Ad7768 Ad7768 4
Ad7768 Ad7768 4
AD7768/AD7768-4
8-/4-Channel, 24-Bit, Simultaneous Sampling ADCs with Power Scaling, 110.8 kHz
Bandwidth
FEATURES ► Linear phase digital filter
► Low latency sinc5 filter
► Precision ac and dc performance
► Wideband brick wall filter: ±0.005 dB pass band ripple to
► 8-/4-channel simultaneous sampling 102.4 kHz
► 256 kSPS maximum ADC ODR per channel ► Analog input precharge and reference precharge buffers
► 108 dB dynamic range ► Power supply
► DC to 110.8 kHz maximum input bandwidth (−3 dB bandwidth) ► AVDD1 = 5.0 V, AVDD2 = 2.25 V to 5.0 V
► −120 dB THD, typical ► IOVDD = 2.5 V to 3.3 V or IOVDD = 1.8 V
► ±2 ppm of full-scale range (FSR) integral nonlinearity (INL), ► 64-lead LQFP, no exposed pad
±50 µV offset error, ±30 ppm of FSR gain error
► Temperature range: −40°C to +105°C
► Optimized power dissipation vs. noise vs. input bandwidth
► Selectable power, speed, and input bandwidth APPLICATIONS
► Fast (highest speed): 110.8 kHz bandwidth, 51.5 mW per
channel ► Data acquisition systems: USB/PXI/Ethernet
► Median (half speed): 55.4 kHz bandwidth, 27.5 mW per ► Instrumentation and industrial control loops
channel ► Audio testing and measurement
► Low power (lowest power): 13.8 kHz bandwidth, 9.375 mW ► Vibration and asset condition monitoring
per channel ► 3-phase power quality analysis
► Per channel digital filter ► Sonar
► Programmable input bandwidth/sampling rates ► High precision medical electroencephalogram (EEG)/electro-
► CRC error checking on data interface myography (EMG)/electrocardiogram (ECG)
► Daisy-chaining
FUNCTIONAL BLOCK DIAGRAM
Figure 1.
Rev. D
Information furnished by Analog Devices is believed to be accurate and reliable "as is". However, no responsibility is assumed by Analog Devices
DOCUMENT FEEDBACK 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
TECHNICAL SUPPORT trademarks are the property of their respective owners. All Analog Devices products contained herein are subject to release and availability.
Data Sheet AD7768/AD7768-4
TABLE OF CONTENTS
REVISION HISTORY
4/2025—Rev. C to Rev. D
Changed Master to Main and Slave to Subordinate (Throughout).................................................................. 1
Changes to Table 1.......................................................................................................................................... 5
Changes to Table 2........................................................................................................................................ 13
Changes to Clock Selection Section..............................................................................................................60
Change to Table 63........................................................................................................................................94
Changes to Table 69...................................................................................................................................... 99
8/2022—Rev. B to Rev. C
Changes to Features Section.......................................................................................................................... 1
Changes to General Description Section.........................................................................................................4
Changes to Specifications Section.................................................................................................................. 5
Changes to Table 9........................................................................................................................................ 24
Changes to Table 10...................................................................................................................................... 28
Changes to Figure 28 Caption....................................................................................................................... 34
Change to AC Common-Mode Rejection Ratio (AC CMRR) Section............................................................ 42
Changes to Power Supplies Section..............................................................................................................48
Changes to Recommended Power Supply Configuration Section.................................................................48
Changes to Analog Supply Internal Connectivity Section..............................................................................49
Changes to Pin Control Section.....................................................................................................................49
Change to SPI Control Functionality Section.................................................................................................53
Changes to Reset over SPI Control Interface Section...................................................................................54
Changes to CRC Protection Section..............................................................................................................55
Changes to ADC Synchronization over SPI Section......................................................................................55
Changes to Clock Selection Section..............................................................................................................60
Change to Table 28 Title................................................................................................................................ 62
Change to Table 29 Title................................................................................................................................ 64
Changes to Modulator Saturation Point Section............................................................................................ 66
Changes to Gain Adjustment Section............................................................................................................ 67
Changes to Daisy-Chaining Section.............................................................................................................. 73
Changes to Table 43...................................................................................................................................... 83
Change to Interface Configuration Register Section......................................................................................85
Changes to Digital Filter RAM Built In Self Test (BIST) Register Section...................................................... 85
Changes to Status Register Section.............................................................................................................. 86
Changes to Offset Registers Section.............................................................................................................90
Changes to Table 63...................................................................................................................................... 94
Changes to Table 67...................................................................................................................................... 98
Changes to Interface Configuration Register Section..................................................................................100
Changes to Digital Filter RAM Built In Self Test (BIST) Register Section.................................................... 101
Changes to Status Register Section............................................................................................................ 101
The AD7768/AD7768-4 are 8-channel and 4-channel 24-bit, simul- Within these filter options, the user can improve the dynamic range
taneous sampling, sigma-delta (Σ-Δ) analog-to-digital converters by selecting from decimation rates of ×32, ×64, ×128, ×256, ×512,
(ADCs) with power scaling and 110.8 kHz bandwidth. The devices and ×1024. The ability to vary the decimation filtering optimizes
have a Σ-Δ modulator and digital filter per channel, which enables noise performance to the required input bandwidth.
synchronized sampling of ac and dc signals.
Embedded analog functionality on each ADC channel makes de-
The AD7768/AD7768-4 achieve 108 dB dynamic range at a max- sign easier, such as a precharge buffer on each analog input that
imum input bandwidth of 110.8 kHz, combined with a typical per- reduces analog input current and a precharge reference buffer per
formance of ±2 ppm integral nonlinearity (INL), ±50 µV offset error, channel that reduces input current and glitches on the reference
and ±30 ppm gain error. input terminals.
The AD7768/AD7768-4 user can trade off input bandwidth, output The device operates with a 5 V AVDD1A and AVDD1B supply, a
data rate, and power dissipation, and select one of three power 2.25 V to 5.0 V AVDD2A and AVDD2B supply, and a 2.5 V to 3.3 V
modes to optimize for noise targets and power consumption. The or 1.8 V IOVDD supply (see the 1.8 V IOVDD Operation section for
flexibility of the AD7768/AD7768-4 allows them to become reusable specific requirements for operating at 1.8 V IOVDD).
platforms for low power dc and high performance ac measurement
modules. The device requires an external reference. The absolute input
reference voltage range is 1 V to AVDD1 − AVSS.
The AD7768/AD7768-4 have three modes: fast mode (256 kSPS
maximum, 110.8 kHz input bandwidth, 51.5 mW per channel), For the purposes of clarity in this data sheet, the AVDD1A and
median mode (128 kSPS maximum, 55.4 kHz input bandwidth, AVDD1B supplies are referred to as AVDD1, and the AVDD2A
27.5 mW per channel) and low power mode (32 kSPS maximum, and AVDD2B supplies are referred to as AVDD2. For the negative
13.8 kHz input bandwidth, 9.375 mW per channel). supplies, AVSS refers to the AVSS1A, AVSS1B, AVSS2A, AVSS2B,
and AVSS pins.
The AD7768/AD7768-4 offer extensive digital filtering capabilities,
such as a wideband, low ±0.005 dB pass-band ripple, antialiasing The specified operating temperature range is −40°C to +105°C. The
low-pass filter with sharp roll-off, and 105 dB attenuation at the device is housed in a 10 mm × 10 mm, 64-lead low profile quad flat
Nyquist frequency. package (LQFP) with a 12 mm × 12 mm printed circuit board (PCB)
footprint.
Frequency domain measurements can use the wideband linear
phase filter. This filter has a flat pass band (±0.005 dB ripple) from Throughout this data sheet, multifunction pins, such as XTAL2/
dc to 102.4 kHz at 256 kSPS, from dc to 51.2 kHz at 128 kSPS, or MCLK, are referred to either by the entire pin name or by a single
from dc to 12.8 kHz at 32 kSPS. function of the pin, for example MCLK, when only that function is
relevant.
The AD7768/AD7768-4 also offer sinc response via a sinc5 filter, a
low latency path for low bandwidth, and low noise measurements.
The wideband and sinc5 filters can be selected and run on a per
channel basis.
AVDD1A = AVDD1B = 4.5 V to 5.5 V, AVDD2A = AVDD2B = 2.0 V to 5.5 V, IOVDD = 2.25 V to 3.6 V, AVSS = DGND = 0 V, REFx+ = 4.096 V
and REFx− = 0 V, MCLK = 32.768 MHz, analog input precharge buffers on, reference precharge buffers off, wideband filter, chopping frequency
(fCHOP) = fMOD/32, TA = −40°C to +105°C, unless otherwise noted. See Table 2 for specifications at 1.8 V IOVDD.
Table 1.
Parameter Test Conditions/Comments Min Typ Max Unit
ADC SPEED AND PERFORMANCE
Output Data Rate (ODR), per Fast mode 8 256 kSPS
Channel1
Median mode 4 128 kSPS
Low power mode 1 32 kSPS
−3 dB Bandwidth Fast mode, wideband filter 110.8 kHz
Median mode, wideband filter 55.4 kHz
Low power mode, wideband filter 13.8 kHz
Data Output Coding Twos complement, MSB first
No Missing Codes2 24 Bits
DYNAMIC PERFORMANCE
Fast Mode Decimation by 32, 256 kSPS ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
Signal-to-Noise Ratio (SNR) 1 kHz, −0.5 dBFS, sine wave input
Sinc5 filter 109 111 dB
Wideband filter 106 107.8 dB
Signal-to-Noise-and-Distortion 1 kHz, −0.5 dBFS, sine wave input 104.7 107.5 dB
Ratio (SINAD)
Total Harmonic Distortion 1 kHz, −0.5 dBFS, sine wave input −120 −107 dB
(THD)
Spurious-Free Dynamic Range 128 dBc
(SFDR)
Median Mode Decimation by 32, 128 kHz ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
SNR Sinc5 filter, 1 kHz, −0.5 dBFS, sine wave input 109 111 dB
Wideband filter, 1 kHz, −0.5 dBFS, sine wave input 106 107.8 dB
SINAD 1 kHz, −0.5 dBFS, sine wave input 105.8 107.5 dB
THD 1 kHz, −0.5 dBFS, sine wave input −120 −113 dB
SFDR 128 dBc
Low Power Mode Decimation by 32, 32 kHz ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
SNR Sinc5 filter, 1 kHz, −0.5 dBFS, sine wave input 109 111 dB
Wideband filter, 1 kHz, −0.5 dBFS, sine wave input 106 107.8 dB
SINAD 1 kHz, −0.5 dBFS, sine wave input 105.8 107.5 dB
THD 1 kHz, −0.5 dBFS, sine wave input −120 −113 dB
SFDR 128 dBc
INTERMODULATON DISTORTION fINA = 9.7 kHz, fINB = 10.3 kHz
(IMD)3
Second-order −125 dB
Third-order −125 dB
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
ACCURACY
INL Endpoint method ±2 ±7 ppm of FSR
Offset Error4 DCLK frequency ≤ 24 MHz ±50 ±115 µV
24 MHz to 32.768 MHz DCLK frequency2 ±75 ±150 µV
Offset Error Drift DCLK frequency ≤ 24 MHz ±250 nV/°C
24 MHz to 32.768 MHz DCLK frequency ±750 nV/°C
Gain Error4 TA = 25°C ±30 ±70 ppm of FSR
Gain Drift vs. Temperature2 ±0.5 ±1 ppm/°C
VCM PIN
Output With respect to AVSS (AVDD1 − V
AVSS)/2
Load Regulation Change in output voltage to change in load current (∆VOUT/∆IL) 400 µV/mA
Voltage Regulation Applies to the following VCM output options only: common-mode 5 µV/V
voltage (VCM) = ∆VOUT/∆(AVDD1 − AVSS)/2, VCM = 1.65 V, and
VCM = 2.5 V
Short-Circuit Current 30 mA
ANALOG INPUTS See the Analog Inputs section
Differential Input Voltage Range VREF = (REFx+) − (REFx−) −VREF +VREF V
Input Common-Mode Range2 AVSS AVDD1 V
Absolute Analog Input Voltage AVSS AVDD1 V
Limits2
Analog Input Current
Unbuffered Differential component ±48 µA/V
Common-mode component +17 µA/V
Precharge Buffer On5 −20 µA
Input Current Drift
Unbuffered ±5 nA/V/°C
Precharge Buffer On ±31 nA/°C
EXTERNAL REFERENCE
Reference Voltage VREF = (REFx+) − (REFx−) 1 AVDD1 − V
AVSS
Absolute Reference Voltage Lim- Precharge reference buffers off AVSS − AVDD1 + V
its2 0.05 0.05
Precharge reference buffer on AVSS AVDD1 V
Average Reference Current Fast mode, see Figure 63
Precharge reference buffers off ±72 µA/V/channel
Precharge reference buffers on ±16 µA/V/channel
Average Reference Current Drift Fast mode, see Figure 63
Precharge reference buffers off ±1.7 nA/V/°C
Precharge reference buffers on ±49 nA/V/°C
Common-Mode Rejection 95 dB
DIGITAL FILTER RESPONSE
Low Ripple Wideband Filter FILTER pin = 0
Decimation Rate Up to six selectable decimation rates 32 1024
Group Delay Latency 34/ODR sec
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Settling Time Complete settling 68/ODR sec
Pass-Band Ripple2 From dc to 102.4 kHz at 256 kSPS ±0.005 dB
Pass Band ±0.005 dB bandwidth 0.4 × ODR Hz
−0.1 dB bandwidth 0.409 × ODR Hz
−3 dB bandwidth 0.433 × ODR Hz
Stop Band Frequency Attenuation > 105 dB 0.499 × ODR Hz
Stop Band Attenuation 105 dB
Sinc5 Filter FILTER pin = 1
Decimation Rate Up to six selectable decimation rates 32 1024
Group Delay Latency 3/ODR sec
Settling Time Complete settling 7/ODR sec
Pass Band −3 dB bandwidth 0.204 × ODR Hz
REJECTION
AC Power Supply Rejection Ratio Input voltage (VIN) = 0.1 V, AVDD1 = 5 V, AVDD2 = 5 V, IOVDD =
(PSRR) 2.5 V
AVDD1 90 dB
AVDD2 100 dB
IOVDD 75 dB
DC PSRR VIN = 1 V
AVDD1 100 dB
AVDD2 118 dB
IOVDD 90 dB
Analog Input Common-Mode Re-
jection Ratio (CMRR)
DC VIN = 0.1 V 95 dB
AC Up to 10 kHz 95 dB
Crosstalk −0.5 dBFS input on adjacent channels −120 dB
CLOCK See the Clocking Selections section for performance functionality
Crystal Frequency 8 32.768 34 MHz
External Clock (MCLK) 32.768 MHz
Duty Cycle 50:50 %
MCLK Pulse Width2
Logic Low 12.2 ns
Logic High 12.2 ns
CMOS Clock Input Voltage See the Logic Inputs parameter
High (VINH)
Low (VINL)
LVDS Clock2 Load resistance (RL) = 100 Ω provided externally
Differential Input Voltage 100 650 mV
Common-Mode Input Voltage 800 1575 mV
Absolute Input Voltage 1.88 V
ADC RESET2
ADC Start-Up Time After Reset6 Time to first DRDY, fast mode, decimation by 32 1.58 1.66 ms
Minimum RESET Low Pulse Width tMCLK = 1/MCLK 2 × tMCLK
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
LOGIC INPUTS
Input Voltage2
High (VINH) 0.65 × V
IOVDD
Low (VINL) 0.7 V
Hysteresis2 0.04 0.09 V
Leakage Current −10 +0.03 +10 µA
RESET pin7 −10 +10 µA
LOGIC OUTPUTS See Table 2 for 1.8 V operation
Output Voltage2
High (VOH) Source current (ISOURCE) = 200 μA 0.8 × V
IOVDD
Low (VOL) Sink current (ISINK) = 400 µA 0.4 V
Leakage Current Floating state −10 +10 µA
Output Capacitance Floating state 10 pF
SYSTEM CALIBRATION2
Full-Scale Calibration Limit 1.05 × VREF V
Zero-Scale Calibration Limit −1.05 × V
VREF
Input Span 0.4 × VREF 2.1 × VREF V
POWER REQUIREMENTS
Power Supply Voltage
AVDD1 − AVSS 4.5 5.0 5.5 V
AVDD2 − AVSS 2.0 2.25 to 5.0 5.5 V
AVSS − DGND −2.75 0 V
IOVDD − DGND See Table 2 for 1.8 V operation 2.25 2.5 to 3.3 3.6 V
POWER SUPPLY CURRENTS Maximum output data rate, CMOS MCLK, eight DOUTx signals,
all supplies at maximum voltages, all channels in Channel Mode
A
AD7768 Eight channels active
Fast Mode
AVDD1 Current Analog input precharge ON 36 40 mA
Analog input precharge ON and positive reference buffer ON 46.9 mA
Analog input precharge ON, and positive and negative reference 57.5 64 mA
buffers ON
AVDD2 Current 37.5 40 mA
IOVDD Current Wideband filter 63 67 mA
Sinc5 filter 27 29 mA
Median Mode
AVDD1 Current Analog input precharge ON 18.5 20.5 mA
Analog input precharge ON and positive reference buffer ON 24.2 mA
Analog input precharge ON, and positive and negative reference 29 32.5 mA
buffers ON
AVDD2 Current 21.3 23 mA
IOVDD Current Wideband filter 34 37 mA
Sinc5 filter 16 18 mA
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Low Power Mode
AVDD1 Current Analog input precharge ON 5.1 5.8 mA
Analog input precharge ON and positive reference buffer ON 6.6 mA
Analog input precharge ON, and positive and negative reference 8 9 mA
buffers ON
AVDD2 Current 9.3 10.1 mA
IOVDD Current Wideband filter 12.5 13.7 mA
Sinc5 filter 8 9 mA
AD7768-4 Four channels active
Fast Mode
AVDD1 Current Analog input precharge ON 18.2 20.3 mA
Analog input precharge ON and positive reference buffer ON 24.5 mA
Analog input precharge ON, and positive and negative reference 28.8 32.5 mA
buffers ON
AVDD2 Current 18.8 20.3 mA
IOVDD Current Wideband filter2 43.5 46.8 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 37 40 mA
filter8
Sinc5 filter2 17 18.6 mA
Median Mode
AVDD1 Current Analog input precharge ON 9.3 10.5 mA
Analog input precharge ON and positive reference buffer ON 13 mA
Analog input precharge ON, and positive and negative reference 14.7 16.6 mA
buffers ON
AVDD2 Current 10.7 11.7 mA
IOVDD Current Wideband filter2 24.4 26.4 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 21 23 mA
filter8
Sinc5 filter2 11 12.3 mA
Low Power Mode
AVDD1 Current Analog input precharge ON 2.7 3.1 mA
Analog input precharge ON and positive reference buffer ON 3.48 mA
Analog input precharge ON, and positive and negative reference 4.1 4.7 mA
buffers ON
AVDD2 Current 4.7 5.3 mA
IOVDD Current Wideband filter2 10 11.1 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 9 10 mA
filter8
Sinc5 filter2 6.5 7.6 mA
AD7768 and AD7768-4—Two Serial peripheral interface (SPI) control mode only, see the Chan-
Channels Active4 nel Standby section for details on disabling channels
Fast Mode
AVDD1 Current Analog input precharge ON 9.3 10.5 mA
Analog input precharge ON and positive reference buffer ON 11.92 mA
Analog input precharge ON, and positive and negative reference 14.7 16.6 mA
buffers ON
AVDD2 Current 9.5 10.5 mA
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
IOVDD Current Wideband filter 33.7 36.3 mA
Wideband filter, disabled channels in Channel Mode A, and set to 23.4 25.5 mA
sinc5 filter mode8
Sinc5 filter 11.9 13.3 mA
Median Mode
AVDD1 Current Analog input precharge ON 4.8 5.5 mA
Analog input precharge ON and positive reference buffer ON 6.2 mA
Analog input precharge ON, and positive and negative reference 7.5 8.6 mA
buffers ON
AVDD2 Current 5.5 6.2 mA
IOVDD Current Wideband filter 19.4 21.1 mA
Wideband filter, disabled channels in Channel Mode A, and set to 14.1 15.5 mA
sinc5 filter mode8
Sinc5 filter 8.5 9.6 mA
Low Power Mode
AVDD1 Current Analog input precharge ON 1.52 1.77 mA
Analog input precharge ON and positive reference buffer ON 1.84 mA
Analog input precharge ON, and positive and negative reference 2.2 2.6 mA
buffers ON
AVDD2 Current 2.4 3 mA
IOVDD Current Wideband filter 8.6 9.7 mA
Wideband filter, disabled channels in Channel Mode A, and set to 7.2 8 mA
sinc5 filter mode8
Sinc5 filter 5.8 6.7 mA
Standby Mode All channels disabled (sinc5 filter enabled) 6.5 8 mA
Sleep Mode2 Full power-down (SPI control mode only) 0.73 1.2 mA
Crystal Excitation Current Extra current in IOVDD when using an external crystal compared 540 µA
to using the CMOS MCLK
POWER DISSIPATION External CMOS MCLK, all channels active, MCLK = 32.768 MHz,
all channels in Channel Mode A except where otherwise speci-
fied
Full Operating Mode—AD7768 Analog precharge buffers on
Wideband Filter
Fast Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 412 446 mW
buffers off2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 600 645 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 631 681 mW
reference buffers off
Median Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 220 240 mW
buffers off2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 320 345 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 341 372 mW
reference buffers off
Low Power Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 75 85 mW
buffers off2
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 107 118 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 124 137 mW
reference buffers off
Sinc5 Filter
Fast Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 325 355 mW
buffers off2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 475 525 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 501 545 mW
reference buffers off
Median Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 175 195 mW
buffers off2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 260 285 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 277 304 mW
reference buffers off
Low Power Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 65 72 mW
buffers off2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 95 105 mW
buffers on2
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 108 120 mW
reference buffers off
Full Operating Mode—AD7768-4
Wideband Filter
Fast Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 235 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 336 mW
buffers on
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 360 392 mW
reference buffers off2
SPI mode only; AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 337 368 mW
V, precharge reference buffers off, Channel Mode A set to sinc5
filter8
Median Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 127 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 181 mW
buffers on
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 198 218 mW
reference buffers off2
SPI mode only; AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 186 205 mW
V, precharge reference buffers off, Channel Mode A set to sinc5
filter8
Low Power Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 49 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 66 mW
buffers on
Table 1. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 77 87 mW
reference buffers off2
SPI mode only; AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 73 83 mW
V, precharge reference buffers off, Channel Mode A set to sinc5
filter8
Sinc5 Filter
Fast Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 168 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 248 mW
buffers on
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 265 291 mW
reference buffers off
Median Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 94 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 137 mW
buffers on
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 150 167 mW
reference buffers off
Low Power Mode AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, precharge reference 40 mW
buffers off
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V, precharge reference 55 mW
buffers on
AVDD1 = 5.5 V, AVDD2 = 5.5 V, IOVDD = 3.6 V, precharge 64 74 mW
reference buffers off
Standby Mode All channels disabled (sinc5 filter enabled), AVDD1 = 5 V, 18 mW
AVDD2 = IOVDD = 2.5 V2
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V2 26 mW
AVDD1 = AVDD2 = 5.5 V, IOVDD = 3.6 V 29 mW
Sleep Mode2 Full power-down (SPI control mode), AVDD1 = 5 V, AVDD2 = 1.8 4 mW
IOVDD = 2.5 V
AVDD1 = 5 V, AVDD2 = IOVDD = 3.3 V 2.5 5 mW
AVDD1 = AVDD2 = 5.5 V, IOVDD = 3.6 V 2.7 6.5 mW
1 The output data rate ranges refer to the programmable decimation rates available on the AD7768/AD7768-4 for a fixed MCLK rate of 32.768 MHz. Varying MCLK rates
allow users a wider variation of ODR.
2 These specifications are not production tested but are supported by characterization data at initial product release.
3 See the Terminology section for more information about the fa and fb input frequencies.
4 Following a system zero-scale calibration, the offset error is in the order of the noise for the programmed output data rate selected. A system full-scale calibration reduces
the gain error to the order of the noise for the programmed output data rate.
5 −25 µA is measured when the analog input is close to either the AVDD1 or AVSS rail. The input current reduces as the common-mode voltage approaches (AVDD1 −
AVSS)/2. The analog input current scales with the MCLK frequency and device power mode. See Figure 85 and Figure 86 for more details on how the analog input current
scales with input voltage.
6 For lower MCLK rates or higher decimation rates, use Table 28 and Table 29 to calculate any additional delay before the first DRDY pulse.
7 The RESET pin has an internal pull-up device to IOVDD.
8 Configuring Channel Mode A to the sinc5 filter and/or assigning disabled channels to Channel Mode A allows a lower power consumption to be achieved. To do this, the
user must be operating in SPI control mode because it requires assigning channels to different channel modes (only possible in SPI control mode). If using pin control
mode, all channels, whether active or in standby, are assigned to the same channel group and use the same filter type. This means that, in pin control mode, a higher
current consumption is seen from disabled channels than can be achieved in SPI mode. See the Channel Modes section for more details.
Table 2.
Parameter Test Conditions/Comments Min Typ Max Unit
DYNAMIC PERFORMANCE For dynamic range and SNR across all decimation rates, see
Table 12 and Table 13
Fast Mode Decimation by 32, 256 kSPS ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
SNR Sinc5 filter, 1 kHz, −0.5 dBFS, sine wave input 109 111 dB
Wideband filter, 1 kHz, −0.5 dBFS, sine wave input 106 107.8 dB
SINAD1 1 kHz, −0.5 dBFS, sine wave input 103.8 107.5 dB
THD 1 kHz, −0.5 dBFS, sine wave input −120 −107 dB
SFDR 128 dBc
Median Mode Decimation by 32, 128 kHz ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
SNR 1 kHz, −0.5 dBFS, sine wave input
Sinc5 filter 109 111 dB
Wideband filter 106 107.8 dB
SINAD 1 kHz, −0.5 dBFS, sine wave input 105.8 107.5 dB
THD 1 kHz, −0.5 dBFS, sine wave input −120 −113 dB
SFDR 128 dBc
Low Power Mode Decimation by 32, 32 kHz ODR
Dynamic Range Shorted input, wideband filter 106.2 108 dB
SNR Sinc5 filter, 1 kHz, −0.5 dBFS, sine wave input 109 111 dB
Wideband filter, 1 kHz, −0.5 dBFS, sine wave input 106 107.8 dB
SINAD 1 kHz, −0.5 dBFS, sine wave input 105.8 107.5 dB
THD 1 kHz, −0.5 dBFS, sine wave input −120 −113 dB
SFDR 128 dBc
ACCURACY1
INL Endpoint method ±2 ±7 ppm of FSR
Offset Error2 DCLK frequency ≤ 24 MHz ±50 ±115 µV
24 MHz to 32.768 MHz DCLK frequency ±75 ±170 µV
Offset Error Drift DCLK frequency ≤ 24 MHz ±250 nV/°C
24 MHz to 32.768 MHz DCLK frequency ±750 nV/°C
Gain Error2 TA = 25°C ±60 ±120 ppm/FSR
Gain Drift vs. Temperature ±0.5 ±2 ppm/°C
LOGIC INPUTS
Input Voltage1
VINH 0.65 × IOVDD V
VINL 0.4 V
Hysteresis1 0.04 0.2 V
Table 2. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Leakage Current −10 +0.03 +10 µA
RESET pin −10 +10 µA
LOGIC OUTPUTS
Output Voltage1
VOH ISOURCE = 200 µA 0.8 × IOVDD V
VOL ISINK = 400 µA 0.4 V
Leakage Current Floating state −10 +10 µA
Output Capacitance Floating state 10 pF
POWER REQUIREMENTS
Power Supply Voltage
AVDD1 − AVSS 4.5 5.0 5.5 V
AVDD2 − AVSS 2.0 2.25 to 5.0 5.5 V
AVSS − DGND −2.75 0 V
IOVDD − DGND DREGCAP shorted to IOVDD 1.72 1.8 1.88 V
POWER SUPPLY CURRENTS1 Maximum output data rate, CMOS MCLK, eight DOUTx signals,
all supplies at maximum voltages, all channels in Channel
Mode A except where otherwise specified
AD7768 Eight channels active
Fast Mode
AVDD1 Current Analog input precharge ON 36 40 mA
Analog input precharge ON and positive reference buffer ON 46.9 mA
Analog input precharge ON, and positive and negative 57.5 64 mA
reference buffers ON
AVDD2 Current 37.5 40 mA
IOVDD Current Wideband filter 63 69 mA
Sinc5 filter 26 28.4 mA
Median Mode
AVDD1 Current Analog input precharge ON 18.5 20.5 mA
Analog input precharge ON and positive reference buffer ON 24.2 mA
Analog input precharge ON, and positive and negative 29 32.5 mA
reference buffers ON
AVDD2 Current 21.3 23 mA
IOVDD Current Wideband filter 34 36.8 mA
Sinc5 filter 15 16.8 mA
Low Power Mode
AVDD1 Current Analog input precharge ON 5.1 5.8 mA
Analog input precharge ON and positive reference buffer ON 6.6 mA
Analog input precharge ON, and positive and negative 8 9 mA
reference buffers ON
AVDD2 Current 9.3 10.1 mA
IOVDD Current Wideband filter 11.6 12.9 mA
Sinc5 filter 7 8.1 mA
AD7768-4 Four channels active
Fast Mode
AVDD1 Current Analog input precharge ON 18.2 20.3 mA
Analog input precharge ON and positive reference buffer ON 24.5 mA
Table 2. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Analog input precharge ON, and positive and negative 28.8 32.5 mA
reference buffers ON
AVDD2 Current 18.8 20.3 mA
IOVDD Current Wideband filter 43.9 47.7 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 36.8 41 mA
filter3
Sinc5 filter 16 17.7 mA
Median Mode
AVDD1 Current Analog input precharge ON 9.3 10.5 mA
Analog input precharge ON and positive reference buffer ON 13 mA
Analog input precharge ON, and positive and negative 14.7 16.6 mA
reference buffers ON
AVDD2 Current 10.7 11.7 mA
IOVDD Current Wideband filter 24 26.1 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 20.4 22.7 mA
filter3
Sinc5 filter 10 11.3 mA
Low Power Mode
AVDD1 Current Analog input precharge ON 2.7 3.1 mA
Analog input precharge ON and positive reference buffer ON 3.48 mA
Analog input precharge ON, and positive and negative 4.1 4.7 mA
reference buffers ON
AVDD2 Current 4.7 5.3 mA
IOVDD Current Wideband filter 9 10.2 mA
Wideband filter, SPI mode only, Channel Mode A set to sinc5 8.1 9.2 mA
filter3
Sinc5 filter 5.5 6.5 mA
AD7768 and AD7768-4—Two SPI control mode only, see the Channel Standby section for
Channels Active details on disabling channels
Fast Mode
AVDD1 Current Analog input precharge ON 9.3 10.5 mA
Analog input precharge ON and positive reference buffer ON 11.92 mA
Analog input precharge ON, and positive and negative 14.7 16.6 mA
reference buffers ON
AVDD2 Current 9.5 10.5 mA
IOVDD Current Wideband filter 33.8 36.7 mA
Wideband filter, SPI mode only, disabled channels in Channel 23.1 25.6 mA
Mode A, and set to sinc5 filter3
Sinc5 filter 11 12.3 mA
Median Mode
AVDD1 Current Analog input precharge ON 4.8 5.5 mA
Analog input precharge ON and positive reference buffer ON 6.2 mA
Analog input precharge ON, and positive and negative 7.5 8.6 mA
reference buffers ON
AVDD2 Current 5.5 6.2 mA
IOVDD Current Wideband filter 18.9 20.6 mA
Table 2. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Wideband filter, SPI mode only; disabled channels in Channel 13.4 15.1 mA
Mode A, and set to sinc5 filter3
Sinc5 filter 7.4 8.6 mA
Low Power Mode
AVDD1 Current Analog input precharge ON 1.52 1.77 mA
Analog input precharge ON and positive reference buffer ON 1.84 mA
Analog input precharge ON, and positive and negative 2.2 2.6 mA
reference buffers ON
AVDD2 Current 2.4 3 mA
IOVDD Current Wideband filter 7.6 8.8 mA
Wideband filter, SPI mode only, disabled channels in Channel 6.3 7.2 mA
Mode A, and set to sinc5 filter3
Sinc5 filter 4.8 5.8 mA
Standby Mode All channels disabled (sinc5 filter enabled) 6.5 8 mA
Sleep Mode Full power-down (SPI control mode) 0.73 1.2 mA
Crystal Excitation Current Extra current in IOVDD when using an external crystal com- 540 µA
pared to using the CMOS MCLK
POWER DISSIPATION1 External CMOS MCLK, all channels active, AVDD1 = AVDD2 =
5.5 V, IOVDD = 1.88 V, MCLK = 32.768 MHz, all channels in
Channel Mode A except where otherwise noted
Full Operating Mode—AD7768 Analog precharge buffers on, eight channels active
Wideband Filter
Fast Mode Reference precharge buffers off 524 571 mW
Positive and negative reference precharge buffers ON 638 704 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 422.9
ON
Median Mode Reference precharge buffers off 284 309 mW
Positive and negative reference precharge buffers ON 342 375 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 224.8
ON
Low Power Mode Reference precharge buffers off 98.5 109 mW
Positive and negative reference precharge buffers ON 118 130 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 74.1
ON
Sinc5 Filter
Fast Mode Reference precharge buffers off 455 495 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 358.1
ON
Median Mode Reference precharge buffers off 248 271 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 192.4
ON
Low Power Mode Reference precharge buffers off 94 105 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 66
ON
Full Operating Mode—AD7768-4 Four channels active
Wideband Filter
Fast Mode Reference precharge buffers off 287 314 mW
Table 2. (Continued)
Parameter Test Conditions/Comments Min Typ Max Unit
Positive and negative reference precharge buffers ON 345 381 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 238
ON
Median Mode Reference precharge buffers off 156 172 mW
Positive and negative reference precharge buffers ON 185 206 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 129.6
ON
Low Power Mode Reference precharge buffers off 58 66 mW
Reference precharge buffers on 66 75 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 43
ON
Sinc5 Filter
Fast Mode Reference precharge buffers off 234 257 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 190.7
ON
Median Mode Reference precharge buffers off 129 144 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 104.4
ON
Low Power Mode Reference precharge buffers off 51 59 mW
AVDD1 = 5 V, AVDD2 = 2 V, positive reference precharge buffer 36.7
ON
Standby Mode All channels disabled (sinc5 filter enabled) 17 mW
Sleep Mode Full power-down (SPI control mode) 1.5 4.5 mW
1 These specifications are not production tested but are supported by characterization data at initial product release.
2 Following a system zero-scale calibration, the offset error is in the order of the noise for the programmed output data rate selected. A system full-scale calibration reduces
the gain error to the order of the noise for the programmed output data rate.
3 This configuration of setting Channel Mode A to the sinc5 filter and/or assigning disabled channels to Channel Mode A allows a lower power consumption to be achieved
due to the disabling of internal clocks on the disabled only and sinc5 only channel modes. This configuration requires assigning sinc5 and wideband filters to different
channels, or channel modes, and is only available in SPI control mode. In pin control mode, all channels, whether active or in standby, effectively use the same channel
mode. See the Channel Modes section for more details.
TIMING SPECIFICATIONS
AVDD1A = AVDD1B = 5 V, AVDD2A = AVDD2B = 5 V, IOVDD = 2.25 V to 3.6 V, Input Logic 0 = DGND, Input Logic 1 = IOVDD, CLOAD = 10
pF on the DCLK pin, CLOAD = 20 pF on the other digital outputs, REFx+ = 4.096 V, TA = −40°C to +105°C. See Table 5 and Table 6 for timing
specifications at 1.8 V IOVDD.
1 These specifications are not production tested but are supported by characterization data at initial product release.
1 These specifications are not production tested but are supported by characterization data at initial product release.
1 These specifications are not production tested but are supported by characterization data at initial product release.
1 These specifications are not production tested but are supported by characterization data at initial product release.
Timing Diagrams
1 AI is analog input, P is power, DI/O is digital input/output, DI is digital input, DO is digital output, and AO is analog output.
AVDD1 = 5 V, AVDD2 = 2.5 V, AVSS = 0 V, IOVDD = 2.5 V, VREF = 4.096 V, TA = 25°C, wideband filter, decimation = ×32, MCLK = 32.768 MHz,
analog input precharge buffers on, precharge reference buffers off, unless otherwise noted.
Figure 12. Fast Fourier Transform (FFT), Fast Mode, Wideband Filter, Figure 15. FFT, Fast Mode, Wideband Filter, −6 dBFS
−0.5 dBFS
Figure 18. FFT, Fast Mode, Sinc5 Filter, −0.5 dBFS Figure 21. FFT, Fast Mode, Sinc5 Filter, −6 dBFS
Figure 19. FFT, Median Mode, Sinc5 Filter, −0.5 dBFS Figure 22. FFT, Median Mode, Sinc5 Filter, −6 dBFS
Figure 20. FFT, Low Power Mode, Sinc5 Filter, −0.5 dBFS Figure 23. FFT, Low Power Mode, Sinc5 Filter, −6 dBFS
Figure 24. FFT One Shot Mode, Sinc5 Filter, Median Mode, Decimation = ×64, Figure 27. Shorted Noise, Sinc5 Filter
−0.5 dBFS, SYNC_IN Frequency = MCLK/4000
Figure 30. RMS Noise vs. Temperature, Median Mode Figure 33. Crosstalk
Figure 34. SNR, Dynamic Range, THD, and THD + N vs. MCLK Frequency
Figure 31. RMS Noise vs. Temperature, Low Power Mode
Figure 35. THD vs. Input Frequency, Three Power Modes, Wideband Filter
Figure 32. RMS Noise per Channel for Various VREF Values
Figure 36. THD vs. Input Frequency, Three Power Modes, Sinc5 Filter Figure 39. SNR vs. Input Amplitude
Figure 37. THD and THD + N vs. Input Amplitude, Wideband Filter Figure 40. INL Error vs. Input Voltage for Various VREF Levels, Fast Mode
Figure 38. THD and THD + N vs. Input Amplitude, Sinc5 Filter Figure 41. INL Error vs. Input Voltage for Various VREF Levels, Median Mode
Figure 42. INL Error vs. Input Voltage for Various VREF Levels, Low Power Figure 45. Offset Error Distribution, DCLK = 24 MHz
Mode
Figure 44. INL Error vs. Input Voltage for Various Temperatures, Fast Mode
Figure 48. Offset Error Drift, DCLK = 32 MHz Figure 51. Gain Error Distribution
Figure 49. Offset Drift vs. DCLK Frequency Figure 52. Channel to Channel Gain Error Matching
Figure 50. Channel Offset Error Matching Figure 53. AC CMRR vs. Input Frequency
Figure 54. AC PSRR vs. Frequency, AVDD1 Figure 57. Amplitude vs. Normalized Input Frequency (fIN/fODR), Wideband
Filter Profile
Figure 55. AC PSRR vs. Frequency, AVDD2 Figure 58. Step Response, Wideband Filter
Figure 56. AC PSRR vs. Frequency, IOVDD Figure 59. Wideband Filter Ripple
Figure 60. Amplitude vs. Normalized Input Frequency (fIN/fODR), Sinc5 Filter Figure 63. Reference Input Current vs. Temperature, Reference Precharge
Profile Buffers On/Off
Figure 62. Analog Input Current vs. Temperature, Analog Input Precharge
Buffers On/Off Figure 65. Supply Current vs. Temperature, AVDD1
INL error refers to the deviation of each individual code from a Signal-to-Noise-and-Distortion Ratio (SINAD)
line drawn from negative full scale through positive full scale. The SINAD is the ratio of the rms value of the actual input signal to
point used as negative full scale occurs ½ LSB before the first code the rms sum of all other spectral components below the Nyquist
transition. Positive full scale is defined as a level 1½ LSB greater frequency, including harmonics but excluding dc. The value for
than the last code transition. The deviation is measured from the SINAD is expressed in decibels.
middle of each code to the true straight line.
Spurious-Free Dynamic Range (SFDR)
Intermodulation Distortion (IMD)
SFDR is the difference, in decibels, between the rms amplitude of
With inputs consisting of sine waves at two frequencies, fa and fb, the input signal and the peak spurious signal (excluding the first five
any active device with nonlinearities creates distortion products at harmonics).
the sum and difference frequencies of mfa and nfb, where m, n =
0, 1, 2, 3, and so on. Intermodulation distortion terms are those for Total Harmonic Distortion (THD)
which neither m or n are equal to 0. For example, the second-order THD is the ratio of the rms sum of the first five harmonic compo-
terms include (fa + fb) and (fa − fb), and the third-order terms nents to the rms value of a full-scale input signal and is expressed
include (2fa + fb), (2fa − fb), (fa + 2fb), and (fa − 2fb). in decibels.
The AD7768/AD7768-4 are tested using the CCIF standard, where
two input frequencies near to each other 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 fre-
quency close to the input frequencies. As a result, the second-order
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.
The Σ-Δ conversion technique is an oversampled architecture. This CLOCKING, SAMPLING TREE, AND POWER
oversampled approach spreads the quantization noise over a wide SCALING
frequency band (see Figure 69). To reduce the quantization noise
in the signal band, the high order modulator shapes the noise The AD7768/AD7768-4 include multiple ADC cores. Each of these
spectrum so that most of the noise energy is shifted out of the ADCs receives the same main clock signal, MCLK. The MCLK
band of interest (see Figure 70). The digital filter that follows the signal can be sourced from one of three options: a CMOS clock,
modulator removes the large out of band quantization noise (see a crystal connected between the XTAL1 pin and XTAL2 pin, or
Figure 71). in the form of an LVDS signal. The MCLK signal received by the
AD7768/AD7768-4 defines the modulator clock rate, fMOD, and, in
For further information on the basics as well as more advanced turn, the sampling frequency of the modulator of 2 × fMOD. The
concepts of Σ-Δ ADCs, see the MT-022 Tutorial and the MT-023 same MCLK signal is also used to define the digital output clock,
Tutorial. DCLK. The fMOD and DCLK internal signals are synchronous with
Digital filtering has certain advantages over analog filtering. First, MCLK.
it is insensitive to component tolerances and the variation of com- Figure 72 illustrates the clock tree from the MCLK input to the
ponent parameters over time and temperature. Because digital modulator, the digital filter, and the DCLK output. There are divider
filtering on the AD7768/AD7768-4 occurs after the analog to digital settings for MCLK and DCLK. These dividers in conjunction with
conversion, digital filtering can remove some of the noise injected the power mode and digital filter decimation settings are key to
during the conversion process. Analog filtering cannot remove AD7768/AD7768-4 operation.
noise injected during conversion. Second, the digital filter combines
low pass-band ripple with a steep roll-off, and high stop band The AD7768/AD7768-4 have the ability to scale power consump-
attenuation, while also maintaining a linear phase response, which tion vs. the input bandwidth or noise desired. The user controls
is difficult to achieve in an analog filter implementation. two parameters to achieve this: MCLK division and power mode.
Combined, these two settings determine the clock frequency of the
modulator (fMOD) and the bias current supplied to each modulator.
The power mode (fast, median, or low power) sets the noise, speed
capability, and current consumption of the modulator. The power
mode is the dominant control for scaling the power consumption of
the ADC. All settings of MCLK division and power mode apply to all
ADC channels.
Figure 69. Σ-Δ ADC Quantization Noise (Linear Scale X-Axis)
Figure 72. Sampling Structure, Defined by MCLK, DCLK_DIV, and MCLK_DIV Settings
The modulator clock frequency (fMOD) is determined by selecting highest resolution. This choice is due to an overlap in the coverage
one of three clock divider settings: MCLK/4, MCLK/8, or MCLK/32. of each power mode. The devices offer the ability to balance the
MCLK division ratio with the rate of decimation (averaging) set
Although the MCLK division and power modes are independent in the digital filter. Lower power can be achieved by using lower
settings, there are restrictions that must be adhered to. A valid modulator clock frequencies. Conversely, the highest resolution
range of modulator frequencies exists for each power mode. Table can be achieved by using higher modulator clock frequencies and
11 describes this recommended range, which allows the device to maximizing the amount of oversampling.
achieve the best performance while minimizing power consumption.
The AD7768/AD7768-4 specifications do not cover the performance As an example, consider a system constraint with a maximum
and function beyond the maximum fMOD for a given power mode. available MCLK of 16 MHz. The system is targeting a measurement
bandwidth of approximately 25 kHz with the wideband filter, setting
For example, in fast mode, to maximize the speed of conversion the output data rate of the AD7768/AD7768-4 to 62.5 kHz. Because
or input bandwidth, an MCLK of 32.768 MHz is required and of the low MCLK frequency available and system power budget,
MCLK_DIV = 4 must be selected for a modulator frequency of median power mode is used.
8.192 MHz.
Table 11. Recommended fMOD Range for Each Power Mode
In median power mode, this 25 kHz input bandwidth can be ach-
ieved by setting the MCLK division and decimation ratio to balance,
Power Mode Recommended fMOD (MHz) Range, MCLK = 32.768 MHz using two configurations. This flexibility is possible in SPI control
Low Power 0.036 to 1.024 mode only.
Median 1.024 to 4.096
Fast 4.096 to 8.192
Configuration A
To maximize the dynamic range, use the following settings:
Control of the settings for power mode, the modulator frequency
and the data clock frequency differs in pin control mode vs. SPI ► MCLK = 16 MHz
control mode. ► Median power
In SPI control mode, the user can program the power mode, MCLK ► fMOD = MCLK/4
divider (MCLK_DIV), and DCLK frequency using Register 0x04 and ► Decimation = ×64 (digital filter setting)
Register 0x07 (see Table 42 and Table 45 for register information ► ODR = 62.5 kHz
for the AD7768 or Table 68 and Table 71 for the AD7768-4). Inde-
pendent selection of the power mode and MCLK_DIV allows full This configuration maximizes the available decimation rate (or over-
freedom in the MCLK speed selection to achieve a target modulator sampling ratio) for the bandwidth required and MCLK rate available.
frequency. The decimation averages the noise from the modulator, maximizing
the dynamic range.
In pin control mode, the MODEx pins determine the power mode,
modulator frequency, and DCLK frequency. The modulator frequen- Configuration B
cy tracks the power mode. This means that fMOD is fixed at
MCLK/32 for low power mode, MCLK/8 for median mode, and To minimize power, use the following settings:
MCLK/4 for fast mode (see Table 20). ► MCLK = 16 MHz
► Median power
Example of Power vs. Noise Performance
► fMOD = MCLK/8
Optimization
► Decimation = ×32 (digital filter setting)
Depending on the bandwidth of interest for the measurement, the ► ODR = 62.5 kHz
user can choose a strategy of either lowest current consumption or
This configuration reduces the clocking speed of the modulator and the AD7768-4. Therefore, the intended minimum decimation and
the digital filter. desired DCLK_DIV setting must be understood prior to choosing
the setting of the FORMATx pins.
Compared to Configuration A, Configuration B saves 48 mW of
power. The trade-off in the case of Configuration B is that the digital NOISE PERFORMANCE AND RESOLUTION
filter must run at a 2× lower decimation rate. This 2× reduction in
decimation rate (or oversampling ratio) results in a 3 dB reduction in Table 12 and Table 13 show the noise performance for the wide-
the dynamic range vs. Configuration A. band and sinc5 digital filters of the AD7768/AD7768-4 for various
output data rates and power modes. The noise values and dynamic
Clocking Out the ADC Conversion Results range specified are typical for the bipolar input range with an
external 4.096 V reference (VREF). The rms noise is measured with
(DCLK)
shorted analog inputs, which are driven to (AVDD1 − AVSS)/2 using
The AD7768/AD7768-4 DCLK is a divided version of the main clock the on-board VCM buffer output.
input. As shown in Figure 72, the DCLK_DIV setting determines the The dynamic range is calculated as the ratio of the rms shorted
speed of the DCLK. DCLK is a continuous clock. input noise to the rms full-scale input signal range.
The user can set the DCLK frequency rate to one of four divisions Dynamic Range (dB) = 20log10((2 × VREF/2√2)/(RMS Noise)
of MCLK: MCLK/1, MCLK/2, MCLK/4, and MCLK/8. Because there
are eight channels and 32 bits of data per conversion, the conver- The LSB size with 4.096 V reference is 488 nV, and is calculated as
sion time and the setting of DCLK directly determine the number follows:
of data output lines that are required via the FORMAT0 pin and
FORMAT1 pin settings on the AD7768, or the FORMAT0 pin on LSB (V) = (2 × VREF)/224
Table 12. Wideband Filter Noise: Performance vs. Output Data Rate (VREF = 4.096 V)
Output Data Rate (kSPS) −3 dB Bandwidth (kHz) Shorted Input Dynamic Range (dB) RMS Noise (µV)
Fast Mode
256 110.8 107.96 11.58
128 55.4 111.43 7.77
64 27.7 114.55 5.42
32 13.9 117.58 3.82
16 6.9 120.56 2.72
8 3.5 123.5 1.94
Median Mode
128 55.4 108.13 11.36
64 27.7 111.62 7.6
32 13.9 114.75 5.3
16 6.9 117.79 3.74
8 3.5 120.8 2.64
4 1.7 123.81 1.87
Low Power Mode
32 13.9 108.19 11.28
16 6.9 111.69 7.54
8 3.5 114.83 5.25
4 1.7 117.26 3.71
2 0.87 120.88 2.62
1 0.43 123.88 1.85
Table 13. Sinc5 Filter Noise: Performance vs. Output Data Rate (VREF = 4.096 V)
Output Data Rate (kSPS) −3 dB Bandwidth (kHz) Shorted Input Dynamic Range (dB) RMS Noise (µV)
Fast Mode
256 52.224 111.36 7.83
128 26.112 114.55 5.43
64 13.056 117.61 3.82
32 6.528 120.61 2.71
16 3.264 123.52 1.93
8 1.632 126.39 1.39
Median Mode
128 26.112 111.53 7.68
64 13.056 114.75 5.3
32 6.528 117.81 3.72
16 3.264 120.82 2.64
8 1.632 123.82 1.87
4 0.816 126.79 1.33
Low Power Mode
32 6.528 111.57 7.65
16 3.264 114.82 5.26
8 1.632 117.88 3.7
4 0.816 120.9 2.61
2 0.408 123.91 1.85
1 0.204 126.89 1.31
The AD7768/AD7768-4 offer users a multichannel platform meas- ► Control of reference and analog input precharge buffers on a per
urement solution for ac and dc signal processing. channel basis.
Flexible filtering allows the AD7768/AD7768-4 to be configured to ► Wideband, low ripple, digital filter for ac measurement.
simultaneously sample ac and dc signals on a per channel basis. ► Fast sinc5 filter for precision low frequency measurement.
Power scaling allows users to trade off the input bandwidth of the ► Two channel modes, defined by the user selected filter choice,
measurement vs. the current consumption. This ability, coupled with and decimation ratios, can be defined for use on different ADC
the flexibility of the digital filtering, allows the user to optimize the channels. This enables optimization of the input bandwidth ver-
energy efficiency of the measurement, while still meeting power, sus the signal of interest.
bandwidth, and performance targets. ► Option of SPI or pin strapped control and configuration.
Key capabilities that allow users to choose the AD7768/ ► Offset, gain, and phase calibration registers per channel.
AD7768-4as their platform high resolution ADC are as follows: ► Common-mode voltage output buffer for use by driver amplifier.
► On-board AVDD2 and IOVDD LDOs for the low power, 1.8 V,
► Eight fully differential or pseudo differential analog inputs on the
internal circuitry.
AD7768 (four channels on the AD7768-4).
► Fast throughput simultaneous sampling ADCs catering for input Refer to Figure 73 and Table 14 for the typical connections
signals up to 110.8 kHz. and minimum requirements to get started using the AD7768/
► Three selectable power modes (fast, median, and low power) for AD7768-4.
scaling the current consumption and input bandwidth of the ADC Table 15 shows the typical power and performance of the AD7768/
for optimal measurement efficiency. AD7768-4 for the available power modes, for each filter type.
► Analog input precharge and reference precharge buffers reduce
the drive requirements of external amplifiers.
Table 15. Speed, Dynamic Range, THD, and Power Overview, Eight Channels Active, Decimate by 321
1 Analog precharge buffers on, reference precharge buffers and VCM disabled, typical values, AVDD1 = 5 V, AVDD2 = IOVDD = 2.5 V, VREF = 4.096 V, MCLK = 32.768 MHz,
DCLK = MCLK/4, TA = 25°C.
POWER SUPPLIES (minimum) to 1.88 V (maximum), with respect to DGND. See the
1.8 V IOVDD Operation section for more information on operating
The AD7768/AD7768-4 have three independent power supplies: the AD7768/AD7768-4 at 1.8 V IOVDD.
AVDD1 (AVDD1A and AVDD1B), AVDD2 (AVDD2A and AVDD2B),
and IOVDD. Recommended Power Supply Configuration
The reference potentials for these supplies are AVSS and DGND. Analog Devices, Inc., has a wide range of power management
Tie all the AVSS supply pins (AVSS1A, AVSS1B, AVSS2A, products to meet the requirements of most high performance signal
AVSS2B, and AVSS) to the same potential with respect to DGND. chains.
AVDD1A, AVDD1B, AVDD2A, and AVDD2B are referenced to this
AVSS rail. IOVDD is referenced to DGND. An example of a power solution that uses the ADP7118 is shown in
Figure 74. The ADP7118 provides positive supply rails for optimal
The supplies can be powered within the following ranges: converter performance, creating either a single 5 V, 3.3 V, or
► AVDD1 = 5 V ± 10%, relative to AVSS dual AVDD1x and AVDD2x/IOVDD supply rail, depending on the
► AVDD2 = 2 V to 5.5 V, relative to AVSS required supply configuration. The ADP7118 can operate from input
► IOVDD (with internal regulator) = 2.25 V to 3.6 V, relative to voltages of up to 20 V.
DGND
► IOVDD (bypassing regulator) = 1.72 V to 1.88 V, relative to
DGND
► AVSS = −2.75 V to 0 V, relative to DGND
The AVDD1A and AVDD1B (AVDD1) supplies power the analog
front end, reference input, and common-mode output circuitry. Figure 74. Power Supply Configuration
AVDD1 is referenced to AVSS, and all AVDD1 supplies must be tied
Alternatively, the ADP7112 or ADP7104 can be selected for power-
to the same potential with respect to AVSS. If AVDD1 supplies are
ing the AD7768/AD7768-4. Refer to the AN-1120 Application Note
used in a ±2.5 V split supply configuration, the ADC inputs are truly
for more information regarding low noise LDO performance and
bipolar. When using split supplies, reference the absolute maximum
power supply filtering.
ratings, which apply to the voltage allowed between AVSS and
IOVDD supplies.
1.8 V IOVDD Operation
The AVDD2A and AVDD2B (AVDD2) supplies connect to internal
1.8 V analog LDO regulators. The regulators power the ADC core. The AD7768/AD7768-4 contain an internal 1.8 V LDO on the
AVDD2 is referenced to AVSS, and all AVDD2 supplies must be tied IOVDD supply to regulate the IOVDD down to the operating voltage
to the same potential with respect to AVSS. The voltage on AVDD2 of the digital core. This internal LDO allows the internal logic to
can range from 2 V (minimum) to 5.5 V (maximum), with respect to operate efficiently at 1.8 V and the input/output logic to operate at
AVSS. the level set by IOVDD. The IOVDD supply is rated from 2.25 V to
3.6 V for normal operation, and 1.8 V for LDO bypass setup.
IOVDD powers the internal 1.8 V digital LDO regulator. This regula-
tor powers the digital logic of the ADC. IOVDD also sets the voltage
levels for the SPI interface of the ADC. IOVDD is referenced to
DGND, and the voltage on IOVDD can vary from 2.25 V (minimum)
to 3.6 V (maximum), with respect to DGND. IOVDD can also be
configured to run at 1.8 V. In this case, IOVDD and DREGCAP
must be tied together and must be within the range of 1.72 V
analog.com Rev. D | 48 of 108
Data Sheet AD7768/AD7768-4
APPLICATIONS INFORMATION
For control, the device can be configured in one of two modes. The
two modes of configuration are as follows:
► Pin control mode: pin strapped digital logic inputs (which allows a
subset of the configurability options)
► SPI control mode: over a 3-wire or 4-wire SPI interface (complete
configurability)
On power-up, the state of the PIN/SPI pin determines the mode
used. Immediately after power-up, the user must apply a soft or
hard reset to the device when using either control mode.
pin. The chosen decimation rate is used on all ADC channels. Table the user to reduce the DCLK frequency for lower, less demanding
17 shows the truth table for the DECx pins. power modes and selecting either the one-shot or standard conver-
Table 17. Decimation Rate Control Pins Truth Table
sion modes.
DEC1 DEC0 Decimation Rate See Table 20 for the complete selection of operating modes that are
0 0 ×32
available via the MODEx pins in pin control mode.
0 1 ×64 The power mode setting automatically scales the bias currents of
1 0 ×128 the ADC and divides the applied MCLK signal to the correct setting
1 1 ×1024
for that mode. Note that this is not the same as using SPI control,
where separate bit fields exist to control the bias currents of the
ADC and MCLK division.
Operating Mode
In pin control mode, the modulator rate is fixed for each power
The MODE3 pin to MODE0 pin determine the configuration of all mode to achieve the best performance. Table 19 shows the modu-
channels when using pin control mode. The variables controlled lator division for each power mode.
by the MODEx pins are shown in Table 18. The user selects how
much current the device consumes, the sampling speed of the Table 19. Modulator Rate, Pin Control Mode
ADC (power mode), how fast the ADC result is received by the Power Mode Modulator Rate (fMOD)
digital host (DCLK_DIV), and how the ADC conversion is initiated Fast MCLK/4
(conversion operation). Figure 76 illustrates the inputs used to
Median MCLK/8
configure the AD7768 in pin control mode, and Figure 77 illustrates
the inputs used to configure the AD7768-4 in pin control mode. Low Power MCLK/32
The MODEx pins map to 16 distinct settings. The settings are se-
lected to optimize the use cases of the AD7768/AD7768-4, allowing
Configuration Example Minimizing the DCLK frequency means selecting DCLK = MCLK/8,
which results in a 4 MHz DCLK signal. The period of DCLK in this
In the example shown in Table 23, the lowest current consumption case is 1/4 MHz = 250 ns. The data conversion on each DOUTx
is used, and the AD7768/AD7768-4 are connected to an FPGA. pin is 32 bits long. The conversion data takes 32 × 250 ns = 8
The FORMATx pins are set such that all eight data outputs, DOUT0 µs to be output. All 32 bits must be output within the ODR period
to DOUT7, connect to the FPGA. For the lowest power, the lowest of 1/16 kHz, which is approximately 64 µs. In this case, the 8 µs
DCLK frequency is used. The input bandwidth is set through the required to read out the conversion data is well within the 64 µs
combination of selecting decimation by 64 and selecting the wide- between conversion outputs. Therefore, this combination, which is
band filter. summarized in Table 23, is viable for use.
ODR = fMOD ÷ Decimation Ratio
Channel Standby
where:
fMOD is MCLK/32 for low power mode (see Table 19). Table 21 and Table 23 show how the user can put channels into
MCLK = 32.768 MHz. standby mode. Set either ST0 or ST1 to Logic 1 to place banks
Decimation Ratio = 64. of four channels into standby mode. When in standby mode, the
channels are disabled but still hold their position in the output data
Therefore, for this example, where MCLK = 32.768 MHz, stream. The 8-bit header and 24-bit conversion result are set to all
ODR = (32.768 MHz/32) ÷ 64 = 16 kHz zeros when the ADC channels are set to standby.
The VCM voltage output is associated with the Channel 0 circuitry. CS clocks out the MSB, the falling edge of SCLK is the drive edge,
If Channel 0 is put into standby mode, the VCM voltage output is and the rising edge of SCLK is the sample edge. This means that
also disabled for maximum power savings. Channel 0 must be ena- data is clocked out on the falling/drive edge and data is clocked in
bled while VCM is being used externally to the AD7768/AD7768-4. on the rising/sample edge.
The crystal excitation circuitry is associated with the Channel 4
(Channel 2 on the AD7768-4) circuitry. If Channel 4 (Channel 2 on
the AD7768-4) is put into standby mode, the crystal circuitry is also
disabled for maximum power savings. Channel 4 must be enabled
while the external crystal is used on the AD7768. Channel 2 must
be enabled while the external crystal is used on the AD7768-4.
Figure 78. SPI Mode 0 SCLK Edges
Table 21. Truth Table for the AD7768 ST0 Pin and ST1 Pin
ST1 ST0 Function Accessing the ADC Register Map
0 0 All channels operational.
To use SPI control mode, set the PIN/SPI pin to logic high. The SPI
0 1 Channel 0 to Channel 3 in standby.
control mode operates as a 16-bit, 4-wire interface, allowing read
Channel 4 to Channel 7 operational.
and write access. Figure 80 shows the interface format between the
1 0 Channel 4 to Channel 7 in standby. AD7768/AD7768-4 and the digital host.
Channel 0 to Channel 3 operational.
1 1 All channels in standby.
The SPI serial control interface of the AD7768 is an independent
path for controlling and monitoring the AD7768. There is no direct
Table 22. Truth Table for the AD7768-4 ST0 Pin link to the data interface. The timing of MCLK and DCLK is not
ST0 Function directly related to the timing of the SPI control interface. However,
0 All channels operational.
the user must ensure that the SPI reads and writes satisfy the
minimum t30 specification (see Table 4 and Table 6) so that the
1 Channel 0 to Channel 3 in standby.
AD7768/AD7768-4 can detect changes to the register map.
SPI CONTROL SPI access is ignored during the period immediately after a reset.
The AD7768/AD7768-4 have a 4-wire SPI interface that is compati- Allow the full ADC start-up time after reset (see Table 1) to elapse
ble with QSPI™, MICROWIRE®, and DSPs. The interface operates before accessing the AD7768/AD7768-4 over the SPI interface.
in SPI Mode 0. In SPI Mode 0, SCLK idles low, the falling edge of
Table 23. MODEx Example Selection
Mode Hex MODE3 MODE2 MODE1 MODE0 Power Mode DCLK Frequency Data Conversion
0x3 0 0 1 1 Low power MCLK/8 Standard
one channel assigned to Channel Mode A. If all eight channels Sleep Mode
of the AD7768 are assigned to Channel Mode B, conversion data
is not output on the data interface for any of the channels. This Sleep mode puts the AD7768/AD7768-4 into their lowest power
consideration does not affect the AD7768-4. mode. In sleep mode, all ADCs are disabled and a large portion of
the digital core is inactive.
On the AD7768-4, it is recommended that Channel Mode A be set
to the sinc5 filter whenever possible. There is a small power saving The AD7768/AD7768-4 SPI remains active and is available to the
in IOVDD current when Channel Mode A is set to the sinc5 filter user when in sleep mode. Write to Register 0x04, Bit 7 to exit sleep
compared to setting Channel Mode A to the wideband filter. mode. For the lowest power consumption, select the sinc5 filter
before entering sleep mode.
For example, to assign two channels of the AD7768-4 to the
wideband filter, and the remaining two channels to the sinc5 filter, it Channel Standby
is recommended to assign the two sinc5 filter channels to Channel
Mode A. Set Channel Mode A to the sinc5 filter, set Channel Mode For efficient power usage, users can place the selected channels
B to the wideband filter, and assign the two wideband filter channels into standby mode, effectively disabling them, when not in use. Set-
to Channel Mode B. Similarly, to assign all four channels of the ting the bits in Register 0x00 disables the corresponding channel
AD7768-4 to wideband filter, assign all four channels to Channel (see Table 38 for the AD7768 or Table 64 for the AD7768-4). For
Mode B. Set Channel Mode B to the wideband filter, and keep maximum power savings, switch disabled channels to the sinc5
Channel Mode A set to the sinc5 filter. Assigning the channels in filter using the channel mode configurations, which disables some
this way ensures that the lowest IOVDD current is achieved. clocks associated with the wideband filters of those channels.
Table 24. Channel Mode A/Channel Mode B, Register 0x01 and Register 0x02 For highest power savings when disabling channels on the
Bits Bit Name Setting Description Reset Access AD7768-4, set Channel Mode A to the sinc5 filter, and assign
the disabled channels to Channel Mode A, while keeping any active
3 FILTER_TYPE_x Filter output 0x1 RW channels in Channel Mode B.
0 Wideband filter
The VCM voltage output is associated with the Channel 0 circuitry.
1 Sinc5 filter
If Channel 0 is put into standby mode, the VCM voltage output is
[2:0] DEC_RATE_x Decimation rate 0x5 RW also disabled for maximum power savings. Channel 0 must be ena-
000 to ×32 to ×1024 bled while VCM is being used externally to the AD7768/AD7768-4.
101
The crystal excitation circuitry is associated with the Channel 4
Table 25. Channel Mode Selection, Register 0x03 (Channel 2 on the AD7768-4) circuitry. If Channel 4 (Channel 2 on
Bits Bit Name Setting Description Reset Access the AD7768-4) is put into standby mode, the crystal circuitry is also
[7:0] CH_x_MODE Channel x 0x0 RW disabled for maximum power savings. Channel 4 must be enabled
while the external crystal is used on the AD7768. Channel 2 must
0 Mode A
be enabled while the external crystal is used on the AD7768-4.
1 Mode B
Clocking Selections
Reset over SPI Control Interface
The internal fMOD that is used by each of the ADCs in the AD7768/
Two successive commands must be written to the AD7768/ AD7768-4 is derived from the externally applied MCLK signal. The
AD7768-4 data control register to initiate a full reset of the device MCLK division bits allow the user to control the ratio between the
over the SPI interface. This action fully resets all registers to the MCLK frequency and the internal modulator clock frequency. This
default conditions. Details of the commands and their sequence are control allows the user to select the division ratio that is best for
shown in Table 44 for the AD7768 or Table 70 for the AD7768-4. their configuration.
After a reset over the SPI control interface, the AD7768/AD7768-4 The appropriate clock configuration depends on the power mode,
respond to the first command sent to them with 0x0E00. This the decimation rate, and the base MCLK frequency available in
response, in addition to the fact that all registers have assumed the system. See the Clocking, Sampling Tree, and Power Scaling
their default values, indicates that the software reset succeeded. section for further information on setting MCLK_DIV correctly.
After a reset, it is recommended to wait for the specified ADC
start-up time after reset time to elapse before issuing an SPI write MCLK Source Selection
command.
The following clocking options are available as the MCLK input
source in SPI control mode:
► LVDS
Analog Input Precharge Buffers Normal ADC conversion is disrupted when this test is run. A syn-
chronization pulse is required after this test is complete to resume
The AD7768/AD7768-4 contain precharge buffers on each analog normal ADC operation.
input to ease the drive requirements on the external amplifier. Each
analog input precharge buffer can be enabled or disabled using the Revision Identification Number
analog input precharge buffer registers (see Table 52 and Table 53
for the AD7768 or Table 78 and Table 79 for the AD7768-4). When The AD7768/AD7768-4 contain an identification register that can be
writing to these registers, the user must write the inverse of the accessed in SPI control mode, the revision identification register.
required bit settings. For example, to clear Bit 1 of this register, the This register is an excellent way to verify the correct operation of
user must write 0x01 to the register. This clears Bit 1 and sets all the serial control interface. Register information is available in the
other bits. If the user reads the register again after writing 0x01, the Revision Identification Register section.
data read is 0xFE, as required.
Diagnostic Meter Mode
Reference Precharge Buffers
The diagnostic metering mode can be used to verify the functionali-
The AD7768/AD7768-4 contain reference precharge buffers on ty of each ADC by internally passing a positive full-scale, midscale,
each reference input to ease the drive requirements on the external or negative full-scale voltage to the ADC. The user can then read
reference and help to settle any nonlinearity on the reference the resulting ADC conversion result to determine that the ADC is
inputs. Each reference precharge buffer can be enabled or disa- operating correctly. To configure ADC conversion diagnostics, see
bled using the reference precharge buffer registers (see Table 54 the ADC Diagnostic Receive Select Register section and the ADC
and Table 55 for the AD7768 or Table 80 and Table 81 for the Diagnostic Control Register section.
AD7768-4).
GPIOs
The AD7768/AD7768-4 have five general-purpose input/output
(GPIO) pins available when operating in SPI control mode. For
further information on GPIO configuration, see the GPIO Function-
ality section.
SPI CONTROL MODE EXTRA DIAGNOSTIC
FEATURES
CORE SIGNAL CHAIN power dissipation for the AD7768/AD7768-4. Table 11 shows the
recommended fMOD frequencies for each power mode, Table 42
Each ADC channel on the AD7768/AD7768-4 has an identical shows the register information for the AD7768, and Table 68 shows
signal path from the analog input pins to the data interface. Figure the register information for the AD7768-4.
83 shows a top level implementation of the core signal chain. Each
ADC channel has its own Σ-Δ modulator that oversamples the
analog input and passes the digital representation to the digital filter
block. The fMOD ranges are explained in the Clocking, Sampling
Tree, and Power Scaling section. The data is filtered, scaled for
gain and offset (depending on user settings), and then output on
the data interface. Control of the flexible settings for the signal
chain is provided by either using the pin control or the SPI control
set at power-up by the state of the PIN/SPI input pin.
The AD7768/AD7768-4 can use up to a 5 V reference and con-
verts the differential voltage between the analog inputs (AINx+ and
AINx−) into a digital output. The analog inputs can be configured
as either differential or pseudo differential inputs. As a pseudo
differential input, either AINx+ or AINx− can be connected to a
constant input voltage (such as 0 V, GND, AVSS, or some other ref-
erence voltage). The ADC converts the voltage difference between Figure 82. ADC Ideal Transfer Functions (FS is Full Scale)
the analog input pins into a digital code on the output. Using a
Table 26. Output Codes and Ideal Input Voltages
common-mode voltage of AVDD1 ÷ 2 for the analog inputs, AINx+
and AINx−, maximizes the ADC input range. The 24-bit conversion Analog Input (AINx+ − Digital Output Code,
result is in twos complement, MSB first, format. Figure 82 shows Description (AINx−)) VREF = 4.096 V Twos Complement (Hex.)
the ideal transfer functions for the AD7768/AD7768-4. FS − 1 LSB 4.095999512 V 0x7FFFFF
Midscale + 1 LSB 488 nV 0x000001
ADC Power Modes Midscale 0V 0x000000
The AD7768/AD7768-4 have three selectable power modes. In pin Midscale − 1 LSB −488 nV 0xFFFFFF
control mode, the modulator rate and power mode are tied together −FS + 1 LSB −4.095999512 V 0x800001
for best performance. In SPI control mode, the user can select the −FS −4.096 V 0x800000
power mode and modulator MCLK divider settings. The choice of
power modes gives more flexibility to control the bandwidth and
ANALOG INPUTS
Figure 84 shows the AD7768/AD7768-4 analog front end. The
ESD protection diodes that are designed to protect the ADC from
some short duration overvoltage and ESD events are shown on
the signal path. The analog input is sampled at twice the fMOD,
which is derived from MCLK. By default, the ADC internal sampling
capacitors, CS1 and CS2, are driven by a per channel analog input
precharge buffer to ease the driving requirement of the external
network.
Figure 85. Analog Input Current (IAIN) vs. Input Voltage, Analog Input
Precharge Buffer Off, VCM = 2.5 V, fMOD = 8.192 MHz
For example, if the precharge buffers are off, with AIN1+ = 5 V, and used for the AD7768/AD7768-4 for most amplifier pairings. The
AIN1− = 0 V, estimate the current in each input pin as follows: RC network performs a variety of tasks. C1 and C2 are charge
reservoirs to the ADC, providing the ADC with fast charge current
AIN1+ = 5 V × 48 µA/V + 5 V × 17 µA/V = 325 µA to the sampling capacitors. Capacitor C3 removes common-mode
AIN1− = −5 V × 48 µA/V + 0 V × 17 µA/V = −240 µA errors between the AINx+ and AINx− inputs. These capacitors, in
combination with input resistance (RIN), form a low-pass filter to
When the precharge buffers are enabled, the absolute voltage filter out glitches related to the input switching. The input resistance
with respect to AVSS determines the majority of the current. The also stabilizes the amplifier when driving large capacitor loads and
maximum input current of approximately −25 µA is measured when prevents the amplifier from oscillating.
the analog input is close to either the AVDD1 or AVSS rails.
The optimum driver amplifiers for each of these power, perform-
With either precharge buffers enabled or disabled, the analog input ance, and supply requirements are as follows:
current scales linearly with the modulator clock rate. The analog
input current vs. input voltage is shown in Figure 85. ► The ADA4805-2 is suited for low power, particularly in low power
mode.
Full settling of the analog inputs to the ADC requires the use
► The ADA4940-1 is suited for single-supply operation and is also
of an external amplifier. Pair amplifiers such as the ADA4805-2
the recommended fully differential amplifier to drive the AD7768/
for low power mode, the ADA4807-2 or ADA4940-1/ADA4940-2 for
AD7768-4.
median mode, and the ADA4807-2 or ADA4896-2 for fast mode
► For optimum performance in fast power mode, the ADA4896-2
with the AD7768/AD7768-4 (see Figure 87 for details). Running the
AD7768/AD7768-4 in median and low power modes or reducing performs best, although the device does not consume the same
the MCLK rate reduces the load and speed requirements of the power as the ADA4899-1. The ADA4896-2 is also suitable for
amplifier. Therefore, lower power amplifiers can be paired with the a general-purpose DAQ module, which can be configured for all
analog inputs to achieve the optimum signal chain efficiency. three power modes.
There is a resistor/capacitor (RC) network between the amplifier For more details, refer to the AN-1384 Application Note.
output and the ADC input. Figure 87 shows a typical RC network
VCM
The AD7768/AD7768-4 provide a buffered common-mode voltage
output on Pin 59. This output can bias up analog input signals. By
incorporating the VCM buffer into the ADC, the AD7768/AD7768-4
reduce component count and board space. In pin control mode, the Figure 88. Typical Reference Input Configuration Diagram
VCM potential is fixed to (AVDD1 − AVSS)/2, and is enabled by
default. CLOCK SELECTION
In SPI control mode, configure the VCM potential using the general The AD7768/AD7768-4 have an internal oscillator that is used for
configuration register (Register 0x05). The output can be enabled initial power-up of the device. After the AD7768/AD7768-4 have
or disabled, and set to (AVDD1 − AVSS)/2, 1.65 V, 2.14 V, or 2.5 V, completed their start-up routine, the devices normally transfer con-
with respect to AVSS. trol of the internal clocking to the externally applied MCLK. The
The VCM voltage output is associated with the Channel 0 circuitry. AD7768/AD7768-4 count the falling edges of the external MCLK
If Channel 0 is put into standby mode, the VCM voltage output is over a given number of internal clock cycles to determine if the
also disabled for maximum power savings. Channel 0 must be ena- clock is valid and at least a frequency of 1.15 MHz. If there is a
bled while VCM is being used externally to the AD7768/AD7768-4. fault with the external MCLK, the transfer of control does not occur,
the AD7768/AD7768-4 output an error in the status header, and the
REFERENCE INPUT clock error bit is set in the device status register. No conversion
The AD7768/AD7768-4 have two differential reference input pairs. data is output and a reset is required to exit this error state.
On the AD7768 REF1+ and REF1− are the reference inputs for Three clock source input options are available to the AD7768/
Channel 0 to Channel 3, and REF2+ and REF2− are for Channel AD7768-4: external CMOS, crystal oscillator, or LVDS. The clock
4 to Channel 7. On the AD7768-4 REF1+ and REF1− are the is selected on power-up and is determined by the state of the
reference inputs for Channel 0 and Channel 1, and REF2+ and CLK_SEL pin.
REF2− are for Channel 2 and Channel 3. The absolute input
reference voltage range is 1 V to AVDD1 − AVSS. If CLK_SEL = 0, the CMOS clock option is selected and the clock
is applied to Pin 32 (Pin 31 is tied to DGND). Follow the conditions
Like the analog inputs, the reference inputs have a precharge below when using CMOS clock:
buffer option. Each ADC has an individual buffer for each REFx+
and REFx−. The precharge buffers help reduce the burden on the ► CLK_SEL (Pin 58) must be set to 0.
external reference circuitry. ► Connect XTAL1 (Pin 31) to DGND.
► The CMOS clock is applied to Pin 32.
In pin control mode, the reference precharge buffers are off by
default. In SPI control mode, the user can enable or disable the ref- ► The CMOS logic level in this case is between IOVDD and
erence precharge buffers. In the case of unipolar analog supplies, DGND.
in SPI control mode, the user can achieve the best performance If CLK_SEL = 1, the crystal or LVDS option is selected and the
and power efficiency by enabling only the REFx+ buffers. The crystal or LVDS is applied to Pin 31 and Pin 32. The LVDS option
reference input current scales linearly with the modulator clock rate. is available only in SPI control mode. Channel 4 on the AD7768
For 32 MHz MCLK and MCLK/4 fast mode, the differential input must be enabled in the channel standby register to use the crystal
current is ~72 µA/V per channel unbuffered, and ~16 µA/V per because this is linked to the crystal excitation circuitry. On the
channel with the precharge buffers enabled. AD7768-4, Channel 2 must be enabled in the channel standby
register.
With the precharge buffers off, REFx+ = 5 V, and REFx− = 0 V,
To use an external crystal, it is necessary to adhere to the following
REFx± = 5 V × 72 µA/V = 360 µA steps:
With the precharge buffers on, REFx+ = 5 V, and REFx− = 0 V, 1. Channel 4 of the ADC must be enabled. This is because Chan-
REFx± = 5 V × 16 µA/V = 80 µA nel 4 is linked to the crystal excitation circuitry. Powering down
this channel also powers down the crystal excitation circuitry.
For the best performance and headroom, it is recommended to use 2. The CLK_SEL pin must be set to Logic 1.
a 4.096 V reference such as the ADR444 or the ADR4540. 3. The crystal is applied between Pin 31 and Pin 32. Capacitance
For the best performance at high sampling rates, it is recommended of approximately 20 pF is required on each of these pins. This
to use an external reference drive amplifier such as the ADA4841-1 may vary depending on the crystal that is selected.
or the AD8031. See Figure 88 for the configuration diagram of the 4. If the AD7768/AD7768-4 is in SPI control mode then the FIL-
reference connection. TER/GPIO4 pin (Pin 11) must be set to Logic 1.
To enable the LVDS clock, there are two options. Set GPIO4
to an output, then writing to the LVDS bit field in the POWER_
CLOCK register enables the LVDS clock. Or, set GPIO4 to an
input. Then GPIO4 must be tied to Logic 0. An SPI write to Bit 3
of Register 0x04 enables the LVDS clock option and disables the
crystal excitation circuitry.
The following steps must be carried out to enable LVDS clocking to
be used:
1. Set the CLK_SEL pin to Logic 1.
2. Set the FILTER/GPIO4 pin (Pin 11) to Logic 0. This disables the
crystal excitation circuitry.
3. Power up the AD7768 with no external clock applied. Initially, it
will run off the internal clock for SPI writes until Step 4 and Step
5 have been carried out. Figure 89. Sinc5 Filter Frequency Response (Decimation = ×32)
4. Enable this option by writing to the LVDS_Enable bit (Bit 3) in
the Power_Mode register (Register 0x04). The settling times for the AD7768/AD7768-4 when using the sinc5
5. Apply the LVDS signals on Pin 31 and Pin 32. filter are shown in Figure 89.
Delay from First MCLK Rise Delay from First MCLK Rise After SYNC_IN Rise
After SYNC_IN Rise to First to Earliest Settled Data DRDY Rise
Table 28. Wideband Filter SYNC_IN to Settled Data (DCLK = MCLK) (Continued)
Delay from First MCLK Rise Delay from First MCLK Rise After SYNC_IN Rise
After SYNC_IN Rise to First to Earliest Settled Data DRDY Rise
Table 29. Sinc5 Filter SYNC_IN to Settled Data (DCLK = MCLK) (Continued)
Delay from First MCLK Rise Delay from First MCLK Rise After SYNC_IN Rise to
After SYNC_IN Rise to First Earliest Settled Data DRDY Rise
decimation rate), input frequencies, and in particular harmonics of chop aliasing, fMOD/16 for modulator saturation, and 2 × fMOD for the
input frequencies, that may fall close to fODR/2, do not fold back into first zone with 0 dB attenuation. It assumes the corner frequency of
the pass band of the AD7768/AD7768-4. the antialiasing filter is at fMOD/64, which is just above the maximum
input bandwidth that the AD7768/AD7768-4 digital filter can pass
when using a decimate by 32 filter setting.
Table 31. External Antialiasing Filter Attenuation
RC Filter fMOD/32 (dB) fMOD/16 (dB) fMOD/8 (dB) 2 × fMOD (dB)
First Order −6 −12 −18 −42
Second Order −12 −24 −36 −84
Third Order −18 −36 −54 −126
SETTING THE FORMAT OF DATA OUTPUT However, there is a trade-off against ADC offset performance with
higher DCLK frequencies. For the best offset and offset drift per-
The data interface format is determined by setting the FORMATx formance, use the lowest DCLK frequency possible. The user can
pins. The logic state of the FORMATx pins are read on power-up choose to reduce the DCLK frequency by an appropriate selection
and determine how many data lines (DOUTx) the ADC conversions of MCLK frequency, DCLK divider, and/or the number of DOUTx
are output on. lines used. Table 1 and Table 2 give the offset and offset drift
Because the FORMATx pins are read on power-up of the AD7768 specifications for ranges of DCLK frequency, and Figure 49 shows
and the device remains in this output configuration, this function the typical offset drift over a range of DCLK frequencies.
must always be hardwired and cannot be altered dynamically. Table 33. FORMATx Truth Table for the AD7768
Table 33, Figure 95, Figure 96, and Figure 98 show the formatting FORMAT1 FORMAT0 Description
configuration for the digital output pins on the AD7768.
0 0 Each ADC channel outputs on its own dedicat-
Calculate the minimum required DCLK rate for a given data inter- ed pin. DOUT0 to DOUT7 are in use.
face configuration as follows: 0 1 The ADCs share the DOUT0 and DOUT1 pins:
DCLK (Minimum) = Output Data Rate × Channels per DOUTx × 32 Channel 0 to Channel 3 output on DOUT0.
Channel 4 to Channel 7 output on DOUT1. The
where MCLK ≥ DCLK. ADC channels share data pins in time division
For example, if MCLK = 32.768 MHz, with two DOUTx lines, multiplexed (TDM) output. DOUT0 and DOUT1
are in use.
DCLK (Minimum) = 256 kSPS × 4 channels per DOUTx × 32 = 1 X All channels output on the DOUT0 pin, in TDM
32.768 Mbps output. Only DOUT0 is in use.
Therefore, DCLK = MCLK/1. Table 34. FORMAT0 Truth Table for the AD7768-4
Alternatively, if MCLK = 32.768 MHz, with eight DOUTx lines, FORMAT0 Description
DCLK (Minimum) = 256 kSPS × 1 channel per DOUTx × 32 = 8.192 0 Each ADC channel outputs on its own dedicated pin.
Mbps DOUT0 to DOUT3 are in use.
1 All channels output on the DOUT0 pin, in TDM output. Only
Therefore, DCLK = MCLK/4. DOUT0 is in use.
Higher DCLK rates make it easier to receive the conversion data
from the AD7768/AD7768-4 with a lower number of DOUTx lines.
Figure 98. AD7768 FORMATx = 10 or 11, or AD7768-4 FORMAT0 = 1, One Data Output Pin
ADC CONVERSION OUTPUT: HEADER AND unexpectedly changed state, or an internal CRC error has been
DATA detected.
The AD7768 data is output on the DOUT0 pin to DOUT7 pin, In the case where an external clock is not detected, the conversion
depending on the FORMATx pins. The AD7768-4 data is output results are output as all zeros regardless of the analog input
on the DOUT0 pin to DOUT3 pin, depending on the FORMAT0 voltages applied to the ADC channels.
pin. The actual structure of the data output for each ADC result
is shown in Figure 99. Each ADC result is comprised of 32 bits. Filter Not Settled
The first eight bits are the header status bits, which contain status After power-up, reset, or synchronization, the AD7768/AD7768-4
information and the channel number. The names of each of the clear the digital filters and begins conversion. Due to the weighting
header status bits are shown in Table 35, and their functions are of the digital filters, there is a delay from the first conversion to fully
explained in the subsequent sections. This header is followed by a settled data. The settling times for the AD7768/AD7768-4 when
24-bit ADC output in twos complement coding, MSB first. using the wideband and sinc5 filters are shown in Table 28 and
Table 29, respectively. This bit is set if this settling delay has not yet
elapsed.
Repeated Data
If different channels use different decimation rates, data outputs
Figure 99. ADC Output: 8-Bit Header, 24-Bit ADC Conversion Data are repeated for the slower speed channels. In these cases, the
header is output as normal with the repeated data bit set to 1,
Table 35. Header Status Bits
and the following repeated ADC result is output as all zeros. This
Bit Bit Name bit indicates that the conversion result of all zeros is not real. The
7 ERROR_FLAGGED bit indicates that there is a repeated data condition because two
6 Filter not settled different decimation rates are selected. This condition can only
occur during SPI control of the AD7768/AD7768-4.
5 Repeated data
4 Filter type
Filter Type
3 Filter saturated
[2:0] Channel ID[2:0] In pin control mode, all channels operate using one filter selection.
The filter selected in pin control mode is determined by the logic
level of the FILTER pin. In SPI control mode, the digital filters can
ERROR_FLAGGED
be selected on a per channel basis using the mode registers. This
The error flagged bit indicates that a serious error has occurred. If header bit is 0 for channels using the wideband filter, and 1 for
this bit is set, a reset is required to clear this bit. This bit indicates channels using the sinc5 filter.
that the external clock is not detected, a memory map bit has
Filter Saturated Each DRDY falling edge starts the output of the new ADC conver-
sion data. The first eight bits output after the DRDY falling edge are
The filter saturated bit indicates that the filter output is clipping at the header bits. The last 24 bits are the ADC conversion result.
either positive or negative full scale. The digital filter clips if the
signal goes beyond the specification of the filter, it does not wrap. Figure 100, Figure 101, Figure 102, and Figure 103 are distinct
The clipping may be caused by the analog input exceeding the examples of the impact of the FORMATx pins on the AD7768 out-
analog input range, or by a step change in the input, which may put operating in standard conversion operation. Figure 104, Figure
cause overshoot in the digital filter. Clipping may also occur when 105, and Figure 106 show examples of the AD7768-4 interface
the combination of the analog input signal and the channel gain configuration.
register setting cause the signal seen by the filter to be higher than Figure 100 through Figure 103 represent running the AD7768 at
the analog input range. maximum data rate for the three FORMATx options.
Channel ID Figure 100 shows FORMATx = 00 and each ADC has its own
data out (DOUT) pin running at the MCLK/4 bit rate. In pin control
The channel ID bits indicate the ADC channel from which the mode, this is achieved by selecting Mode 0xA (fast mode, DCLK
succeeding conversion data originates (see Table 36). = MCLK/4, standard conversion, see Table 20) with the decimation
Table 36. Channel ID vs. Channel Number rate set as ×32.
Channel Channel ID 2 Channel ID 1 Channel ID 0 Figure 101 shows FORMATx = 01 share DOUT1 at the maximum
Channel 0 0 0 0 bit rate. In pin control mode, this is achieved by selecting Mode 0x8
Channel 1 0 0 1 (fast mode, DCLK = MCLK/1, standard conversion) with a decima-
Channel 2 0 1 0
tion rate of ×32.
Channel 3 0 1 1 If running in pin control mode, the example shown in Figure 103
Channel 4 1 0 0 represents Mode 0x4 (median mode, DCLK = MCLK/1, standard
Channel 5 1 0 1 conversion) with a decimation rate of ×32, giving the maximum
output data capacity possible on one DOUTx pin.
Channel 6 1 1 0
Channel 7 1 1 1 Figure 102 (AD7768) and Figure 106 (AD7768-4) show examples
of one configuration where there can be long periods in which
Data Interface: Standard Conversion Operation no data is output by the AD7768. This configuration depends on
the FORMATx, MCLK, and decimation settings. In Figure 102,
In standard mode operation, the AD7768/AD7768-4 operate as the FORMATx = 01, meaning the channels share DOUT0 and DOUT1.
main and stream data to the DSP or FPGA. The AD7768/AD7768-4 In Figure 106, FORMAT0 = 1, meaning all channels share the
supply the data, the data clock (DCLK), and a falling edge framing DOUT0 pin. For both Figure 102 and Figure 106, DCLK = MCLK/4
signal (DRDY) to the subordinate device. All of these signals are and the decimation rate is 512. In pin control mode, this setup is
synchronous. The data interface connections to DSP/FPGA are achieved by selecting Mode 0x0A (fast mode, DCLK = MCLK/4,
shown in Figure 107. The FORMATx pins determine how the data standard conversion mode). With a decimation rate of 512, the ratio
is output from the AD7768/AD7768-4. of ODR to DCLK rate is high enough to show that only ¼ of the or
ODR period is used with output data, and the other ¾ of the period
Figure 100 through Figure 103 show the data interface operating in
DOUTx is low.
standard mode at the maximum data rate. In all instances, DRDY is
asserted one clock cycle before the MSB of the data conversion is
made available on the data pin.
Figure 100. AD7768 FORMATx = 00: Each ADC Has a Dedicated Data Output Pin, Maximum Data Rate
Figure 101. AD7768 FORMATx = 01: Channel 0 to Channel 3 Share DOUT0, and Channel 4 to Channel 7 Share DOUT1, Maximum Data Rate
Figure 102. AD7768 FORMATx = 01: Channel 0 to Channel 3 Share DOUT0, and Channel 4 to Channel 7 Share DOUT1, Decimation = 512
Figure 103. AD7768 FORMATx = 11 or 10: Channel 0 to Channel 7 Output on DOUT0 Only, Maximum Data Rate
Figure 104. AD7768-4 FORMAT0 = 0: Each ADC Has a Dedicated Data Output Pin, Maximum Data Rate
Figure 105. AD7768-4 FORMAT0 = 1: Channel 0 to Channel 3 Output on DOUT0 Only, Maximum Data Rate
Figure 106. AD7768-4 FORMAT0 = 1: Channel 0 to Channel 3 Output on DOUT0 Only, Decimation = 512
Figure 107. Data Interface: Standard Conversion Operation, AD7768 = Main, DSP/FPGA = Subordinate
Data Interface: One-Shot Conversion Operation Mode 0xF when in pin control mode. In SPI control mode, set Bit
4 (one shot) of Register 0x06, the data control register. Figure 108
One shot mode is available in both SPI and pin control modes. shows the device operating in one shot mode.
This conversion mode is available by selecting one of Mode 0xC to
In one shot mode, the AD7768/AD7768-4 are pseudo subordinates. For the AD7768/AD7768-4, this connection can be implemented by
Conversions occur on request by the main device, for example, the cascading DOUT0 and DOUT1 through a number of devices, or
DSP or FPGA. The SYNC_IN pin initiates the conversion request. just using DOUT0. Whether two data output pins or only one data
In one shot mode, all ADCs run continuously. However, the rising output pin is enabled depends on the FORMATx pins. The ability to
edge of the SYNC_IN pin controls the point in time from which data daisy-chain devices and the limit on the number of devices that can
is output. be handled by the chain is dependent on the power mode, DCLK,
and the decimation rate employed.
To receive data, the main must pulse the SYNC_IN pin to reset
the filter and force DRDY low. DRDY subsequently goes high The maximum usable DCLK frequency allowed when daisy-chain-
to indicate to the main device that the device has valid settled ing devices is limited by the combination of timing specifications in
data available. Unlike standard mode, DRDY remains high for the Table 3 or Table 5, as well as by the propagation delay of the data
number of clock periods of valid data before it goes low again. between devices and any skew between the MCLK signals at each
Therefore, in this conversion mode, it is an active high frame of the AD7768/AD7768-4 device. The propagation delay and MCLK skew
data. are dependent on the PCB layout and trace lengths.
When the main pulses SYNC_IN and the AD7768/AD7768-4 re- This feature is especially useful for reducing component count and
ceive the rising edge of this signal, the digital filter is reset and the wiring connections, for example, in isolated multiconverter applica-
full settling time of the filter elapses before the data is available. tions or for systems with a limited interfacing capacity.
The duration of the settling time depends on the filter path and
decimation rate. Running one-shot mode with the sinc5 filter allows When daisy-chaining, on the AD7768, DOUT6 and DOUT7 become
the fastest throughput, because this filter has a lower settling time serial data inputs, and DOUT0 and DOUT1 remain as serial data
than the wideband filter. outputs under the control of the FORMATx pins. For the AD7768-4
the DIN pin is the daisy chain serial data input pin and DOUT0 is
As soon as settled data is available on any channel, the device the serial data output pin.
outputs data from all channels. The contents of Bit 6 of the channel
header status bits indicates whether the data is fully settled.
The period before the data is settled on all channels (tSETTLE) is
shown in Figure 108. The settling time (tSETTLE) for the AD7768 in
one shot mode is equivalent to the number of clock cycles specified
as Delay from the First MCLK Rise after SYNC_IN Rise to Earliest
Settled Data, DRDY Rise in Table 30. After the data has settled
on all channels, DRDY is asserted high and the device outputs the
required settled data on all channels before DRDY is asserted low.
If the user configures the same filter and decimation rate on each
ADC, the data is settled for all channels on the first DRDY output
frame, which avoids a period of unsettled data prior to the settled
data and ensures that all data is output at the same time on all
ADCs. The device then waits for another SYNC_IN signal before
outputting more data.
Figure 109. Daisy-Chaining Multiple AD7768 Devices
Because all the ADCs are sampling continuously, one shot mode
affects the sampling theory of the AD7768/AD7768-4. Particularly, Figure 109 shows an example of daisy-chaining AD7768 devices
a user periodically sending a SYNC_IN pulse to the device is a when FORMATx = 01. In this case, the DOUT0 and DOUT1 pins
form of subsampling of the ADC output. The subsampling occurs of the AD7768 devices are cascaded to the DOUT6 and DOUT7
at the rate of the SYNC_IN pulses. The SYNC_IN pulse must be pins, respectively, of the next device in the chain. Data readback is
synchronous with the main clock to ensure coherent sampling and analogous to clocking a shift register where data is clocked on the
to reduce the effects of jitter on the frequency response. rising edge of DCLK.
The scheme operates by passing the output data of the DOUT0
Daisy-Chaining pin and DOUT1 pin of an AD7768 upstream device to the DOUT6
Daisy-chaining devices allows numerous devices to use the same and DOUT7 inputs, respectively, of the next AD7768 device down-
data interface lines by cascading the outputs of multiple ADCs from stream in the chain. The data then continues through the chain
separate AD7768/AD7768-4 devices. Only one ADC device has its until it is clocked onto the DOUT0 pin and DOUT1 pin of the final
data interface in direct connection with the digital host. downstream device in the chain.
The devices in the chain must be synchronized by using one of the
following methods:
analog.com Rev. D | 73 of 108
Data Sheet AD7768/AD7768-4
DATA INTERFACE
► Applying a synchronous signal to the SYNC_IN pin of all devices chronization pulse that is truly synchronous with the base MCLK
in the chain signal.
► By routing the SYNC_OUT pin of the first device to the SYNC_IN
Two synchronization pulses are required in a system of more than
pin of that same device and to the SYNC_IN pins of all other one AD7768/AD7768-4 device sharing a single MCLK signal, to
devices in the chain and applying an asynchronous signal to the ensure that all devices are in close phase alignment, or where the
START input. DRDY pin of only one device is used to detect new data.
► Issuing an SPI_SYNC command over the SPI control interface.
If the user cannot provide a signal that is synchronous to the base
Figure 109 shows the configuration where an asynchronous signal MCLK signal, one of the following two methods can be employed:
is applied to the START pin, and the SYNC_OUT pin of the first
device is connected to the SYNC_IN pins of all devices in the chain ► Apply a START pulse to the first AD7768 or AD7768-4 device.
The first AD7768 or AD7768-4 device samples the asynchronous
Daisy-chaining can be achieved in a similar manner on the START pulse and generates a pulse on SYNC_OUT of the first
AD7768/AD7768-4 when using only the DOUT0 pin. In this case, device related to the base MCLK signal for distribution locally.
only Pin 21 of the AD7768/AD7768-4 is used as the serial data ► Use synchronization over SPI (only available in SPI control
input pin. mode) to write a synchronization command to the first AD7768
In a daisy-chained system of AD7768/AD7768-4 devices, two suc- or AD7768-4 device. Similarly to the START pin method, the SPI
cessive synchronization pulses must be applied to guarantee that sync generates a pulse on SYNC_OUT of the first device related
all ADCs are synchronized. It is recommended to wait at least 16 to the base MCLK signal for distribution locally.
MCLK pulses between issuing the first and second synchronization In both cases, route the SYNC_OUT pin of the first device to the
pulses. Two synchronization pulses are also required in a system SYNC_IN pin of that same device and to the SYNC_IN pins of all
of more than one AD7768/AD7768-4 device sharing a single MCLK other devices that are to be synchronized (see Figure 110). The
signal, where the DRDY pin of only one device is used to detect SYNC_OUT pins of the other devices must remain open circuit. Tie
new data. all unused START pins to a Logic 1 through pull-up resistors.
The maximum DCLK frequency that can be used when daisy-chain-
ing devices is a function of the AD7768/AD7768-4 timing specifica-
tions (t4 and t11 in Table 3 and Table 5) and any timing differences
between the AD7768/AD7768-4 devices due to layout and spacing
of devices on the PCB.
Use the following formula to aid in determining the maximum
operating frequency of the interface:
1
fMAX = 2 × (t11 + t4 + tP + tSKEW) (2)
where:
fMAX is the maximum useable DCLK frequency.
t11 and t4 are the AD7768/AD7768-4 timing specifications (see
Table 3 and Table 5).
tP is the maximum propagation delay of the data between succes-
Figure 110. Synchronizing Multiple AD7768/AD7768-4 Devices Using
sive AD7768/AD7768-4 devices in the chain.
SYNC_OUT
tSKEW is the maximum skew in the MCLK signal seen by any pair of
AD7768/AD7768-4 devices in the chain. If the user can provide a signal that is synchronous to the base
MCLK, this signal can be applied directly to the SYNC_IN pin.
Synchronization Route the signal from a star point and connect it directly to the
The basic provision for synchronizing multiple devices is that each SYNC_IN pin of each AD7768/AD7768-4 device (see Figure 111).
device is clocked with the same base MCLK signal and that the The signal is sampled on the rising MCLK edge. Setup and hold
user can provide a synchronization signal to at least one of the times are associated with the SYNC_IN input and are relative to the
devices by one of the methods described in this section. AD7768/AD7768-4 MCLK rising edge.
The AD7768/AD7768-4 offer three options to allow ease of system In this case, tie the START pin to Logic 1 through a pull-up resistor.
synchronization. Choosing between the options depends on the SYNC_OUT is not used and can remain open circuit.
system, but is determined by whether the user can supply a syn-
GPIO FUNCTIONALITY
The AD7768/AD7768-4 have additional GPIO functionality when
operated in SPI mode. This fully configurable mode allows the
device to operate five GPIOs. The GPIOx pins can be set as inputs
or outputs (read or write) on a per pin basis.
In write mode, these GPIO pins can be used to control other circuits
such as switches, multiplexers, buffers, over the same SPI interface
as the AD7768/AD7768-4. Sharing the SPI interface in this way
allows the user to use a lower overall number of data lines from
the controller compared to a system where multiple control signals
are required. This sharing is especially useful in systems where
reducing the number of control lines across an isolation barrier is
important. See Figure 113 and Figure 114 for details of the GPIO
pin options available on the AD7768 and AD7768-4, respectively. Figure 114. AD7768-4 GPIO Functionality
Similarly, a GPIO read is a useful feature because it allows a
Configuration control and readback of the GPIOx pins are set in
peripheral device to send information to the input GPIO and then
Register 0x0E, Register 0x0F, and Register 0x10 (see Table 49,
this information can be read from the SPI interface of the AD7768/
Table 50, and Table 51 for more information for the AD7768, and
AD7768-4.
Table 75, Table 76, and Table 77 for the AD7768-4).
STATUS REGISTER
Address: 0x09, Reset: 0x00, Name: Device Status
Table 55. Bit Descriptions for Negative Reference Precharge Buffer (Continued)
Bits Bit Name Settings Description Reset
6 CH6_REFN_BUF 0 Off 0x0
1 On
5 CH5_REFN_BUF 0 Off 0x0
1 On
4 CH4_REFN_BUF 0 Off 0x0
1 On
3 CH3_REFN_BUF 0 Off 0x0
1 On
2 CH2_REFN_BUF 0 Off 0x0
1 On
1 CH1_REFN_BUF 0 Off 0x0
1 On
0 CH0_REFN_BUF 0 Off 0x0
1 On
OFFSET REGISTERS
The CHx_OFFSET_MSB, CHx_OFFSET_MID, and CHx_OFFSET_LSB registers at addresses 0x1E to 0x35 form 24-bit signed twos
complement registers for channel offset adjustment. If the channel gain setting is at its ideal nominal value of 0x555555, an LSB of offset
register adjustment changes the digital output by −4/3 LSBs. For example, changing the offset register from 0 to 100 changes the digital output
by −133 LSBs. As offset adjustment occurs before gain adjustment, the ratio of 4/3 changes linearly with gain adjustment via the CHx_GAIN_x
registers. After a reset or power cycle, the register values revert to the default factory setting.
Table 56. Per Channel 24-Bit Offset Registers, Three 8-Bit Registers for Each Channel, Split Up as MSB, MID, and LSB
Address Reset
MSB Mid LSB Name Description MSB Mid LSB Access
0x1E 0x1F 0x20 Channel 0 Offset Channel 0 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x21 0x22 0x23 Channel 1 Offset Channel 1 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x24 0x25 0x26 Channel 2 Offset Channel 2 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x27 0x28 0x29 Channel 3 Offset Channel 3 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x2A 0x2B 0x2C Channel 4 Offset Channel 4 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x2D 0x2E 0x2F Channel 5 Offset Channel 5 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x30 0x31 0x32 Channel 6 Offset Channel 6 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x33 0x34 0x35 Channel 7 Offset Channel 7 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
GAIN REGISTERS
Each ADC channel has an associated gain coefficient. The coefficient is stored in three single-byte registers split up as MSB, MID, and LSB.
Each of the gain registers are factory programmed. Nominally, this gain is around the value 0x555555 (for an ADC channel). The user may
overwrite the gain register setting. However, after a reset or power cycle, the gain register values revert to the hard coded programmed factory
setting.
Table 57. Per Channel 24-Bit Gain Registers, 3 8-Bit Registers for Each Channel, Split Up as MSB, MID, and LSB
Address Reset
MSB Mid LSB Name Description MSB Mid LSB Access
0x36 0x37 0x38 Channel 0 Gain Channel 0 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x39 0x3A 0x3B Channel 1 Gain Channel 1 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
Table 57. Per Channel 24-Bit Gain Registers, 3 8-Bit Registers for Each Channel, Split Up as MSB, MID, and LSB (Continued)
Address Reset
MSB Mid LSB Name Description MSB Mid LSB Access
0x3C 0x3D 0x3E Channel 2 Gain Channel 2 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x3F 0x40 0x41 Channel 3 Gain Channel 3 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x42 0x43 0x44 Channel 4 Gain Channel 4 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x45 0x46 0x47 Channel 5 Gain Channel 5 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x48 0x49 0x4A Channel 6 Gain Channel 6 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x4B 0x4C 0x4D Channel 7 Gain Channel 7 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
STATUS REGISTER
Address: 0x09, Reset: 0x00, Name: Device Status
Table 81. Bit Descriptions for Negative Reference Precharge Buffer (Continued)
Bits Bit Name Settings Description Reset
1 CH1_REFN_BUF 0 Off 0x0
1 On
0 CH0_REFN_BUF 0 Off 0x0
1 On
OFFSET REGISTERS
The CHx_OFFSET_MSB, CHx_OFFSET_MID, and CHx_OFFSET_LSB registers in addresses 0x1E to 0x35 form 24-bit, signed twos
complement registers for channel offset adjustment. If the channel gain setting is at its ideal nominal value of 0x555555, an LSB of offset
register adjustment changes the digital output by −4/3 LSBs. For example, changing the offset register from 0 to 100 changes the digital output
by −133 LSBs. As offset adjustment occurs before gain adjustment, the ratio of 4/3 changes linearly with gain adjustment via register addresses
0x36 to 0x4D. After a reset or power cycle, the register values revert to the default factory setting.
Table 82. Per Channel 24-Bit Offset Registers, Three 8-Bit Registers for Each Channel, Split Up as MSB, MID, and LSB
Address Reset
MSB Mid LSB Name Description MSB Mid LSB Access
0x1E 0x1F 0x20 Channel 0 Offset Channel 0 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x21 0x22 0x23 Channel 1 Offset Channel 1 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x2A 0x2B 0x2C Channel 2 Offset Channel 2 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
0x2D 0x2E 0x2F Channel 3 Offset Channel 3 offset registers: upper, middle, and lower bytes (24 bits in total) 0x00 0x00 0x00 RW
GAIN REGISTERS
Each ADC channel has an associated gain coefficient. The coefficient is stored in three single-byte registers split up as MSB, MID, and LSB.
Each of the gain registers are factory programmed. Nominally, this gain is around the value 0x555555 (for an ADC channel). The user may
overwrite the gain register setting. However, after a reset or power cycle, the gain register values revert to the hard coded programmed factory
setting.
Table 83. Per Channel 24-Bit Gain Registers, Three 8-Bit Registers for Each Channel, Split Up as MSB, MID, and LSB
Address Reset
MSB Mid LSB Name Description MSB Mid LSB Access
0x36 0x37 0x38 Channel 0 Gain Channel 0 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x39 0x3A 0x3B Channel 1 Gain Channel 1 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x42 0x43 0x44 Channel 2 Gain Channel 2 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
0x45 0x46 0x47 Channel 3 Gain Channel 3 gain registers: upper, middle, and lower bytes (24 bits in total) 0xXX 0xXX 0xXX RW
For the latest package outline information and land patterns (footprints), go to Package Index.
Updated: July 14, 2022
ORDERING GUIDE
Model1 Temperature Range Package Description Packing Quantity Package Option
AD7768-4BSTZ −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Tray, 160 ST-64-2
AD7768-4BSTZ-RL −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Reel, 1500 ST-64-2
AD7768-4BSTZ-RL7 −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Reel, 300 ST-64-2
AD7768BSTZ −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Tray, 160 ST-64-2
AD7768BSTZ-RL −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Reel, 1500 ST-64-2
AD7768BSTZ-RL7 −40°C to +105°C 64-Lead LQFP (10 mm × 10 mm) Reel, 300 ST-64-2
1 Z = RoHS Compliant Part.
EVALUATION BOARDS
Model1 Description
EVAL-AD7768FMCZ Evaluation Board
EVAL-AD7768-4FMCZ AD7768-4 Evaluation Board
EVAL-SDP-CH1Z Controller Board
©2016-2025 Analog Devices, Inc. All rights reserved. Trademarks and Rev. D | 108 of 108
registered trademarks are the property of their respective owners.
One Analog Way, Wilmington, MA 01887-2356, U.S.A.