MAX ÷2 ÷4 ÷8 ÷16 The DAC PLL lock and lost signals are available as IRQ events. Use Register 0x01F[5:4] to enable these signals, and then use Register 0x023[5:4] to read back their statuses and reset the IRQ signals. See the Interrupt Request Operation section for more information. C1 RETIMER UP C2 C3 R1 DOWN VCO LDO LC VCO 6GHz TO 12GHz ÷2 R3 ÷2 ÷2 I Q ÷2 I Q I Q ALC CAL FO CAL CAL CONTROL BITS 0.1mA TO 6.4mA MUX/SELECTABLE BUFFERS LODivFactor = 4, 8, 16 ÷2 B COUNTER BCount (INTEGER FEEDBACK DIVIDER) RANGE = 6 TO 127 MAXIMUM FREQUENCY = 1.6GHz Figure 82. Device Clock PLL Block Diagram Rev. A | Page 74 of 117 DAC CLOCK 420MHz TO 2.8GHz 12578-064 35MHz TO 1GHz DAC PLL IRQ <750MHz CHARGE PUMP RefDivFactor = 1, 2, 4, 8, 16 fREF Register 0x1BB Setting 0x07 0x06 0x06 750MHz TO 1.5GHz Register 0x1B6 Setting 0x03 0x03 0x13 >1.5GHz Register 0x1B4 Setting 0x08 0x09 0x09 VCO Frequency Range (GHz) fVCO < 6.3 6.3 ≤ fVCO < 7.25 fVCO ≥ 7.25 Program the registers in the DAC PLL Fixed Register Writes section. Determine the VCO frequency based on the DAC frequency requirements. Determine the VCO divider ratio to achieve the desired DAC frequency. Program the VCO divider ratio in Register 0x08B[1:0]. Determine the BCount ratio to achieve the desired PLL reference frequency (35 MHz to 80 MHz). Program the BCount ratio in Register 0x085[7:0]. Determine the reference divider ratio to achieve the desired PLL reference frequency. Program the reference divider ratio in Register 0x08C[2:0]. Based on the fVCO found in Step 2, write the temperature tracking registers as shown in Table 73. Enable the DAC PLL synthesizer by setting Register 0x083[4] to 1. Data Sheet AD9135/AD9136 ANALOG OUTPUTS TRANSMIT DAC OPERATION 28 Figure 83 shows a simplified block diagram of the transmit path DACs. The DAC core consists of a current source array, a switch core, digital control logic, and full-scale output current control. The DAC full-scale output current (IOUTFS) is nominally 20.48 mA. The output currents from the OUTx± pins are complementary, meaning that the sum of the two currents always equals the fullscale current of the DAC. The digital input code to the DAC determines the effective differential current delivered to the load. 26 FSC (mA) 18 16 OUT1+ 12 0 OUT1– 256 384 512 640 768 896 1024 Figure 84. DAC Full-Scale Current (IOUTFS) vs. Gain DAC Code DAC0 DAC0 FULL-SCALE ADJUST OUT0+ Transmit DAC Transfer Function OUT0– The output currents from the OUTx+ and OUTx− pins are complementary, meaning that the sum of the positive and negative currents always equals the full-scale current of the DAC. The digital input code to the DAC determines the effective differential current delivered to the load. OUTx± provides the maximum output current when all bits are high for binary data. The output currents vs. DACCODE for the DAC outputs using binary format are expressed as 12578-371 4kΩ Figure 83. Simplified Block Diagram of DAC Core The DAC has a 1.2 V band gap reference. A 4 kΩ external resistor, RSET, must be connected from the I120 pin to the ground plane. This resistor, along with the reference control amplifier, sets up the correct internal bias currents for the DAC. Because the fullscale current is inversely proportional to this resistor, the tolerance of RSET is reflected in the full-scale output amplitude. DACFSC_x (where x is either 0 or 1, corresponding to DAC0 or DAC1) is a 10-bit twos complement value that controls the fullscale current of each of the four DAC outputs. These values are stored in Register 0x040 to Register 0x041 and Register 0x044 to Register 0x045 as shown in Table 74. The typical full-scale current for each DAC is given by I OUTP Table 74. DAC Full-Scale Current Registers Description DAC0 MSB gain code DAC0 LSB gain code DAC1 MSB gain code DAC1 LSB gain code (5) (6) where DACCODEBIN is the 11-/16-bit input to the DAC in unsigned binary. DACCODEBIN has a range of 0 to 2N − 1. If the data format is twos complement, the output currents are expressed as I OUTP For nominal values of VREF (1.2 V), RSET (4 kΩ), and DACFSC_x (0, which is midscale in twos complement), the full-scale current of the DAC is nominally 20.48 mA. The DAC full-scale current can be adjusted from 13.9 mA to 27.0 mA, by programming the appropriate DACFSC_x values in Register 0x040, Register 0x041, and Register 0x044, and Register 0x045. Analog output full-scale current vs. gain DAC code is shown in Figure 84. DACCODEBIN I OUTFS 2N – 1 IOUTN = IOUTFS − IOUTP IOUTFS = 20.45 + (DACFSC_x × 6.55 mA)/2(10 − 1) Value DACFSC_0[9:8] DACFSC_0[7:0] DACFSC_1[9:8] DACFSC_1[7:0] 128 GAIN DAC CODE CURRENT SCALING Address 0x040[1:0] 0x041[7:0] 0x044[1:0] 0x045[7:0] 20 14 DAC1 FULL-SCALE ADJUST DAC1 I120 22 12578-066 1.2V 24 DACCODETWOS 2N – 1 2N – 1 IOUTN = IOUTFS − IOUTP IOUTFS (7) (8) where DACCODETWOS is the 11-/16-bit input to the DAC in twos complement. DACCODETWOS has a range of −2N − 1 to 2N − 1 − 1. Powering Down Unused DACs Power down any unused DAC outputs to avoid burning excess power. The DAC power downs are located in Register 0x011. Register 0x011, Bit 6 corresponds to DAC0, and Bit 4 to DAC1. Write a 1 to each bit to power down the appropriate DACs. Register 0x011, Bit 7 and Bit 2, must stay low to enable the band gap and DAC master bias, respectively. Rev. A | Page 75 of 117 AD9135/AD9136 Data Sheet –40 Self Calibration 4 –80 –100 0 50 100 200 250 300 Figure 87. Pre-Calibration and Post-Calibration, HD2 and HD4 –60 CALIBRATION OFF CALIBRATION ON –65 –75 –80 –85 0 –90 –1 –100 –3 0 50 100 150 200 250 –4 0 10k 20k 30k 40k 50k 60k 70k DAC GAIN CODE 12578-088 fOUT (MHz) Figure 88. Pre-Calibration and Post-Calibration, IMD2 To calibrate, follow the routine in Table 75. Figure 85. Pre-Calibration and Post-Calibration, INL 4 Table 75. Device Self Calibration Procedure CALIBRATION OFF CALIBRATION ON Addr. 0x0E7 3 2 Bit [7:0] SPI Data Byte 0x38 3 2 1 0 [7:0] [7:0] [7:0] [7:0] 0 0b0 or 0b1 0 0b0 or 0b1 0xA2 0x01 0x03 0x30 0x0E8 1 0 –2 20k 30k 40k 50k 60k DAC GAIN CODE Figure 86. Pre-Calibration and Post-Calibration, DNL 70k 12578-089 –1 10k 300 12578-096 –95 –2 0 150 fOUT (MHz) 12578-095 –90 IMD2 (dBc) 1 INL (LSB) –70 –70 2 DNL (LSB) SECOND HARMONIC FOURTH HARMONIC –60 CALIBRATION OFF CALIBRATION ON 3 CALIBRATION OFF CALIBRATION ON –50 SFDR (dBc) The AD9135/AD9136 have a self calibration feature that improves the DAC dc and ac linearity in zero or low IF applications. The performance improvement includes the INL/DNL, second and fourth harmonic distortions (HD2 and HD4), and second-order intermodulation distortion (IMD2) of the device. Figure 85 and Figure 86 show the typical DAC INL and DNL before and after the calibration. Figure 87 and Figure 88 show the calibration effect on the HD2, HD4, and IMD2 performance. The improvement from calibration decreases with the DAC output frequency. For improvement in HD2 and HD4, it is recommended to run the calibration routine when the desired output frequency is below 100 MHz. For improvement in IMD2, it is recommended to run the routine when the desired output frequency is below 200 MHz. A single run of the routine is sufficient to obtain the desired performance for both ac and dc performance. 0x0ED 0x0E9 0x0E9 0x0E7 Description Use highest comparator speed and set calibration clock divider Select DACs to calibrate Set this bit to 0 1 if DAC1 is enabled Set this bit to 0 1 if DAC0 is enabled Configure initial value Enable calibration Start calibration Disable calibration clock For each DAC that is calibrated, verify the calibration status by writing a 1 in the corresponding bit of CAL_PAGE (Register 0x0E8) and reading Register 0x0E9. If the calibration completed correctly, CAL_FIN (Register 0x0E9[7]) = 1 to indicate that calibration is complete, and Register 0x0E9[6:4] = 0 to indicate that no errors have occurred. Rev. A | Page 76 of 117 Data Sheet AD9135/AD9136 –60 The post-calibration result is a function of operating temperature. A set of calibration coefficients obtained at one temperature may not be the optimal setting for a different temperature. Figure 89 and Figure 90 show the typical temperature drift effect after a single run calibration. –75 –80 –85 –90 –95 –100 SFDR (dBc) –70 –80 50 100 150 fOUT (MHz) 200 250 300 12578-378 –90 0 50 100 150 200 250 300 Figure 90. Post-Calibration IMD2 over Temperature, Calibrated at 25°C SECOND HARMONIC FOURTH HARMONIC –60 –100 0 fOUT (MHz) –40°C +25°C +85°C –50 –70 Figure 89. Post-Calibration HD2 and HD4 over Temperature, Calibrated at 25°C Rev. A | Page 77 of 117 12578-379 –40 –65 IMD2 (dBc) For optimal performance, run the calibration again when the operating temperature changes significantly. Note that it is recommended to power down the DAC outputs when running the calibration routine. If continuous transmission is required in the system, running the calibration again during the operation may not be an option. In this case, it is recommended to perform a calibration at the average temperature of the operating temperature range and to use the same set of coefficients during the operation. This results in the best overall performance over temperature. –40°C +25°C +85°C AD9135/AD9136 Data Sheet DEVICE POWER DISSIPATION The AD9135/AD9136 have eight supply rails, AVDD33, DVDD12, SVDD12, SIOVDD33, CVDD12, IOVDD, VTT, and PVDD12, which can be driven from five regulators to achieve optimum performance, as shown in Figure 67. The AVDD33 supply powers the DAC core circuitry. The power dissipation of the AVDD33 supply rail is independent of the digital operating mode and sample rate. The current drawn from the AVDD33 supply rail is typically 68 mA (225 mW) when the full-scale current of DAC0 and DAC1 are set to the nominal value of 20.48 mA. PVDD12 powers the DAC PLLs and varies depending on the DAC sample rate. CVDD12 can be combined with the PVDD12 regulator, but requires proper bypass capacitor networks near the pins. CVDD12 powers the clock tree, and the current varies directly with the DAC sample rate. DVDD12 powers the DSP core, and the current draw depends on the number of DSP functions and the DAC sample rate used. SVDD12 supplies the SERDES lanes and associated circuitry including the equalizers, SERDES PLL, PHY, and up to the input of the DSP. The current depends on the number lanes and the lane bit rate. IOVDD powers the SPI circuit and draws a very small current. TEMPERATURE SENSOR The AD9135/AD9136 have a band gap temperature sensor for monitoring the temperature changes of the AD9135/AD9136. The temperature must be calibrated against a known temperature to remove the device-to-device variation on the band gap circuit used to sense the temperature. To monitor temperature change, the user must take a reading at a known ambient temperature for a single-point calibration of each AD9135/AD9136 device. Tx = TREF + 7.3 × (CODE_X − CODE_REF)/1000 where: CODE_X is the readback code at the unknown temperature, Tx. CODE_REF is the readback code at the calibrated temperature, TREF. To use the temperature sensor, it must be enabled by setting Register 0x12F[0] to 1. The user must write a 1 to Register 0x134[0] before reading back the die temperature from Register 0x132 and Register 0x133. SIOVDD33 powers the equalizers for the SERDES lanes. The VTT termination voltage draws a very small current of <5 mA. Rev. A | Page 78 of 117 Data Sheet AD9135/AD9136 START-UP SEQUENCE Table 76 through Table 83 show the register writes needed to set up the AD9135/AD9136 with fDAC = 1474.56 MHz, 1× interpolation, and the DAC PLL enabled with a 368.64 MHz reference clock. The JESD204B interface is configured in Mode 11, single-link mode, Subclass 1, and scrambling is enabled with all eight SERDES lanes running at 7.3728 Gbps, inputting twos complement formatted data. No remapping of lanes with the crossbar is used in this example. The sequence of steps to properly start up the AD9135/AD9136 is as follows: 1. 2. 3. 4. 5. 6. Set up the SPI interface, power up necessary circuit blocks, make required writes to the configuration register, and set up the DAC clocks (see the Step 1: Start Up the DAC section). Set the digital features of the AD9135/AD9136 (see the Step 2: Digital Datapath section). Set up the JESD204B links (see the Step 3: Transport Layer section). Set up the physical layer of the SERDES interface (see the Step 4: Physical Layer section). Set up the data link layer of the SERDES interface. This procedure is for quick startup or debug only and does not guarantee deterministic latency (see the Step 5: Data Link Layer section). Check for errors on Link 0 and Link 1 (see the Step 6: Error Monitoring section). These steps are outlined in detail in the following sections in tables that list the required register write and read commands. Configure the DAC PLL Table 78. Configure DAC PLL Command W Address 0x087 Value 0x62 W 0x088 0xC9 W 0x089 0x0E W 0x08A 0x12 W 0x08D 0x7B W 0x1B0 0x00 W 0x1B9 0x24 W 0x1BC 0x0D W 0x1BE 0x02 W 0x1BF 0x8E W 0x1C0 0x2A W 0x1C1 0x2A W 0x1C4 0x7E W 0x08B 0x02 W 0x08C 0x03 W 0x085 0x10 W 0x1B5 0x09 W 0x1BB 0x13 W 0x1C5 0x06 W R 0x083 0x084 0x10 0x01 STEP 1: START UP THE DAC Power-Up and DAC Initialization Table 76. Power-Up and DAC Initialization Command W W W Address 0x000 0x000 0x011 Value 0xBD 0x3C 0x28 W W 0x080 0x081 0x00 0x00 Description Soft reset Deassert reset, set 4-wire SPI Enable reference, DAC channels, and master DAC Power up all clocks Power up SYSREF± receiver, disable hysteresis Required Device Configurations Table 77. Required Device Configurations Command W W W W W Address 0x12D 0x146 0x2A4 0x232 0x333 Value 0x8B 0x01 0xFF 0xFF 0x01 Description Digital datapath configuration Digital datapath configuration Clock configuration SERDES interface configuration SERDES interface configuration Description Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL loop filter settings Optimal DAC PLL charge pump settings Optimal DAC LDO settings for DAC PLL Power DAC PLL blocks when power machine disabled Optimal DAC PLL charge pump settings Optimal DAC PLL VCO control settings Optimal DAC PLL VCO power control settings Optimal DAC PLL VCO calibration settings Optimal DAC PLL lock counter length setting Optimal DAC PLL charge pump setting Optimal DAC PLL varactor settings Set the VCO LO divider to 8 so that 6 GHz ≤ fVCO = fDAC × 2(LODivMode + 1) ≤ 12 GHz Set the reference clock divider to 8 so that the reference clock into the PLL is less than 80 MHz Set the B counter to 16 to divide the DAC clock down to 2× the reference clock PLL lookup value from Table 25 for fVCO 7.25 GHz PLL lookup value from Table 25 for fVCO 7.25 GHz PLL lookup value from Table 25 for fVCO 7.25 GHz Enable DAC PLL Verify that Bit 1 reads back high for PLL locked STEP 2: DIGITAL DATAPATH Table 79. Digital Datapath Command W W Rev. A | Page 79 of 117 Address 0x112 0x110 Value 0x00 0x00 Description Set the interpolation to 1× Set twos complement data format AD9135/AD9136 Data Sheet STEP 3: TRANSPORT LAYER STEP 4: PHYSICAL LAYER Table 80. Link 0 Transport Layer Table 81. Physical Layer Command W W W Address 0x200 0x201 0x300 Value 0x00 0x00 0x08 Command W Address 0x2AA Value 0xB7 W 0x2AB 0x87 W 0x450 0x00 W 0x2B1 0xB7 W 0x451 0x00 W 0x2B2 0x87 W 0x452 0x00 W 0x453 0x83 W W W W 0x2A7 0x2AE 0x314 0x230 0x01 0x01 0x01 0x28 W W W 0x454 0x455 0x456 0x00 0x1F 0x00 W W W 0x206 0x206 0x289 0x00 0x01 0x04 W W 0x457 0x458 0x0F 0x2F W 0x459 0x21 W W W W 0x284 0x285 0x286 0x287 0x62 0xC9 0x0E 0x12 W W W W W 0x45A 0x45D 0x46C 0x476 0x47D 0x80 0x45 0xFF 0x01 0xFF W W 0x28A 0x28B 0x7B 0x00 W 0x290 0x89 W 0x294 0x24 W W W 0x296 0x297 0x299 0x03 0x0D 0x02 W 0x29A 0x8E W 0x29C 0x2A W 0x29F 0x78 W 0x2A0 0x06 W R 0x280 0x281 0x01 0x01 W 0x268 0x62 1 Description Power up the interface Enable all lanes Bit 3 = 0 for single link, Bit 2 = 0 to access Link 0 registers Set the device ID to match Tx (0x00 in this example) Set the bank ID to match Tx (0x00 in this example) Set the lane ID to match Tx (0x00 in this example) Set descrambling and L to 4 (in n − 1 notation) (L = 8 on transmit side)1 Set F = 1 (in n − 1 notation) Set K = 32 (in n − 1 notation) Set M to 1 (in n − 1 notation) (M = 2 on transmit side)1 Set N = 16 (in n − 1 notation) Set Subclass 1 and NP = 16 (in n − 1 notation) Set JESD204B Version and S = 2 (in n − 1 notation) Set HD = 1 Set checksum for Lane 0 Deskew Lane 0 to Lane 7 Set F (not in n − 1 notation) Enable Lane 0 to Lane 7 Note that for Mode 11 through Mode 13, the M and L the parameters programmed on the receive side do not match the parameters on the transmit side. The parameters on the transmit side reflect the true number of converters and lanes per link. Rev. A | Page 80 of 117 Description SERDES interface termination setting SERDES interface termination setting SERDES interface termination setting SERDES interface termination setting Autotune PHY setting Autotune PHY setting SERDES SPI configuration Configure CDRs in half rate mode Resets CDR logic Release CDR logic reset Configure PLL divider to 1 along with PLL required configuration Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL loop filter Optimal SERDES PLL charge pump Optimal SERDES PLL VCO LDO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO Optimal SERDES PLL VCO Optimal SERDES PLL configuration Optimal SERDES PLL VCO varactor Optimal SERDES PLL charge pump Optimal SERDES PLL VCO varactor Optimal SERDES PLL VCO varactor Enable SERDES PLL Verify that Bit 0 reads back high for SERDES PLL lock Set EQ mode to low power Data Sheet AD9135/AD9136 STEP 5: DATA LINK LAYER STEP 6: ERROR MONITORING Note that this procedure does not guarantee deterministic latency. Link 0 Checks Table 82. Data Link Layer—Does Not Guarantee Deterministic Latency Command W W W W Address 0x301 0x304 0x305 0x306 Value 0x01 0x00 0x00 0x0A W 0x307 0x0A W W W SYSREF± Signal W 0x03A 0x03A 0x03A 0x01 0x81 0xC1 0x300 0x01 Description Set subclass = 1 Set the LMFC delay setting to 0 Set the LMFC delay setting to 0 Set the LMFC receive buffer delay to 10 Set the LMFC receive buffer delay to 10 Set sync mode to one-shot sync Enable the sync machine Arm the sync machine Ensure that at least one SYSREF± edge is sent to the device Bit 0 = 1 to enable Link 0, Bit 2 = 0 to access Link 0 Confirm that the registers in Table 83 read back as noted and that system tasks are completed as described. Table 83. Link 0 Checks Command R Address 0x470 Value 0xFF SYNCOUT0± Signal SERDINx± Signals R 0x471 0xFF R R 0x472 0x473 0xFF 0xFF Rev. A | Page 81 of 117 Description Acknowledge that four consecutive K28.5 characters have been detected on Lane 0 to Lane 3. Confirm that SYNCOUT0± is high. Apply ILAS and data to SERDES input pins. Check for frame sync on all lanes. Check for good checksum. Check for ILAS. AD9135/AD9136 Data Sheet REGISTER MAPS AND DESCRIPTIONS In the following tables, register addresses (Reg. column) and reset (Reset column) values are hexadecimal and in the read/write (R/W) column, R means read only, W means write only, R/W means read/write, and N/A means not applicable. All values in the register address and reset columns are hexadecimal numbers. DEVICE CONFIGURATION REGISTER MAP Table 84. Device Configuration Register Map Reg. Name Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 Reset R/W 0x000 SPI_INTFCONFA SOFT RESET_M LSBFIRST_ M ADDRINC_M SDOACTIVE_M SDOACTIVE ADDRINC LSBFIRST SOFTRESET 0x00 R/W 0x003 CHIPTYPE CHIPTYPE 0x04 R 0x004 PRODIDL PRODIDL 0x44 R 0x005 PRODIDH 0x006 CHIPGRADE 0x008 SPI_PAGEINDX 0x00A SCRATCH_PAD 0x011 PWRCNTRL0 PRODIDH PROD_GRADE DEV_REVISION RESERVED DAC_PAGE SCRATCHPAD PD_BG PD_DAC_0 RESERVED PD_DAC_1 0x012 TXENMASK 0x013 PWRCNTRL3 0x014 GROUP_DLY 0x01F IRQEN_ STATUSMODE0 0x020 IRQEN_ STATUSMODE1 0x021 IRQEN_ STATUSMODE2 IRQEN_ SMODE_ PDPERR0 RESERVED IRQEN_ SMODE_ BLNKDONE0 RESERVED 0x022 IRQEN_ STATUSMODE3 IRQEN_ SMODE_ PDPERR1 RESERVED IRQEN_ SMODE_ BLNKDONE1 RESERVED 0x023 IRQ_STATUS0 CALPASS CALFAIL DACPLLLOST DACPLLLOCK RESERVED PD_DACM RESERVED RESERVED TX_PROTECT_ RESERVED OUT DAC1_MASK RESERVED IRQEN_ SMODE_ CALPASS IRQEN_ SMODE_ CALFAIL R R 0x03 R/W 0x00 R/W 0x7C R/W DAC0_MASK 0x00 R/W RESERVED SPI_PROTECT_OUT SPI_PROTECT 0x91 0x46/ 0x66 RESERVED GROUP_DLY 0x20 R/W 0x88 R/W IRQEN_SMODE_ SERPLLLOCK RESERVED IRQEN_ SMODE_ LANEFIFOERR 0x00 R/W IRQEN_SMODE_ PRBS1 RESERVED IRQEN_ SMODE_ PRBS0 0x00 R/W IRQEN_SMODE_ SYNC_LOCK0 IRQEN_SMODE_ SYNC_ROTATE0 IRQEN_ SMODE_ SYNC_ WLIM0 0x00 IRQEN_ SMODE_ SYNC_TRIP0 R/W IRQEN_SMODE_ SYNC_LOCK1 IRQEN_SMODE_ SYNC_ROTATE1 IRQEN_ SMODE_ SYNC_ WLIM1 0x00 IRQEN_ SMODE_ SYNC_TRIP1 R/W SERPLLLOST SERPLLLOCK LANEFIFOERR RESERVED 0x00 R IRQEN_ IRQEN_SMODE_ IRQEN_SMODE_ SMODE_ DACPLLLOCK SERPLLLOST DACPLLLOST RESERVED 0x024 IRQ_STATUS1 PRBS1 RESERVED PRBS0 0x00 R 0x025 IRQ_STATUS2 PDPERR0 RESERVED BLNKDONE0 RESERVED RESERVED SYNC_LOCK0 SYNC_ROTATE0 SYNC_ WLIM0 SYNC_TRIP0 0x00 R 0x026 IRQ_STATUS3 PDPERR1 RESERVED BLNKDONE1 RESERVED SYNC_LOCK1 SYNC_ROTATE1 SYNC_ WLIM1 SYNC_TRIP1 0x00 R 0x030 JESD_CHECKS ERR_JESDBAD ERR_KUNSUPP ERR_ SUBCLASS ERR_INTSUPP 0x00 R 0x034 SYNC_ ERRWINDOW 0x038 SYNC_LASTERR_L 0x039 SYNC_LASTERR_H LASTUNDER LASTOVER 0x03A SYNC_CONTROL SYNCENABLE SYNCARM 0x03B SYNC_STATUS SYNC_ BUSY 0x03C SYNC_CURRERR_L 0x03D SYNC_CURRERR_H CURRUNDER RESERVED ERR_DLYOVER ERR_WINLIMIT RESERVED ERRWINDOW RESERVED LASTERROR RESERVED SYNCCLRSTKY SYNCCLRLAST RESERVED SYNCMODE SYNC_LOCK SYNC_ ROTATE RESERVED SYNC_WLIM CURRERROR CURROVER RESERVED Rev. A | Page 82 of 117 SYNC_ TRIP 0x00 R/W 0x00 R 0x00 R 0x00 R/W 0x00 R 0x00 R 0x00 R Data Sheet Reg. Name 0x040 DACGAIN0_1 0x041 DACGAIN0_0 0x044 DACGAIN1_1 0x045 DACGAIN1_0 0x080 CLKCFG0 AD9135/AD9136 Bit 7 Bit 6 PD_CLK0 0x083 DACPLLCNTRL RECAL_ DACPLL 0x084 DACPLLSTATUS DACPLL_ OVERRANGE_H 0x085 DACINTEGERWORD0 0x087 DACLOOPFILT1 0x08A DACCPCNTRL 0x08B DACLOGENCNTRL 0x08C DACLDOCNTRL1 Bit 1 Bit 0 DACFSC_0[9:8] DACFSC_1[9:8] DACFSC_1[7:0] SYSREF_ACTRL0 DACLOOPFILT2 Bit 2 RESERVED SYSREF_ACTRL1 DACLOOPFILT3 Bit 3 DACFSC_0[7:0] 0x082 0x089 Bit 4 RESERVED 0x081 0x088 Bit 5 PD_CLK1 PD_CLK_DIG PD_SERDES_ PCLK RESERVED PD_CLK_REC PD_SYSREF RESERVED HYS_ON SYSREF_RISE HYS_CNTRL1 DACPLL_ OVERRANGE_L LF_R1_WORD LF_ LF_BYPASS_ BYPASS_R1 C2 LF_BYPASS_C1 RESERVED R 0x08 R/W LF_C1_WORD 0x88 R/W LF_C3_WORD 0x88 R/W LF_R3_WORD 0x08 R/W LO_DIV_MODE RESERVED DACLDOCNTRL2 CAL_CTRL_ GLOBAL 0x0E7 CAL_CLKDIV RESERVED 0x0E8 CAL_PAGE RESERVED 0x0E9 CAL_CTRL REF_DIV_MODE DAC_LDO RESERVED CAL_FIN CAL_ ACTIVE CAL_START_ AVG CAL_CLK_EN CAL_EN_ AVG RESERVED CAL_PAGE CAL_ERRHI CAL_ERRLO R/W 0x00 RESERVED CP_CURRENT 0x0E2 R/W 0xF8 R/W RESERVED 0x08D R/W 0x00 0x00 B_COUNT LF_ BYPASS_ R3 0x00 R/W DACPLL_ LOCK LF_C2_WORD R/W R/W RESERVED DACPLL_ CAL_VALID 0x00 0x10 RESERVED ENABLE_ DACPLL R/W R/W 0x00 HYS_CNTRL0 RESERVED Reset 0x00 RESERVED CAL_START CAL_EN CAL_INIT 0x20 R/W 0x02 R/W 0x01 R/W 0x2B R/W 0x00 R/W 0x30 R/W 0x0F R/W 0x00 R/W A6 R/W 00 R/W 0xA0 R/W 0x01 R/W 0x83 R/W 0x0ED CAL_INIT 0x110 DATA_FORMAT BINARY_ FORMAT 0x111 DATAPATH_CTRL INVSINC_ ENABLE 0x112 INTERP_MODE 0x11F TXEN_SM_0 0x121 TXEN_RISE_ COUNT_0 RISE_COUNT_0 0x0F R/W 0x122 TXEN_RISE_ COUNT_1 RISE_COUNT_1 0x00 R/W 0x123 TXEN_FALL_ COUNT_0 FALL_COUNT_0 0xFF R/W 0x124 TXEN_FALL_ COUNT_1 FALL_COUNT_1 0xFF R/W 0x12D DEVICE_CONFIG_ REG_0 DEVICE_CONFIG_0 0x46 R/W 0x12F DIE_TEMP_CTRL0 0x20 R/W 0x132 DIE_TEMP0 DIE_TEMP[7:0] 0x00 R 0x133 DIE_TEMP1 DIE_TEMP[15:8] 0x00 R 0x134 DIE_TEMP_ UPDATE 0x00 R/W RESERVED RESERVED RESERVED DIG_GAIN_ ENABLE RESERVED FALL_COUNTERS INTERP_MODE RISE_COUNTERS RESERVED RESERVED RESERVED Rev. A | Page 83 of 117 PROTECT_OUT_ INVERT RESERVED AUXADC_ ENABLE DIE_TEMP_ UPDATE AD9135/AD9136 Reg. Name 0x135 DC_OFFSET_CTRL 0x136 DAC_DC_ OFFSET_1PART0 0x137 DAC_DC_ OFFSET_1PART1 0x13A DAC_DC_ OFFSET_2PART 0x13C DAC_DIG_GAIN0 0x13D DAC_DIG_GAIN1 0x140 GAIN_RAMP_UP_ STEP0 0x141 GAIN_RAMP_ UP_STEP1 0x142 GAIN_RAMP_ DOWN_STEP0 0x143 GAIN_RAMP_ DOWN_STEP1 0x146 DEVICE_CONFIG_ REG_1 0x147 BSM_STAT 0x14B PRBS 0x14C PRBS_ERROR 0x1B0 DACPLLT0 0x1B5 DACPLLT5 0x1B9 DACPLLT9 0x1BB DACPLLTB 0x1BC DACPLLTC 0x1BE DACPLLTE Data Sheet Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Reset R/W DC_OFFSET_ 0x00 ON R/W LSB_OFFSET[7:0] 0x00 R/W LSB_OFFSET[15:8] 0x00 R/W 0x00 R/W RESERVED RESERVED Bit 0 SIXTEENTH_OFFSET DAC_DIG_GAIN[7:0] 0xEA R/W 0x0A R/W 0x04 R/W 0x00 R/W 0x09 R/W 0x00 R/W 0x00 R/W 0x00 R 0x10 R/W PRBS_COUNT 0x00 R DAC_PLL_PWR 0xFA R/W 0x83 R/W 0x34 R/W 0x0C R/W DAC_PLL_VCO_CTRL 0x00 R/W DAC_PLL_VCO_PWR 0x00 R/W RESERVED DAC_DIG_GAIN[11:8] GAIN_RAMP_UP_STEP[7:0] RESERVED GAIN_RAMP_UP_STEP[11:8] GAIN_RAMP_DOWN_STEP[7:0] RESERVED GAIN_RAMP_DOWN_STEP[11:8] DEVICE_CONFIG_1 SOFTBLANKRB RESERVED RESERVED PRBS_ GOOD RESERVED PRBS_MODE RESERVED PRBS_RESET PRBS_EN VCO_VAR DAC_PLL_CP1 RESERVED VCO_BIAS_TCF VCO_BIAS_REF 0x1BF DACPLLTF DAC_PLL_VCOCAL 0x8D R/W 0x1C0 DACPLLT10 DAC_PLL_LOCK_CNTR 0x2E R/W 0x1C1 DACPLLT11 DAC_PLL_CP2 0x24 R/W 0x1C4 DACPLLT17 DAC_PLL_VAR1 0x33 R/W 0x1C5 DACPLLT18 DAC_PLL_VAR2 0x08 R/W 0x200 MASTER_PD 0x01 R/W 0x201 PHY_PD 0x203 GENERIC_PD 0x206 CDR_RESET 0x230 CDR_OPERATING_ MODE_REG_0 0x232 DEVICE_CONFIG_ REG_3 0x268 EQ_BIAS_REG 0x280 SERDESPLL_ ENABLE_CNTRL 0x281 PLL_STATUS 0x284 LOOP_FILTER_1 LOOP_FILTER_1 0x77 R/W 0x285 LOOP_FILTER_2 LOOP_FILTER_2 0x87 R/W 0x286 LOOP_FILTER_3 LOOP_FILTER_3 0x08 R/W RESERVED SPI_PD_ MASTER SPI_PD_PHY RESERVED SPI_ SYNC1_PD RESERVED RESERVED ENHALFRATE RESERVED CDR_ OVERSAMP 0x00 R/W SPI_ SYNC2_PD 0x00 R/W SPI_CDR_ RESETN 0x01 R/W RESERVED 0x28 R/W 0x0 R/W 0x62 R/W 0x00 R/W DEVICE_CONFIG_3 EQ_POWER_MODE RESERVED RESERVED RESERVED RECAL_ SERDESPLL SERDES_PLL_ SERDES_PLL_ OVERRANGE_ OVERRANGE_L H SERDES_PLL_CAL_ VALID_RB Rev. A | Page 84 of 117 RESERVED RESERVED ENABLE_ SERDESPLL SERDES_PLL_ 0x00 LOCK_RB R Data Sheet Reg. Name 0x287 SERDES_PLL_CP1 0x289 REF_CLK_ DIVIDER_LDO 0x28A VCO_LDO 0x28B SERDES_PLL_PD1 0x290 0x294 AD9135/AD9136 Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 Reset R/W 0x3F R/W 0x00 R/W SERDES_PLL_VCO_LDO 0x2B R/W SERDES_PLL_PD1 0x7F R/W SERDESPLL_VAR1 SERDES_PLL_VAR1 0x83 R/W SERDES_PLL_CP2 SERDES_PLL_CP2 0xB0 R/W 0x296 SERDESPLL_VCO1 SERDES_PLL_VCO1 0x0C R/W 0x297 SERDESPLL_VCO2 SERDES_PLL_VCO2 0x00 R/W 0x299 SERDES_PLL_PD2 SERDES_PLL_PD2 0x00 R/W 0x29A SERDESPLL_VAR2 SERDES_PLL_VAR2 0xFE R/W SERDES_PLL_CP1 RESERVED SERDES_PLL_DIV_MODE DEVICE_ CONFIG_4 0x29C SERDES_PLL_CP3 SERDES_PLL_CP3 0x17 R/W 0x29F SERDESPLL_VAR3 SERDES_PLL_VAR3 0x33 R/W 0x2A0 SERDESPLL_VAR4 SERDES_PLL_VAR4 0x08 R/W 0x2A4 DEVICE_CONFIG_ REG_8 DEVICE_CONFIG_8 0x4B R/W 0x2A5 SYNCOUTB_ SWING RESERVED SYNCOUTB_ SWING_MD 0x00 R/W 0x2A7 TERM_BLK1_ CTRLREG0 RESERVED RCAL_ TERMBLK1 0x00 R/W 0x2AA DEVICE_CONFIG_ REG_9 DEVICE_CONFIG_9 0xC3 R/W 0x2AB DEVICE_CONFIG_ REG_10 DEVICE_CONFIG_10 0x93 R/W 0x2AE TERM_BLK2_ CTRLREG0 0x00 R/W 0x2B1 DEVICE_CONFIG_ REG_11 DEVICE_CONFIG_11 0xC3 R/W 0x2B2 DEVICE_CONFIG_ REG_12 DEVICE_CONFIG_12 0x93 R/W 0x300 GENERAL_JRX_ CTRL_0 0x00 R/W 0x301 GENERAL_JRX_ CTRL_1 0x01 R/W 0x302 DYN_LINK_ LATENCY_0 RESERVED DYN_LINK_LATENCY_0 0x00 R 0x303 DYN_LINK_ LATENCY_1 RESERVED DYN_LINK_LATENCY_1 0x00 R 0x304 LMFC_DELAY_0 RESERVED LMFC_DELAY_0 0x00 R/W 0x305 LMFC_DELAY_1 RESERVED LMFC_DELAY_1 0x00 R/W 0x306 LMFC_VAR_0 RESERVED LMFC_VAR_0 0x06 R/W 0x307 LMFC_VAR_1 RESERVED LMFC_VAR_1 0x06 R/W 0x308 XBAR_LN_0_1 RESERVED LOGICAL_LANE0_SRC 0x08 R/W 0x309 XBAR_LN_2_3 RESERVED LOGICAL_LANE3_SRC LOGICAL_LANE2_SRC 0x1A R/W 0x30A XBAR_LN_4_5 RESERVED LOGICAL_LANE5_SRC LOGICAL_LANE4_SRC 0x2C R/W 0x30B XBAR_LN_6_7 RESERVED LOGICAL_LANE7_SRC LOGICAL_LANE6_SRC 0x3E R/W 0x30C FIFO_STATUS_ REG_0 LANE_FIFO_FULL 0x00 R 0x30D FIFO_STATUS_ REG_1 LANE_FIFO_EMPTY 0x00 R RESERVED RESERVED CHECKSUM _MODE RESERVED RCAL_ TERMBLK2 LINK_MODE LINK_PAGE RESERVED RESERVED LINK_EN SUBCLASSV_LOCAL LOGICAL_LANE1_SRC 0x312 SYNCB_GEN_1 0x00 R/W 0x314 SERDES_SPI_REG SYNCB_ERR_DUR SERDES_SPI_CONFIG RESERVED 0x00 R/W 0x315 PHY_PRBS_TEST_ EN PHY_TEST_EN 0x00 R/W Rev. A | Page 85 of 117 AD9135/AD9136 Data Sheet Reg. Name 0x316 PHY_PRBS_TEST_ RESERVED CTRL Bit 7 Bit 6 0x317 PHY_PRBS_TEST_ THRESHOLD_ LOBITS 0x318 Bit 5 Bit 4 Bit 3 Bit 1 Bit 0 Reset R/W PHY_TEST_ START PHY_TEST_ RESET 0x00 R/W PHY_PRBS_THRESHOLD[7:0] 0x00 R/W PHY_PRBS_TEST_ THRESHOLD_ MIDBITS PHY_PRBS_THRESHOLD[15:8] 0x00 R/W 0x319 PHY_PRBS_TEST_ THRESHOLD_ HIBITS PHY_PRBS_THRESHOLD[23:16] 0x00 R/W 0x31A PHY_PRBS_TEST_ ERRCNT_LOBITS PHY_PRBS_ERR_CNT[7:0] 0x00 R 0x31B PHY_PRBS_TEST_ ERRCNT_MIDBITS PHY_PRBS_ERR_CNT[15:8] 0x00 R 0x31C PHY_PRBS_TEST_ ERRCNT_HIBITS PHY_PRBS_ERR_CNT[23:16] 0x00 R 0x31D PHY_PRBS_TEST_ STATUS PHY_PRBS_PASS 0xFF R 0x32C SHORT_TPL_ TEST_0 0x32D SHORT_TPL_ TEST_1 0x32E PHY_SRC_ERR_CNT RESERVED Bit 2 PHY_PRBS_PAT_SEL SHORT_TPL_SP_SEL SHORT_TPL_DAC_SEL SHORT_TPL_ 0x00 TEST_EN R/W SHORT_TPL_REF_SP_LSB 0x00 R/W SHORT_TPL_ TEST_2 SHORT_TPL_REF_SP_MSB 0x00 R/W 0x32F SHORT_TPL_ TEST_3 RESERVED 0x00 R 0x333 DEVICE_CONFIG_ REG_13 DEVICE_CONFIG_13 0x00 R/W 0x334 JESD_BIT_ INVERSE_CTRL JESD_BIT_INVERSE 0x00 R/W 0x400 DID_REG DID_RD 0x00 R 0x401 BID_REG 0x00 R 0x402 LID0_REG RESERVED ADJDIR_RD PHADJ_RD LID0_RD 0x00 R 0x403 SCR_L_REG SCR_RD L-1_RD 0x00 R 0x404 F_REG 0x00 R 0x405 K_REG 0x406 M_REG SHORT_ TPL_FAIL ADJCNT_RD BID_RD RESERVED F-1_RD RESERVED K-1_RD M-1_RD R 0x00 R CS_N_REG 0x408 NP_REG 0x409 S_REG 0x40A HD_CF_REG 0x40B RES1_REG 0x40C RES2_REG RES2_RD 0x00 R 0x40D CHECKSUM_REG FCHK0_RD 0x00 R 0x40E COMPSUM0_REG FCMP0_RD 0x00 R 0x412 LID1_REG 0x415 CHECKSUM1_REG COMPSUM1_REG 0x41A LID2_REG 0x41D CHECKSUM2_REG 0x41E COMPSUM2_REG 0x422 LID3_REG 0x425 CHECKSUM3_REG HD_RD RESERVED 0x00 0x407 0x416 CS_RD SHORT_TPL_ TEST_RESET N-1_RD 0x00 R SUBCLASSV_RD NP-1_RD 0x00 R JESDV_RD S-1_RD 0x00 R CF_RD 0x00 R 0x00 R RESERVED RES1_RD RESERVED LID1_RD FCHK1_RD FCMP1_RD RESERVED LID2_RD FCHK2_RD FCMP2_RD RESERVED LID3_RD FCHK3_RD Rev. A | Page 86 of 117 0x00 R 0x00 R 0x00 R 0x00 R 0x00 R 0x00 R 0x00 R 0x00 R Data Sheet Reg. Name 0x426 COMPSUM3_REG 0x42A LID4_REG AD9135/AD9136 Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 FCMP3_RD RESERVED LID4_RD Reset R/W 0x00 R 0x00 R 0x42D CHECKSUM4_REG FCHK4_RD 0x00 R 0x42E COMPSUM4_REG FCMP4_RD 0x00 R 0x432 LID5_REG 0x00 R 0x435 CHECKSUM5_REG FCHK5_RD 0x00 R 0x436 COMPSUM5_REG FCMP5_RD 0x00 R 0x43A LID6_REG 0x00 R 0x43D CHECKSUM6_REG FCHK6_RD 0x00 R 0x43E COMPSUM6_REG FCMP6_RD 0x00 R 0x442 LID7_REG 0x00 R 0x445 CHECKSUM7_REG FCHK7_RD 0x00 R 0x446 COMPSUM7_REG FCMP7_RD 0x00 R 0x450 ILS_DID 0x451 ILS_BID 0x452 ILS_LID0 RESERVED ADJDIR 0x453 ILS_SCR_L SCR 0x454 ILS_F 0x455 ILS_K 0x456 ILS_M 0x457 ILS_CS_N 0x458 ILS_NP 0x459 ILS_S 0x45A ILS_HD_CF RESERVED LID5_RD RESERVED LID6_RD RESERVED LID7_RD DID ADJCNT BID PHADJ RESERVED 0x00 R/W L-1 0x83 R/W 0x00 R/W 0x1F R/W 0x01 R/W 0x0F R/W K-1 M-1 CS HD RESERVED R/W R/W LID0 F-1 RESERVED 0x00 0x00 N-1 SUBCLASSV NP-1 0x2F R/W JESDV S-1 0x20 R/W CF 0x80 R/W RESERVED 0x45B ILS_RES1 RES1 0x00 R/W 0x45C ILS_RES2 RES2 0x00 R/W FCHK0 0x45 R/W READERRORCNTR 0x00 R 0x00 R/W 0x45D ILS_CHECKSUM 0x46B ERRCNTRMON_RB 0x46B ERRCNTRMON 0x46C LANEDESKEW 0x46D BADDISPARITY_RB 0x46D BADDISPARITY 0x46E NIT_RB 0x46E NIT_W 0x46F UNEXPECTEDCONTROL_RB 0x46F UNEXPECTEDCONTROL_W 0x470 CODEGRPSYNCFLG RESERVED LANESEL RESERVED CNTRSEL LANEDESKEW 0x0F R/W BADDIS 0x00 R 0x00 R/W 0x00 R 0x00 R/W 0x00 R 0x00 R/W CODEGRPSYNC 0x00 R/W RST_IRQ_ DISABLE_ RST_ERR_ DIS ERR_CNTR_ CNTR_DIS DIS RESERVED RST_IRQ_ DISABLE_ RST_ERR_ NIT ERR_CNTR_ CNTR_NIT NIT RESERVED LANE_ADDR_DIS NIT LANE_ADDR_NIT UCC RST_IRQ_ DISABLE_ RST_ERR_ UCC ERR_CNTR_ CNTR_UCC UCC RESERVED LANE_ADDR_UCC 0x471 FRAMESYNCFLG FRAMESYNC 0x00 R/W 0x472 GOODCHKSUMFLG GOODCHECKSUM 0x00 R/W 0x473 INITLANESYNCFLG INITIALLANESYNC 0x00 R/W 0x476 CTRLREG1 0x477 CTRLREG2 0x478 KVAL F ILAS_ MODE RESERVED THRESHOLD_ MASK_EN KSYNC Rev. A | Page 87 of 117 RESERVED 0x01 R/W 0x00 R/W 0x01 R/W AD9135/AD9136 Bit 7 Data Sheet Reg. Name Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Reset R/W 0x47A IRQVECTOR_MASK BADDIS_ MASK NIT_MASK UCC_ MASK RESERVED INITIALLANESYNC_ MASK BADCHECKSUM FRAMESYNC_ _MASK MASK 0x00 CODEGRPSYNC_MASK R/W 0x47A IRQVECTOR_FLAG BADDIS_ FLAG NIT_FLAG UCC_FLAG RESERVED INITIALLANESYNC_ FLAG BADCHECKSUM FRAMESYNC_ _FLAG FLAG CODEGRPSYNC_FLAG 0x47B SYNCASSERTIONMASK UCC_S CMM CMM_ENABLE 0x47C ERRORTHRES 0x47D LANEENABLE 0x47E RAMP_ENA 0x520 DIG_TEST0 0x521 DC_TEST_VALUE0 0x522 DC_TEST_VALUE1 BADDIS_S NIT_S Bit 1 Bit 0 RESERVED ETH LANE_ENA 0x00 R 0x008 R/W 0xFF R/W 0x0F R/W ENA_RAMP_ 0x00 CHECK R/W RESERVED 0x1C R/W DC_TEST_VALUE[7:0] 0x00 R/W DC_TEST_VALUE[15:8] 0x00 R/W RESERVED RESERVED DC_TEST_ MODE DEVICE CONFIGURATION REGISTER DESCRIPTIONS Table 85. Device Configuration Register Descriptions Address 0x000 Name SPI_INTFCONFA Bit No. 7 6 5 4 3 2 Bit Name SOFTRESET_M LSBFIRST_M ADDRINC_M SDOACTIVE_M SDOACTIVE ADDRINC Settings 1 0 1 LSBFIRST 1 0 0 SOFTRESET 1 0x003 CHIPTYPE [7:0] CHIPTYPE 0x004 0x005 0x006 PRODIDL PRODIDH CHIPGRADE [7:0] [7:0] [7:4] PRODIDL PRODIDH PROD_GRADE 0x008 SPI_PAGEINDX [3:0] [7:2] [1:0] DEV_REVISION RESERVED DAC_PAGE 0b01 0b10 0b11 Description Soft Reset (Mirror). LSB First (Mirror). Address Increment (Mirror). SDO Active (Mirror). SDO Active. Address Increment. Controls whether addresses are incremented or decremented during multibyte data transfers. Addresses are incremented during multibyte data transfers Addresses are decremented during multibyte data transfers LSB First. Controls whether input and output data are oriented as LSB first or MSB first. Shift LSB in first Shift MSB in first Soft Reset. Setting this bit initiates a reset. This bit is autoclearing after the soft reset is complete. Assert soft reset The product type is “High Speed DAC”, which is represented by a code of 0x04. Product Identification Low. Product Identification High. Product Grade. AD9136 AD9135 Device Revision. Reserved. DAC Paging. Selects which DAC is accessed and written to when changing digital features, such as digital gain, dc offset. This paging affects Register 0x013 to Register 0x014, Register 0x034 to Register 0x03D, Register 0x110 to Register 0x124, and Register 0x135 to Register 0x14C. Read and write DAC0 Read and write DAC1 Write both DACs; read DAC0 Rev. A | Page 88 of 117 Reset 0x0 0x0 0x0 0x0 0x0 0x0 Access R R R R R/W R/W 0x0 R/W 0x0 R/W 0x4 R 0x44 0x91 R R R R R R R R/W 0x6 0x4 0x6 0x0 0x3 Data Sheet AD9135/AD9136 Address 0x00A Name SCRATCH_PAD Bit No. [7:0] Bit Name SCRATCHPAD Settings 0x011 PWRCNTRL0 7 PD_BG 6 PD_DAC_0 5 4 RESERVED PD_DAC_1 3 2 RESERVED PD_DACM [1:0] [7:2] 1 RESERVED RESERVED DAC1_MASK 0 DAC0_MASK [7:6] 5 4 3 RESERVED TX_PROTECT_OUT RESERVED SPI_PROTECT_ OUT SPI_PROTECT RESERVED RESERVED GROUP_DLY 1 1 1 1 0x012 TXENMASK 1 1 0x013 PWRCNTRL3 0x014 GROUP_DLY 0x01F IRQEN_ STATUSMODE0 2 [1:0] [7:4] [3:0] 7 IRQEN_SMODE_ CALPASS 1 1 1 0 6 IRQEN_SMODE_ CALFAIL 1 0 5 IRQEN_SMODE_ DACPLLLOST 1 0 Description This register does not affect any functions in the part and can be used for testing SPI communication with the part. Any value written to this register will be read back to reflect the change unless a reset or powercycle occurs. Reference Power-Down. Powers down the band gap reference for the entire chip. Circuits will not be provided with bias currents. Power down reference Powers Down DAC0. Powers down the I-channel DAC. Powers down DAC0 Reserved. Powers Down DAC1. Powers down the Q-channel DAC. Powers down DAC1 Reserved. Powers Down the DAC Master Bias. The master bias cell provides currents and DAC full-scale adjustments to the four DACs. With the DAC master bias powered down, the DACs are inoperative. Powers down the DAC master bias Reserved. Reserved. DAC1 TXEN1 Mask. Power down DAC1 on a falling edge of TXEN1. If TXEN1 is low, power down DAC1 DAC0 TXEN0 Mask. Power down DAC0 on a falling edge of TXEN0. If TXEN0 is low, power down DAC0 Reserved. TX_PROTECT triggers PROTECT_OUTx. Reserved. SPI_PROTECT triggers PROTECT_OUTx. Reset 0x00 Access R/W 0x0 R/W 0x1 R/W 0x0 0x1 R R/W 0x0 0x1 R R/W 0x0 0x0 0x0 R R R/W 0x0 R/W 0x0 0x1 0x0 0x0 R R/W R R/W SPI_PROTECT Reserved. Reserved. Group Delay Control. Delays the selected DAC channel output per the paging register. 0 = minimum delay. 15 = maximum delay. The range of the delay is −4 to +3.5 DAC clock periods, and the resolution is 1/2 DAC clock period. Calibration Pass Detection Status Mode. If CALPASS goes high, it latches and pulls IRQ low CALPASS shows current status Calibration Fail Detection Status Mode. If CALFAIL goes high, it latches and pulls IRQ low CALFAIL shows current status DAC PLL Lost Detection Status Mode. If DACPLLLOST goes high, it latches and pulls IRQ low DACPLLLOST shows current status 0x0 0x0 0x8 0x8 R/W R R R/W 0x0 R/W 0x0 R/W 0x0 R/W Rev. A | Page 89 of 117 AD9135/AD9136 Address Name Data Sheet Bit No. 4 Bit Name IRQEN_SMODE_ DACPLLLOCK Settings 1 0 3 IRQEN_SMODE_ SERPLLLOST 1 0 2 IRQEN_SMODE_ SERPLLLOCK 1 0 1 IRQEN_SMODE_ LANEFIFOERR 1 0 0x020 IRQEN_ STATUSMODE1 0 [7:3] RESERVED RESERVED 2 IRQEN_SMODE_ PRBS1 1 0 1 0 RESERVED IRQEN_SMODE_ PRBS0 0 IRQEN_ STATUSMODE2 7 IRQEN_SMODE_ PDPERR0 1 0 6 5 RESERVED IRQEN_SMODE_ BLNKDONE0 1 0 4 3 RESERVED IRQEN_SMODE_ SYNC_LOCK0 1 0 2 IRQEN_SMODE_ SYNC_ROTATE0 1 0 1 IRQEN_SMODE_ SYNC_WLIM0 1 0 0 IRQEN_SMODE_ SYNC_TRIP0 DAC1 PRBS Error Status Mode. 1 0 Reset 0x0 Access R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 0x0 R R 0x0 R/W 0x0 0x0 R/W R/W 0x0 R/W 0x0 0x0 R R/W 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 R/W If PRBS1 goes high, it latches and pulls IRQ low PRBS1 shows current status DAC0 PRBS Error Status Mode. 1 0x021 Description DAC PLL Lock Detection Status Mode. If DACPLLLOCK goes high, it latches and pulls IRQ low DACPLLLOCK shows current status SERDES PLL Lost Detection Status Mode. If SERPLLLOST goes high, it latches and pulls IRQ low SERPLLLOST shows current status SERDES PLL Lock Detection Status Mode. If SERPLLLOCK goes high, it latches and pulls IRQ low SERPLLLOCK shows current status Lane FIFO Error Detection Status Mode. If LANEFIFOERR goes high, latches and pulls IRQ low LANEFIFOERR shows current status Reserved. Reserved. If PRBS0 goes high, it latches and pulls IRQ low PRBS0 shows current status DAC0 PDP Error. If PDPERR0 goes high, it latches and pulls IRQ low PDPERR0 shows current status Reserved. DAC0 Blanking Done Status Mode. If BLNKDONE0 goes high, it latches and pulls IRQ low BLNKDONE0 shows current status Reserved. DAC0 Alignment Locked Status Mode. If SYNC_LOCK0 goes high, it latches and pulls IRQ low SYNC_LOCK0 shows current status DAC0 Alignment Rotate Status Mode. If SYNC_ROTATE0 goes high, it latches and pulls IRQ low SYNC_ROTATE0 shows current status DAC0 Outside Window Status Mode. If SYNC_WLIM0 goes high, it latches and pulls IRQ low SYNC_WLIM0 shows current status DAC0 Alignment Tripped Status Mode. If SYNC_TRIP0 goes high, it latches and pulls IRQ low SYNC_TRIP0 shows current status Rev. A | Page 90 of 117 Data Sheet Address 0x022 Name IRQEN_ STATUSMODE3 AD9135/AD9136 Bit No. 7 Bit Name IRQEN_SMODE_ PDPERR1 Settings 1 0 6 5 RESERVED IRQEN_SMODE_ BLNKDONE1 1 0 4 3 RESERVED IRQEN_SMODE_ SYNC_LOCK1 1 0 2 IRQEN_SMODE_ SYNC_ROTATE1 1 0 1 IRQEN_SMODE_ SYNC_WLIM1 1 0 0 IRQEN_SMODE_ SYNC_TRIP1 1 0 0x023 IRQ_STATUS0 7 CALPASS 6 CALFAIL 5 DACPLLLOST 4 DACPLLLOCK 3 SERPLLLOST 1 1 1 1 1 Description DAC1 PDP Error. If PDPERR1 goes high, it latches and pulls IRQ low PDPERR1 shows current status Reserved. DAC1 Blanking Done Status Mode. If BLNKDONE1 goes high, it latches and pulls IRQ low BLNKDONE1 shows current status Reserved. DAC1 Alignment Locked Status Mode. If SYNC_LOCK1 goes high, it latches and pulls IRQ low SYNC_LOCK1 shows current status DAC1 Alignment Rotate Status Mode. If SYNC_ROTATE1 goes high, it latches and pulls IRQ low SYNC_ROTATE1 shows current status DAC1 Outside Window Status Mode. If SYNC_WLIM1 goes high, it latches and pulls IRQ low SYNC_WLIM1 shows current status DAC1 Alignment Tripped Status Mode. If SYNC_TRIP1 goes high, it latches and pulls IRQ low SYNC_TRIP1 shows current status Calibration Pass Status. If IRQEN_SMODE_CALPASS is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. Calibration passed Calibration Fail Detection Status. If IRQEN_SMODE_CALFAIL is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. Calibration failed DAC PLL Lost Status. If IRQEN_SMODE_DACPLLLOST is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC PLL lock was lost DAC PLL Lock Status. If IRQEN_SMODE_DACPLLLOCK is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC PLL locked SERDES PLL Lost Status. If IRQEN_SMODE_SERPLLLOST is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. SERDES PLL lock was lost Rev. A | Page 91 of 117 Reset 0x0 Access R/W 0x0 0x0 R R/W 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R 0x0 R 0x0 R 0x0 R 0x0 R AD9135/AD9136 Address Name Data Sheet Bit No. 2 Bit Name SERPLLLOCK 1 LANEFIFOERR 0 [7:3] 2 RESERVED RESERVED PRBS1 1 0 RESERVED PRBS0 7 PDPERR0 6 5 RESERVED BLNKDONE0 4 3 RESERVED SYNC_LOCK0 2 SYNC_ROTATE0 Settings 1 1 0x024 IRQ_STATUS1 1 1 0x025 IRQ_STATUS2 1 1 1 1 Description SERDES PLL Lock Status. If IRQEN_SMODE_SERPLLLOCK is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. SERDES PLL locked Lane FIFO Error Status. If IRQEN_SMODE_LANEFIFOERR is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. A lane FIFO error occurs when there is a full or empty condition on any of the FIFOs between the deserializer block and the core digital. This error requires a link disable and reenable to remove it. The status of the lane FIFOs can be found in Register 0x30C (FIFO full), and Register 0x30D (FIFO empty). Lane FIFO error Reserved. Reserved. DAC1 PRBS Error Status. If IRQEN_SMODE_PRBS1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 failed PRBS Reserved. DAC0 PRBS Error Status. If IRQEN_SMODE_PRBS0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 failed PRBS DAC0 PDP Error. If IRQEN_SMODE_PAERR0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. Data into DAC0 over power threshold Reserved. DAC0 Blanking Done Status. If IRQEN_SMODE_BLNKDONE0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 blanking done Reserved DAC0 LMFC Alignment Locked Status. If IRQEN_SMODE_SYNC_LOCK0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 LMFC alignment locked DAC0 LMFC Alignment Rotate Status. If IRQEN_SMODE_SYNC_ROTATE0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 LMFC alignment rotated Rev. A | Page 92 of 117 Reset 0x0 Access R 0x0 R 0x0 0x0 0x0 R R R 0x0 0x0 R R 0x0 R 0x0 0x0 R R 0x0 0x0 R R 0x0 R Data Sheet Address Name AD9135/AD9136 Bit No. 1 Bit Name SYNC_WLIM0 0 SYNC_TRIP0 7 PDPERR1 6 5 RESERVED BLNKDONE1 4 3 RESERVED SYNC_LOCK1 2 SYNC_ROTATE1 1 SYNC_WLIM1 0 SYNC_TRIP1 [7:6] 5 RESERVED ERR_DLYOVER 4 ERR_WINLIMIT Settings 1 1 0x026 IRQ_STATUS3 1 1 1 1 1 1 0x030 JESD_CHECKS 1 1 3 ERR_JESDBAD 1 2 ERR_KUNSUPP 1 Description DAC0 Outside Window Status. If IRQEN_SMODE_SYNC_WLIM0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 LMFC phase outside of window DAC0 LMFC Alignment Tripped Status. If IRQEN_SMODE_SYNC_TRIP0 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC0 LMFC alignment tripped DAC1 PDP Error. If IRQ_SMODE_PDPERR1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. Data into DAC1 over power threshold Reserved. DAC1 Blanking Done Status. If IRQEN_SMODE_BLNKDONE1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 blanking done Reserved. DAC1 LMFC Alignment Locked Status. If IRQEN_SMODE_SYNC_LOCK1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 LMFC alignment locked DAC1 LMFC Alignment Rotate Status. If IRQEN_SMODE_SYNC_ROTATE1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 LMFC alignment rotated DAC1 Outside Window Status. If IRQEN_SMODE_SYNC_WLIM1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 LMFC phase outside of window DAC1 LMFC Alignment Tripped Status. If IRQEN_SMODE_SYNC_TRIP1 is low, this bit shows current status. If not, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. DAC1 LMFC alignment tripped Reserved. Error: LMFC_Delay > JESD_K Parameter. LMFC_Delay > JESD_K Unsupported Window Limit. Unsupported SYSREF window limit Unsupported M/L/S/F Selection. This JESD combination is not supported Unsupported K Values. 16 and 32 are supported. K value unsupported Rev. A | Page 93 of 117 Reset 0x0 Access R 0x0 R 0x0 R 0x0 0x0 R R 0x0 0x0 R R 0x0 R 0x0 R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R 0x0 R AD9135/AD9136 Address Name Data Sheet Bit No. 1 Bit Name ERR_SUBCLASS 0 ERR_INTSUPP Settings 1 1 0x034 SYNC_ERRWINDOW [7:2] [1:0] RESERVED ERRWINDOW 0x038 SYNC_LASTERR_L [7:4] [3:0] RESERVED LASTERROR 0x039 SYNC_LASTERR_H 7 LASTUNDER 1 6 LASTOVER 1 0x03A SYNC_CONTROL [5:0] 7 RESERVED SYNCENABLE 1 0 6 SYNCARM 1 5 SYNCCLRSTKY 4 SYNCCLRLAST [3:0] SYNCMODE 0b0001 0b0010 0b1000 0b1001 0x03B SYNC_STATUS 7 SYNC_BUSY 1 [6:4] 3 RESERVED SYNC_LOCK 2 SYNC_ROTATE 1 1 1 SYNC_WLIM 1 0 SYNC_TRIP 1 Description Unsupported Subclass Value. 0 and 1 are supported. Unsupported subclass value Unsupported Interpolation Rate Factor. 1, 2, 4, 8 are supported. Unsupported interpolation rate factor Reserved. LMFC Sync Error Window. The error window allows the SYSREF sample phase to vary within the confines of the window without triggering a clock adjustment. This is useful if SYSREF cannot be guaranteed to always arrive in the same period of the device clock associated with the target phase. Error window tolerance = ± ERRWINDOW Reserved. LMFC Sync Last Alignment Error. 4-bit twos complement value that represents the phase error (in number of DAC clock cycles) when the clocks were last adjusted. LMFC Sync Last Error Under Flag. Last phase error was beyond lower window tolerance boundary LMFC Sync Last Error Over Flag. Last phase error was beyond upper window tolerance boundary Reserved. LMFC Sync Logic Enable. Enable sync logic Disable sync logic LMFC Sync Arming Strobe. Sync one-shot armed LMFC Sync Sticky Bit Clear. On a rising edge, this bit clears SYNC_ROTATE and SYNC_TRIP. LMFC Sync Clear Last Error. On a rising edge, this bit clears LASTERROR, LASTUNDER, LASTOVER. LMFC Sync Mode. Sync one-shot mode Sync continuous mode Sync monitor only mode Sync one-shot, then monitor LMFC Sync Machine Busy. Sync logic SM is busy Reserved. LMFC Sync Alignment Locked. Sync logic aligned within window LMFC Sync Rotated. Sync logic rotated with SYSREF (sticky) LMFC Sync Alignment Limit Range. Phase error outside window threshold LMFC Sync Tripped After Arming. Sync received SYSREF pulse (sticky) Rev. A | Page 94 of 117 Reset 0x0 Access R 0x0 R 0x0 0x0 R R/W 0x0 R R 0x0 R 0x0 R 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R 0x0 0x0 R R 0x0 R 0x0 R 0x0 R Data Sheet AD9135/AD9136 Address 0x03C Name SYNC_CURRERR_L Bit No. [7:4] [3:0] Bit Name RESERVED CURRERROR 0x03D SYNC_CURRERR_H 7 CURRUNDER Settings 1 6 CURROVER 1 [5:0] [7:2] [1:0] RESERVED RESERVED DACFSC_0[9:8] DACGAIN0_0 DACGAIN1_1 [7:0] [7:2] [1:0] DACFSC_0[7:0] RESERVED DACFSC_1[9:8] DACGAIN1_0 CLKCFG0 [7:0] 7 DACFSC_1[7:0] PD_CLK0 6 PD_CLK1 5 PD_CLK_DIG 4 PD_SERDES_PCLK 3 PD_CLK_REC [2:0] RESERVED 0x040 DACGAIN0_1 0x041 0x044 0x045 0x080 Description Reserved. LMFC Sync Alignment Error. 4-bit twos complement value that represents the phase error in number of DAC clock cycles (that is, number of DAC clocks between LMFC edge and SYSREF edge). When an adjustment of the clocks is made on any given SYSREF, the value of the phase error is placed into SYNC_ LASTERR, and SYNC_CURRERR is forced to 0. LMFC Sync Current Error Under Flag. Current phase error is beyond lower window tolerance boundary LMFC Sync Current Error Over Flag. Current phase error is beyond upper window tolerance boundary Reserved. Reserved. 2 MSBs of I-Channel DAC Gain DAC0. A 10-bit twos complement value that is mapped to analog full-scale current for DAC0 as shown: 01111111111 = 27.0 mA 0000000000 = 20.48 mA 1000000000 = 13.9 mA 8 LSBs of I-Channel DAC Gain DAC0. Reserved. 2 MSBs of Q-Channel DAC Gain DAC1. A 10-bit twos complement value that is mapped to analog full-scale current for DAC as shown in Register 0x040. 01111111111 = 27.0 mA 0000000000 = 20.48 mA 1000000000 = 13.9 mA 8 LSBs of Q-Channel DAC Gain DAC1. Power-Down Clock for DAC0. This bit disables the digital and analog clocks for DAC0. Power-Down Clock for DAC1. This bit disables the digital and analog clocks for DAC1. Power-Down Clocks to all DACs. This bit disables the digital and analog clocks for both duals. This includes all reference clocks, PCLK, DAC clocks, and digital clocks. Serdes PLL Clock Power-Down. This bit disables the reference clock to the SERDES PLL, which is needed to have an operational serial interface. Clock Receiver Power-Down. This bit powers down the analog DAC clock receiver block. With this bit set, clocks are not passed to internal nets. Reserved. Rev. A | Page 95 of 117 Reset 0x0 0x0 Access R R 0x0 R 0x0 R 0x0 0x0 0x0 R R R/W 0x0 0x0 0x0 R/W R R/W 0x0 0x1 R/W R/W 0x1 R/W 0x1 R/W 0x1 R/W 0x1 R/W 0x0 R AD9135/AD9136 Address 0x081 Name SYSREF_ACTRL0 Data Sheet Bit No. [7:5] 4 Bit Name RESERVED PD_SYSREF 3 HYS_ON 2 SYSREF_RISE Settings 0 1 0x082 0x083 0x084 SYSREF_ACTRL1 DACPLLCNTRL DACPLLSTATUS [1:0] HYS_CNTRL1 [7:0] 7 HYS_CNTRL0 RECAL_DACPLL [6:5] 4 RESERVED ENABLE_DACPLL [3:0] 7 RESERVED DACPLL_ OVERRANGE_H 6 DACPLL_ OVERRANGE_L 5 DACPLL_CAL_ VALID [4:2] 1 RESERVED DACPLL_LOCK 0x085 DACINTEGERWORD0 0 [7:0] RESERVED B_COUNT 0x087 DACLOOPFILT1 [7:4] LF_C2_WORD [3:0] LF_C1_WORD [7:4] LF_R1_WORD [3:0] LF_C3_WORD 0x088 DACLOOPFILT2 Description Reserved. Power-Down SYSREF Buffer. This bit powers down the SYSREF receiver. For Subclass 1 operation to work, this buffer must be enabled. Hysteresis Enabled. This bit enables the programmable hysteresis control for the SYSREF receiver. Using hysteresis gives some noise resistance, but delays the SYSREF± edge an amount depending on HYS_CNTRL and the SYSREF± edge rate. The SYSREF± KOW is not guaranteed when using hysteresis. Select DAC Clock Edge to Sample SYSREF. Use falling edge of DAC clock to sample SYSREF for alignment Use rising edge of DAC clock to sample SYSREF for alignment Hysteresis Control Bits[9:8]. HYS_CNTRL is a 10-bit thermometer-coded number. Each bit set adds 10 mV of differential hysteresis to the SYSREF receiver. Hysteresis Control Bits[7:0]. Recalibrate DAC PLL. On a rising edge of this bit, recalibrate the DAC PLL. Reserved. Synthesizer Enable. This bit enables and calibrates the DAC PLL. Reserved. DAC PLL High Overrange. This bit indicates that the DAC PLL hit the upper edge of its operating band. Recalibrate. DAC PLL Low Overrange. This bit indicates that the DAC PLL hit the lower edge of its operating band. Recalibrate. DAC PLL Calibration Valid. This bit indicates that the DAC PLL has been successfully calibrated. Reserved. DAC PLL Lock Bit. This bit is set high by the PLL when it has achieved lock. Reserved. Integer Division Word. This bit controls the integer feedback divider for the DAC PLL. Determine the frequency of the DAC clock by the following equations (see the Clock Multiplication section for more details): fDAC = fREF/(REF_DIVRATE) × 2 × B_COUNT fVCO = fREF/(REF_DIVRATE) × 2 × B_COUNT × LO_DIV_MODE Minimum value is 6. C2 Control Word. Set this control to 0x6 for optimal performance. C1 Control Word. Set this control to 0x2 for optimal performance. R1 Control Word. Set this control to 0xC for optimal performance. C3 Control Word. Set this control to 0x9 for optimal performance. Rev. A | Page 96 of 117 Reset 0x0 0x1 Access R R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 0x0 R/W R/W 0x0 0x0 R R/W 0x0 0x0 R R 0x0 R 0x0 R 0x0 0x0 R R 0x0 0x8 R R/W 0x8 R/W 0x8 R/W 0x8 R/W 0x8 R/W Data Sheet Address 0x089 Name DACLOOPFILT3 AD9135/AD9136 Bit No. 7 Bit Name LF_BYPASS_R3 6 LF_BYPASS_R1 5 LF_BYPASS_C2 4 LF_BYPASS_C1 [3:0] LF_R3_WORD 0x08A DACCPCNTRL [7:6] [5:0] RESERVED CP_CURRENT 0x08B DACLOGENCNTRL [7:2] [1:0] RESERVED LO_DIV_MODE Settings 01 10 11 0x08C DACLDOCNTRL1 [7:3] [2:0] RESERVED REF_DIV_MODE 000 001 010 011 100 0x08D DACLDOCNTRL2 [7:0] DAC_LDO 0x0E2 CAL_CTRL_GLOBAL [7:2] 1 RESERVED CAL_START_AVG 0 CAL_EN_AVG 1 Description Bypass R3 Resistor. When this bit is set, bypass the R3 capacitor (set to 0 pF) when R3_WORD is set to 0. Set this control to 0x0 for optimal performance. Bypass R1 Resistor. When this bit is set, bypass the R1 capacitor (set to 0 pF) when R1_WORD is set to 0. Set this control to 0x0 for optimal performance. Bypass C2 Capacitor. When this bit is set, bypass the C2 capacitor (set to 0 pF) when C2_WORD is set to 0. Set this control to 0x0 for optimal performance. Bypass C1 Capacitor. When this bit is set, bypass the C1 capacitor (set to 0 pF) when C1_WORD is set to 0. Set this control to 0x0 for optimal performance. R3 Control Word. Set this control to 0xE for optimal performance. Reserved. Charge Pump Current Control. Set this control to 0x12 for optimal performance. Reserved. This range controls the RF clock divider between the VCO and DAC clock rates. The options are 4×, 8×, or 16× division. Choose the LO_DIV_MODE so that 6 GHz < fVCO < 12 GHz (see the Clock Multiplication section for more details): DAC clock = VCO/4 DAC clock = VCO/8 DAC clock = VCO/16 Reserved. Reference Clock Division Ratio. This field controls the amount of division that is done to the input clock at the CLK+/CLK− pins before it is presented to the PLL as a reference clock. The reference clock frequency must be between 35 MHz and 80 MHz, but the CLK+/CLK− input frequency can range from 35 MHz to 1 GHz. The user sets this division to achieve a 35 MHz to 80 MHz PLL reference frequency. For more details see the Clock Multiplication section. 1 2 4 8 16 DAC PLL LDO setting. This register must be written to 0x7B for optimal performance. Reserved. Averaged Calibration Start. On rising edge, calibrate the DACs. Only use if calibrating all DACs. Averaged Calibration Enable. Set prior to starting calibration with CAL_START_AVG. While this bit is set, calibration can be performed, and the results are applied. Enable averaged calibration Rev. A | Page 97 of 117 Reset 0x0 Access R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x8 R/W 0x0 0x20 R R/W 0x0 0x2 R R/W 0x0 0x1 R R/W 0x2B R/W 0x0 0x0 R R/W 0x0 R/W AD9135/AD9136 Address 0x0E7 Name CAL_CLKDIV Data Sheet Bit No. [7:4] Bit Name RESERVED 3 CAL_CLK_EN Settings 1 0 0x0E8 CAL_PAGE [2:0] [7:4] [3:0] RESERVED RESERVED CAL_PAGE 0x0E9 CAL_CTRL 7 CAL_FIN 6 CAL_ACTIVE 5 CAL_ERRHI 4 CAL_ERRLO [3:2] 1 RESERVED CAL_START 1 1 1 1 0 1 0 CAL_EN 0 1 Description Must write the default value for proper operation. Enable Self Calibration Clock. Enable calibration clock Disable calibration clock Reserved. Reserved. DAC Calibration Paging. Selects which of the DACs are being accessed for calibration or calibration readback. This paging affects Register 0x0E9 and Register 0x0ED. Calibration: any number of DACs can be accessed simultaneously to write and calibrate. Write a 1 to Bit 0 to include DAC0. Write a 1 to Bit 2 to include DAC1. Readback: only one DAC at a time can be accessed when reading back CAL_CTRL (Register 0x0E9). Write a 1 to Bit 0 to read from DAC0 or write a 1 to Bit 2 to read from DAC1 (the other bits must be 0). Calibration finished. This bit is high when the calibration has completed. If the calibration completes and either CAL_ERRHI or CAL_ ERRLO is high, then the calibration cannot be considered valid and are considered a timeout event. Calibration ran and is finished Calibration Active. This bit is high while the calibration is in progress. Calibration is running SAR Data Error: Too High. This bit is set at the end of a calibration cycle if any of the calibration DACs has overranged to the high side. This typically means that the algorithm adjusts the calibration preset of the calibration DACs and runs another cycle. Data saturated high SAR Data Error: Too Low. This bit is set at the end of a calibration cycle if any of the calibration DACs has overranged to the low side. This typically means that the algorithm adjusts the calibration preset of the calibration DACs and runs another cycle. Data saturated low Reserved. Calibration Start. The rising edge of this bit kicks off a calibration sequence for the DACs that have been selected in the CAL_INDX register. Normal operation Start calibration state machine Calibration Enable. Enable the calibration DAC of the converter. Enable to calibration engine and machines. Prepare for a calibration start. For calibration coefficients to be applied to the calibrated DACs, this bit must be high. Do not use calibration DACs Use calibration DACs Rev. A | Page 98 of 117 Reset 0x3 Access R/W 0x0 R/W 0x0 0x0 0xF R R R/W 0x0 R 0x0 R 0x0 R 0x0 R 0x0 0x0 R R/W 0x0 R/W Data Sheet AD9135/AD9136 Address 0x0ED Name CAL_INIT Bit No. [7:0] Bit Name CAL_INIT Settings 0x110 DATA_FORMAT 7 BINARY_FORMAT 0 1 0x111 DATAPATH_CTRL [6:0] 7 RESERVED INVSINC_ENABLE 1 0 6 5 RESERVED DIG_GAIN_ENABLE 1 0 0x112 INTERP_MODE [4:0] [7:3] [2:0] RESERVED RESERVED INTERP_MODE 000 001 011 100 0x11F TXEN_SM_0 [7:6] FALL_COUNTERS [5:4] RISE_COUNTERS 3 2 RESERVED PROTECT_OUT_ INVERT 0 1 [1:0] RESERVED 0x121 TXEN_RISE_COUNT_0 [7:0] RISE_COUNT_0 0x122 TXEN_RISE_COUNT_1 [7:0] RISE_COUNT_1 0x123 TXEN_FALL_ COUNT_0 [7:0] FALL_COUNT_0 0x124 TXEN_FALL_ COUNT_1 [7:0] FALL_COUNT_1 0x12D DEVICE_CONFIG_ REG_0 DIE_TEMP_CTRL0 [7:0] DEVICE_CONFIG_0 [7:1] RESERVED 0 AUXADC_ENABLE 0x12F 0 1 Description Initialize Calibration. Must be written to 0xA2 before starting calibration or averaged calibration. Binary or Twos Complementary Format on the Data Bus. Input data is twos complement Input data is offset binary Reserved. Enable Inverse Sinc Filter. Enable inverse sinc filter Disable inverse sinc filter Reserved. Enable Digital Gain. Enable digital gain function Disable digital gain function Reserved Reserved. Interpolation Mode. 1× mode 2× mode 4× mode 8× mode Fall Counters. The number of counters to use to delay TX_PROTECT fall from TXENx falling edge. Must be set to 1 or 2. Rise Counters. The number of counters to use to delay TX_PROTECT rise from TXENx rising edge. Reserved. PROTECT_OUTx Invert. PROTECT_OUTx is high when output is valid. Suitable for enabling downstream components during transmission PROTECT_OUTx is high when output is invalid. Suitable for disabling downstream components when not transmitting Must write the default value for proper operation. First counter used to delay TX_PROTECT rise from TXENx rising edge. Delays by 32 × RISE_COUNT_0 DAC clock cycles. Second counter used to delay TX_PROTECT rise from TXENx rising edge. Delays by 32 × RISE_COUNT_1 DAC clock cycles. First counter used to delay TX_PROTECT fall from TXENx falling edge. Delays by 32 × FALL_COUNT_0 DAC clock cycles. Must be set to a minimum of 0x12. Second counter used to delay TX_PROTECT fall from TXENx falling edge. Delays by 32 × FALL_COUNT_1 DAC clock cycles. Must be set to 0x8B for proper digital datapath configuration. Must write the default value for proper operation. Enables the AUX ADC Block. AUX ADC disable AUX ADC enable Rev. A | Page 99 of 117 Reset 0xA6 Access R/W 0x0 R/W 0x0 0x1 R R/W 0x0 0x1 R R/W 0x0 0x0 0x1 R R R/W 0x2 R/W 0x0 R/W 0x0 0x0 R R/W 0x3 R/W 0xF R/W 0x0 R/W 0xFF R/W 0xFF R/W 0x46 R/W 0x10 R/W 0x0 R/W AD9135/AD9136 Data Sheet Address 0x132 0x133 0x134 Name DIE_TEMP0 DIE_TEMP1 DIE_TEMP_UPDATE Bit No. [7:0] [7:0] [7:1] 0 0x135 DC_OFFSET_CTRL [7:1] 0 Bit Name DIE_TEMP[7:0] DIE_TEMP[15:8] RESERVED DIE_TEMP_ UPDATE RESERVED DC_OFFSET_ON 0x136 DAC_DC_OFFSET_1 PART0 [7:0] LSB_OFFSET[7:0] 0x137 DAC_DC_OFFSET_ 1PART1 [7:0] LSB_OFFSET[15:8] 0x13A DAC_DC_OFFSET_ 2PART [7:5] [4:0] RESERVED SIXTEENTH_ OFFSET 0x13C DAC_DIG_GAIN0 [7:0] DAC_DIG_ GAIN[7:0] 0x13D DAC_DIG_GAIN1 [7:4] [3:0] 0x140 GAIN_RAMP_UP_ STEP0 [7:0] RESERVED DAC_DIG_ GAIN[11:8] GAIN_RAMP_UP_ STEP[7:0] Settings 1 x 0x0 0xFFF 0x141 0x142 GAIN_RAMP_UP_ STEP1 GAIN_RAMP_DOWN_ STEP0 [7:4] RESERVED [3:0] GAIN_RAMP_UP_ STEP[11:8] GAIN_RAMP_ DOWN_STEP[7:0] [7:0] 0 0xFFF 0x143 0x146 0x147 GAIN_RAMP_ DOWN_STEP1 DEVICE_CONFIG_ REG_1 BSM_STAT [7:4] RESERVED [3:0] [7:0] GAIN_RAMP_ DOWN_STEP[11:8] DEVICE_CONFIG_1 [7:6] SOFTBLANKRB 00 01 10 11 [5:0] RESERVED Description Aux ADC Readback Value. Aux ADC Readback Value. Reserved. Die Temperature Update. On a rising edge, a new temperature code is generated. Reserved. DC Offset On. Enables dc offset module 8 LSBs of DC Offset. LSB_OFFSET is a 16-bit twos complement number that is added to incoming data. Applies to the DAC selected by DAC_PAGE (Register 0x008 [1:0]). 8 MSBs of DC Offset. LSB_OFFSET is a 16-bit twos complement number that is added to incoming data. Applies to the DAC selected by DAC_PAGE (Register 0x008 [1:0]). Reserved. SIXTEENTH_OFFSET is a 5-bit twos complement number in 16ths of an LSB that is added to incoming I data. x/16 LSB DC offset 8 LSBs of DAC Digital Gain. DAC_DIG_GAIN is the digital gain of the DAC selected by DAC_PAGE (Register 0x008 [1:0]). The digital gain is a multiplier from 0 to 4095/2048 in steps of 1/2048. Reserved. 4 MSBs of DAC Digital Gain 8 LSBs of Gain Ramp Up Step. GAIN_RAMP_UP_STEP controls the amplitude step size of the BSM’s ramping feature when the gain is being ramped to its assigned value. Smallest ramp up step size Largest ramp up step size Reserved. 4 MSBs of Gain Ramp Up Step. See Register 0x140 for description. 8 LSBs of Gain Ramp Down Step. GAIN_RAMP_DOWN_STEP controls the amplitude step size of the BSM’s ramping feature when the gain is being ramped to zero. Smallest ramp down step size Largest ramp down step size Reserved. 4 MSBs of Gain Ramp Down Step. See Register 0x142 for description. Must be set to 0x01 for proper digital datapath configuration. Blanking State. Data is fully blanked Ramping from data process to full blanking Ramping from fully blanked to data process Data is being processed Reserved. Rev. A | Page 100 of 117 Reset 0x0 0x0 0x0 0x0 Access R R R R/W 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 0x0 R R/W 0xEA R/W 0x0 0xA R R/W 0x4 R/W 0x0 R 0x0 R/W 0x9 R/W 0x0 R 0x0 R/W 0x0 R/W 0x0 R 0x0 R Data Sheet Address 0x14B Name PRBS AD9135/AD9136 Bit No. 7 6 Bit Name RESERVED PRBS_GOOD Settings 0 1 [5:3] 2 RESERVED PRBS_MODE 0 1 1 PRBS_RESET 0 1 0 PRBS_EN 0 1 0x14C 0x1B0 PRBS_ERROR DACPLLT0 [7:0] [7:0] PRBS_COUNT DAC_PLL_PWR 0x1B5 DACPLLT5 [7:4] RESERVED [3:0] VCO_VAR 0x1B9 DACPLLT9 [7:0] DAC_PLL_CP1 0x1BB DACPLLTB [7:5] [4:3] RESERVED VCO_BIAS_TCF [2:0] VCO_BIAS_REF 0x1BC DACPLLTC [7:0] DAC_PLL_VCO_ CTRL 0x1BE DACPLLTE [7:0] DAC_PLL_VCO_ PWR 0x1BF DACPLLTF [7:0] DAC_PLL_VCOCAL 0x1C0 DACPLLT10 [7:0] 0x1C1 DACPLLT11 [7:0] DAC_PLL_LOCK_ CNTR DAC_PLL_CP2 0x1C4 DACPLLT17 [7:0] DAC_PLL_VAR1 0x1C5 DACPLLT18 [7:0] DAC_PLL_VAR2 0x200 MASTER_PD [7:1] 0 RESERVED SPI_PD_MASTER 0x201 PHY_PD [7:0] SPI_PD_PHY Description Reserved. Good Data Indicator. Incorrect sequence detected Correct PRBS sequence detected Reserved. Polynomial Select 7-bit: x7 + x6 + 1 15-bit: x15 + x14 + 1 Reset Error Counters. Normal operation Reset counters Enable PRBS Checker. Disable Enable Error Count Value. DAC PLL PD settings. This register must be written to 0x00 for optimal performance. Must write the default value for proper operation. Varactor KVO Setting. See Table 73 for optimal settings based on the fVCO being used. DAC PLL Charge Pump settings. This register must be written to 0x24 for optimal performance. Reserved. Temperature Coefficient for VCO Bias. See Table 73 for optimal settings based on the fVCO being used. VCO Bias Control. See Table 73 for optimal settings based on the fVCO being used. DAC PLL VCO control settings. This register must be written to 0x0D for optimal performance. DAC PLL VCO power control settings. This register must be written to 0x02 for optimal performance. DAC PLL VCO calibration settings. This register must be written to 0x8E for optimal performance. This register must be written to 0x2A for optimal performance. This register must be written to0x2A for optimal performance. DAC PLL Varactor setting. Must be set to 0x7E for proper DAC PLL configuration. DAC PLL Varactor setting. See Table 73 for optimal settings based on the fVCO being used. Reserved. Power Down the Entire JESD Receiver Analog (All Eight Channels Plus Bias). SPI Override to Power Down the Individual PHYs. Set Bit x to power down the corresponding SERDINx± PHY Rev. A | Page 101 of 117 Reset 0x0 0x0 Access R R 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 0xFA R R/W 0x8 R/W 0x3 R/W 0x34 R/W 0x0 0x1 R R/W 0x4 R/W 0x00 R/W 0x00 R/W 0x8D R/W 0x2E R/W 0x24 R/W 0x33 R/W 0x08 R/W 0x0 0x1 R R/W 0x0 R/W AD9135/AD9136 Address 0x203 Name GENERIC_PD 0x206 CDR_RESET Data Sheet Bit No. [7:2] 1 Bit Name RESERVED SPI_SYNC1_PD Settings Description Reserved. Power down LVDS buffer for SYNCOUT0±. 0x232 0x268 CDR_OPERATING_ MODE_REG_0 DEVICE_CONFIG_ REG_3 EQ_BIAS_REG SPI_SYNC2_PD Power down LVDS buffer for SYNCOUT1±. 0x0 R/W [7:1] 0 RESERVED SPI_CDR_RESETN Reserved. Resets the Digital Control Logic for All PHYs. Hold CDR in reset Enable CDR Reserved. Enables Half-Rate CDR Operation. Set to 1 when 5.75 Gbps ≤ lane rate ≤ 10.64. Must write the default value for proper operation. Enables Oversampling of the Input Data. Set to 1 when 1.44 Gbps ≤ lane rate ≤ 2.76 Gbps. Reserved. Must be set to 0xFF for proper JESD interface configuration. Control the Equalizer Power/Insertion Loss Capability. Normal mode Low power mode Must write the default value for proper operation. Reserved. Recalibrate SERDES PLL. On a rising edge, recalibrate the SERDES PLL. Reserved. Enable the SERDES PLL. Setting this bit enables and calibrates the SERDES PLL. Reserved. SERDES PLL High Overrange. This bit indicates that the SERDES PLL hit the lower edge of its operating band. Recalibrate. SERDES PLL Low Overrange. This bit indicates that the SERDES PLL hit the lower edge of its operating band. Recalibrate. SERDES PLL Calibration Valid. This bit indicates that the SERDES PLL has been successfully calibrated. Reserved. SERDES PLL Lock. This bit is set high by the PLL when it has achieved lock. SERDES PLL loop filter setting. This register must be written to 0x62 for optimal performance. SERDES PLL loop filter setting. This register must be written to 0xC9 for optimal performance. SERDES PLL loop filter setting. This register must be written to 0x0E for optimal performance. SERDES PLL charge pump setting. This register must be written to 0x12 for optimal performance. 0x0 0x1 R R/W 0x0 0x1 R R/W 0x2 R/W 0x0 R/W 0x0 0x0 R R/W 0x1 R/W 0x22 R/W 0x0 0x0 R R/W 0x0 0x0 R R/W 0x0 0x0 R R 0x0 R 0x0 R 0x0 0x0 R R 0x77 R/W 0x87 R/W 0x08 R/W 0x3F R/W [7:6] 5 RESERVED ENHALFRATE [4:2] RESERVED 1 CDR_OVERSAMP 0 [7:0] RESERVED DEVICE_CONFIG_3 [7:6] EQ_POWER_ MODE 00 01 0x280 0x281 SERDESPLL_ ENABLE_CNTRL PLL_STATUS Access R R/W 0 0 1 0x230 Reset 0x0 0x0 [5:0] RESERVED [7:3] 2 RESERVED RECAL_SERDESPLL 1 0 RESERVED ENABLE_ SERDESPLL RESERVED SERDES_PLL_ OVERRANGE_H [7:6] 5 4 SERDES_PLL_ OVERRANGE_L 3 SERDES_PLL_CAL_ VALID_RB [2:1] 0 0x284 LOOP_FILTER_1 [7:0] RESERVED SERDES_PLL_ LOCK_RB LOOP_FILTER_1 0x285 LOOP_FILTER_2 [7:0] LOOP_FILTER_2 0x286 LOOP_FILTER_3 [7:0] LOOP_FILTER_3 0x287 SERDES_PLL_CP1 [7:0] SERDES_PLL_CP1 Rev. A | Page 102 of 117 Data Sheet Address 0x289 Name REF_CLK_DIVIDER_ LDO AD9135/AD9136 Bit No. [7:3] 2 Bit Name RESERVED DEVICE_CONFIG_4 [1:0] SERDES_PLL_DIV_ MODE Settings 00 01 10 0x28A VCO_LDO [7:0] SERDES_PLL_ VCO_LDO 0x28B SERDES_PLL_PD1 [7:0] SERDES_PLL_PD1 0x290 SERDESPLL_VAR1 [7:0] SERDES_PLL_VAR1 0x294 SERDES_PLL_CP2 [7:0] SERDES_PLL_CP2 0x296 SERDESPLL_VCO1 [7:0] 0x297 SERDESPLL_VCO2 [7:0] 0x299 SERDES_PLL_PD2 [7:0] SERDES_PLL_ VCO1 SERDES_PLL_ VCO2 SERDES_PLL_PD2 0x29A SERDESPLL_VAR2 [7:0] SERDES_PLL_VAR2 0x29C SERDES_PLL_CP3 [7:0] SERDES_PLL_CP3 0x29F SERDESPLL_VAR3 [7:0] SERDES_PLL_VAR3 0x2A0 SERDESPLL_VAR4 [7:0] SERDES_PLL_VAR4 0x2A4 DEVICE_CONFIG_ REG_8 SYNCOUTB_SWING [7:0] DEVICE_CONFIG_8 [7:1] 0 RESERVED SYNCOUTB_ SWING_MD 0x2A5 0 1 0x2A7 0x2AA 0x2AB TERM_BLK1_ CTRLREG0 DEVICE_CONFIG_ REG_9 DEVICE_CONFIG_ REG_10 [7:1] RESERVED 0 RCAL_TERMBLK1 [7:0] DEVICE_CONFIG_ 9 DEVICE_CONFIG_ 10 [7:0] Description Reserved. Must be set to 1 for proper SERDES PLL configuration. SERDES PLL Reference Clock Division Factor. This field controls the division of the SERDES PLL reference clock before it is fed into the SERDES PLL Phase Frequency Detector (PFD). It must be set so fREF/DivFactor is between 35 MHz and 80 MHz. Divide by 4 for 5.75 Gbps to 10.64 Gbps lane rate Divide by 2 for 2.88 Gbps to 5.52 Gbps lane rate Divide by 1 for 1.44 Gbps to 2.76 Gbps lane rate SERDES PLL VCO LDO setting. This register must be written to 0x7B for optimal performance. SERDES PLL PD setting. This register must be written to 0x00 for optimal performance. SERDES PLL Varactor setting. This register must be written to 0x89 for optimal performance. SERDES PLL Charge Pump setting. This register must be set to 0x24 for optimal performance. SERDES PLL VCO setting. This register must be set to 0x03 for optimal performance. SERDES PLL VCO setting. This register must be set to 0x0D for optimal performance. SERDES PLL PD setting. This register must be set to 0x02 for optimal performance. SERDES PLL Varactor setting. This register must be set to 0x8E for optimal performance. SERDES PLL Charge Pump setting. Must be set to 0x2A for proper SERDES PLL configuration. SERDES PLL Varactor setting. Must be set to 0x78 for proper SERDES PLL configuration. SERDES PLL Varactor setting. This register must be set to 0x06 for optimal performance. Must be set to 0xFF for proper clock configuration. Reserved. SYNCOUTx± Swing Mode. Sets the output differential swing mode for the SYNCOUTx± pins. See Table 8 for details. Normal Swing Mode High Swing Mode Reserved. Reset 0x0 0x0 Access R R/W 0x0 R/W 0x2B R/W 0x7F R/W 0x83 R/W 0xB0 R/W 0x0C R/W 0x00 R/W 0x00 R/W 0xFE R/W 0x17 R/W 0x33 R/W 0x08 R/W 0x4B R/W 0x0 0x0 R R/W 0x0 R Termination Calibration. The rising edge of this bit calibrates PHY0, PHY1, PHY6, and PHY7 terminations to 50 Ω. Must be set to 0xB7 for proper JESD interface termination configuration. Must be set to 0x87 for proper JESD interface termination configuration. 0x0 R/W 0xC3 R/W 0x93 R/W Rev. A | Page 103 of 117 AD9135/AD9136 Data Sheet Address 0x2AE Name TERM_BLK2_ CTRLREG0 Bit No. [7:1] 0 Bit Name RESERVED RCAL_TERMBLK2 0x2B1 DEVICE_CONFIG_ REG_11 DEVICE_CONFIG_ REG_12 GENERAL_JRX_ CTRL_0 [7:0] DEVICE_CONFIG_ 11 DEVICE_CONFIG_ 12 RESERVED CHECKSUM_MODE 0x2B2 0x300 [7:0] 7 6 Settings 0 1 [5:4] 3 RESERVED LINK_MODE 0 1 2 LINK_PAGE 0 1 [1:0] LINK_EN 0b00 0b01 0b10 0b11 0x301 GENERAL_JRX_CTRL_1 [7:3] [2:0] RESERVED SUBCLASSV_ LOCAL 000 001 0x302 DYN_LINK_LATENCY_0 [7:5] [4:0] RESERVED DYN_LINK_ LATENCY_0 0x303 DYN_LINK_LATENCY_1 [7:5] [4:0] RESERVED DYN_LINK_ LATENCY_1 Description Reserved. Terminal Calibration. The rising edge of this bit calibrates PHY2, PHY3, PHY4 and PHY5 terminations to 50 Ω. Must be set to 0xB7 for proper JESD interface termination configuration. Must be set to 0x87 for proper JESD interface termination configuration. Reserved. Checksum Mode. This bit controls the locally generated JESD204B link parameter checksum method. The value is stored in the FCMP registers (Register 0x40E, Register 0x416, Register 0x41E, Register 0x426, Register 0x42E, Register 0x436, Register 0x43E, and Register 0x446). Checksum is calculated by summing the individual fields in the link configuration table as defined in Section 8.3, Table 20 of the JESD204B standard Checksum is calculated by summing the registers containing the packed link configuration fields (Σ[0x400:0x40A] modulo 256). Reserved. Link Mode. This register selects either singlelink or dual-link mode. Single-link mode Dual-link mode Link Paging. Selects which link’s register map is used. This paging affects Registers 0x401 to 0x47E. Use Link 0 register map Use Link 1 register map Link Enable. These bits bring up the JESD204B receiver digital circuitry: Bit 0 for Link 0 and Bit 1 for Link 1. Enable the link only after the following has occurred: all JESD204B parameters are set, the DAC PLL is enabled and locked (Register 0x084[1] = 1), and the JESD204B PHY is enabled (Register 0x200 = 0x00) and calibrated (Register 0x281[2] = 0). Disable both JESD Link 1 and JESD Link 0 Disable JESD Link 1, enable JESD Link 0 Enable JESD Link 1, disable JESD Link 0 Enable both JESD Link 1 and JESD Link 0 Reserved. JESD204B Subclass. Subclass 0 Subclass 1 Reserved. Dynamic Link Latency: Link 0. Latency between the LMFCRx for Link 0 and the last arriving LMFC boundary in units of PCLK cycles. See the Deterministic Latency section. Reserved. Dynamic Link Latency: Link 1. Latency between the LMFCRx for Link 1 and the last arriving LMFC boundary in units of PCLK cycles. See the Deterministic Latency section. Rev. A | Page 104 of 117 Reset 0x0 0x0 Access R R/W 0xC3 R/W 0x93 R/W 0x0 0x0 R R/W 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 0x1 R R/W 0x0 0x0 R R 0x0 0x0 R R Data Sheet AD9135/AD9136 Address 0x304 Name LMFC_DELAY_0 Bit No. [7:5] [4:0] Bit Name RESERVED LMFC_DELAY_0 Settings 0x305 LMFC_DELAY_1 [7:5] [4:0] RESERVED LMFC_DELAY_1 0x306 LMFC_VAR_0 [7:5] [4:0] RESERVED LMFC_VAR_0 0x307 LMFC_VAR_1 [7:5] [4:0] RESERVED LMFC_VAR_1 0x308 XBAR_LN_0_1 [7:6] [5:3] RESERVED LOGICAL_LANE1_ SRC [2:0] LOGICAL_LANE0_ SRC [7:6] [5:3] RESERVED LOGICAL_LANE3_ SRC [2:0] LOGICAL_LANE2_ SRC [7:6] [5:3] RESERVED LOGICAL_LANE5_ SRC [2:0] LOGICAL_LANE4_ SRC [7:6] [5:3] RESERVED LOGICAL_LANE7_ SRC [2:0] LOGICAL_LANE6_ SRC [7:0] LANE_FIFO_FULL x x 0x309 XBAR_LN_2_3 x x 0x30A XBAR_LN_4_5 x x 0x30B XBAR_LN_6_7 x x 0x30C FIFO_STATUS_REG_0 Description Reserved. LMFC Delay: Link 0 Delay from the LMFC to LMFCRx for Link 0. In units of frame clock cycles for subclass 1 and PCLK cycles for subclass 0. See the Deterministic Latency section. Reserved. LMFC Delay: Link 1. Delay from the LMFC to LMFCRx for Link 1. In units of frame clock cycles for subclass 1 and PCLK cycles for subclass 0. See the Deterministic Latency section. Reserved. Variable Delay Buffer: Link 0. Sets when data is read from a buffer to be consistent across links and power cycles. In units of PCLK cycles. See the Deterministic Latency section. This setting must not be more than 10. Reserved. Variable Delay Buffer: Link 1. Sets when data is read from a buffer to be consistent across links and power cycles. In units of PCLK cycles. See the Deterministic Latency section. This setting must not be more than 10. Reserved. Logical Lane 1 Source. Selects a physical lane to be mapped onto Logical Lane 1. Data is from SERDINx Logical Lane 0 Source. Selects a physical lane to be mapped onto Logical Lane 0. Data is from SERDINx Reserved. Logical Lane 3 Source. Selects a physical lane to be mapped onto Logical Lane 3. Data is from SERDINx Logical Lane 2 source. Selects a physical lane to be mapped onto Logical Lane 2. Data is from SERDINx Reserved. Logical Lane 5 Source. Selects a physical lane to be mapped onto Logical Lane 5. Data is from SERDINx Logical Lane 4 Source. Selects a physical lane to be mapped onto Logical Lane 4. Data is from SERDINx Reserved. Logical Lane 7 Source. Selects a physical lane to be mapped onto Logical Lane 7. Data is from SERDINx Logical Lane 6 Source. Selects a physical lane to be mapped onto Logical Lane 6. Data is from SERDINx FIFO Full Flags for Each Logical Lane. A full FIFO indicates an error in the JESD204B configuration or with a system clock. If the FIFO for Lane x is full, Bit x in this register will be high. Rev. A | Page 105 of 117 Reset 0x0 0x0 Access R R/W 0x0 0x0 R R/W 0x0 0x6 R R/W 0x0 0x6 R R/W 0x0 0x1 R R/W 0x0 R/W 0x0 0x3 R R/W 0x2 R/W 0x0 0x5 R R/W 0x4 R/W 0x0 0x7 R R/W 0x6 R/W 0x0 R AD9135/AD9136 Data Sheet Address 0x30D Name FIFO_STATUS_REG_1 Bit No. [7:0] Bit Name LANE_FIFO_EMPTY 0x312 SYNCB_GEN_1 [7:6] [5:4] RESERVED SYNCB_ERR_DUR Settings 0 1 2 0x314 SERDES_SPI_REG [3:0] [7:0] 0x315 PHY_PRBS_TEST_EN [7:0] RESERVED SERDES_SPI_ CONFIG PHY_TEST_EN 0x316 PHY_PRBS_TEST_CTRL 7 [6:4] RESERVED PHY_SRC_ERR_CNT [3:2] PHY_PRBS_PAT_SEL x 00 01 10 1 PHY_TEST_START 0 1 0 PHY_TEST_RESET 0 1 0x317 0x318 0x319 0x31A 0x31B 0x31C 0x31D PHY_PRBS_TEST_ THRESHOLD_LOBITS PHY_PRBS_TEST_ THRESHOLD_ MIDBITS PHY_PRBS_TEST_ THRESHOLD_HIBITS PHY_PRBS_TEST_ ERRCNT_LOBITS [7:0] PHY_PRBS_TEST_ ERRCNT_MIDBITS PHY_PRBS_TEST_ ERRCNT_HIBITS PHY_PRBS_TEST_ STATUS [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] PHY_PRBS_ THRESHOLD[7:0] PHY_PRBS_ THRESHOLD[15:8] PHY_PRBS_ THRESHOLD[23:16] PHY_PRBS_ERR_ CNT[7:0] PHY_PRBS_ERR_ CNT[15:8] PHY_PRBS_ERR_ CNT[23:16] PHY_PRBS_PASS Description FIFO Empty Flags for Each Logical Lane. An empty FIFO indicates an error in the JESD204B configuration or with a system clock. If the FIFO for Logical Lane x is empty, Bit x in this register will be high. Reserved. Duration of SYNCOUTx± Low for Error. The duration applies to both SYNCOUT0 and SYNCOUT1. A sync error is asserted at the end of a multiframe whenever one or more disparity, not in table or unexpected control character errors are encountered. ½ PCLK cycle 1 PCLK cycle 2 PCLK cycles Reserved. SERDES SPI Configuration. Must be written to 0x01 as part of the Physical Layer setup step. PHY Test Enable. Enables the PHY BER test. Set Bit x to enable the PHY test for Lane x. Reserved. PHY Error Count Source. Selects which PHY errors are being reported in Register 0x31A to Register 0x31C. Report Lane x error count PHY PRBS Pattern Select. Selects the PRBS pattern for PHY BER test. PRBS7 PRBS15 PRBS31 PHY PRBS Test Start. Starts and stops the PHY PRBS test. Test stopped Test in progress PHY PRBS Test Reset. Resets the PHY PRBS test state machine and error counters. Enable PHY PRBS test state machine Hold PHY PRBS test state machine in reset 8 LSBs of PHY PRBS Error Threshold. Reset 0x0 Access R 0x0 R/W 0x0 0x0 R/W R/W 0x0 R/W 0x0 0x0 R R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R/W 8 ISBs of PHY PRBS Error Threshold. 0x0 R/W 8 MSBs of PHY PRBS Error Threshold. 0x0 R/W 8 LSBs of PHY PRBS Error Count. Reported PHY BERT error count from lane selected using Register 0x316[6:4]. 8 ISBs of PHY PRBS Error Count. 0x0 R 0x0 R 8 MSBs of PHY PRBS Error Count. 0x0 R PHY PRBS Test Pass/Fail. Bit x corresponds to PHY PRBS pass/fail for Physical Lane x. The bit is set to 1 while the error count for Physical Lane x is less than PHY_PRBS_THRESHOLD. 0xFF R Rev. A | Page 106 of 117 Data Sheet Address 0x32C Name SHORT_TPL_TEST_0 AD9135/AD9136 Bit No. [7:6] [5:4] Bit Name RESERVED SHORT_TPL_SP_ SEL [3:2] SHORT_TPL_DAC_ SEL Settings x 0 2 1 SHORT_TPL_TEST_ RESET 0 1 0 SHORT_TPL_TEST_ EN 0 1 0x32D SHORT_TPL_TEST_1 [7:0] SHORT_TPL_REF_ SP_LSB 0x32E SHORT_TPL_TEST_2 [7:0] SHORT_TPL_REF_ SP_MSB 0x32F SHORT_TPL_TEST_3 [7:1] 0 RESERVED SHORT_TPL_FAIL 0 1 0x333 [7:0] DEVICE_CONFIG_ 13 JESD_BIT_INVERSE 0x400 DEVICE_CONFIG_ REG_13 JESD_BIT_INVERSE_ CTRL DID_REG [7:0] DID_RD 0x401 BID_REG [7:4] ADJCNT_RD [3:0] BID_RD 7 6 RESERVED ADJDIR_RD 5 PHADJ_RD [4:0] LID0_RD 0x334 0x402 LID0_REG [7:0] Description Reserved. Short Transport Layer Sample Select. Selects which sample to check from the DAC selected via Bits[3:2]. Sample x Short Transport Layer Test DAC Select. Selects which DAC to sample. Sample from DAC0 Sample from DAC1 Short Transport Layer Test Reset. Resets the result of short transport layer test. Not reset Reset Short Transport Layer Test Enable. See the Subclass 0 section for details on how to perform this test. Disable Enable Short Transport Layer Test Reference, Sample LSB. This is the lower eight bits of the expected DAC sample. It is used to compare with the received DAC sample at the output of the JESD204B receiver. Short Transport Layer Test Reference, Sample MSB. This is the upper eight bits of the expected DAC sample. It is used to compare with the received DAC sample at the output of the JESD204B receiver. Reserved. Short Transport Layer Test Fail. This bit shows whether the selected DAC sample matches the reference sample. If they match, it is a test pass, otherwise it is a test fail. Test pass Test fail Must be set to 0x01 for proper JESD interface configuration. Logical Lane Invert. Set Bit x high to invert the JESD deserialized data on Logical Lane x. Device Identification Number. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Adjustment Resolution to DAC LMFC. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be 0. Bank Identification: Extension to DID. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Reserved. Direction to Adjust DAC LMFC. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be 0. Phase Adjustment Request to DAC Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be 0. Lane Identification for Lane 0. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Rev. A | Page 107 of 117 Reset 0x0 0x0 Access R R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x0 0x0 R R 00 R/W 0x0 R/W 0x0 R 0x0 R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R AD9135/AD9136 Address 0x403 Name SCR_L_REG Data Sheet Bit No. 7 Bit Name SCR_RD Settings 0 1 [6:5] [4:0] RESERVED L-1_RD 0 1 3 7 0x404 F_REG [7:0] F-1_RD 0 1 3 0x405 K_REG [7:5] [4:0] RESERVED K-1_RD 0x0F 0x1F 0x406 M_REG [7:0] M-1_RD 0 1 0x407 CS_N_REG [7:6] CS_RD 5 [4:0] RESERVED N-1_RD [7:5] SUBCLASSV_RD [4:0] NP-1_RD 0x0F 0x408 NP_REG 0x0F Description Transmit Scrambling Status. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Scrambling is disabled Scrambling is enabled Reserved. Number of Lanes per Converter Device. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. One lane per converter Two lanes per converter Four lanes per converter Eight lanes per converter (single link only) Number of Octets per Frame. Settings of 1, 2 and 4 octets per frame are valid. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. (One octet per frame) per lane (Two octets per frame) per lane (Four octets per frame) per lane Reserved. Number of Frames per Multiframe. Settings of 16 or 32 are valid. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. 16 frames per multiframe 32 frames per multiframe Number of converters per device. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be 0 or 1. One converter per device Two converters per device Number of Control Bits per Sample. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. CS must be 0. Reserved. Converter Resolution. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Converter resolution must be 16. Converter resolution of 16 Device Subclass Version. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Total Number of Bits per Sample. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be 16 bits per sample. 16 bits per sample. Rev. A | Page 108 of 117 Reset 0x0 Access R 0x0 0x0 R R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R Data Sheet Address 0x409 Name S_REG AD9135/AD9136 Bit No. [7:5] Bit Name JESDV_RD Settings 000 001 [4:0] S-1_RD 0 1 0x40A HD_CF_REG 7 HD_RD 0 1 [6:5] [4:0] RESERVED CF_RD 0x40B RES1_REG [7:0] RES1_RD 0x40C RES2_REG [7:0] RES2_RD 0x40D CHECKSUM_REG [7:0] FCHK0_RD 0x40E COMPSUM0_REG [7:0] FCMP0_RD 0x412 LID1_REG [7:5] [4:0] RESERVED LID1_RD 0x415 CHECKSUM1_REG [7:0] FCHK1_RD 0x416 COMPSUM1_REG [7:0] FCMP1_RD 0x41A LID2_REG 0x41D 0x41E CHECKSUM2_REG COMPSUM2_REG [7:5] [4:0] [7:0] [7:0] RESERVED LID2_RD FCHK2_RD FCMP2_RD 0x422 LID3_REG 0x425 0x426 CHECKSUM3_REG COMPSUM3_REG [7:5] [4:0] [7:0] [7:0] RESERVED LID3_RD FCHK3_RD FCMP3_RD Description JESD204 Version. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. JESD204A JESD204B Number of Samples per Converter per Frame Cycle. Settings of one and two are valid. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. One sample per converter per frame Two samples per converter per frame High Density Format. See Section 5.1.3 of the JESD294B standard. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Low density mode High density mode: link information received on Lane 0 as specified in Section 8.3 of JESD204B Reserved. Number of Control Words per Frame Clock Period per Link. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Bits[4:0] must be 0. Reserved Field 1. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Reserved Field 2. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Checksum for Link Lane 0. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Computed Checksum for Link Lane 0. The JESD204B receiver computes the checksum of the link information received on Lane 0 as specified in Section 8.3 of JESD204B. The computation method is set by the CHECKSUM_MODE bit (Address 0x300[6]) and must match the likewise calculated checksum in Register 0x40D. Reserved. Lane Identification for Link Lane 1.Link information received on Lane 0 as specified in section 8.3 of JESD204B. Checksum for Link Lane 1. Link information received on Lane 0 as specified in Section 8.3 of JESD204B. Computed Checksum for Link Lane 1. See the description for Register 0x40E. Reserved. Lane Identification for Link Lane 2. Checksum for Link Lane 2. Computed Checksum for Link Lane 2 (see the description for Register 0x40E). Reserved. Lane Identification for Link Lane 3. Checksum for Link Lane 3. Computed Checksum for Link Lane 3 (see the description for Register 0x40E). Rev. A | Page 109 of 117 Reset 0x0 Access R 0x0 R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R 0x0 R 0x0 R 0x0 0x0 R R 0x0 R 0x0 R 0x0 0x0 0x0 0x0 R R R R 0x0 0x0 0x0 0x0 R R R R AD9135/AD9136 Address 0x42A Name LID4_REG 0x42D 0x42E CHECKSUM4_REG COMPSUM4_REG 0x432 LID5_REG 0x435 0x436 CHECKSUM5_REG COMPSUM5_REG 0x43A LID6_REG 0x43D 0x43E CHECKSUM6_REG COMPSUM6_REG 0x442 LID7_REG 0x445 0x446 Data Sheet Bit No. [7:5] [4:0] [7:0] [7:0] Bit Name RESERVED LID4_RD FCHK4_RD FCMP4_RD [7:5] [4:0] [7:0] [7:0] RESERVED LID5_RD FCHK5_RD FCMP5_RD [7:5] [4:0] [7:0] [7:0] RESERVED LID6_RD FCHK6_RD FCMP6_RD CHECKSUM7_REG COMPSUM7_REG [7:5] [4:0] [7:0] [7:0] RESERVED LID7_RD FCHK7_RD FCMP7_RD 0x450 ILS_DID [7:0] DID 0x451 ILS_BID [7:4] ADJCNT [3:0] BID 7 6 5 RESERVED ADJDIR PHADJ [4:0] LID0 7 SCR 0x452 0x453 ILS_LID0 ILS_SCR_L Settings 0 1 [6:5] [4:0] RESERVED L-1 0 1 3 7 0x454 ILS_F [7:0] F-1 0 1 3 0x455 ILS_K [7:5] [4:0] RESERVED K-1 0x0F 0x1F Description Reserved. Lane Identification for Link Lane 4. Checksum for Link Lane 4. Computed Checksum for Link Lane 4 (see the description for Register 0x40E). Reserved. Lane Identification for Link Lane 5. Checksum for Link Lane 5. Computed Checksum for Link Lane 5 (see the description for Register 0x40E). Reserved. Lane Identification for Link Lane 6. Checksum for Link Lane 6. Computed Checksum for Link Lane 6 (see the description for Register 0x40E). Reserved. Lane Identification for Link Lane 7. Checksum for Link Lane 7. Computed Checksum for Link Lane 7 (see the description for Register 0x40E). Device Identification Number. Link information received on Link Lane 0 as specified in Section 8.3 of JESD204B. Must be set to value read in Register 0x400. Adjustment Resolution to DAC LMFC Must be set to 0. Bank Identification: Extension to DID Must be set to value read in Register 0x401[3:0]. Reserved. Direction to Adjust DAC LMFC. Must be set to 0. Phase Adjustment Request to DAC. Must be set to 0. Lane Identification for Link Lane 0. Must be set to the value read in Register 0x402[4:0]. Receiver Descrambling Enable. Descrambling is disabled Descrambling is enabled Reserved. Number of Lanes per Converter Device. See Table 34 and Table 35. One lane per converter Two lanes per converter Four lanes per converter Eight lanes per converter (single link only) Number of Octets per Lane per Frame. Settings of 1, 2, and 4 (octets per lane) per frame are valid. See Table 34 and Table 35. (One octet per lane) per frame (Two octets per lane) per frame (Four octets per lane) per frame Reserved. Number of Frames per Multiframe. Settings of 16 or 32 are valid. Must be set to 32 when F = 1 (Register 0x476). 16 frames per multiframe 32 frames per multiframe Rev. A | Page 110 of 117 Reset 0x0 0x0 0x0 0x0 Access R R R R 0x0 0x0 0x0 0x0 R R R R 0x0 0x0 0x0 0x0 R R R R 0x0 0x0 0x0 0x0 R R R R 0x0 R/W 0x0 R/W 0x0 R/W 0x0 0x0 0x0 R R/W R/W 0x0 R/W 0x1 R/W 0x0 0x3 R R/W 0x0 R/W 0x0 0x1F R R/W Data Sheet Address 0x456 Name ILS_M AD9135/AD9136 Bit No. [7:0] Bit Name M-1 Settings 0 1 0x457 ILS_CS_N [7:6] CS 5 [4:0] RESERVED N-1 [7:5] SUBCLASSV 0 0x0F 0x458 ILS_NP 0 1 [4:0] NP-1 0xF 0x459 ILS_S [7:5] JESDV 000 001 [4:0] S-1 0 1 0x45A ILS_HD_CF 7 HD 0 1 [6:5] [4:0] RESERVED CF 0x45B 0x45C 0x45D ILS_RES1 ILS_RES2 ILS_CHECKSUM [7:0] [7:0] [7:0] RES1 RES2 FCHK0 0x46B ERRCNTRMON_RB [7:0] READERRORCNTR 0x46B ERRCNTRMON 7 [6:4] RESERVED LANESEL [3:2] [1:0] RESERVED CNTRSEL x 00 01 10 0x46C 0x46D LANEDESKEW BADDISPARITY_RB [7:0] [7:0] LANEDESKEW BADDIS Description Number of Converters per Device. See Table 34 and Table 35. One converter per link Two converters per link Number of Control Bits per Sample. Must be set to 0. Control bits are not supported. Zero control bits per sample Reserved. Converter Resolution. Must be set to 16 bits of resolution. Converter resolution of 16. Device Subclass Version. Subclass 0 Subclass 1 Total Number of Bits per Sample. Must be set to 16 bits per sample. 16 bits per sample. JESD204 Version. JESD204A JESD204B Number of Samples per Converter per Frame Cycle. Settings of one and two are valid. One sample per converter per frame Two samples per converter per frame High Density Format. If F = 1, HD must be set to 1. Otherwise, HD must be set to 0. See Section 5.1.3 of JESD204B standard. Low density mode High density mode Reserved. Number of Control Words per Frame Clock Period per Link. Must be set to 0. Control bits are not supported. Reserved Field 1. Reserved Field 2. Checksum for Link Lane 0. Calculated checksum. Calculation depends on 0x300[6]. Read JESD204B Error Counter. After selecting the lane and error counter by writing to LANESEL and CNTRSEL (both in this same register), the selected error counter is read back here. Reserved. Link Lane select for JESD204B error counter. Selects the lane whose errors are read back in this register. Selects Link Lane x Reserved. JESD204B Error Counter Select. Selects the type of error that are read back in this register. BADDISCNTR: bad running disparity counter NITCNTR: not in table error counter UCCCNTR: Unexpected control character counter Lane Deskew. Setting Bit x deskews Link Lane x Bad Disparity Character Error (BADDIS). Bit x is set when the bad disparity error count for Link Lane x reaches the threshold in Register 0x47C. Rev. A | Page 111 of 117 Reset 0x1 Access R/W 0x0 R/W 0x0 0xF R R/W 0x1 R/W 0xF R/W 0x1 R/W 0x0 R/W 0x1 R/W 0x0 0x0 R R/W 0x0 0x0 0x45 R/W R/W R/W 0x0 R 0x0 0x0 R W 0x0 0x0 R W 0xF 0x0 R/W R AD9135/AD9136 Address 0x46D Name BADDISPARITY Data Sheet Bit No. 7 Bit Name RST_IRQ_DIS Settings 6 DISABLE_ERR_ CNTR_DIS 5 RST_ERR_CNTR_DIS [4:3] [2:0] RESERVED LANE_ADDR_DIS 0x46E NIT_RB [7:0] NIT 0x46E NIT_W 7 RST_IRQ_NIT 6 DISABLE_ERR_ CNTR_NIT 5 RST_ERR_CNTR_NIT [4:3] [2:0] RESERVED LANE_ADDR_NIT 0x46F UNEXPECTEDCONTROL_RB [7:0] UCC 0x46F UNEXPECTEDCONTROL_W 7 RST_IRQ_UCC 6 DISABLE_ERR_ CNTR_UCC 5 [4:3] [2:0] RST_ERR_CNTR_ UCC RESERVED LANE_ADDR_UCC [7:0] CODEGRPSYNC 0x470 CODEGRPSYNCFLG 0 1 0x471 FRAMESYNCFLG [7:0] FRAMESYNC 0 1 0x472 GOODCHKSUMFLG [7:0] GOODCHECKSUM 0 1 Description BADDIS IRQ Reset. Reset BADDIS IRQ for lane selected via Bits[2:0] by writing 1 to this bit. BADDIS Error Counter Disable. Disable the BADDIS error counter for lane selected via Bits[2:0] by writing 1 to this bit. BADDIS Error Counter Reset. Reset BADDIS error counter for lane selected via Bits[2:0] by writing 1 to this bit. Reserved. Link Lane Address for Functions Described in Bits[7:5]. Not in table Character Error (NIT). Bit x is set when Link Lane x’s NIT error count reaches the threshold in Register 0x47C. IRQ Reset. Reset IRQ for lane selected via Bits[2:0] by writing 1 to this bit. Disable Error Counter. Disable the error counter for lane selected via Bits[2:0] by writing 1 to this bit. Reset Error Counter. Reset error counter for lane selected via Bits[2:0] by writing 1 to this bit. Reserved. Link Lane Address for Functions Described in Bits[7:5]. Unexpected Control Character Error (UCC). Bit x is set when Link Lane x’s UCC error count reaches the threshold in Register 0x47C. IRQ Reset. Reset IRQ for lane selected via Bits[2:0] by writing 1 to this bit. Disable Error Counter. Disable the error counter for lane selected via Bits[2:0] by writing 1 to this bit. Reset Error Counter. Reset error counter for lane selected via Bits[2:0] by writing 1 to this bit. Reserved. Link Lane Address for Functions Described in Bits[7:5]. Code Group Sync Flag (from Each Instantiated Lane). Writing 1 to Bit 7 resets the IRQ. The associated IRQ flag is located in Register 0x47A[0]. A loss of CODEGRPSYNC triggers sync request assertion. See the SYNCOUTx±, SYSREF±, and CLK± Signals section and the Deterministic Latency section. Synchronization is lost Synchronization is achieved Frame Sync Flag (from Each Instantiated Lane). This register indicates the live status for each lane. Writing 1 to Bit 7 resets the IRQ. A loss of frame sync automatically initiates a synchronization sequence. Synchronization is lost Synchronization is achieved Good Checksum Flag (from Each Instantiated Lane). Writing 1 to Bit 7 resets the IRQ. The associated IRQ flag is located in Register 0x47A[2]. Last computed checksum is not correct Last computed checksum is correct Rev. A | Page 112 of 117 Reset 0x0 Access W 0x0 W 0x0 W 0x0 0x0 R W 0x0 R 0x0 W 0x0 W 0x0 W 0x0 0x0 R W 0x0 R 0x0 W 0x0 W 0x0 W 0x0 0x0 R W 0x0 R/W 0x0 R/W 0x0 R/W Data Sheet AD9135/AD9136 Address 0x473 Name INITLANESYNCFLG Bit No. [7:0] Bit Name INITIALLANESYNC 0x476 CTRLREG1 [7:0] F Settings 1 2 4 0x477 CTRLREG2 7 ILAS_MODE 1 0 [6:4] 3 0x478 KVAL [2:0] [7:0] RESERVED THRESHOLD_ MASK_EN RESERVED KSYNC 0x47A IRQVECTOR_MASK 7 BADDIS_MASK x 1 6 NIT_MASK 1 5 UCC_MASK 1 4 3 2 1 0 0x47A IRQVECTOR_FLAG 7 RESERVED INITIALLANESYNC_ MASK 1 BADCHECKSUM_ MASK 1 FRAMESYNC_ MASK 1 CODEGRPSYNC_ MASK 1 BADDIS_FLAG 1 6 NIT_FLAG 1 Description Initial Lane Sync Flag (from Each Instantiated Lane). Writing 1 to Bit 7 resets the IRQ. The associated IRQ flag is located in Register 0x47A[3]. Loss of synchronization is also reported on SYNCOUT1± or SYNCOUT0±. See the SYNCOUTx±, SYSREF±, and CLK± Signals section and the Deterministic Latency section. Number of Octets per Frame. Settings of 1, 2, and 4 are valid. See Table 34 and Table 35. One octet per frame Two octets per frame Four octets per frame ILAS Test Mode. Defined in Section 5.3.3.8 of JESD204B specification. JESD204B receiver is constantly receiving ILAS frames Normal link operation Reserved. Threshold Mask Enable. Set this bit if using SYNC_ASSERTION_MASK (Register 0x47B[7:5]). Reserved. Number of K Multiframes During ILAS (Divided by Four). Sets the number of multiframes to send initial lane alignment sequence. Cannot be set to 0. 4× multiframes during ILAS Bad Disparity Mask. If the bad disparity count reaches ERRORTHRESH on any lane, IRQ is pulled low. Not in table Mask. If the not in table character count reaches ERRORTHRESH on any lane, IRQ is pulled low. Unexpected Control Character Mask. If the unexpected control character count reaches ERRORTHRESH on any lane, IRQ is pulled low. Reserved. Initial Lane Sync Mask. If initial lane sync (0x473) fails on any lane, IRQ is pulled low. Bad Checksum Mask. If there is a bad checksum (0x472) on any lane, IRQ is pulled low. Frame Sync Mask If frame sync (0x471) fails on any lane, IRQ is pulled low. Code Group Sync Machine Mask. If code group sync (0x470) fails on any lane, IRQ is pulled low. Bad Disparity Error Count. Bad disparity character count reached ERRORTHRESH (0x47C) on at least one lane. Read Register 0x46D to determine which lanes are in error. Not in table Error Count Not in table character count reached ERRORTHRESH (0x47C) on at least one lane. Read Register 0x46E to determine which lanes are in error. Rev. A | Page 113 of 117 Reset 0x0 Access R/W 0x1 R/W 0x0 R/W 0x0 0x0 R R/W 0x0 0x1 R R/W 0x0 W 0x0 W 0x0 W 0x0 0x0 R W 0x0 W 0x0 W 0x0 W 0x0 R 0x0 R AD9135/AD9136 Address Name Data Sheet Bit No. 5 Bit Name UCC_FLAG Settings 1 4 3 2 1 0 0x47B SYNCASSERTIONMASK 7 RESERVED INITIALLANESYNC_ FLAG 1 BADCHECKSUM_ FLAG 1 FRAMESYNC_ FLAG 1 CODEGRPSYNC_ FLAG 1 BADDIS_S 1 6 NIT_S 1 5 UCC_S 1 4 CMM 1 3 CMM_ENABLE 1 0 0x47C ERRORTHRES [2:0] [7:0] 0x47D LANEENABLE [7:0] RESERVED ETH LANE_ENA Description Unexpected Control Character Error Count Unexpected control character count reached ERRORTHRESH (0x47C) on at least one lane. Read Register 0x46F to determine which lanes are in error. Reserved. Initial Lane Sync Flag. Initial lane sync failed on at least one lane. Read Register 0x473 to determine which lanes are in error Bad Checksum Flag. Bad checksum on at least one lane. Read Register 0x472 to determine which lanes are in error. Frame Sync Flag. Frame sync failed on at least one lane. Read Register 0x471 to determine which lanes are in error. Code Group Sync Flag. Code group sync failed on at least one lane. Read Register 0x470 to determine which lanes are in error Bad Disparity Error on Sync. Asserts a sync request on SYNCOUTx± when the bad disparity character count reaches the threshold in Register 0x47C Not in table Error on Sync. Asserts a sync request on SYNCOUTx± when the not in table character count reaches the threshold in Register 0x47C Unexpected Control Character Error on Sync. Asserts a sync request on SYNCOUTx± when the unexpected control character count reaches the threshold in Register 0x47C Configuration Mismatch IRQ. If CMM_ENABLE is high, this bit latches on a rising edge and pull IRQ low. When latched, write a 1 to clear this bit. If CMM_ENABLE is low, this bit is non-functional. Link Lane 0 configuration registers (Register 0x450 to Register 0x45D) do not match the JESD204B transmit settings (Register 0x400 to Register 0x40D) Configuration Mismatch IRQ Enable. Enables IRQ generation if a configuration mismatch is detected Configuration mismatch IRQ disabled Reserved. Error Threshold. Bad disparity, not in table, and unexpected control character errors are counted and compared to the error threshold value. When the count reaches the threshold, either an IRQ is generated or the SYNCOUTx± signal is asserted per the mask register settings, or both. Function is performed in all lanes. Lane Enable. Setting Bit x enables Link Lane x. This register must be programmed before receiving the code group pattern for proper operation. Rev. A | Page 114 of 117 Reset 0x0 Access R 0x0 0x0 R R 0x0 R 0x0 R 0x0 R 0x0 R/W 0x0 R/W 0x0 R/W 0x0 R/W 0x1 R/W 0x0 0xFF R R/W 0xF R/W Data Sheet Address 0x47E Name RAMP_ENA AD9135/AD9136 Bit No. [7:1] 0 Bit Name RESERVED ENA_RAMP_ CHECK Settings 0 1 0x520 DIG_TEST0 [7:2] RESERVED 0x521 DC_TEST_VALUE0 1 0 [7:0] 0x522 DC_TEST_VALUE1 [7:0] DC_TEST_MODE RESERVED DC_TEST_ VALUE[7:0] DC_TEST_ VALUE[15:8] Description Reserved. Enable Ramp Checking at the Beginning of ILAS. Disable ramp checking at beginning of ILAS; ILAS data need not be a ramp Enable ramp checking; ILAS data needs to be a ramp starting at 00-01-02; otherwise, the ramp ILAS fails and the device does not start up Must write default value for proper operation. DC Test Mode Reserved. DC Value LSB of DC Test Mode for DAC0 and DAC1. DC value MSB of DC Test Mode for DAC0 and DAC1. Rev. A | Page 115 of 117 Reset 0x0 0x0 Access R W 0x7 R/W 0x0 0x0 0x0 R/W R/W R/W 0x0 R/W AD9135/AD9136 Data Sheet OUTLINE DIMENSIONS 12.10 12.00 SQ 11.90 0.28 0.23 0.18 0.60 MAX 0.60 MAX 67 66 88 PIN 1 INDICATOR 1 PIN 1 INDICATOR 11.85 11.75 SQ 11.65 0.50 BSC 0.50 0.40 0.30 22 23 45 44 TOP VIEW BOTTOM VIEW 10.50 REF 0.70 0.65 0.60 FOR PROPER CONNECTION OF THE EXPOSED PAD, REFER TO THE PIN CONFIGURATION AND FUNCTION DESCRIPTIONS SECTION OF THIS DATA SHEET. 0.045 0.025 0.005 COPLANARITY 0.08 0.20 REF SEATING PLANE 08-10-2012-A 12° MAX 0.90 0.85 0.80 7.55 7.40 SQ 7.25 EXPOSED PAD COMPLIANT TO JEDEC STANDARDS MO-220-VRRD Figure 91. 88-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 12 mm × 12 mm Body, Very Thin Quad (CP-88-6) Dimensions shown in millimeters 12.10 12.00 SQ 11.90 0.30 0.25 0.20 0.60 MAX 0.60 MAX 67 88 66 1 PIN 1 INDICATOR PIN 1 INDICATOR 11.85 11.75 SQ 11.65 0.50 BSC 7.55 7.40 SQ 5.25 EXPOSED PAD 0.65 0.55 0.45 22 44 0.90 0.85 0.80 PKG-004598 SEATING PLANE 0.70 0.65 0.60 12° MAX 0.190~0.245 REF 0.50 0.40 0.30 45 23 BOTTOM VIEW 10.50 REF 0.045 0.025 0.005 COPLANARITY 0.08 COMPLIANT TO JEDEC STANDARDS MO-220 Figure 92. 88-Lead Lead Frame Chip Scale Package [LFCSP_VQ] (Variable Lead Length) 12 mm × 12 mm Body, Very Thin Quad (CP-88-9) Dimensions shown in millimeters Rev. A | Page 116 of 117 08-04-2014-A TOP VIEW 1.00 0.90 0.80 0.80 0.70 0.60 Data Sheet AD9135/AD9136 ORDERING GUIDE Model1 AD9135BCPZ AD9135BCPZRL AD9135BCPAZ AD9135BCPAZRL AD9136BCPZ AD9136BCPZRL AD9136BCPAZ AD9136BCPAZRL AD9136-EBZ AD9136-FMC-EBZ AD9135-EBZ AD9135-FMC-EBZ 1 Temperature Range −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C Package Description 88-Lead LFCSP_VQ 88-Lead LFCSP_VQ 88-Lead LFCSP_VQ (Variable Lead Length) 88-Lead LFCSP_VQ (Variable Lead Length) 88-Lead LFCSP_VQ 88-Lead LFCSP_VQ 88-Lead LFCSP_VQ (Variable Lead Length) 88-Lead LFCSP_VQ (Variable Lead Length) DPG3 Evaluation Board FMC Evaluation Board DPG3 Evaluation Board FMC Evaluation Board Z = RoHS Compliant Part. ©2014–2015 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D12578-0-7/15(A) Rev. A | Page 117 of 117 Package Option CP-88-6 CP-88-6 CP-88-9 CP-88-9 CP-88-6 CP-88-6 CP-88-9 CP-88-9