14-Bit, 2400 MSPS RF DAC with 4-Channel Signal Processing AD9789 The on-chip rate converter supports a wide range of baud rates with a fixed DAC clock. The digital upconverter can place the channels from 0 to 0.5 × fDAC. This permits four contiguous channels to be synthesized and placed anywhere from dc to fDAC/2. FEATURES DOCSIS 3.0 performance: 4 QAM carriers ACLR over full band (47 MHz to 1 GHz) −75 dBc @ fOUT = 200 MHz −72 dBc @ fOUT = 800 MHz (noise) −67 dBc @ fOUT = 800 MHz (harmonics) Unequalized MER = 42 dB On chip and bypassable 4 QAM encoders with SRRC filters, 16× to 512× interpolation, rate converters, and modulators Flexible data interface: 4, 8, 16, or 32 bits wide with parity Power: 1.6 W (IFS = 20 mA, fDAC = 2.4 GHz, LVDS interface) Direct to RF synthesis support with fS mix mode Built-in self-test (BIST) support Input connectivity check Internal random number generator The AD9789 includes a serial peripheral interface (SPI) for device configuration and status register readback. The flexible digital interface can be configured for data bus widths of 4, 8, 16, and 32 bits. It can accept real or complex data. The AD9789 operates from 1.5 V, 1.8 V, and 3.3 V supplies for a total power consumption of 1.6 W. It is supplied in a 164-ball chip scale package ball grid array for lower thermal impedance and reduced package parasitics. No special power sequencing is required. The clock receiver powers up muted to prevent start-up noise. PRODUCT HIGHLIGHTS APPLICATIONS 1. Broadband communications systems CMTS/DVB Cellular infrastructure Point-to-point wireless 2. 3. GENERAL DESCRIPTION 4. The AD9789 is a flexible QAM encoder/interpolator/upconverter combined with a high performance, 2400 MSPS, 14-bit RF digitalto-analog converter (DAC). The flexible digital interface can accept up to four channels of complex data. The QAM encoder supports constellation sizes of 16, 32, 64, 128, and 256 with SRRC filter coefficients for all standards. 5. Highly integrated and configurable QAM mappers, interpolators, and upconverters for direct synthesis of one to four DOCSIS- or DVB-C-compatible channels in a block. Low noise and intermodulation distortion (IMD) performance enable high quality synthesis of signals up to 1 GHz. Flexible data interface supports LVDS for improved SFDR or CMOS input data for less demanding applications. Interface is configurable from 4-bit nibbles to 32-bit words and can run at up to 150 MHz CMOS or 150 MHz LVDS double data rate (DDR). Manufactured on a CMOS process, the AD9789 uses a proprietary switching technique that enhances dynamic performance. FUNCTIONAL BLOCK DIAGRAM DCO FS RETIMER DATA FORMATTER/ ASSEMBLER 150MHz LVDS/CMOS CMOS 16 TO 31 LVDS FALL DATA QAM/ FILTER/ NCO DATA QAM/ FILTER/ NCO DATA QAM/ FILTER/ NCO DATA QAM/ FILTER/ NCO 16× INTERPOLATOR AND BPF + SCALARS 14-BIT 2.4GSPS DAC SPI IRQ RS 07852-001 32 INPUT PINS AND 2 PARITY PINS CMOS 0 TO 15 LVDS RISE Figure 1. Rev. A Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.461.3113 ©2009-2011 Analog Devices, Inc. All rights reserved. AD9789 TABLE OF CONTENTS Features .............................................................................................. 1 SPI Register Descriptions.......................................................... 29 Applications....................................................................................... 1 Theory of Operation ...................................................................... 39 General Description ......................................................................... 1 Datapath Signal Processing....................................................... 39 Product Highlights ........................................................................... 1 Digital Block Upconverter ........................................................ 43 Functional Block Diagram .............................................................. 1 Digital Interface Modes ............................................................. 45 Revision History ............................................................................... 2 Analog Modes of Operation ..................................................... 54 Detailed Functional Block Diagrams............................................. 3 Analog Control Registers .......................................................... 55 Specifications..................................................................................... 4 Voltage Reference ....................................................................... 56 DC Specifications ......................................................................... 4 DAC Output Stages .................................................................... 56 Digital Specifications ................................................................... 5 Clocking the AD9789 ................................................................ 57 AC Specifications.......................................................................... 6 Mu Delay Controller.................................................................. 58 Absolute Maximum Ratings............................................................ 8 Interrupt Requests...................................................................... 61 Thermal Resistance ...................................................................... 8 Recommended Start-Up Sequence .......................................... 62 ESD Caution.................................................................................. 8 Customer BIST Modes................................................................... 63 Pin Configurations and Function Descriptions ........................... 9 Using the Internal PRN Generator to Test QAM Output AC Performance................................................................................ 63 Typical Performance Characteristics ........................................... 12 Terminology .................................................................................... 22 Serial Control Port.......................................................................... 23 Serial Control Port Pin Descriptions ....................................... 23 General Operation of Serial Control Port............................... 23 Instruction Word (16 Bits) ........................................................ 24 MSB/LSB First Transfers............................................................ 24 Using the Internal Built-In Self-Test (BIST) to Test for Digital Data Input Connectivity............................................................. 63 QAM Constellation Maps ............................................................. 65 Channelizer Mode Pin Mapping for CMOS and LVDS............ 68 Outline Dimensions ....................................................................... 74 Ordering Guide .......................................................................... 74 SPI Register Map............................................................................. 27 REVISION HISTORY 7/11—Rev. 0 to Rev. A Changes to Table 2, DAC Clock Input (CLKP, CLKN): Added DAC Clock Rate Parameter............................................................. 5 Changes to Table 3, Dynamic Performance, DAC Update Rate Parameter; Added Adjusted DAC Update Rate Parameter......... 6 Changes to Captions for Figure 42, Figure 44, Figure 46, Figure 49 .......................................................................................... 18 Changes to Digital 16x Tunable Band-Pass Filter Section, Third Paragraph......................................................................................... 44 Changes to Retimer and Latency Look-Up Tables Section, Second Paragraph ........................................................................... 50 Changes to Captions for Figure 122, Figure 124, Figure 125.... 65 4/09—Revision 0: Initial Version Rev. A | Page 2 of 76 AD9789 DETAILED FUNCTIONAL BLOCK DIAGRAMS 32 INPUT PINS 4 TO RETIMER 32 BITS LVDS/CMOS CMOS 16 TO 31 LVDS FALL DATA FORMATTER/ ASSEMBLER DATAPATH 0 UP TO 32 BITS DATAPATH 1 UP TO 32 BITS P0 P1 FS DCO UP TO 32 BITS UP TO 32 BITS CLK CTL DATAPATH 2 16× INTERPOLATOR BPF fC = 0 TO fDAC/2 SUM SCALE BPF DATAPATH 3 fC Figure 2. Digital Signal Processing Functional Block Diagram 24-BIT NCO 0 TO fDAC /16 QAM MAPPER SRRC 2N (N = 0 TO 5) INPUT SCALE BYPASS QAM BYPASS SRRC RATE CONVERTER P/Q 24-BIT (P/Q = 0.5 TO 1) CH GAIN 0× TO 2× Figure 3. Channel 0 Through Channel 3 Datapath Block Detail (I and Q Paths Are Identical So Only One Is Shown) Rev. A | Page 3 of 76 07852-003 2 07852-002 CMOS 0 TO 15 LVDS RISE AD9789 SPECIFICATIONS DC SPECIFICATIONS AVDD33 = DVDD33 = 3.3 V, CVDD18 = DVDD18 = 1.8 V, DVDD15 = 1.5 V, fDAC = 2.4 GHz, IFS = 20 mA, unless otherwise noted. Table 1. Parameter DAC RESOLUTION ANALOG OUTPUTS Offset Error Gain Error (with Internal Reference) Full-Scale Output Current (Monotonicity Guaranteed) Output Compliance Range Output Resistance Output Capacitance TEMPERATURE DRIFT Gain Reference Voltage REFERENCE Internal Reference Voltage Output Resistance 1 ANALOG SUPPLY VOLTAGES AVDD33 CVDD18 DIGITAL SUPPLY VOLTAGES DVDD33 DVDD18 DVDD15 SUPPLY CURRENTS AND POWER DISSIPATION fDAC = 2.4 GSPS, fOUT = 930 MHz, IFS = 25 mA, Four Channels Enabled IAVDD33 IDVDD18 ICVDD18 IDVDD33 CMOS Interface LVDS Interface IDVDD15 fDAC = 2.0 GSPS, fOUT = 70 MHz, IFS = 20 mA, CMOS Interface IAVDD33 IDVDD18 ICVDD18 IDVDD33 IDVDD15 (Four Channels Enabled, All Signal Processing Enabled) IDVDD15 (One Channel Enabled, 16× Interpolation Only) Power Dissipation fDAC = 2.4 GSPS, fOUT = 930 MHz, IFS = 25 mA, Four Channels Enabled CMOS Interface LVDS Interface 1 Min 8.66 −1.0 Typ 14 6.5 3.5 20.2 Max Unit Bits 70 1 % FSR % FSR mA V Ω pF 135 25 ppm/°C ppm/°C 1.2 5 V kΩ 31.66 +1.0 3.14 1.71 3.3 1.8 3.47 1.89 V V 3.14 1.71 1.43 3.3 1.8 1.5 3.47 1.89 1.58 V V V Use an external amplifier to drive any external load. Rev. A | Page 4 of 76 45 72 180 mA mA mA 42 16 640 mA mA mA 37.4 67.3 155.4 40.3 517 365 1.7 1.63 38.5 70.5 180 50.7 556 391 mA mA mA mA mA mA W W AD9789 DIGITAL SPECIFICATIONS AVDD33 = DVDD33 = 3.3 V, CVDD18 = DVDD18 = 1.8 V, DVDD15 = 1.5 V, fDAC = 2.4 GHz, IFS = 20 mA, LVDS drivers and receivers are compliant with the IEEE Std 1596.3-1996 reduced range link, unless otherwise noted. Table 2. Parameter CMOS DATA INPUTS (D[31:0], P0, P1) Input Voltage High, VIH Input Voltage Low, VIL Input Current High, IIH Input Current Low, IIL Input Capacitance Setup Time, CMOS Data Input to CMOS_DCO 1 Hold Time, CMOS Data Input to CMOS_DCO1 CMOS OUTPUTS (CMOS_FS, CMOS_DCO) Output Voltage High, VOH Output Voltage Low, VOL Output Current High, IOH Output Current Low, IOL Maximum Clock Rate (CMOS_DCO) CMOS_DCO to CMOS_FS Delay LVDS DATA INPUTS (D[15:0]P, D[15:0]N, PARP, PARN) Input Voltage Range, VIA or VIB Input Differential Threshold, VIDTH Input Differential Hysteresis, VIDTHH, VIDTHL Input Differential Input Impedance, RIN Maximum LVDS Input Rate Setup Time, LVDS Differential Input Data to Differential DCOx 2 Hold Time, LVDS Differential Input Data to Differential DCOx2 LVDS OUTPUTS (DCOP, DCON, FSP, FSN) DCOP, FSP = VOA; DCON, FSN = VOB; 100 Ω Termination Output Voltage High, VOA or VOB Output Voltage Low, VOA or VOB Output Differential Voltage, |VOD| Output Offset Voltage, VOS Output Impedance, Single Ended, RO RO Mismatch Between A and B, ∆RO Change in |VOD| Between 0 and 1, |∆VOD| Change in VOS Between 0 and 1, ∆VOS Output Current—Driver Shorted to Ground, ISA, ISB Output Current—Drivers Shorted Together, ISAB Power-Off Output Leakage, |IXA|, |IXB| Maximum Clock Rate (DCOP, DCON) DCOx to FSx Delay DAC CLOCK INPUT (CLKP, CLKN) 3 Differential Peak Voltage Common-Mode Voltage DAC Clock Rate SERIAL PERIPHERAL INTERFACE Maximum Clock Rate (fSCLK, 1/tSCLK) Minimum Pulse Width High, tPWH Minimum Pulse Width Low, tPWL Minimum SDIO and CS to SCLK Setup, tDS Min Typ 2.0 3.3 0 −10 −10 Max 0.8 +10 +10 2 5.3 −1.4 2.4 0 3.3 0.4 12 12 150 0.28 0.85 825 −100 1575 +100 25 80 150 1.41 0.24 120 1375 1025 150 1150 40 200 150 0.12 1.4 Rev. A | Page 5 of 76 V V mA mA MHz ns mV mV mV Ω MSPS ns ns 0.37 2400 V mV MHz 250 1250 140 10 25 25 20 4 10 25 10 V V μA μA pF ns ns mV mV mV mV Ω % mV mV mA mA mA MHz ns 1.8 900 20 20 Unit MHz ns ns ns AD9789 Parameter Minimum SCLK to SDIO Hold, tDH Maximum SCLK to Valid SDIO and SDO, tDV Minimum SCLK to Invalid SDIO and SDO, tDNV INPUTS (SDIO, SCLK, CS) Input Voltage High, VIH Input Voltage Low, VIL Input Current High, IIH Input Current Low, IIL OUTPUTS (SDO, SDIO) Output Voltage High, VOH Output Voltage Low, VOL Output Current High, IOH Output Current Low, IOL 1 2 3 Min Typ 5 20 5 2.0 3.3 0 Max −10 −10 0.8 +10 +10 2.4 0 3.6 0.4 Unit ns ns ns V V μA μA V V mA mA 4 4 See the CMOS Interface Timing section for more information. See the LVDS Interface Timing section for more information. See the Clock Phase Noise Effects on AC Performance section for more information. AC SPECIFICATIONS AVDD33 = DVDD33 = 3.3 V, CVDD18 = DVDD18 = 1.8 V, DVDD15 = 1.5 V, fDAC = 2.4 GHz, IFS = 20 mA, digital scale = 0 dBFS, unless otherwise noted. Table 3. Parameter DYNAMIC PERFORMANCE DAC Update Rate Adjusted DAC Update Rate 1 Output Settling Time (tST) SPURIOUS-FREE DYNAMIC RANGE (SFDR) fDAC = 2000 MSPS fOUT = 100 MHz fOUT = 316 MHz fOUT = 550 MHz fDAC = 2400 MSPS fOUT = 100 MHz fOUT = 316 MHz fOUT = 550 MHz fOUT = 850 MHz TWO-TONE INTERMODULATION DISTORTION (IMD) fDAC = 2000 MSPS fOUT = 100 MHz fOUT = 316 MHz fOUT = 550 MHz fDAC = 2400 MSPS fOUT = 100 MHz fOUT = 316 MHz fOUT = 550 MHz fOUT = 850 MHz NOISE SPECTRAL DENSITY (NSD) 1-Channel QAM fOUT = 100 MHz fOUT = 316 MHz fOUT = 550 MHz fOUT = 850 MHz Test Conditions/Comments To 0.025% Min Typ Max Unit 2400 150 13 MSPS MSPS ns 70 63 58 dBc dBc dBc 70 70 60 60 dBc dBc dBc dBc 86 73 62 dBc dBc dBc 86 74 66 66 dBc dBc dBc dBc −167 −166.5 −166.5 −166.5 dBm/Hz dBm/Hz dBm/Hz dBm/Hz fOUT2 = fOUT1 + 1.25 MHz fDAC = 2400 MSPS POUT = −14.5 dBm POUT = −15.5 dBm POUT = −18 dBm POUT = −18.5 dBm Rev. A | Page 6 of 76 AD9789 Parameter ADJACENT CHANNEL LEAKAGE RATIO (ACLR) 1-Channel QAM fOUT = 200 MHz (Harmonics) fOUT = 200 MHz (Noise Floor) fOUT = 500 MHz (Harmonics) fOUT = 500 MHz (Noise Floor) fOUT = 800 MHz (Harmonics) fOUT = 800 MHz (Noise Floor) 2-Channel QAM fOUT = 200 MHz (Harmonics) fOUT = 200 MHz (Noise Floor) fOUT = 500 MHz (Harmonics) fOUT = 500 MHz (Noise Floor) fOUT = 800 MHz (Harmonics) fOUT = 800 MHz (Noise Floor) 4-Channel QAM fOUT = 200 MHz (Harmonics) fOUT = 200 MHz (Noise Floor) fOUT = 500 MHz (Harmonics) fOUT = 500 MHz (Noise Floor) fOUT = 800 MHz (Harmonics) fOUT = 800 MHz (Noise Floor) WCDMA ACLR Single Carrier First Adjacent Channel Second Alternate Channel Third Alternate Channel Single Carrier First Adjacent Channel Second Alternate Channel Third Alternate Channel Four Carrier First Adjacent Channel Second Alternate Channel Third Alternate Channel 1 Test Conditions/Comments fDAC = 2293.76 MSPS measured in 6 MHz channels Min Typ Max Unit −76 −82 −74.5 −78 −69 −78 dBc dBc dBc dBc dBc dBc −77.5 −81 −68 −76 −66 −76 dBc dBc dBc dBc dBc dBc −75 −76 −69 −72 −67 −72 dBc dBc dBc dBc dBc dBc −70 −72.5 −74 dBc dBc dBc −68 −70.4 −72.7 dBc dBc dBc −63.5 −65.1 −66.9 dBc dBc dBc fDAC = 2304 MSPS, mix mode second Nyquist zone fOUT = 1850 MHz fOUT = 2100 MHz fOUT = 2100 MHz Adjusted DAC update rate is calculated as fDAC divided by the minimum required interpolation factor. For the AD9789, the minimum interpolation factor is 16. Thus, with fDAC = 2400 MSPS, FDACadj = 2400 MSPS/16 = 150 MSPS. Rev. A | Page 7 of 76 AD9789 ABSOLUTE MAXIMUM RATINGS THERMAL RESISTANCE Table 4. Parameter AVDD33 to AVSS DVDD18 to DVSS DVDD33 to DVSS DVDD15 to DVSS CVDD18 to AVSS AVSS to DVSS CLKP, CLKN to AVSS FS, DCO to DVSS CMOS and LVDS Data Inputs to DVSS IOUTN, IOUTP to AVSS I120, VREF, IPTAT to AVSS IRQ, CS, SCLK, SDO, SDIO, RESET to DVSS Junction Temperature Storage Temperature Range Rating −0.3 V to +3.6 V −0.3 V to +1.98 V −0.3 V to +3.6 V −0.3 V to +1.98 V −0.3 V to +1.98 V −0.3 V to +0.3 V −0.3 V to CVDD18 + 0.3 V −0.3 V to DVDD33 + 0.3 V −0.3 V to DVDD33 + 0.3 V θJA is specified for the worst-case conditions, that is, a device soldered in a circuit board for surface-mount packages. −1.0 V to AVDD33 + 0.3 V −0.3 V to AVDD33 + 0.3 V −0.3 V to DVDD33 + 0.3 V ESD CAUTION Table 5. Thermal Resistance Package Type 164-Ball CSP_BGA 150°C −65°C to +150°C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. Rev. A | Page 8 of 76 θJA 25.5 24.4 19.0 17.2 θJB 14.4 θJC 6.8 Unit °C/W °C/W °C/W °C/W Notes 4-layer board, no vias 4-layer board, 4 PCB vias 8-layer board, 4 PCB vias 8-layer board, 16 PCB vias AD9789 PIN CONFIGURATIONS AND FUNCTION DESCRIPTIONS CVDD18 1 2 3 IOUTN 4 5 A + + B + 6 IOUTP 7 8 – + 9 AVDD33 10 11 12 13 14 1 2 3 4 + + + + X 5 6 7 8 + + 9 10 11 12 13 14 A X X NC B + X X I120 C CLKN C N + + X X VREF D CLKP D P + + X X IPTAT E E F F G DVDD18 G + + + + X X X X H H J J K K CS L X X X SB NC NC L SCLK M CK NC NC M SDO N DO R NC NC N SDIO P IO I NC NC AVSS DVDD18 RESET Figure 4. Clock and Analog Pins (Top View) 1 2 3 4 5 6 7 8 9 10 11 1 12 13 14 A A B C C D D E E F F G G H H J PARP J K PARN K P1 31 27 23 19 15 M P0 11 7 DVSS + DVDD15 X DVDD33 Figure 6. Digital Supply and SPI Pins (Top View) B L IRQ NC NO CONNECT 07852-006 X AVDD33 2 3 4 5 6 7 8 9 10 11 12 13 14 FSP 3 BU CMOS_BUS L P+ 15 13 11 9 P– 15 13 7 5 3 1 FS FSN 6 2 CT CMOS_CTRL M 11 9 7 5 3 1 FS DCOP 29 25 21 17 13 9 5 1 FS CMOS_FS N 14 12 10 8 6 4 2 0 DC DCON P 28 24 20 16 12 8 4 0 DC CMOS_DCO P 14 12 10 8 6 4 2 0 DC D[31:0] CMOS DATA INPUTS PARITY AND CONTROL INPUTS 07852-005 30 26 22 18 14 10 N Figure 5. CMOS Mode Data Input Pins (Top View) 14 +LVDS 14 –LVDS Figure 7. LVDS Mode Data Input Pins (Top View) Rev. A | Page 9 of 76 07852-007 + CVDD18 07852-004 P AD9789 Table 6. Pin Function Descriptions Pin No. A1, A2, A3, A6, A9, A10, A11, B1, B2, B3, B6, B7, B8, B9, B10, B11, C2, C3, C6, C7, C8, C9, C10, C11, D2, D3, D6, D7, D8, D9, D10, D11, E1, E2, E3, E4, E13, E14, F1, F2, F3, F4, F11, F12, F13, F14 A4, A5, B4, B5, C4, C5, D4, D5 A7 A8 A12, A13, B12, B13, C12, C13, D12, D13 A14 B14 C1 C14 Mnemonic AVSS Description Analog Supply Ground. CVDD18 IOUTN IOUTP AVDD33 1.8 V Clock Supply. DAC Negative Output Current. DAC Positive Output Current. 3.3 V Analog Supply. NC I120 CLKN VREF DVDD18 DVDD15 No Connect. Leave floating. Tie this pin to analog ground with a 10 kΩ resistor to generate a 120 μA reference current. Negative DAC Clock Input (DACCLK). Band Gap Voltage Reference I/O. Decouple to analog ground with a 1 nF capacitor. Output impedance is approximately 5 kΩ. Positive DAC Clock Input (DACCLK). Factory Test Pin. Output current, proportional to absolute temperature, is approximately 10 μA at 25°C with a slope of approximately 20 nA/°C. 1.8 V Digital Supply. 1.5 V Digital Supply. D1 D14 CLKP IPTAT E11, E12 G1, G2, G3, G4, G7, G8, G11, G12, G13, G14 H1, H2, H3, H4, H7, H8, H11, H12, H13, H14, J1, J2, J3, J4, J11, J12, J13, J14 K1, K2, K3, K4, K11, K12, K13, K14 L1 L2, L3, M2, M3, N3, N4, P3, P4 L4 L5 L6 L7 L8 L9 L10 L11 L12 L13 L14 DVSS Digital Supply Ground. DVDD33 3.3 V Digital Supply. CS NC P1/PARP D31/D15P D27/D13P D23/D11P D19/D9P D15/D7P D11/D5P D7/D3P D3/D1P FSP CMOS_BUS Active Low Chip Select for SPI. Not Used. Leave unconnected. CMOS/LVDS Parity Bit. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. Positive LVDS Frame Sync (FSP) for Data Bus. Active High Input. Configures data bus for CMOS inputs. Low input configures data bus to accept LVDS inputs. Qualifying Clock for SPI. CMOS/LVDS Parity Bit. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. Negative LVDS Frame Sync (FSN) for Data Bus. M1 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 SCLK P0/PARN D30/D15N D26/D13N D22/D11N D18/D9N D14/D7N D10/D5N D6/D3N D2/D1N FSN Rev. A | Page 10 of 76 AD9789 Pin No. M14 Mnemonic CMOS_CTRL N1 N2 N5 N6 N7 N8 N9 N10 N11 N12 N13 N14 P1 P2 SDO RESET D29/D14P D25/D12P D21/D10P D17/D8P D13/D6P D9/D4P D5/D2P D1/D0P DCOP CMOS_FS SDIO IRQ P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 D28/D14N D24/D12N D20/D10N D16/D8N D12/D6N D8/D4N D4/D2N D0/D0N DCON CMOS_DCO Description Active High Input. Enables CMOS_DCO and CMOS_FS signals and disables DCOP/DCON and FSP/FSN signals. Low input disables CMOS_DCO and CMOS_FS signals and enables DCOP/DCON and FSP/FSN signals. Serial Data Output for SPI. Active High Input. Resets the AD9789. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. Positive LVDS Data Clock Output (DCOP) for Data Bus. CMOS Frame Sync for Data Bus. Serial Data Input/Output for SPI. Active Low, Open-Drain Interrupt Request Output. Pull up to DVDD33 with a 10 kΩ resistor. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. CMOS/LVDS Data Input. Negative LVDS Data Clock Output (DCON) for Data Bus. CMOS Data Clock Output for Data Bus. Rev. A | Page 11 of 76 AD9789 –40 –45 –45 –50 –50 –55 –55 –60 –60 SFDR (dBc) –40 –65 –70 0 200 400 600 fOUT (MHz) 800 1000 –85 1200 –90 –45 –50 –50 HARMONIC LEVEL (dBc) –45 –55 –60 –65 –70 –75 0dBFS –3dBFS –6dBFS –12dBFS 0 200 400 600 fOUT (MHz) 800 1000 1200 1000 1200 –55 –60 –65 –70 –75 –90 0dBFS –3dBFS –6dBFS –12dBFS 0 200 400 600 800 1000 1200 fOUT (MHz) Figure 12. Third-Order Harmonic vs. fOUT over Digital Full Scale, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Temperature = 25°C –50 –45 –55 –50 –60 –55 –65 –60 –70 SFDR (dBc) –40 –65 –70 –75 –75 –80 –85 32mA 20mA 8mA –80 –85 0 200 400 600 800 1000 +85°C +25°C –40°C –90 –95 1200 fOUT (MHz) 07852-011 SFDR (dBc) 800 –85 Figure 9. Second-Order Harmonic vs. fOUT over Digital Full Scale, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Temperature = 25°C –90 600 –80 07852-009 HARMONIC LEVEL (dBc) –40 –90 400 Figure 11. SFDR vs. fOUT over Digital Full Scale, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Temperature = 25°C –40 –85 200 fOUT (MHz) Figure 8. SFDR vs. fOUT over fDAC, Full-Scale Current = 20 mA, Digital Scale = 0 dBFS, Temperature = 25°C –80 0 Figure 10. SFDR vs. fOUT over Full-Scale Current, fDAC = 2.4 GHz, Digital Scale = 0 dBFS, Temperature = 25°C –100 0 200 400 600 800 1000 1200 fOUT (MHz) Figure 13. SFDR vs. fOUT over Temperature, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Digital Scale = 0 dBFS Rev. A | Page 12 of 76 07852-008 –90 0dBFS –3dBFS –6dBFS –12dBFS –80 07852-010 –85 –70 –75 2.4GHz 2.2GHz 2GHz 1.6GHz 1GHz –80 –65 07852-012 –75 07852-013 SFDR (dBc) TYPICAL PERFORMANCE CHARACTERISTICS AD9789 90 100 80 90 80 70 IMD (dBc) 60 50 0 100 200 300 400 500 600 700 fOUT (MHz) 800 900 1000 1100 30 100 90 90 80 200 300 400 500 600 700 800 900 1000 1100 70 70 IMD (dBc) 60 60 50 50 32mA 20mA 8mA 0 100 200 300 400 500 600 700 fOUT (MHz) 800 +85°C +25°C –40°C 40 900 1000 1100 30 07852-038 40 Figure 15. Third-Order IMD vs. fOUT over Full-Scale Current, fDAC = 2.4 GHz, Digital Scale = 0 dBFS, Temperature = 25°C 0 –157 –159 –159 –161 –161 NSD (dBm/Hz) –157 –167 –169 400 500 600 700 800 900 1000 1100 –163 –165 –167 –169 2.4GHz 2.0GHz 1.6GHz –173 0 200 400 600 fOUT (MHz) 800 1000 +85°C +25°C –40°C –171 –173 1200 –175 07852-016 –171 –175 300 Figure 18. Third-Order IMD vs. fOUT over Temperature, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Digital Scale = 0 dBFS –155 –165 200 fOUT (MHz) –155 –163 100 07852-041 IMD (dBc) 100 Figure 17. Third-Order IMD vs. fOUT over Digital Full Scale, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Temperature = 25°C 80 NSD (dBm/Hz) 0 fOUT (MHz) Figure 14. Third-Order IMD vs. fOUT over fDAC, Full-Scale Current = 20 mA, Digital Scale = 0 dBFS, Temperature = 25°C 30 0dBFS –3dBFS –6dBFS –12dBFS 40 07852-034 40 30 50 2.4GHz 2.0GHz 1.6GHz 1.0GHz 07852-037 60 Figure 16. NSD vs. fOUT over fDAC, 1-Channel QAM, Full-Scale Current = 20 mA 0 200 400 600 fOUT (MHz) 800 1000 1200 07852-019 IMD (dBc) 70 Figure 19. NSD vs. fOUT over Temperature, 1-Channel QAM, fDAC = 2.4 GHz, Full-Scale Current = 20 mA Rev. A | Page 13 of 76 AD9789 –5 DOCSIS3 –40°C 0°C +25°C +85°C ACLR (dBc) –25 –35 –45 –45 –55 –65 –65 –75 –75 –85 50 250 450 650 FREQUENCY (MHz) 850 –85 50 Figure 20. ACLR Performance over Temperature, 1-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 20 mA, fOUT = 200 MHz, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc; Harmonic Exception Is −63 dBc) –55 –60 –60 HARMONIC LEVEL (dBc) –55 –65 –70 –75 DOCSIS3 25°C 65°C 85°C –80 –85 0 100 200 300 400 500 600 fOUT (MHz) 700 800 900 250 350 450 550 650 FREQUENCY (MHz) 750 1000 850 950 –65 –70 –75 DOCSIS3 25°C 65°C 85°C –80 Figure 21. Second-Order Harmonic Performance vs. fOUT over Temperature, 1-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc; Harmonic Exception Is −63 dBc) –85 0 100 200 300 400 500 600 700 800 900 fOUT (MHz) 1000 Figure 24. Third-Order Harmonic Performance vs. fOUT over Temperature, 1-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc; Harmonic Exception Is −63 dBc) –55 –5 DOCSIS3 25°C 65°C 85°C –60 150 Figure 23. ACLR Performance over Temperature, 1-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 20 mA, fOUT = 800 MHz, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc) 07852-014 HARMONIC LEVEL (dBc) –35 –55 07852-015 ACLR (dBc) –25 DOCSIS3 –40°C 0°C +25°C +85°C –15 07852-017 –15 07852-018 –5 –15 DOCSIS3 2.3GHz 2.2GHz 2.4GHz –25 ACLR (dBc) ACLR (dBc) –65 –70 –35 –45 –55 –75 –65 –80 100 200 300 400 500 600 fOUT (MHz) 700 800 900 1000 –85 50 Figure 22. Noise Floor vs. fOUT over Temperature (ACLR Measured Beyond 30 MHz), 1-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc) 250 450 650 FREQUENCY (MHz) 850 1050 07852-039 0 07852-031 –85 –75 Figure 25. ACLR Performance over fDAC, 1-Channel QAM, fOUT = 850 MHz, Full-Scale Current = 20 mA, Temperature = 25°C, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc) Rev. A | Page 14 of 76 AD9789 0 –5 DOCSIS3 CMOS LVDS –10 –20 DOCSIS3 25°C 65°C 85°C –15 –25 ACLR (dBc) –40 –50 –65 –70 100 200 300 400 500 600 700 FREQUENCY (MHz) 800 900 1000 Figure 26. ACLR Performance for CMOS and LVDS Interfaces, 1-Channel QAM, fOUT = 840 MHz, fDAC = 2.4 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 (DOCSIS SPEC Is −73 dBc) 450 650 FREQUENCY (MHz) 850 1050 Figure 29. ACLR Performance over Temperature, 2-Channel QAM, fOUT = 200 MHz, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 32 (DOCSIS SPEC Is −70 dBc; Harmonic Exception Is −63 dBc) DOCSIS3 25°C 65°C 85°C –60 HARMONIC LEVEL (dBc) –25 –35 –45 –55 –65 –65 –70 –75 DOCSIS3 25°C 65°C 85°C –80 –75 250 450 650 FREQUENCY (MHz) 850 –85 07852-042 –85 50 1050 0 100 200 300 400 500 600 700 800 900 1000 fOUT (MHz) Figure 27. ACLR Performance over Temperature, 2-Channel QAM, fOUT = 800 MHz, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 32 (DOCSIS SPEC Is −70 dBc) 07852-045 –15 Figure 30. Second Harmonic Performance vs. fOUT over Temperature, 2-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 32 (DOCSIS SPEC Is −70 dBc; Harmonic Exception Is −63 dBc) –55 –55 –60 –65 –65 ACLR (dBc) –60 –70 –75 DOCSIS3 25°C 65°C 85°C –70 –75 DOCSIS3 25°C 65°C 85°C 0 100 200 300 400 500 600 fOUT (MHz) 700 800 900 1000 –80 –85 07852-043 –80 –85 250 –55 –5 ACLR (dBc) –85 50 07852-040 0 07852-044 –75 –80 HARMONIC LEVEL (dBc) –45 –55 –60 –90 –35 Figure 28. Third-Order Harmonic Performance vs. fOUT over Temperature, 2-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 32 (DOCSIS SPEC Is −70 dBc; Harmonic Exception Is −63 dBc) 0 100 200 300 400 500 600 fOUT (MHz) 700 800 900 1000 07852-046 ACLR (dBc) –30 Figure 31. Noise Floor vs. fOUT over Temperature (ACLR Measured Beyond 30 MHz), 2-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 32 (DOCSIS SPEC Is −70 dBc) Rev. A | Page 15 of 76 AD9789 0 0 –10 DOCSIS3 –40°C 0°C +25°C +85°C ACLR (dBc) –40 –50 –40 –50 –60 –60 –70 –70 –80 50 250 450 650 FREQUENCY (MHz) 850 1050 –80 50 Figure 32. ACLR Performance over Temperature, 4-Channel QAM, fOUT = 200 MHz, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc; Harmonic Exception Is −63 dBc) –55 –60 –60 HARMONIC LEVEL (dBc) –55 –65 –70 –75 DOCSIS3 25°C 65°C 85°C –80 0 100 200 300 400 500 600 fOUT (MHz) 700 800 900 1000 1050 –65 –70 –75 –85 DOCSIS3 25°C 65°C 85°C 0 100 200 300 400 500 600 700 800 900 1000 fOUT (MHz) Figure 36. Third-Order Harmonic Performance vs. fOUT over Temperature, 4-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc; Harmonic Exception Is −63 dBc) 0 –55 –10 –60 –20 ACLR (dBc) –65 ACLR (dBc) 850 –80 Figure 33. Second-Order Harmonic Performance vs. fOUT over Temperature, 4-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc; Harmonic Exception Is −63 dBc) –70 –75 100 200 300 400 500 600 fOUT (MHz) 700 800 900 1000 –30 –40 –50 –70 –80 50 07852-028 0 DOCSIS3 2.3GHz 2.2GHz 2.4GHz –60 DOCSIS3 25°C 65°C 85°C –80 –85 450 650 FREQUENCY (MHz) Figure 34. Noise Floor vs. fOUT over Temperature (ACLR Measured Beyond 30 MHz), 4-Channel QAM, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc) 250 450 650 FREQUENCY (MHz) 850 1050 07852-047 –85 250 Figure 35. ACLR Performance over Temperature, 4-Channel QAM, fOUT = 800 MHz, fDAC = 2.3 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc) 07852-026 HARMONIC LEVEL (dBc) –30 07852-029 –30 DOCSIS3 –40°C 0°C +25°C +85°C –20 07852-027 ACLR (dBc) –20 07852-030 –10 Figure 37. ACLR Performance over fDAC, 4-Channel QAM, fOUT = 850 MHz, Full-Scale Current = 25 mA, Temperature = 25°C, Sum Scale = 20 (DOCSIS SPEC Is −67 dBc) Rev. A | Page 16 of 76 AD9789 RMS RESULTS CARRIER POWER –18.10dBm/ 6.00000MHz VBW 560kHz FREQ. OFFSET 3.375MHz 6.375MHz 12.00MHz 18.00MHz REF BW 750.0kHz 5.250MHz 6.000MHz 6.000MHz SPAN 42MHz SWEEP 39.12ms (601 PTS) LOWER dBc dBm –65.57 –83.66 –75.01 –93.11 –76.83 –94.92 –77.17 –95.26 UPPER dBc dBm –68.98 –87.07 –74.62 –92.71 –76.46 –94.55 –76.56 –94.66 CENTER 840.00MHz RES BW 30kHz Figure 38. 1-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 48, Full-Scale Current = 20 mA, Span = 42 MHz CENTER 840.00MHz RES BW 30kHz RMS RESULTS CARRIER POWER –21.75dBm/ 6.00000MHz VBW 300kHz FREQ. OFFSET 3.375MHz 6.375MHz 12.00MHz 18.00MHz REF BW 750.0kHz 5.250MHz 6.000MHz 6.000MHz SPAN 18MHz SWEEP 58.4ms (601 PTS) Figure 40. 1-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 48, Full-Scale Current = 20 mA, Span = 18 MHz REF –35.91dBm SPAN 42MHz SWEEP 136.2ms (601 PTS) LOWER UPPER dBc dBm dBc dBm –71.64 –93.39 –72.50 –94.25 –73.71 –95.47 –66.72 –88.47 –73.58 –95.33 0.50 –21.10 –73.70 –95.45 –66.72 –88.48 VBW 300kHz FREQ. LOWER UPPER OFFSET REF BW dBc dBm dBc dBm RMS RESULTS CARRIER POWER 3.375MHz 750.0kHz –73.99 –91.97 –74.93 –92.91 6.375MHz 5.250MHz –74.94 –92.92 –75.35 –93.33 –17.98dBm/ 6.00000MHz ATTEN 2dB START 831.00MHz RES BW 30kHz 07852-061 REF –35.91dBm ATTEN 2dB RMS RESULTS CARRIER POWER –21.29dBm/ 6.00000MHz Figure 39. 2-Channel QAM ACLR, fOUT = 840 MHz, Sum Scale = 32, Full-Scale Current = 25 mA, Span = 42 MHz, Channel 1 ATTEN 2dB VBW 300kHz FREQ. OFFSET 3.375MHz 6.375MHz 12.00MHz 18.00MHz STOP 873MHz SWEEP 136.2ms (601 PTS) LOWER REF BW dBc dBm 750.0kHz –70.07 –92.16 5.250MHz –69.05 –90.34 6.000MHz –0.49 –21.78 6.000MHz –66.61 –87.90 UPPER dBc dBm –73.20 –94.49 –73.87 –95.16 –73.29 –94.58 –73.98 –95.27 Figure 41. 2-Channel QAM ACLR, fOUT = 840 MHz, Sum Scale = 32, Full-Scale Current = 25 mA, Span = 42 MHz, Channel 2 Rev. A | Page 17 of 76 07852-066 CENTER 840.00MHz RES BW 56kHz REF –32.76dBm 07852-023 ATTEN 2dB 07852-020 REF –32.76dBm AD9789 VBW 300kHz SPAN 18MHz SWEEP 58.4ms (601 PTS) FREQ. LOWER UPPER OFFSET REF BW dBc dBm dBc dBm RMS RESULTS CARRIER POWER 3.375MHz 750.0kHz –75.37 –96.93 –75.56 –97. 11 –21.56dBm/ 6.375MHz 5.250MHz –73.85 –95.41 –72.54 –94.10 6.00000MHz CENTER 852.00MHz RES BW 30kHz Figure 42. Zoomed 2-Channel QAM ACLR, fOUT = 840 MHz, Sum Scale = 32, Full-Scale Current = 25 mA, Span = 18 MHz, Channel 1 CENTER 834.00MHz RES BW 30kHz VBW 300kHz RMS RESULTS CARRIER POWER –23.63dBm/ 6.00000MHz FREQ. OFFSET 3.375MHz 6.375MHz 12.00MHz 18.00MHz REF BW 750.0kHz 5.250MHz 6.000MHz 6.000MHz SPAN 42MHz SWEEP 136.2ms (601 PTS) ATTEN 2dB CENTER 852.00MHz VBW 300kHz RES BW 30kHz Figure 43. 4-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 20, Full-Scale Current = 25 mA, Span = 42 MHz, Channel 1 SPAN 18MHz SWEEP 58.4ms (601 PTS) Figure 44. Zoomed 2-Channel QAM ACLR, fOUT = 840 MHz, Sum Scale = 32, Full-Scale Current = 25 mA, Span = 18 MHz, Channel 2 REF –35.96dBm LOWER UPPER dBc dBm dBc dBm –70.33 –93.96 –11.07 –34.70 –69.04 –92.67 –0.49 –24.12 –70.38 –94.01 0.00 –23.63 –71.02 –94.65 0.43 –23.20 VBW 300kHz FREQ. LOWER UPPER OFFSET REF BW dBc dBm dBc dBm RMS RESULTS CARRIER POWER 3.375MHz 750.0kHz –75.51 –96.54 –75.17 –96.20 –21.03dBm/ 6.375MHz 5.250MHz –72.55 –93.58 –73.90 –94.93 6.00000MHz ATTEN 2dB 07852-021 REF –35.96dBm ATTEN 2dB RMS RESULTS CARRIER POWER –23.23dBm/ 6.00000MHz FREQ. OFFSET 3.375MHz 6.375MHz 12.00MHz 18.00MHz REF BW 750.0kHz 5.250MHz 6.000MHz 6.000MHz SPAN 42MHz SWEEP 136.2ms (601 PTS) LOWER dBc dBm –11.10 –34.32 –0.75 –23.98 –0.59 –23.81 –0.35 –23.58 UPPER dBc dBm –72.19 –95.42 –68.97 –92.20 –70.32 –93.55 –70.70 –93.93 07852-022 CENTER 840.00MHz RES BW 30kHz REF –35.91dBm 07852-067 ATTEN 2dB 07852-065 REF –35.91dBm Figure 45. 4-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 20, Full-Scale Current = 25 mA, Span = 42 MHz, Channel 4 Rev. A | Page 18 of 76 AD9789 CENTER 834.00MHz RES BW 30kHz VBW 300kHz REF –35.96dBm SPAN 18MHz SWEEP 58.4ms (601 PTS) CENTER 852.00MHz RES BW 30kHz FREQ. LOWER UPPER OFFSET REF BW dBc dBm dBc dBm RMS RESULTS CARRIER POWER 3.375MHz 750.0kHz –72.95 –96.56 –10.86 –34.48 6.375MHz 5.250MHz –69.38 –92.99 –0.51 –24.13 –23.62dBm/ 6.00000MHz Figure 49. Zoomed 4-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 20, Full-Scale Current = 25 mA, Span = 18 MHz, Channel 4 50 50 48 48 46 46 44 44 42 42 40 38 38 36 +25°C +85°C –40°C 32 150 250 350 450 550 650 fOUT (MHz) 750 850 950 +25°C +85°C –40°C 34 32 30 50 07852-032 34 150 250 350 450 550 650 750 850 950 fOUT (MHz) Figure 47. Modulation Error Ratio, Equalized, 1-Channel 256-QAM, fDAC = 2.29376 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 (Equalization Filter from Demodulation Toolbox on Spectrum Analyzer Used) Figure 50. Modulation Error Ratio, Equalized, 4-Channel 256-QAM, fDAC = 2.29376 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 (Equalization Filter from Demodulation Toolbox on Spectrum Analyzer Used) 50 50 48 48 46 46 44 44 42 42 MER (dB) 40 38 36 40 38 36 +25°C +85°C –40°C 32 150 250 350 450 550 650 fOUT (MHz) 750 850 950 +25°C +85°C –40°C 34 32 30 50 07852-033 34 150 250 350 450 550 fOUT (MHz) Figure 48. Modulation Error Ratio, Unequalized, 1-Channel 256-QAM, fDAC = 2.29376 GHz, Full-Scale Current = 20 mA, Sum Scale = 48 650 750 850 950 07852-036 MER (dB) 40 07852-035 36 30 50 SPAN 18MHz SWEEP 58.4ms (601 PTS) VBW 300kHz FREQ. LOWER UPPER OFFSET REF BW dBc dBm dBc dBm RMS RESULTS CARRIER POWER 3.375MHz 750.0kHz –11.20 –34.40 –74.44 –97.64 6.375MHz 5.250MHz –0.77 –23.96 –69.07 –92.26 –23.20dBm/ 6.00000MHz MER (dB) MER (dB) Figure 46. Zoomed 4-Channel QAM ACLR, fOUT = 840 MHz, Temperature = 25°C, Sum Scale = 20, Full-Scale Current = 25 mA, Span = 18 MHz, Channel 1 30 50 ATTEN 2dB 07852-025 ATTEN 2dB 07852-024 REF –35.96dBm Figure 51. Modulation Error Ratio, Unequalized, 4-Channel 256-QAM, fDAC = 2.29376 GHz, Full-Scale Current = 25 mA, Sum Scale = 20 Rev. A | Page 19 of 76 AD9789 REF –32.62dBm 80 ATTEN 0dB 75 70 65 60 SFDR (dBc) 55 50 45 40 35 30 25 20 CENTER 2.100GHz RES BW 30kHz Figure 52. SFDR vs. fOUT in Mix Mode, fDAC = 2.4 GHz, Full-Scale Current = 20 mA (Second Nyquist Zone Performance) 90 RMS RESULTS CARRIER POWER –19.95dBm/ 3.84000MHz 85 VBW 300kHz FREQ. OFFSET 5.000MHz 10.00MHz 15.00MHz 20.00MHz 25.00MHz REF BW 3.840MHz 3.840MHz 3.840MHz 3.840MHz 3.840MHz SPAN 53.84MHz SWEEP 174.6ms (601 PTS) LOWER dBc dBm –68.93 –88.88 –71.31 –91.26 –73.43 –93.37 –75.12 –95.07 –75.60 –95.55 UPPER dBc dBm –67.99 –87.94 –70.42 –90.37 –72.68 –92.63 –74.89 –94.84 –76.51 –96.46 07852-092 10 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 fOUT (MHz) 07852-068 15 Figure 55. One-Carrier WCDMA ACLR in Mix Mode, fOUT = 2.1 GHz, fDAC = 2304 MHz, Full-Scale Current = 20 mA 80 75 IMD (dBc) 70 REF –38.62dBm 65 ATTEN 2dB 60 55 50 45 40 30 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 fOUT (MHz) 07852-076 35 Figure 53. IMD vs. fOUT in Mix Mode, fDAC = 2.4 GHz, Full-Scale Current = 20 mA (Second Nyquist Zone Performance) –45 –50 CENTER 2.102 50GHz RES BW 30kHz VBW 300kHz FIRST ADJACENT CHANNE L SECOND ADJACENT CHANNE L THIRD ADJACENT CHANNE L FIFTH ADJACENT CHANNE L RMS RESULTS CARRIER POWER –26.06dBm/ 3.84000MHz ACLR (dBc) –55 –60 –65 REF BW 3.840MHz 3.840MHz 3.840MHz 3.840MHz 3.840MHz 3.840MHz LOWER dBc dBm –0.25 –26.31 –0.42 –26.48 –64.07 –90.13 –65.36 –91.42 –66.86 –92.92 –67.83 –93.89 UPPER dBc dBm –0.42 –26.47 –63.50 –89.56 –65.13 –91.18 –66.97 –93.03 –68.70 –94.76 –68.64 –94.70 Figure 56. Four-Carrier WCDMA ACLR in Mix Mode, fOUT = 2.1 GHz, fDAC = 2304 MHz, Full-Scale Current = 20 mA –70 –75 07852-075 –80 –85 1150 1250 1350 1450 1550 1650 1750 1850 1950 2050 2150 2250 fOUT (MHz) FREQ. OFFSET 5.000MHz 10.00MHz 15.00MHz 20.00MHz 25.00MHz 30.00MHz SPAN 63.84MHz SWEEP 207ms (601 PTS) Figure 54. ACLR vs. fOUT in Mix Mode with One-Carrier WCDMA, fDAC = 2304 MHz, Full-Scale Current = 20 mA (Second Nyquist Zone Performance) Rev. A | Page 20 of 76 07852-093 –40 AD9789 1100 2000 AVDD33 DVDD33 (LVDS) DVDD33 (CMOS) DVDD18 DVDD15 CVDD18 1600 700 600 500 400 300 1000 800 600 100 200 1.2 1.4 1.6 1.8 fDAC (GHz) 2.0 2.2 2.4 0 1.0 500 1200 POWER DISSIPATION (mW) 600 400 300 200 1.6 1.8 2.0 2.2 2.4 TOTAL (CMOS) TOTAL (LVDS) 1000 800 600 400 200 1.2 1.4 1.6 1.8 2.0 2.2 2.4 fDAC (GHz) 0 1.0 07852-095 0 1.0 180 160 140 AVDD33 120 100 80 60 40 14 16 18 20 22 24 26 FULL-SCALE CURRENT (mA) 28 30 32 07852-098 20 12 1.6 1.8 2.0 2.2 2.4 Figure 61. Total Power Dissipation vs. fDAC, 16× Interpolation, One Channel Enabled, fOUT = 70 MHz, Full-Scale Current = 20 mA 200 10 1.4 fDAC (GHz) Figure 58. Power Dissipation by Supply vs. fDAC, 16× Interpolation, One Channel Enabled, fOUT = 70 MHz, Full-Scale Current = 20 mA 8 1.2 Figure 59. AVDD33 Power Dissipation vs. Full-Scale Current Rev. A | Page 21 of 76 07852-097 100 0 1.4 Figure 60. Total Power Dissipation vs. fDAC, 4-Channel DOCSIS, fOUT = 915 MHz, Full-Scale Current = 25 mA (Datapath Configuration: QAM Encoder On, SRRC Filter On, Four 2× Interpolation Filters On) 1400 AVDD33 DVDD33 (LVDS) DVDD33 (CMOS) DVDD18 DVDD15 CVDD18 1.2 fDAC (GHz) 700 POWER DISSIPATION (mW) 1200 400 Figure 57. Power Dissipation by Supply vs. fDAC, 4-Channel DOCSIS, fOUT = 915 MHz, Full-Scale Current = 25 mA (Datapath Configuration: QAM Encoder On, SRRC Filter On, Four 2× Interpolation Filters On) POWER DISSIPATION (mW) 1400 200 0 1.0 TOTAL (CMOS) TOTAL (LVDS) 07852-096 800 07852-094 POWER DISSIPATION (mW) 900 1800 POWER DISSIPATION (mW) 1000 AD9789 TERMINOLOGY Monotonicity A DAC is monotonic if the output either increases or remains constant as the digital input increases. Offset Error Offset error is the deviation of the output current from the ideal of 0. For IOUTP, 0 mA output is expected when all inputs are set to 0. For IOUTN, 0 mA output is expected when all inputs are set to 1. Spurious-Free Dynamic Range (SFDR) SFDR is the difference, in dB, between the peak amplitude of the output signal and the peak spurious signal over the specified bandwidth. Noise Spectral Density (NSD) NSD is the converter noise power per unit of bandwidth. NSD is usually specified in dBm/Hz in the presence of a 0 dBm fullscale signal. Gain Error Gain error is the difference between the actual and ideal output span. The actual span is determined by the output when all inputs are set to 1s minus the output when all inputs are set to 0s. Adjacent Channel Leakage Ratio (ACLR) The adjacent channel leakage (power) ratio is the ratio, in dBc, between the measured power within a channel relative to its adjacent channels. Temperature Drift Modulation Error Ratio (MER) Modulated signals create a discrete set of output values referred to as a constellation. Each symbol creates an output signal corresponding to one point on the constellation. MER is a measure of the discrepancy between the average output symbol magnitude and the rms error magnitude of the individual symbol. Temperature drift is specified as the maximum change from the ambient (25°C) value to the value at either TMIN or TMAX. For offset, gain, and reference drift, the drift is reported in ppm per °C. Power Supply Rejection (PSR) PSR is the maximum change in the full-scale output as the supplies are varied from nominal to minimum and maximum specified voltages. Output Compliance Range The output compliance range is the range of allowable voltage at the output of a current output DAC. Operation beyond the maximum compliance limits may cause either output stage saturation or breakdown, resulting in nonlinear performance. Intermodulation Distortion (IMD) IMD is the result of two or more signals at different frequencies mixing together. Many products are created according to the formula af1 ± bf2, where a and b are integer values. Rev. A | Page 22 of 76 AD9789 SERIAL CONTROL PORT The AD9789 serial control port is a flexible, synchronous serial communications port that allows an easy interface to many industry-standard microcontrollers and microprocessors. The AD9789 serial control port is compatible with most synchronous transfer formats, including both the Motorola SPI® and Intel® SSR protocols. The serial control port allows read/write access to all registers that configure the AD9789. Single- or multiple-byte transfers are supported, as well as MSB first or LSB first transfer formats. The AD9789 serial control port can be configured for a single bidirectional I/O pin (SDIO only) or for two unidirectional I/O pins (SDIO/SDO). By default, the AD9789 is in unidirectional long instruction mode (long instruction mode is the only instruction mode supported). SERIAL CONTROL PORT PIN DESCRIPTIONS The SCLK (serial clock) pin is the serial shift clock. This pin is an input. SCLK is used to synchronize serial control port reads and writes. Write data bits are registered on the rising edge of this clock, and read data bits are registered on the falling edge. This pin is internally pulled down by a 30 kΩ resistor to ground. SDIO (serial data input/output) is a dual-purpose pin that acts as an input only (unidirectional mode) or as both an input and an output (bidirectional mode). The AD9789 defaults to the unidirectional I/O mode (Register 0x00[7] = 0). The SDO (serial data output) pin is used only in the unidirectional I/O mode as a separate output pin for reading back data. CS (chip select bar) is an active low control that gates the read and write cycles. When CS is high, SDO and SDIO are in a high impedance state. This pin is internally pulled up by a 30 kΩ resistor to DVDD33. M1 CS L1 SDO N1 SDIO P1 AD9789 SERIAL CONTROL PORT 07852-048 SCLK Figure 62. Serial Control Port GENERAL OPERATION OF SERIAL CONTROL PORT A write or read operation to the AD9789 is initiated by pulling CS low. CS stall high is supported in modes where three or fewer bytes of data (plus the instruction data) are transferred (see Table 7). In these modes, CS can temporarily return high on any byte boundary, allowing time for the system controller to process the next byte. CS can go high on byte boundaries only and can go high during either part (instruction or data) of the transfer. During CS stall high mode, the serial control port state machine enters a wait state until all data is sent. If the system controller decides to abort the transfer before all of the data is sent, the state machine must be reset by either completing the remaining transfers or by returning CS low for at least one complete SCLK cycle (but less than eight SCLK cycles). Raising CS on a nonbyte boundary terminates the serial transfer and flushes the buffer. In streaming mode (see Table 7), any number of data bytes can be transferred in a continuous stream. The register address is automatically incremented or decremented (see the MSB/LSB First Transfers section). CS must be raised at the end of the last byte to be transferred, thereby ending streaming mode. Communication Cycle—Instruction Plus Data There are two parts to a communication cycle with the AD9789. In the first part, a 16-bit instruction word is written to the AD9789, coincident with the first 16 SCLK rising edges. The instruction word provides the AD9789 serial control port with information regarding the data transfer, which is the second part of the communication cycle. The instruction word defines whether the upcoming data transfer is a read or a write, the number of bytes in the data transfer, and the starting register address for the first byte of the data transfer. Write If the instruction word is for a write operation, the second part of the communication cycle is the transfer of data into the serial control port buffer of the AD9789. Data bits are registered on the rising edge of SCLK. The length of the transfer (one, two, or three bytes or streaming mode) is indicated by two bits (N1 and N0) in the instruction byte. When the transfer is one, two, or three bytes (but not streaming mode), CS can be raised after each sequence of eight bits to stall the bus, except after the last byte, where it ends the cycle. When the bus is stalled, the serial transfer resumes when CS is lowered. Raising CS on a nonbyte boundary resets the serial control port. During a write, streaming mode does not skip reserved or blank registers; therefore, the user must know what bit pattern to write to the reserved registers to preserve proper operation of the part. It does not matter what data is written to blank registers. Most writes to the control registers immediately reconfigure the device. However, Register 0x16 through Register 0x1D do not directly control device operation. They provide data to internal logic that must perform additional operations on the data before it is downloaded and the device configuration is changed. For any updates to Register 0x16 through Register 0x1D to take effect, the FREQNEW bit (Register 0x1E[7]) must be set to 1 (this bit is self-clearing). Any number of bytes of data can be changed before updating registers. Setting the FREQNEW bit simultaneously updates Register 0x16 through Register 0x1D. In a similar fashion, any changes to Register 0x22 and Register 0x23 require PARMNEW (Register 0x24[7]) to be toggled from a low state to a high state before the new values take effect. Unlike the FREQNEW bit, PARMNEW is not self-clearing. Rev. A | Page 23 of 76 AD9789 Read If the instruction word is for a read operation, the next N × 8 SCLK cycles clock out the data from the address specified in the instruction word, where N is 1 to 3 as determined by Bits[N1:N0]. If N = 4, the read operation is in streaming mode, continuing until CS is raised. Streaming mode does not skip over reserved or blank registers. The readback data is valid on the falling edge of SCLK. The default mode of the AD9789 serial control port is the unidirectional mode. In unidirectional mode, the readback data appears on the SDO pin. It is also possible to set the AD9789 to bidirectional mode using the SDIO_DIR bit (Register 0x00[7]). In bidirectional mode, both the sent data and the readback data appear on the SDIO pin. A readback request reads the data that is in the serial control port buffer area or the data in the active registers (see Figure 63). The AD9789 supports only the long instruction mode; therefore, Register 0x00[4:3] reads 11 (this register uses mirrored bits). Long instruction mode is the default at power-up or reset, and writing to these bits has no effect. SDO CS SERIAL CONTROL PORT FREQNEW WRITE REGISTER 0x1E = 0x10 TO UPDATE REGISTERS 07852-049 SDIO ACTIVE REGISTERS SCLK BUFFER REGISTERS The AD9789 uses Register Address 0x00 to Register Address 0x55. Figure 63. Relationship Between Serial Control Port Buffer Registers and Active Registers of the AD9789 INSTRUCTION WORD (16 BITS) The MSB of the instruction word is R/W, which indicates whether the instruction is a read or a write. The next two bits, N1 and N0, indicate the length of the transfer in bytes. The final 13 bits (Bits[A12:A0]) are the address at which to begin the read or write operation. For a write, the instruction word is followed by the number of bytes of data indicated by Bits[N1:N0] (see Table 7). Table 7. Byte Transfer Count N1 0 0 1 1 N0 0 1 0 1 Bytes to Transfer 1 2 3 Streaming mode Bits[A12:A0] select the address within the register map that is written to or read from during the data transfer portion of the communication cycle. Only Bits[A6:A0] are needed to cover the range of the 0x55 registers used by the AD9789. Bits[A12:A7] must always be 0. For multibyte transfers, this address is the starting byte address. In MSB first mode, subsequent bytes increment the address. MSB/LSB FIRST TRANSFERS The AD9789 instruction word and byte data can be MSB first or LSB first. Any data written to Register 0x00 must be mirrored, the upper four bits (Bits[7:4]) with the lower four bits (Bits [3:0]). This makes it irrelevant whether LSB first or MSB first is in effect. As an example of this mirroring, the default setting for Register 0x00[7:0] is 0x18, which mirrors Bit 4 and Bit 3. These bits set the long instruction mode (the default and the only mode supported). The default for the AD9789 is MSB first. When LSB first is set by Register 0x00[1] and Register 0x00[6], it takes effect immediately. In multibyte transfers, subsequent bytes reflect any changes in the serial port configuration. When MSB first mode is active, the instruction and data bytes must be written from MSB to LSB. Multibyte data transfers in MSB first format start with an instruction byte that includes the register address of the most significant data byte. Subsequent data bytes must follow in order from the high address to the low address. In MSB first mode, the serial control port internal address generator decrements for each data byte of the multibyte transfer cycle. When LSB first mode is active, the instruction and data bytes must be written from LSB to MSB. Multibyte data transfers in LSB first format start with an instruction byte that includes the register address of the least significant data byte followed by multiple data bytes. The internal byte address generator of the serial control port increments for each byte of the multibyte transfer cycle. The AD9789 serial control port register address decrements from the register address just written toward 0x00 for multibyte I/O operations if the MSB first mode is active (default). If the LSB first mode is active, the register address of the serial control port increments from the address just written toward 0x55 for multibyte I/O operations. Streaming mode always terminates when it reaches Address 0x2F. Note that unused addresses are not skipped during multibyte I/O operations. Table 8. Streaming Mode (No Addresses Are Skipped) Write Mode LSB First MSB First Rev. A | Page 24 of 76 Address Direction Increment Decrement Stop Sequence 0x02D, 0x02E, 0x02F, stop 0x001, 0x000, 0x02F, stop AD9789 Table 9. Serial Control Port, 16-Bit Instruction Word, MSB First MSB I15 I14 I13 I12 I11 I10 I9 I8 I7 I6 I5 I4 I3 I2 I1 LSB I0 R/W N1 N0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 CS SCLK DON'T CARE SDIO DON'T CARE R/W N1 N0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 16-BIT INSTRUCTION HEADER REGISTER (N) DATA DON'T CARE REGISTER (N – 1) DATA 07852-050 DON'T CARE Figure 64. Serial Control Port Write—MSB First, 16-Bit Instruction, Two Bytes of Data CS SCLK DON'T CARE SDIO DON'T CARE D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 16-BIT INSTRUCTION HEADER REGISTER (N) DATA REGISTER (N – 1) DATA REGISTER (N – 2) DATA REGISTER (N – 3) DATA DON'T CARE 07852-051 SDO DON'T CARE R/W N1 N0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 Figure 65. Serial Control Port Read—MSB First, 16-Bit Instruction, Four Bytes of Data tDS tHI tS tDH CS DON'T CARE SDIO DON'T CARE DON'T CARE N1 R/W N0 A12 A11 A10 A9 A8 A7 A6 A5 D4 D3 D2 D1 D0 DON'T CARE 07852-052 SCLK tC tCLK tLO Figure 66. Serial Control Port Write—MSB First, 16-Bit Instruction, Timing Measurements CS tDV SDIO SDO DATA BIT N DATA BIT N – 1 07852-053 SCLK Figure 67. Timing Diagram for Serial Control Port Register Read CS SCLK DON'T CARE A0 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 N0 N1 R/W D0 D1 D2 D3 D4 D5 D6 D7 D0 16-BIT INSTRUCTION HEADER REGISTER (N) DATA D1 D2 D3 D4 D5 D6 D7 REGISTER (N + 1) DATA Figure 68. Serial Control Port Write—LSB First, 16-Bit Instruction, Two Bytes of Data Rev. A | Page 25 of 76 DON'T CARE 07852-054 SDIO DON'T CARE DON'T CARE AD9789 tS tC CS tCLK tHI SCLK tLO tDS SDIO BIT N BIT N + 1 Figure 69. Serial Control Port Timing—Write Table 10. Serial Control Port Timing Parameter tDS tDH tCLK tS tC tHI tLO tDV Description Setup time between data and rising edge of SCLK Hold time between data and rising edge of SCLK Period of the clock Setup time between CS falling edge and SCLK rising edge (start of communication cycle) Setup time between SCLK rising edge and CS rising edge (end of communication cycle) Minimum period that SCLK should be in a logic high state Minimum period that SCLK should be in a logic low state SCLK to valid SDIO and SDO (see Figure 67) Rev. A | Page 26 of 76 07852-055 tDH AD9789 SPI REGISTER MAP Do not write to the following registers unless instructed otherwise: Register 0x34, Register 0x35, Register 0x37, Register 0x3B, Register 0x3F, or Register 0x40 through Register 0x55. Table 11. Register Map Addr 0x00 0x01 0x02 0x03 0x04 0x05 0x06 0x07 0x08 0x09 0x0A 0x0B 0x0C 0x0D 0x0E 0x0F 0x10 0x11 0x12 0x13 0x14 0x15 0x16 0x17 0x18 0x19 0x1A 0x1B 0x1C 0x1D 0x1E 0x1F 0x20 0x21 0x22 0x23 0x24 0x25 0x26 0x27 0x28 0x29 Register Name SPI control Saturation counter Parity counter Interrupt enable Interrupt status/clear Channel enable Bypass QAM/SRRC configuration Summing node scalar Input scalar NCO 0 frequency tuning word Bit 7 SDIO_DIR Bit 6 LSBFIRST Bit 5 RESET PARERR PARERR BISTDONE BISTDONE PARMSET PARMSET QAM SRRC Reserved PARCNT[7:0] PARMCLR LOCKACQ PARMCLR LOCKACQ Reserved Reserved ALPHA[1:0] Reserved Bit 2 Bit 1 Bit 0 LOCKLOST LOCKLOST SATERR SATERR Reserved Reserved CHANEN[3:0] INT[4:0] MAPPING[2:0] NCO 3 frequency tuning word Rate converter denominator (Q) Rate converter numerator (P) Interpolating BPF center frequency FREQNEW CMOS_BUS BIN Reserved PARMNEW Reserved CMOS_CTRL Reserved BUSWDTH[1:0] DCODIV[2:0] DSCPHZ[3:0] DCO_INV IF_MODE DATWDTH CMPLX Reserved CHAN0GAIN[7:0] CHAN1GAIN[7:0] CHAN2GAIN[7:0] CHAN3GAIN[7:0] Reserved Rev. A | Page 27 of 76 0x00 0x00 0x00 0x0D INSCALE[7:0] FTW0[7:0] FTW0[15:8] FTW0[23:16] FTW1[7:0] FTW1[15:8] FTW1[23:16] FTW2[7:0] FTW2[15:8] FTW2[23:16] FTW3[7:0] FTW3[15:8] FTW3[23:16] Q[7:0] Q[15:8] Q[23:16] P[7:0] P[15:8] P[23:16] FC[7:0] FC[15:8] Reserved NCO 2 frequency tuning word Default 0x18 0x00 0x00 0x00 0x01 SUMSCALE[7:0] NCO 1 frequency tuning word Frequency update Hardware version Interface configuration Data control DCO frequency Internal clock phase adjust Parameter update Channel 0 gain Channel 1 gain Channel 2 gain Channel 3 gain Spectrum shaping Bit 4 Bit 3 LNG_INST SATCNT[7:0] VER[3:0] CHANPRI PAR[1:0] LTNCY[2:0] ONES[3:0] SNCPHZ[3:0] 0x20 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x80 0x00 0x00 0x80 0x00 0x00 0x00 0x03 0xC8 0x61 0x1F 0x85 SPEC_INV 0x00 0x80 0x80 0x80 0x80 0x00 AD9789 Addr 0x2F 0x30 0x31 0x32 0x33 0x34 0x35 0x36 0x37 0x38 0x39 0x3A 0x3B 0x3C 0x3D 0x3E 0x3F 0x40 0x41 0x42 0x43 0x44 0x45 0x46 0x47 0x48 0x49 0x4A 0x4B 0x4C 0x4D 0x50 0x51 0x52 0x53 0x54 0x55 Register Name Mu Delay Control 1 Mu control duty cycle Clock Receiver 1 Clock Receiver 2 Mu Delay Control 2 Reserved Reserved DAC bias Reserved DAC decoder Mu Delay Control 3 Mu Delay Control 4 Reserved Full-Scale Current 1 Full-Scale Current 2 Phase detector control Reserved BIST control BIST status BIST zero padding length Bit 7 SEARCH_ TOL Duty cycle correct enable CLK_DIS MU_CLKDIS Bit 6 SEARCH_ERR Bit 5 TRACK_ ERR BIST Signature 1 Bit 2 Bit 1 GUARDBAND[4:0] Bit 0 MANUAL_ADJ[5:0] (Factory test only) CLKN_CML[3:0] Reserved PSIGN SLOPE MODE[1:0] NSIGN MU_EN MSEL[1:0] DAC decoder mode SEARCH_DIR[1:0] MUPHZ[4:0] Reserved CMP_BST CLKSHDN BDONE INPUTSEL Reserved S0ENABL S1ENABL BCLKDIV[3:0] S0RDEN S0PRNG S0CLKDIV[3:0] S1RDEN S1PRNG S1CLKDIV[3:0] 0xF0 0x3F 0x42 0x00 0xCA 0x03 0x00 0x00 0x40 MUDLY[8:1] 0x00 Reserved FSC[7:0] 0x00 0x00 Reserved PHZ_PD Default 0x0B 0x40 Reserved CLKP_CML[3:0] MUSAMP GAIN[1:0] Reserved Reserved Reserved Reserved Reserved BIST vector length BIST clock adjust Sign 0 control Sign 0 clock adjust Sign 1 control Sign 1 clock adjust RegFnl0Freq RegFnl1Freq BIST Signature 0 Bit 3 INC_DEC (Factory) PDBIAS MUDLY[0] Bit 4 FSC[9:8] AUTO_CAL PHZ_DET_BIAS[3:0] Reserved BENABLE BSTATUS[6:0] PADLEN[7:0] PADLEN[15:8] VECTLEN[7:0] VECTLEN[15:8] VECTLEN[23:16] S0ZERO S0NEG S1ZERO S1NEG Final Rate/Offset Control 0 [7:0] Final Rate/Offset Control 1 [7:0] SGN0[7:0] SGN0[15:8] SGN0[23:16] SGN1[7:0] SGN1[15:8] SGN1[23:16] Rev. A | Page 28 of 76 0x02 0x18 BCLKPHZ[3:0] S0FNLCH S0SEL[1:0] S0CLKPHZ[3:0] 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 S1FNLCH S1SEL[1:0] S1CLKPHZ[3:0] 0x00 0x00 BMODE[3:0] 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 AD9789 SPI REGISTER DESCRIPTIONS Table 12. SPI Control Register (Address 0x00) Bit 7 Bit Name SDIO_DIR 6 LSBFIRST 5 RESET 4 [3:0] LNG_INST Description This bit configures the SDIO pin as an input-only pin or as a bidirectional input/output pin. Both choices conform to the SPI standard. 0 = input only. 1 = bidirectional (input/output). This bit configures the SPI interface for MSB first or LSB first mode. Both choices conform to the SPI standard. 0 = MSB first. 1 = LSB first. When set to 1, this bit resets the part. After the part is reset, 0 is written to this bit on the next cycle. 0 = no reset. 1 = software reset. This bit sets the SPI to long instruction mode; 1 is the only valid value. These bits should mirror Bits[7:4]. Bit 3 should mirror Bit 4, Bit 2 should mirror Bit 5, Bit 1 should mirror Bit 6, and Bit 0 should mirror Bit 7. Table 13. Saturation Counter Register (Address 0x01) Bit [7:0] Bit Name SATCNT[7:0] Description This read-only register contains the saturation counter. This register reflects the number of samples at the output of the SUMSCALE gain block that overrange the datapath and are digitally clipped. The count is cleared by writing a 1 to Register 0x04, Bit 1. Table 14. Parity Counter Register (Address 0x02) Bit [7:0] Bit Name PARCNT[7:0] Description This read-only register contains the input data parity error counter. The count is cleared by writing a 1 to Register 0x04, Bit 7. Table 15. Interrupt Enable Register (Address 0x03) Bit 7 Name PARERR 6 BISTDONE 5 PARMSET 4 PARMCLR 3 LOCKACQ 2 LOCKLOST 1 SATERR 0 Reserved Description Setting this bit to 1 enables a PARERR flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 7 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a BISTDONE flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 6 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a PARMS_SET flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 5 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a PARMS_CLR flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 4 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a LOCKACQ flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 3 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a LOCKLOST flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 2 being set in Register 0x04 and the IRQ pin going low. Setting this bit to 1 enables a SATERR (overflow into 16× interpolator) flag to generate an interrupt request. Generating an interrupt request results in Interrupt Bit 1 being set in Register 0x04 and the IRQ pin going low. Reserved. Rev. A | Page 29 of 76 AD9789 Table 16. Interrupt Status/Clear Register (Address 0x04) Bit 7 6 5 Name PARERR BISTDONE PARMSET 4 PARMCLR 3 2 LOCKACQ LOCKLOST 1 SATERR 0 Reserved Description If this bit is set to 1, one or more parity errors has occurred. Writing a 1 to this bit clears the interrupt. If this bit is set to 1, the BIST has reached the terminal state. Writing a 1 to this bit clears the interrupt. If this bit is set to 1, the parameter update register (Address 0x24) has been updated. Writing a 1 to this bit clears the interrupt. If this bit is set to 1, the parameter update register (Address 0x24) has been cleared. Writing a 1 to this bit clears the interrupt. If this bit is set to 1, proper data handoff between the digital engine and the DAC core is occurring. If this bit is set to 1, proper data handoff between the digital engine and the DAC core has been lost. Writing a 1 to this bit clears the interrupt. If this bit is set to 1, one or more saturation errors (overflow into 16× interpolator) has occurred. Writing a 1 to this bit clears the interrupt. Reserved. Table 17. Channel Enable Register (Address 0x05) Bit [7:4] [3:0] Bit Name Reserved CHANEN[3:0] Description Reserved. A Logic 1 in any bit position enables the corresponding channel; 0000 means that all channels are disabled. Setting Channels Enabled 0000 All channels disabled. 0001 Channel 0 enabled. 0010 Channel 1 enabled. 0011 Channel 0 and Channel 1 enabled. … … 1110 Channel 1, Channel 2, and Channel 3 enabled. 1111 All channels enabled. Table 18. Bypass Register (Address 0x06) Bit 7 6 5 [4:0] Bit Name QAM SRRC Reserved INT[4:0] Description If this bit is set to 1, the QAM mappers are bypassed. If this bit is set to 1, the square root raised cosine (SRRC) filters are bypassed. Reserved. A Logic 1 in any bit position bypasses the corresponding interpolation filter. The preferred order for bypassing interpolation filters is to first bypass Filter 0, then Filter 1, and so on. Setting Interpolation Filters Bypassed 00000 All interpolation filters enabled. 00001 Interpolation Filter 0 bypassed. 00010 Interpolation Filter 1 bypassed. 00011 Interpolation Filter 0 and Interpolation Filter 1 bypassed. … … 01111 Interpolation Filter 0, Interpolation Filter 1, Interpolation Filter 2, and Interpolation Filter 3 bypassed. … … 11111 All interpolation filters bypassed. Rev. A | Page 30 of 76 AD9789 Table 19. QAM/SRRC Configuration Register (Address 0x07) Bit [7:6] [5:4] Bit Name Reserved ALPHA[1:0] 3 [2:0] Reserved MAPPING[2:0] Description Reserved. These bits set the SRRC filter alpha. Setting Alpha Filter 00 0.12 01 0.18 10 0.15 11 0.13 Reserved. These bits set the QAM encoding. Setting QAM Encoding 000 DOCSIS 64-QAM 001 DOCSIS 256-QAM 010 DVB-C 16-QAM 011 DVB-C 32-QAM 100 DVB-C 64-QAM 101 DVB-C 128-QAM 110 DVB-C 256-QAM 111 Unused Table 20. Summing Node Scalar Register (Address 0x08) Bit [7:0] Bit Name SUMSCALE[7:0] Description This register sets the value of the 2.6 multiplier that is applied to the output of the channel summing node. Setting 2.6 Multiplier 00000000 0 00000001 0.015625 00000010 0.03125 … … 00001101 0.203125 (default) … … 11111110 3.96875 11111111 3.984375 Table 21. Input Scalar Register (Address 0x09) Bit [7:0] Bit Name INSCALE[7:0] Description This register sets the value of the 3.5 multiplier that is applied to the input data. This scaling block is in parallel with the QAM encoder block and is used when the QAM encoder block is bypassed. Setting 3.5 Multiplier 00000000 0 00000001 0.03125 00000010 0.0625 … … 00100000 1 (default) … … 11111110 7.9375 11111111 7.96875 Rev. A | Page 31 of 76 AD9789 The three NCO 0 frequency tuning word registers together compose the 24-bit frequency tuning word for NCO 0. For more information about programming these registers, see the Baseband Digital Upconverter section. Table 22. NCO 0 Frequency Tuning Word Registers (Address 0x0A to Address 0x0C) Address 0x0A 0x0B 0x0C Bit Name FTW0[7:0] FTW0[15:8] FTW0[23:16] Description Frequency tuning word for NCO 0, Bits[7:0] Frequency tuning word for NCO 0, Bits[15:8] Frequency tuning word for NCO 0, Bits[23:16] The three NCO 1 frequency tuning word registers together compose the 24-bit frequency tuning word for NCO 1. For more information about programming these registers, see the Baseband Digital Upconverter section. Table 23. NCO 1 Frequency Tuning Word Registers (Address 0x0D to Address 0x0F) Address 0x0D 0x0E 0x0F Bit Name FTW1[7:0] FTW1[15:8] FTW1[23:16] Description Frequency tuning word for NCO 1, Bits[7:0] Frequency tuning word for NCO 1, Bits[15:8] Frequency tuning word for NCO 1, Bits[23:16] The three NCO 2 frequency tuning word registers together compose the 24-bit frequency tuning word for NCO 2. For more information about programming these registers, see the Baseband Digital Upconverter section. Table 24. NCO 2 Frequency Tuning Word Registers (Address 0x10 to Address 0x12) Address 0x10 0x11 0x12 Bit Name FTW2[7:0] FTW2[15:8] FTW2[23:16] Description Frequency tuning word for NCO 2, Bits[7:0] Frequency tuning word for NCO 2, Bits[15:8] Frequency tuning word for NCO 2, Bits[23:16] The three NCO 3 frequency tuning word registers together compose the 24-bit frequency tuning word for NCO 3. For more information about programming these registers, see the Baseband Digital Upconverter section. Table 25. NCO 3 Frequency Tuning Word Registers (Address 0x13 to Address 0x15) Address 0x13 0x14 0x15 Bit Name FTW3[7:0] FTW3[15:8] FTW3[23:16] Description Frequency tuning word for NCO 3, Bits[7:0] Frequency tuning word for NCO 3, Bits[15:8] Frequency tuning word for NCO 3, Bits[23:16] The three rate converter denominator (Q) registers together compose the 24-bit denominator for the rate converter decimation ratio. For more information about programming these registers, see the Sample Rate Converter section. Table 26. Rate Converter Denominator (Q) Registers (Address 0x16 to Address 0x18) Address 0x16 0x17 0x18 Bit Name Q[7:0] Q[15:8] Q[23:16] Description Rate converter denominator, Bits[7:0] Rate converter denominator, Bits[15:8] Rate converter denominator, Bits[23:16] The three rate converter numerator (P) registers together compose the 24-bit numerator for the rate converter decimation ratio. For more information about programming these registers, see the Sample Rate Converter section. Table 27. Rate Converter Numerator (P) Registers (Address 0x19 to Address 0x1B) Address 0x19 0x1A 0x1B Bit Name P[7:0] P[15:8] P[23:16] Description Rate converter numerator, Bits[7:0] Rate converter numerator, Bits[15:8] Rate converter numerator, Bits[23:16] Rev. A | Page 32 of 76 AD9789 The two interpolating BPF center frequency registers together compose the 16-bit center frequency of the 16× band-pass interpolation filter. For more information about programming these registers, see the Digital 16× Tunable Band-Pass Filter section. Table 28. Interpolating BPF Center Frequency Registers (Address 0x1C and Address 0x1D) Address 0x1C 0x1D Bit Name FC[7:0] FC[15:8] Description Center frequency, Bits[7:0] Center frequency, Bits[15:8] Table 29. Frequency Update Register (Address 0x1E) Bit 7 Name FREQNEW [6:0] Reserved Description Setting this bit to 1 updates the derived registers in the AD9789. This bit must be set for changes to Register 0x16 through Register 0x1D to take effect. This self-clearing bit is reset to 0 after the derived registers are updated. Reserved. Table 30. Hardware Version Register (Address 0x1F) Bit [7:4] [3:0] Name Reserved VER[3:0] Description Reserved. This read-only register indicates the version of the chip (0011). Table 31. Interface Configuration Register (Address 0x20) Bit 7 6 5 4 3 Bit Name CMOS_BUS CMOS_CTRL Reserved DCO_INV IF_MODE 2 CHANPRI [1:0] PAR[1:0] Description This bit reflects the state of the CMOS_BUS pin (L14). This bit reflects the state of the CMOS_CTRL pin (M14). Reserved. When set to 1, the DCO pin is inverted. This bit sets the data interface mode. 0 = channelizer mode. Supports all available interface widths and 8- and 16-bit word widths. Supports maximum fBAUD of fDAC/48. 1 = quadrature digital upconverter (QDUC) mode. Supports 32-bit interface, 16-bit word mode only. Supports maximum fBAUD of fDAC/16. This bit selects the channel prioritization value (used in channelizer mode only). 0 = device expects input samples only for those channels that are enabled. 1 = device expects data for all four channels. Data for disabled channels is expected and must be sent, but this data is discarded by the AD9789. These bits set the parity checking. For more information, see the Parity section. Setting Parity Checking 00 Parity checking deactivated 01 IQ parity (a value of 0 is expected on the I channel and a value of 1 is expected on the Q channel) 10 Even parity 11 Odd parity Rev. A | Page 33 of 76 AD9789 Table 32. Data Control Register (Address 0x21) Bit 7 Bit Name BIN [6:5] BUSWDTH[1:0] 4 DATWDTH 3 CMPLX [2:0] LTNCY[2:0] Description This bit selects the coding for the device. 0 = twos complement coding. 1 = straight binary coding. These bits set the input data bus width for the device. Setting Input Bus Width 00 4 bits 01 8 bits 10 16 bits 11 32 bits This bit sets the data-word width that is sent to the datapaths. 0 = 8-bit data-word. 1 = 16-bit data-word. This bit configures the datapath for real or complex data. 0 = real data. 1 = complex data. These bits set the turnaround latency from the FS pulse to the internal data sampling time. For more information, see the Latency Register section. Setting Latency 000 Input data begins to be sampled at approximately the first rising edge of DCO after FS goes low. 001 Input data begins to be sampled at approximately the second rising edge of DCO after FS goes low. … … 111 Input data begins to be sampled at approximately the eighth rising edge of DCO after FS goes low. Table 33. DCO Frequency Register (Address 0x22) Bit 7 [6:4] Bit Name Reserved DCODIV[2:0] [3:0] ONES[3:0] Description Reserved. These bits configure the data clock output (DCO) frequency. Setting DCO Clock Frequency 000 DCO clock disabled 001 fDACCLK/16 010 fDACCLK/32 011 Invalid 100 fDACCLK/64 101 Invalid 11x Invalid These bits always read back 1111. Rev. A | Page 34 of 76 AD9789 Table 34. Internal Clock Phase Adjust Register (Address 0x23) Bit [7:4] Bit Name DSCPHZ[3:0] [3:0] SNCPHZ[3:0] Description The data sampling clock (DSC) is an internal clock that is used to sample the input data. This clock can occur on 1 of 16 phases to optimize the setup and hold timing of the data interface. Setting Selected Phase 0000 Earliest clock phase 0001 Second earliest clock phase that occurs 1/16 of a DSC cycle later … … 1111 Last available clock phase The synchronization clock (SNC) is an internal clock that is used to synchronize the digital datapath clock with the DAC clock. This clock can occur on 1 of 16 phases to optimize the DAC-to-datapath timing. Setting Selected Phase 0000 Earliest clock phase 0001 Second earliest clock phase that occurs 1/16 of a DSC cycle later … … 1111 Last available clock phase Table 35. Parameter Update Register (Address 0x24) Bit 7 Name PARMNEW [6:0] Reserved Description This bit must transition from 0 to 1 for changes to Register 0x22 and Register 0x23 to take effect. Assuming that this bit was previously set to 0, writing a 1 to this bit causes the readback value of the bit to reflect the state of the chip. (The state of the chip is updated very quickly; for this reason, users with slow SPI implementations may never read back a 0 after an update.) 0 = values have not been updated. 1 = values have been updated. Reserved. Table 36. Channel Gain Registers (Address 0x25 to Address 0x28) Address 0x25 0x26 0x27 0x28 Register Name Channel 0 gain Channel 1 gain Channel 2 gain Channel 3 gain Bit Name CHAN0GAIN[7:0] CHAN1GAIN[7:0] CHAN2GAIN[7:0] CHAN3GAIN[7:0] Description These registers configure a value for the 1.7 multiplier applied to each individual channel just prior to the SUMSCALE block. The range of the channel gain is 0 to 1.9921875 with a step size of 0.0078125. To mute an individual channel, set the scale factor to 0. Setting Channel Gain 00000000 0 00000001 0.0078125 … … 11111111 1.9921875 Table 37. Spectrum Shaping Register (Address 0x29) Bit [7:1] 0 Name Reserved SPEC_INV Description Reserved. Setting this bit to 1 spectrally inverts the signal, effectively multiplying the Q data by −1. Rev. A | Page 35 of 76 AD9789 Table 38. Mu Delay Control 1 Register (Address 0x2F) Bit 7 Bit Name SEARCH_TOL 6 SEARCH_ERR 5 TRACK_ERR [4:0] GUARDBAND[4:0] Description This bit specifies the accuracy of the phase search. The optimal value for this bit is 1. 0 = not exact: the search can find a phase within two values of the desired phase. 1 = exact: the search finds the exact phase specified. This bit configures the search behavior when an error is encountered. 0 = stop on error. 1 = retry on error. This bit configures the track behavior if the controller does not find the desired phase. The optimal value for this bit is 0. 0 = continue on error. 1 = reset on error. These bits set the guard band value. The guard band is defined as follows: GUARDBAND[4:0] × 8 = number of mu delay codes of guard band from the endpoints If the search mode is alternating, the search proceeds in both directions until the guard band is reached in one direction. When the guard band is reached, the search continues only in the opposite direction. If the desired phase is not found before the guard band is reached in the second direction, the search reverts to the alternating mode and continues looking within the guard band. The search fails if the mu delay reaches the endpoints. For more information, see the Mu Delay Controller section. Setting Guard Band 00000 0 … … 01011 11 (default) … … 11111 31 Table 39. Mu Control Duty Cycle Register (Address 0x30) Bit 7 Bit Name Duty cycle correct enable 6 [5:0] INC_DEC MANUAL_ADJ[5:0] Description Setting this bit to 1 turns on the mu control duty cycle correction circuitry. Turn on this function before enabling the mu controller. Along with the phase comparator boost (enabled in Register 0x3E[5]), this function allows for more robust operation of the mu controller over the entire operating speed of the part. Reserved (factory use only). Reserved (factory use only). Table 40. Clock Receiver 1 Register (Address 0x31) Bit [7:4] Bit Name CLKN_CML[3:0] [3:0] Reserved Description These bits adjust the common-mode level at the CLKN pin. The recommended value for these bits and the CLKP_CML[3:0] bits is 0xF. For more information, see the Optimizing the Clock Common-Mode Voltage section. Reserved. Table 41. Clock Receiver 2 Register (Address 0x32) Bit 7 Bit Name CLK_DIS 6 5 Reserved PSIGN [4:1] CLKP_CML[3:0] 0 NSIGN Description This bit disables or enables the clock receiver. When the AD9789 powers up, this bit is set to 0 to prevent severe output noise that occurs on power-up with no clock. When the DAC clock is stable, set this bit to 1. 0 = disabled. 1 = enabled. Reserved (factory use only; leave at default value). This bit specifies the sign for the CLKP_CML bits. 0 = negative (recommended). 1 = positive. These bits adjust the common-mode level at the CLKP pin. The recommended value for these bits and the CLKN_CML[3:0] bits is 0xF. For more information, see the Optimizing the Clock Common-Mode Voltage section. This bit specifies the sign for the CLKN_CML bits. 0 = negative (recommended). 1 = positive. Rev. A | Page 36 of 76 AD9789 Table 42. Mu Delay Control 2 Register (Address 0x33) Bit 7 Bit Name MU_CLKDIS 6 SLOPE [5:4] MODE[1:0] 3 MUSAMP [2:1] GAIN[1:0] 0 MU_EN Description This bit disables or enables the clock to the mu delay controller. 0 = enabled. 1 = disabled. This bit configures the desired slope for the phase measurement of the mu delay. When the desired phase is measured, the slope of the phase measurement is calculated and compared to the value of this bit. For optimal ac performance, the best setting for the search is a positive slope and a phase value of 14. 0 = negative. 1 = positive. These bits configure the mode of operation for the mu controller. 00 = search and track (recommended). 01 = track only. 10 = search only. 11 = invalid. Transitioning this bit from 0 to 1 enables the user to read back the mu delay value that the controller locked to (the MUDLY bits in Register 0x39 and Register 0x3A), as well as the phase that it locked to (the MUPHZ bits in Register 0x39). 0 = no action. 1 = transition from 0 to 1 captures the readback of the mu controller phase and delay. These bits set the tracking rate of the mu controller. 00 = slowest tracking. 01 = nominal tracking (recommended). 10 = fastest tracking. 11 = invalid (do not use). This bit enables or disables the mu controller. Before enabling the mu controller, turn on both the phase comparator boost (Register 0x3E[5]) and the mu control duty cycle correction circuitry (Register 0x30[7]). Both of these functions allow for more robust operation of the mu controller over the entire operating speed of the part. 0 = mu controller off (manual mode). 1 = mu controller on (auto mode). Table 43. DAC Bias Register (Address 0x36) Bit 7 [6:2] [1:0] Bit Name PDBIAS Reserved MSEL[1:0] Description Setting this bit to 1 powers down the DAC circuitry. Reserved. These bits set the mirror roll-off frequency control, which can be used to adjust the noise contribution of the internal current mirror to optimize the 1/f noise. 00 = bypass the mirror roll-off frequency control. 01 = narrowest bandwidth. 10 = medium bandwidth. 11 = widest bandwidth. Table 44. DAC Decoder Register (Address 0x38) Bit [7:2] [1:0] Bit Name Reserved DAC decoder mode Description Reserved. These bits set the decoder mode for the DAC. It is recommended that normal mode (the default) be used. 00 = normal mode. 01 = return to zero mode. 10 = mix mode. 11 = invalid. Rev. A | Page 37 of 76 AD9789 Table 45. Mu Delay Control 3 Register (Address 0x39) Bit 7 Bit Name MUDLY[0] [6:5] SEARCH_DIR[1:0] [4:0] MUPHZ[4:0] Description This bit is the LSB of the mu delay value. Along with Bits[7:0] in Register 0x3A, this bit configures the programmable mu delay; the search algorithm begins at this specified mu delay value. In manual mode, the MUDLY bits can be written to. In tracking mode, the sampled MUDLY value can be read back. Even though there are 9 bits of resolution for this delay line value, the maximum allowable mu delay is 431 (0x1AF). The optimal point to begin the search is in the middle of the delay line, or approximately 216 (0xD8). These bits configure the search direction, starting at the selected mu delay value. 00 = search down. 01 = search up. 10 = search up and down (optimal). 11 = invalid. These bits specify the phase to be measured with the maximum allowable phase being 16 (10000). If a value larger than 16 is loaded, the controller will not lock. When the desired phase is measured, the slope of the phase measurement is calculated and compared to the configured slope, which is specified by the SLOPE bit in Register 0x33[6]. For optimal ac performance, the best setting for the search is for a positive slope and a phase value of 14 (01110). Table 46. Mu Delay Control 4 Register (Address 0x3A) Bit [7:0] Bit Name MUDLY[8:1] Description Along with Bit 7 in Register 0x39, these bits configure the programmable mu delay; the search algorithm begins at this specified mu delay value. In manual mode, the MUDLY bits can be written to. In tracking mode, the sampled MUDLY value can be read back. Even though there are 9 bits of resolution for this delay line value, the maximum allowable mu delay is 431 (0x1AF). The optimal point to begin the search is in the middle of the delay line, or approximately 216 (0xD8). Table 47. Full-Scale Current 1 Register (Address 0x3C) Bit [7:0] Bit Name FSC[7:0] Description Along with Bits[1:0] in Register 0x3D, this register sets the full-scale current for the DAC. For more information, see the Voltage Reference section. Setting (Includes Register 0x3D[1:0]) Full-Scale Current (mA) 0000000000 8.6 … … 1000000000 20 (default) … … 1011010000 25 … … 1111111111 32.1 Table 48. Full-Scale Current 2 Register (Address 0x3D) Bit [7:2] [1:0] Bit Name Reserved FSC[9:8] Description Reserved. Along with the FSC[7:0] bits in Register 0x3C, these bits set the full-scale current for the DAC. For more information, see Table 47 and the Voltage Reference section. Table 49. Phase Detector Control Register (Address 0x3E) Bit 7 6 5 4 [3:0] Bit Name PHZ_PD Reserved CMP_BST AUTO_CAL PHZ_DET_BIAS[3:0] Description Powers down the phase detector. This bit is for factory use only; this bit should be set to 0. Reserved. Comparator boost. This bit is for factory use only; this bit should always be set to 1. This bit is for factory use only; this bit should always be set to 1. These bits display the binary weighted current. Do not write to these bits (factory use only). Rev. A | Page 38 of 76 AD9789 THEORY OF OPERATION QAM Encoder The QAM encoder supports seven different standards-compliant mappings. (For illustrations of the supported mappings, see the QAM Constellation Maps section.) The QAM encoder receives input data-words of 8 bits in width and maps them into 16, 32, 64, 128, or 256 point constellations. It outputs 5-bit complex QAM modulated samples. The mode in which the QAM encoder runs is selected via the QAM/SRRC configuration register (Register 0x07[2:0]). 5 FROM INPUT INTERFACE CMOS 16 TO 31 LVDS FALL FS 16× INTERPOLATOR AND BPF + SCALARS QAM/ DATA FILTER/ NCO QAM/ DATA FILTER/ NCO SPI 5 I Q Table 50 lists the available QAM mapper modes along with the corresponding input bits and output range. The operation of the QAM encoder when configured in DOCSIS 64-QAM mode is described in this section. The operation of the QAM encoder in the other modes is conceptually the same; only the input data bit encoding and scale factors are different. QAM/ DATA FILTER/ NCO QAM/ DATA FILTER/ NCO QAM ENCODER Figure 72. QAM Encoder I/O 14-BIT 2.4GSPS DAC IRQ RS The DOCSIS 64-QAM constellation diagram is shown in Figure 73. The constellation diagram shows how the QAM encoder input is mapped into the QAM constellation. For example, an input data-word of 111111 maps to the constellation point in the upper right corner of the 64-QAM constellation. 07852-099 DCO CMOS 0 TO 15 LVDS RISE RETIMER DATA FORMATTER/ASSEMBLER 32 INPUT PINS AND 2 PARITY PINS 150MHz LVDS/CMOS Control of the AD9789 functions is via a serial peripheral interface (SPI). 8 07852-056 The AD9789 is a flexible digital signal processing (DSP) engine combined with a high performance, 2400 MSPS, 14-bit DAC (Figure 70). The DSP blocks include a QAM encoder, a 2× upsampling square root raised cosine (SRRC) filter, selectable interpolation from 16× to 512×, a rate converter, and a complex modulator. The digital interface can accept up to four channels of complex data. The QAM encoder supports constellation sizes of 16, 32, 64, 128, and 256. The on-chip rate converter allows fine resolution of baud rates with a fixed DAC sampling clock. The digital upconverters can place the input signals from dc to 0.5 × fDAC. An analog mix mode extends the output spectrum into the second and third DAC Nyquist zones. Figure 70. Top Level Functional Block Diagram DATAPATH SIGNAL PROCESSING The DSP blocks included on the AD9789 can be grouped into two sections. The first is the datapath signal processing. Four identical datapaths, or channels, can be used. A block diagram of a single channel is shown in Figure 71. Enabling and disabling each DSP block within the datapath takes effect on all channels. There is independent control of the scaling and the frequency placement of each channel. SRRC 110,111 111,011 010,111 011,011 100,101 101,111 110,101 111,111 110,100 111,000 010,100 011,000 100,000 101,010 110,000 111,010 100,111 101,011 000,111 001,011 000,101 001,111 010,101 011,111 100,100 101,000 000,100 001,000 000,000 001,010 010,000 011,010 2 2N (N = 0 TO 5) INSCALE C5 C4 C3, C2 C1 C0 24-BIT NCO 0 TO fDAC /16 RATE CONVERTER P/Q 24-BIT I 010,011 011,001 000,011 001,001 000,001 001,101 100,001 101,101 (P/Q = 0.5 TO 1) BYPASS QAM BYPASS SRRC CH GAIN 0× TO 2× 07852-129 QAM MAPPER Q Figure 71. Datapath Block Diagram 010,110 011,100 000,110 001,100 000,010 001,110 100,010 101,110 110,011 111,001 100,011 101,001 010,001 011,101 110,001 111,101 The following sections describe each of the DSP blocks included in the datapath. 07852-057 110,110 111,100 100,110 101,100 010,010 011,110 110,010 111,110 Figure 73. DOCSIS 64-QAM Constellation Rev. A | Page 39 of 76 AD9789 Table 50. QAM Mapper Input and Output Range vs. Mode ITU-T J.83 Annex B B A A A and C A and C A and C Bit Range at Output −14 to +14 −15 to +15 −15 to +15 −15 to +15 −14 to +14 −11 to +11 −15 to +15 Input Bits B7 B6 B5 B4 B3 B2 B1 B01 X X C5 C4 C3 C2 C1 C0 C7 C6 C5 C4 C3 C2 C1 C0 X X X X C3 C2 C1 C0 X X X C4 C3 C2 C1 C0 X X C5 C4 C3 C2 C1 C0 X C6 C5 C4 C3 C2 C1 C0 C7 C6 C5 C4 C3 C2 C1 C0 X = don’t care. Each constellation point corresponds to an I and Q coordinate pair, as shown in Figure 74. In the figure, two symbols are highlighted in a 64-QAM constellation: I = 14, Q = 14 (Pair 1) and I = 6, Q = −10 (Pair 2). To represent the I and Q coordinate points, 5-bit, twos complement numbers are used. For example, an input of 011101 into the QAM encoder maps to the I = 6, Q = −10 position of the QAM-64 constellation and results in output samples of I = 00110, Q = 10110. SYMBOL I = 14, Q = 14 I = 01110, Q = 01110 Q 14 Input Scalar The input scalar block is active only when the QAM mapper is bypassed. The value of INSCALE[7:0] is programmed in Register 0x09[7:0]. The scale factor applied to the input data is calculated as follows: ScaleFactor INSCALE[7:0] 32 This factor provides a scaling range of the input data from 0 to 7.96875 in steps of 0.03125. The default value of 0x20 provides a scale factor of 1. As shown in Figure 76, the output of the input scalar block is rounded to the nearest 16-bit value. If the output exceeds the maximum or minimum value, it is clipped to either positive or negative full scale (0x7FFF or 0x8000). 10 ROUND SATURATE 6 8 INSCALE 2 –14 –10 –6 –2 2 6 10 14 Figure 76. Input Scalar Block Diagram I –2 SRRC Filter The square root raised cosine (SRRC) filter performs a 2× interpolation and filtering operation on the input data. The SRRC filter has a pass band, transition band, and stop band requirement as per the DOCSIS, Euro-DOCSIS, and DVB-C standards. –6 –10 07852-058 –14 SYMBOL I = 6, Q = –10 I = 00110, Q = 10110 Figure 74. I and Q Symbol Mapping 8 16 QAM MAPPER X 5 The SRRC filter accepts only five bits at its input and can be bypassed (Register 0x06[6]). If the SRRC filter is the first block enabled in the datapath, these five bits are the five MSBs of the 8-bit data-word. SRRC 5 16 16 16 2 To cover all the standards, the value of alpha can be set to 0.12, 0.13, 0.15, or 0.18. This value is programmed in Register 0x07[5:4]. The frequency, fN, is determined by the input data baud rate. The response of the SRRC filter is illustrated in Figure 77. BYPASS QAM BYPASS SRRC 07852-059 16 INSCALE 07852-100 1 SPI Register 0x07, MAPPING[2:0] Bits 000 001 010 011 100 101 110 111 Description DOCSIS 64-QAM DOCSIS 256-QAM DVB-C 16-QAM DVB-C 32-QAM DVB-C 64-QAM DVB-C 128-QAM DVB-C 256-QAM Unused Figure 75. QAM Mapper and SRRC Filter Detail (I and Q Paths Are Identical So Only One Is Shown) Rev. A | Page 40 of 76 AD9789 10 0 –10 <0.4dB FREQUENCY –20 MAGNITUDE (dB) <0.4dB –3.01dB <–43dB –50 –60 (1 +α) fN –80 –90 Figure 77. SRRC Filter Characteristics –100 –2.0 If the SRRC filter is used, at least four of the 2× interpolation filters must be enabled. The reason for this is that the SRRC filter requires a minimum of 12 clock cycles at the fDAC/16 rate per sample to function properly. –1.5 –1.0 –0.5 0 0.5 1.0 FREQUENCY × fINPUT (Hz) 1.5 2.0 07852-103 fN Figure 80. 2× Half-Band Interpolation Filter 1 Response 10 Half-Band Interpolation Filters 0 –10 –20 MAGNITUDE (dB) The AD9789 can provide from 1× to 32× interpolation through the datapath using five bypassable half-band interpolation filters. The half-band interpolation filters are controlled via Register 0x06[4:0]. The preferred order in terms of power savings for bypassing these filters is to bypass Filter 0 first, then Filter 1, and so on. The frequency response of the low-pass filters is shown in Figure 79 through Figure 82. All of the filters have a pass band of 0.8 × fINPUT, where fINPUT is the data rate at the input of each filter. The pass band is flat to within 0.01 dB for all filters. The stop-band attenuation exceeds 85 dB in Filter 0, Filter 1, and Filter 2, and 75 dB in Filter 3 and Filter 4. 2 –40 –70 07852-060 (1 –α) fN –30 –30 –40 –50 –60 –70 –80 –90 –100 –2.0 –1.5 –1.0 0 –0.5 0 0.5 1.0 FREQUENCY × fINPUT (Hz) 1.5 2.0 07852-104 0dB Figure 81. 2× Half-Band Interpolation Filter 2 Response 1 07852-101 10 BYPASS REGISTER 0x06[4:0] 0 –10 Figure 78. Conceptual Block Diagram of 2× Half-Band Interpolation Filters 0 –10 –20 –30 –40 –50 –60 –80 –50 –100 –1.875 –70 –80 –1.250 –0.625 0 0.625 FREQUENCY × fINPUT (Hz) 1.250 1.875 07852-105 –90 –60 Figure 82. 2× Half-Band Interpolation Filter 3 and Filter 4 Response –90 –100 –2.0 –30 –70 –40 –1.5 –1.0 –0.5 0 0.5 1.0 FREQUENCY × fINPUT (Hz) 1.5 2.0 07852-102 MAGNITUDE (dB) –20 MAGNITUDE (dB) 10 Figure 79. 2× Half-Band Interpolation Filter 0 Response Rev. A | Page 41 of 76 AD9789 Sample Rate Converter Example The purpose of the sample rate converter (SRC) is to provide increased flexibility in the ratio of the input baud rate to the DAC update rate. Each of the four channelization datapaths contains a sample rate converter (SRC) that provides a data rate conversion in the range of 0.5 to 1.0 inclusive. The rate conversion factor is set by the ratio of two 24-bit values, P and Q. Figure 83 is a conceptual block diagram of the SRC. It can be thought of as an interpolation block, followed by filtering and decimation blocks. A DOCSIS application has a master system clock that runs at a frequency of fMASTER. Several channel baud rates are supported, all of which are fractions of the master clock and can be represented by the following equation: Q 24 P 24 Q 07852-106 P f BAUD = f DAC = 224 × f MASTER = 2293.76 MHz The values of P and Q are set by programming the P[23:0] and Q[23:0] registers at Address 0x16 through Address 0x1B. Denominator (Q) Register 0x18 Register 0x17 Register 0x16 224 × f MASTER = I × P × 16 × f BAUD Q (1) P N 7 = × Q M 16 f BAUD = P ≤ 1.0 Q Q[23] = 1 (7) 401 × 10.24 MHz = 5.0569 MHz 812 (8) P and Q can then be calculated from the numerator and denominator of Equation 9. P 812 7 5684 0 x1634 = × = = Q 401 16 6416 0 x1910 The values of P and Q must be selected within the following constraints: 0.5 ≤ (6) Recall that N and M are given by the required baud rate. For example, assume a baud rate of 5.0569 MHz, which results from M = 401 and N = 812. where I is the total interpolation ratio of the SRRC filter and the five half-band interpolation filters. If Equation 1 is satisfied, the long-term baud rate, fBAUD, is exactly maintained. No residual frequency offset errors are introduced by the rate conversion process. P M × 16 × × f MASTER Q N Enabling the SRRC filter and four of the half-band interpolation filters would result in the total interpolation factor, I, being equal to 32. Substituting 32 for I and simplifying Equation 6 results in Equation 7. The values of P and Q should be selected to satisfy the following equation for the desired baud rate (fBAUD) and DAC clock frequency (fDAC). f DAC = I × (5) Inserting Equation 4 and Equation 5 into Equation 1 results in Equation 6. Table 51. Register Locations for Sample Rate Converter Numerator (P) Register 0x1B Register 0x1A Register 0x19 (4) Equation 1 must be satisfied for fBAUD to be exactly maintained. To facilitate this, the DAC sampling frequency is selected to be a multiple of fMASTER that satisfies the signal bandwidth and output frequency requirements. For fMASTER = 10.24 MHz, a signal bandwidth requirement of 32 MHz or greater, and a supported output frequency band of up to 1 GHz, the following DAC sampling frequency can be selected: Figure 83. Conceptual Block Diagram of the Sample Rate Converter Bits [23:16] (Byte 2) [15:8] (Byte 1) [7:0] (Byte 0) M × f MASTER N (2) (3) Equation 3 states that the value of Q must be shifted so that the MSB of Q is set. In most systems, the baud rate is a given, and the DAC sample rate is selected so that it is high enough to support the signal bandwidth and output frequency requirements. In many cases, it is desirable to set the DAC clock rate to a multiple of a system clock rate. The following example shows how P and Q can be selected in such a system. (9) Because the value of Q must be MSB justified, both numbers can be shifted by 11 bits, resulting in the final P and Q values of 0xB1A000 and 0xC80000, respectively. Baseband Digital Upconverter The digital upconverter enables each baseband channel to be placed anywhere from dc to fDAC/16. The center frequency for each of the four channels is register programmable through the 24-bit frequency tuning words, FTW 0 through FTW 3. For the desired center frequency of each individual channel, the FTW can be calculated as follows: Rev. A | Page 42 of 76 FTW = f CENTER × (2 24 − 1) ⎛ f DAC ⎞ ⎜ ⎟ ⎝ 16 ⎠ AD9789 Channel 0 Reg. 0x0C Reg. 0x0B Reg 0x0A Channel 1 Reg. 0x0F Reg. 0x0E Reg 0x0D Channel 2 Reg. 0x12 Reg. 0x11 Reg 0x10 Channel 3 Reg. 0x15 Reg. 0x14 Reg 0x13 DATAPATH 0 The FTW sets the frequency of the sine and cosine signals generated by the numerically controlled oscillator (NCO). The complex output from the NCO is multiplied by the input datapath signal to modulate the signal to the desired output frequency. A conceptual block diagram of the baseband digital upconverter is shown in Figure 84. Individual Channel Scalar The last block in the datapath is an 8-bit scalar (Register 0x25 to Register 0x28) intended for compensating out any sampling and hardware roll-offs that may be encountered. The scale factor applied to each channel is calculated as follows: CHANxGAIN[7 :0] 128 The range of the channel gain is 0 to 1.9921875 with a step size of 0.0078125. An individual channel can be easily and quickly muted, if desired, by setting the scale factor to 0. Table 53. Register Locations for Channel Gain Scalar CHANxGAIN [7:0] Channel 0 Reg. 0x25 Channel 1 Reg. 0x26 DATAPATH 3 Channel 2 Reg. 0x27 Channel 3 Reg. 0x28 ScaleFactor = fC SUMSCALE[7 :0] 64 This factor provides a scaling range of the input data from 0 to 3.984375 with a step size of 0.015625. The default value of 0x0D provides a scale factor of 0.203125. Note that when the channels are summed, they are clipped at the output of the summing junction scalar block if the value exceeds the maximum or minimum full-scale value (0x7FFF or 0x8000). If the full 16-bit range of each individual channel is used, the sum scalar should be set to 0x10 (0.25) to avoid the possibility of clipping. TO SATURATION COUNTER SATERR REGISTER 0x03[1] ROUND SATURATE 8 ROUND SATURATE SUMSCALE REGISTER 0x08 07852-108 CHANxGAIN[7:0] BPF The summing junction scalar block operates on the sum of the four channels. The value of SUMSCALE[7:0] is programmed in Register 0x08. The scale factor applied to the data is calculated as follows: The default value of the channel gain provides a scale factor of 1. As shown in Figure 85, the output of the input scalar block is rounded to the nearest 16-bit value. If the output exceeds the maximum or minimum value, it is clipped to either positive or negative full scale (0x7FFF or 0x8000). 8 SUM SCALE BPF fC = 0 TO fDAC/2 Summing Junction Scalar Figure 84. Conceptual Block Diagram of the Baseband Digital Upconverter ScaleFactor = DATAPATH 2 Each block of the digital block upconverter is described in more detail in the following sections. 07852-107 FTW NCO FREQUENCY TUNING WORD DATAPATH 1 Figure 86. Functional Block Diagram of the Digital Block Upconverter SIN COS 24 DIGITAL BLOCK UPCONVERTER Figure 85. Individual Channel Gain Control 07852-110 FTW [23:16] [15:8] [7:0] The second half of the DSP engine on the AD9789 combines the outputs of the four datapaths into one block, scales the block of channels, interpolates by 16× to the full DAC rate, and performs a band-pass filter operation allowing the block of channels to be placed anywhere in the Nyquist bandwidth of the DAC. 07852-109 Table 52. Register Locations of FTWs for Each Channel DIGITAL BLOCK UPCONVERTER 16× INTERPOLATOR The calculated FTW for each channel should be entered into the register locations listed in Table 52. Figure 87. Block Diagram of the Summing Junction Scalar In practice, the signal-to-noise ratio (SNR) of the channel can be improved by increasing the sum scale factor and permitting a small amount of clipping. The larger signal amplitude can improve the SNR if the clipping is brief and infrequent. Rev. A | Page 43 of 76 AD9789 should be taken to appropriately filter the desired signal with the interpolation filters prior to the input of the BPF.. 0 –20 MAGNITUDE (dB) Table 54 shows recommended sum scale values for each QAM mapper mode. The criteria used to determine the recommended sum scale values were MER/EVM measurements and spectral purity. Because clipping results in impulsive noise, it can be observed in the output spectrum as a transient increase in the output noise floor. These sum scale values were chosen such that the transient increases in the noise floor were minimal. These tests were completed for one, two, three, and four carrier outputs at approximately 850 MHz. Because clipping can occur in the RF chain following the DAC, further verification of these values should be performed at the system level by adding BER tests to the sum scale selection criteria. –40 –60 –80 Table 54. Recommended Sum Scale Values for all QAM Mapper Modes and Channel Count 54 34 26 20 54 34 26 20 80 50 38 30 54 34 26 20 54 34 26 20 54 34 26 20 0.5 1.0 1.5 FREQUENCY (GHz) 07852-062 1 Channel 48 0 2.0 Figure 89. Band-Pass Filter Response at 200 MHz, fDAC = 2.4 GHz 0 –20 MAGNITUDE (dB) –40 –60 –80 Digital 16× Tunable Band-Pass Filter –100 The digital band-pass filter works in conjunction with a fixed 16× interpolator (see Figure 88). The 16× interpolation filter creates 16 images of the baseband signal in the Nyquist band of the DAC. The digital band-pass filter must then be tuned to reject the 15 undesired images. The center frequency of the band-pass filter can be placed anywhere from dc to fDAC/2. The tuning word for the band-pass filter center frequency can be calculated as follows: f CENTER 216 1 f DAC The resulting tuning word is a 16-bit value where the most significant byte is written to Register 0x1D[7:0] and the least significant byte is written to Register 0x1C[7:0]. 0.5 1.0 1.5 FREQUENCY (GHz) 2.0 Figure 90. Band-Pass Filter Response at 1 GHz, fDAC = 2.4 GHz 0 –2 MAGNITUDE (dB) BPF _ Center _ Freq 0 07852-063 QAM Mode DVB-C 16-QAM DVB-C 32-QAM DVB-C 64-QAM DVB-C 128-QAM DVB-C 256-QAM DOCSIS 64-QAM DOCSIS 256-QAM –100 Sum Scale Value (Decimal) 2 Channels 3 Channels 4 Channels 28 22 16 –4 –6 –8 –10 fC 0 07852-111 16 20 40 FREQUENCY (MHz) 60 80 Figure 91. Band-Pass Filter Pass-Band Detail, fDAC = 2.4 GHz Figure 88. Conceptual Block Diagram of 16× Tunable Band-Pass Filter The width of the filter’s stop band is fixed at approximately fDAC/64. The effective FLAT pass band is fDAC/64. As can be inferred from Figure 89 to Figure 91, mistuning of the BPF center frequency can result in unwanted images appearing. Care Rev. A | Page 44 of 76 07852-064 16 AD9789 fDAC DIGITAL INTERFACE MODES In QDUC mode (Register 0x20[3] = 1), the interface is fixed at a 32-bit bus width and one channel of complex data. The available signal processing methods are interpolation (16× to 512×), rate conversion (0.5 to 1.0), and complex modulation. The maximum baud rate supported in QDUC mode is fDAC/16. In both channelizer and QDUC modes, the input data bus can be configured to accept LVDS or CMOS data via the CMOS_BUS pin (L14). If CMOS_BUS is pulled to 3.3 V, the data bus is configured to accept CMOS inputs (D[31:0], P0, and P1). If CMOS_BUS is pulled to 0 V, the bus is configured to accept LVDS inputs (D[15:0]P, D[15:0]N, PARP, and PARN). Two output signals are used to source data into the AD9789. The first is the data clock output signal (DCO), which is provided to clock data from the digital data source. DCO is a divided-down version of DACCLK. The second is the frame sync signal (FS), which is provided to request a new data-word. The average frequency of the FS signal is equal to the symbol rate or baud rate of the data. As with the input data bus, the DCO and FS signals can be configured as LVDS or CMOS outputs via the CMOS_CTRL pin (M14). If CMOS_CTRL is pulled to 3.3 V, DCO and FS are output as CMOS signals on the P14 and N14 pins (CMOS_DCO and CMOS_FS), respectively. If CMOS_CTRL is pulled to 0 V, DCO and FS are output as LVDS signals on the N13, P13, L13, and M13 pins (DCOP, DCON, FSP, and FSN), respectively. Channelizer Mode In channelizer mode, the digital interface has programmable bus width, data width, and data format. The bus width, which is the physical width of the digital data bus at the input of the AD9789, can be set to a 4-, 8-, 16-, or 32-bit wide interface. The data width, which is the internal width of the data at the input to the digital datapath, can be set to an 8-bit or 16-bit word. The data format can be programmed for real or complex data. A list of supported interface modes is shown in Table 55. FS DCO CLK CTL DATAUP TO 32 BITS PATH 2 UP TO DATA32 BITS PATH 3 fDAC fDAC DATAUP TO 32 BITS PATH 1 16 32 SUM SCALE 16× INTERPOLATOR CMOS 16 TO 31 LVDS FALL DATAUP TO PATH 32 BITS 0 BPF fC = 0 TO fDAC/2 BPF fC Figure 92. Channelizer Mode Table 55. Interface Configurations Supported in Channelizer Mode First Input Block Enabled QAM Encoder SRRC Filter Interpolation Filter Bus Width Reg. 0x21[6:5] 32 bits 16 bits 8 bits 4 bits 32 bits 16 bits 8 bits 4 bits 32 bits 16 bits 8 bits Data Width Reg. 0x21[4] 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits 16 bits 16 bits 16 bits Data Format Reg. 0x21[3] Real Real Real Real Complex Complex Complex Complex Complex Complex Complex If the QAM encoder is the first block enabled in the datapath, the data width should be set to an 8-bit word and real data format. If the SRRC filter is the first block enabled in the datapath, the data width should be set to an 8-bit word and complex data format. If both the QAM encoder and the SRRC filters are bypassed, the data width should be set to a 16-bit word and complex data format. Pin Mapping in Channelizer Mode In CMOS mode (CMOS_BUS and CMOS_CTRL pins = 3.3 V), the various interface width options are mapped to the AD9789 input pins as shown in Table 56. Table 56. CMOS Pin Assignments for Various Interface Widths Interface Width 4 bits 8 bits 16 bits 32 bits Rev. A | Page 45 of 76 Pin Assignments D[3:0] D[7:0] D[15:0] D[31:0] BUSWDTH[1:0] 00 01 10 11 07852-112 P0 P1 RETIMER In channelizer mode (Register 0x20[3] = 0), the interface can be configured for 4- to 32-bit bus widths and can accept up to four channels of complex data. Any of the signal processing blocks in the digital datapath can be used. The maximum baud rate supported in channelizer mode is fDAC/48. LVDS/CMOS Channelizer mode Quadrature digital upconverter (QDUC ) mode 32 INPUT PINS • • CMOS 0 TO 15 LVDS RISE PROGRAMMABLE DATA FORMATTER/ASSEMBLER The AD9789 can be configured for two main digital interface modes of operation: 4 TO 32 BITS 16 TO 1024 AD9789 In LVDS mode, the various interface width options are mapped to the AD9789 input pins as shown in Table 57. When the interface width is set to 32 bits in LVDS mode, the interface becomes double data rate (DDR). In DDR mode, the first 16 bits are sampled on the rising edge of the data sampling clock (DSC, which is synchronous to DCO), and the second 16 bits are sampled on the falling edge of DSC. All other interface widths are single data rate (SDR), where the input data is sampled on the falling edge of DSC. Table 57. LVDS Pin Assignments for Various Interface Widths Interface Width 4 bits 8 bits 16 bits 32 bits Pin Assignments D[3:0]P, D[3:0]N D[7:0]P, D[7:0]N D[15:0]P, D[15:0]N D[15:0]P, D[15:0]N rising edge and falling edge BUSWDTH[1:0] 00 01 10 11 In nibble or byte loading, the most significant nibble or byte should be loaded first. Data for Channel 0 should be loaded first followed by Channel 1, Channel 2, and Channel 3. In complex data format, the in-phase part should be loaded before the quadrature part of the data-word. The data bus is LSB justified when the data for each channel is assembled internally. A few examples of how the interface maps for different configurations follow. For more information on how a particular configuration is mapped, see the Channelizer Mode Pin Mapping for CMOS and LVDS section. Example 1 For a CMOS interface with a 32-bit bus width, 8-bit data width, real data format, and four channels enabled, the data in Table 58 is expected on the input port after data is requested. Table 58. CMOS Pin Mapping for Bus Width = 32 Bits, Data Width = 8 Bits, Data Format = Real, Four Channels1 DCO 1 1 D[31:24] R3 D[23:16] R2 D[15:8] R1 D[7:0] R0 R represents the real data loaded to a given channel; the channel number follows R. Example 2 For a CMOS interface with a 32-bit bus width, 8-bit data width, complex data format, and four channels enabled, the data in Table 59 is expected on the input port after data is requested. Table 59. CMOS Pin Mapping for Bus Width = 32 Bits, Data Width = 8 Bits, Data Format = Complex, Four Channels1 DCO 1 2 1 D[31:24] Q1 Q3 D[23:16] I1 I3 D[15:8] Q0 Q2 D[7:0] I0 I2 I represents the in-phase term and Q represents the quadrature term of the complex data loaded to a given channel; the channel number follows I or Q. Example 3 For an LVDS interface with a 16-bit bus width, 8-bit data width, complex data format, and four channels enabled, the data in Table 60 is expected on the input port after data is requested. Table 60. LVDS Pin Mapping for Bus Width = 16 Bits, Data Width = 8 Bits, Data Format = Complex, Four Channels1 DCO 1 2 3 4 1 D[15:8]P, D[15:8]N Q0 Q1 Q2 Q3 D[7:0]P, D[7:0]N I0 I1 I2 I3 I represents the in-phase term and Q represents the quadrature term of the complex data loaded to a given channel; the channel number follows I or Q. Example 4 For an LVDS interface with a 32-bit bus width, 8-bit data width, complex data format, and four channels enabled, the data in Table 61 is expected on the input port after data is requested. Table 61. LVDS Pin Mapping for Bus Width = 32 Bits, Data Width = 8 Bits, Data Format = Complex, Four Channels1 DCO2 1 rise 1 fall 2 rise 2 fall D[15:8]P, D[15:8]N Q0 Q1 Q2 Q3 D[7:0]P, D[7:0]N I0 I1 I2 I3 1 I represents the in-phase term and Q represents the quadrature term of the complex data loaded to a given channel; the channel number follows I or Q. 2 “Rise” means that the data is sourced on the rising edge of DCOx; “fall” means that the data is sourced on the falling edge of DCOx. DCO and FS Rates in Channelizer Mode The DCO signal is a data clock output provided to clock data from the digital data source. The DCO is a divided version of the DAC clock. The FS signal is an output provided to request a new data-word. The average frequency of the FS signal (fFS) is exactly equal to the symbol rate or baud rate (fBAUD) of the data. FS is intended as a request line; timing should be taken from the DCO. The frequencies of the DCO signal (fDCO), the baud rate (fBAUD), and the DAC clock (fDAC) are related as shown by the following two equations: f DAC = I × P × 16 × f BAUD Q fDCO = fDAC /(16 × N) (1) (2) where: I is the interpolation factor, which can range from 1 to 64. P/Q is the rate conversion factor (0.5 to 1.0, inclusive). N is a programmable DCO divide factor set using the DCODIV[2:0] bits in Register 0x22[6:4]. Set DCODIV[2:0] to 1, 2, or 4. A value of 0 disables the DCO. A DCODIV value of 3 is not functional. The frequency of the DSC signal is always equal to DCO. Rev. A | Page 46 of 76 AD9789 Before choosing an interface configuration, divide the frequency of DCO by the highest frequency baud rate that will be used in the system and truncate it. The result is the number of available DCO cycles (cyclesAVAIL) between FS pulses. ⎞ ⎟ ⎟ ⎠ An example of channel prioritization set to 0 is shown in Table 62. In this example, the data interface is configured for CMOS with 32-bit bus width, 8-bit data width, and real data format. Each interface configuration requires a particular number of DCO cycles between FS pulses to successfully load data into all channels. This number can be calculated using the following formula: DW BW where: N is the number of channels enabled (1 to 4). N is always equal to 4 if channel prioritization is set to 1 (see the Channel Prioritization section). F represents the data format. If the data format is real, F = 1; if the data format is complex, F = 2. DW is the data width in number of bits (8 or 16). BW is the bus width in number of bits (4, 8, 16, or 32). For a successful interface design, the number of DCO cycles between FS pulses must be greater than the number of DCO cycles required by the interface. Design Example In this example, a system has the baud rate fFS = 6.4 MHz. If a 4-bit-wide interface is desired for four channels with real data format and a data width of 8 bits, the selected fDCO should be at least 8 × fFS. First, using Equation 1 and Equation 2, evaluate the interface speed with N = 1, P/Q = 0.7, and I = 32. Channels 4 Channels Enabled Channel 0 Disabled Channel 0, Channel 2 Disabled [D31:D24] Channel 3 CMOS Bit Mapping [D23:D16] [D15:D8] Channel 2 Channel 1 [D7:D0] Channel 0 Channel 3 Channel 2 Channel 1 Channel 3 Channel 1 The same example behaves differently when channel prioritization is set to 1, as shown in Table 63. Table 63. Input Mapping vs. Enabled Channels, Channel Prioritization = 1 Channels 4 Channels Enabled Channel 0 Disabled Channel 0, Channel 2 Disabled [D31:D24] Channel 3 CMOS Bit Mapping [D23:D16] [D15:D8] Channel 2 Channel 1 Channel 3 Channel 2 [D7:D0] Channel 0 Channel 1 Channel 3 Channel 1 Quadrature Digital Upconverter (QDUC) Mode fDAC = 32 × 0.7 × 16 × 6.4 MHz = 2293.76 MHz fDCO = 2293.76 MHz/(16 × 1) = 143.36 MHz The fDCO/fBAUD ratio = 22.4. If a value of N = 2 is selected, the number of available DCO cycles is reduced to 11; this option may not be feasible when the latency values are taken into account. See the Latency Effects on Channelizer Mode section for more information about latency. In QDUC mode (Register 0x20[3] = 1), the data interface is fixed at a 32-bit bus width, 16-bit data width, and complex data format. In QDUC mode, only one channel should be enabled. If more than one channel is enabled, identical I and Q data is sent to each enabled channel. Within the datapath, the QAM mapper and the SRRC filter must be bypassed (Register 0x06[7:6] = 11). Channel Prioritization 32 INPUT PINS P0 P1 FS DCO Rev. A | Page 47 of 76 LVDS/CMOS When channels are enabled and disabled, the input interface mapping can be affected. If channel prioritization (Register 0x20[2]) is set to 0, the device expects input samples for only the channels that are enabled. In this configuration, the physical channel mapping at the DUT input can move around based on the number of channels enabled, where Channel 0 has highest priority (it never moves location when enabled). If channel prioritization is set to 1, data is expected for all four channels but the data is ignored internally if the channel is disabled. This method is recommended because enabling and disabling channels does not shift the input data bus. CMOS 0 TO 15 LVDS RISE I CMOS 16 TO 31 Q LVDS FALL I AND Q 16 BITS ON I AND Q 16 BITS OFF I AND Q 16 BITS OFF I AND Q 16 BITS OFF BPF fC = 0 TO fDAC /2 BPF Figure 93. QDUC Mode fC 07852-069 cycles INTERFACE = N × F × Table 62. Input Mapping vs. Enabled Channels, Channel Prioritization = 0 16× INTERPOLATOR ⎛ f DCO cycles AVAIL = floor ⎜⎜ ⎝ max f BAUD If the number of channels enabled is always less than four and the user does not plan to enable and disable channels dynamically, setting channel prioritization to 0 is the best choice because fewer clocks and/or pins are required to transfer the input data. AD9789 Pin Mapping in QDUC Mode 1 2 3 4 5 6 7 8 9 10 11 12 13 14 A In CMOS mode, the AD9789 input pins are mapped as shown in Table 64. B C Table 64. Pin Mapping in QDUC Mode for CMOS Interface D Data Bit D31 D16 D15 D0 P1 P0 E 2 3 4 5 6 7 8 9 10 11 Pin No. L5 P8 L9 P12 L4 M4 F G H PARP J PARN K 12 13 14 A FSP L P+ 15 13 11 9 7 5 3 1 FS FSN M P– 15 13 11 9 7 5 3 1 FS DCOP N 14 12 10 8 6 4 2 0 DC DCON P 14 12 10 8 6 4 2 0 DC B 14 +LVDS C Figure 95. LVDS Data Input Pin Mapping D DCO and FS Rates in QDUC Mode E In QDUC mode, DCODIV should always be set to 1 (Register 0x22[6:4] = 001). The clock period of DCO is equal to 16 DAC clock periods. When only 16× interpolation is required and the rate converter is not used, the data rate of the interface is equal to fDCO. F G H J K L P1 31 27 23 19 15 11 7 3 BU CMOS_BUS M P0 30 26 22 18 14 10 6 2 CT CMOS_CTRL N 29 25 21 17 13 9 5 1 FS CMOS_FS P 28 24 20 16 12 8 4 0 DC CMOS_DCO If further interpolation or rate conversion is enabled in the datapath, the data rate of the interface is fBAUD. The average rate of FS, fFS, is equal to the baud rate, fBAUD. The baud rate can be specified by the following equation: 07852-113 D[31:0] CMOS DATA INPUTS PARITY AND CONTROL INPUTS f BAUD = Figure 94. CMOS Data Input Pin Mapping In LVDS mode, the AD9789 input pins are mapped as shown in Table 65. Table 65. Pin Mapping in QDUC Mode for LVDS Interface1 Data Bit D15P, D15N rising D0P, D0N rising D15P, D15N falling D0P, D0N falling PARP, PARN rising PARP, PARN falling 1 14 –LVDS 07852-114 1 Description MSB of I data LSB of I data MSB of Q data LSB of Q data Parity for D[31:16] Parity for D[15:0] Description MSB of I data LSB of I data MSB of Q data LSB of Q data Parity for D[15:0]P, D[15:0]N rising Parity for D[15:0]P, D[15:0]N falling Pin No. L5, M5 N12, P12 L5, M5 N12, P12 L4, M4 f DCO P 2 × Q N where: N is the number of 2× interpolation filters enabled. P/Q is the rate converter ratio. The FS signal becomes a request for data that effectively gates the DCO clock and ensures that data is sent at the correct baud rate. If P/Q = 1 and N = 0, DCO occurs at the baud rate and FS is not required. In this case, FS is inactive (always high). The DCO signal can be used as a constant rate clock to request samples from the data source. L4, M4 “Rising” means that the data is sourced on the rising edge of DCOx; “falling” means that the data is sourced on the falling edge of DCOx. Rev. A | Page 48 of 76 AD9789 DCO 7 DCO CYCLES FS 6 DCO CYCLES tPD 7 DCO CYCLES tPD tPD SAMPLE 0 SAMPLE 1 SAMPLE 2 07852-115 D[31:0] DSC Figure 96. QDUC Mode Interface Timing Diagram for Design Example When FS Is Active Design Example LVDS DATA In this example, a system has a DAC rate of 1600 MHz and a baud rate of 15 MHz. Because fDCO = fDAC/16 = 100 MHz, the ratio of fDCO/fFS = 6.667. To satisfy the requirement that P/Q be between 0.5 and 1.0, an additional interpolation factor of 8× must be applied, so N = 3. Solving for P/Q results in 5/6. Retimer Operation The AD9789 uses a three-register retimer. The first two registers are clocked from any one of 16 phases derived from the DAC clock. The clock for the last register is fixed to Phase 15. The programmable register clocks are the digital sample clock (DSC) and the synchronizer clock (SNC). By choosing different phases, fine adjustment of the sampling time can be made to adjust for delays in the data source. Register 0x23[7:4] sets the DSC phase (DSCPHZ) and Register 0x23[3:0] sets the SNC phase (SNCPHZ) to any one of the 16 phases. The last register in the chain is always clocked from Phase 15. The parity counters can aid in identifying the edges of the data valid windows. Operation in CMOS mode is quite similar to operation in LVDS mode, as can be seen in Figure 97 and Figure 98. CMOS DATA 32 32 Q D CLK 32 Q D CLK Q D CLK BITS 0 TO 31 DSC Φ 0 TO 15 07852-070 SNC Φ 0 TO 13 PHZ Φ 15 Figure 97. CMOS Retiming Registers Q D CLK 16 Q D CLK 16 Q D CLK 16 Q D CLK Q D CLK BITS 0 TO 15 Q D CLK BITS 16 TO 31 16 Q D CLK DSC Φ 0 TO 15 SNC Φ 0 TO 13 07852-071 Therefore, three out of every 20 DCO clock edges should result in data samples being loaded into the device (the ratio of fFS/fDCO = 3/20). Figure 96 shows a timing diagram that illustrates the operation of the interface in this example. In the timing diagram, tPD corresponds to the propagation delay between the rising edge of FS and when the first sample in a given transmission is sampled into the AD9789. Note that tPD can vary by more than 1 DCO cycle. 16 16 PHZ Φ 15 Figure 98. LVDS Rearranges the DSC Register Register 0x23 and Register 0x21[2:0] can provide timing adjustments with very low jitter penalty, but they can also be set to the following recommended safe values: • • In LVDS mode, DSCPHZ = 0, SNCPHZ = 3, LTNCY = 0 (see the Latency Register section) In CMOS mode, DSCPHZ = 0, SNCPHZ = 7, LTNCY = 0 (see the Latency Register section) Timing adjustments can then be made in an FPGA or other data source. Note that selecting Phase 14 or Phase 15 for SNCPHZ results in a timing violation. In CMOS mode, setting DSCPHZ one step behind or at SNCPHZ also results in a timing violation. Latency Register A latency register, controlled via Register 0x21[2:0], follows the three-register retimer and can delay the data up to seven DCO clocks in steps of one DCO clock. The critical retiming is already done in the first three registers, so an incorrect latency value does not result in a timing violation. The latency value determines which data sample is the first sample in a transmission and routes that sample to the appropriate channel. Latency is affected by the round-trip delay from when FS goes high to when the first data sample is output from the retimer. If the latency value programmed into the part is incorrect, the input data samples will not be assembled properly. Rev. A | Page 49 of 76 AD9789 0123 4 56 7 8 16 24 32 40 48 56 64 72 SAMPLE SAMPLE SAMPLE SAMPLE SAMPLE SAMPLE SAMPLE SAMPLE DCO FS LVDS DDR SAMPLE SAMPLE LVDS SDR SAMPLE SAMPLE SAMPLE SAMPLE SAMPLE CMOS SAMPLE SAMPLE SAMPLE SAMPLE 07852-116 SAMPLE Figure 99. Sampling Points at Delay = 0 Retimer and Latency Look-Up Tables For LVDS DDR, In practice, the retimer and latency parameters can be reduced to a single verified and guaranteed table that provides delays at optimum sample points from 0 to over 100 DAC clocks. The sampling points for LVDS DDR, LVDS SDR, and CMOS interface modes are given in Figure 99 for delay = 0. The number scale above the DCO signal in Figure 99 corresponds to the delay value in DAC clock cycles in Table 66 and Table 67. The delay of the pins should be taken into account. This delay is 800 ps for the output delay and 800 ps for the input delay, for a total of 1.6 ns. This delay is included in the following formulas. See Table 66 for a complete set of recommended retimer settings for all delay values. Note: for LVDS DDR, zero (0) measured delay results in a retimer setting of 20, while for .LVDS SDR or CMOS, the zero(0) delay corresponds to a retimer setting of 12 at fDAC = 2.4 GHz. To use Table 66 and Table 67, probe the FS, DCO, and data input signals at the AD9789. While viewing these signals on an oscilloscope, measure the delay between the rising edge of FS and the start of the first data sample and add 1.6 ns from the delay of the pins to this value. Normalize this total delay to one DAC clock period. The optimum sampling point in number of DAC clock cycles, which corresponds to the delay number in Table 66 and Table 67, can be found from this measured value for each interface mode. Delay OPTIMAL = Delay MEASURED + 1.6 ns t DCO / 16 + 16 For LVDS SDR, Delay OPTIMAL = Delay MEASURED + 1.6 ns t DCO / 16 +8 For CMOS, Delay OPTIMAL = Delay MEASURED + 1.6 ns t DCO / 16 +8 For a maximum valid sampling window, the sampling point should be fine-tuned based on the data input setup and hold times. If the setup and hold times are symmetric about the DCO edge, choosing a sampling point at the center of the data window results in the maximum valid sampling window. For more information on the input data setup and hold times, refer to the CMOS Interface Timing section or the LVDS Interface Timing section. The LAT, SNC, and DSC values for the optimal sampling point in Table 66 or Table 67 should be written to the LTNCY[2:0] bits in Register 0x21[2:0], the SNCPHZ[3:0] bits in Register 0x23[3:0], and the DSCPHZ[3:0] bits in Register 0x23[7:4], respectively. Rev. A | Page 50 of 76 AD9789 Table 66. Recommended Retimer Settings for All Delay Values, LVDS Mode Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC 0 0 7 8 8 1 3 0 16 1 7 8 24 2 3 0 32 2 7 8 40 3 3 0 48 3 7 8 56 4 3 0 64 4 7 8 72 5 3 0 80 5 7 8 88 6 3 0 1 0 8 9 9 1 4 1 17 1 8 9 25 2 4 1 33 2 8 9 41 3 4 1 49 3 8 9 57 4 4 1 65 4 8 9 73 5 4 1 81 5 8 9 89 6 4 1 2 0 9 10 10 1 4 2 18 1 9 10 26 2 4 2 34 2 9 10 42 3 4 2 50 3 9 10 58 4 4 2 66 4 9 10 74 5 4 2 82 5 9 10 90 6 4 2 3 0 9 11 11 1 5 3 19 1 9 11 27 2 5 3 35 2 9 11 43 3 5 3 51 3 9 11 59 4 5 3 67 4 9 11 75 5 5 3 83 5 9 11 91 6 5 3 4 0 10 12 12 1 5 4 20 1 10 12 28 2 5 4 36 2 10 12 44 3 5 4 52 3 10 12 60 4 5 4 68 4 10 12 76 5 5 4 84 5 10 12 92 6 5 4 5 0 10 13 13 1 6 5 21 1 10 13 29 2 6 5 37 2 10 13 45 3 6 5 53 3 10 13 61 4 6 5 69 4 10 13 77 5 6 5 85 5 10 13 93 6 6 5 6 1 2 14 14 1 6 6 22 2 2 14 30 2 6 6 38 3 2 14 46 3 6 6 54 4 2 14 62 4 6 6 70 5 2 14 78 5 6 6 86 6 2 14 94 6 6 6 7 1 3 15 15 1 7 7 23 2 3 15 31 2 7 7 39 3 3 15 47 3 7 7 55 4 3 15 63 4 7 7 71 5 3 15 79 5 7 7 87 6 3 15 95 6 7 7 Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC 96 6 7 8 104 7 3 0 112 7 7 8 97 6 8 9 105 7 4 1 113 7 8 9 98 6 9 10 106 7 4 2 114 7 9 10 99 6 9 11 107 7 5 3 115 7 9 11 100 6 10 12 108 7 5 4 116 7 10 12 101 6 10 13 109 7 6 5 117 7 10 13 102 7 2 14 110 7 6 6 X X X X 103 7 3 15 111 7 7 7 X X X X Table 67. Recommended Retimer Settings for All Delay Values, CMOS Mode Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Rev. A | Page 51 of 76 0 0 7 0 8 1 3 8 16 1 7 0 24 2 3 8 32 2 7 0 40 3 3 8 48 3 7 0 56 4 3 8 64 4 7 0 1 0 8 1 9 1 4 9 17 1 8 1 25 2 4 9 33 2 8 1 41 3 4 9 49 3 8 1 57 4 4 9 65 4 8 1 2 0 8 2 10 1 4 10 18 1 8 2 26 2 4 10 34 2 8 2 42 3 4 10 50 3 8 2 58 4 4 10 66 4 8 2 3 0 9 3 11 1 5 11 19 1 9 3 27 2 5 11 35 2 9 3 43 3 5 11 51 3 9 3 59 4 5 11 67 4 9 3 4 0 9 4 12 1 5 12 20 1 9 4 28 2 5 12 36 2 9 4 44 3 5 12 52 3 9 4 60 4 5 12 68 4 9 4 5 0 2 5 13 1 6 13 21 1 2 5 29 2 6 13 37 2 2 5 45 3 6 13 53 3 2 5 61 4 6 13 69 4 2 5 6 1 2 6 14 1 6 14 22 2 2 6 30 2 6 14 38 3 2 6 46 3 6 14 54 4 2 6 62 4 6 14 70 5 2 6 7 1 3 7 15 1 7 15 23 2 3 7 31 2 7 15 39 3 3 7 47 3 7 15 55 4 3 7 63 4 7 15 71 5 3 7 AD9789 Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC Delay LAT SNC DSC 72 5 3 8 80 5 7 0 88 6 3 8 96 6 7 0 104 7 3 8 112 7 7 0 73 5 4 9 81 5 8 1 89 6 4 9 97 6 8 1 105 7 4 9 113 7 8 1 74 5 4 10 82 5 8 2 90 6 4 10 98 6 8 2 106 7 4 10 114 7 8 2 75 5 5 11 83 5 9 3 91 6 5 11 99 6 9 3 107 7 5 11 115 7 9 3 76 5 5 12 84 5 9 4 92 6 5 12 100 6 9 4 108 7 5 12 116 7 9 4 77 5 6 13 85 5 2 5 93 6 6 13 101 6 2 5 109 7 6 13 117 7 2 5 78 5 6 14 86 6 2 6 94 6 6 14 102 7 2 6 110 7 6 14 X X X X 79 5 7 15 87 6 3 7 95 6 7 15 103 7 3 7 111 7 7 15 X X X X The timing of the input data is referenced to DCO for a given phase of DSC. The CMOS data input timing over temperature is shown in Table 68 for DCO_INV = 0 (Register 0x20[4]), DSCPHZ = 0 (Register 0x23[7:4]), and DCODIV = 1 (Register 0x22[6:4]). Table 68 also shows the data valid window (DVW). The data valid window is the sum of the setup and hold times of the interface. DVW is the minimum amount of time that valid data must be presented to the device to ensure proper sampling. Table 68. CMOS Data Input Timing with Respect to DCO Temperature −40°C +25°C +85°C −40°C to +85°C Min tS (ns) 4.9 5.1 5.3 5.3 Min DVW (ns) 3.5 3.5 3.6 3.9 For any value of DSCPHZ greater than 0, the setup and hold times shift by increments of tDCO/16, where tDCO is the period of the data clock. tS = 5.3 ns − ((tDCO/16) × DSCPHZ) tH = 0.24 ns + ((tDCO/16) × DSCPHZ) DCO tS Latency Effects on Channelizer Mode Min tH (ns) −1.4 −1.6 −1.7 −1.4 tH When selecting an interface configuration in channelizer mode, the number of DCO cycles between FS pulses (cyclesAVAIL) must be greater than the number of DCO cycles required by the interface configuration (cyclesINTERFACE). Latency consumes some of these available DCO cycles between FS. This decrease in available DCO cycles is a result of the round-trip propagation delay from the FS output of the AD9789 to the respective data sample at the input of the AD9789 (LTNCY[2:0]) in addition to the internal latency of the device. For a successful interface design, the following condition must be met: 07852-117 INPUT DATA DSC Figure 100. CMOS Input Timing In some interface modes, the delay from the rising edge of DCO to the rising edge of FS needs to be known. This delay is summarized over temperature in Table 69. DCO tD FS 07852-118 cyclesAVAIL ≥ cyclesINTERFACE + LTNCY[2:0] + 2 DSC CMOS Interface Timing When the AD9789 is configured with a CMOS interface (CMOS_CTRL = CMOS_BUS = 3.3 V), a CMOS data clock output signal, DCO, is provided to drive data from the data source. The output signal operates at the input data rate, which is equal to fDAC/16 when DCODIV = 1. CMOS data on the bus is sampled on the rising edge of an internal sampling clock (DSC). Note that the frequency of DCO is equal to the frequency of DSC and the phase relationship between DCO and DSC is determined by DSCPHZ (Register 0x23[7:4]). Figure 101. CMOS_DCO to CMOS_FS Delay Table 69. Timing Delay Between CMOS_DCO and CMOS_FS Temperature −40°C +25°C +85°C −40°C to +85°C Rev. A | Page 52 of 76 tD, MAX DCO to FS (ns) 0.64 0.71 0.85 0.85 tD, MIN DCO to FS (ns) 0.28 0.4 0.49 0.28 AD9789 When the AD9789 is configured with an LVDS interface (CMOS_CTRL = CMOS_BUS = 0 V), an LVDS data clock output signal, DCO, is provided to drive data from the data source. The LVDS interface may be single data rate (SDR) or double data rate (DDR) depending on the bus width configuration. In SDR, data is sampled into the part only on the falling edge of the internal sampling clock (DSC). Note that the frequency of DCO is equal to the frequency of DSC, so the effective data rate is equal to the DCO frequency. The phase relationship between DCO and DSC is determined by DSCPHZ (Register 0x23[7:4]). In DDR, data is sampled into the part on both the rising and falling edges of DSC, so the effective data rate is equal to twice the DCO frequency. The interface is DDR only when the bus width is equal to 32 bits. The DCO frequency is equal to fDAC/16 when DCODIV = 1. The timing of the input data is referenced to DCO for a given phase of DSC. The LVDS input data timing over temperature is shown in Table 70 for DCO_INV = 0 (Register 0x20[4]), DSCPHZ = 0 (Register 0x23[7:4]), and DCODIV = 1 (Register 0x22[6:4]). Table 70. LVDS Data Input Timing with Respect to DCO Temperature −40°C +25°C +85°C −40°C to +85°C Min tS (ns) 1.04 1.23 1.41 1.41 Min tH (ns) 0.24 0.16 0.03 0.24 Min DVW (ns) 1.28 1.39 1.44 1.65 In DDR mode, these setup and hold times must be applied to both edges of DCO. In SDR mode, these setup and hold times must be applied to the falling edge of DCO. DCO tD FS DSC Figure 103. LVDS DCO to FS Delay Table 71. Timing Delay Between LVDS DCO and FS Temperature −40°C +25°C +85°C −40°C to +85°C The AD9789 supports parity checking on the input data bus. There are three parity checking modes: even parity, odd parity, and IQ parity. In IQ parity mode, a value of 0 is always expected on the I channel and a value of 1 is always expected on the Q channel. Note that IQ parity mode is generally useful only when the LVDS interface is used. These modes are controlled via Register 0x20[1:0]. SINGLE DATA RATE (SDR) DCO tH INPUT DATA Recall that the LVDS interface can be single data rate (SDR) or double data rate (DDR), depending on the bus width configuration. The interface is DDR only when the bus width is equal to 32 bits. DSC DOUBLE DATA RATE (DDR) DCO tS tH tS tH 07852-119 INPUT DATA DSC Register 0x20[1:0] 00 01 10 11 If parity checking is used, each data-word that is transferred into the AD9789 should have a parity bit accompanying it, regardless of FS. In other words, parity must be valid for every DCO edge. The parity bits are located at Pin L4 and Pin M4. When operating the interface in CMOS mode, the input parity bits are referred to as P1 and P0, respectively. When operating the interface in LVDS mode, the input parity bits are referred to as PARP and PARN, respectively. tH = 0.24 ns + ((tDCO/16) × DSCPHZ) tH tD, MIN DCO to FS (ns) 0.21 0.16 0.12 0.12 Parity Parity Mode Deactivates Parity Checking IQ Parity Even Parity Odd Parity tS = 1.41 ns − ((tDCO/16) × DSCPHZ) tS tD, MAX DCO to FS (ns) 0.37 0.35 0.32 0.37 Table 72. Parity Mode SPI Settings For any value of DSCPHZ greater than 0, the setup and hold times shift by increments of tDCO/16, where tDCO is the period of the data clock. tS In some interface modes, the delay from the rising edge of DCO to the rising edge of FS needs to be known. This delay is summarized over temperature in Table 71. 07852-120 LVDS Interface Timing Figure 102. LVDS Input Timing, SDR vs. DDR Rev. A | Page 53 of 76 AD9789 In QDUC mode, where the interface is fixed at a 32-bit bus width, the parity behavior is straightforward (see Table 73). If a parity error occurs, the parity counter (Register 0x02[7:0]) is incremented. The parity counter continues to accumulate until it is cleared or until it reaches a maximum value of 255. The count can be cleared by writing a 1 to Register 0x04[7]. Table 73. Parity Behavior in QDUC Mode Interface CMOS Bus Width 32 bits LVDS1 (DDR) 32 bits 1 Even/Odd Parity P1 checks D[31:16] P0 checks D[15:0] [PARP, PARN] rising checks D[15:0]P, D[15:0]N rising [PARP, PARN] falling checks D[15:0]P, D[15:0]N falling IQ Parity P1 = 0 P0 = 1 PARP rising = 0 PARN rising = 1 PARP falling = 1 PARN falling = 0 “Rising” corresponds to the data sampled on the rising edge of DSC; “falling” corresponds to the data sampled on the falling edge of DSC. An IRQ can be enabled to trigger when a parity error occurs by writing a 1 to Register 0x03[7]. The status of IRQ can be measured via Register 0x04[7] or by using the IRQ pin (Pin P2). If using the IRQ pin and more than one IRQ is enabled, the user must check Register 0x04 when an IRQ event occurs to determine whether the IRQ was caused by a parity error. The IRQ can also be cleared by writing a 1 to Register 0x04[7]. ANALOG MODES OF OPERATION In channelizer mode, where the interface is configurable for different bus widths, data widths, and data formats, the parity bits check the data-word on the bus. The AD9789 uses a quad-switch architecture that can be configured to operate in one of three modes via the serial peripheral interface: normal mode, RZ mode, and mix mode. For example, consider a configuration in channelizer mode where the bus width is 4, the data width is 8, and the data format is real. In this case, eight clock cycles are required to transfer all of the baud rate data to represent four channels. In the even parity or odd parity mode, one parity bit and four data bits are sent on each clock; the parity bit checks the four data bits to verify that all of the data was sent over the interface. The quad-switch architecture masks the code-dependent glitches that occur in a conventional two-switch DAC. Figure 104 shows the waveforms for a conventional DAC and the quad-switch DAC. In the two-switch architecture with D1 and D2 in different states, a switch transition results in a glitch. However, if D1 and D2 are at the same state, the switch does not create a glitch. This codedependent glitching causes an increased amount of distortion in the DAC. In the quad-switch architecture, two switches are always transitioning at each half clock cycle, regardless of the code; therefore, code-dependent glitches are eliminated, but a constant glitch at 2 × fDAC is created. Table 74. Parity Behavior in Channelizer Mode Interface CMOS Bus Width 4 bits CMOS 8 bits CMOS 16 bits CMOS 32 bits LVDS (SDR)1 LVDS (SDR)1 LVDS (SDR)1 LVDS (DDR)1 4 bits 1 8 bits 16 bits 32 bits Even/Odd Parity P1 ignored P0 checks D[3:0] P1 ignored P0 checks D[7:0] P1 ignored P0 checks D[15:0] P1 checks D[31:16] P0 checks D[15:0] [PARP, PARN] falling checks D[3:0]P, D[3:0]N falling [PARP, PARN] falling checks D[7:0]P, D[7:0]N falling [PARP, PARN] falling checks D[15:0]P, D[15:0]N falling [PARP, PARN] rising checks D[15:0]P, D[15:0]N rising [PARP, PARN] falling checks D[15:0]P, D[15:0]N falling IQ Parity P1 = 0 P0 = 1 P1 = 0 P0 = 1 P1 = 0 P0 = 1 P1 = 0 P0 = 1 Not supported Not supported INPUT DATA D1 D2 D3 D4 D5 D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D6 D7 D8 D9 D10 DACCLK 2-SWITCH DAC OUTPUT 4-SWITCH DAC OUTPUT (NORMAL MODE) t D6 D1 D2 D3 D4 D7 D8 D9 D10 t D5 Not supported Figure 104. Two-Switch and Quad-Switch DAC Waveforms PARP rising = 0 PARN rising = 1 PARP falling = 1 PARN falling = 0 “Rising” corresponds to the data sampled on the rising edge of DSC; “falling” corresponds to the data sampled on the falling edge of DSC. The quad-switch architecture can also be easily configured to perform an analog mix or return-to-zero (RZ) function. In mix mode, the output is effectively chopped at the DAC sample rate. The RZ mode is similar to mix mode, except that the intermediate data samples are replaced with midscale values instead of inverting values. Figure 105 shows the DAC waveforms for both mix mode and RZ mode. Rev. A | Page 54 of 76 07852-072 Table 74 summarizes the behavior of the two parity pins and how they interact with the data in all interface modes. AD9789 INPUT DATA D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 ANALOG CONTROL REGISTERS DACCLK D2 4-SWITCH DAC OUTPUT (fS MIX MODE) –D7 D4 D1 D5 –D9 –D10 –D6 t –D1 –D5 D6 –D2 –D4 D8 D3 D4 D7 D8 D9 D10 t D5 –110 07852-073 D2 Using the MSEL[1:0] bits (Register 0x36[1:0]), the user can adjust the noise contribution of the internal current mirror to optimize the 1/f noise. Figure 107 shows MSEL vs. the 1/f noise with 20 mA full-scale current into a 50 Ω resistor. D9 D7 D6 D1 Mirror Roll-Off Frequency Control D10 –D3 4-SWITCH DAC OUTPUT (RETURN-TOZERO MODE) The AD9789 includes registers for optimizing its analog performance. These registers include noise reduction in the output current mirror and output current mirror headroom adjustments. –D8 D3 –115 Switching between analog modes reshapes the sinc roll-off inherent at the DAC output. The performance and maximum amplitude in all three Nyquist zones is affected by this sinc roll-off depending on where the carrier is placed, as shown in Figure 106. FIRST NYQUIST ZONE SECOND NYQUIST ZONE 0 NOISE (dBm/Hz) Figure 105. Mix Mode and RZ Mode DAC Waveforms THIRD NYQUIST ZONE MSEL = 11 –125 MSEL = 10 –130 MSEL = 00 –140 1 10 FREQUENCY (kHz) Figure 107. 1/f Noise with Respect to MSEL Bits –15 NORMAL MODE –20 –25 0 0.5 1.0 1.5 2.0 FREQUENCY (Hz) 2.5 3.0 07852-074 –30 Figure 106. Sinc Roll-Off for Each Analog Operating Mode (fS = 2 × DACCLK) The RZ mode, with its lower but flat response, can be quite useful for quick checks of system frequency response. Rev. A | Page 55 of 76 100 07852-083 MIX MODE –10 AMPLITUDE (dBm) MSEL = 01 –135 –5 RZ MODE –35 –120 AD9789 VOLTAGE REFERENCE The AD9789 output current is set by a combination of digital control bits and the I120 reference current, as shown in Figure 108. AD9789 FSC[9:0] VBG 1.2V VREF CURRENT SCALING + FULL-SCALE CURRENT 07852-084 10kΩ AVSS IPTAT (Pin D14) is used for factory testing and can simply be left floating. IPTAT is an output current that is proportional to absolute temperature. At 25°C, the output current is approximately 10 μA and follows a slope of approximately 20 nA/°C. – I120 1nF DAC I120 VREF (Pin C14) must be bypassed to ground with a 1 nF capacitor. The band gap voltage is present on this pin and can be buffered for use in external circuitry. The typical output impedance is near 5 kΩ. If desired, an external reference can be used to overdrive the internal reference by connecting it to the VREF pin. For optimal DOCSIS 3.0 ACLR performance, the full-scale output current settings provided in Table 75 are recommended. Figure 108. Voltage Reference Circuit The reference current is obtained by forcing the band gap voltage across an external 10 kΩ resistor from I120 (Pin B14) to ground. The 1.2 V nominal band gap voltage, VREF (Pin C14), generates a 120 μA reference current in the 10 kΩ resistor. This current is adjusted digitally by FSC[7:0] (Register 0x3C[7:0]) and FSC[9:8] (Register 0x3D[1:0]) to set the output full-scale current, IFS, in milliamperes. IFS = 0.023 × FSC[9:0] + 8.58 The full-scale output current range is approximately 8.6 mA to 32.1 mA for register values from 0x000 to 0x3FF. The default value of 0x200 generates 20 mA full scale. The typical range is shown in Figure 109. 35 Table 75. Recommended Full-Scale Current Settings vs. Number of QAM Channels Number of QAM Channels 1 2 3 4 Recommended IFS (mA) 20 25 25 25 FSC[9:0] 512 720 720 720 DAC OUTPUT STAGES To properly evaluate the AD9789 in the lab, three distinct output coupling circuits were used. Figure 110 shows the optimal output network when measuring traditional DAC performance specifications such as SFDR and IMD performance with sine waves. 30 IOUTP 90Ω JTX-2-10T 90Ω 07852-121 70Ω 20 IOUTN 15 Figure 110. Recommended Transformer Output Stage for Single-Tone/Multitone Measurements 10 0 0 200 400 600 800 DAC GAIN CODE 1000 07852-085 5 Figure 111 shows the optimal output network when measuring signals in mix mode (second or third Nyquist zone). The bandwidth of the center tap transformer is not sufficient to support mix mode outputs, so the best solution is to use a wideband balun by itself. Figure 109. Full-Scale Current vs. DAC Gain Code 90Ω IOUTP Always connect a 10 kΩ resistor from the I120 pin to ground and use the digital controls to adjust the full-scale current. The AD9789 is not a multiplying DAC. Applying an analog signal to I120 is not supported. 70Ω IOUTN MABACT0039 90Ω 07852-122 IFS (mA) 25 Figure 111. Recommended Transformer Output Stage for Mix Mode Rev. A | Page 56 of 76 AD9789 Finally, when measuring performance for CMTS and other digital TV applications, it is advantageous to insert a 1 dB, 1.2 GHz Chebyshev low-pass filter between the DAC and the transformer to better control the impedance seen at the DAC core. This helps to decrease the folded back harmonics for higher frequency outputs. The optimal transformer for CMTS measurements is the JTX-2-10T, which consists of a balun and center-tapped transformer in a single package. This output stage is shown in Figure 112. VCC 5.6nH IOUTP 90Ω 90Ω VREF JTX-2-10T 2.2pF 50Ω 07852-123 IOUTN 4.7pF Traces from the DAC to the transformer should be 50 Ω impedance to ground each in Figure 110 and Figure 112 and 25 Ω to ground each in Figure 111 to avoid unnecessary parasitics. CLOCKING THE AD9789 To provide the required signal swing for the internal clock receiver of the AD9789, it is necessary to use an external clock buffer chip to drive the CLKP and CLKN inputs. These high level, high slew rate signals should not be routed any distance on a PCB. The recommended clock buffer for this application is the ADCLK914. This ultrafast clock buffer is capable of providing 1.9 V out of each side into a 50 Ω load terminated to VCC (3.3 V) for a total differential swing of 3.8 V. GND 2 3 4 5 C82 0.01µF GND R15 49.9Ω GND VEE VCC 1 2 3 4 VT VREF GND D D NC NC 5 6 7 8 Q D VEE Figure 113. ADCLK914 Functional Block Diagram The internal 50 Ω resistors shown at the ADCLK914 inputs are rated to carry currents from PECL or CML drivers. The VT pin can be connected to VCC, a PECL current sink, or the internal VREF, or it can be left floating depending on the source. The common-mode input range of the ADCLK914 does not include LVDS voltage levels, so ac coupling is required in that case. VCC33 R13 49.9Ω R14 49.9Ω U3 ADCLK914 C99 2400pF C0803H50 CLKP R17 100Ω R0402 C102 2400pF C0803H50 12 Q 11 Q NC 10 NC 9 CLKN ADCLK914 SUPPLY DECOUPLING NC NC VEE VCC 1 C81 0.01µF 50Ω Q VCC33 C83 0.01µF 16 15 14 13 50Ω 50Ω D Figure 112. Recommended Transformer Output Stage for CMTS Measurements J3 PSTRNKPE4117 ADCLK914 VT 5.6nH 07852-124 70Ω VCC33 VCC33 GND C31 0.1µF C0402 C32 0.01µF C0402 C33 0.1µF C0402 C34 0.01µF C0402 GND GND GND GND Figure 114. ADCLK914/AD9789 Interface Circuit for Use with a Lab Generator Rev. A | Page 57 of 76 07852-125 4.7pF The buffer, in turn, can be easily driven from lower level signals such as CML or attenuated PECL that might be encountered on a PCB. This buffer also provides very low, 100 fs added random jitter, which is important to obtain the optimal ac performance from the AD9789. A functional block diagram of the ADCLK914 is shown in Figure 113. Figure 114 shows the recommended schematic for the ADCLK914/AD9789 interface. Refer to the ADCLK914 data sheet for more information. Any time that the noise floor from the DAC cannot meet the specifications in this data sheet, the clock should be examined. AD9789 Optimizing the Clock Common-Mode Voltage In addition to the system that optimizes the handoff timing, an additional system sets the common-mode voltage of the clock. This system can be used to properly align the crossing point of the CLKP and CLKN signals to ensure that the duty cycle of the clock is set properly. Figure 115 shows how the common-mode voltage of CLKP and CLKN is set. There are eight switches controlled by the CLKP_CML bits (Register 0x32[4:1]) and the CLKN_CML bits (Register 0x31[7:4]) for both the CLKP and CLKN signals. The direction of the adjustment is determined by the PSIGN and NSIGN bits (Register 0x32, Bit 5 and Bit 0). If PSIGN and NSIGN are low, the common-mode voltage decreases with CLKP_CML/CLKN_CML values. If PSIGN and NSIGN are high, the common-mode voltage increases with CLKP_CML/ CLKN_CML values, as shown in Figure 116. With both CLKP_CML and CLKN_CML set to 0, the feedback path forces the common-mode voltage to be set to approximately 0.9 V. The optimal ac performance occurs at a setting of −15 on both the CLKP and CLKN offset bits. Table 76. Four-Carrier DOCSIS Close-In ACLR Performance at 900 MHz for Various Phase Noise Profiles Band 750 kHz to 6 MHz 6 MHz to 12 MHz 12 MHz to 18 MHz CLKx_CML SIGN = 1 −70.9 −70.3 −67 −63.8 −63 −71 −70.8 −70.8 −70.8 −65 Table 77. Phase Noise Summary for Each Profile Phase Noise (dBc/Hz) Offset1 2 kHz 20 kHz 200 kHz 2 MHz 20 MHz Profile 1 −114.8 −117.8 −128.3 −148.5 −152.5 Profile 2 −112.8 −115.5 −118.9 −127.9 −149.9 Profile 3 −111.7 −114.6 −118.3 −122.2 −148 Profile 4 −111.2 −113.8 −116.8 −117.9 −145.7 At offsets less than 500 kHz, the measurement instrument dominates the phase noise performance. MU DELAY CONTROLLER 07852-081 The mu delay adjusts timing between the digital and analog blocks. The mu delay controller receives phase relational information between the digital and analog clock domains. The control system continuously adjusts the mu delay to maintain the desired phase relationship between the digital and analog sections. A top level diagram of the mu delay within the DAC is shown in Figure 117. CVDD18 Figure 115. Clock Common-Mode Control 1.10 CLKN 1.05 1.00 16-BIT DATA 0.95 16 DIGITAL CIRCUITRY 14 14-BIT 2.4GSPS DAC 0.90 DAC CLOCK 0.85 MU DELAY MU Φ DET MU Φ CONTROL 0.80 0.75 5 7 9 11 13 15 Figure 117. Mu Delay Controller Block Diagram 07852-082 0.70 –15 –13 –11 –9 –7 –5 –3 –1 1 3 OFFSET CODE 07852-077 COMMON-MODE VOLTAGE (V) Spec −60 To meet the close-in ACLR requirements for four-carrier DOCSIS, the phase noise found in Profile 3 is the minimum requirement necessary. CLKP/CLKN CLKP Phase Noise (dBc) Profile 2 Profile 3 Profile 4 −67.2 −62.4 −59.1 Table 77 shows the phase noise at various offsets for each profile. (All phase noise numbers are specified in dBc/Hz.) 1 CLKx_CML SIGN = 0 Profile 1 −71 Figure 116. Common-Mode Voltage with Respect to CLKP_CML/CLKN_CML and PSIGN/NSIGN Clock Phase Noise Effects on AC Performance The quality of the clock source driving the ADCLK914 determines the achievable ACLR performance of the AD9789. Table 76 summarizes the close-in ACLR for a four-carrier DOCSIS signal at 900 MHz with respect to various phase noise profiles. (All ACLR values are specified in dBc.) The mu controller has two modes of operation: initial phase search and phase tracking. In the phase search mode, the controller looks for the initial mu delay value to use before going into tracking mode. In tracking mode, the controller makes adjustments to the initial mu delay value to keep the phase at the desired value. The initial phase search is required because multiple mu delay settings may result in the desired phase, but the device may not operate correctly at all of those mu delay values. Rev. A | Page 58 of 76 AD9789 The mu controller is enabled via Register 0x33[0]. Enabling the controller sets in motion the phase search mode. Before enabling the controller, it is important to turn on both the phase comparator boost (Register 0x3E[5]) and the mu control duty cycle correction circuitry (Register 0x30[7]). Both of these functions allow for more robust operation of the mu controller over the entire operating speed of the part. The three modes of operation for the mu controller are specified by the MODE[1:0] bits in Register 0x33[5:4] as follows: Not exact (0): can find a phase within two values of the desired phase Exact (1): finds the exact phase specified (optimal setting) Figure 118 shows a typical plot of mu phase vs. mu delay line value at 2.4 GSPS. Starting at the selected mu delay value, the search direction can be specified via the SEARCH_DIR[1:0] bits in Register 0x39[6:5]. The three possible choices for the search are as follows: • • • 18 DESIRED PHASE AND SLOPE 16 14 GUARD BAND 12 GUARD BAND 10 8 6 4 2 0 SEARCH STARTING LOCATION 0 40 80 120 160 200 240 280 MU DELAY 320 360 400 440 Figure 118. Typical Mu Phase Characteristics @ 2.4 GSPS To determine whether the search is on the correct slope, the controller measures the slope by first incrementing and then decrementing the mu delay value until any of the following events happens: • • • • • The phase changes by 2. The phase is equal to 16 (the maximum value). The phase is equal to 0 (the minimum value). The mu delay is 431 (the maximum value). The mu delay is 0 (the minimum value). After incrementing and then decrementing the mu delay value, the values of the measured phases are compared to determine whether the slope matches the desired slope. To consider the slope valid, the positive direction phase and the negative direction phase must be on opposite sides of the desired phase. Examples of valid and invalid phase choices are shown in Figure 119 and Figure 120. 15 14 Down only (00) Up only (01) Alternating up and down (10) (optimal setting) 9 8 13 7 12 DESIRED If the search direction is alternating, the search proceeds in both directions until a programmable guard band is reached in one of the directions, specified by the GUARDBAND[4:0] bits in Register 0x2F[4:0]. When the guard band is reached, the search continues only in the opposite direction. If the desired phase is not found before the guard band is reached in the second direction, the search reverts to the alternating mode and continues looking within the guard band. 11 10 POSITIVE SLOPE Rev. A | Page 59 of 76 DESIRED 6 5 4 NEGATIVE SLOPE Figure 119. Valid Positive and Negative Slope Phase Examples 07852-079 • Continue (0): continues to search (optimal setting) Reset (1) Search and track (00) (optimal setting) Track only (01) Search only (10) The search algorithm begins at a specified mu delay value set using the MUDLY[8:0] bits, where the LSB is located in Register 0x39[7] and the MSBs are located in Register 0x3A[7:0]. Even though there are nine bits of resolution for this delay line value, the maximum allowable mu delay is 431 (decimal). The optimal point to begin the search is in the middle of the delay line, or approximately 216. The initial search algorithm works by sweeping through different mu delay values until the desired phase is measured; this phase is specified using the MUPHZ[4:0] bits in Register 0x39[4:0], with the maximum allowable phase being 16. If values larger than 16 are loaded, the controller will not lock. When the desired phase is measured, the slope of the phase measurement is calculated and compared to the desired slope, which is specified by the SLOPE bit in Register 0x33[6]. For optimal ac performance, the best setting for the search is a positive slope and a phase value of 14. If the phase and slope match the configured values, the search algorithm is finished. The SEARCH_TOL bit (Register 0x2F[7]) can be used to specify the accuracy of the search as follows: • • • MU PHASE • • • The search fails if the mu delay reaches the endpoints. If the controller does not find the desired phase during the search, the TRACK_ERR bit (Register 0x2F[5]) determines the corrective action as follows: 07852-078 Operating the Mu Controller in Auto Mode AD9789 13 12 DESIRED 3 3 14 13 2 2 DESIRED 1 07852-080 14 4 15 15 1 Figure 120. Invalid Slope Phase Examples Table 78 lists register writes and reads to lock to the controller. The program assumes that the clock receiver is already enabled and that a clean lock is provided. The typical locking time for the mu controller is approximately 180,000 DAC cycles (at 2 GSPS, ~75 μs). Table 78. AD9789 Mu Delay Controller Routine When the initial mu delay value has been found by the search algorithm, the tracking mode is enabled. In tracking mode, a simple control loop is used to increment by 1, decrement by 1, or not change the mu delay value depending on the measured phase. The control loop uses the desired slope to determine whether the mu delay should be incremented or decremented. No attempt is made to determine whether the actual slope has changed or is still valid. Address 0x30 0x31 Data 0x80 0xF0 R/W Write Write 0x32 0x9E Write Two status bits, LOCKACQ (Register 0x04[3]) and LOCKLOST (Register 0x04[2]) are available to signal proper operation of the control loop. If the current phase is more than five steps away from the desired phase and the LOCKACQ bit was previously set, the LOCKACQ bit is cleared and the LOCKLOST interrupt bit is set. Furthermore, if lock is lost, the controller can remain in the tracking loop, or it can be reset to start the search again. 0x3E 0x38 Write 0x24 0x24 0x2F 0x00 0x80 0xCE Write Write Write 0x33 0x39 0x42 0x4E Write Write 0x3A 0x6C Write 0x03 0x04 0x03 0x33 0x00 0xFE 0x0C 0x43 Write Write Write Write 0x33 0x33 0x04 0x4B 0x43 Write Write Read By setting the MUSAMP bit high (Register 0x33[3]) from a low state, the user can read back the mu delay value that the controller locked to by reading the MUDLY bits (Register 0x39[7] and Register 0x3A[7:0]), as well as the phase it locked to by reading back the MUPHZ[4:0] bits (Register 0x39[4:0]). These bits will no longer read back the value that the search started at or the desired phase, but instead will read back the mu delay line value and phase that the controller is locked to. 0x39 Rev. A | Page 60 of 76 Read Description Enable duty cycle correction. Set common-mode level of CLKN: CLKN_CML = 0xF. Set common-mode level of CLKP: CLKP_CML = 0xF. Set direction of CLKP_CML and CLKN_CML: PSIGN = 0; NSIGN = 0. Enable clock receiver: CLK_DIS = 1. Set phase comparator boost (AUTO_CAL must be set to its default value, 1). Enable digital clocks. Search for exact phase with a guard band of 98 codes from endpoints. Set search slope to positive. Set search phase to 14, search up and down. Set start point of search to midpoint of mu delay line (Code 216). Disable lock and lock lost indicators. Clear lock and lock lost indicators. Enable lock and lock lost indicators. Enable mu delay controller and start search/track routine. Set mu phase read bit high. Set mu phase read bit low. Check lock and lock lost bits: LOCKACQ should be on. LOCKLOST should be off. Check phase readback (should be equal to 14). AD9789 Operating the Mu Controller in Manual Mode INTERRUPT REQUESTS In manual mode, the user must sweep through all the mu delay values and record the phase value at each value of MUDLY as shown in Figure 118. Every time that the MUDLY value is stepped, the MUSAMP bit must be toggled from low to high to read the corresponding phase for the specified mu delay line value. It is not possible to keep read high and continuously read back the phase value. As with auto mode, the optimal ac performance occurs at a positive slope and a phase of 14; therefore, when the curve is complete, choose the MUDLY value that corresponds to this condition and write that value to the MUDLY[8:0] bits (Register 0x39[7] and Register 0x3A). The following interrupt (IRQ) requests can be used for additional information and verification of the status of various functional blocks: • Calculating Mu Delay Line Step Size • Stepping through all of the mu delay line values and plotting mu phase vs. mu delay not only allows the user to find the optimal mu delay value, but can also allow the user to determine the mu delay line step size. To calculate the step size, take one full cycle of the mu phase curve and divide the period of the DAC clock by this delta. From Figure 118, the two transition points are approximately 56 and 270, providing a delta of approximately 214 steps. Therefore, the mu delay line step size would be approximately 2 ps/step, as shown in the following equation: ⎛ ⎞ 1 ⎜ ⎟ ⎜ 2.4 GHz ⎟ ⎝ ⎠ = 1.95 ps 214 If the mu controller is enabled, this value allows the user to calculate (in picoseconds) how much drift is in their system with respect to the DAC clock period over temperature. • • • • PARERR—triggered when one or more parity errors occurs on the data bus PARMSET—triggered when PARMNEW is set and internally registered PARMCLR—triggered when PARMNEW is cleared and internally registered LOCKACQ—triggered when the mu controller is locked to the user-defined phase LOCKLOST—triggered when the mu controller loses lock (if the LOCKACQ bit was previously set) SATERR—triggered when one or more saturation errors occurs Each IRQ is enabled using the enable bits in the interrupt enable register, Register 0x03. The status of the IRQ can be measured in one of the following ways: via the SPI bits found in the interrupt status/clear register (Register 0x04) or using the IRQ pin (Pin P2). If the pin is used to determine that an interrupt has occurred, it is necessary to check Register 0x04 to determine which bit caused the interrupt because the pin indicates only that an interrupt has occurred. To clear an IRQ, it is necessary to write a 1 to the bit in Register 0x04 that corresponds to the interrupt. Rev. A | Page 61 of 76 AD9789 RECOMMENDED START-UP SEQUENCE The steps necessary to optimize the performance of the part and generate an output waveform are listed in Table 79. Table 79. Recommended System Start-Up Sequence Step 0 0 1 1 2 3 4 4 4 4 5 6 7 8 9 9 9 9 9 9 9 10 11 12 13 14 15 1 Description Power up the AD9789. Apply the clock. Enable the clock receiver and set the clock CML. Enable duty cycle correction. Enable digital clocks. Set up mu controller. Disable all interrupts. Clear all interrupts. Enable mu control interrupts. Enable mu delay controller. Set up digital datapath. Set up rate converter. Set up BPF center frequency. Set up interface. Set up channel gains. Set up spectral invert. Set up full-scale current. Wait until mu delay controller is locked (SPI read) 1 . Update rate converter and BPF. Update interface clocks. Enable channels. Enable other interrupts if desired. Register Data 0x32 0x30 0x24 0x24 0x2F 0x33 0x39 0x3A 0x03 0x04 0x03 0x33 0x06 to 0x15 0x16 to 0x1B 0x1C to 0x1D 0x20 to 0x23 0x25 to 0x28 0x29 0x3C to 0x3D 0x04 0x1E 0x24 0x24 0x05 0x03 0x9E 0x80 0x00 0x80 0xCE 0x42 0x4E 0x6C 0x00 0xFE 0x0C 0x43 Typical lock time of the mu controller is approximately 180,000 DAC cycles (at 2 GSPS, ~75 μs). Rev. A | Page 62 of 76 0x08 0x80 0x00 0x80 AD9789 CUSTOMER BIST MODES USING THE INTERNAL PRN GENERATOR TO TEST QAM OUTPUT AC PERFORMANCE The AD9789 can be configured to enable an on-chip pseudorandom number (PRN) generator. The PRN output is connected to the front end of the datapath and disconnects the datapath from the input pins. In this way, the PRN generator can be used in conjunction with the on-chip QAM encoder to generate a QAM output. The PRN generator allows the user to measure the ac performance of a QAM signal at the DAC output without an external data source. To enable the internal PRN generator via the serial port, follow these steps. 1. Ensure that the clock is enabled and that the clock common-mode level is set to its optimal value by setting the registers in Table 80 to the values shown in the table. After the PRN generator is started, users can freely configure the datapath for their desired test configuration as long as Register 0x40 to Register 0x55 are not modified. To disable the PRN generator, write 0x00 to Register 0x40. USING THE INTERNAL BUILT-IN SELF-TEST (BIST) TO TEST FOR DIGITAL DATA INPUT CONNECTIVITY The AD9789 includes an internal built-in self-test (BIST) engine that processes incoming data and creates a signature that can be read back via the serial port. This BIST feature can be configured to observe the static state of the digital data input pins (L4 to L12, M4 to M12, N5 to N12, and P5 to P12) and to reflect the state of these pins via the signature registers (Register 0x50 to Register 0x55). In this way, the user can verify digital data input connectivity. Table 80. Register Settings to Configure the Clock Testing Connectivity for LVDS Interface Mode Register 0x30 0x31 Data 0x80 0xF0 To test the connectivity of the digital data input pins in LVDS interface mode, follow these steps. 0x32 0x9E 2. Description Enable duty cycle correction. Set the common-mode level of CLKN: CLKN_CML = 0xF. Set the common-mode level of CLKP: CLKP_CML = 0xF. Set PSIGN = 0, NSIGN = 0. Enable clock receiver (CLK_DIS = 1). Configure BIST mode for PRN generation and disconnect the inputs by setting the registers in Table 81 to the values shown in the table. 1. Ensure that the clock is enabled and that the clock common-mode level is set to its optimal value by setting the registers in Table 83 to the values shown in the table. Table 83. Register Settings to Configure the Clock Register 0x30 0x31 Data 0x80 0xF0 0x32 0x9E Table 81. Register Settings to Configure PRN Generation Register 0x42 0x43 0x44 0x45 0x46 0x47 0x49 0x4B 0x4C 0x4D 0x05 3. 4. Setting 0x10 0x00 0x10 0x00 0x00 0x10 0x16 0x17 0x4E 0x1F 0x0F 2. 3. Cycle the PARMNEW bit to ensure that the digital clocks are active by first setting Register 0x24 to 0x00, and then setting Register 0x24 to 0x80. Configure the LVDS interface for high speed, 16-bit bus width, 16-bit data width operation by setting the registers in Table 84 to the values shown in the table. Table 84. Register Settings for LVDS Interface Cycle the PARMNEW bit to ensure that the digital clocks are active by first setting Register 0x24 to 0x00, and then setting Register 0x24 to 0x80. Start PRN generation by setting the registers in Table 82 to the values shown in the table. Register 0x20 0x21 0x22 0x23 Table 82. Register Settings to Start PRN Generation Register 0x48 0x4A 0x40 Description Enable duty cycle correction. Set the common-mode level of CLKN: CLKN_CML = 0xF. Set the common-mode level of CLKP: CLKP_CML = 0xF. Set PSIGN = 0, NSIGN = 0. Enable clock receiver (CLK_DIS = 1). Setting 0xAB 0xAB 0x56 Rev. A | Page 63 of 76 Setting 0x08 0x41 0x1F 0x87 AD9789 4. Configure pin mode by setting the registers in Table 85 to the values shown in the table. Table 85. Register Settings to Configure Pin Modes Register 0x42 0x43 0x44 0x45 0x46 0x47 0x49 0x4B 0x4C 0x4D 5. 6. 7. 3. Setting 0x00 0x08 0x00 0x08 0x00 0x10 0x1C 0x1C 0x00 0x00 Register 0x20 0x21 0x22 0x23 4. Table 86. Register Settings for BIST Pin Test 8. Setting 0x80 0x80 0x55 Read back the signature registers (Register 0x50 to Register 0x55) to determine the pin states (see Table 87). 5. Ensure that the clock is enabled and that the clock common-mode level is set to its optimal value by setting the registers in Table 88 to the values shown in the table. Table 88. Register Settings to Configure the Clock Data 0x80 0xF0 0x32 0x9E Cycle the PARMNEW bit to ensure that the interface configuration was updated by first setting Register 0x24 to 0x00, and then setting Register 0x24 to 0x80. Apply static CMOS data to the input ports. Enable the BIST pin test by setting the registers in Table 91 to the values shown in the table. Register 0x48 0x4A 0x40 To test the connectivity of the digital data input pins in CMOS interface mode, follow these steps. Register 0x30 0x31 Setting 0x00 0x08 0x00 0x08 0x00 0x10 0x1C 0x1C 0x00 0x00 Table 91. Register Settings for BIST Pin Test Testing Connectivity for CMOS Interface Mode 1. Configure pin mode by setting the registers in Table 90 to the values shown in the table. Register 0x42 0x43 0x44 0x45 0x46 0x47 0x49 0x4B 0x4C 0x4D 6. 7. Associated LVDS Pairs Data bits D[7:0] Data bits D[15:8] Parity PAR Data bits D[7:0] (repeated) Data bits D[15:8] (repeated) Parity PAR (repeated) Setting 0x08 0x61 0x1F 0x87 Table 90. Register Settings to Configure Pin Modes Table 87. Signature Register Settings Register 0x50 0x51 0x52 0x53 0x54 0x55 Cycle the PARMNEW bit to ensure that the digital clocks are active by first setting Register 0x24 to 0x00, and then setting Register 0x24 to 0x80. Configure the CMOS interface for high speed, 32-bit bus width, 16-bit data width operation by setting the registers in Table 89 to the values shown in the table. Table 89. Register Settings for CMOS Interface Cycle the PARMNEW bit to ensure that the interface configuration was updated by first setting Register 0x24 to 0x00, and then setting Register 0x24 to 0x80. Apply static LVDS data to the input ports. Enable the BIST pin test by setting the registers in Table 86 to the values shown in the table. Register 0x48 0x4A 0x40 2. Description Enable duty cycle correction. Set the common-mode level of CLKN: CLKN_CML = 0xF. Set the common-mode level of CLKP: CLKP_CML = 0xF. Set PSIGN = 0, NSIGN = 0. Enable clock receiver (CLK_DIS = 1). 8. Setting 0x80 0x80 0x55 Read back the signature registers (Register 0x50 to Register 0x55) to determine the pin states (see Table 92). Table 92. Signature Register Settings Register 0x50 0x51 0x52 0x53 0x54 0x55 Rev. A | Page 64 of 76 Associated CMOS Pairs Data bits D[23:16] Data bits D[31:24] Parity P1 Data bits D[7:0] Data bits [D15:8] Parity P0 AD9789 QAM CONSTELLATION MAPS Q IKQK = 10 Q 1001 0010 0011 1010 1000 0000 0001 1101 1100 0100 0110 1111 1110 0101 0111 IKQK = 11 10011 00110 00010 10010 10101 10001 00100 00101 00111 10110 10100 10000 00000 00001 00011 I 11011 11001 11000 01000 01100 01110 IKQK = 01 11111 11101 11100 01001 01101 01010 IKQK ARE THE TWO MSBs IN EACH QUADRANT. 11010 11110 01011 01111 IKQK = 11 I IKQK = 01 07852-087 1011 IKQK = 00 10111 IKQK = 00 07852-086 IKQK = 10 IKQK ARE THE TWO MSBs IN EACH QUADRANT. Figure 121. DVB-C 16-QAM Constellation Figure 123. DVB-C 32-QAM Constellation 11 Q IKQK = 10 101100 101110 100110 100100 001000 001001 001101 001100 101101 101111 100111 100101 001010 001011 001111 001110 101001 101011 100011 100001 9 IKQK = 00 IKQK = 10 7 π/2 ROTATION 000010 000011 000111 000110 11010 11011 01011 01010 11000 11001 01001 01000 10000 10001 10101 10100 11100 11101 10010 10011 10111 10110 11110 11111 00010 00011 00111 00110 01110 01111 00000 00001 00101 00100 01100 01101 1 3 5 7 9 11 IKQK = 00 101000 101010 100010 100000 000000 000001 000101 000100 110100 110101 110001 110000 010000 010010 011010 011000 110110 110111 110011 110010 010001 010011 011011 011001 111110 111111 111011 111010 010101 010111 011111 011101 111100 111101 111001 111000 010100 010110 011110 011100 3 I 1 IKQK = 01 IKQK ARE THE TWO MSBs IN EACH QUADRANT. IKQK = 11 IKQK = 01 π ROTATION 3π/2 ROTATION 07852-088 IKQK = 11 IKQK ARE THE TWO MSBs IN EACH QUADRANT. Figure 124. DVB -C128-QAM Constellation3 Figure 122. DVB-C 64-QAM Constellation Rev. A | Page 65 of 76 07852-089 5 AD9789 13 IKQK = 10 π/2 ROTATION 11 9 7 5 3 1 0000 0001 0101 0100 0100 0101 0001 0000 0010 0011 0111 0110 0110 0111 0011 0010 1011 1111 1110 1110 1111 1011 1010 1010 10 11 1101 1001 1000 1000 1001 1101 1100 1100 1101 1001 1000 1010 1011 1111 1110 1110 1111 1011 1010 0010 0011 0111 0110 0110 0111 0011 0010 0101 0001 0000 11 13 15 10 1000 1001 1101 1100 11 1100 00 00 IKQK = 00 01 0000 0001 0101 0100 01 0100 1 3 5 7 9 IKQK = 11 IKQK = 01 π ROTATION 3π/2 ROTATION 07852-090 15 IKQK ARE THE TWO MSBs IN EACH QUADRANT. Figure 125. DVB-C 256-QAM Constellation Q C5 C4 C3, C2 C1 C0 110,111 111,011 010,111 011,011 100,101 101,111 110,101 111,111 110,100 111,000 010,100 011,000 100,000 101,010 110,000 111,010 100,111 101,011 000,111 001,011 000,101 001,111 010,101 011,111 100,100 101,000 000,100 001,000 000,000 001,010 010,000 011,010 I 010,011 011,001 000,011 001,001 000,001 001,101 100,001 101,101 010,110 011,100 000,110 001,100 000,010 001,110 100,010 101,110 110,011 111,001 100,011 101,001 010,001 011,101 110,001 111,101 07852-091 110,110 111,100 100,110 101,100 010,010 011,110 110,010 111,110 Figure 126. DOCSIS 64-QAM Constellation Rev. A | Page 66 of 76 AD9789 Q 1110, 1111, 1110, 1111, 1110, 1111, 1110, 1111, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1111 1101 1011 1001 0111 0101 0011 0001 1111 1111 1111 1111 1111 1111 1111 1111 C7 C6 C5 C4, C3 C2 C1 C0 1100, 1101, 1100, 1101, 1100, 1101, 1100, 1101, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1110 1100 1010 1000 0110 0100 0010 0000 1100 1100 1100 1100 1100 1100 1100 1100 1010, 1011, 1010, 1011, 1010, 1011, 1010, 1011, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1111 1101 1011 1001 0111 0101 0011 0001 1011 1011 1011 1011 1011 1011 1011 1011 1000, 1001, 1000, 1001, 1000, 1001, 1000, 1001, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1110 1100 1010 1000 0110 0100 0010 0000 1000 1000 1000 1000 1000 1000 1000 1000 0110, 0111, 0110, 0111, 0110, 0111, 0110, 0111, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1111 1101 1011 1001 0111 0101 0011 0001 0111 0111 0111 0111 0111 0111 0111 0111 0100, 0101, 0100, 0101, 0100, 0101, 0100, 0101, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1110 1100 1010 1000 0110 0100 0010 0000 0100 0100 0100 0100 0100 0100 0100 0100 0010, 0011, 0010, 0011, 0010, 0011, 0010, 0011, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1111 1101 1011 1001 0111 0101 0011 0001 0011 0011 0011 0011 0011 0011 0011 0011 0000, 0001, 0000, 0001, 0000, 0001, 0000, 0001, 0000, 0011, 0100, 0111, 1000, 1011, 1100, 1111, 1110 1100 1010 1000 0110 0100 0010 0000 0000 0000 0000 0000 0000 0000 0000 0000 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 0000, 0001, 0000, 0001, 0000, 0001, 0000, 0001, 0001 0001 0001 0001 0001 0001 0001 0001 0001 0011 0101 0111 1001 1011 1101 1111 I 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 0010, 0011, 0010, 0011, 0010, 0011, 0010, 0011, 0010 0010 0010 0010 0010 0010 0010 0010 0000 0010 0100 0110 1000 1010 1100 1110 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 0100, 0101, 0100, 0101, 0100, 0101, 0100, 0101, 0101 0101 0101 0101 0101 0101 0101 0101 0001 0011 0101 0111 1001 1011 1101 1111 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 0110, 0111, 0110, 0111, 0110, 0111, 0110, 0111, 0110 0110 0110 0110 0110 0110 0110 0110 0000 0010 0100 0110 1000 1010 1100 1110 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 1000, 1001, 1000, 1001, 1000, 1001, 1000, 1001, 1001 1001 1001 1001 1001 1001 1001 1001 0001 0011 0101 0111 1001 1011 1101 1111 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 1010, 1011, 1010, 1011, 1010, 1011, 1010, 1011, 1010 1010 1010 1010 1010 1010 1010 1010 0000 0010 0100 0110 1000 1010 1100 1110 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 1110, 1111, 1110, 1111, 1110, 1111, 1110, 1111, 1110 1110 1110 1110 1110 1110 1110 1110 0000 0010 0100 0110 1000 1010 1100 1110 Figure 127. DOCSIS 256-QAM Constellation Rev. A | Page 67 of 76 07852-126 1110, 1101, 1010, 1001, 0110, 0101, 0010, 0001, 1100, 1101, 1100, 1101, 1100, 1101, 1100, 1101, 1101 1101 1101 1101 1101 1101 1101 1101 0001 0011 0101 0111 1001 1011 1101 1111 AD9789 CHANNELIZER MODE PIN MAPPING FOR CMOS AND LVDS Table 94 and Table 95, along with Figure 128 and Figure 129, describe CMOS and LVDS data input pin mapping. CMOS mode is always single data rate and samples on the rising edge of DSC. LVDS mode is single data rate (SDR) for bus widths of 4 bits through 16 bits and double data rate (DDR) for a bus width of 32 bits. Table 93. Data Input Configurations for Channelizer Mode Bus Width 4 4 8 8 8 16 16 16 32 32 32 Data Width 8 8 8 8 16 8 8 16 8 8 16 Data Format Real Complex Real Complex Complex Real Complex Complex Real Complex Complex 1 2 3 4 5 6 7 8 9 10 11 12 13 14 A B C D E F G H J K L P1 31 27 23 19 15 11 7 3 BU CMOS_BUS M P0 30 26 22 18 14 10 6 2 CT CMOS_CTRL N 29 25 21 17 13 9 5 1 FS CMOS_FS P 28 24 20 16 12 8 4 0 DC CMOS_DCO D[31:0] CMOS DATA INPUTS 07852-127 Table 93 lists the available combinations of data input configuration parameters when the AD9789 is in channelizer mode. Many of these configurations require multiple clocks to load all channels. All of these configurations are described in detail in Table 96 and Table 97. PARITY AND CONTROL INPUTS Figure 128. CMOS Data Input Pin Mapping 1 2 3 4 5 6 7 8 9 10 11 12 13 14 A B C D E F Table 94. CMOS Pin Assignments for Various Interface Widths Pin Assignments D[3:0] D[7:0] D[15:0] D[31:0] BUSWDTH[1:0] 00 01 10 11 H PARP J PARN K Table 95. LVDS Pin Assignments for Various Interface Widths Interface Width 4 bits 8 bits 16 bits 32 bits Pin Assignments D[3:0]P, D[3:0]N D[7:0]P, D[7:0]N D[15:0]P, D[15:0]N D[15:0]P, D[15:0]N rising edge and falling edge BUSWDTH[1:0] 00 01 10 11 Rev. A | Page 68 of 76 FSP L P+ 15 13 11 9 7 5 3 1 FS FSN M P– 15 13 11 9 7 5 3 1 FS DCOP N 14 12 10 8 6 4 2 0 DC DCON P 14 12 10 8 6 4 2 0 DC 14 +LVDS 14 –LVDS Figure 129. LVDS Data Input Pin Mapping 07852-128 Interface Width 4 bits 8 bits 16 bits 32 bits G AD9789 In Table 96, “R” represents real data loaded to a given channel, “I” represents the in-phase term, and “Q” represents the quadrature term of complex data. The channel number follows R, I, or Q. Table 96. Channelizer Mode Configurations and Channel Construction: CMOS Interface, Channel Prioritization = 1 Datapath Configuration BW DW Format 4 8 Real Datapath Configuration BW DW Format 4 8 Complex Datapath Configuration BW DW Format 8 8 Real Datapath Configuration BW DW Format 8 8 Complex DCO 1 2 3 4 5 6 7 8 [D31:D28] [D27:D24] CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] [D11:D8] [D7:D4] [D3:D0] R0 R0 R1 R1 R2 R2 R3 R3 [D11:D8] [D7:D4] [D3:D0] I0 I0 Q0 Q0 I1 I1 Q1 Q1 I2 I2 Q2 Q2 I3 I3 Q3 Q3 [D11:D8] [D7:D4] DCO 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 [D31:D28] [D27:D24] CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] DCO 1 2 3 4 [D31:D28] [D27:D24] CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] DCO 1 2 3 4 5 6 7 8 [D31:D28] [D3:D0] R0 R1 R2 R3 [D27:D24] CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] [D11:D8] [D7:D4] [D3:D0] I0 Q0 I1 Q1 I2 Q2 I3 Q3 Rev. A | Page 69 of 76 AD9789 Datapath Configuration BW DW Format 8 16 Complex Datapath Configuration BW DW Format 16 8 Real Datapath Configuration BW DW Format 16 8 Complex Datapath Configuration BW DW Format 16 16 Complex Datapath Configuration BW DW Format 32 8 Real Datapath Configuration BW DW Format 32 8 Complex Datapath Configuration BW DW Format 32 16 Complex DCO 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 [D31:D28] DCO 1 2 [D31:D28] DCO 1 2 3 4 DCO 1 2 3 4 5 6 7 8 DCO 1 DCO 1 2 DCO 1 2 3 4 [D27:D24] CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] [D11:D8] [D7:D4] [D3:D0] I0 I0 Q0 Q0 I1 I1 Q1 Q1 I2 I2 Q2 Q2 I3 I3 Q3 Q3 CMOS Pin Mapping [D23:D20] [D19:D16] [D15:D12] [D11:D8] R1 R3 CMOS Pin Mapping [D31:D28] [D27:D24] [D23:D20] [D19:D16] [D15:D12] [D11:D8] Q0 Q1 Q2 Q3 CMOS Pin Mapping [D31:D28] [D27:D24] [D23:D20] [D19:D16] [D15:D12] [D11:D8] I0 Q0 I1 Q1 I2 Q2 I3 Q3 CMOS Pin Mapping [D31:D28] [D27:D24] [D23:D20] [D19:D16] [D15:D12] [D11:D8] R3 R2 R1 CMOS Pin Mapping [D31:D28] [D27:D24] [D23:D20] [D19:D16] [D15:D12] [D11:D8] Q1 I1 Q0 Q3 I3 Q2 CMOS Pin Mapping [D31:D28] [D27:D24] [D23:D20] [D19:D16] [D15:D12] [D11:D8] Q0 I0 Q1 I1 Q2 I2 Q3 I3 [D27:D24] Rev. A | Page 70 of 76 [D7:D4] [D3:D0] R0 R2 [D7:D4] [D3:D0] I0 I1 I2 I3 [D7:D4] [D3:D0] [D7:D4] [D3:D0] R0 [D7:D4] [D3:D0] I0 I2 [D7:D4] [D3:D0] AD9789 In DDR mode, “rise” corresponds to data sampled on the rising edge of DSC; “fall” corresponds to data sampled on the falling edge of DSC. Table 97. Channelizer Mode Configurations and Channel Construction: LVDS Interface, Channel Prioritization = 1 Datapath Configuration BW DW Format 4 8 Real Datapath Configuration BW DW Format 4 8 Complex Datapath Configuration BW DW Format 8 8 Real Datapath Configuration BW DW Format 8 8 Complex [D15:D12] LVDS Pin Mapping [D11:D8] [D7:D4] DCO 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 [D15:D12] LVDS Pin Mapping [D11:D8] [D7:D4] DCO 1 2 3 4 [D15:D12] LVDS Pin Mapping [D11:D8] [D7:D4] DCO 1 2 3 4 5 6 7 8 [D15:D12] DCO 1 2 3 4 5 6 7 8 [D3:D0] R0 R0 R1 R1 R2 R2 R3 R3 [D3:D0] I0 I0 Q0 Q0 I1 I1 Q1 Q1 I2 I2 Q2 Q2 I3 I3 Q3 Q3 [D3:D0] R0 R1 R2 R3 LVDS Pin Mapping [D11:D8] [D7:D4] [D3:D0] I0 Q0 I1 Q1 I2 Q2 I3 Q3 Rev. A | Page 71 of 76 AD9789 Datapath Configuration BW DW Format 8 16 Complex Datapath Configuration BW DW Format 16 8 Real Datapath Configuration BW DW Format 16 8 Complex Datapath Configuration BW DW Format 16 16 Complex Datapath Configuration BW DW Format 32 8 Real Datapath Configuration BW DW Format 32 8 Complex DCO 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 [D15:D12] DCO 1 2 [D15:D12] LVDS Pin Mapping [D11:D8] [D7:D4] [D3:D0] I0 I0 Q0 Q0 I1 I1 Q1 Q1 I2 I2 Q2 Q2 I3 I3 Q3 Q3 LVDS Pin Mapping [D11:D8] [D7:D4] R1 R3 DCO 1 2 3 4 [D15:D12] DCO 1 2 3 4 5 6 7 8 [D15:D12] DCO 1 rise 1 fall [D15:D12] DCO 1 rise 1 fall 2 rise 2 fall [D15:D12] LVDS Pin Mapping [D11:D8] [D7:D4] Q0 Q1 Q2 Q3 [D3:D0] I0 I1 I2 I3 LVDS Pin Mapping [D11:D8] [D7:D4] I0 Q0 I1 Q1 I2 Q2 I3 Q3 LVDS Pin Mapping [D11:D8] [D7:D4] R1 R3 [D3:D0] [D3:D0] R0 R2 LVDS Pin Mapping [D11:D8] [D7:D4] Q0 Q1 Q2 Q3 Rev. A | Page 72 of 76 [D3:D0] R0 R2 [D3:D0] I0 I1 I2 I3 AD9789 Datapath Configuration BW DW Format 32 16 Complex DCO 1 rise 1 fall 2 rise 2 fall 3 rise 3 fall 4 rise 4 fall [D15:D12] Rev. A | Page 73 of 76 LVDS Pin Mapping [D11:D8] [D7:D4] I0 Q0 I1 Q1 I2 Q2 I3 Q3 [D3:D0] AD9789 OUTLINE DIMENSIONS 12.00 BSC SQ A1 BALL CORNER 14 13 12 11 10 9 8 7 6 5 4 3 2 1 A B C D E 10.40 BSC SQ F G H J 0.80 BSC K L M N P *1.30 1.22 1.14 0.80 REF BOTTOM VIEW DETAIL A 0.65 REF DETAIL A 0.38 0.33 0.28 0.24 REF SEATING PLANE 0.96 0.89 0.82 0.53 COPLANARITY 0.08 0.48 0.43 BALL DIAMETER *COMPLIANT TO JEDEC STANDARDS MO-219 WITH THE EXCEPTION TO PACKAGE HEIGHT. 111808-A TOP VIEW Figure 130. 164-Ball Chip Scale Package Ball Grid Array [CSP_BGA] (BC-164-1) Dimensions shown in millimeters ORDERING GUIDE Model 1 AD9789BBCZ AD9789BBCZRL AD9789BBC AD9789BBCRL AD9789-EBZ AD9789-MIX-EBZ 1 Temperature Range −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C Package Description 164-Ball Chip Scale Package Ball Grid Array (CSP_BGA) 164-Ball Chip Scale Package Ball Grid Array (CSP_BGA) 164-Ball Chip Scale Package Ball Grid Array (CSP_BGA) 164-Ball Chip Scale Package Ball Grid Array (CSP_BGA) Evaluation Board for CMTS and Normal Mode Evaluation Evaluation Board for Mix Mode Evaluation Z = RoHS Compliant Part. Rev. A | Page 74 of 76 Package Option BC-164-1 BC-164-1 BC-164-1 BC-164-1 AD9789 NOTES Rev. A | Page 75 of 76 AD9789 NOTES ©2009-2011 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D07852-0-7/11(A) Rev. A | Page 76 of 76