2.5 MSPS, 24-Bit, 100 dB Sigma-Delta ADC with On-Chip Buffer AD7760 FEATURES FUNCTIONAL BLOCK DIAGRAM VIN- VIN+ MULTIBIT - MODULATOR DIFF VREF+ AVDD1 AVDD2 AVDD3 RECONSTRUCTION BUF AVDD4 DECAPA/B RBIAS AD7760 MCLK PROGRAMMABLE DECIMATION RESET CS VDRIVE DVDD CONTROL LOGIC I/O OFFSET AND GAIN REGISTERS SYNC AGND DGND FIR FILTER ENGINE RD/WR DRDY DB0 TO DB15 04975-001 120 dB dynamic range at 78 kHz output data rate 100 dB dynamic range at 2.5 MHz output data rate 112 dB SNR at 78 kHz output data rate 100 dB SNR at 2.5 MHz output data rate 2.5 MHz maximum fully filtered output word rate Programmable oversampling rate (8x to 256x) Fully differential modulator input On-chip differential amplifier for signal buffering Low-pass finite impulse response (FIR) filter with default or user-programmable coefficients Modulator output mode Overrange alert bit Digital offset and gain correction registers Filter bypass modes Low power and power-down modes Synchronization of multiple devices via SYNC pin Figure 1. APPLICATIONS Data acquisition systems Vibration analysis Instrumentation GENERAL DESCRIPTION The AD7760 is a high performance, 24-bit - analog-to-digital converter (ADC). It combines wide input bandwidth and high speed with the benefits of - conversion to achieve a performance of 100 dB SNR at 2.5 MSPS, making it ideal for high speed data acquisition. Wide dynamic range combined with significantly reduced antialiasing requirements simplify the design process. An integrated buffer to drive the reference, a differential amplifier for signal buffering and level shifting, an overrange flag, internal gain and offset registers, and a low-pass digital FIR filter make the AD7760 a compact, highly integrated data acquisition device requiring minimal peripheral component selection. In addition, the device offers programmable decimation rates, and the digital FIR filter can be adjusted if the default characteristics are not appropriate for the application. The AD7760 is ideal for applications demanding high SNR without a complex front-end signal processing design. The differential input is sampled at up to 40 MSPS by an analog modulator. The modulator output is processed by a series of lowpass filters, with the final filter having default or user-programmable coefficients. The sample rate, filter corner frequencies, and output word rate are set by a combination of the external clock frequency and the configuration registers of the AD7760. The reference voltage supplied to the AD7760 determines the analog input range. With a 4 V reference, the analog input range is 3.2 V differential biased around a common mode of 2 V. This common-mode biasing can be achieved using the on-chip differential amplifier, further reducing the external signal conditioning requirements. The AD7760 is available in an exposed paddle, 64-lead TQFP and is specified over the industrial temperature range from -40C to +85C. Table 1. Related Devices Part No. AD7762 AD7763 Description 24-bit, 625 kSPS, 109 dB, - parallel interface 24-bit, 625 kSPS, 109 dB, - serial interface Rev. A Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.461.3113 (c)2006 Analog Devices, Inc. All rights reserved. AD7760 TABLE OF CONTENTS Features .............................................................................................. 1 Writing to the AD7760 .............................................................. 23 Applications....................................................................................... 1 Clocking the AD7760 .................................................................... 24 Functional Block Diagram .............................................................. 1 Buffering the MCLK signal....................................................... 24 General Description ......................................................................... 1 MCLK Jitter Requirements ....................................................... 24 Revision History ............................................................................... 3 Driving the AD7760....................................................................... 26 Specifications..................................................................................... 4 Using the AD7760 ...................................................................... 27 Timing Specifications .................................................................. 6 Decoupling and Layout Recommendations................................ 28 Timing Diagrams.......................................................................... 7 Supply Decoupling ..................................................................... 29 Absolute Maximum Ratings............................................................ 8 Additional Decoupling .............................................................. 29 ESD Caution.................................................................................. 8 Reference Voltage Filtering ....................................................... 29 Pin Configuration and Function Descriptions............................. 9 Differential Amplifier Components ........................................ 29 Terminology .................................................................................... 11 Bias Resistor Selection ............................................................... 29 Typical Performance Characteristics ........................................... 12 Layout Considerations............................................................... 29 Theory of Operation ...................................................................... 18 Exposed Paddle........................................................................... 29 Modulator Data Output Mode...................................................... 19 Programmable FIR Filter............................................................... 30 Modulator Inputs........................................................................ 19 Downloading a User-Defined Filter ............................................ 31 Modulator Data Output Scaling ............................................... 19 Example Filter Download ......................................................... 31 Modulator Data Output Mode Interface ..................................... 20 AD7760 Registers ........................................................................... 33 Clock Divide-by-1 Mode (CDIV = 1) ..................................... 20 Control Register 1--Address 0x0001 ...................................... 33 Clock Divide-by-2 Mode (CDIV = 0) ..................................... 20 Control Register 2--Address 0x0002 ...................................... 33 Using the AD7760 in Modulator Output Mode..................... 21 Status Register (Read Only) ...................................................... 34 AD7760 Interface............................................................................ 22 Offset Register--Address 0x0003............................................. 34 Reading Data............................................................................... 22 Gain Register--Address 0x0004............................................... 34 Reading Status and Other Registers......................................... 22 Overrange Register--Address 0x0005..................................... 34 Sharing the Parallel Bus ............................................................. 22 Outline Dimensions ....................................................................... 35 Synchronization.......................................................................... 22 Ordering Guide .......................................................................... 35 Rev. A | Page 2 of 36 AD7760 REVISION HISTORY 8/06--Rev. 0 to Rev. A Updated Package Option................................................... Universal Change to Features............................................................................1 Changes to Specifications.................................................................4 Changes to Absolute Maximum Ratings........................................8 Changes to Terminology Section ..................................................11 Added Figure 36 Through Figure 39 ............................................17 Added Modulator Data Output Mode Section ...........................19 Added Figure 41 Through Figure 47 ............................................19 Added Modulator Data Output Mode Interface Section...........20 Changes to Reading Data Section.................................................22 Added Synchronization Section....................................................22 Changes to Clocking the AD7760 Section...................................24 Added Buffering the MCLK Signal Section.................................24 Added MCLK Jitter Requirements Heading ...............................24 Changes to Driving the AD7760 Section.....................................26 Changes to Figure 51 ......................................................................26 Added Figure 52 ..............................................................................26 Changes to Figure 55 ......................................................................28 Changes to Figure 56 ......................................................................29 Added Exposed Paddle Section.....................................................29 Change to Control Register 2--Address 0x0002 Section ..........33 Changes to Status Register (Read Only) Section ........................34 7/05--Revision 0: Initial Version Rev. A | Page 3 of 36 AD7760 SPECIFICATIONS AVDD1 = DVDD = VDRIVE = 2.5 V, AVDD2 = AVDD3 = AVDD4 = 5 V, VREF+ = 4.096 V, MCLK amplitude = 5 V, TA = 25C, normal mode, using the on-chip amplifier with components as shown in Table 8, unless otherwise noted. 1 Table 2. Parameter DYNAMIC PERFORMANCE Decimate by 256 Dynamic Range Signal-to-Noise Ratio (SNR) 2 Spurious-Free Dynamic Range (SFDR) Total Harmonic Distortion (THD) Decimate by 32 Dynamic Range Signal-to-Noise Ratio (SNR)2 Spurious-Free Dynamic Range (SFDR) Total Harmonic Distortion (THD) Decimate by 8 Dynamic Range Signal-to-Noise Ratio (SNR)2 Spurious-Free Dynamic Range (SFDR) Total Harmonic Distortion (THD) IMD Second Order IMD Third Order DC ACCURACY Resolution Differential Nonlinearity Integral Nonlinearity Zero Error Gain Error Zero Error Drift Gain Error Drift DIGITAL FILTER RESPONSE Decimate by 8 Group Delay Decimate by 32 Group Delay Decimate by 256 Group Delay Test Conditions/Comments Specification Unit 119 120.5 112 59 126 77 -105 -106 -75 dB min dB typ dB typ dB typ dBc typ dBc typ dB typ dB typ dB typ 108 109.5 107 120 -105 -106 dB min dB typ dB typ dBc typ dB typ dB typ 99 100.5 100 99 98 120 114 -103 -102 -115 -89 dB min dB typ dB typ dB typ dB typ dBc typ dBc typ dB typ dB typ dB typ dB typ 24 Bits 0.00076 0.014 0.02 0.016 0.00001 0.0002 % typ % typ % max % typ % FS/C typ % FS/C typ MCLK = 40 MHz 12 s typ MCLK = 40 MHz 47 s typ MCLK = 40 MHz 358 s typ MCLK = 40 MHz, ODR = 78 kHz, fIN = 1 kHz Modulator inputs shorted Input amplitude = -0.5 dBFS Input amplitude = -60 dBFS Nonharmonic, input amplitude = -6 dBFS Input amplitude = -60 dBFS Input amplitude = -0.5 dBFS Input amplitude = -6 dBFS Input amplitude = -60 dBFS MCLK = 40 MHz, ODR = 625 kHz, fIN =100 kHz Modulator inputs shorted Input amplitude = -0.5 dBFS Nonharmonic, input amplitude = -6 dBFS Input amplitude = -0.5 dBFS Input amplitude = -6 dBFS MCLK = 40 MHz, ODR = 2.5 MHz Modulator inputs shorted fIN = 1 kHz, input amplitude = -0.5 dBFS fIN = 100 kHz, input amplitude = -0.5 dBFS fIN = 1 MHz, input amplitude = -0.5 dBFS Nonharmonic, fIN = 100 kHz, input amplitude = -6 dBFS Nonharmonic, fIN = 1 MHz, input amplitude = -6 dBFS Input amplitude = -0.5 dBFS, fIN = 100 kHz Input amplitude = -6 dBFS, fIN = 100 kHz fIN A = 989.95 kHz, fIN B = 999.95 kHz fIN A = 989.95 kHz, fIN B = 999.95 kHz Guaranteed monotonic to 24 bits Rev. A | Page 4 of 36 AD7760 Parameter ANALOG INPUT Differential Input Voltage Input Capacitance REFERENCE INPUT/OUTPUT VREF Input Voltage VREF Input DC Leakage Current VREF Input Capacitance POWER DISSIPATION Total Power Dissipation Standby Mode POWER REQUIREMENTS AVDD1 (Modulator Supply) AVDD2 (General Supply) AVDD3 (Differential Amplifier Supply) AVDD4 (Reference Buffer Supply) DVDD VDRIVE Normal Mode AIDD1 (Modulator) AIDD2 (General) 3 AIDD4 (Reference Buffer) Low Power Mode AIDD1 (Modulator) AIDD2 (General)3 AIDD4 (Reference Buffer) AIDD3 (Differential Amplifier) DIDD DIGITAL I/O MCLK Input Amplitude 4 Input Capacitance Input Leakage Current Three-State Leakage Current (D15:D0) VINH VINL VOH 5 VOH 6 VOL4 Test Conditions/Comments Specification Unit VIN(+) - VIN(-), VREF = 2.5 V VIN(+) - VIN(-), VREF = 4.096 V At internal buffer inputs At modulator inputs 2 3.25 5 55 V p-p V p-p pF typ pF typ VDD3 = 3.3 V 5% VDD3 = 5 V 5% +2.5 +4.096 6 5 V max V max A max pF max Normal mode Low power mode Clock stopped 958 661 6.35 mW max mW max mW max 5% 5% +2.5 +5 +3.15/+5.25 +3.15/+5.25 +2.5 +1.65/+2.7 V V V min/max V min/max V V min/max AVDD4 = 5 V 49/51 40/42 34/36 mA typ/max mA typ/max mA typ/max AVDD4 = 5 V AVDD3 = 5 V, both modes Both modes 26/28 20/23 9/10 41/44 63/70 mA typ/max mA typ/max mA typ/max mA typ/max mA typ/max 5 7.3 5 5 0.7 x VDRIVE 0.3 x VDRIVE 1.5 2.4 0.1 V typ pF typ A max A max V min V max V min V typ V max 5% 1 See the Terminology section. SNR specifications in decibels are referred to a full-scale input, FS. Tested with an input signal at 0.5 dB below full scale, unless otherwise specified. Current scales with ICLK frequency. See the Typical Performance Characteristics section. 4 Although the AD7760 can function with an MCLK amplitude of less than 5 V, this is the recommended amplitude to achieve the performance as stated. 5 Tested using the minimum VDRIVE voltage of 1.65 V with a 400 A load current. 6 Tested using VDRIVE = 2.5 V with a 400 A load current. 2 3 Rev. A | Page 5 of 36 AD7760 TIMING SPECIFICATIONS AVDD1 = DVDD = VDRIVE = 2.5 V, AVDD2 = AVDD3 = AVDD4 = 5 V, TA = 25C, normal mode, unless otherwise noted. Table 3. Parameter fMCLK fICLK t1 1, 2 t2 t3 t4 t5 t6 t7 t8 t92 t102 t11 t12 3, 4 t133, 4 t14 t15 t16 t17 t18 t194, 5 t204, 5 Limit at TMIN, TMAX 1 40 500 20 0.5 x tICLK 10 3 (0.5 x tICLK) + 16 ns tICLK tICLK 3 11 0.5 x tICLK 0.5 x tICLK (0.5 x tICLK) + 16 ns 23 19 11 4 x tICLK 4 x tICLK 5 0 23 19 Unit MHz min MHz max kHz min MHz max typ ns min ns min max min min ns min ns max typ typ max ns min ns min ns max min min ns min ns min ns min ns min Description Applied master clock frequency Internal modulator clock derived from MCLK DRDY pulse width DRDY falling edge to CS falling edge RD/WR setup time to CS falling edge Data access time CS low read pulse width CS high pulse width between reads RD/WR hold time to CS rising edge Bus relinquish time DRDY high period DRDY low period Data access time Data valid prior to DRDY rising edge Data valid after DRDY rising edge Bus relinquish time CS low write pulse width CS high period between address and data Data setup time Data hold time Data valid prior to MCLK falling edge while DRDY is logic low Data valid after MCLK falling edge while DRDY is logic low 1 tICLK = 1/fICLK. When ICLK = MCLK, DRDY pulse width depends on the mark-space ratio of applied MCLK. 3 Valid when using the modulator output mode with CDIV = 1. 4 See the Modulator Data Output Mode section for timing diagrams. 5 Valid when using the modulator output mode with CDIV = 0. 2 Rev. A | Page 6 of 36 AD7760 TIMING DIAGRAMS DRDY t5 t1 CS t6 t2 t7 t3 RD/WR DATA MSW 04975-002 t8 t4 D[0:15] LSW + STATUS Figure 2. Filtered Output--Parallel Interface Timing Diagram CS t15 t16 t17 D[0:15] t18 REGISTER ADDRESS Figure 3. AD7760 Register Write Rev. A | Page 7 of 36 REGISTER DATA 04975-004 RD/WR AD7760 ABSOLUTE MAXIMUM RATINGS TA = 25C, unless otherwise noted. Table 4. Parameters AVDD1 to GND AVDD2:AVDD4 to GND DVDD to GND VDRIVE to GND VIN+, VIN- to GND1 VINA+, VINA- to GND1 Digital Input Voltage to GND2 MCLK to MCLKGND VREF+ to GND3 AGND to DGND Input Current to Any Pin Except Supplies4 Operating Temperature Range Commercial Storage Temperature Range Junction Temperature TQFP Exposed Paddle Package JA Thermal Impedance JC Thermal Impedance Lead Temperature, Soldering Vapor Phase (60 sec) Infrared (15 sec) ESD Rating -0.3 V to +3 V -0.3 V to +6 V -0.3 V to +3 V -0.3 V to +3 V -0.3 V to +6 V -0.3 V to +6 V -0.3 V to DVDD + 0.3 V -0.3 V to +6 V -0.3 V to AVDD4 + 0.3 V -0.3 V to +0.3 V 10 mA Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those listed in the operational sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. ESD CAUTION -40C to +85C -65C to +150C 150C 92.7C/W 5.1C/W 215C 220C 600 V 1 Absolute maximum voltage for VIN-, VIN+ and VINA-, VINA+ is 6.0 V or AVDD3 + 0.3 V, whichever is lower. 2 Absolute maximum voltage on digital inputs is 3.0 V or DVDD + 0.3 V, whichever is lower. 3 Absolute maximum voltage on VREF+ input is 6.0 V or AVDD4 + 0.3 V, whichever is lower. 4 Transient currents of up to 200 mA do not cause SCR latch-up. Rev. A | Page 8 of 36 AD7760 64 63 62 61 60 59 58 DB11 DB10 DB9 DB8 DGND DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0 DGND VDRIVE DGND PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 57 56 55 54 53 52 51 50 49 DGND 1 48 DB12 47 DB13 MCLK 3 46 DB14 AVDD2 4 45 DB15 AGND2 5 44 VDRIVE AVDD1 6 43 DGND 42 DGND 41 DVDD 40 CS VREF+ 10 39 RD/WR AGND4 11 38 DRDY AVDD4 12 37 RESET AGND2 13 36 SYNC AVDD2 14 35 DGND AVDD2 15 34 AGND1 AGND2 16 33 AVDD1 PIN 1 MCLKGND 2 AGND1 7 AD7760 DECAPA 8 TOP VIEW (Not to Scale) REFGND 9 04975-005 AGND3 AGND3 DECAPB AGND3 AGND2 AVDD2 VIN- VIN+ AVDD3 AGND3 VOUTA+ VOUTA- VINA- VINA+ RBIAS AGND2 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 4. 64-Lead TQFP Pin Configuration Table 5. Pin Function Descriptions Pin No. 6, 33 Mnemonic AVDD1 4, 14, 15, 27 AVDD2 24 AVDD3 12 AVDD4 7, 34 5, 13, 16, 18, 28 23, 29, 31, 32 11 9 41 AGND1 AGND2 AGND3 AGND4 REFGND DVDD 44, 63 VDRIVE 1, 35, 42, 43, 53, 62, 64 19 20 21 22 25 26 10 DGND VINA+ VINA- VOUTA- VOUTA+ VIN+ VIN- VREF+ 8 DECAPA Description 2.5 V Power Supply for Modulator. These pins should be decoupled to AGND1 (Pin 7 and Pin 34, respectively) with 100 nF and 10 F capacitors on each pin. See the Decoupling and Layout Recommendations section for details. 5 V Power Supply. These pins should be decoupled to AGND2 (Pin 5 and Pin 13, with 100 nF capacitors on each of Pin 4, Pin 14, and Pin 15). Pin 27 should be connected to Pin 14 via a 15 nH inductor. See the Decoupling and Layout Recommendations section for details. 3.3 V to 5 V Power Supply for Differential Amplifier. This pin should be decoupled to AGND3 (Pin 23) with a 100 nF capacitor. See the Decoupling and Layout Recommendations section for details. 3.3 V to 5 V Power Supply for Reference Buffer. This pin should be decoupled to Pin 9 with a 10 nF capacitor in series with a 10 resistor. Power Supply Ground for Analog Circuitry Powered by AVDD1. Power Supply Ground for Analog Circuitry Powered by AVDD2. Power Supply Ground for Analog Circuitry Powered by AVDD3. Power Supply Ground for Analog Circuitry Powered by AVDD4. Reference Ground. Ground connection for the reference voltage. 2.5 V Power Supply for Digital Circuitry and FIR Filter. This pin should be decoupled to DGND with a 100 nF capacitor. Logic Power Supply Input, 1.8 V to 2.5 V. The voltage supplied at these pins determines the operating voltage of the logic interface. Both of these pins must be connected together and tied to the same supply. Each pin should also be decoupled to DGND with a 100 nF capacitor. Ground Reference for Digital Circuitry. Positive Input to Differential Amplifier. Negative Input to Differential Amplifier. Negative Output from Differential Amplifier. Positive Output from Differential Amplifier. Positive Input to the Modulator. Negative Input to the Modulator. Reference Input. The input range of this pin is determined by the reference buffer supply voltage (AVDD4). See the Reference Voltage Filtering section for more details. Decoupling Pin. A 100 nF capacitor must be inserted between this pin and AGND. Rev. A | Page 9 of 36 AD7760 Pin No. 30 17 Mnemonic DECAPB RBIAS 45 to 52, 54 to 61 DB15:DB8, DB7:DB0 37 RESET 3 MCLK 2 36 MCLKGND SYNC 39 RD/WR 38 DRDY 40 CS Description Decoupling Pin. A 33 pF capacitor must be inserted between this pin and AGND3. Bias Current Setting Pin. A resistor must be inserted between this pin and AGND. For more details, see the Bias Resistor Selection section. 16-Bit Bidirectional Data Bus. These are three-state pins that are controlled by the CS pin and the RD/WR pin. The operating voltage for these pins is determined by the VDRIVE voltage. See the Modulator Data Output Mode and AD7760 Interface sections for more details. A falling edge on this pin resets all internal digital circuitry and powers down the part. Holding this pin low keeps the AD7760 in a reset state. Master Clock Input. A low jitter, buffered digital clock must be applied to this pin. The output data rate depends on the frequency of this clock. See the Clocking the AD7760 section for more details. Master Clock Ground Sensing Pin. Synchronization Input. A falling edge on this pin resets the internal filter. This can be used to synchronize multiple devices in a system. See the Synchronization section for more details. Read/Write Input. This pin, in conjunction with the chip select pin, is used to read and write data to and from the AD7760. If this pin is low when CS is low, a read takes place. If this pin is high when CS is low, a write occurs. See the Modulator Data Output Mode and AD7760 Interface sections for more details. Data Ready Output. Each time new conversion data is available, an active low pulse, 1/2 ICLK period wide, is produced on this pin. See the Modulator Data Output Mode and AD7760 Interface sections for more details. Chip Select Input. Used in conjunction with the RD/WR pin to read and write data from and to the AD7760. See the Modulator Data Output Mode and AD7760 Interface sections for more details. Rev. A | Page 10 of 36 AD7760 TERMINOLOGY Signal-to-Noise Ratio (SNR) SNR is the ratio of the rms value of the actual input signal to the rms sum of all other spectral components below the Nyquist frequency, excluding harmonics and dc. The value for SNR is expressed in decibels. Total Harmonic Distortion (THD) THD is the ratio of the rms sum of harmonics to the fundamental. For the AD7760, it is defined as THD (dB ) = 20 log V 22 + V 32 + V 42 + V52 + V62 V1 Integral Nonlinearity (INL) INL is the maximum deviation from a straight line passing through the endpoints of the ADC transfer function. Differential Nonlinearity (DNL) DNL is the difference between the measured and the ideal 1 LSB change between any two adjacent codes in the ADC. Zero Error Zero error is the difference between the ideal midscale input voltage (when both inputs are shorted together) and the actual voltage producing the midscale output code. where: V1 is the rms amplitude of the fundamental. V2, V3, V4, V5, and V6 are the rms amplitudes of the second to the sixth harmonics. Zero Error Drift Zero error drift is the change in the actual zero error value due to a temperature change of 1C. It is expressed as a percentage of full scale at room temperature. Nonharmonic Spurious-Free Dynamic Range (SFDR) SFDR is the ratio of the rms signal amplitude to the rms value of the peak spurious spectral component, excluding harmonics. Gain Error The first transition (from 100 ... 000 to 100 ... 001) should occur for an analog voltage 1/2 LSB above the nominal negative full scale. The last transition (from 011 ... 110 to 011 ... 111) should occur for an analog voltage 11/2 LSB below the nominal full scale. The gain error is the deviation of the difference between the actual level of the last transition and the actual level of the first transition, from the difference between the ideal levels. Dynamic Range Dynamic range is the ratio of the rms value of the full scale to the rms noise measured with the inputs shorted together. The value for the dynamic range is expressed in decibels. Intermodulation Distortion With inputs consisting of sine waves at two frequencies, fa and fb, any active device with nonlinearities creates distortion products at sum and difference frequencies of mfa nfb, where m, n = 0, 1, 2, 3, and so on. Intermodulation distortion terms are those for which neither m nor n are equal to 0. For example, the secondorder terms include (fa + fb) and (fa - fb), and the third-order terms include (2fa + fb), (2fa - fb), (fa + 2fb), and (fa - 2fb). Gain Error Drift Gain error drift is the change in the actual gain error value due to a temperature change of 1C. It is expressed as a percentage of full scale at room temperature. The AD7760 is tested using the CCIF standard, where two input frequencies near the top end of the input bandwidth are used. In this case, the second-order terms are usually distanced in frequency from the original sine waves, and the third-order terms are usually at a frequency close to the input frequencies. As a result, the second- and third-order terms are specified separately. The calculation of the intermodulation distortion is as per the THD specification, where it is the ratio of the rms sum of the individual distortion products to the rms amplitude of the sum of the fundamentals expressed in decibels. Rev. A | Page 11 of 36 AD7760 TYPICAL PERFORMANCE CHARACTERISTICS 0 -25 -25 -50 -50 -75 -100 -125 -75 -100 -125 -150 -150 -175 -175 -200 0 4000 8000 12000 16000 20000 -200 24000 04975-009 AMPLITUDE (dB) 0 04975-006 AMPLITUDE (dB) AVDD1 = DVDD = VDRIVE = 2.5 V, AVDD2 = AVDD3 = AVDD4 = 5 V, VREF = 4.096 V, TA = 25C, normal mode, unless otherwise noted. All FFTs are generated from 65,536 samples using a 7-term Blackman-Harris window. 0 4000 Figure 5. Normal Mode FFT, 1 kHz, -0.5 dB Input Tone, 256x Decimation 24000 -50 AMPLITUDE (dB) -75 -100 -125 -75 -100 -125 -150 -150 -175 04975-007 -175 0 4000 8000 12000 16000 20000 04975-010 AMPLITUDE (dB) 20000 -25 -50 -200 0 24000 4000 -25 -25 -50 -50 AMPLITUDE (dB) 0 -75 -100 -125 -125 -175 04975-008 -175 16000 24000 -100 -150 12000 20000 -75 -150 8000 16000 Figure 9. Low Power FFT, 1 kHz, -6 dB Input Tone, 256x Decimation 0 4000 12000 20000 04975-011 Figure 6. Normal Mode FFT, 1 kHz, -6 dB Input Tone, 256x Decimation 0 8000 FREQUENCY (Hz) FREQUENCY (Hz) AMPLITUDE (dB) 16000 0 -25 -200 12000 Figure 8. Low Power FFT, 1 kHz, -0.5 dB Input Tone, 256x Decimation 0 -200 8000 FREQUENCY (Hz) FREQUENCY (Hz) -200 24000 0 FREQUENCY (Hz) 4000 8000 12000 16000 20000 24000 FREQUENCY (Hz) Figure 7. Normal Mode FFT, 1 kHz, -60 dB Input Tone, 256x Decimation Figure 10. Low Power FFT, 1 kHz, -60 dB Input Tone, 256x Decimation Rev. A | Page 12 of 36 0 -25 -25 -50 -50 -75 -100 -125 -75 -100 -125 -150 -150 -175 -175 -200 0 250 500 750 1000 04975-015 AMPLITUDE (dB) 0 04975-012 AMPLITUDE (dB) AD7760 -200 1250 0 250 FREQUENCY (kHz) 0 -25 -25 -50 -50 -75 -100 -125 -125 -175 -175 1000 -200 1250 0 500 750 1000 1250 Figure 15. Low Power FFT, 100 kHz, -6 dB Input Tone, 8x Decimation 0 -25 -25 -50 -50 AMPLITUDE (dB) 0 -75 -100 -125 -150 -75 -100 -125 -150 04975-014 0 250 500 750 1000 -175 -200 1250 FREQUENCY (kHz) 04975-017 -175 -200 250 FREQUENCY (kHz) Figure 12. Normal Mode FFT, 100 kHz, -6 dB Input Tone, 8x Decimation AMPLITUDE (dB) -100 -150 500 750 FREQUENCY (kHz) 1250 -75 -150 250 1000 04975-016 AMPLITUDE (dB) 0 0 750 Figure 14. Low Power FFT, 100 kHz, -0.5 dB Input Tone, 8x Decimation 04975-013 AMPLITUDE (dB) Figure 11. Normal Mode FFT, 100 kHz, -0.5 dB Input Tone, 8x Decimation -200 500 FREQUENCY (kHz) 0 250 500 750 1000 1250 FREQUENCY (kHz) Figure 13. Normal Mode FFT, 1 MHz, -0.5 dB Input Tone, 8x Decimation Figure 16. Low Power FFT, 1 MHz, -0.5 dB Input Tone, 8x Decimation Rev. A | Page 13 of 36 0 -25 -25 -50 -50 -75 -100 -125 -75 -100 -125 -150 -150 -175 -175 -200 0 250 500 750 1000 -200 1250 04975-021 AMPLITUDE (dB) 0 04975-018 AMPLITUDE (dB) AD7760 0 250 FREQUENCY (kHz) Figure 17. Normal Mode FFT, 1 MHz, -6 dB Input Tone, 8x Decimation -75 -100 -125 -75 -100 -125 -150 -150 -175 -175 0 250 500 750 1000 -200 1250 04975-022 AMPLITUDE (dB) -50 04975-019 AMPLITUDE (dB) 1250 TONE A: 999.75kHz TONE B: 1.00025MHz -25 -50 0 250 FREQUENCY (kHz) 500 750 1000 1250 FREQUENCY (kHz) Figure 18. Normal Mode IMD, 1 MHz Center Frequency, 8x Decimation Figure 21. Low Power IMD, 1 MHz Center Frequency, 8x Decimation 0 0 TONE A: 999.75kHz TONE B: 1.00025MHz SECOND-ORDER IMD: -105.6dB -20 TONE A: 999.75kHz TONE B: 1.00025MHz SECOND-ORDER IMD: -115.7dB -20 -40 -60 -80 -100 -60 -80 -100 -120 -140 -140 04975-020 -120 0 2000 4000 6000 8000 -160 10000 FREQUENCY (kHz) 04975-023 AMPLITUDE (dB) -40 AMPLITUDE (dB) 1000 0 TONE A: 999.75kHz TONE B: 1.00025MHz -25 -160 750 Figure 20. Low Power FFT, 1 MHz, -6 dB Input Tone, 8x Decimation 0 -200 500 FREQUENCY (kHz) 0 2000 4000 6000 8000 10000 FREQUENCY (kHz) Figure 19. Normal Mode IMD, 1 MHz Center Frequency, 8x Decimation Figure 22. Low Power IMD, 1 MHz Center Frequency, 8x Decimation Rev. A | Page 14 of 36 AD7760 0 0 -80 -100 -60 -80 -100 -120 -120 -140 -140 997 999 1001 1003 -160 995 1005 04975-027 AMPLITUDE (dB) -40 -60 -160 995 TONE A: 999.75kHz TONE B: 1.00025MHz THIRD-ORDER IMD: -87.67dB -20 04975-024 AMPLITUDE (dB) TONE A: 999.75kHz TONE B: 1.00025MHz -20 THIRD-ORDER IMD: -89.15dB -40 997 999 FREQUENCY (kHz) 1001 1003 1005 FREQUENCY (kHz) Figure 23. Normal Mode IMD, 1 MHz Center Frequency, 8x Decimation Figure 26. Normal Mode IMD, 1 MHz Center Frequency, 8x Decimation 100.5 -105 NORMAL MODE 100.0 -107 NORMAL MODE -109 99.0 THD (dBc) SNR (dBFS) 99.5 98.5 98.0 -111 -113 LOW POWER MODE LOW POWER MODE 97.5 04975-025 96.5 0 10 20 30 -117 40 04975-028 -115 97.0 0 10 MCLK FREQUENCY (MHz) 20 30 Figure 24. SNR vs. MCLK Frequency, 8x Decimation, -6 dB, 1 kHz Input Tone Figure 27. THD vs. MCLK Frequency, 8x Decimation, -6 dB, 1 kHz Input Tone 120 116 -60dB -60dB -6dB 115 112 -6dB SNR (dBFS) 105 104 100 04975-026 100 -0.5dB 108 0 64 128 192 96 256 DECIMATION RATE 04975-029 SNR (dBFS) -0.5dB 110 95 40 MCLK FREQUENCY (MHz) 0 64 128 192 256 DECIMATION RATE Figure 25. Normal Mode SNR vs. Decimation Rate, 1 kHz Input Tone Figure 28. Low Power SNR vs. Decimation Rate, 1 kHz Input Tone Rev. A | Page 15 of 36 AD7760 4500 3000 4000 2500 3000 OCCURRENCE OCCURRENCE 3500 2500 2000 1500 2000 1500 1000 1000 04975-030 0 8385222 8385238 8385254 0 8383530 83835246 8383562 8385270 24-BIT CODE 04975-032 500 500 8383578 8383594 8383610 24-BIT CODE Figure 29. Normal Mode, 24-Bit Histogram, 256x Decimation Figure 32. Low Power, 24-Bit Histogram, 256x Decimation 600 450 400 500 400 OCCURRENCE OCCURRENCE 350 300 200 300 250 200 150 100 100 8385116 8385216 8385316 8385416 04975-033 04975-031 0 8385016 50 0 8383236 8385516 8383386 24-BIT CODE 8383536 8383686 8383836 24-BIT CODE Figure 30. Normal Mode, 24-Bit Histogram, 8x Decimation Figure 33. Low Power, 24-Bit Histogram, 8x Decimation 0.0010 0.0015 +85C +85C 0.0010 0.0005 +25C INL (%) INL (%) 0.0005 0 +25C 0 -40C -0.0005 -40C 0 4194304 8388608 12582912 16777216 -0.0010 24-BIT CODE 04975-036 -0.0010 04975-034 -0.0005 0 4194304 8388608 12582912 24-BIT CODE Figure 31. 24-Bit INL, Normal Mode Figure 34. 24-Bit INL, Low Power Mode Rev. A | Page 16 of 36 16777216 AD7760 0.6 50 45 0.4 40 35 DIDD (mA) 0 -0.2 25 20 15 10 04975-035 -0.4 -0.6 30 0 4194304 8388608 12582912 04975-059 DNL (LSB) 0.2 5 0 16777216 0 2 4 24-BIT CODE 6 8 10 12 14 ICLK FREQUENCY (MHz) 16 18 20 Figure 38. Decimate x 32, DIDD vs. ICLK Frequency (DVDD = 2.5 V) Figure 35. 24-Bit DNL 45 40 40 35 35 30 DIDD (mA) 20 15 20 10 5 5 0 2 4 6 8 10 12 14 ICLK FREQUENCY (MHz) 16 18 0 20 60 50 40 30 20 04975-058 10 0 2 4 6 8 10 12 14 ICLK FREQUENCY (MHz) 16 0 2 4 6 8 10 12 14 ICLK FREQUENCY (MHz) 16 18 20 Figure 39. Decimate x 256, DIDD vs. ICLK Frequency (DVDD = 2.5 V) Figure 36. AIDD2 vs. ICLK Frequency (AVDD2 = 5 V) 0 04975-060 10 0 DIDD (mA) 25 15 04975-057 AIDD2 (mA) 30 25 18 20 Figure 37. Decimate x 8, DIDD vs. ICLK Frequency (DVDD = 2.5 V) Rev. A | Page 17 of 36 AD7760 THEORY OF OPERATION The AD7760 employs a - conversion technique to convert the analog input into an equivalent digital word. The modulator samples the input waveform and outputs an equivalent digital word to the digital filter at a rate equal to ICLK. The third filter has a fixed decimation rate of 2x, is user programmable, and has a default configuration. It is described in detail in the Programmable FIR Filter section. This filter can also be bypassed. By employing oversampling, the quantization noise is spread across a wide bandwidth from 0 to fICLK. This means that the noise energy contained in the signal band of interest is reduced (see Figure 40a). To further reduce the quantization noise in the signal band of interest, a high order modulator is employed to shape the noise spectrum so that most of the noise energy is shifted out of the signal band (see Figure 40b). Table 6 shows some characteristics of the default filter. The group delay of the filter is defined to be the delay to the center of the impulse response and is equal to the computation plus the filter delays. The delay until valid data is available (the DVALID status bit is set) is equal to twice the filter delay plus the computation delay. a. The digital filtering that follows the modulator removes the large out-of-band quantization noise (see Figure 40c) while also reducing the data rate from fICLK at the input of the filter to fICLK/8 or less at the output of the filter, depending on the decimation rate used. QUANTIZATION NOISE fICLK\2 BAND OF INTEREST b. Digital filtering has certain advantages over analog filtering: It does not introduce significant noise or distortion and can be made perfectly linear in terms of phase. NOISE SHAPING The AD7760 employs three FIR filters in series. By using different combinations of decimation ratios, filter selection, and bypassing, data can be obtained from the AD7760 at a large range of data rates. Multibit data from the modulator can be obtained at the ICLK rate (see Modulator Data Output Mode section). The first filter receives the data from the modulator at a maximum frequency of 20 MHz and decimates it by 4 to output the data at 5 MHz. The partially filtered data can be output at this stage. The second filter allows the decimation rate to be chosen from 2x to 32x or to be completely bypassed. fICLK\2 BAND OF INTEREST c. fICLK\2 BAND OF INTEREST 04975-037 DIGITAL FILTER CUTOFF FREQUENCY Figure 40. - ADC Table 6. Configuration with Default Filter ICLK Frequency 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 20 MHz 12.288 MHz 12.288 MHz 12.288 MHz 12.288 MHz Filter 1 Bypassed 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x 4x Filter 2 Bypassed Bypassed Bypassed 2x 2x 4x 4x 8x 8x 16x 16x 32x 32x 8x 16x 32x 32x Filter 3 Bypassed Bypassed 2x Bypassed 2x Bypassed 2x Bypassed 2x Bypassed 2x Bypassed 2x 2x 2x Bypassed 2x Data State Unfiltered Partially filtered Fully filtered Partially filtered Fully filtered Partially filtered Fully filtered Partially filtered Fully filtered Partially filtered Fully filtered Partially filtered Fully filtered Fully filtered Fully filtered Partially filtered Fully filtered Computation Delay 0 0.325 s 1.075 s 1.35 s 1.625 s 1.725 s 1.775 s 2.6 s 2.25 s 4.175 s 3.1 s 7.325 s 4.65 s 3.66 s 5.05 s 11.92 s 7.57 s Rev. A | Page 18 of 36 Filter Delay 0 1.2 s 10.8 s 3.6 s 22.8 s 6 s 44.4 s 10.8 s 87.6 s 20.4 s 174 s 39.6 s 346.8 s 142.6 s 283.2 s 64.45 s 564.5 s Pass-Band Bandwidth 10 MHz 1.35 MHz 1 MHz 562.5 kHz 500 kHz 281.25 kHz 250 kHz 140.625 kHz 125 kHz 70.3125 kHz 62.5 kHz 35.156 kHz 31.25 kHz 76.8 kHz 38.4 kHz 21.6 kHz 19.2 kHz Output Data Rate (ODR) 20 MHz 5 MHz 2.5 MHz 2.5 MHz 1.25 MHz 1.25 MHz 625 kHz 625 kHz 312.5 kHz 312.5 kHz 156.25 kHz 156.25 kHz 78.125 kHz 192 kHz 96 kHz 96 kHz 48 kHz AD7760 MODULATOR DATA OUTPUT MODE Operating the AD7760 in modulator output mode enables the output of data directly from the - modulator. This mode of operation bypasses the AD7760 on-board digital filtering capabilities, outputting data in its unfiltered form. scaled to 15 bits. The transfer function in Figure 42 shows the scaling involved for the 16 data bits output from Modulator Pins D[15:0] vs. the maximum differential voltage input allowed for the modulator inputs (VIN+ and VIN-). D[15:0] 0011 1111 1111 1111 0011 0011 0011 0010 +4.096V +3.275V = MODULATOR FULL SCALE = 80% OF +4.096V VIN+ = 2.048V VIN- = 2.048V 0000 0000 0000 0000 1100 1100 1100 1100 80% OF +4.096V = MODULATOR FULL SCALE = -3.275V MODULATOR INPUTS The maximum voltage input to each differential modulator input pin is 0.8 x 4.096 V 3.275 V (80% of VREF), which must sit on a common mode of VREF/2. This maximum differential input voltage is shown as the conditioned output of the AD7760's on-board differential amplifier in Figure 52 in the Driving the AD7760 section. Further details on the signal conditioning implemented by the AD7760's on-board differential amplifier and the recommended external circuitry that accompanies it is described in the Driving the AD7760 section. -20 -30 -40 AMPLITUDE (dB) VIN+ = 0.4105V VIN- = 3.6855V -4.096V Figure 42. Modulator Output Data Scaling As the nature of the modulator output is coarse relative to the fully filtered output of the AD7760 (due to the associated quantization noise of the modulator output), Bits D[3:0] of the modulator output are zero when operating in modulator data output mode. Thus, the data outputs for the calculations listed in Example 1 and Example 2 for inputs to the modulator pins VIN+ and VIN- show Bits D[3:0] of the modulator output as zero. Example 1 VIN+ = 3.5 V VIN- = 0.595 V Modulator Output Code = ([VIN(+) - VIN(-)]/4.096 V) x 16384 = [(3.5 V - 0.595 V)/4.096 V] x 16384 = +11620 Direct Scaling: [0010 1101 0110 0100] Value Output on Data Output Pins D[15:0]: D [15:0] = [0010 1101 0110 0000]. 0 -10 -50 -60 -70 -80 -90 -100 -110 Example 2 -120 0 1 2 3 4 5 6 7 8 9 10 FREQUENCY (MHz) 04975-048 -130 -140 -150 -160 1100 0000 0000 0000 VIN+ = 3.6855V VIN- = 0.4105V 04975-049 As discussed in the Theory of Operation section, the AD7760 operates using oversampling, which spreads quantization noise over a wide bandwidth. The decrease in the quantization noise energy in the resulting signal band is illustrated in Figure 40a. By coupling the use of oversampling with the use of a high order, multibit - modulator, the AD7760 further reduces the quantization noise in the signal band. Figure 41 is an FFT of unfiltered data output from the AD7760 when it is used in modulator output mode. This clearly demonstrates the shaping of the quantization noise performed by the AD7760's - modulator. Figure 41. FFT of Data Output by the AD7760 in Modulator Output Mode MODULATOR DATA OUTPUT SCALING VIN+ = 0.595 V VIN- = 3.5 V Modulator Output Code = ([VIN(+) - VIN(-)]/4.096 V) x 16384 = [(0.595 V - 3.5 V)/4.096 V] x 16384 = -11620 Direct Scaling: [1101 0010 1001 1100] Value Output on Data Output Pins D[15:0]: D [15:0] = [1101 0010 1001 0000]. In modulator output mode, data is output in a 16-bit twos complement format on Pins D [15:0]; however, this data is Rev. A | Page 19 of 36 AD7760 MODULATOR DATA OUTPUT MODE INTERFACE output mode when operating with CDIV = 1 (that is, ICLK = MCLK). A DRDY pulse is generated for each word. The data on each of the 16 data output pins, D [15:0], is valid on the rising edge of the DRDY pulse. The DRDY pulse can be used to latch the modulator data into a FIFO or as a DMA control signal. Shortly after the RD/WR and CS lines return high, the AD7760 stops outputting data and the data bus returns to high impedance. The AD7760 can be configured in modulator data output mode (bypassing the default decimation filtering) by writing 0 to each of the bits contained in Control Register 1: BYP F1, BYP F3, and DEC [2:0]. This will bypass all digital decimation filtering offered by the AD7760. See the AD7760 Registers section for further details. When the AD7760 is operating in modulator data output mode, a different parallel interfacing scheme than that used for configurations, where the AD7760's data output is filtered is necessary. CLOCK DIVIDE-BY-2 MODE (CDIV = 0) When operating in modulator output mode with CDIV = 0 (that is, ICLK = MCLK/2), the frequency of the DRDY signal created is half that of the MCLK frequency input to the device. The timing scheme that is used when CDIV = 0 depends on the number of MCLK cycles that occur between RESET and SYNC. If the number of MCLK cycles (n) between the rising edge of RESET and the rising edge of SYNC (see Figure 44) is an even value, use the interface timing shown in Figure 43. If n is an odd value, use the interface timing shown in Figure 45. The data output rate depends on the clock divider ratio that is used. When the CDIV bit in Control Register 2 is set to logic high, data is output at the MCLK frequency. If the CDIV bit is set to logic low, data is output at a frequency of MCLK/2. See the Clocking the AD7760 section. CLOCK DIVIDE-BY-1 MODE (CDIV = 1) When obtaining data from the AD7760 in modulator output mode, both the RD/WR and CS lines must be held low. This brings the data bus out of its high impedance state. Figure 43 shows the timing diagram for reading data in the modulator data t9 t10 DRDY t13 D[0:15] t14 t12 t11 INVALID DATA MOD DATA M MOD DATA M + 1 MOD D... 04975-050 CS, RD/WR Figure 43. AD7760 Modulator Output Mode (CDIV = 1) and (CDIV = 0, n is even) RESET MCLK 04975-051 n x tMCLK SYNC Figure 44. AD7760 Relative Timing Between RESET and SYNC in Modulator Output Mode CDIV = 0 Rev. A | Page 20 of 36 AD7760 t9 t10 DRDY t20 D[0:15] INVALID DATA MOD DATA M MOD DATA M + 1 MOD D... t19 MCLK t14 04975-052 t11 CS, RD/WR Figure 45. AD7760 Modulator Output Mode (CDIV = 0, n is odd) In the case where n is an odd number of MCLK cycles, the modulator data output on Pins D [15:0] is output on the rising edge of DRDY. In this case, the modulator data should be read on the falling edge of MCLK when DRDY is logic low. Figure 45 shows timing details to be used when reading the modulator output data where CDIV = 0 and there is an odd number of MCLK cycles between the rising edge of RESET and the rising edge of SYNC. The edge of MCLK that should be used under these conditions is illustrated in Figure 45 by arrows on the MCLK falling edges in question. USING THE AD7760 IN MODULATOR OUTPUT MODE The following is the recommended sequence for powering up and using the AD7760: 1. Apply power. 2. Start the clock oscillator, applying MCLK. 3. Take RESET low for a minimum of one MCLK cycle. 4. Wait a minimum of two MCLK cycles after the rising edge of RESET. 5. Write to Control Register 2 to power up the ADC and the differential amplifier as required. The correct clock divider (CDIV) ratio should be programmed at this time. 6. Write to Control Register 1 to set the bypass filter bits, BYP F1 and BYP F3, and the decimation rate bits, DEC [2:0], to 0. 7. Wait a minimum of six MCLK cycles after the rising edge of CS has been released. 8. Take SYNC low for a minimum of four MCLK cycles, if required, to synchronize multiple parts. Using this sequence results in an even number of MCLK cycles between the rising edge of RESET and the rising edge of SYNC. Therefore, when using this sequence with CDIV = 0, the interface timing shown in Figure 43 should be implemented. Note that whether the number of MCLK cycles between the rising edge of RESET and SYNC is odd or even is irrelevant when the AD7760 is operated with CDIV = 1. When using the AD7760 in modulator output mode, the offset, gain, and overrange registers are not operational. The only registers that can be used are Control Register 1 and Control Register 2. Rev. A | Page 21 of 36 AD7760 AD7760 INTERFACE When the AD7760 is outputting data at a 5 MHz output data rate or less, the interface operates in a conventional mode, as shown in Figure 2, using a 16-bit bidirectional parallel interface. This interface is controlled by the RD/WR and CS pins. The 24-bit conversion data is output in twos complement format. When a new conversion result is available, an active low pulse is output on the DRDY pin. To read a conversion result from the AD7760, two 16-bit read operations are performed. The DRDY pulse indicates that a new conversion result is available. Both RD/WR and CS go low to perform the first read operation. Shortly after both lines go low, the data bus becomes active and the 16 most significant bits (MSBs) of the conversion result are output. The RD/WR and CS lines must return high for a full ICLK period before the second read is performed. This second read contains the eight least significant bits (LSBs) of the conversion result along with six status bits. These status bits are shown in Table 7. Descriptions of the other status bits are found in Table 17. Table 7. Status Bits During Data Read MSB D7 DVALID D6 OVR D5 UFILT D4 LPWR D3 FILTOK D2 DLOK D1 0 LSB D0 0 Shortly after RD/WR and CS return high, the data bus returns to a high impedance state. Both read operations must be completed before a new conversion result is available because the new result overwrites the contents of the output register. If a DRDY pulse occurs during a read operation, the data read is invalid. If multiple synchronized AD7760 parts that share a properly distributed common MCLK signal exist in a system, these parts can share a common bus without being isolated from each other. This bus can then be isolated from the system bus by a single latch or buffer. SYNCHRONIZATION The SYNC input to the AD7760 provides a synchronization function that allows the user to begin gathering samples of the analog front-end input from a known point in time. The SYNC function allows multiple AD7760s, operated from the same MCLK, RESET, and SYNC signals, to be synchronized so that each ADC simultaneously updates its output register. The distribution of the signals that are common to each of the devices that are to be synchronized is extremely important in ensuring that the timing of each of the AD7760 devices is correct, that is, that each AD7760 device sees the same digital edges synchronously. The SYNC signal is sensed on the falling edge of MCLK. On the first falling edge of MCLK after SYNC goes logic low, the digital filter sequencer is reset to 0. The filter is held in a reset state until a falling edge of the MCLK senses SYNC logic high. The SYNC signal must remain logic low for a minimum of four MCLK cycles. Figure 46 shows the recommended timing for the SYNC signal with respect to MCLK. DEVICE SYNCHRONIZED FROM THIS POINT IN TIME MCLK SYNC READING STATUS AND OTHER REGISTERS The AD7760 features a number of programmable registers. To read back the contents of these registers or the status register, the user must first write to the control register of the device, setting a bit that corresponds to the register to be read. The next read operation outputs the contents of the selected register instead of a conversion result. The AD7760 Registers section provides more information on the relevant bits in the control register. SHARING THE PARALLEL BUS By its nature, the high accuracy of the AD7760 makes it sensitive to external noise sources. These include digital activity on the parallel bus. For this reason, it is recommended that the AD7760 data lines be isolated from the system data bus by means of a latch or buffer to ensure all digital activity on the D0 to D15 pins is controlled by the AD7760. MIN SYNC LOGIC LOW 4 x tMCLK 04975-053 READING DATA Figure 46. Recommended SYNC Timing The rising edge of SYNC should be coincident with the rising edge of MCLK. Thus, the next falling edge of MCLK senses SYNC logic high and takes the filter out of its reset state. By applying this signal scheme to multiple ADCs using the same MCLK and SYNC signals, all of the devices will gather input samples synchronously. Following a SYNC signal, the digital filter needs time to settle before valid data can be read from the AD7760. The DVALID status bit (D7 in Table 7) output with each conversion indicates when valid data is being output by the converter. The time from the rising edge of SYNC until the DVALID bit is asserted is dependent on the filter configuration used. See the Theory of Operation section and the values listed in Table 6 for details on calculating the time until DVALID is asserted. Rev. A | Page 22 of 36 AD7760 WRITING TO THE AD7760 There are many features and parameters that the user can change by writing to the AD7760 device. See the Using the AD7760 section, which details the writing sequence needed to initialize the operation of the part. The AD7760 has programmable registers that are 16 bits wide. This means that two write operations are required to program a register. The first write contains the register address, and the second write contains the register data. An exception is when a user-defined filter is being downloaded to the AD7760. This is described in detail in the Downloading a User-Defined Filter section. The AD7760 Registers section contains the register addresses and details. Figure 3 shows a write operation to the AD7760. The RD/WR line is held high while the CS line is brought low for a minimum of four ICLK periods. The register address is latched during this period. The CS line is brought high again for a minimum of four ICLK periods before the register data is put onto the data bus. If a read operation occurs between the writing of the register address and the register data, the register address is cleared and the next write must be the register address. This also provides a method to revert back to a known situation if the user forgets whether the next write is an address or data. Generally, the AD7760 is written to and configured on powerup and very infrequently, if at all, after that. Following any write operation, the full group delay of the filter must elapse before valid data is output from the AD7760. Rev. A | Page 23 of 36 AD7760 CLOCKING THE AD7760 The AD7760 requires an external low jitter clock source. This signal is applied to the MCLK pin, and the MCLKGND pin is used to sense the ground from the clock source. An internal clock signal (ICLK) is derived from the MCLK input signal. The ICLK controls all internal operations of the AD7760. The maximum ICLK frequency is 20 MHz, but due to an internal clock divider, a range of MCLK frequencies can be used. There are two ways to generate the ICLK: ICLK = MCLK (CDIV = 1) ICLK = MCLK/2 (CDIV = 0) These options are selected from the control register (see the AD7760 Registers section for more details). On power-up, the default is ICLK = MCLK/2 to ensure that the part can handle the maximum MCLK frequency of 40 MHz. For output data rates equal to those used in audio systems, a 12.288 MHz ICLK frequency can be used. As shown in Table 6, output data rates of 192 kHz, 96 kHz, and 48 kHz are achievable with this ICLK frequency. As mentioned previously, this ICLK frequency can be derived from different MCLK frequencies. It is recommended that the MCLK signal applied to the AD7760 has a 50-50 mark-space ratio. When operating in clock divide-by-1 mode (that is, CDIV = 1), using higher mark-space ratios reduces the maximum MCLK frequency that can be applied to the AD7760 yielding maximum performance. For example, using a mark-space ratio of 60-40 (with CDIV = 1) reduces the maximum MCLK frequency that will yield the maximum INL and THD performance to 16 MHz. BUFFERING THE MCLK SIGNAL The MCLK signal for the AD7760 must be buffered before being input to the MCLK pin on the AD7760 device. This can be done simply by routing the MCLK signal to both inputs of an AND gate (see Figure 47). The recommended buffer is the NC7SZ08M5, which is a twoinput AND gate from Fairchild Semiconductor. Using the buffer with a supply voltage of 5 V is advised to achieve optimum performance from the AD7760. AD7760 3 NC7SZ08M5 (AND GATE) The MCLK jitter requirements depend on a number of factors and are given by t j(rms ) = 2 x x f IN OSR SNR (dB) x 10 20 where: OSR = oversampling ratio = fICLK/ODR. fIN = maximum input frequency. SNR(dB) = target SNR. Example 1 This example can be taken from Table 6, where: ODR = 2.5 MHz. fICLK = 20 MHz. fIN (max) = 1 MHz. SNR = 108 dB. t j (rms) = 8 = 1.79 ps 2 x x 10 6 x 10 5.4 This is the maximum allowable clock jitter for a full-scale, 1 MHz input tone with the given ICLK and output data rate. Example 2 Take a second example from Table 6, where: ODR = 48 kHz. fICLK = 12.288 MHz. fIN (max) = 19.2 kHz. SNR = 120 dB. t j(rms ) = 256 = 133 ps 2 x x 19.2 x 103 x 106 The input amplitude also has an effect on these jitter figures. For example, if the input level was 3 dB below full-scale, the allowable jitter would be increased by a factor of 2, increasing the first example to 2.53 ps rms. This happens when the maximum slew rate is decreased by a reduction in amplitude. Figure 48 and Figure 49 illustrate this point, showing the maximum slew rate of a sine wave of the same frequency but with different amplitudes. MCLK 04975-054 MCLK SOURCE MCLK JITTER REQUIREMENTS Figure 47. Buffering the MCLK Signal Using the NC7SZ08M5 AND Gate Rev. A | Page 24 of 36 1.0 0.5 0.5 0 0 -0.5 -0.5 -1.0 04975-038 1.0 -1.0 Figure 48. Maximum Slew Rate of Sine Wave with Amplitude of 2 V p-p 04975-039 AD7760 Figure 49. Maximum Slew Rate of Sine Wave (with the Same Frequency as in Figure 48) with Amplitude of 1 V p-p Rev. A | Page 25 of 36 AD7760 DRIVING THE AD7760 The AD7760 has an on-chip differential amplifier that operates with a supply voltage (AVDD3) within the 3.15 V to 5.25 V range. For a 4.096 V reference, the supply voltage must be 5 V. To achieve the specified performance in normal mode, the differential amplifier should be configured as a first-order antialias filter, as shown in Figure 50. Any additional filtering should be carried out in previous stages using low noise, high performance op amps, such as the AD8021. Suitable component values for the first-order filter are listed in Table 8. Using the values in the table as an example yields a 10 dB attenuation at the first alias point of 19 MHz. CFB signal to sit on the optimum common mode of VREF/2, in this case 2.048 V. The signal is also scaled to give the maximum allowable voltage swing with this reference value. This is calculated as 80% of VREF, that is, 0.8 x 4.096 V 3.275 V p-p on each input. With a 4.096 V reference, a 5 V supply must be provided to the reference buffer (AVDD4). With a 2.5 V reference, a 3.3 V supply must be provided to AVDD4. Figure 51 shows the transfer function in terms of the 24-bit digital output codes (twos complement coding) of the AD7760 vs. the voltage signals VA and VB applied to the on-board differential amplifier A1, as shown in Figure 52. 24 BITS RFB 011...111 011...110 RM VIN- A1 B RM RIN 000...010 VIN+ AD7760 24-BIT OUTPUT RFB 04975-040 111...110 Figure 50. Differential Amplifier Configuration 100...001 100...000 Table 8. Normal Mode Component Values RIN 1 k RFB 655 000...000 111...111 CFB VREF 4.096 V 000...001 RM 18 CS 5.6 pF CFB 33 pF Figure 52 shows the signal conditioning that occurs using the circuit shown in Figure 50 with a 2.5 V input signal biased around ground and the component values and conditions listed in Table 8. The differential amplifier always biases the output B = +2.5V A = -2.5V A = 0V B = 0V Figure 51. Transfer Function for the AD7760 Filtered Output Where VA and VB are Inputs to the On-Board Differential Amplifier A1 +2.5V +3.685V 0V +2.048V VIN+ A -2.5V +0.410V +2.5V +3.685V B 0V +2.048V -2.5V +0.410V INPUTS TO THE AD7760 DIFFERENTIAL AMPLIFIER A = +2.5V B = -2.5V 04975-056 CS VIN- OUTPUTS OF THE AD7760 DIFFERENTIAL AMPLIFIER Figure 52. Differential Amplifier Signal Conditioning Rev. A | Page 26 of 36 04975-055 A RIN AD7760 To obtain maximum performance from the AD7760, it is advisable to drive the ADC with differential signals. Figure 53 shows how a bipolar, single-ended signal biased around ground can drive the AD7760 with the use of an external op amp, such as the AD8021. USING THE AD7760 The following is the recommended sequence for powering up and using the AD7760: 1. Apply power. CFB 2. Start the clock oscillator, applying MCLK. RFB 3. Take RESET low for a minimum of one MCLK cycle. 4. Wait a minimum of two MCLK cycles after RESET has been released. 5. Write to Control Register 2 to power up the ADC and the differential amplifier as required. The correct clock divider (CDIV) ratio should be programmed at this time. Figure 53. Single-Ended-to-Differential Conversion 6. Write to Control Register 1 to set the output data rate. The AD7760 employs a double-sampling front end, as shown in Figure 54. For simplicity, only the equivalent input circuit for VIN+ is shown. The equivalent input circuitry for VIN- is the same. 7. Wait a minimum of five MCLK cycles after CS has been released. 8. Take SYNC low for a minimum of four MCLK cycles, if required, to synchronize multiple parts. 2R 2R RIN AD8021 RM VIN- CS A1 R RIN RM VIN+ RFB 04975-042 VIN CFB VIN+ CS1 SS1 SH3 CPA SH1 CPB1 Data can then be read from the part using the default filter, offset, gain, and overrange threshold values. The conversion data read is not valid, however, until the group delay of the filter has elapsed. Once this has occurred, the DVALID bit read with the data LSW is set, indicating that the data is indeed valid. SS3 ANALOG MODULATOR CS2 SS2 The user can then download a different filter if required (see the Downloading a User-Defined Filter section). Values for gain, offset, and overrange threshold registers can be written or read at this stage. SH4 CPB2 SS4 04975-043 SH2 Figure 54. Equivalent Input Circuit Sampling Switches SS1 and SS3 are driven by ICLK, whereas Sampling Switches SS2 and SS4 are driven by ICLK. When ICLK is high, the analog input voltage is connected to CS1. On the falling edge of ICLK, the SS1 and SS3 switches open and the analog input is sampled on CS1. Similarly, when ICLK is low, the analog input voltage is connected to CS2. On the rising edge of ICLK, the SS2 and SS4 switches open and the analog input is sampled on CS2. Capacitors CPA, CPB1, and CPB2 represent parasitic capacitances that include the junction capacitances associated with the MOS switches. Table 9. Equivalent Component Values Mode Normal Low Power CS1 (pF) 51 13 CS2 (pF) 51 13 CPA (pF) 12 12 CPB1/2 (pF) 20 5 Rev. A | Page 27 of 36 AD7760 DECOUPLING AND LAYOUT RECOMMENDATIONS 41 VDRIVE VDRIVE DVDD PIN 41 44 63 AVDD3 AVDD2 PIN 44 PIN 63 24 27 AVDD1 AVDD1 PIN 24 PIN 27 6 33 PIN 6 PIN 33 12 AVDD4 PIN 12 4 14 15 19 VINA+ 20 VINA- 21 VOUTA- 22 VOUTA+ U2 8 DECAPA 30 DECAPB C7 100nF 25 VIN+ 26 VIN- VIN+ VIN- C64 33pF AD7760BSV 10 VREF+ 9 REFGND AVDD2 DGND DGND DGND DGND DGND DGND DGND AVDD4 L1 PIN 4 (RHS) C48 100nF L3 PIN 15 (VBIAS) C50 100nF PIN 12 (VBUF) L2 PIN 14 (LHS) C62 100nF L9 PIN 27 AVDD3 RESET SYNC DRDY 3 2 MCLK MCLKGND 11 AGND4 AGND3 AGND3 AGND3 AGND3 23 29 31 32 AVDD1 CS RD/WR 37 36 38 RESET SYNC DRDY AGND2 AGND2 AGND2 AGND2 AGND2 R19 160k 1 35 42 43 53 62 64 DB0 DB1 DB2 DB3 DB4 DB5 DB6 DB7 DB8 DB9 DB10 DB11 DB12 DB13 DB14 DB15 40 39 CS RD/WR RBIAS 5 13 16 18 28 17 7 AGND1 34 AGND1 VREFX DB [0:15] 61 60 59 58 57 56 55 54 52 51 50 49 48 47 46 45 DB0 DB1 DB2 DB3 DB4 DB5 DB6 DB7 DB8 DB9 DB10 DB11 DB12 DB13 DB14 DB15 MCLK VDRIVE DVDD L4 L5 R38 10 C59 10nF PIN 5 (VMOD1) C52 100nF L11 PIN 33 (VMOD2) L6 PIN 24 (VDIF1) C53 100nF Figure 55. Simplified Connection Diagram Rev. A | Page 28 of 36 C54 100nF L7 PIN 44 (VDRV1) C56 100nF L8 L12 PIN 63 (VDRV2) C57 100nF PIN 41 (DVDD) C58 100nF 04975-046 VINA+ VINA- VOUTA- VOUTA+ AVDD2 AVDD2 AVDD2 PIN 4 PIN 14 PIN 15 Due to the high performance nature of the AD7760, correct decoupling and layout techniques are required to obtain the performance as stated within this data sheet. Figure 55 shows a simplified connection diagram for the AD7760. AD7760 SUPPLY DECOUPLING BIAS RESISTOR SELECTION Every supply pin must be connected to the appropriate supply via a ferrite bead and decoupled to the correct ground pin with a 100 nF, 0603 case size, X7R dielectric capacitor. There are two exceptions to this: The AD7760 requires a resistor to be connected between the RBIAS and AGND pins. The value of this resistor is dependent on the reference voltage being applied to the device. The resistor value should be selected to produce a current of 25 A through the resistor to ground. For a 2.5 V reference voltage, the correct resistor value is 100 k, and for a 4.096 V reference, the correct resistor value is 160 k. * Pin 12 (AVDD4) must have a 10 resistor inserted between the pin and a 10 nF decoupling capacitor, which is connected to ground at Pin 9. * Pin 27 (AVDD2) does not require a separate decoupling capacitor or a direct connection to the supply, but instead is connected to Pin 14 via a 15 nH inductor. LAYOUT CONSIDERATIONS While using the correct components is essential to achieve optimum performance, the correct layout is just as important. The AD7760 product page on the Analog Devices website contains the Gerber files for the AD7760 evaluation board. These files should be used as a reference when designing any system using the AD7760. ADDITIONAL DECOUPLING There are two other decoupling pins on the AD7760--Pin 8 (DECAPA) and Pin 30 (DECAPB). Pin 8 should be decoupled with a 100 nF capacitor, and Pin 30 requires a 33 pF capacitor. The location and orientation of some of the components mentioned in previous sections of this data sheet are critical, and particular attention must be paid to the components that are located close to the AD7760. Locating these components farther away from the device can have a direct impact on the achievable maximum performance. REFERENCE VOLTAGE FILTERING A low noise reference source, such as the ADR431 (2.5 V) or ADR434 (4.096 V), is suitable for use with the AD7760. The reference voltage supplied to the AD7760 should be decoupled and filtered as shown in Figure 56. The use of ground planes should also be carefully considered. To ensure that the return currents through the decoupling capacitors are flowing to the correct ground pin, the ground side of the capacitors should be as close as possible to the ground pin associated with that supply. A ground plane should not be relied on as the sole return path for decoupling capacitors because the return current path using ground planes is not easily predictable. The recommended scheme for the reference voltage supply is a 100 series resistor connected to a 100 F tantalum capacitor, followed by a series resistor of 10 , and finally a 10 nF capacitor placed as close as possible to the VREF+ pin, decoupling this capacitor to the associated ground pin, Pin 11. U3 2 7.5V C15 10F + C9 100nF VIN VOUT GND 4 6 R30 100 C11 100F R17 10 + C46 10nF EXPOSED PADDLE PIN 10 04975-047 ADR434 Figure 56. Reference Connection DIFFERENTIAL AMPLIFIER COMPONENTS The correct components for use around the on-chip differential amplifier are detailed in Table 8. Matching the components on both sides of the differential amplifier is important to minimize distortion of the signal applied to the amplifier. A tolerance of 0.1% or better is required for these components. Symmetrical routing of the tracks on both sides of the differential amplifier also assists in achieving the stated performance. The AD7760 64-lead TQFP employs a 6 mm x 6 mm exposed paddle (see Figure 59). The paddle reduces the thermal resistance of the package by providing a path for heat energy to flow between the package and the PCB and, in turn, increases the heat transfer efficiency from the AD7760 package. Connecting the exposed paddle to the AGND plane of the PCB is essential in creating the conditions that allow the AD7760 package to perform to the highest specifications possible. The exposed paddle should not be connected directly to any of the ground pins on the AD7760 and should only be connected to the analog ground plane. Best practice is to use multiple vias connecting the exposed paddle to the AGND plane of the PCB. Rev. A | Page 29 of 36 AD7760 PROGRAMMABLE FIR FILTER No. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Dec Value 53656736 25142688 -4497814 -11935847 -1313841 6976334 3268059 -3794610 -3747402 1509849 3428088 80255 -2672124 -1056628 1741563 1502200 -835960 -1528400 93626 1269502 411245 -864038 -664622 434489 Hex Value 332BCA0 17FA5A0 444A196 4B62067 4140C31 6A734E 31DDDB 439E6B2 4392E4A 1709D9 344EF8 1397F 428C5FC 4101F74 1A92FB 16EBF8 40CC178 4175250 16DBA 135EFE 6466D 40D2F26 40A242E 6A139 No. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Dec Value 700847 -70922 -583959 -175934 388667 294000 -183250 -302597 16034 238315 88266 -143205 -128919 51794 121875 16426 -90524 -63899 45234 114720 102357 52669 15559 1963 Hex Value AB1AF 401150A 408E917 402AF3E 5EE3B 47C70 402CBD2 4049E05 3EA2 3A2EB 158CA 4022F65 401F797 CA52 1DC13 402A 401619C 400F99B B0B2 1C020 18FD5 CDBD 3CC7 7AB The default filter should be sufficient for most applications. It is a standard brick wall filter with a symmetrical impulse response. The default filter has a length of 96, is nonaliasing, and provides 120 dB of attenuation at Nyquist. This filter not only performs signal antialiasing, but also suppresses out-of-band quantization noise produced by the analog-to-digital conversion process. Any significant relaxation in the stop-band attenuation or transition bandwidth relative to the default filter can result in a failure to meet the SNR specifications. * The filter must be an even, symmetrical FIR. * The coefficients are in sign-and-magnitude format, with 26 magnitude bits and sign coded as positive = 0. * The filter length must be between 12 taps and 96 taps in steps of 12. * Because the filter is symmetrical, the number of coefficients that must be downloaded is half the filter length. The default filter coefficients exemplify this with only 48 coefficients listed for a 96-tap filter. * Coefficients are written from the center of the impulse response (adjacent to the point of symmetry) outwards. * The coefficients are scaled so that the in-band gain of the filter is equal to 134,217,726, with the coefficients rounded to the nearest integer. For a low-pass filter, this is the equivalent of having the coefficients summed arithmetically (including sign) to a +67,108,863 (0x3FF FFFF) positive value over the half-impulse-response coefficient set (a maximum of 48 coefficients). Any deviation from this introduces a gain error. 0 PASS-BAND RIPPLE = 0.05dB -0.1dB FREQUENCY = 1.004MHz -3dB FREQUENCY = 1.06MHz STOP BAND = 1.25MHz -20 -40 -60 -80 -100 -120 -140 -160 0 500 1000 1500 2000 2500 FREQUENCY (kHz) 04975-044 Table 10. Default Filter Coefficients To create a filter, note the following: AMPLITUDE (dB) As previously mentioned, the third FIR filter on the AD7760 is user programmable. The default coefficients that are loaded upon reset are given in Table 10, and the frequency responses are shown in Figure 57. The frequencies quoted in Figure 57 scale directly with the output data rate. Figure 57. Default Filter Frequency Response (2.5 MHz ODR) The procedure for downloading a user-defined filter is detailed in the Downloading a User-Defined Filter section. The default filter characteristics scale with both the MCLK frequency applied and the decimation rate chosen by the user. Rev. A | Page 30 of 36 AD7760 DOWNLOADING A USER-DEFINED FILTER EXAMPLE FILTER DOWNLOAD The following is an example of downloading a short userdefined filter with 24 taps. The frequency response is shown in Figure 58. 10 0 -10 AMPLITUDE (dB) As previously mentioned, the filter coefficients are 27 bits in length--one sign and 26 magnitude bits. Because the AD7760 has a 16-bit parallel bus, the coefficients are padded with 5 MSB 0s to generate a 32-bit word, split into two 16-bit words for downloading. The first 16-bit word for each coefficient becomes (00000, sign bit, Magnitude [25:16]), whereas the second word becomes (Magnitude [15:0]). To ensure that a filter is downloaded correctly, a checksum must also be generated and then downloaded following the final coefficient. The checksum is a 16-bit word generated by splitting each 32-bit word into four bytes and summing the bytes from all coefficients up to a maximum of 192 bytes (48 coefficients x four bytes). The same checksum is generated internally in the AD7760 and compared with the downloaded checksum. The DL_OK bit in the status register is set if these two checksums agree. -20 -30 -40 -50 -60 -80 1. Write to Control Register 1, setting the DL_FILT bit and the correct filter length bits corresponding to the length of the filter to be downloaded (see Table 11). 0 100 200 300 400 500 FREQUENCY (kHz) 600 04975-045 -70 To download a user filter Figure 58. 24-Tap FIR Frequency Response 2. Write the first half of the current coefficient data (00000, sign bit, Magnitude [25:16]). The first coefficient to be written must be the one adjacent to the point of filter symmetry. The coefficients for the filter are listed in Table 12 and are shown from the center of symmetry outwards. The raw coefficients were generated using a commercial filter design tool and were scaled appropriately so that their sum equals 67,108,863 (0x3FF FFFF). 3. Write the second half of the current coefficient data (Magnitude [15:0]). Table 12. 24-Tap FIR Coefficients 4. Repeat Step 2 and Step 3 for each coefficient. 5. Write the 16-bit checksum. 6. Use the following methods to verify that the filter coefficients are downloaded correctly: a. Read the status register, checking the DL_OK bit. b. Read data and observe the status of the DL_OK bit. Note that because the user coefficients are stored in RAM, they are cleared after a RESET operation or a loss of power. Coefficient 1 2 3 4 5 6 7 8 9 10 11 12 Table 11. Filter Length Values FLEN [3:0] 0000 0001 0011 0101 0111 1001 1011 1101 1111 Number of Coefficients Default 6 12 18 24 30 36 42 48 Filter Length Default 12 24 36 48 60 72 84 96 Rev. A | Page 31 of 36 Raw 0.365481974 0.201339905 0.009636604 -0.075708848 -0.042856209 0.019944246 0.036437914 0.007592007 -0.021556583 -0.024888355 -0.012379538 -0.001905756 Scaled 53188232 29300796 1402406 -11017834 -6236822 2902466 5302774 1104856 -3137108 -3621978 -1801582 -277343 AD7760 Table 13 shows the hexadecimal values (in sign-and-magnitude format) that are downloaded to the AD7760 to realize this filter. The table is also split into the bytes that are summed to produce the checksum. The checksum generated from these coefficients is 0x0E6B. Table 14 lists the 16-bit words the user would write to the AD7760 to set up the ADC and download this filter, assuming an output data rate of 1.25 MHz has been selected. Table 13. Filter Hexadecimal Values Word 0x0001 0x8079 Coefficient 1 2 3 4 5 6 7 8 9 10 11 12 Word 1 Byte 1 Byte 2 03 2B 01 BF 00 15 04 A8 04 5F 00 2C 00 50 00 10 04 2F 04 37 04 1B 04 04 Word 2 Byte 3 Byte 4 96 88 18 3C 66 26 1E 6A 2A 96 49 C2 E9 F6 DB D8 DE 54 44 5A 7D 6E 3B 5F Table 14. Sequence of Write Instructions to Set Up Device and Download the Filter Example 0x032B 0x9688 0x01BF 0x183C ... 0x0404 0x3B5F 0x0E6B 0x0001 0x0879 Rev. A | Page 32 of 36 Description Address of Control Register 1. Control register data. DL filter: set filter length = 24, set output data rate = 1.25 MHz. First coefficient, Word 1. First coefficient, Word 2. Second coefficient, Word 1. Second coefficient, Word 2. Other coefficients. Twelfth (final) coefficient, Word 1. Final coefficient, Word 2. Checksum. Wait (0.5 x tICLK x number of unused coefficients) for AD7760 to write 0s to the remaining unused coefficients. Address of control register. Control register data. Set read status and maintain filter length and decimation settings. Read contents of status register. Check Bit 7 (DL_OK) to determine if the filter was downloaded correctly. AD7760 AD7760 REGISTERS The AD7760 has a number of user-programmable registers. The control registers are used to set the decimation rate, the filter configuration, the clock divider, and so on. There are also digital gain, offset, and overrange threshold registers. Writing to these registers involves writing the register address first, then a 16-bit data-word. Register addresses, details of individual bits, and default values are given in this section. CONTROL REGISTER 1--ADDRESS 0x0001 Default Value 0x001A MSB D15 D14 D13 D12 D11 D10 D9 D8 D7 D6 D5 D4 D3 D2 D1 LSB D0 DL_ FILT RD OVR RD GAIN RD OFF RD STAT 0 SYNC FLEN3 FLEN2 FLEN1 FLEN0 BYP F3 BYP F1 DEC2 DEC1 DEC0 Table 15. Bit Descriptions of Control Register 1 Bit 15 Mnemonic DL_FILT 1 14 RD OVR1, 2 13 12 11 10 9 RD GAIN1, 2 RD OFF1, 2 RD STAT1, 2 0 SYNC1 8 to 5 4 3 FLEN [3:0] BYP F3 BYP F1 2 to 0 DEC [2:0] 1 2 Description Download Filter. Before downloading a user-defined filter, this bit must be set. The filter length bits must also be set at this time. The write operations that follow are interpreted as the user coefficients for the FIR filter until all the coefficients and the checksum have been written. Read Overrange. If this bit has been set, the next read operation outputs the contents of the overrange threshold register instead of a conversion result. Read Gain. If this bit has been set, the next read operation outputs the contents of the digital gain register. Read Offset. If this bit has been set, the next read operation outputs the contents of the digital offset register. Read Status. If this bit has been set, the next read operation outputs the contents of the status register. 0 must be written to this bit. Synchronize. Setting this bit initiates an internal synchronization routine. Setting this bit simultaneously on multiple devices synchronizes all filters. Filter Length Bits. These bits must be set when the DL_FILT bit is set before a user-defined filter is downloaded. Bypass Filter 3. If this bit is 0, Filter 3 (programmable FIR) is bypassed. Bypass Filter 1. If this bit is 0, Filter 1 is bypassed. This should only occur when the user requires unfiltered modulator data to be output. Decimation Rate. These bits set the decimation rate of Filter 2. All 0s implies that the filter is bypassed. A value of 1 corresponds to 2x decimation, a value of 2 corresponds to 4x decimation, and so on, up to the maximum value of 5, corresponding to 32x decimation. Bit 15 to Bit 9 are self-clearing bits. Only one of the bits from Bit 14 to Bit 11 can be set in any write operation because it determines the contents of the next read operation. CONTROL REGISTER 2--ADDRESS 0x0002 Default Value After RESET: 0x009B Recommended register setting for power-up and normal operation using clock divide-by-2 (CDIV = 0) mode: 0x0002 MSB D15 D14 D13 D12 D11 D10 D9 D8 D7 D6 0 0 0 0 0 0 0 0 0 0 D5 CDIV D4 D3 D2 D1 LSB D0 0 PD LPWR 1 D1PD Table 16. Bit Descriptions of Control Register 2 Bit Mnemonic Description 5 CDIV 3 2 PD LPWR 1 0 1 D1PD Clock Divider Bit. This sets the divide ratio of the MCLK signal to produce the internal ICLK. Setting CDIV = 0 divides the MCLK by 2. If CDIV = 1, the ICLK frequency is equal to the MCLK. Power Down. Setting this bit powers down the AD7760, reducing the power consumption to 6.35 mW. Low Power. If this bit is set, the AD7760 is operating in a low power mode. The power consumption is reduced for a 6 dB reduction in noise performance. Write 1 to this bit. Differential Amplifier Power Down. Setting this bit powers down the on-chip differential amplifier. Rev. A | Page 33 of 36 AD7760 STATUS REGISTER (READ ONLY) MSB D15 D14 D13 D12 D11 D10 D9 D8 D7 D6 D5 PART 1 PART 0 DIE 2 DIE 1 DIE 0 0 LPWR OVR DL_OK FILTOK UFILT D4 BYP F3 D3 BYP F1 D2 D1 LSB D0 DEC2 DEC1 DEC0 Table 17. Bit Descriptions of Status Register Bit 15, 14 13 to 11 10 9 8 7 Mnemonic PART [1:0] DIE [2:0] 0 LPWR OVR DL_OK 6 FILTOK 5 4 3 2 to 0 UFILT BYP F3 BYP F1 DEC [2:0] Comment Part Number. These bits are constant for the AD7760. Die Number. These bits reflect the current AD7760 die number for identification purposes within a system. This bit is set to 0. Low Power. If the AD7760 is operating in low power mode, this bit is set to 1. If the current analog input exceeds the current overrange threshold, this bit is set. When downloading a user filter to the AD7760, a checksum is generated. This checksum is compared to the one downloaded following the coefficients. If these checksums agree, this bit is set. When a user-defined filter is in use, a checksum is generated when the filter coefficients pass through the filter. This generated checksum is compared to the one downloaded. If they match, this bit is set. If a user-defined filter is in use, this bit is set. Bypass Filter 3. If Filter 3 is bypassed by setting the relevant bit in Control Register 1, this bit is also set. Bypass Filter 1. If Filter 1 is bypassed by setting the relevant bit in Control Register 1, this bit is also set. Decimation Rate. These bits correspond to the bits set in Control Register 1. OFFSET REGISTER--ADDRESS 0x0003 OVERRANGE REGISTER--ADDRESS 0x0005 Non-bit-mapped, Default Value 0x0000 Non-bit-mapped, Default Value 0xCCCC The offset register uses twos complement notation and is scaled such that 0x7FFF (maximum positive value) and 0x8000 (maximum negative value) correspond to an offset of +0.78125% and -0.78125%, respectively. Offset correction is applied after any gain correction. Using the default gain value of 1.25 and assuming a reference voltage of 4.096 V, the offset correction range is approximately 25 mV. The overrange register value is compared with the output of the first decimation filter to obtain an overload indication with minimum propagation delay. This is prior to any gain scaling or offset adjustment. The default value is 0xCCCC, which corresponds to 80% of VREF (the maximum permitted analog input voltage). Assuming VREF = 4.096 V, the bit is then set when the input voltage exceeds approximately 6.55 V p-p differential. Once the overrange bit is set, the DVALID bit in the status bits of the AD7760 ouptut is set to zero, providing another indication that an input overrange has occurred. Note that the overrange bit is set immediately if the analog input voltage exceeds 100% of VREF for more than four consecutive samples at the modulator rate. GAIN REGISTER--ADDRESS 0x0004 Non-bit-mapped, Default Value 0xA000 The gain register is scaled such that 0x8000 corresponds to a gain of 1.0. The default value of this register is 1.25 (0xA000). This results in a full-scale digital output when the input is at 80% of VREF, tying in with the maximum analog input range of 80% of VREF p-p. Rev. A | Page 34 of 36 AD7760 OUTLINE DIMENSIONS 0.75 0.60 0.45 12.20 12.00 SQ 11.80 1.20 MAX 64 49 1 49 64 1 48 48 PIN 1 10.20 10.00 SQ 9.80 TOP VIEW (PINS DOWN) 6.00 BSC SQ EXPOSED PAD 0 MIN 1.05 1.00 0.95 0.15 0.05 SEATING PLANE 0.20 0.09 7 3.5 0 0.08 MAX COPLANARITY BOTTOM VIEW 16 33 17 32 VIEW A (PINS UP) 33 32 0.50 BSC LEAD PITCH VIEW A 16 17 0.38 0.32 0.22 ROTATED 90 CCW COMPLIANT TO JEDEC STANDARDS MS-026-ACD-HD Figure 59. 64-Lead Thin Quad Flat Package, Exposed Pad [TQFP_EP] (SV-64-2) Dimensions shown in millimeters ORDERING GUIDE Model AD7760BSVZ 1 AD7760BSVZ-REEL1 EVAL-AD7760EB 1 Temperature Range -40C to +85C -40C to +85C Package Description 64-Lead Thin Quad Flat Package, Exposed Pad [TQFP_EP] 64-Lead Thin Quad Flat Package, Exposed Pad [TQFP_EP] Evaluation Board Z = Pb-free part. Rev. A | Page 35 of 36 Package Option SV-64-2 SV-64-2 AD7760 NOTES (c)2006 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D04975-0-8/06(A) Rev. A | Page 36 of 36