DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 16-BIT, 500 MSPS 2x-8x INTERPOLATING DUAL-CHANNEL DIGITAL-TO-ANALOG CONVERTER (DAC) FEATURES • • • • • • • • • • • • • 500 MSPS Selectable 2x-8x Interpolation On-Chip PLL/VCO Clock Multiplier Full IQ Compensation Including Offset, Gain, and Phase Flexible Input Options: – FIFO With Latch on External or Internal Clock – Even/Odd Multiplexed Input – Single Port Demultiplexed Input Complex Mixer With 32-Bit NCO Fixed Frequency Mixer With Fs/4 and Fs/2 1.8-V or 3.3-V I/O Voltage On-Chip 1.2-V Reference Differential Scalable Output: 2 mA to 20 mA Pin Compatible to DAC5686 High Performance – 81-dBc ACLR WCDMA TM1 at 30.72 MHz – 72-dBc ACLR WCDMA TM1 at 153.6 MHz Package: 100-Pin HTQFP APPLICATIONS • • Cellular Base Transceiver Station Transmit Channel – CDMA: W-CDMA, CDMA2000, TD-SCDMA – TDMA: GSM, IS-136, EDGE/UWC-136 – OFDM: 802.16 Cable Modem Termination System DESCRIPTION The DAC5687 is a dual-channel 16-bit high-speed digital-to-analog converter (DAC) with integrated 2x, 4x, and 8x interpolation filters, a complex numerically controlled oscillator (NCO), on-board clock multiplier, IQ compensation and on-chip voltage reference. The DAC5687 is pin compatible to the DAC5686, requiring only changes in register settings for most applications, and offers additional features and superior linearity, noise, crosstalk, and PLL phase noise performance. The DAC5687 has six signal processing blocks: two interpolate by two digital filters, fine frequency mixer with 32-bit NCO, a quadrature modulation compensation block, another interpolate by two digital filter and a coarse frequency mixer with Fs/2 or Fs/4. The different modes of operation enable or bypass the signal processing blocks. The coarse and fine mixers can be combined to span a wider range of frequencies with fine resolution. The DAC5687 allows both complex or real output. Combining the frequency upconversion and complex output produces a Hilbert Transform pair that is output from the two DACs. An external RF quadrature modulator then performs the final single sideband up-conversion. The IQ compensation feature allows optimization of phase, gain and offset to maximize sideband rejection and minimize LO feedthrough for an analog quadrature modulator. The DAC5687 includes several input options: single port interleaved data, even and odd multiplexing at half rate, and an input FIFO with either external or internal clock to ease the input timing ambiguity when the DAC5687 is clocked at the DAC output sample rate. ORDERING INFORMATION TA Package Device -40°C to 85°C 100 HTQFP (1) (P&P) PowerPAD™ Plastic Quad FlatPack DAC5687IPZP (1) Thermal Pad Size: 6 mm × 6 mm. Please be aware that an important notice concerning availability, standard warranty, and use in critical applications of Texas Instruments semiconductor products and disclaimers thereto appears at the end of this data sheet. PRODUCTION DATA information is current as of publication date. Products conform to specifications per the terms of the Texas Instruments standard warranty. Production processing does not necessarily include testing of all parameters. Copyright © 2005, Texas Instruments Incorporated DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 These devices have limited built-in ESD protection. The leads should be shorted together or the device placed in conductive foam during storage or handling to prevent electrostatic damage to the MOS gates. FUNCTIONAL BLOCK DIAGRAM CLKVDD CLKGND LPF PLLGND PLLVDD PHSTR SLEEP DVDD CLK2 CLK2C A Offset A gain PLLLOCK FIR1 FIR2 FIR3 FIR4 IOUTA1 x2 x2 Quardrature Mod Correction (QMC) x2 Fine Mixer Input FIFO/ Reorder/ Mux/Demux x2 DB[15:0] SIF sin NCO 16-bit DAC IOUTA2 sin(x) IOUTB1 x: x2 cos TXENABLE x2 Coarse Mixer: fs/2 or fs/4 x: DA[15:0] 16-bit DAC sin(x) IOUTB2 B Offset B gain SDIO SDO SDENB SCLK IOGND IOVDD 100-Pin HTQFP QFLAG 2 EXTIO EXTLO BIASJ CLK1C RESETB 1.2-V Reference 2x−8x fdata Internal Clock Generation and 2x−8x PLL Clock Multiplier CLK1 DGND AVDD AGND DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 IOGND DB7 DB6 DB5 79 78 77 76 DGND IOVDD 81 80 DB8 DVDD 83 82 DB10 DB9 85 86 84 DB12 DB11 87 DVDD DGND 89 88 DB14 DB13 91 92 90 DGND DB15 (MSB) 93 RESETB PHSTR 94 96 95 TESTMODE SLEEP 97 DGND QFLAG 98 DVDD 99 100 PINOUT 75 DB4 2 74 DB3 AGND 1 AVDD AVDD 3 73 DB2 AGND 4 72 DB1 IOUTB1 5 71 DB0 (LSB) IOUTB2 6 70 PLLLOCK AGND 7 69 DGND AVDD 8 68 DVDD AGND 9 67 PLLVDD AVDD 10 66 LPF EXTIO 11 65 PLLGND AGND 12 64 CLKGND BIASJ 13 63 CLK2C AVDD 14 62 CLK2 EXTLO 15 61 CLKVDD AVDD 16 60 CLK1C AGND 17 59 CLK1 AVDD 18 58 CLKGND Top View 100 HTQFP DAC5687 47 48 49 50 IOGND DA7 DA6 DA5 45 46 DGND IOVDD 43 44 DA8 41 DA10 DVDD 40 DA11 42 39 DA12 DA9 37 38 DVDD DGND 35 36 DA14 DA13 DA4 34 51 DA15 (MSB) 25 33 AGND TXENABLE DA3 32 DA2 52 DVDD 53 24 31 23 AVDD SDO AVDD 30 DA1 SDIO DA0 (LSB) 54 29 55 22 SCLK 21 AGND 28 IOUTA1 DGND DVDD SDENB DGND 56 26 57 20 27 19 DVDD AGND IOUTA2 TERMINAL FUNCTIONS TERMINAL I/O DESCRIPTION NAME NO. AGND 1, 4, 7, 9, 12, 17, 19, 22, 25 I Analog ground return AVDD 2, 3, 8, 10, 14, 16, 18, 23, 24 I Analog supply voltage BIASJ 13 O Full-scale output current bias 3 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 TERMINAL FUNCTIONS (continued) TERMINAL I/O DESCRIPTION 59 I In PLL clock mode and dual clock modes, provides data input rate clock. In external clock mode, provides optional input data rate clock to FIFO latch. When the FIFO is disabled, CLK1 is not used and can be left unconnected. CLK1C 60 I Complementary input of CLK1. CLK2 62 I External and dual clock mode clock input. In PLL mode, CLK2 is unused and can be left unconnected. NAME NO. CLK1 CLK2C 63 I Complementary of CLK2. In PLL mode, CLK2C is unused and can be left unconnected. CLKGND 58, 64 I Ground return for internal clock buffer CLKVDD 61 I Internal clock buffer supply voltage DA[15..0] 34-36, 39-43, 48-55 I A-Channel Data bits 0 through 15. DA15 is most significant data bit (MSB). DA0 is least significant data bit (LSB). Order can be reversed by register change. DB[0..15] 71-78, 83-87, 90-92 I B-Channel Data bits 0 through 15. DB15 is most significant data bit (MSB). DB0 is least significant data bit (LSB). Order can be reversed by register change. DGND 27, 38, 45, 57, 69, 81, 88, 93, 99 I Digital ground return DVDD 26, 32, 37, 44, 56, 68, 82, 89, 100 I Digital supply voltage EXTIO 11 I/O Used as external reference input when internal reference is disabled (i.e., EXTLO connected to AVDD). Used as internal reference output when EXTLO = AGND, requires a 0.1-µF decoupling capacitor to AGND when used as reference output EXTLO 15 I/O Internal/external reference select. Internal reference selected when tied to AGND, external reference selected when tied to AVDD. Output only when ATEST is not zero (register 0x1B bits 7 to 3). IOUTA1 21 O A-Channel DAC current output. Full scale when all input bits are set 1 IOUTA2 20 O A-Channel DAC complementary current output. Full scale when all input bits are 0 IOUTB1 5 O B-Channel DAC current output. Full scale when all input bits are set 1 IOUTB2 6 O B-Channel DAC complementary current output. Full scale when all input bits are 0 IOGND 47, 79 I Digital I/O ground return IOVDD 46, 80 I Digital I/O supply voltage LPF 66 I PLL loop filter connection PHSTR 94 I Synchronization input signal that can be used to initialize the NCO, course mixer, internal clock divider, and/or FIFO circuits. PLLGND 65 I Ground return for internal PLL PLLVDD 67 I PLL supply voltage. When PLLVDD is 0 V, the PLL is disabled. PLLLOCK 70 O In PLL mode, provides PLL lock status bit or internal clock signal. PLL is locked to input clock when high. In external clock mode, provides input rate clock. QFLAG 98 I When qflag register is 1, the QFLAG pin is used by the user during interleaved data input mode to identify the B sample. High QFLAG indicates B sample. Must be repeated every B sample. RESETB 95 I Resets the chip when low. Internal pull-up SCLK 29 I Serial interface clock SDENB 28 I Active low serial data enable, always an input to the DAC5687 SDIO 30 I/O Bidirectional serial data in 3-pin interface mode, input only in 4 pin interface mode. Three-pin mode is the default after chip reset. SDO 31 O Serial interface data, uni-directional data output, if SDIO is an input. SDO is 3-stated when the 3-pin interface mode is selected (register 0x08 bit 1). SLEEP 96 I Asynchronous hardware power down input. Active High. Internal pull down. 4 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 TERMINAL FUNCTIONS (continued) TERMINAL NAME NO. I/O DESCRIPTION TXENABLE 33 I TXENABLE has two purposes. In all modes, TXENABLE must be high for the DATA to the DAC to be enabled. When TXENABLE is low, the digital logic section is forced to all 0, and any input data presented to DA[15:0] and DB[15:0] is ignored. In interleaved data mode, when the qflag register bit is cleared, TXENABLE is used to synchronizes the data to channels A and B. The first data after the rising edge of TXENABLE is treated as A data, while the next data is treated as B data, and so on. TESTMODE 97 I TESTMODE is DGND for the user ABSOLUTE MAXIMUM RATINGS over operating free-air temperature range (unless otherwise noted) (1) UNIT Supply voltage range AVDD (2) –0.5 V to 4 V DVDD (3) –0.5 V to 2.3 V CLKVDD (2) –0.5 V to 4 V IOVDD (2) –0.5 V to 4 V PLLVDD (2) –0.5 V to 4 V Voltage between AGND, DGND, CLKGND, PLLGND, and IOGND –0.5 V to 0.5 V AVDD to DVDD –0.5 V to 2.6 V DA[15..0] (4) –0.5 V to IOVDD + 0.5 V DB[15..0] (4) –0.5 V to IOVDD + 0.5 V SLEEP (4) Supply voltage range –0.5 V to IOVDD + 0.5 V CLK1/2, CLK1/2C (3) –0.5 V to CLKVDD + 0.5 V RESETB (4) –0.5 V to IOVDD + 0.5 V LPF (4) –0.5 V to PLLVDD + 0.5 V IOUT1, IOUT2 (2) –1.0 V to AVDD + 0.5 V EXTIO, BIASJ (2) –0.5 V to AVDD + 0.5 V EXTLO (2) –0.5 V to IAVDD + 0.5 V Peak input current (any input) 20 mA Peak total input current (all inputs) 30 mA TA Operating free-air temperature range (DAC5687I) –40°C to 85°C Tstg Storage temperature range –65°C to 150°C Lead temperature 1,6 mm (1/16 inch) from the case for 10 seconds (1) (2) (3) (4) 260°C Stresses beyond those listed under absolute maximum ratings may cause permanent damage to the device. These are stress ratings only, and functional operation of these or any other conditions beyond those indicated under recommended operating conditions is not implied. Exposure to absolute-maximum-rated conditions for extended periods may affect device reliability. Measured with respect to AGND. Measured with respect to DGND. Measured with respect to IOGND. THERMAL CHARACTERISTICS (1) over operating free-air temperature range (unless otherwise noted) Thermal Conductivity TJ θJA θJC (1) (2) Junction temperature (2) 100 HTQFP UNIT 105 °C Theta junction-to-ambient (still air) 19.88 °C/W Theta junction-to-ambient (150 lfm) 14.37 °C/W Theta junction-to-case 0.12 °C/W Air flow or heat sinking reduces θJA and is highly recommended. Air flow or heat sinking required for sustained operation at 85°C and maximum operating conditions to maintain junction temperature. 5 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 3.3 V, IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA (unless otherwise noted) DC SPECIFICATIONS PARAMETER TEST CONDITIONS RESOLUTION MIN TYP MAX 16 UNIT Bits DC ACCURACY (1) INL Integral nonlinearity DNL Differential nonlinearity 1 LSB = IOUTFS/216 TMIN to TMAX ±4 LSB ±4 LSB ±0.04 LSB ±3 LSB ANALOG OUTPUT Coarse gain linearity Worst case error from ideal linearity Fine gain linearity Offset error Mid code offset Without internal reference Gain error With internal reference With internal reference, dual DAC, and SSB mode Gain mismatch 0.01 %FSR 1 %FSR 0.7 %FSR -2 2 %FSR current (2) 2 mA Maximum full-scale output current (2) 20 mA Minimum full-scale output Output compliance range (3) IOUTFS = 20 mA AVDD –0.5 V Output resistance Output capacitance AVDD + 0.5 V V 300 kΩ 5 pF REFERENCE OUTPUT Reference voltage 1.14 Reference output current (4) 1.2 1.26 100 V nA REFERENCE INPUT VEXTIO Input voltage range 0.1 Input resistance 1.25 V 1 MΩ Small signal bandwidth 1.4 MHz Input capacitance 100 pF ±1 ppm of FSR/°C TEMPERATURE COEFFICIENTS Offset drift Gain drift Without internal reference ±15 With internal reference ±30 Reference voltage drift ppm of FSR/°C ppm of FSR/°C ±8 POWER SUPPLY AVDD Analog supply voltage DVDD CLKVDD IOVDD I/O supply voltage PLLVDD PLL supply voltage (1) (2) (3) (4) 6 3 3.3 3.6 V Digital supply voltage 1.71 1.8 2.15 V Clock supply voltage 3 3.3 3.6 V 3.6 V 3.6 V 1.71 3 3.3 Measured differential across IOUTA1 and IOUTA2 or IOUTB1 and IOUTB2 with 25 Ω each to AVDD. Nominal full-scale current, IOUTFS , equals 32X the IBIAS current. The lower limit of the output compliance is determined by the CMOS process. Exceeding this limit may result in transistor breakdown, resulting in reduced reliability of the DAC5687 device. The upper limit of the output compliance is determined by the load resistors and full-scale output current. Exceeding the upper limit adversely affects distortion performance and integral nonlinearity. Use an external buffer amplifier with high impedance input to drive any external load. DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS (continued) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 3.3 V, IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA (unless otherwise noted) DC SPECIFICATIONS PARAMETER IAVDD Analog supply current current (5) TEST CONDITIONS MIN TYP MAX UNIT Mode 5 (5) 41 mA Mode 6 (5) 80 mA 6 (5) IDVDD Digital supply 587 mA ICLKVDD Clock supply current (5) Mode 6 (5) 5 mA IPLLVDD PLL supply current (5) Mode 6 (5) 20 mA 6 (5) current (5) Mode IIOVDD IO supply 2 mA IAVDD Sleep mode AVDD supply current Sleep mode (Sleep pin high) CLK2 = 500 MHz 1 mA IDVDD Sleep mode DVDD supply current Sleep mode (Sleep pin high) CLK2 = 500 MHz 2 mA ICLKVDD Sleep mode CLKVDD supply current Sleep mode (Sleep pin high) CLK2 = 500 MHz 0.25 mA IPLLVDD Sleep mode PLLVDD supply current Sleep mode (Sleep pin high) CLK2 = 500 MHz 0.6 mA IIOVDD Sleep mode IOVDD supply current Sleep mode (Sleep pin high) CLK2 = 500 MHz 0.6 mA Mode 1 (5) AVDD = 3.3 V, DVDD = 1.8 V 750 Mode 2 (5) AVDD = 3.3 V, DVDD = 1.8 V 910 PD Power dissipation Mode Mode 3 (5) AVDD = 3.3 V, DVDD = 1.8 V 760 Mode 4 (5) AVDD = 3.3 V, DVDD = 1.8 V 1250 Mode 5 (5) AVDD = 3.3 V, DVDD = 1.8 V 1250 Mode 6 (5) AVDD = 3.3 V, DVDD = 1.8 V 1410 Mode 7 (5) AVDD = 3.3 V, DVDD = 1.8 V 1400 1750 11 20 Sleep mode (Sleep pin high) CLK2 = 500 MHz APSRR DPSRR (5) Power supply rejection ratio mW -0.2 0.2 %FSR/V -0.2 0.2 %FSR/V MODE 1 – MODE 7: a. Mode 1: X2, PLL off, CLK2 = 320 MHz, DACA and DACB on, IF = 5 MHz b. Mode 2: X4 QMC, PLL on, CLK1 = 125 MHz, DACA and DACB on, IF = 5 MHz c. Mode 3: X4 CMIX, PLL off, CLK2 = 500 MHz, DACA off and DACB on, IF = 150 MHz d. Mode 4: X4L FMIX CMIX, PLL off, CLK2 = 500 MHz, DACA off and DACB on, IF = 150 MHz e. Mode 5: X4L FMIX CMIX, PLL on, CLK1 = 125 MHz, DACA off and DACB on, IF = 150 MHz f. Mode 6: X4L FMIX CMIX, PLL on, CLK1 = 125 MHz, DACA on and DACB on, IF = 150 MHz g. Mode 7: X8 FMIX CMIX, PLL on, CLK1 = 62.5 MHz, DACA and DACB on, IF = 150 MHz 7 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS (1) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 0 V (= 3.3 V for PLL Clock Mode), IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA, External Clock Mode, 4:1 transformer output termination, 50-Ω doubly terminated load (unless otherwise noted) AC SPECIFICATIONS PARAMETER TEST CONDITIONS MIN TYP MAX UNIT ANALOG OUTPUT fCLK Maximum output update rate 500 ts(DAC) Output settling time to 0.1% 10.4 ns tpd Output propagation delay 3 ns tr(IOUT) Output rise time 10% to 90% 2 ns tf(IOUT) Output fall time 90% to 10% 2 ns Transition: Code 0x0000 to 0xFFFF MSPS AC PERFORMANCE SFDR Spurious free dynamic range (2) X2, PLL off, CLK2 = 250 MHz, DAC A and DAC B on, IF = 5.1 MHz, First Nyquist Zone < fDATA/2 78 X4, PLL off, CLK2 = 500 MHz, DAC A and DAC B on, IF = 5.1 MHz, First Nyquist Zone < fDATA/2 77 X4, CLK2 = 500 MHz, DAC A and DAC B on, IF = 20.1 MHz, PLL on for Min, PLL off for TYP, First Nyquist Zone < fDATA/2 SNR IMD3 IMD (1) (2) (3) 8 Signal-to-noise ratio Third-order two-tone intermodulation (each tone at –6 dBFS) Four-tone Intermodulation to Nyquist (each tone at –12 dBFS) 68 (3) dBc 76 X4, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, Single tone, 0 dBFS, IF = 20.1 MHz 73 X4 CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, IF = 70.1 MHz 65 X4 CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, Single tone, 0 dBFS, IF = 150.1 MHz 57 X4 FMIX CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, Single tone, 0 dBFS, IF = 180.1 MHz 54 X4, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on,Four tone, each -12 dBFS, IF = 24.7, 24.9, 25.1, 25.3 MHz 73 X4, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, IF = 20.1 and 21.1 MHz 79 X4 CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, IF = 70.1 and 71.1 MHz 73 X4 CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, IF= 150.1 and 151.1 MHz 68 X4 FMIX CMIX, PLL off, CLK2 = 500 MSPS, DAC A and DAC B on, IF = 180.1 and 181.1 MHz 67 X4 CMIX, CLK2 = 500 MHz fOUT = 149.2, 149.6, 150.4, and 150.8 MHz 66 dBc dBc Measured single ended into 50-Ω load. See the Non-Harmonic Clock Related Spurious Signals section for information on spurious products out of band (< fDATA/2). 1:1 transformer output termination. dBc DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS (continued) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 0 V (= 3.3 V for PLL Clock Mode), IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA, External Clock Mode, 4:1 transformer output termination, 50-Ω doubly terminated load (unless otherwise noted) AC SPECIFICATIONS PARAMETER ACLR (4) Adjacent channel leakage ratio Noise Floor (4) TEST CONDITIONS MIN TYP MAX Single carrier, baseband, X4, PLL Clock Mode, CLK1 = 122.88 MHz 78.4 Single carrier, baseband, X4, PLL Clock Mode, CLK2 = 491.52 MHz 78.5 Single carrier, IF = 153.6 MHz, X4 CMIX, External Clock Mode, CLK2 = 491.52 MHz 70.9 Two carrier, IF = 153.6 MHz, X4 CMIX, External Clock Mode, CLK2 = 491.52 MHz 67.8 Four carrier, baseband, X4, External Clock Mode, CLK2 = 491.52 MHz 76.1 Four carrier, IF = 92.16 MHz, X4L, External Clock Mode, CLK2 = 491.52 MHz 66.8 Single carrier, IF = 153.6 MHz, X4 CMIX, External Clock Mode, CLK2 = 491.52 MHz, DVDD = 2.1 V 72.2 Two carrier, IF = 153.6 MHz, X4 CMIX, External Clock Mode, CLK2 = 491.52 MHz, DVDD = 2.1 V 69.3 Four carrier, baseband, X4, External Clock Mode, CLK2 = 491.52 MHz, DVDD = 2.1 V 68.5 Four carrier, IF = 92.16 MHz, X4L, External Clock Mode, CLK2 = 491.52 MHz, DVDD = 2.1 V 66.3 UNIT dBc 50-MHz offset, 1-MHz BW, Single carrier, baseband, X4, External Clock Mode, CLK1 = 122.88 MHz 92 50-MHz offset, 1-MHz BW, Four carrier, baseband, X4, External Clock Mode, CLK1 = 122.88 MHz 81 50-MHz offset, 1-MHz BW, Single carrier, baseband, X4, PLL Clock Mode, CLK2 = 491.52 MHz 88 50-MHz offset, 1-MHz BW, Four carrier, baseband, X4, PLL Clock Mode, CLK2 = 491.52 MHz 81 dBc W-CDMA with 3.84 MHz BW, 5-MHz spacing, centered at IF. TESTMODEL 1, 10 ms ELECTRICAL CHARACTERISTICS (DIGITAL SPECIFICATIONS) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 3.3 V, IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA (unless otherwise noted) DIGITAL SPECIFICATIONS PARAMETER TEST CONDITIONS MIN TYP MAX UNIT CMOS INTERFACE VIH High-level input voltage 2 3 VIL Low-level input voltage 0 0 IIH High-level input current –40 IIL Low-level input current –40 Input capacitance PLLLOCK, SDO, SDIO Iload = –8 mA PLLLOCK, SDO, SDIO Input data rate µA 40 µA pF IOVDD – 0.2 V 0.8 x IOVDD V Iload = 100 µA VOL V 40 5 Iload = –100 µA VOH V 0.8 Iload = 8 mA 0.2 V 0.22 x IOVDD V External or dual-clock modes 16 250 PLL clock mode 16 160 MSPS 9 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS (DIGITAL SPECIFICATIONS) (continued) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 3.3 V, IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA (unless otherwise noted) DIGITAL SPECIFICATIONS PARAMETER TEST CONDITIONS MIN TYP MAX UNIT PLL Phase noise VCO maximum frequency VCO minimum frequency At 600-kHz offset, measured at DAC output, 25 MHz 0-dBFS tone, FDATA = 125 MSPS, 4x interpolation, pll_freq = 1, pll_kv = 0 At 6-MHz offset, measured at DAC output, 25 MHz 0-dBFS tone, 125 MSPS, 4x interpolation, pll_freq = 1, pll_kv = 0 pll_freq = 0, pll_kv = 0 370 pll_freq = 0, pll_kv = 1 480 pll_freq = 1, pll_kv = 0 495 pll_freq = 1, pll_kv = 1 520 133 dBc/Hz 148.5 dBc/Hz MHz pll_freq = 0, pll_kv = 0 225 pll_freq = 0, pll_kv = 1 200 pll_freq = 1, pll_kv = 0 480 pll_freq = 1, pll_kv = 1 480 MHz NCO and QMC BLOCKS QMC clock rate 320 MHz NCO clock rate 320 MHz SERIAL PORT TIMING ts(SDENB) Setup time, SDENB to rising edge of SCLK 20 ns ts(SDIO) Setup time, SDIO valid to rising edge of SCLK 10 ns th(SDIO) Hold time, SDIO valid to rising edge of SCLK 5 ns tSCLK Period of SCLK 100 ns tSCLKH High time of SCLK 40 ns tSCLK Low time of SCLK 40 td(Data) Data output delay after falling edge of SCLK ns 10 ns CLOCK INPUT (CLK1/CLK1C, CLK2/CLK2C) Duty cycle 40% Differential voltage 0.4 60% 1 V TIMING PARALLEL DATA INPUT: CLK1 LATCHING MODES (PLL Mode –Figure 45, Dual Clock Mode FIFO Disabled – See Figure 47, Dual Clock Mode With FIFO Enabled – See Figure 48) ts(DATA) Setup time, DATA valid to rising edge of CLK1 0.5 ns th(DATA) Hold time, DATA valid after rising edge of CLK1 1.5 ns t_align Maximum offset between CLK1 and CLK2 rising edges – Dual Clock Mode with FIFO disabled 1 0.5 ns 2FCLK2 ns Timing Parallel Data Input (External Clock Mode, Latch on PLLLock Rising Edge, CLK2 Clock Input, See Figure 43 ) ts(DATA) Setup time, DATA valid to rising edge of PLLLOCK 72-Ω load on PLLLOCK 0.5 ns th(DATA) Hold time, DATA valid after rising edge of PLLLOCK 72-Ω load on PLLLOCK 1.5 ns tdelay(Plllock) Delay from CLK2 rising edge to PLLLOCK rising edge 72-Ω load on PLLLOCK. Note that PLLLOCK delay increases with a lower impedance load. 4.5 ns Timing Parallel Data Input (External Clock Mode, Latch on PLLLock Falling Edge, CLK2 Clock Input, See Figure 44) ts(DATA) Setup time, DATA valid to falling edge of PLLLOCK High impedance load on PLLLOCK 0.5 ns th(DATA) Hold time, DATA valid after falling edge of PLLLOCK High impedance load on PLLLOCK 1.5 ns 10 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 ELECTRICAL CHARACTERISTICS (DIGITAL SPECIFICATIONS) (continued) over recommended operating free-air temperature range, AVDD = 3.3 V, CLKVDD = 3.3 V, PLLVDD = 3.3 V, IOVDD = 3.3 V, DVDD = 1.8 V, IOUTFS = 19.2 mA (unless otherwise noted) DIGITAL SPECIFICATIONS PARAMETER TEST CONDITIONS Delay from CLK2 rising edge to PLLLOCK rising edge tdelay(Plllock) MIN High impedance load on PLLLOCK. Note that PLLLOCK delay increases with a lower impedance load. TYP MAX 4.5 UNIT ns Typical Characteristics 8 6 6 4 2 2 Error − LSB Error − LSB 4 0 −2 0 −2 −4 −4 −6 −8 −6 0 10000 20000 30000 40000 50000 60000 70000 0 10000 20000 30000 40000 50000 60000 70000 Code Code G001 G002 Figure 1. Integral Nonlinearity Figure 2. Differential Nonlinearity 10 10 fdata = 125 MSPS fin = 20 MHz Real IF = 20 MHz 4 Interpolation PLL Off 0 −10 −10 −20 P − Power − dBm P − Power − dBm −20 fdata = 125 MSPS fin = −30 MHz Complex IF = 95 MHz 4L Interpolation CMIX PLL Off 0 −30 −40 −50 −30 −40 −50 −60 −60 −70 −70 −80 −80 −90 −90 0 50 100 150 200 250 f − Frequency − MHz 0 50 100 150 200 G003 Figure 3. Single Tone Spectral Plot 250 f − Frequency − MHz G004 Figure 4. Single Tone Spectral Plot 11 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) 100 10 0 −10 P − Power − dBm −20 SFDR − Spurious-Free Dynamic Range − dBc fdata = 125 MSPS fin = 30 MHz Real IF = 155 MHz 4L Interpolation HP/HP PLL Off −30 −40 −50 −60 −70 −80 −90 fdata = 125 MSPS 4 Interpolation PLL Off 95 90 −6 dBFS 85 80 75 0 dBFS −12 dBFS 70 65 60 0 50 100 150 200 5 250 f − Frequency − MHz 10 15 20 25 30 G005 G006 Figure 5. Single Tone Spectral Plot Figure 6. In-Band SFDR vs Intermediate Frequency 100 fdata = 125 MSPS 4 Interpolation PLL Off 85 fdata = 125 MSPS 4L Interpolation PLL Off fout = IF 0.5 MHz 95 80 90 75 85 70 IMD3 − dBc SFDR − Spurious-Free Dynamic Range − dBc 90 0 dBFS 65 60 80 75 0 dBFS 70 55 65 50 60 45 55 40 50 0 50 100 150 200 250 IF − Intermediate Frequency − MHz 0 50 100 150 200 250 IF − Intermediate Frequency − MHz G007 Figure 7. Out-of-Band SFDR vs Intermediate Frequency 12 35 IF − Intermediate Frequency − MHz G008 Figure 8. Two Tone IMD vs Intermediate Frequency DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) 90 80 IMD3 − dBC 75 −20 −30 70 65 60 −40 −50 −60 55 −70 50 −80 45 −90 40 −35 −30 −25 −20 −15 fdata = 125 MSPS fin = 20 MHz 0.5 MHz Real IF = 20 MHz 4 Interpolation PLL Off −10 P − Power − dBm 85 0 fdata = 125 MSPS fin = −30 MHz 0.5 MHz Complex IF = 95 MHz 4L Interpolation CMIX PLL Off −10 −5 −100 10 0 15 Amplitude − dBFS 20 25 G009 G010 Figure 9. Two Tone IMD vs Amplitude Figure 10. Two Tone IMD Spectral Plot 90 0 −20 P − Power − dBm −30 fdata = 125 MSPS fin = −30 MHz 0.5 MHz Complex IF = 95 MHz 4L Interpolation CMIX PLL Off fdata = 122.88 MSPS Baseband Input DVDD = 1.8 V 85 80 ACLR − dBc −10 30 f − Frequency − MHz −40 −50 −60 75 PLL Off 70 65 PLL On −70 60 −80 55 −90 −100 85 50 90 95 100 105 f − Frequency − MHz 50 100 150 200 250 IF − Intermediate Frequency − MHz G011 Figure 11. Two Tone IMD Spectral Plot 0 G012 Figure 12. WCDMA ACLR vs Intermediate Frequency 13 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) −10 −20 −30 −20 −30 −40 −50 P − Power − dBm P − Power − dBm −40 −10 Carrier Power: −7.99 dBm ACLR (5 MHz): 81.24 dB ACLR (10 MHz): 83.79 dB fdata = 122.88 MSPS IF = 30.72 MHz 4 Interpolation −60 −70 −80 −90 −50 −60 −70 −80 −90 −100 −100 −110 −110 −120 −120 −130 18 23 28 33 38 Carrier Power: −7.99 dBm ACLR (5 MHz): 75.8 dB ACLR (10 MHz): 80.18 dB fdata = 122.88 MSPS IF = 30.72 MHz 4 Interpolation −130 18 43 23 f − Frequency − MHz 28 33 38 43 f − Frequency − MHz G013 G014 Figure 13. WCDMA TM1 : Single Carrier, PLL Off, DVDD = 1.8 V −20 −30 P − Power − dBm −40 −50 −10 Carrier Power: −8.7 dBm ACLR (5 MHz): 75.97 dB ACLR (10 MHz): 77.47 dB fdata = 122.88 MSPS IF = 92.16 MHz 4 Interpolation CMIX −20 −30 −40 P − Power − dBm −10 Figure 14. WCDMA TM1 : Single Carrier, PLL On, DVDD = 1.8 V −60 −70 −80 −90 −50 −60 −70 −80 −90 −100 −100 −110 −110 −120 −120 −130 80 85 90 95 100 105 f − Frequency − MHz Carrier Power: −8.7 dBm ACLR (5 MHz): 67.43 dB ACLR (10 MHz): 73.21 dB fdata = 122.88 MSPS IF = 92.16 MHz 4 Interpolation CMIX −130 80 85 90 95 100 105 f − Frequency − MHz G015 Figure 15. WCDMA TM1 : Single Carrier, PLL Off, DVDD = 1.8 V 14 G016 Figure 16. WCDMA TM1 : Single Carrier, PLL On, DVDD = 1.8 V DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) −20 −30 P − Power − dBm −40 −50 −10 Carrier Power: −10.35 dBm ACLR (5 MHz): 72.06 dB ACLR (10 MHz): 73.21 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −20 −30 −40 P − Power − dBm −10 −60 −70 −80 −90 −50 −60 −70 −80 −90 −100 −100 −110 −110 −120 −120 −130 141 146 151 156 161 Carrier Power: −10.35 dBm ACLR (5 MHz): 63.12 dB ACLR (10 MHz): 69.17 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −130 141 166 146 f − Frequency − MHz 151 156 161 166 f − Frequency − MHz G017 G018 Figure 17. WCDMA TM1 : Single Carrier, PLL Off, DVDD = 1.8 V −20 −30 P − Power − dBm −40 −50 −10 Carrier Power 1 (Ref): −15.78 dBm ACLR (5 MHz): 68.19 dB ACLR (10 MHz): 69.48 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −20 −30 −40 P − Power − dBm −10 Figure 18. WCDMA TM1 : Single Carrier, PLL On, DVDD = 1.8 V −60 −70 −80 −90 −50 −60 −70 −80 −90 −100 −100 −110 −110 −120 −120 −130 138 143 148 153 158 163 168 f − Frequency − MHz Carrier Power 1 (Ref): −15.78 dBm ACLR (5 MHz): 61.28 dB ACLR (10 MHz): 64.61 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −130 138 143 148 153 158 163 168 f − Frequency − MHz G019 Figure 19. WCDMA TM1 : Two Carrier, PLL Off, DVDD = 1.8 V G020 Figure 20. WCDMA TM1 : Two Carrier, PLL On, DVDD = 1.8 V 15 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) −20 −30 −40 −40 −50 −50 P − Power − dBm P − Power − dBm −30 −20 Carrier Power 1 (Ref): −17.41 dBm ACLR (5 MHz): 69.09 dB ACLR (10 MHz): 69.34 dB −60 −70 −80 −90 −100 Carrier Power 1 (Ref): −17.42 dBm ACLR (5 MHz): 64 dB ACLR (10 MHz): 65.79 dB −60 −70 −80 −90 −100 −110 −110 fdata = 122.88 MSPS IF = 92.16 MHz 4 Interpolation CMIX −120 −130 72 77 82 87 92 97 fdata = 122.88 MSPS IF = 92.16 MHz 4 Interpolation CMIX −120 102 107 −130 72 112 77 f − Frequency − MHz 82 87 92 97 102 107 112 f − Frequency − MHz G021 G022 Figure 21. WCDMA TM1 : Four Carrier, PLL Off, DVDD = 1.8 V −20 −30 P − Power − dBm −40 −50 −10 Carrier Power: −10.35 dBm ACLR (5 MHz): 73.83 dB ACLR (10 MHz): 75.39 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −20 −30 −40 P − Power − dBm −10 Figure 22. WCDMA TM1 : Four Carrier, PLL On, DVDD = 1.8 V −60 −70 −80 −90 −50 −60 −70 −80 −90 −100 −100 −110 −110 −120 −120 −130 141 146 151 156 161 166 f − Frequency − MHz Carrier Power 1 (Ref): −15.77 dBm ACLR (5 MHz): 69.74 dB ACLR (10 MHz): 71.17 dB fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −130 138 143 148 153 158 163 168 f − Frequency − MHz G023 Figure 23. WCDMA TM1 : Single Carrier, PLL Off, DVDD = 2.1 V 16 G024 Figure 24. WCDMA TM1 : Two Carrier, PLL Off, DVDD = 2.1 V DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Typical Characteristics (continued) −20 −30 Carrier Power 1 (Ref): −19.88 dBm ACLR (5 MHz): 66.6 dB ACLR (10 MHz): 65.73 dB −40 P − Power − dBm −50 −60 −70 −80 −90 −100 −110 fdata = 122.88 MSPS IF = 153.6 MHz 4 Interpolation CMIX −120 −130 133 138 143 148 153 158 163 168 173 f − Frequency − MHz G025 Figure 25. WCDMA TM1 : Four Carrier, PLL Off, DVDD = 2.1 V Test Methodology Typical ac specifications in external clock mode were characterized with the DAC5687EVM using the test configuration shown in Figure 26. The DAC sample rate clock fDAC is generated by a HP8665B signal generator. An Agilent 8133A pulse generator is used to divide fDAC for the data rate clock fDATA for the Agilent 16702A pattern generator clock and provide adjustable skew to the DAC input clock. The 8133A fDAC output is set to 1 VPP, equivalent to 2-VPP differential at CLK2/CLK2C pins. Alternatively, the DAC5687 PLLLOCK output can be used for the pattern generator clock. The DAC5687 output is characterized with a Rohde & Schwarz FSQ8 spectrum analyzer. For WCDMA signal characterization, it is important to use a spectrum analyzer with high IP3 and noise subtraction capability so that the spectrum analyzer does not limit the ACPR measurement. For all specifications, both DACA and DACB are measured and the lowest value used as the specification. 17 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Test Methodology (continued) 1.8 V/2.1 V 200 Ω Mini Circuits TCM4−1W CLK2C CLK1 CLK1C PULSE FREQ. = fdata 10 pF DVDD (Pin 56) SLEEP CLK2 DVDD (Not Including Pin 56) PULSE FREQ. = fDAC Ampl. = 1 VPP 1:4 PHSTR 0.01 µF Agilent 8133A Pulse Generator PLLGND Sinusoid FREQ. = fDAC PLLVDD 10 Ω DGND HP8665B Synthesized Signal Generator EXTLO BIASJ 3.3 V RBIAS 1 kΩ PLLLOCK Agilent 16702B Mainframe System With 16702A Pattern Generator Card 16 Rohde & Schwarz FSQ8 Spectrum Analyzer CEXTIO 0.1 µF EXTIO 100 Ω DA[15:0] 3.3 V 1:4 IOUTA1 IOUTA2 16 DB[15:0] IOUTB1 IOUTB2 TxENABLE RESETB IOVDD 3.3 V AGND AVDD LPF Mini Circuits T4−1 3.3 V IOGND CLKGND CLKVDD 3.3 V 100 Ω 330 pF 3.3 V 3.3 V 0.033 µF 93.1 Ω B0039-01 Figure 26. DAC5687 Test Configuration for External Clock Mode PLL clock mode was characterized using the test configuration shown in Figure 27. The DAC data rate clock fDATA is generated by a HP8665B signal generator. An Agilent 8133A pulse generator is used to generate a clock fDATA for the Agilent 16702A pattern generator clock and provide adjustable skew to the DAC input clock. The 8133A fDAC output is set to 1 VPP, equivalent to 2-VPP differential at CLK1/CLK1C pins. Alternatively, the DAC5687 PLLLOCK output can be used for the pattern generator clock. 18 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Test Methodology (continued) 1.8 V/2.1 V 3.3 V 200 Ω Mini Circuits TCM4−1W CLK1C CLK2 CLK2C PULSE FREQ. = fdata 10 pF DVDD (Pin 56) SLEEP CLK1 DVDD (Not Including Pin 56) PULSE FREQ. = fdata Ampl. = 1 VPP 1:4 PHSTR 0.01 µF Agilent 8133A Pulse Generator PLLGND Sinusoid FREQ. = fdata PLLVDD 10 Ω DGND HP8665B Synthesized Signal Generator EXTLO BIASJ 3.3 V RBIAS 1 kΩ PLLLOCK Agilent 16702B Mainframe System With 16702A Pattern Generator Card 16 Rohde & Schwarz FSQ8 Spectrum Analyzer CEXTIO 0.1 µF EXTIO 100 Ω DA[15:0] 3.3 V 1:4 IOUTA1 IOUTA2 16 DB[15:0] IOUTB1 IOUTB2 IOVDD 3.3 V AGND AVDD LPF Mini Circuits T4−1 3.3 V IOGND CLKGND CLKVDD TxENABLE RESETB 100 Ω 330 pF 3.3 V 3.3 V 0.033 µF 93.1 Ω B0039-02 Figure 27. DAC5687 Test Configuration for PLL Clock Mode WCDMA Test Model 1 test vectors for the DAC5687 characterization were generated in accordance with the 3GPP Technical Specification . Chip rate data was generated using the Test Model 1 block in Agilent ADS. For multicarrier signals, different random seeds and PN offsets were used for each carrier. A Matlab script performed pulse shaping, interpolation to the DAC input data rate, frequency offsets, rounding and scaling. Each test vector is 10 ms long, representing one frame. Special care is taken to assure that the end of file wraps smoothly to the beginning of the file. The cumulative complementary distribution function (CCDF) for the 1, 2, and 4 carrier test vectors is shown in Figure 28. The test vectors are scaled such that the absolute maximum data point is 0.95 (-0.45 dB) of full scale. No peak reduction is used. 19 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Cummulative Complementary Distribution Function Test Methodology (continued) 100 10−2 2 Carriers 4 Carriers 1 Carrier 10−4 10−6 3 5 7 9 11 13 15 Peak-to-Average Ratio − dB G041 Figure 28. WCDMA TM1 Cumulative Complementary Distribution Function for 1, 2, and 4 carriers DETAILED DESCRIPTION Modes of Operation The DAC5687 has six digital signal processing blocks: FIR1 and FIR2 (interpolate by two digital filters), FMIX (fine frequency mixer), QMC (quadrature modulation phase correction), FIR3 (interpolate by two digital filter) and CMIX (coarse frequency mixer). The modes of operation, listed in Table 1, enable or bypass the blocks to produce different results. The modes are selected by registers CONFIG1, CONFIG2, and CONFIG3 (0x02, 0x03, and 0x04). Block diagrams for each mode (x2, x4, x4L, and x8) are shown in Figure 29 through Figure 32. Table 1. DAC5687 Modes of Operation MODE FIR1 FIR2 FMIX QMC FULL BYPASS – – – X2 – – – X2 FMIX – – X2 QMC – X2 FMIX QMC – X2 CMIX X2 FMIX CMIX CMIX – – – – ON – ON – ON – – – ON ON – – ON ON ON – – – – – ON ON – – ON – ON ON X2 QMC CMIX – – – ON (1) ON ON X2 FMIX QMC CMIX – – ON ON (1) ON ON ON ON – – – – FMIX (2) ON ON ON – – – X4 QMC (2) ON ON – ON – – X4 FMIX QMC ON ON ON ON – – X4 CMIX ON ON – – – ON X4 X4 (1) (2) 20 FIR3 The QMC phase correction will be eliminated by the CMIX, so the QMC phase should be set to zero. The QMC gain settings can still be used to adjust the signal path gain as needed. fDAC limited to maximum clock rate for the NCO and QMC, (See the AC specs). DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 1. DAC5687 Modes of Operation (continued) FIR1 FIR2 FMIX QMC FIR3 CMIX X4L MODE ON – – – ON – X4L FMIX ON – ON – ON – X4L QMC ON – – ON ON – X4L FMIX QMC ON – ON ON ON – X4L CMIX ON – – – ON ON X4L FMIX CMIX ON – ON ON ON X4L QMC CMIX ON – – ON (2) ON ON X4L FMIX QMC CMIX ON – ON ON (2) ON ON X8 ON ON – – ON – X8 FMIX ON ON ON – ON – X8 QMC ON ON – ON ON – X8 FMIX QMC ON ON ON ON ON – X8 CMIX ON ON – – ON ON X8 FMIX CMIX ON ON ON – ON ON X8 QMC CMIX ON ON – ON (3) ON ON ON ON (3) ON ON X8 FMIX QMC CMIX (3) ON ON The QMC phase correction will be eliminated by the CMIX, so the QMC phase should be set to zero. The QMC gain settings can still be used to adjust the signal path gain as needed. A Offset DB[15:0] FIR3 FIR4 16-bit DAC x2 x sin(x) x sin(x) 16-bit DAC x2 cos Coarse Mixer: fs/2 or fs/4 Quadrature Mod Correction (QMC) Fine Mixer Input Formater DA[15:0] A gain sin NCO IOUTA1 IOUTA2 IOUTB1 IOUTB2 B Offset B gain Figure 29. Block Diagram for X2 Mode 21 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 A Offset DB[15:0] x2 x2 Coarse Mixer: fs/2 or fs/4 x2 FIR4 Fine Mixer Input Formater DA[15:0] FIR2 Quadrature Mod Correction (QMC) FIR1 x2 cos IOUTA1 x sin(x) 16-bit DAC x sin(x) 16-bit DAC sin NCO A. A gain IOUTA2 IOUTB1 B Offset B gain A Offset A gain IOUTB2 FMIX or QMC block cannot be enabled with CMIX block. Figure 30. Block Diagram for X4 Mode (A) Quadrature Mod Correction (QMC) x2 Fine Mixer Input Formater DA[15:0] DB[15:0] x2 x2 x2 cos FIR4 Coarse Mixer: fs/2 or fs/4 FIR3 FIR1 16-bit DAC x sin(x) 16-bit DAC sin NCO Figure 31. Block Diagram for X4L Mode 22 IOUTA1 x sin(x) IOUTA2 IOUTB1 IOUTB2 B Offset B gain DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 A Offset A gain FIR2 DB[15:0] Quadrature Mod Correction (QMC) x2 x2 Fine Mixer Input Formater DA[15:0] x2 x2 FIR3 FIR4 x2 x sin(x) 16-bit DAC x sin(x) 16-bit DAC Coarse Mixer: fs/2 or fs/4 FIR1 x2 cos IOUTA1 IOUTA2 IOUTB1 IOUTB2 sin B Offset B gain NCO Figure 32. Block Diagram for X8 Mode Programming Registers REGISTER MAP Name Address Default Bit 7 (MSB) Bit 6 sleep_daca sleep_dacb VERSION 0x00 0x01 CONFIG0 0x01 0x00 CONFIG1 0x02 0x00 qflag CONFIG2 0x03 0x80 CONFIG3 0x04 SYNC_CNTL SER_DATA_0 Bit 5 Bit 4 Bit 3 unused hpla hplb pll_freq pll_kv interl dual_clk twos nco nco_gain qmc 0x00 sif_4pin dac_ser_dat a half_rate 0x05 0x00 sync_phstr sync_nco sync_cm 0x06 0x00 dac_data(7:0) SER_DATA_1 0x07 0x00 dac_data(15:8) factory use only 0x08 0x00 NCO_FREQ_0 0x09 0x00 freq(7:0) NCO_FREQ_1 0x0A 0x00 freq(15:8) pll_div(1:0) Bit 2 rev_bbus inv_plllock fifo_bypass fir_bypass full_bypass cm_mode(3:0) unused (LSB) Bit 0 version(2:0) interp(1:0) rev_abus Bit 1 invsinc usb counter_mode(2:0) sync_fifo(2:0) unused unused NCO_FREQ_2 0x0B 0x00 freq(23:16) NCO_FREQ_3 0x0C 0x0C freq(31:24) NCO_PHASE_0 0x0D 0x00 phase(7:0) NCO_PHASE_1 0x0E 0x00 phase(15:8) DACA_OFFSET_0 0x0F 0x00 daca_offset(7:0) DACB_OFFSET_0 0x10 0x00 DACA_OFFSET_1 0x11 0x00 daca_offset(12:8) unused unused unused DACB_OFFSET_1 0x12 0x00 dacb_offset(12:8) unused unused unused QMCA_GAIN_0 0x13 0x00 qmc_gain_a(7:0) QMCB_GAIN_0 0x14 0x00 qmc_gain_b(7:0) QMC_PHASE_0 0x15 0x00 QMC_PHASE_GAIN_1 0x16 0x00 DACA_GAIN_0 0x17 0x00 daca_gain(7:0) DACB_GAIN_0 0x18 0x00 dacb_gain(7:0) DACA_DACB_GAIN_1 0x19 0xFF factory use only 0x1A 0x00 atest 0x1B 0x00 dacb_offset(7:0) qmc_phase(7:0) qmc_phase(9:8) qmc_gain_a(10:8) daca_gain(11:8) atest qmc_gain_b(10:8) dacb_gain(11:8) phstr_del(1:0) unused 23 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 REGISTER MAP (continued) Name Address Default DAC_TEST 0x1C 0x00 factory use only 0x1D 0x00 factory use only 0x1E 0x00 factory use only 0x1F 0x00 Bit 7 (MSB) Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 factory use only (LSB) Bit 0 phstr_clkdiv_sel Register Name: VERSION — Address: 0x00, Default = 0x01 BIT 7 BIT 0 sleep_daca sleep_dacb hpla hplb unused 0 0 0 0 0 version(2:0) 0 1 1 sleep_daca: DAC A sleeps when set, operational when cleared. sleep_dacb: DAC B sleeps when set, operational when cleared. hpla: A side first FIR filter in high-pass mode when set, low-pass mode when cleared. hplb: B side first FIR filter in high-pass mode when set, low-pass mode when cleared. version(2:0): A hardwired register that contains the version of the chip. Read Only. Register Name: CONFIG0 — Address: 0x01, Default = 0x00 BIT 7 BIT 0 pll_div(1:0) 0 0 pll_freq pll_kv 0 0 interp(1:0) 0 0 inv_plllock fifo_bypass 0 0 pll_div(1:0): PLL VCO divider; {00 = 1, 01 = 2, 10 = 4, 11 = 8}. pll_freq: PLL VCO center frequency; {0 = low center frequency, 1 = high center frequency}. pll_kv: PLL VCO gain; {0 = high gain, 1 = low gain}. interp(1:0): FIR Interpolation; {00 = X2, 01 = X4, 10 = X4L, 11 = X8}. X4 uses lower power than x4L, but fDAC = 320 MHz max when NCO or QMC are used. inv_plllock: Multi-function bit depending on clock mode. fifo_bypass: When set, the internal 4-sample FIFO is disabled. When cleared, the FIFO is enabled. Table 2. inv_plllock Bit Modes 24 PLLVDD dual_clk inv_pllock fifo_bypass DESCRIPTION 0V 0 0 1 Input data latched on PLLLOCK pin rising edges, FIFO disabled. 0V 0 1 1 Input data latched on PLLLOCK pin falling edges, FIFO disabled. 0V 0 0 0 Input data latched on PLLLOCK pin rising edges, FIFO enabled and must be sync’d. 0V 0 1 0 Input data latched on PLLLOCK pin falling edges, FIFO enabled and must be sync’d. 0V 1 0 1 Input data latched on CLK1/CLK1C differential input. Timing between CLK1 and CLK2 rising edges must be tightly controlled (500 ps max at 500-MHz CLK2). PLLLOCK output signal is always low. The FIFO is always disabled in this mode. 0V 1 1 0 Input data latched on CLK1/CLK1C differential input. No phase relationship required between CLK1 and CLK2. The FIFO is employed to manage the internal handoff between the CLK1 input clock and the CLK2 derived output clock; the FIFO must be sync’d. The PLLLOCK output signal reflects the internally generated FIFO output clock. 0V 1 0 0 Not a valid setting. Do not use. DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 2. inv_plllock Bit Modes (continued) PLLVDD dual_clk inv_pllock fifo_bypass DESCRIPTION 0V 1 1 1 Not a valid setting. Do not use. 3.3 V X X 1 Internal PLL enabled, CLK1/CLK1C input differential clock is used to latch the input data. The FIFO is always disabled in this mode. 3.3 V X X 0 Not a valid setting. Do not use. Register Name: CONFIG1 — Address: 0x02, Default = 0x00 BIT 7 BIT 0 qflag interl dual_clk twos rev_abus rev_bbus fir_bypass full_bypass 0 0 0 0 0 0 0 0 qflag: When set, the QFLAG input pin operates as a B sample indicator when interleaved data is enabled. When cleared, the TXENABLE rising determines the A/B timing relationship. interl: When set, interleaved input data mode is enabled; both A and B data streams are input at the DA(15:0) input pins. dual_clk: Only used when the PLL is disabled. When set, two differential clocks are used to input the data to the chip; CLK1/CLK1C is used to latch the input data into the chip and CLK2/CLK2C is used as the DAC sample clock. twos: When set, input data is interpreted as 2’s complement. When cleared, input data is interpreted as offset binary. rev_abus: When cleared, DA input data MSB to LSB order is DA(15) = MSB and DA(0) = LSB. When set, DA input data MSB to LSB order is reversed, DA(15) = LSB and DA(0) = MSB. rev_bbus: When cleared, DB input data MSB to LSB order is DB(15) = MSB and DB(0) = LSB. When set, DB input data MSB to LSB order is reversed, DB(15) = LSB and DB(0) = MSB. fir_bypass: When set, all interpolation filters are bypassed (interp(1:0) setting has no effect). QMC and NCO blocks are functional in this mode up to fdac = 250 MHz, limited by the input datarate. full_bypass: When set, all filtering, QMC and NCO functions are bypassed. Register Name: CONFIG2 — Address: 0x03, Default = 0x80 BIT 7 BIT 0 nco nco_gain qmc 1 0 0 cm_mode(3:0) 0 0 invsinc 0 nco: When set, the NCO is enabled. nco_gain: When set, the data output of the NCO is increased by 2x. qmc: Quadrature modulator gain and phase correction is enabled when set. 0 0 cm_mode(3:0): Controls fDAC/2 or fDAC/4 mixer modes for the coarse mixer block. Table 3. Coarse Mixer Sequences cm_mode(3:0) Mixing Mode Sequence 00XX No mixing 0100 fDAC/2 DAC A = {-A +A -A +A …} DAC B = {-B +B -B +B …} 0101 fDAC/2 DAC A = {-A +A -A +A …} DAC B = {+B -B +B -B …} 0110 fDAC/2 DAC A = {+A -A +A -A …} DAC B = {-B +B -B +B …} 25 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 3. Coarse Mixer Sequences (continued) invsinc: cm_mode(3:0) Mixing Mode Sequence 0111 fDAC2 DAC A = {+A -A +A -A …} DAC B = {+B -B +B -B …} 1000 fDAC/4 DAC A = {+A -B -A +B …} DAC B = {+B +A -B -A …} 1001 fDAC/4 DAC A = {+A -B -A +B …} DAC B = {-B -A +B +A …} 1010 fDAC/4 DAC A = {-A +B +A -B …} DAC B = {+B +A -B -A …} 1011 fDAC/4 DAC A = {-A +B +A -B …} DAC B = {-B -A +B +A …} 1100 -fDAC/4 DAC A = {+A +B -A -B …} DAC B = {+B -A -B +A …} 1101 -fDAC/4 DAC A = {+A +B -A -B …} DAC B = {-B +A +B -A …} 1110 -fDAC/4 DAC A = {-A -B +A +B …} DAC B = {+B -A -B +A …} 1111 -fDAC/4 DAC A = {-A -B +A +B …} DAC B = {-B +A +B -A …} Enables the invsinc compensation filter when set. Register Name: CONFIG3 — Address: 0x04, Default = 0x00 BIT 7 BIT 0 sif_4pin dac_ser_data half_rate Unused usb 0 0 0 0 0 sif_4pin: counter_mode(2:0) 0 0 0 Four-pin serial interface mode is enabled when set, 3-pin mode when cleared. dac_ser_data: When set, both DAC A and DAC B input data is replaced with fixed data loaded into the 16 bit serial interface ser_data register. half_rate: Enables half-rate input mode. Input data for the DAC A data path is input to the chip at half speed using both the DA(15:0) and DB(15:0) input pins. usb: When set, the data to DACB is inverted to generate upper side band output. counter_mode(2:0): Controls the internal counter that can be used as the {0XX = off; 100 = all 16b; 101 = 7b LSBs; 110 = 5b MIDs; 111 = 5b MSBs} DAC data source. Register Name: SYNC_CNTL — Address: 0x05, Default = 0x00 BIT 7 BIT 0 sync_phstr sync_nco sync_cm 0 0 0 sync_fifo(2:0) 0 0 unused unused 0 0 0 sync_phstr: When set, the internal clock divider logic is initialized with a PHSTR pin low-to-high transition. sync_nco: When set, the NCO phase accumulator is cleared with a PHSTR low-to-high transition. sync_cm: When set, the coarse mixer is initialized with a PHSTR low-to-high transition. sync_fifo(2:0): Sync source selection mode for the FIFO. When a low-to-high transition is detected on the selected sync source, the FIFO input and output pointers are initialized. 26 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 4. Synchronization Source sync_fifo (2:0) Synchronization Source 000 txenable pin 001 phstr pin 010 qflag pin 011 db(15) 100 da(15) first transition (one shot) 101 Software sync using SIF write 110 Sync source disabled (always off) 111 Always on Register Name: SER_DATA_0— Address: 0x06, Default = 0x00 BIT 7 BIT 0 dac_data(7:0) 0 0 0 0 0 0 0 0 dac_data(7:0): Lower 8 bits of DAC data input to the DACs when dac_ser_data is set. Register Name: SER_DATA_1— Address: 0x07, Default = 0x00 BIT 7 BIT 0 dac_data(15:8) 0 0 0 0 0 0 0 0 dac_data(15:8): Upper 8 bits of DAC data input to the DACs when dac_ser_data is set. Register Name: BYPASS_MASK_CNTL— Address: 0x08, Default = 0x00 BIT 7 BIT 0 fast__latch bp_ invsinc bp_fir3 bp_qmc bp_fmix bp_fir2 bp_fir1 nco_only 0 0 0 0 0 0 0 0 These modes are for factory use only – leave as default. Register Name: NCO_FREQ_0— Address: 0x09, Default = 0x00 BIT 7 BIT 0 freq(7:0) 0 freq(7:0): 0 0 0 0 0 0 0 Bits 7:0 of the NCO frequency word. Register Name: NCO_FREQ_1— Address: 0x0A, Default = 0x00 BIT 7 BIT 0 freq(15:8) 0 freq(15:8): 0 0 0 0 0 0 0 Bits 15:8 of the NCO frequency word. Register Name: NCO_FREQ_2— Address: 0x0C, Default = 0x40 BIT 7 BIT 0 freq(23:16) 0 0 0 0 0 0 0 0 27 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 freq(23:16): Bits 23:16 of the NCO frequency word. Register Name: NCO_FREQ_3— Address: 0x0B, Default = 0x00 BIT 7 BIT 0 freq(31:24) 0 1 0 0 0 0 0 0 freq(31:24): Bits 31:24 of the NCO frequency word. Register Name: NCO_PHASE_0— Address: 0x0D, Default = 0x00 BIT 7 BIT 0 Phase(7:0) 0 0 0 0 0 0 0 0 phase(7:0): Bits 7:0 of the NCO phase offset word. Register Name: NCO_PHASE_1— Address: 0x0E, Default = 0x00 BIT 7 BIT 0 Phase(15:8) 0 0 0 0 0 0 0 0 phase(15:8): Bits 15:8 of the NCO phase offset word. Register Name: DACA_OFFSET_0— Address: 0x0F, Default = 0x00 BIT 7 BIT 0 daca_offset(7:0) 0 0 0 0 daca_offset(7:0): Bits 7:0 of the DAC A offset word. 28 0 0 0 0 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Register Name: DACB_OFFSET_0— Address: 0x10, Default = 0x00 BIT 7 BIT 0 dacb_offset(7:0) 0 0 0 0 0 0 0 0 dacb_offset(7:0): Bits 7:0 of the DAC B offset word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: DACA_OFFSET_1— Address: 0x11, Default = 0x00 BIT 7 BIT 0 daca_offset(12:8) 0 0 0 0 unused unused unused 0 0 0 0 daca_offset(12:8): Bits 12:8 of the DAC A offset word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: DACB_OFFSET_1— Address: 0x12, Default = 0x00 BIT 7 BIT 0 dacb_offset(12:8) 0 0 0 0 unused unused unused 0 0 0 0 dacb_offset(12:8): Bits 12:8 of the DAC B offset word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: QMCA_GAIN_0— Address: 0x13, Default = 0x00 BIT 7 BIT 0 qmc_gain_a(7:0) 0 0 0 0 0 0 0 0 qmc_gain_a(7:0): Bits 7:0 of the QMC A path gain word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: QMCB_GAIN_0— Address: 0x14, Default = 0x00 BIT 7 BIT 0 qmc_gain_b(7:0) 0 0 0 0 0 0 0 0 qmc_gain_b(7:0): Bits 7:0 of the QMC B path gain word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: QMC_PHASE_0— Address: 0x15, Default = 0x00 BIT 7 BIT 0 qmc_phase(7:0) 0 0 0 0 0 0 0 0 qmc_phase(7:0): Bits 7:0 of the QMC phase word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: QMC_PHASE_GAIN_1— Address: 0x16, Default = 0x00 BIT 7 BIT 0 qmc_phase(9:8) 0 qmc_gain_a(10:8) 0 0 0 qmc_gain_b(10:8) 0 0 0 0 29 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 qmc_phase(9:8): Bits 9:8 of the QMC phase word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. qmc_gain_a(10:8) : Bits 10:8 of the QMC A path gain word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. qmc_gain_b(10:8): Bits 10:8 of the QMC B path gain word. Updates to this register do not take effect until DACA_OFFSET_0 has been written. Register Name: DACA_GAIN_0— Address: 0x17, Default = 0x00 BIT 7 BIT 0 daca_gain(7:0) 0 0 0 0 0 0 0 0 daca_gain(7:0): Bits 7:0 of the DAC A gain adjustment word. Register Name: DACB_GAIN_0— Address: 0x18, Default = 0x00 BIT 7 BIT 0 dacb_gain(7:0) 0 0 0 0 0 0 0 0 dacb_gain(7:0): Bits 7:0 of the DAC B gain adjustment word. Register Name: DACA_DACB_GAIN_1— Address: 0x19, Default = 0xFF BIT 7 BIT 0 daca_gain(11:8) 1 1 dacb_gain(11:8) 1 1 1 1 1 1 daca_gain(11:8): Four MSBs of gain control for DACA. dacb_gain(11:8): Bits 11:8 of the DAC B gain word. Four MSBs of gain control for DACB. Register Name: DAC_CLK_CNTL— Address: 0x1A, Default = 0x00 BIT 7 BIT 0 factory use only 0 0 0 0 0 0 0 0 Reserved for factory use only. Register Name: ATEST— Address: 0x1B, Default = 0x00 BIT 7 BIT 0 atest(4:0) 0 atest: 30 0 0 phstr_del(1:0) 0 0 0 unused 0 0 Can be used to enable clock output at the PLLLOCK pin according to Table 5. Pin EXTLO must be open when atest (4:0) is not equal to 00000. DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 5. atest(4:0) PLLLOCK Output Signal PLL Enabled (PLLVDD = 3.3 V) PLL Disabled (PLLVDD = 0 V) 11101 fDAC Normal operation 11110 fDAC divided by 2 Normal operation 11111 fDAC divided by 4 Normal operation All others phstr_del: Normal operation Adjusts the initial phase of the fs/2 and fs/4 blocks cmix block after PHSTR. Register Name: DAC_TEST— Address: 0x1C, Default = 0x00 BIT 7 BIT 0 Factory Use Only 0 0 0 0 phstr_clkdiv_sel 0 0 0 0 phstr_clkdiv_sel: Selects the clock used to latch the PHSTR input when restarting the internal dividers. When set, the full DAC sample rate CLK2 signal latches PHSTR and when cleared, the divided down input clock signal latches PHSTR. Address: 0x1D, 0x1E, and 0x1F – Reserved Writes have no effect and reads will be 0x00. Serial Interface The serial port of the DAC5687 is a flexible serial interface which communicates with industry standard microprocessors and microcontrollers. The interface provides read/write access to all registers used to define the operating modes of the DAC5687. It is compatible with most synchronous transfer formats and can be configured as a 3- or 4-pin interface by sif4 in register config_msb. In both configurations, SCLK is the serial interface input clock and SDENB is serial interface enable. For 3-pin configuration, SDIO is a bidirectional pin for both data in and data out. For 4-pin configuration, SDIO is data in only and SDO is data out only. Each read/write operation is framed by signal SDENB (serial data enable bar) asserted low for 2 to 5 bytes, depending on the data length to be transferred (1 – 4 bytes). The first frame byte is the instruction cycle which identifies the following data transfer cycle as read or write, how many bytes to transfer, and what address to transfer the data. Table 6 indicates the function of each bit in the instruction cycle and is followed by a detailed description of each bit. Frame bytes 2 to 5 comprise the data transfer cycle. Table 6. Instruction Byte of the Serial Interface MSB LSB Bit 7 6 5 4 3 2 1 0 Description R/W N1 N0 A4 A3 A2 A1 A0 R/W Identifies the following data transfer cycle as a read or write operation. A high indicates a read operation from the DAC5687 and a low indicates a write operation to the DAC5687. [N1 : N0] Identifies the number of data bytes to be transferred per Table 7. Data is transferred MSB first. Table 7. Number of Transferred Bytes Within One Communication Frame N1 N0 Description 0 0 Transfer 1 Byte 0 1 Transfer 2 Bytes 1 0 Transfer 3 Bytes 1 1 Transfer 4 Bytes 31 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 [A4 : A3 : A2 : A1 : A0] Identifies the address of the register to be accessed during the read or write operation. For multi-byte transfers, this address is the starting address. Note that the address is written to the DAC5687 MSB first. Figure 33 shows the serial interface timing diagram for a DAC5687 write operation. SCLK is the serial interface clock input to the DAC5687. Serial data enable SDENB is an active low input to the DAC5687. SDIO is serial data in. Input data to the DAC5687 is clocked on the rising edges of SCLK. Instruction Cycle SDENB Data Transfer Cycle(s) SCLK SDIO R/W N1 N0 A4 A3 A2 A1 A0 D7 D6 D5 D4 D3 D2 D1 D0 ts(SDENB) t(SCLK) SDENB SCLK SDIO ts(SDIO) th(SDIO) t(SCLKL) t(SCLKH) Figure 33. Serial Interface Write Timing Diagram Figure 34 shows the serial interface timing diagram for a DAC5687 read operation. SCLK is the serial interface clock input to the DAC5687. Serial data enable SDENB is an active low input to the DAC5687. SDIO is serial data in during the instruction cycle. In 3-pin configuration, SDIO is data out from the DAC5687 during the data transfer cycle(s), while SDO is in a high-impedance state. In 4-pin configuration, SDO is data out from the DAC5687 during the data transfer cycle(s). At the end of the data transfer, SDO outputs low on the final falling edge of SCLK until the rising edge of SDENB when it will 3-state. 32 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Data Transfer Cycle(s) Instruction Cycle SDENB SCLK SDIO R/W N1 N0 A4 A3 A2 A1 A0 D7 D6 D5 D4 D3 D2 D1 D0 0 D7 D6 D5 D4 D3 D2 D1 D0 0 SDO 4-Pin Configuration Output 3-Pin Configuration Output SDENB SCLK SDIO SDO Data n Data n−1 td(DATA) Figure 34. Serial Interface Read Timing Diagram FIR Filters Figure 35 shows the magnitude spectrum response for the identical 51-tap FIR1 and FIR3 filters. The transition band is from 0.4 to 0.6 x FIN (the input data rate for the FIR filter) with < 0.002-dB pass-band ripple and > 80-dB stop-band attenuation. Figure 36 shows the region from 0.35 to 0.45 x FIN– Up to 0.44 x FIN there is less than 0.5-dB attenuation. Figure 37 shows the magnitude spectrum response for the 19-tap FIR2 filter. The transition band is from 0.25 to 0.75x FIN (the input data rate for the FIR filter) with < 0.002-dB pass-band ripple and > 80-dB stop-band attenuation. The DAC5687 also has an inverse Sinc filter (FIR4) that runs at the DAC update rate (fDAC) that can be used to flatten the frequency response of the sample and hold output. The DAC sample and hold output set the output current and holds it constant for one DAC clock cycle until the next sample, resulting in the well known Sin(x)/x or Sinc(x) frequency response shown in Figure 38 (black solid line). The inverse sinc filter response (Figure 38, blue dotted line) has the opposite frequency response between 0 to 0.4 x fDAC, resulting in the combined response (Figure 38, red dashed line). Between 0 to 0.4 x fDAC, the inverse sin filter compensates the sample and hold rolloff with less than < 0.03-dB error. The inverse sine filter has a gain > 1 at all frequencies. Therefore, the signal input to FIR4 must be reduced from full scale to prevent saturation in the filter. The amount of backoff required depends on the signal frequency, and is set such that at the signal frequencies the combination of the input signal and filter response is less than 1 (0 dB). For example, if the signal input to FIR4 is at 0.25 × fDAC, the response of FIR4 is 0.9 dB, and the signal will need to be backed off from full scale by 0.9 dB. The gain function in the QMC block can be used to set reduce amplitude of the input signal. The advantage of FIR4 having a positive gain at all frequencies is that the user is then able to optimized backoff of the signal based on the signal frequency. The filter taps for all digital filters are listed in Table 8. Note that the loss of signal amplitude may result in lower SNR due to decrease in signal amplitude. 33 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Figure 35. Magnitude Spectrum for FIR1 and FIR3 Figure 36. FIR1 and FIR3 Transition Band 4 3 2 dB 1 0 −1 Sin(x)/x effect FIR4 response Corrected spectrum −2 −3 −4 0 0.1 Figure 37. Magnitude Spectrum for FIR2 0.2 0.3 Fout/Fdac 0.5 Figure 38. Magnitude Spectrum for Inverse Sinc Filter FIR4 (Versions 1 and 2) Table 8. Digital Filter Taps FIR1 and FIR3 34 0.4 FIR2 FIR4 (Invsinc) Tap Coeff Tap Coeff Tap Coeff 1, 51 8 1, 19 9 1, 9 1 2, 50 0 2, 18 0 2, 8 –4 3, 49 –24 3, 17 –58 3, 7 13 4, 48 0 4, 16 0 4, 6 –50 5, 47 58 5, 15 214 5 592 6, 46 0 6, 14 0 7, 45 –120 7, 13 –638 8, 44 0 8, 12 0 9, 43 221 9, 11 2521 10, 42 0 10 4096 11, 41 –380 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 8. Digital Filter Taps (continued) FIR1 and FIR3 FIR2 Tap Coeff 12, 40 0 13, 39 619 14, 38 0 15, 37 –971 16, 36 0 17, 35 1490 18, 34 0 19, 33 –2288 20, 32 0 21, 31 3649 22, 30 0 23, 29 –6628 24, 28 0 25, 27 20750 26 32768 Tap FIR4 (Invsinc) Coeff Tap Coeff Dual Channel Real Upconversion The DAC5687 can be used in a dual channel mode with real upconversion by mixing with a 1, -1, … sequence in the signal chain to invert the spectrum. This mixing mode maintains isolation of the A and B channels. There are two points of mixing: in X4L mode, the FIR1 output is inverted (high-pass mode) by setting registers hpla and hplb to 1 and the FIR3 output is inverted by setting CMIX to fDAC/2. In X8 mode, the output of FIR1 is inverted by setting hpla and hplb to 1 and the FIR3 output is inverted by setting CMIX to fDAC/2. In X2 and X4 modes, the output of FIR3 is inverted by setting CMIX to fDAC/2. The wide bandwidth of FIR3 (40% passband) in X4L mode provides options for setting four different frequency ranges, listed in Table 9. For example, with fDATA = 125 MSPS (fDAC = 500 MSPS), setting FIR1/FIR3 to High Pass/High Pass respectively will upconvert a signal between 25 and 50 MHz to 150 to 175 MHz. With the High Pass/Low Pass and Low Pass/High Pass setting the upconvertered signal will be spectrally inverted. Table 9. X4L Mode High-Pass/Low-Pass Options FIR1 FIR3 Input Frequency Output Frequency Bandwidth Inverted? Low Pass Low Pass 0 - 0.4 x fDATA 0 - 0.4 x fDATA 0.4 x fDATA No High Pass Low Pass 0.2 to 0.4 x fDATA 0.6 - 0.8 x fDATA 0.2 x fDATA Yes High Pass High Pass 0.2 to 0.4 x fDATA 1.2 - 1.4x fDATA 0.2 x fDATA No Low Pass High Pass 0 - 0.4 x fDATA 1.6 - 2x fDATA 0.4 x fDATA Yes Limitations on Signal BW and Final Output Frequency in X4L and X8 Modes For very wide bandwidth signals, the FIR3 pass-band (0 – 0.4 x FDAC/2) can limit the range of the final output frequency. For example in X4L FMIX CMIX mode (4x interpolation with FMIX after FIR1), at the maximum input data rate FIN = 125 MSPS the input signal can be ±50 MHz before running into the transition band of FIR1. After 2x interpolation, FIR3 limits the signal to ±100 MHz (0.4 x 250 MHz). Therefore, at the maximum signal bandwidth, FMIX can mix up to 50 MHz and still fall within the passband of FIR3. This results in gaps in the final output frequency between FMIX alone (0 MHz to 50 MHz) and FMIX + CMIX with fDAC/4 (75 MHz to 175 MHz) and FMIX + fDAC/2 (200 MHz to 250 MHz). 35 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 In practice, it may be possible to extend the signal into the FIR3 transition band. Referring to Figure 36 in the Digital Filter section above, if 0.5 dB of attenuation at the edge of the signal can be tolerated, then the signal can be extended up to 0.44 x FIN. This would extend the range of FMIX in the example to 60 MHz. Fine Mixer (FMIX) The fine mixer block FMIX uses a numerically controlled oscillator (NCO) with a 32-bit frequency register freq(31:0) and a 16-bit phase register phase(15:0) to provide sin and cos for mixing. The NCO tuning frequency is programmed in registers 0x09 through 0x0C. Phase offset is programmed in registers 0xD and 0xE. A block-diagram of the NCO is shown in Figure 39. 32 16 32 Frequency Register 32 Σ Accumulator 32 16 16 Σ sin Look-Up Table 16 cos CLK RESET 16 fDAC PHSTR Phase Register B0026-01 Figure 39. Block-Diagram of the NCO Synchronization of the NCO occurs by resetting the NCO accumulator to zero with assertion of PHSTR. See the Fine Mixer Synchronization section below. Frequency word freq in the frequency register is added to the accumulator every clock cycle. The output frequency of the NCO is (freq 2 32) ƒ freq ƒ NCO_CLK NCO_CLK 31 ƒ for freq 2 for freq 2 31 ƒ NCO NCO 2 32 2 32 where fNCO_CLK is the clock frequency of the NCO circuit. In X4 mode, the NCO clock frequency is the same as the DAC sample rate fDAC. The maximum clock frequency the NCO can operate at is 320 MHz – in X4 FMIX mode, where FMIX operates at the DAC update rate, the DAC updated rate will be limited to 320 MSPS. In X2, X4L and X8 modes, the NCO circuit is followed by a further 2x interpolation and so fNCO_CLK = fDAC/2 and operates at fDAC = 500 MHz. Treating channels A and B as a complex vector I + I x Q where I(t) = A(t) and Q(t) = B(t), the output of FMIX IOUT(t) and QOUT(t) is IOUT(t) = (IIN(t)cos(2πfNCOt + δ) – QIN(t)sin(2πfNCOt + δ)) x 2(NCO_GAIN – 1) QOUT(t) = (IIN(t)sin(2πfNCOt + δ) + QIN(t)cos(2πfNCOt + δ)) x 2(NCO_GAIN – 1) Where t is the time since the last resetting of the NCO accumulator, δ is the initial accumulator value and NCO_GAIN, bit 6 in register CONFIG2, is either 0 or 1. δ is given by δ = 2π x phase/216. The maximum output amplitude of FMIX occurs if IIN(t) and QIN(t) are simultaneously full scale amplitude and the sine and cosine arguments 2πfNCOt + δ = (2N-1) x π/4 (N = 1, 2, ...). With NCO_GAIN = 0, the gain through FMIX is sqrt(2)/2 or – 3 dB. This loss in signal power is in most cases undesirable, and it is recommended that the gain function of the QMC block be used to increase the signal by 3 dB to 0 dBFS by setting qmca_gain and qmcb_gain each to 1446 (decimal). With NCO_GAIN = 1, the gain through FMIX is sqrt(2) or + 3 dB, which can cause clipping of the signal if IIN(t) and QIN(t) are simultaneously near full scale amplitude and should therefore be used with caution. 36 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Coarse Mixer (CMIX) The coarse mixer block provides mixing capability at the DAC output rate with fixed frequencies of FS/2 or FS/4. The coarse mixer output phase sequence is selected by the cm_mode(3:0) bits in register CONFIG2 and is shown in Table 10. 37 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 10. Coarse Mixer Sequences cm_mode(3:0) Mixing Mode Sequence 00XX No mixing 0100 fDAC/2 DAC A = {-A +A -A +A …} DAC B = {-B +B -B +B …} 0101 fDAC/2 DAC A = {-A +A -A +A …} DAC B = {+B -B +B -B …} 0110 fDAC/2 DAC A = {+A -A +A -A …} DAC B = {-B +B -B +B …} 0111 fDAC/2 DAC A = {+A -A +A -A …} DAC B = {+B -B +B -B …} 1000 fDAC/4 DAC A = {+A -B -A +B …} DAC B = {+B +A -B -A …} 1001 fDAC/4 DAC A = {+A -B -A +B …} DAC B = {-B -A +B +A …} 1010 fDAC/4 DAC A = {-A +B +A -B …} DAC B = {+B +A -B -A …} 1011 fDAC/4 DAC A = {-A +B +A -B …} DAC B = {-B -A +B +A …} 1100 -fDAC/4 DAC A = {+A +B -A -B …} DAC B = {+B -A -B +A …} 1101 -fDAC/4 DAC A = {+A +B -A -B …} DAC B = {-B +A +B -A …} 1110 -fDAC/4 DAC A = {-A -B +A +B …} DAC B = {+B -A -B +A …} 1111 -fDAC/4 DAC A = {-A -B +A +B …} DAC B = {-B +A +B -A …} The output of CMIX is complex. For a real output, either DACA or DACB can be used and the other DAC slept, the difference being the phase sequence. Quadrature Modulator Correction (QMC) The quadrature modulator correction (QMC) block provides a means for changing the phase balance of the complex signal to compensate for I and Q imbalance present in an analog quadrature modulator. The QMC block is limited in operation to a clock rate of 320 MSPS. The block diagram for the QMC block is shown in Figure 40. The QMC block contains three programmable parameters. Registers qma_gain and qmb_gain control the I and Q path gains and are 11 bit values with a range of 0 to approximately 2. Note that the I and Q gain can also be controlled by setting the DAC full-scale output current (see below). Register qm_phase controls the phase imbalance between I and Q and is a 10-bit value with a range of –1/2 to approximately ½. LO feedthrough can be minimized by adjusting the DAC offset feature described below. An example of sideband optimization using the QMC block and gain adjustment is shown in Figure 41. The QMC phase adjustment in combination with the DAC gain adjustment can reduce the unwanted sideband signal from ~40 dBc to > 65 dBc. Note that mixing in the CMIX block after the QMC correction will destroy the I and Q phase compensation information from the QMC block. 38 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 qma_gain/210 {0, 1/210 , …, 2 −1/210 } 11 I(t) Σ X X Q(t) 10 qm_phase/210 {−1/2, −1/2 + 1/210, … , 1/2 −1/210} X 11 qmb_gain/210 {0, 1/2 10 , …, 2 −1/210 } Figure 40. QMC Block Diagram L O L O sideband sideband Uncorrected Corrected Figure 41. Example of Sideband Optimization Using QMC Phase and Gain Adjustments DAC Offset Control Registers qma_offset and qmb_offset control the I and Q path offsets and are 13-bit values with a range of –4096 to 4095. The DAC offset value adds a digital offset to the digital data before digital-to-analog conversion. The qma_gain and qmb_gain registers can be used to backoff the signal before the offset to prevent saturation when the offset value is added to the digital signal. 39 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 qma_offset {−4096, −4095, … , 4095 } 13 I Σ Q Σ 13 qmb_offset {−4096, −4095, … , 4095 } Figure 42. DAC Offset Block Analog DAC Gain The full-scale DAC output current can be set by programming the daca_gain and dacb_gain registers. The DAC gain value controls the full-scale output current. I fullscale 16Vextio (GAINCODE 1) 1 FINEGAIN R 16 3072 biasj where GAINCODE = daca_gain[11:8] or dacb_gain[11:8] is the coarse gain setting ( 0 to 15) and FINEGAIN = daca_gain[7:0] or dacb_gain[7:0] (-128 to 127) is the fine gain setting. Clock Modes In the DAC5687, the internal clocks (1x, 2x, 4x, and 8x as needed) for the logic, FIR interpolation filters, and DAC are derived from a clock at either the input data rate using an internal PLL (PLL clock mode) or DAC output sample rate (external clock mode). Power for the internal PLL blocks (PLLVDD and PLLGND) are separate from the other clock generation blocks power (CLKVDD and CLKGND), thus minimizing phase noise within the PLL. The DAC5687 has three clock modes for generating the internal clocks (1x, 2x, 4x, and 8x as needed) for the logic, FIR interpolation filters, and DACs. The clock mode is set using the PLLVDD pin and dual_clk in register CONFIG1. 1. PLLVDD = 0 V and dual_clk = 0: EXTERNAL CLOCK MODE In EXTERNAL CLOCK MODE, the user provides a clock signal at the DAC output sample rate through CLK2/CLK2C. CLK1/CLK1C and the internal PLL are not used. LPF and CLK1/CLK1C pins can be left unconnected. The input data rate clock and interpolation rate are selected by the bits interp(1:0) in register CONFIG0 and is output through the PLLLOCK pin. The PLLLOCK clock can be used to drive the input data source (such as digital upconverter) that sends the data to the DAC. Note that the PLLLOCK delay relative to the input CLK2 rising edge (td(PLLLOCK)) in Figure 43 and Figure 44) increases with increasing loads. The input data is latched on either the rising (inv_plllock = 0) or falling edge (inv_plllock = 1) of PLLLOCK, which is sensed internally at the output pin. 40 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 PLLLOCK td(PLLLOCK) CLK2 th(DATA) ts(DATA) DA[15:0] A0 A1 A2 A3 AN AN+1 DB[15:0] B0 B1 B2 B3 BN BN+1 Figure 43. Dual Bus Mode Timing Diagram for External Clock Mode (PLLLOCK Rising Edge) PLLLOCK td(PLLLOCK) CLK2 th(DATA) ts(DATA) DA[15:0] A0 A1 A2 A3 AN AN+1 DB[15:0] B0 B1 B2 B3 BN BN+1 Figure 44. Dual Bus Mode Timing Diagram for External Clock Mode (PLLLOCK Falling Edge) 2. PLLVDD = 3.3 V (dual_clk can be 0 or 1 and is ignored): PLL CLOCK MODE In PLL CLOCK MODE, you drive the DAC at the input sample rate (unless the data is mux’d) through CLK1/CLK1C. CLK2/CLK2C is not used. In this case, there is no phase ambiguity on the clock. The DAC generates the higher speed DAC sample rate clock using an internal PLL/VCO. In PLL clock mode, the user provides a differential external reference clock on CLK1/CLK1C. A type four phase-frequency detector (PFD) in the internal PLL compares this reference clock to a feedback clock and drives the PLL to maintain synchronization between the two clocks. The feedback clock is generated by dividing the VCO output by 1x, 2x, 4x, or 8x as selected by the prescaler (div[1:0]). The output of the prescaler is the DAC sample rate clock and is divided down to generate clocks at ÷ 2, ÷4, and ÷ 8. The feedback clock is selected by the registers sel(1:0), which is fed back to the PFD for synchronization to the input clock. The feedback clock is also used for the data input rate, so the ratio of DAC output clock to feedback clock sets the interpolation rate of the DAC5687. The PLLLOCK pin is an output indicating when the PLL has achieved lock. An external RC low-pass PLL filter is provided by the user at pin LPF. See the Low-Pass Filter section for filter setting calculations. This is the only mode where the LPF filter applies. 41 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 CLK2 th(DATA) ts(DATA) DA[15:0] A0 A1 A2 A3 AN AN+1 DB[15:0] B0 B1 B2 B3 BN BN+1 Figure 45. Dual Bus Mode Timing Diagram (PLL Mode) DIV[1:0] LPF CLK1 CLK1C CLK2 CLK2C CLK Buffer PFD Charge Pump VCO /1 00 /2 01 /4 10 /8 11 PLLVDD 1 Fdac CLK Buffer 0 00 1 ´2 01 0 ´1 10 11 Data Latch PLLLOCK PLLVDD Fdac/2 ´2 Fdac/4 ´4 /2 0 1 /2 Fdac/4 ´4L /2 Fdac/8 ´8 Data Lock D[15:0] INTERL SEL[1:0] B0053-02 Figure 46. Clock Generation Architecture in PLL Mode Power for the internal PLL blocks (PLLVDD and PLLGND) are separate from the other clock generation blocks power (CLKVDD and CLKGND), thus minimizing PLL phase noise. 3) PLLVDD = 0 V and dual_clk = 1: DUAL CLOCK MODE In DUAL CLOCK MODE, the DAC is driven at the DAC sample rate through CLK2/CLK2C and the input data rate through CLK1/CLK1C. There are two options in dual clock mode: with FIFO (inv_plllock set) and without FIFO (inv_plllock clear). If the FIFO is not used, the CLK1/CLK1C input is used to set the phase of the internal clock divider. In this case, the edges of CLK1 and CLK2 must be aligned to within ±t_align (Figure 47), defined as 42 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 f align 1 2F 0.5 ns CLK2 where fCLK2 is the clock frequency at CLK2. For example, talign = 0.5 ns at fCLK2 = 500 MHz and 1.5 ns at fCLK2 = 250 MHz. If the FIFO is enabled (inv_plllock set) in dual clock mode, then CLK1 is only used as an input latch (Figure 48) and is independent from the internal divided clock generated from CLK2/CLK2C and there is no alignment specification. However, the FIFO needs to be synchronized by one of the methods listed in SYNC_CNTL register and the latency of the DAC can be up to one clock cycle different depending on the phase relationship between CLK1 and the internally divided clock. CLK2 CLK1 ∆ < talign DA[15:0] DB[15:0] th ts Figure 47. Dual Clock Mode Without FIFO CLK1 DA[15:0] DB[15:0] th ts Figure 48. Dual Clock Mode With FIFO The CDC7005 from Texas Instruments is recommended for providing phase aligned clocks at different frequencies for this application. Input FIFO In DAC clock mode, where the DAC5687 is clocked at the DAC update rate, the DAC5687 has an optional input FIFO that allows latching of DA[15:0], DB[15:0] and PHSTR based on a user provided CLK1/CLK1C input or the input data rate clock provided to the PLLLOCK pin. The FIFO can be bypassed by setting register fifo_bypass in CONFIG0 to 1. The input interface FIFO incorporates a four sample register file, an input pointer, and an output pointer. Initialization of the FIFO pointers can be programmed to one of seven different sources. 43 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 DA(15:0), DB(15:0), PHSTR q_a 0 1 q_in D Q D Q S in_sel_a 0 1 q_b D Q MUX S in_sel_b 1 0 D Q q_out resynchonized da(15:0), db(15:0) and phstr S 0 1 q_c D Q fifo_bypass S in_sel_c 0 1 q_d D Q sel_q_a sel_q_b S in_sel_d input pointer generation MUX sync sel_q_c sel_q_d output pointer generation PLLLOCK PLL VCO clk_out clk_in clock generator CLK2 CLK1 CLK1C {PLLVDD, inv_plllock,, dual_clk} sync source {TXENABLE, PHSTR, QFLAG, DB(15), DA(15) oneshot oneshot, SIF write, always off} CLK2C Figure 49. DAC5687 Input FIFO Logic Initialization of the FIFO block involves selecting and asserting a synchronization source. Initialization causes the input and output pointers to be forced to an offset of 2; the input pointer will be forced to the in_sel_a state while the output pointer will be forced to the sel_q_c state. This initialization of the input and output pointers can cause discontinuities in a data stream and should therefore be handled at startup. Table 11. Synchronization Source Selection 44 sync_fifo (2:0) Synchronization Source 000 txenable pin 001 phstr pin 010 qflag pin 011 db(15) 100 da(15) first transition (one shot) 101 sync now with SIF write (always on) 110 sync source disabled (always off) 111 sync now with SIF write (always on) DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 All possible sync sources are registered with clk_in and then passed through a synchronous rising edge detector. D Q TXENABLE MUX 000 001 D Q PHSTR 010 resync_fifo_in 011 1 101 0 110 1 111 D Q D Q resync_fifo_out D Q D Q DB(15) D Q D Q 100 D Q QFLAG D Q sync D Q D Q D Q sync_fifo(2:0) D DA(15) Q D Q MUX PLLLOCK PLL VCO clk_in sync_fifo = “100” DA(15) first rising edge clk_out clock generator CLK2 CLK1 CLK2C CLK1C {PLLVDD, inv_plllock, dual_clk} Figure 50. DAC5687 FIFO Synchronization Source Logic For example, if TXENABLE is selected as the sync source, a low-to-high transition on the TXENABLE pin causes the pointers to be initialized. Once initialized, the FIFO input pointer advances using clk_in and the output pointer advances using clk_out, providing an elastic buffering effect. The phase relationship between clk_in and clk_out can wander or drift until the output pointer overruns the input pointer or vice versa. Even/Odd Input Mode The DAC5687 has a double data rate input mode that allows both input ports to be used to multiplex data onto one DAC channel (A). In the Even/Odd mode, the FIR3 filter can be used to interpolate the data by 2x. The even/odd input mode is enabled by setting half_rate in CONFIG3. The maximum input rate for each port is 250 MSPS, for a combined rate of 500 MSPS. Synchronization The DAC5687 has several digital circuits that can be synchronized to a known state. The circuits that can be synchronized are the fine mixer (NCO), coarse mixer (fixed fs/2 or fs/4 mixer), the FIFO input and output pointers, and the internal clock divider. 45 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 12. Synchronization in Different Clock Modes Clock Mode PLLVDD Pin Serial Interface Register Bits Single External Clock without FIFO 0V Single External Clock with FIFO 0V Dual External Clock without FIFO 0V 1 Dual External Clock with FIFO 0V PLL Enabled 3.3 V DA, DB, PHSTR, and TXENABLE Latch Description Signal at the PLLLOCK output pin is used to clock the PHSTR signal into the chip. The PLLLOCK output clock is generated by dividing the CLK2/CLK2C input signal by the programmed interpolation and interface settings. fifo_bypass dual_clk inv_plllock 1 0 0 PLLLOCK rising edge 1 PLLLOCK falling edge 0 PLLLOCK rising edge 1 PLLLOCK falling edge 1 0 CLK1/CLK1C The CLK1/CLK1C input signal is used to clock in the PHSTR signal. CLK1/CLK1C and CLK2/CLK2C are both input to the chip and the phase relationship must be tightly controlled 0 1 1 CLK1/CLK1C The CLK1/CLK1C input signal is used to clock in the PHSTR signal. CLK1/CLK1C and CLK2/CLK2C are both input to the chip, but no phase relationship is required. The FIFO input circuits are used to manage the clock domain transfers. The FIFO must be initialized in this mode. 1 0 0 CLK1/CLK1C The CLK1/CLK1C input signal is used to clock in the PHSTR signal. The FIFO must be bypassed when the PLL is enabled. 0 0 Signal at the PLLLOCK output pin is used to clock the PHSTR signal into the chip. The PLLLOCK output clock is generated by dividing the CLK2/CLK2C input signal by the programmed interpolation and interface settings. Enabling the FIFO allows the chip to function with large loads on the PLLLOCK output pin at high input rates. The FIFO must be initialized first in this mode. NCO Synchronization The phase accumulator in the NCO block (see the Fine Mixer (FMIX) section and Figure 39 for a description of the NCO) can be synchronously reset when PHSTR is asserted. The PHSTR signal passes through the input FIFO block, using the input clock associated with the clocking mode. If the FIFO is enabled, there can be some uncertainty in the exact instant the PHSTR synchronization signal arrives at the NCO accumulator due to the elastic capabilities of the FIFO. For example, in dual-clock mode with the FIFO enabled, the internal clock generator divides down the CLK2/CLK2C input signal to generate the FIFO output clock. The phase of this generated clock will be unknown externally, resulting in an uncertainty of the exact PHSTR instant of as much as a few input clock cycles. FIFO PHSTR D Q D Q phstr sync to NCO D Q D Q MUX clk_out PLLLOCK clk_in D Q D Q clk_nco clock generator PLL VCO CLK2 CLK2C CLK1 CLK1C {PLLVDD, inv_plllock,, dual_clk}} Figure 51. Logic Path for PHSTR Synchronization Signal to NCO 46 phase accumulator reset DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 The serial interface includes a sync_nco bit in register SYNC_CNTL, which needs to be set for the PHSTR input signal to initialize the phase accumulator. The NCO uses a rising edge detector to perform the synchronous reset of the phase accumulator. Due to the pipelined nature of the NCO, the latency from the phstr sync signal at the FIFO output to the instant the phase accumulator is cleared is 13 fNCO clock cycles (fNCO = fDAC in X4 mode, fNCO = fDAC/2 in X2, X4L, and X8 modes). In 2x interpolation mode with the inverse sinc filter disabled, overall latency from PHSTR input to DAC output is ~100 input clock cycles. PHSTR only needs to be asserted for one clk_in period clk_in Input delay line + FIFO delay clk_out phstr at FIFO output clk_nco phase accumulator reset phase_accum 13 clk_nco cycles Figure 52. NCO Phase Accumulator Reset Synchronization Timing Coarse Mixer (CMIX) Synchronization The coarse mixer implements the fDAC/2 and fDAC/4 (and – fDAC/4) fixed complex mixing operation using simple complements of the datapath signals to create the proper sequences. The sequences are controlled using a simple counter and this counter can be synchronously reset using the PHSTR signal. Similar to the NCO, the PHSTR signal used by the coarse mixer is from the FIFO output. This introduces the same uncertainty effect due to the FIFO input to output pointer relationship. Bypassing the FIFO and using the dual external clock mode without FIFO eliminates this uncertainty for systems using multiple DAC5687 devices when this cannot be tolerated. Using the internal PLL, as with the NCO, allows the complete control and synchronization of the coarse mixer. 47 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 sync_cm D Q PHSTR D Q D Q FIFO phstr sync to coarse mixer D Q D Q D Q Sequencer Reset MUX clk_cmix PLLLOCK clk_in clk_out PLL VCO Clock Generator CLK1 CLK2 CLK2C CLK1C {PLLVDD, inv_plllock, dual_clk} Figure 53. Logic Path for PHSTR Synchronization Signal to CMIX Reset To enable the PHSTR synchronous reset, the serial interface bit sync_cm in register SYNC_CNTL must be set. The coarse mixer sequence counter will be held reset when PHSTR is low and operates when PHSTR is high. Only needs to be high for one clk_in period PHSTR clk_in Input delay line + FIFO delay clk_out phstr at FIFO Output clk_cmix Sequencer Reset Sequencer fs/2 0 180 0 180 Sequencer fs/4 0 90 180 270 0 90 Figure 54. CMIX Reset Synchronization Timing In addition to the reset function provided by the PHSTR signal, the phstr_del(1:0) bits in register ATEST allow the user to select the initial (reset) state. Changing the cm_mode lower 2 bits produces the same phase shift results. Table 13. Initial State of CMIX After Reset 48 Fix Mix Selection phstr_del(1:0) fS/2 00 and 10 Initial State at PHSTR Normal fS/2 01 and 11 180 degree shift fS/4 00 Normal fS/4 01 90 degree shift DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 13. Initial State of CMIX After Reset (continued) Fix Mix Selection phstr_del(1:0) Initial State at PHSTR fS4 10 180 degree shift fS/4 11 270 degree shift Input Clock Synchronization of Multiple DAC5687s For applications where multiple DAC5687 chips are used, clock synchronization is best achieved by using dual-clock mode with the FIFO disabled or the PLL-clock mode. In the dual-clock mode with FIFO disabled, an appropriate clock PLL such as the CDC7005 is required to provide the DAC and input rate clocks that meet the skew requirement t_align (see Figure 47). An example for synchronizing multiple DAC5687 devices in dual clock mode with two CDC7005 is shown in Figure 55. When using the internal PLL-clock mode, synchronization of multiple using PHSTR is completely deterministic due to the phase/frequency detector in the PLL feedback loop. All chips using the same CLK1/CLK1C input clock will have identical internal clocking phases. Ref fINPUT CDC7005 #1 Ref CDC7005 #2 Y0 Y0 Y1 Y1 fINPUT CLK1 CLK1C fDAC CLK2 CLK2C Y2 Y2 Y3 Y3 fINPUT CLK1 CLK1C fDAC CLK2 CLK2C Y0 Y0 Y1 Y1 fINPUT CLK1 CLK1C fDAC CLK2 CLK2C Y2 Y2 Y3 Y3 fINPUT CLK1 CLK1C fDAC CLK2 CLK2C DAC5687 #1 DAC5687 #2 DAC5687 #3 DAC5687 #4 Figure 55. Block Diagram for Clock Synchronization of Multiple DAC5687 Devices in Dual-Clock Mode Reference Operation The DAC5687 comprises a bandgap reference and control amplifier for biasing the full-scale output current. The full-scale output current is set by applying an external resistor RBIAS to pin BIASJ. The bias current IBIAS through resistor RBIAS is defined by the on-chip bandgap reference voltage and control amplifier. The full-scale output current equals 16 times this bias current. The full-scale output current IOUTFS can thus be expressed as: IOUTFS = 16 × IBIAS = 16 × VEXTIO / RBIAS where VEXTIO is the voltage at terminal EXTIO. The bandgap reference voltage delivers an accurate voltage of 1.2 V. This reference is active when terminal EXTLO is connected to AGND. An external decoupling capacitor CEXT of 0.1 µF should be connected externally to terminal EXTIO for compensation. The bandgap reference can additionally be used for external reference operation. In that case, an external buffer with high impedance input should be applied in order to limit the bandgap load current to a maximum of 100 nA. The internal reference can be disabled and overridden by an external reference by connecting EXTLO to AVDD. Capacitor CEXT may hence be omitted. Terminal EXTIO thus serves as either input or output node. 49 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 The full-scale output current can be adjusted from 20 mA down to 2 mA by varying resistor RBIAS or changing the externally applied reference voltage. The internal control amplifier has a wide input range, supporting the full-scale output current range of 20 mA. DAC Transfer Function The CMOS DAC’s consist of a segmented array of NMOS current sinks, capable of sinking a full-scale output current up to 20 mA. Differential current switches direct the current of each current source through either one of the complementary output nodes IOUT1 or IOUT2. Complementary output currents enable differential operation, thus canceling out common mode noise sources (digital feed-through, on-chip and PCB noise), dc offsets, even order distortion components, and increasing signal output power by a factor of two. The full-scale output current is set using external resistor RBIAS in combination with an on-chip bandgap voltage reference source (+1.2 V) and control amplifier. Current IBIAS through resistor RBIAS is mirrored internally to provide a full-scale output current equal to 16 times IBIAS. The full-scale current IOUTFS can be adjusted from 20 mA down to 2 mA. The relation between IOUT1 and IOUT2 can be expressed as: IOUT1 = -IOUTFS– IOUT2 We will denote current flowing into a node as - current and current flowing out of a node as + current. Since the output stage is a current sink the current can only flow from AVDD into the IOUT1 and IOUT2 pins. If IOUT2 = 5 mA and IO(FS) = 20 mA then: IOUT1 = -20 – (-5) = -15 mA The output current flow in each pin driving a resistive load can be expressed as: IOUT1 = IOUTFS × CODE / 65536 IOUT2 = IOUTFS × (65535 – CODE) / 65536 where CODE is the decimal representation of the DAC data input word. For the case where IOUT1 and IOUT2 drive resistor loads RL directly, this translates into single ended voltages at IOUT1 and IOUT2: VOUT1 = AVDD – I IOUT1 I × RL VOUT2 = AVDD – I IOUT2 I × RL Assuming that the data is full scale (65535 in offset binary notation) and the RL is 25 Ω, the differential voltage between pins IOUT1 and IOUT2 can be expressed as: VOUT1 = AVDD – I -20 mA I x 25 Ω = 2.8 V VOUT2 = AVDD – I -0 mA I x 25 Ω = 3.3 V VDIFF = VOUT1 – VOUT2 = 0.5 V Note that care should be taken not to exceed the compliance voltages at node IOUT1 and IOUT2, which would lead to increased signal distortion. Analog Current Outputs Figure 56 shows a simplified schematic of the current source array output with corresponding switches. Differential switches direct the current of each individual NMOS current source to either the positive output node IOUT1 or its complementary negative output node IOUT2. The output impedance is determined by the stack of the current sources and differential switches, and is typically >300 kΩ in parallel with an output capacitance of 5 pF. The external output resistors are referred to an external ground. The minimum output compliance at nodes IOUT1 and IOUT2 is limited to AVDD – 0.5 V, determined by the CMOS process. Beyond this value, transistor breakdown may occur resulting in reduced reliability of the DAC5687 device. The maximum output compliance voltage at nodes IOUT1 and IOUT2 equals AVDD + 0.5 V. Exceeding the minimum output compliance voltage adversely affects distortion performance and integral non-linearity. The optimum distortion performance for a single-ended or differential output is achieved when the maximum full-scale signal at IOUT1 and IOUT2 does not exceed 0.5 V. 50 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 AVDD RLOAD RLOAD IOUT1 IOUT2 S(1) S(1)C S(2) S(2)C S(N) S(N)C S0032-01 Figure 56. Equivalent Analog Current Output The DAC5687 can be easily configured to drive a doubly terminated 50-Ω cable using a properly selected RF transformer. Figure 57 and Figure 58 show the 50-Ω doubly terminated transformer configuration with 1:1 and 4:1 impedance ratio, respectively. Note that the center tap of the primary input of the transformer has to be connected to AVDD to enable a dc-current flow. Applying a 20-mA full-scale output current would lead to a 0.5 VPP for a 1:1 transformer and a 1-VPP output for a 4:1 transformer. The low dc-impedance between the IOUT1 or IOUT2 and the transformer center tap sets the center of the ac-signal at AVDD, so the 1-VPP output for the 4:1 transformer results in an output between AVDD + 0.5 V and AVDD – 0.5 V. AVDD (3.3 V) 50 Ω 1:1 IOUT1 RLOAD 50 Ω 100 Ω IOUT2 50 Ω AVDD (3.3 V) S0033-01 Figure 57. Driving a Doubly Terminated 50-Ω Cable Using a 1:1 Impedance Ratio Transformer 51 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 AVDD (3.3 V) 100 Ω 4:1 IOUT1 RLOAD 50 Ω IOUT2 100 Ω AVDD (3.3 V) S0033-02 Figure 58. Driving a Doubly Terminated 50-Ω Cable Using a 4:1 Impedance Ratio Transformer Combined Output Termination The DAC5687 DAC A and DAC B outputs can be summed together as shown in Figure 59 to provide a 40-mA full-scale output for increased output power. 1:1 IOUTA1 RLOAD 50 Ω IOUTA2 IOUTB1 AVDD (3.3 V) IOUTB2 S0069-01 Figure 59. Combined Output Termination Using a 1:1 Impedance Ratio Transformer into 50-Ω Load For the case where the digital codes for the two DACs are identical, the termination results in a full scale swing of 2 VPP into the 50-Ω load, or 10 dBm. This is 6 dB higher than the 4:1 output termination recommended for a single DAC output. There are two methods to produce identical DAC codes. In modes where there is mixing between digital channels A and B, i.e., when channels A and B are isolated, the identical data can be sent to both input ports to produce identical DAC codes. Channels A and B are isolated when FMIX is disabled, the QMC is disabled or enabled with QMC phase register set to 0, and CMIX is disabled or set to fDAC/2. Note that frequency upconversion is still possible using the high-pass filter setting and CMIX fDAC/2. Alternatively, by applying the input data on one input port only and setting the other input port to mid-scale (zero), the NCO can be used to duplicate the output of the active input channel in the other channel by setting the frequency to zero, phase to 8192 and NCO_GAIN = 1 and QMC gain = 1446. Assuming I(t) is the wanted signal and Q(t) = 0, this is demonstrated by the simplification of the NCO equations in the Fine Mixer (FMIX) section: IOUT(t) = (IIN(t)cos(2π x 0 x t + π/4) – 0 x sin((2π x 0 x t + π/4)) x 2(1 – 1) = IIN(t)cos(π/4) = IIN(t)/2½ QOUT(t) = (IIN(t)sin(2π x 0 x t + π/4) + 0 x cos(2π x 0 x t + π/4)) x 2(1 – 1) = IIN(t)sin(π/4) = IIN(t)/2½ 52 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Applying the QMC gain of 1446, equivalent to 2½, increases the signal back to unity gain through the FMIX and the QMC blocks. Note that with this termination, the DAC side of the transformer is not 50-Ω terminated and therefore may result in reflections when used with a cable output. Digital Inputs Figure 60 shows a schematic of the equivalent CMOS digital inputs of the DAC5687. DA[0..15], DB[0..15], SLEEP, PHSTR, TXENABLE, QFLAG, SDIO, SCLK, and SDENB have pull-down resistors and RESETB has a pull-up resistor internal to the DAC5687. See the specification table for logic thresholds. IOVDD IOVDD DA[15:0] DB[15:0] SLEEP PHSTR TxENABLE QFLAG SDIO SCLK SDENB Internal Digital In Internal Digital In RESETB IOGND IOGND Figure 60. CMOS/TTL Digital Equivalent Input Clock Inputs Figure 61 shows an equivalent circuit for the clock input. CLKVDD CLKVDD R1 10 kΩ Internal Digital In CLKVDD R1 10 kΩ CLK CLKC R2 10 kΩ R2 10 kΩ CLKGND S0028-01 Figure 61. Clock Input Equivalent Circuit Figure 62, Figure 63, and Figure 64 show various input configurations for driving the differential clock input (CLK/CLKC). 53 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Optional, May Be Bypassed for Sine Wave Input Swing Limitation CAC 0.1 µF 1:4 CLK RT 200 Ω CLKC Termination Resistor S0029-01 Figure 62. Preferred Clock Input Configuration Ropt 22 Ω CAC 0.01 µF Ropt 22 Ω 1:1 TTL/CMOS Source TTL/CMOS Source CLK Optional, Reduces Clock Feedthrough CLKC CLK CLKC 0.01 µF Node CLKC Internally Biased to CLKVDD2 S0030-01 Figure 63. Driving the DAC5687 With a Single-Ended TTL/CMOS Clock Source CAC 0.1 µF Differential ECL or (LV)PECL Source CLK + CAC 0.1 µF – 100 Ω CLKC RT 130 Ω RT 130 Ω RT 82.5 Ω RT 82.5 Ω VTT S0031-01 Figure 64. Driving the DAC5687 With Differential ECL/PECL Clock Source Power Up Sequence In all conditions, bring up DVDD first. If PLLVDD is powered (PLL on), CLKVDD should be powered before or simultaneously with PLLVDD. AVDD, CLKVDD, and IOVDD can be powered simultaneously or in any order. Within AVDD, the multiple AVDD pins should be powered simultaneously. 54 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 There are no specific requirements on the ramp rate for the supplies. Sleep Mode The DAC5687 features a power-down mode that turns off the output current and reduces the supply current to less than 5 mA over the supply range of 3 V to 3.6 V and temperature range. The power-down mode is activated by applying a logic level 1 to the SLEEP pin (e.g., by connecting pin SLEEP to AVDD). An internal pull-down circuit at node SLEEP ensures that the DAC5687 is enabled if the input is left disconnected. Power-up and power-down activation times depend on the value of external capacitor at node EXTIO. For a nominal capacitor value of 0.1-µF, power down takes less than 5 µs and approximately 3 ms to power back up. Application Information Designing the PLL Loop Filter Table 14. Optimum DAC5687 PLL Settings fDAC (MHZ) pll_freq pll_kv pll_div (1:0) fVCO / fDAC Estimated GVCO (MHz/V) 25 to 28.125 0 1 11 8 380 28.125 to 46.25 0 0 11 8 250 46.25 to 60 0 1 11 8 300 60 to 61.875 1 0 11 8 130 61.875 to 65 1 1 11 8 225 65 to 92.5 0 0 10 4 250 92.5 to 120 0 1 10 4 300 120 to 123.75 1 0 10 4 130 123.75 to 130 1 1 10 4 225 130 to 185 0 0 01 2 250 185 to 240 0 1 01 2 300 240 to 247.5 1 0 01 2 130 247.5 to 260 1 1 01 2 225 260 to 370 0 0 00 1 250 370 to 480 0 1 00 1 300 480 to 495 1 0 00 1 130 495 to 520 1 1 00 1 225 The optimized DAC5687 PLL settings based on the VCO frequency MIN and MAX values (see the digital specifications) as a function of fDAC are listed in Table 14. To minimize phase noise at a given fDAC, pll_freq, pll_kv, and the pll_div have been chosen so GVCO is minimized and within the MIN and MAX frequency for a given setting. For example, if fDAC = 245.76 MHz, pll_freq is set to 1, pll_kv is set to 0 and pll_div(1:0) is set to 01 (divide by 2) to lock the VCO at 491.52 MHz. The external loop filter components C1, C2, and R1 are set by the GVCO, N = fVCO/fDATA = fVCO x Interpolation/fDAC, the loop phase margin φd and the loop bandwidth ωd. Except for applications where abrupt clock frequency changes require a fast PLL lock time, it is suggested that ωd be set to at least 80 degrees for stable locking and suppression of the phase noise side lobes. Phase margins of 60 degrees or less can be sensitive to board layout and decoupling details. C1, C2, and R1 are then calculated by the following equations 55 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 C1 1 1 2 3 C2 1 2 3 where, K Kvco 1 d 2 tan sec d d d R1 3 2 1 (3 2) 1 2 tan d sec d (1) 3 d tan sec d d d (2) and charge pump current: iqp = 1 mA vco gain: KVCO = 2πxGVCO rad/V FVCO/FDATA: N = {2, 4, 8, 16, 32} phase detector gain: Kd = iqp × (2 × N) – 1 A/rad An Excel spreadsheet is provided by Texas Instruments for automatically calculating the values for C1, C2, and R. Completing the example given above with: Parameter Value Units GVCO 1.30E+02 MHz/V ωd 0.50E+00 MHz N 4 φd 80 Degrees C1 (F) C2 (F) R (Ω) 3.74E-08 2.88E-10 9.74E+01 The component values are: As the PLL characteristics are not sensitive to these components, the closest 20% tolerance capacitor and 1% tolerance resistor values can be used. If the calculation results in a negative value for C2 or an unrealistically large value for C1, then the phase margin may need to be reduced slightly. DAC5687 Passive Interface-to-Analog Quadrature Modulators The DAC5687 has a maximum 20-mA full-scale output and a compliance range of AVDD ±0.5 V. The TRF3701 or TRF3702 analog quadrature modulators (AQM) require a common-mode of approximately 3.7 V and 1.5 V to 2-VPP differential swing. A resistive network as shown in Figure 65 can be used to translate the common mode voltage between the DAC5687 and TRF3701 or TRF3702. The voltage at the DAC output pins for a full-scale sine wave is centered at approximately AVDD with a 1-VPP single ended (2-VPP differential). The voltage at the TRF3701/2 input pins is centered at 3.7 V and swings 0.76-VPP single ended (1.56-VPP differential), or 2.4 dB of insertion loss. 56 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 GND 5V 205 Ω 205 Ω 50 Ω 50 Ω 15.4 Ω 15.4 Ω TRF3701 TRF3702 15.4 Ω 15.4 Ω 205 Ω 205 Ω 50 Ω 50 Ω 5V GND B0046-01 Figure 65. DAC5687 Passive Interface to TRF3701/2 Analog Quadrature Modulator Changing the voltage levels and resistor values enable other common-mode voltages at the analog quadrature modulator input. For example, the network shown in Figure 66 can produce a 1.5-V common mode at the analog quadrature modulator input, with a 0.78-VPP single-ended swing (1.56-VPP differential swing), or 0.2-dB insertion loss. 5V 69.8 Ω −5 V 69.8 Ω 453 Ω 453 Ω 124 Ω 124 Ω AQM 124 Ω 124 Ω 69.8 Ω 69.8 Ω 5V 453 Ω 453 Ω −5 V B0046-02 Figure 66. DAC5687 Passive Interface With 1.5-V Common Mode at AQM Input Non-Harmonic Clock-Related Spurious Signals In interpolating DACs, imperfect isolation between the digital and DAC clock circuits generate spurious signals at frequencies related to the DAC clock rate. The digital interpolation filters in these DACs run at sub-harmonic frequencies of the output rate clock, where these frequencies are fDAC/2N, N = 1 - 3. For example, for X2 mode there is only one interpolation filter running at fDAC/2; for X4 and X4L modes, on the other hand, there are two interpolation filters running at fDAC/2 and fDAC/4. In X8 mode, there are three interpolation filters running at fDAC/2, fDAC/4, and fDAC/8. These lower-speed clocks for the interpolation filter mix with the DAC clock circuit and create spurious images of the wanted signal and second Nyquist-zone image at offsets of fDAC/2N. 57 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 To calculate the non-harmonic clock related spurious signals for a particular condition, we first determine the location of the spurious signals and then the amplitude. Location of the Spurious Signals The location of the spurious signals is determined by the DAC5687 output frequency (fSIG) and whether the output is used as a dual output complex signal to be fed to an analog quadrature modulator (AQM) or as a real IF signal from a single DAC output. Figure 67 shows the location of spurious signals for X2 mode as a function of fSIG/fDAC. For complex outputs, the spurious frequencies cover a range of -0.5 x fDAC to 0.5 x fDAC, with the negative complex frequency indicating that the spurious signal will fall in the opposite sideband at the output of the QAM from the wanted signal. For the real output, the phase information for the spurious signal is lost, and therefore what was a negative frequency for the complex output is a positive frequency for a real output. For the X2 mode, there is one spurious frequency with an absolute frequency less than 0.5 x fDAC. For a complex output in X2 mode, the spurious signal will always be offset fDAC/2 from the wanted signal at fSIG - fDAC/2. For a real output, as fSIG approaches fDAC/4, the spurious signal frequency falls at fDAC/2-fSIG, which will also approach fDAC/4. (a) Complex Output in X2 Mode (b) Real Output in X2 Mode 0.5 0.50 0.4 0.45 0.40 fSIG 0.2 Spurious Frequency/fDAC Spurious Frequency/fDAC 0.3 0.1 −0.0 −0.1 −0.2 fSIG − fDAC/2 −0.3 0.30 fSIG − fDAC/2 0.25 0.20 0.15 0.10 −0.4 −0.5 0.0 fSIG 0.35 0.05 0.1 0.2 0.3 fSIG/fDAC 0.4 0.5 G026 0.00 0.0 0.1 0.2 0.3 fSIG/fDAC 0.4 0.5 G027 Figure 67. Frequency of Clock Mixing Spurious Images in X2 Mode Figure 68 shows the location of spurious signals for X4 and X4L mode as a function of fSIG/fDAC. The addition of the fDAC/4 clock frequency for the first interpolation filter creates three new spurious signals. For a complex output, the nearest spurious signals are fDAC/4 offset from fSIG. For a real output, the signal due to fSIG - fDAC/4 and fSIG - fDAC x 3/4 falls in band as fSIG approaches fDAC/8 and fDAC x 3/8. This creates optimum real output frequencies fSIG = fDAC x N/16 (N = 1, 3, 5, and 7), where the minimum spurious product offset from fSIG is fDAC/8. 58 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 (a) Complex Output in X4 and X4L Modes (b) Real Output in X4 and X4L Modes 0.5 0.4 0.50 fSIG + fDAC/4 0.45 0.2 0.1 fSIG − fDAC/4 0.0 −0.1 fSIG − fDAC/2 −0.2 −0.3 0.35 fSIG 0.30 0.25 fSIG − fDAC/4 0.20 fSIG − fDAC/2 0.15 0.10 fSIG − fDAC*3/4 −0.4 −0.5 0.0 fSIG − fDAC*3/4 0.40 fSIG Spurious Frequency/fDAC Spurious Frequency/fDAC 0.3 fSIG + fDAC/4 0.1 0.2 0.05 0.3 0.4 fSIG/fDAC 0.00 0.0 0.5 0.1 0.2 0.3 0.4 fSIG/fDAC G028 0.5 G029 Figure 68. Frequency of Clock Mixing Spurious Images in X4 and X4L Modes Figure 69 shows the location of spurious signals for X8 mode as a function of fSIG/fDAC. The addition of the fDAC/8 clock frequency for the first interpolation filter creates four new spurious signals. For a complex output, the nearest spurious signals are fDAC/8 offset from fSIG. For a real output, the optimum real output frequencies fSIG = fDAC x N/16 (N = 3 and 5), where the minimum spurious product offset from fSIG is fDAC/8. (a) Complex Output in X8 Mode (b) Real Output in X8 Mode 0.5 0.50 fSIG + fDAC/4 0.45 0.3 0.40 0.2 fSIG fSIG + fDAC/8 Spurious Frequency/fDAC Spurious Frequency/fDAC fSIG + fDAC/4 0.4 0.1 fSIG − fDAC/4 −0.0 −0.1 fSIG − fDAC/8 fSIG − fDAC/2 −0.2 −0.3 0.35 0.30 fSIG fSIG + fDAC/8 0.25 0.20 fSIG − fDAC*3/4 fSIG − fDAC/4 0.15 0.10 fSIG − fDAC/8 fSIG − fDAC*3/4 −0.4 −0.5 0.0 fSIG − fDAC*7/8 0.05 fSIG − fDAC*7/8 0.1 0.2 0.3 fSIG/fDAC 0.4 0.5 G030 0.00 0.0 fSIG − fDAC/2 0.1 0.2 0.3 fSIG/fDAC 0.4 0.5 G031 Figure 69. Frequency of Clock Mixing Spurious Images in X4 and X4L Modes Amplitude of the Spurious Signals The spurious signal amplitude is sensitive to factors such as temperature, voltage, and process. Typical worst case estimates to account for the variation over these factors are provided below as design guidelines. 59 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Figure 70 and Figure 71 show the typical worst case spurious signal amplitudes vs fDAC for a signal frequency fSIG = 11 x fDAC/32 in each mode for PLL on (PLL clock mode) and PLL off (external and dual clock modes). Each spurious signal (fDAC/2, fDAC/4 and fDAC/8) has its own curve. The spurious signal amplitudes can then be adjusted for the exact signal frequency fSIG by applying the amplitude adjustment factor shown in Figure 72. The amplitude adjustment factor is the same for each spurious signal (fDAC/2, fDAC/4, and fDAC/8) and is normalize for fSIG = 11 x fDAC/32. (b) X4L Mode 100 80 90 80 70 Spurious Amplitude − dBc Spurious Amplitude − dBc (a) X2 Mode 90 fDAC/2 60 50 40 30 20 70 fDAC/4 60 fDAC/2 50 40 30 20 10 10 0 0 0 100 200 300 400 fDAC − MHz 500 0 100 G032 400 100 100 90 90 80 80 70 fDAC/2 60 fDAC/4 fDAC x 3/4 40 30 G033 fDAC/4 fDAC x 3/4 70 60 50 fDAC/2 fDAC/8 fDAC x 7/8 40 30 20 20 10 10 0 0 0 100 200 300 fDAC − MHz 400 500 G034 0 100 200 300 400 fDAC − MHz Figure 70. Clock Related Spurious Signal Amplitude With PLL Off for fSIG = 11 x fDAC / 32 60 500 (d) X8 Mode Spurious Amplitude − dBc Spurious Amplitude − dBc 300 fDAC − MHz (c) X4 Mode 50 200 500 G035 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 (b) X4L Mode 60 50 50 Spurious Amplitude − dBc Spurious Amplitude − dBc (a) X2 Mode 60 40 fDAC/2 30 20 10 fDAC/2 40 30 fDAC/4 fDAC x 3/4 20 10 0 0 0 100 200 300 400 fDAC − MHz 500 0 100 300 400 fDAC − MHz G036 (c) X4 Mode 500 G037 (d) X8 Mode 70 70 60 60 Spurious Amplitude − dBc Spurious Amplitude − dBc 200 50 fDAC/2 40 30 fDAC/4 20 10 fDAC/8 fDAC x 7/8 fDAC/2 50 40 fDAC/4 fDAC x 3/4 30 20 10 0 0 0 100 200 300 fDAC − MHz 400 500 G038 0 100 200 300 400 fDAC − MHz 500 G039 Figure 71. Clock Related Spurious Signal Amplitude With PLL On for fSIG = 11 x fDAC / 32 61 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 40 Amplitude Adjustment − dB 30 20 10 0 −10 −20 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 fSIG/fDAC G040 Figure 72. Amplitude Adjustment Factor for fSIG The steps for calculating the non-harmonic spurious signals are: 1. Find the spurious signal frequencies for the appropriate mode from Figure 67, Figure 68, or Figure 69. 2. Find the amplitude for each spurious frequency for the appropriate mode from Figure 70 or Figure 71. 3. Adjust the amplitude of the spurious signals for fSIG using the adjustment factor in Figure 72. Consider Example 1 with the following conditions: 1. X4 Mode 2. Pll off 3. Complex Output 4. fDAC = 500 MHz 5. fSIG = 160 MHz = 0.32 x fDAC First, the location of the spurious signals is found for the x4 complex output in Figure 68(a). Three spurious signals are present in the range -0.5 x fDAC to 0.5 x fDAC: two from fDAC/4 (35 MHz and -215 MHz) and one from fDAC/2 (-90 MHz). Consulting Figure 70©), the raw amplitudes for fDAC/2 and fDAC/4 are 60 and 58 dBc, respectively. From Figure 72, the amplitude adjustment factor for fSIG = 0.32 x fDAC is estimated at ~ 1 dB and so the fDAC/2 and fDAC/4 are adjusted to 61 and 59 dBc. Table 15. Example # 1 for Calculating Spurious Signals Spurious Signal Frequency/fDAC Frequency (MHz) Raw Amplitude (dBc) Adjusted Amplitude (dBc) fDAC/4 0.7 35 58 59 fDAC/2 -0.18 -90 60 61 fDAC/4 -0.43 -215 58 59 Now consider Example 2 with the following conditions: 1. X2 Mode 2. PLL on 3. Real output 4. fDAC = 400 MHz 5. fSIG = 70 MHz = 0.175 x fDAC 62 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 First, the location of the spurious signals is found for the X2 real output in Figure 67(b). One spurious signals is present in the range 0 to 0.5 × fDAC at 0.325 × fDAC (see Table 16). Consulting Figure 71(a), the raw amplitude for fDAC/2 is 47 dBc. From Figure 72, the amplitude adjustment factor for fSIG = 0.175 x fDAC is estimated at ~ 6 dB, and so the fDAC/2 and fDAC/4 is adjusted to 53 dBc. Table 16. Example # 2 for Calculating Spurious Signals Spurious Signal Frequency/fDAC Frequency (MHz) Raw Amplitude (dBc) Adjusted Amplitude (dBc) fDAC/2 0.325 130 47 53 Schematic and Layout Recommendations The DAC5687 clock is sensitive to fast transitions of input data on pins DA0, DA1, and DA2 (55, 54, and 53) due to coupling to DVDD pin 56. The noise-like spectral energy of the DA[0-2] couples into the DAC clock resulting in increased jitter. This significantly improves by using a 10-Ω resistor between DVDD and pin 56 in addition to 10-pF capacitor to DGND, as implemented on the DAC5687EVM (see the DAC5687 EVM user's guide SLWU017). Pin 56 draws only approximately 2 mA of current and the 0.02-V voltage drop across the resistor is acceptable for DVDD voltages within the MINIMUM and MAXIMUM specifications. It is also recommended that the transition rate of the input lines be slowed by inserting series resistors near the data source. The optimized value of the series resistor depends on the capacitance of the trace between the series resistor and DAC5687 input pin. For a 2-3 inch trace, a 22-Ω to 47-Ω resistor would be recommended. The effect of DAC clock jitter on the DAC output signal is worse for signals at higher signal frequencies. For low IF (< 75 MHz) or baseband signals, there is little degradation of the output signal. However, for high IF (> 75 MHz) the DAC clock jitter may result in an elevated noise floor, which often appears as broad humps in the DAC output spectrum. It is recommended for signals above 75 MHz that the inputs to DA0 and DA1, which are the two LSBs if input DA[0-15] is not reversed, not be connected to input data to prevent coupling to the DAC rate clock. The decrease in resolution to 14-bits and increase in quantization noise will not significantly affect the DAC5687 SNR for signals > 75 MHz. Application Examples Application Example: Real IF Radio An system example of the DAC5687 used for a flexible real IF radio is shown in Figure 73. A complex baseband input to the DAC would be generated by a digital upconverter such as Texas Instruments GC4116, GC5016, or GC5316. The DAC5687 would be used to increase the data rate through interpolation and flexibly place the output signal using the FMIX and/or CMIX blocks. Although the DAC5687 X4 mode is shown, any of the modes (x2, x4L, or x8) would be appropriate. 63 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 DAC5687 DUC I 2 2 NCO RF Processing DAC DUC TRF3750 Q 2 GC4116 GC5016 GC5316 2 CDC7005 B0040-01 Figure 73. System Diagram of a Real IF System Using the DAC5687 With the DAC5687 in external clock mode, a low phase noise clock for the DAC5687 at the DAC sample rate would be generated by a VCXO and PLL such as Texas Instruments CDC7005, which can also provide other system clocks at the VCXO frequency divided by 2-n (n = 0 to 4). In this mode, the DAC5687 PLLLOCK pin output would typically be used to clock the digital upconverter. With the DAC in PLL clock mode, the same input rate clock would be used for the DAC clock and digital upconverter and the DAC internal PLL/VCO would generate the DAC sample rate clock. Note that the internal PLL/VCO phase noise may degrade the quality of the DAC output signal, and will also have higher non-harmonic clock-related spurious signals (see the Non-Harmonic Clock Related Spurious Signals section). Either DACA or DACB outputs can be used (with the other DAC put into sleep mode) and would typically be terminated with a transformer (see the Analog Current Output section). An IF filter, either LC or SAW, is used to suppress the DAC Nyquist zone images and other spurious signals before being mixed to RF with a mixer. An alternative architecture uses the DAC5687 in a dual-channel mode to create a dual-channel system with real IF input and output. This would be used for narrower signal bandwidth and at the expense of less output frequency placement flexibility (see Figure 74). Frequency upconversion can be accomplished by using the high-pass filter and CMIX fDAC/2 mixing features. 64 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 DAC5687 RF Processing DUC Ch2 DAC 2 DUC 2 1, −1, 1, ... 1, −1, 1, ... RF Processing DUC Ch1 DAC 2 DUC 2 GC4116 GC5016 GC5316 TRF3750 CDC7005 B0041-01 Figure 74. System Diagram of a Dual Channel Real IF Radio The outputs of multiple DAC5687s can be phase synchronized for multiple antenna/beamforming applications. Application Example: Complex IF to RF Conversion Radio An alternative to a real IF system is to use a complex IF DAC output with analog quadrature modulator, as shown in Figure 75. The same complex baseband input as the real IF system in Figure 73 is used. The DAC5687 would be used to increase the data rate through interpolation and flexibly place the output signal using the FMIX and/or CMIX blocks. Although the DAC5687 X4 mode is shown, any of the modes (x2, x4L, or x8) would be appropriate. TRF3701 TRF3702 DAC5687 I DAC 2 2 NCO RF Processing CMIX Q DAC 2 2 CDC7005 TRF3750 B0042-01 Figure 75. Complex IF System Using the DAC5687 in X4L Mode Instead of only using one DAC5687 output as for the real IF output, both DAC5687 outputs are used for a complex IF Hilbert transform pair. The DAC5687 outputs can be expressed as: 65 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 A(t) = I(t)cos(ωct) - Q(t)sin(ωct) = m(t) B(t) = I(t)sin(ωct) + Q(t)cos(ωct) = mh(t) where m(t) and mh(t) connote a Hilbert transform pair and ωc is the sum of the NCO and CMIX frequencies. The complex DAC5687 output is input to an analog quadrature modulator (AQM) such as the TRF3701 or TRF3702. A passive (resistor only) interface is recommended between the DAC5687 and TRF3701/2 (See the Passive Interface to TRF3701/2 section). Upper single-sideband up-conversion is achieved at the output of the analog quadrature modulator, whose output is expressed as: RF(t) = I(t)cos(ωc + ωLO)t - Q(t)sin(ωc + ωLO)t Flexibility is provided to the user by allowing for the selection of -B(t) out, which results in lower-sideband up-conversion. This option is selected by usb in the CONFIG3 register. Note that the process of complex mixing in FMIX and CMIX to translate the signal frequency from 0 Hz means that the analog quadrature modulator IQ imbalance produces a side-band and LO feedthrough that falls outside the signal. This is shown in Figure 76, which is the RF analog quadrature modulator (AQM) output of an asymmetric three carrier WCDMA signal with the properties in Table 17. The wanted signal is offset from the LO frequency by the DAC5687 complex IF, in this case 122.88 MHz. The nearest spurious signals are ~ 100 MHz away from the wanted signal (due to non-harmonic clock-related spurious signals generated by the fDAC/4 digital clock), providing 200 MHz of spurious free bandwidth. The AQM phase and gain imbalance produce a lower sideband product, which does not affect the quality of the wanted signal. Unlike the real IF architecture, the non-harmonic clock-related spurious signals generated by the fDAC/2 digital clock fall ±245.76-MHz offset from the wanted rather than falling inband. As a consequence, in the complex IF system it may be possible that no AQM phase, gain and offset correction is needed, instead relying on RF filtering to remove the LO feedthrough, sideband, and other spurious products. LO lower sideband 200 MHz C001 Figure 76. Analog Quadrature Modulator Output for a Complex IF System 66 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Table 17. Signal and System Properties for Complex IF System Example in Figure 76 Signal Three WCDMA Carriers, Test Model 1 Baseband Carrier Offsets -7.5 MHz, 2.5 MHz, 7.5 MHz DAC5687 Input Rate 122.88 MSPS DAC5687 Output Rate 491.52 MSPS (4x Interpolation) DAC5687 Mode X4 CMIX DAC5687 Complex IF 122.88 MHz (fDAC/4) LO Frequency 2140 MHz The complex IF has several advantages over the real IF architecture such as: 1. Uncalibrated side-band suppression ~ 35 dBc compared to 0 dBc for real IF architecture. 2. Direct DAC - Complex mixer connection - no amplifiers 3. Non-harmonic clock-related spurious signals fall out-of-band 4. DAC 2nd Nyquist zone image is offset fDAC compared with fDAC - 2 x IF for a real IF architecture, reducing the need for filtering at the DAC output. 5. Uncalibrated LO feed through for AQM is ~ 35 dBc and calibration can reduce or completely remove the LO feed through. Application Example: Wide Bandwidth Direct Baseband to RF Conversion A system example of the DAC5687 used in a wide bandwidth direct baseband to RF conversion is shown in Figure 77. The DAC input would typically be generated by a crest factor reduction processor such as Texas Instruments GC1115 and digital predistortion processor. With a complex baseband input, the DAC5687 would be used to increase the data rate through interpolation. In addition, phase, gain and offset correction of the IQ imbalance is possible using the QMC block, DAC gain and DAC offset features. The correction could be done one time during manufacturing (see the TRF3701 data sheet (SLWS145) and the TRF3702 data sheet (SLWS149)) for the variation with temperature, supply, LO frequency, etc. after calibration at nominal conditions) or during operation with a separate feedback loop measuring imbalance in the RF signal. TRF3701 TRF3702 DAC5687 I DAC 2 2 Phase/ Gain/ Offset Adjust GC1115 and DPD Processor RF Processing Q DAC 2 2 TRF3750 B0043-01 Figure 77. Direct Conversion System Using DAC5687 in X4L Mode Operating at baseband has the advantage that the DAC5687 output is insensitive to DAC sample clock phase noise, so using the DAC PLL clock mode will have similar spectral performance to the External clock mode. In addition, the non-harmonic clock-related spurious signals will be small due to the low DAC output frequency. With a complex input rate specified up to 250 MSPS, the DAC5687 is capable of producing signals with up to 200-MHz bandwidth for systems such as digital predistortion (DPD). 67 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Application Example: CMTS/VOD Transmitter The DAC5687's exceptional SNR enables a dual-cable modem termination system (CMTS) or video on demand (VOD) QAM transmitter in excess of the stringent DOCSIS specification, with > 74 dBc and 75 dBc in the adjacent and alternate channels. A typical system using the DAC5687 for a cost optimized dual channel two QAM transmitter is shown in Figure 78. A GC5016 would take four separate symbol rate inputs and provide pulse shaping and interpolation to ~ 128 MSPS. The four QAM carriers would be combined into two groups of two QAM carriers with intermediate frequencies of approximately 30 MHz to 40 MHz. The GC5016 would output two real data streams to one DAC5687. The DAC5687 would function as a dual-channel device and provide 2x interpolation to increase the frequency of the 2nd Nyquist zone image. The two signals are then output through the two DAC outputs, through a transformer and to an RF upconverter. DAC5687 QAM1 DUC Ch2 QAM2 DUC QAM3 DUC QAM4 DUC DAC 2 2 2 2 Ch1 DAC GC5016 CDC7005 B0044-01 Figure 78. Dual Channel Two QAM CMTS Transmitter System Using DAC5687 The DAC5687 output for a two QAM256 carrier signal at 33-MHz and 39-MHz IF with the signal and system properties listed in Table 18 is shown in Figure 70. The low DAC5687 noise floor provides better than 75 dBc (equal bandwidth normalized to one QAM256 power) at > 6-MHz offset. Table 18. Signal and System Properties for Complex IF System Example in Figure 79 Signal 68 QAM256, 5.36 MSPS, α = 0.12 IF Frequencies 33 MHz and 39 MHz DAC5687 Input Rate 5.36 MSPS x 24 = 128.64 MSPS DAC5687 Output Rate 257.28 MSPS (2x Interpolation) DAC5687 Mode X2 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 C002 Figure 79. Two QAM256 Carriers With at 36-MHz IF Application Example: High-Speed Arbitrary Waveform Generator The DAC5687's flexible input allows use of the dual input ports with demultiplexed odd/even samples at a combined rate of up to 500 MSPS. Combined with the DAC's 16-bit resolution, the DAC5687 allows wideband signal generation for test and measurement applications. DAC5687 Odd Digital Pattern Generator Input Multiplexer DAC Even B0045-01 Figure 80. DAC5687 in Odd/Even Input Mode 69 DAC5687 www.ti.com SLWS164B – FEBRUARY 2005 – REVISED JUNE 2005 Revision History DATE REV PAGE (1) 29 JUN 05 B 1 Ordering Information Added thermal pad dimensions Reversed "External Clock Mode" and "PLL Clock Mode" in Noise floor test SECTION DESCRIPTION 9 AC specifications 31 Register Name: ATEST Changed PLLLOCK Output Signal for PLLVDD = 0 to "Normal Operation" in Table 5 41 Clock Modes Reversed ts(DATA) and th(DATA) in Figures 43 and 44 42 Clock Modes Reversed ts(DATA) and th(DATA) in Figure 45 42 Clock Modes Updated Figure 46 26 MAR 04 A – – – 12 FEB 03 * – – Original version (1) 70 Page numbers for previous versions may differ from page numbers in the current version. IMPORTANT NOTICE Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications, enhancements, improvements, and other changes to its products and services at any time and to discontinue any product or service without notice. Customers should obtain the latest relevant information before placing orders and should verify that such information is current and complete. All products are sold subject to TI’s terms and conditions of sale supplied at the time of order acknowledgment. TI warrants performance of its hardware products to the specifications applicable at the time of sale in accordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TI deems necessary to support this warranty. Except where mandated by government requirements, testing of all parameters of each product is not necessarily performed. TI assumes no liability for applications assistance or customer product design. Customers are responsible for their products and applications using TI components. To minimize the risks associated with customer products and applications, customers should provide adequate design and operating safeguards. TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right, copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or process in which TI products or services are used. Information published by TI regarding third-party products or services does not constitute a license from TI to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property of the third party, or a license from TI under the patents or other intellectual property of TI. Reproduction of information in TI data books or data sheets is permissible only if reproduction is without alteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproduction of this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable for such altered documentation. Resale of TI products or services with statements different from or beyond the parameters stated by TI for that product or service voids all express and any implied warranties for the associated TI product or service and is an unfair and deceptive business practice. TI is not responsible or liable for any such statements. Following are URLs where you can obtain information on other Texas Instruments products and application solutions: Products Applications Amplifiers amplifier.ti.com Audio www.ti.com/audio Data Converters dataconverter.ti.com Automotive www.ti.com/automotive DSP dsp.ti.com Broadband www.ti.com/broadband Interface interface.ti.com Digital Control www.ti.com/digitalcontrol Logic logic.ti.com Military www.ti.com/military Power Mgmt power.ti.com Optical Networking www.ti.com/opticalnetwork Microcontrollers microcontroller.ti.com Security www.ti.com/security Telephony www.ti.com/telephony Video & Imaging www.ti.com/video Wireless www.ti.com/wireless Mailing Address: Texas Instruments Post Office Box 655303 Dallas, Texas 75265 Copyright 2005, Texas Instruments Incorporated