AD AD9695BCPZRL7-1300 Dual analog-to-digital converter Datasheet

14-Bit, 1300 MSPS/625 MSPS, JESD204B,
Dual Analog-to-Digital Converter
AD9695
Data Sheet
FEATURES
APPLICATIONS
JESD204B (Subclass 1) coded serial digital outputs
Lane rates up to 16 Gbps
1.6 W total power at 1300 MSPS
800 mW per ADC channel
SNR = 65.6 dBFS at 172 MHz (1.59 V p-p input range)
SFDR = 78 dBFS at 172.3 MHz (1.59 V p-p input range)
Noise density
−153.9 dBFS/Hz (1.59 V p-p input range)
−155.6 dBFS/Hz (2.04 V p-p input range)
0.95 V, 1.8 V, and 2.5 V supply operation
No missing codes
Internal ADC voltage reference
Flexible input range
1.36 V p-p to 2.04 V p-p (1.59 V p-p typical)
2 GHz usable analog input full power bandwidth
>95 dB channel isolation/crosstalk
Amplitude detect bits for efficient AGC implementation
2 integrated digital downconverters per ADC channel
48-bit NCO
Programmable decimation rates
Differential clock input
SPI control
Integer clock divide by 2 and divide by 4
Flexible JESD204B lane configurations
On-chip dithering to improve small signal linerarity
Communications
Diversity multiband, multimode digital receivers
3G/4G, TD-SCDMA, WCDMA, GSM, LTE
General-purpose software radios
Ultrawideband satellite receiver
Instrumentation
Oscilloscopes
Spectrum analyzers
Network analyzers
Integrated RF test solutions
Radars
Electronic support measures, electronic counter measures,
and electronic counter-counter measures
High speed data acquisition systems
DOCSIS 3.0 CMTS upstream receive paths
Hybrid fiber coaxial digital reverse path receivers
Wideband digital predistortion
FUNCTIONAL BLOCK DIAGRAM
ADC
CORE
VREF
ADC
CORE
BUFFER
SPIVDD
(1.8V)
14
DIGITAL DOWNCONVERTER
DIGITAL DOWNCONVERTER
JESD204B
LINK
AND
Tx
OUTPUTS
SERDOUT0±
SERDOUT1±
SERDOUT2±
SERDOUT3±
SYNCINB±
PDWN/STBY
SYSREF±
4
JESD204B
SUBCLASS 1
CONTROL
CLOCK
DISTRIBUTION
FD_A/GPIO_A0
GPIO MUX
CLK+
CLK–
÷2
SPI AND
CONTROL
REGISTERS
FD_B/GPIO_B0
AD9695
÷4
AGND
SDIO SCLK
CSB
DRGND
DGND
15660-001
VIN+B
VIN–B
DRVDD2
(1.8V)
14
SIGNAL
MONITOR
FAST
DETECT
DRVDD1
(0.95V)
CROSSBAR MUX
BUFFER
DVDD
(0.95V)
CROSSBAR MUX
VIN+A
VIN–A
AVDD3 AVDD1_SR
(0.95V)
(2.5V)
AVDD2
(1.8V)
PROGRAMMABLE
FIR FILTER
AVDD1
(0.95V)
Figure 1.
Rev. 0
Document Feedback
Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other
rights of third parties that may result from its use. Specifications subject to change without notice. No
license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
Trademarks and registered trademarks are the property of their respective owners.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700
©2017 Analog Devices, Inc. All rights reserved.
Technical Support
www.analog.com
AD9695* PRODUCT PAGE QUICK LINKS
Last Content Update: 09/22/2017
COMPARABLE PARTS
DESIGN RESOURCES
View a parametric search of comparable parts.
• AD9695 Material Declaration
• PCN-PDN Information
EVALUATION KITS
• Quality And Reliability
• AD9695 Evaluation Board
• Symbols and Footprints
DOCUMENTATION
DISCUSSIONS
Data Sheet
View all AD9695 EngineerZone Discussions.
• AD9695: 14-Bit, 1300 MSPS/625 MSPS, JESD204B, Dual
Analog-to-Digital Converter Data Sheet
SAMPLE AND BUY
TOOLS AND SIMULATIONS
Visit the product page to see pricing options.
• AD9695 IBIS Model
TECHNICAL SUPPORT
• AD9695 S-Parameters
Submit a technical question or find your regional support
number.
DOCUMENT FEEDBACK
Submit feedback for this data sheet.
This page is dynamically generated by Analog Devices, Inc., and inserted into this data sheet. A dynamic change to the content on this page will not
trigger a change to either the revision number or the content of the product data sheet. This dynamic page may be frequently modified.
AD9695
Data Sheet
TABLE OF CONTENTS
Features .............................................................................................. 1
DDC General Description ........................................................ 42
Applications ....................................................................................... 1
DDC Frequency Translation ..................................................... 45
Functional Block Diagram .............................................................. 1
DDC Decimation Filters ........................................................... 53
Revision History ............................................................................... 3
DDC Gain Stage ......................................................................... 60
General Description ......................................................................... 4
DDC Complex to Real Conversion ......................................... 61
Product Highlights ........................................................................... 4
DDC Mixed Decimation Settings ............................................ 62
Specifications..................................................................................... 5
DDC Example Configurations ................................................. 64
DC Specifications ......................................................................... 5
Signal Monitor ................................................................................ 67
AC Specifications—1300 MSPS.................................................. 6
SPORT Over JESD204B ............................................................ 68
AC Specifications—625 MSPS .................................................... 8
Digital Outputs ............................................................................... 70
Digital Specifications ................................................................... 9
Introduction to the JESD204B Interface ................................. 70
Switching Specifications ............................................................ 10
JESD204B Overview .................................................................. 70
Timing Specifications ................................................................ 11
Functional Overview ................................................................. 71
Absolute Maximum Ratings.......................................................... 13
JESD204B Link Establishment ................................................. 71
Thermal Characteristics ............................................................ 13
Physical Layer (Driver) Outputs .............................................. 73
ESD Caution ................................................................................ 13
Setting Up the AD9695 Digital Interface ................................. 74
Pin Configuration and Function Descriptions ........................... 14
Deterministic Latency.................................................................... 80
Typical Performance Characteristics ........................................... 16
Subclass 0 Operation.................................................................. 80
1300 MSPS ................................................................................... 16
Subclass 1 Operation.................................................................. 80
625 MSPS ..................................................................................... 21
Multichip Synchronization............................................................ 82
Equivalent Circuits ......................................................................... 25
Normal Mode.............................................................................. 82
Theory of Operation ...................................................................... 27
Timestamp Mode ....................................................................... 82
ADC Architecture ...................................................................... 27
SYSREF± Input ........................................................................... 84
Analog Input Considerations.................................................... 27
SYSREF± Setup/Hold Window Monitor ................................. 86
Voltage Reference ....................................................................... 30
Latency ............................................................................................. 88
DC Offset Calibration ................................................................ 30
End to End Total Latency .......................................................... 88
Clock Input Considerations ...................................................... 30
Example Latency Calculations.................................................. 88
Power-Down/Standby Mode..................................................... 33
LMFC Referenced Latency........................................................ 88
Temperature Diode .................................................................... 33
Test Modes ....................................................................................... 90
ADC Overrange and Fast Detect .................................................. 34
ADC Test Modes ........................................................................ 90
ADC Overrange .......................................................................... 34
JESD204B Block Test Modes .................................................... 91
Fast Threshold Detection (FD_A and FD_B) ........................ 34
Serial Port Interface (SPI) .............................................................. 93
ADC Application Modes and JESD204B Tx Converter Mapping
........................................................................................................... 35
Configuration Using the SPI ..................................................... 93
Programmable Finite Impulse Response (FIR) Filters .............. 37
SPI Accessible Features .............................................................. 93
Supported Modes........................................................................ 37
Memory Map .................................................................................. 94
Programming Instructions ........................................................ 39
Reading the Memory Map Register Table............................... 94
Digital Downconverter (DDC) ..................................................... 42
Memory Map Registers ............................................................. 95
DDC I/Q Input Selection .......................................................... 42
Applications Information ............................................................ 133
DDC I/Q Output Selection ....................................................... 42
Power Supply Recommendations........................................... 133
Hardware Interface..................................................................... 93
Rev. 0 | Page 2 of 135
Data Sheet
AD9695
Layout GuideLines ................................................................... 134
Ordering Guide .........................................................................135
AVDD1_SR (Pin 57) and AGND_SR (Pin 56 and Pin 60) .... 134
Outline Dimensions ..................................................................... 135
REVISION HISTORY
9/2017—Revision 0: Initial Version
Rev. 0 | Page 3 of 135
AD9695
Data Sheet
GENERAL DESCRIPTION
The AD9695 is a dual, 14-bit, 1300 MSPS/625 MSPS analog-todigital converter (ADC). The device has an on-chip buffer and a
sample-and-hold circuit designed for low power, small size, and
ease of use. This product is designed to support communications
applications capable of direct sampling wide bandwidth analog
signals of up to 2 GHz. The −3 dB bandwidth of the ADC input
is 2 GHz. The AD9695 is optimized for wide input bandwidth,
high sampling rate, excellent linearity, and low power in a small
package.
The dual ADC cores feature a multistage, differential pipelined
architecture with integrated output error correction logic. Each
ADC features wide bandwidth inputs supporting a variety of
user-selectable input ranges. An integrated voltage reference
eases design considerations. The analog input and clock signals
are differential inputs. The ADC data outputs are internally
connected to four digital downconverters (DDCs) through a
crossbar mux. Each DDC consists of multiple signal processing
stages: a 48-bit frequency translator (numerically controlled
oscillator (NCO)), and decimation filters. The NCO has the option
to select up to 16 preset bands over the general-purpose input/
output (GPIO) pins, or use a coherent fast frequency hopping
mechanism for band selection. Operation of the AD9695 between
the DDC modes is selectable via SPI-programmable profiles.
In addition to the DDC blocks, the AD9695 has several functions
that simplify the automatic gain control (AGC) function in a
communications receiver. The programmable threshold detector
allows monitoring of the incoming signal power using the fast
detect control bits in Register 0x0245 of the ADC. If the input
signal level exceeds the programmable threshold, the fast detect
indicator goes high. Because this threshold indicator has low
latency, the user can quickly turn down the system gain to avoid
an overrange condition at the ADC input. In addition to the fast
detect outputs, the AD9695 also offers signal monitoring
capability. The signal monitoring block provides additional
information about the signal being digitized by the ADC.
The user can configure the Subclasss 1 JESD204B-based high
speed serialized output using either one lane, two lanes, or four
lanes, depending on the DDC configuration and the acceptable
lane rate of the receiving logic device. Multidevice synchronization
is supported through the SYSREF± and SYNCINB± input pins.
The AD9695 has flexible power-down options that allow
significant power savings when desired. All of these features can
be programmed using a 3-wire serial port interface (SPI) and or
PDWN/STBY pin.
The AD9695 is available in a Pb-free, 64-lead LFCSP and is
specified over the −40°C to +105°C junction temperature range.
This product may be protected by one or more U.S. or
international patents.
Note that, throughout this data sheet, multifunction pins, such
as FD_A/GPIO_A0, are referred to either by the entire pin
name or by a single function of the pin, for example, FD_A,
when only that function is relevant.
PRODUCT HIGHLIGHTS
1.
2.
3.
4.
5.
6.
7.
Rev. 0 | Page 4 of 135
Low power consumption per channel.
JESD204B lane rate support up to 16 Gbps.
Wide, full power bandwidth supports intermediate
frequency (IF) sampling of signals up to 2 GHz.
Buffered inputs ease filter design and implementation.
Four integrated wideband decimation filters and NCO
blocks supporting multiband receivers.
Programmable fast overrange detection.
On-chip temperature diode for system thermal management.
Data Sheet
AD9695
SPECIFICATIONS
DC SPECIFICATIONS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, and sample rate = 625 MSPS
(AD9695-625 speed grade), sample rate = 1300 MSPS (AD9695-1300 speed grade), DCS on (AD9695-1300 speedgrade), DCS off (AD9695-625
speed grade), unless otherwise noted. Minimum and maximum specifications are guaranteed for the full operating junction temperature (TJ)
range of −40°C to +105°C. Typical specifications represent performance at TJ = 35°C (TA = 25°C for the AD9695-625 speed grade) and TJ = 40°C
(TA = 25°C for the AD9695-1300 speed grade).
Table 1.
Parameter
RESOLUTION
ACCURACY
No Missing Codes
Offset Error1
Offset Matching
Gain Error
Gain Matching
Differential Nonlinearity (DNL)
Integral Nonlinearity (INL)
TEMPERATURE DRIFT
Offset Error
Gain Error
INTERNAL VOLTAGE REFERENCE
Voltage
INPUT-REFERRED NOISE
ANALOG INPUTS
Differential Input Voltage Range
Common-Mode Voltage (VCM)
Differential Input Resistance
Differential Input Capacitance
Analog Full-Power Bandwidth
POWER SUPPLY
AVDD1
AVDD2
AVDD3
AVDD1_SR
DVDD
DRVDD1
DRVDD2
SPIVDD2
IAVDD1
IAVDD2
IAVDD3
IAVDD1_SR
IDVDD
IDRVDD13
IDRVDD2
ISPIVDD
POWER CONSUMPTION
Total Power Dissipation (Including Output Drivers)4
Power-Down Dissipation
Standby5
Min
14
−0.48
−2.9
−2.64
−0.7
−7
1300 MSPS
Typ
Guaranteed
5
0
±1
±0.18
625 MSPS
Max
Min
14
Typ
Max
Unit
Bits
Guaranteed
5
0
+0.25
±2.22
+2.6
±0.18
+2.5
+0.8
±2
+5
Codes
% FSR
% FSR
% FSR
LSB
LSB
±9
69
±6
123
ppm/°C
ppm/°C
0.5
3.8
0.5
2.7
V
LSB rms
±1
+0.48
+2.9
+2.64
0.8
5
−0.25
−2.6
−2.5
−0.8
−5
1.36
1.59
1.41
200
1.75
2
2.04
1.36
1.7
1.41
200
1.75
2
2.04
V p-p
V
Ω
pF
GHz
0.93
1.71
2.44
0.93
0.93
0.93
1.71
1.71
0.95
1.8
2.5
0.95
0.95
0.95
1.8
1.8
304
450
55
15
218
146
25
2
0.98
1.89
2.56
0.98
0.98
0.98
1.89
1.89
383
500
61
27
400
229
29
5
0.93
1.71
2.44
0.93
0.93
0.93
1.71
1.71
0.95
1.8
2.5
0.95
0.95
0.95
1.8
1.8
182
267
29
9
103
103
28
2
0.98
1.89
2.56
0.98
0.98
0.98
1.89
1.89
257
292
35
15
293
176
35
5
V
V
V
V
V
V
V
V
mA
mA
mA
mA
mA
mA
mA
mA
1.39
1.6
215
890
2
0.86
0.98
200
740
1.35
W
mW
mW
1
DC offset calibration on (Register 0x0701, Bit 7 = 1 and Register 0x073B, Bit 7 = 0).
The voltage level on the SPIVDD rail and on the DRVDD2 rail must be the same.
3
All lanes running. Power dissipation on DRVDD changes with lane rate and number of lanes used.
4
Default mode. No DDCs used.
5
Can be controlled by SPI.
2
Rev. 0 | Page 5 of 135
AD9695
Data Sheet
AC SPECIFICATIONS—1300 MSPS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, and sample rate = 1300 MSPS,
DCS on, buffer current settings specified in Table 11, unless otherwise noted. Minimum and maximum specifications are guaranteed for the
full operating junction temperature (TJ) range of −40°C to +105°C. Typical specifications represent performance at TJ = 40°C (TA = 25°C for the
AD9695-1300 speed grade).
Table 2.
Parameter1
ANALOG INPUT FULL SCALE
NOISE DENSITY2
SIGNAL-TO-NOISE RATIO (SNR)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
fIN = 1400 MHz
fIN = 1700 MHz
fIN = 1980 MHz
SIGNAL-TO-NOISE-AND-DISTORTION
RATIO (SINAD)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
fIN = 1400 MHz
fIN = 1700 MHz
fIN = 1980 MHz
EFFECTIVE NUMBER OF BITS (ENOB)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
fIN = 1400 MHz
fIN = 1700 MHz
fIN = 1980 MHz
SPURIOUS FREE DYNAMIC RANGE
(SFDR)
fIN = 10.3 MHz
fIN = 172.3MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
fIN = 1400 MHz
fIN = 1700 MHz
fIN = 1980 MHz
Analog Input Full
Scale = 1.36 V p-p
Min
Typ
Max
1.36
−152.6
64.4
64.4
64.3
64.0
63.8
63.2
62.7
62.3
64.3
64.3
64.2
63.9
63.6
63.1
62.6
62.1
10.3
10.3
10.3
10.3
10.2
10.1
10.1
10.0
81
81
80
83
82
80
80
81
Analog Input Full Scale =
1.59 V p-p
Min
Typ
Max
1.59
−153.9
64.5
64.3
10.3
74
Analog Input Full Scale =
2.04 V p-p
Min
Typ
Max
2.04
−155.6
Unit
V p-p
dBFS/Hz
65.7
65.6
65.6
65.2
64.9
64.2
63.6
63.0
67.5
67.5
67.3
66.6
66.1
65.2
64.5
63.9
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
65.4
65.4
65.3
65.0
64.7
63.8
63.4
62.8
66.1
66.2
65.7
65.5
65.7
62.9
64.2
61.8
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
10.5
10.5
10.5
10.5
10.4
10.3
10.2
10.1
10.6
10.7
10.6
10.5
10.6
10.1
10.3
9.9
Bits
Bits
Bits
Bits
dBFS
dBFS
dBFS
dBFS
79
78
77
80
81
76
80
79
73
72
71
72
79
67
78
68
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
Rev. 0 | Page 6 of 135
Data Sheet
Parameter1
WORST OTHER, EXCLUDING 2ND OR
3RD HARMONIC
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
fIN = 1400 MHz
fIN = 1700 MHz
fIN = 1980 MHz
TWO-TONE INTERMODULATION
DISTORTION (IMD), AIN1 AND AIN2 =
−7.0 dBFS
fIN1 = 170.8 MHz, fIN2 = 173.8 MHz
fIN1 = 343.5 MHz, fIN2 = 346.5 MHz
CROSSTALK3
Overrange Condition4
ANALOG INPUT BANDWIDTH, FULL
POWER5
AD9695
Analog Input Full
Scale = 1.36 V p-p
Min
Typ
Max
Analog Input Full Scale =
1.59 V p-p
Min
Typ
Max
−96
−95
−98
−95
−96
−90
−91
−90
−94
−96
−99
−95
−93
−89
−90
−90
−84
−83
>95
>95
2
−84
−82
>95
>95
2
1
−85
Analog Input Full Scale =
2.04 V p-p
Min
Typ
Max
Unit
−101
−95
−98
−92
−91
−86
−84
−77
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
dBFS
−83
−81
>95
>95
2
dBFS
dBFS
dB
dB
GHz
See the AN-835 Application Note, Understanding High Speed ADC Testing and Evaluation, for definitions and for details on how these tests were completed.
Noise density is measured at a low analog input frequcency (10 MHz).
3
Crosstalk is measured at 10 MHz with a −1.0 dBFS analog input on one channel, and no input on the adjacent channel.
4
The overrange condition is specified with 3 dB of the full-scale input range.
5
Full power bandwidth is the bandwidth of operation to achieve proper ADC performance.
2
Rev. 0 | Page 7 of 135
AD9695
Data Sheet
AC SPECIFICATIONS—625 MSPS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, and sample rate = 625 MSPS, DCS
off, buffer current setting specified in Table 11, unless otherwise noted. Minimum and maximum specifications are guaranteed for the full
operating junction temperature (TJ) range of −40°C to +105°C. Typical specifications represent performance at TJ = 35°C (TA = 25°C for the
AD9695-625 speed grade).
Table 3.
Parameter1
ANALOG INPUT FULL SCALE
NOISE DENSITY2
SIGNAL-TO-NOISE RATIO (SNR)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
SIGNAL-TO-NOISE-AND-DISTORTION
RATIO (SINAD)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
EFFECTIVE NUMBER OF BITS (ENOB)
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
SPURIOUS FREE DYNAMIC RANGE
(SFDR)
fIN = 10.3 MHz
fIN = 172.3MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
WORST OTHER, EXCLUDING 2ND OR
3RD HARMONIC
fIN = 10.3 MHz
fIN = 172.3 MHz
fIN = 340 MHz
fIN = 750 MHz
fIN = 1000 MHz
TWO-TONE INTERMODULATION
DISTORTION (IMD), AIN1 AND AIN2 =
−7.0 dBFS
fIN1 = 170.8 MHz, fIN2 = 173.8 MHz
fIN1 = 343.5 MHz, fIN2 = 346.5 MHz
CROSSTALK3
Overrange Condition4
Analog Input Full
Scale = 1.36 V p-p
Min
Typ
Max
1.36
−150.5
65.5
65.4
65.4
65.0
64.8
65.5
65.4
65.2
64.9
64.6
10.6
10.6
10.5
10.5
10.4
88
88
79
83
85
Analog Input Full Scale =
1.7 V p-p
Min
Typ
Max
1.7
−152.3
65.5
66.3
10.6
75
Analog Input Full Scale =
2.04 V p-p
Min
Typ
Max
2.04
−153.5
Unit
V p-p
dBFS/Hz
67.3
67.2
67.1
66.6
66.3
68.6
68.5
68.3
67.7
67.3
dBFS
dBFS
dBFS
dBFS
dBFS
66.9
67.0
67.0
65.4
65.0
67.2
68.0
67.9
67.0
67.0
dBFS
dBFS
dBFS
dBFS
dBFS
10.8
10.8
10.8
10.6
10.5
10.9
11.0
11.0
10.8
10.8
Bits
Bits
Bits
Bits
Bits
79
89
80
84
83
74
78
77
77
82
dBFS
dBFS
dBFS
dBFS
dBFS
−99
−99
−98
−100
−100
dBFS
dBFS
dBFS
dBFS
dBFS
−83
−84
>95
>95
dBFS
dBFS
dB
dB
−100
−101
−100
−98
−100
−101
−97
−102
−98
−98
−88
−89
>95
>95
−88
−89
>95
>95
Rev. 0 | Page 8 of 135
−90
Data Sheet
Parameter1
ANALOG INPUT BANDWIDTH, FULL
POWER5
AD9695
Analog Input Full
Scale = 1.36 V p-p
Min
Typ
Max
2
Analog Input Full Scale =
1.7 V p-p
Min
Typ
Max
2
Analog Input Full Scale =
2.04 V p-p
Min
Typ
Max
2
Unit
GHz
1
See the AN-835 Application Note, Understanding High Speed ADC Testing and Evaluation, for definitions and for details on how these tests were completed.
Noise density is measured at a low analog input frequcency (10 MHz).
Crosstalk is measured at 10 MHz with a −1.0 dBFS analog input on one channel, and no input on the adjacent channel.
4
The overrange condition is specified with 3 dB of the full-scale input range.
5
Full power bandwidth is the bandwidth of operation to achieve proper ADC performance.
2
3
DIGITAL SPECIFICATIONS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, and sample rate = 625 MSPS
(AD9695-625 speed grade), sample rate = 1300 MSPS (AD9695-1300 speed grade), DCS on (AD9695-1300 speedgrade), DCS off (AD9695-625
speed grade), unless otherwise noted. Minimum and maximum specifications are guaranteed for the full operating junction temperature
(TJ) range of −40°C to +105°C. Typical specifications represent performance at TJ = 35°C (TA = 25°C for the AD9695-625 speed grade) and
TJ = 40°C (TA = 25°C for the AD9695-1300 speed grade).
Table 4.
Parameter
CLOCK INPUTS (CLK+, CLK−)
Logic Compliance
Differential Input Voltage
Input Common-Mode Voltage
Input Resistance (Differential)
Input Capacitance (Differential)
SYSREF INPUTS (SYSREF+, SYSREF−)
Logic Compliance
Differential Input Voltage
Input Common-Mode Voltage
Input Resistance (Differential)
Input Capacitance (Differential)
LOGIC INPUTS (SDIO, SCLK, CSB, PDWN/STBY, FD_A/GPIO_A0,
FD_B/GPIO_B0)
Logic Compliance
Logic 1 Voltage
Logic 0 Voltage
Input Resistance
LOGIC OUTPUT (SDIO, FD_A, FD_B)
Logic Compliance
Logic 1 Voltage (IOH = 4 mA)
Logic 0 Voltage (IOL = 4 mA)
SYNCIN INPUTS (SYNCINB−, SYNCINB+)
Logic Compliance
Differential Input Voltage
Input Common-Mode Voltage
Input Resistance (Differential)
Input Capacitance (Single-Ended per Pin)
DIGITAL OUTPUTS (SERDOUTx±, x = 0 TO 3)
Logic Compliance
Differential Output Voltage
Differential Termination Impedance
Min
400
Typ
LVDS/LVPECL
800
0.65
32
Max
Unit
1600
mV p-p
V
kΩ
pF
0.9
400
LVDS/LVPECL
800
0.65
18
1
1800
2
mV p-p
V
kΩ
pF
CMOS
0.75 × SPIVDD
0
0.35 × SPIVDD
30
V
V
kΩ
CMOS
SPIVDD − 0.45
0
400
0.45
LVDS/LVPECL/CMOS
800
1800
0.65
2
18
1
V
V
mV p-p
V
kΩ
pF
SST
360
80
Rev. 0 | Page 9 of 135
520
100
770
1200
mV p-p
Ω
AD9695
Data Sheet
SWITCHING SPECIFICATIONS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, and sample rate = 625 MSPS
(AD9695-625 speed grade), sample rate = 1300 MSPS (AD9695-1300 speed grade), DCS on (AD9695-1300 speedgrade), DCS off (AD9695-625
speed grade), unless otherwise noted. Minimum and maximum specifications are guaranteed for the full operating junction temperature (TJ)
range of −40°C to +105°C. Typical specifications represent performance at TJ = 35°C (TA = 25°C for the AD9695-625 speed grade) and
TJ = 40°C (TA = 25°C for the AD9695-1300 speed grade).
Table 5.
Parameter
CLOCK
Clock Rate (at CLK+/CLK− Pins)
Maximum Sample Rate1
Minimum Sample Rate2
Clock Pulse Width3
High
Low
OUTPUT PARAMETERS
Unit Interval (UI)4
Rise Time (tR) (20% to 80% into 100 Ω Load)
Fall Time (tF) (20% to 80% into 100 Ω Load)
Phase-Locked Loop (PLL) Lock Time
Data Rate per Channel (NRZ)5
LATENCY6
Pipeline Latency
Fast Detect Latency
Wake-Up Time7
Standby
Power-Down
APERTURE
Aperture Delay (tA)
Aperture Uncertainty (Jitter, tJ)
Out of Range Recovery Time
Min
1300 MSPS
Typ
Max
0.24
1400
240
1.28
156.25
156.25
62.5
1.6875
Min
625 MSPS
Typ
Max
0.24
640
240
1.28
156.25
156.25
76.9
28
28
5
13
62.5
16
1.6875
Unit
GHz
MSPS
MSPS
ps
ps
160
28
28
5
6.25
16
ps
ps
ps
ms
Gbps
56
26
56
26
Clock cycles
Clock cycles
400
15
400
15
us
ms
192
43
1
159.5
49.2
1
ps
fs rms
Clock cycles
1
The maximum sample rate is the clock rate after the divider.
The minimum sample rate operates at 240 MSPS. See SPI Register 0x011A to reduce the threshold of the clock detect circuit.
3
Clock duty stabilizer (DCS) on. See SPI Register 0x011C and 0x011E to enable DCS.
4
Baud rate = 1/UI. A subset of this range can be supported.
5
Default L = 4. This number can change based on the sample rate and decimation ratio.
6
No DDCs used. L = 4, M = 2, and F = 1.
7
Wake-up time is defined as the time required to return to normal operation from power-down mode.
2
Rev. 0 | Page 10 of 135
Data Sheet
AD9695
TIMING SPECIFICATIONS
Table 6.
Parameter
CLK+ to SYSREF+ TIMING REQUIREMENTS
tSU_SR
tH_SR
SPI TIMING REQUIREMENTS
tDS
tDH
tCLK
tS
tH
tHIGH
tLOW
tACCESS
Test Conditions/Comments
See Figure 3
Device clock to SYSREF+ setup time
Device clock to SYSREF+ hold time
See Figure 4
Setup time between the data and the rising edge of SCLK
Hold time between the data and the rising edge of SCLK
Period of the SCLK
Setup time between CSB and SCLK
Hold time between CSB and SCLK
Minimum period that SCLK must be in a logic high state
Minimum period that SCLK must be in a logic low state
Maximum time delay between falling edge of SCLK and output
data valid for a read operation
Time required for the SDIO pin to switch from an output to an
input relative to the CSB rising edge (not shown in Figure 4)
tDIS_SDIO
Min
Typ
10
N – 56
N+1
N – 54
SAMPLE N
N – 53
N–1
CLK–
CLK+
CLK–
SERDOUT0–
SERDOUT1–
SERDOUT1+
SERDOUT2–
SERDOUT2+
SERDOUT3–
SERDOUT3+
A
B
C
D
E
F
G
H
I
J
CONVERTER0
SAMPLE N – 56 MSB
A
B
C
D
E
F
G
H
I
J
CONVERTER0
SAMPLE N – 56 LSB
A
B
C
D
E
F
G
H
I
J
CONVERTER0
SAMPLE N – 55 MSB
A
B
C
D
E
F
G
H
I
J
CONVERTER0
SAMPLE N – 55 LSB
SAMPLE N – 56 AND N – 55
ENCODED INTO ONE
8-BIT/10-BIT SYMBOL
15660-002
CLK+
SERDOUT0+
Figure 2. Data Output Timing Diagram
CLK–
CLK+
tSU_SR
tH_SR
15660-003
SYSREF–
SYSREF+
Figure 3. SYSREF± Setup and Hold Timing Diagram
Rev. 0 | Page 11 of 135
ps
ps
6
ns
ns
ns
ns
ns
ns
ns
ns
10
ns
APERTURE DELAY
N – 55
Unit
−70
120
4
2
40
2
2
10
10
Timing Diagrams
ANALOG
INPUT
SIGNAL
Max
AD9695
Data Sheet
tDS
tS
tHIGH
tCLK
tDH
SDIO
DON’T CARE
DON’T CARE
R/W
A14
A13
A12
A11
A10
A9
A8
A7
Figure 4. SPI Timing Diagram
Rev. 0 | Page 12 of 135
D5
D4
D3
D2
D1
D0
DON’T CARE
15660-004
DON’T CARE
tH
tLOW
CSB
SCLK
tACCESS
Data Sheet
AD9695
ABSOLUTE MAXIMUM RATINGS
THERMAL CHARACTERISTICS
Table 7.
Parameter
Electrical
AVDD1 to AGND
AVDD1_SR to AGND
AVDD2 to AGND
AVDD3 to AGND
DVDD to DGND
DRVDD1 to DRGND
DRVDD2 to DRGND
SPIVDD to DGND
AGND to DRGND
AGND to DGND
DGND to DRGND
VIN±x to AGND
CLK± to AGND
SCLK, SDIO, CSB to DGND
PDWN/STBY to DGND
SYSREF± to AGND
SYNCINB± to DRGND
Junction Temperature Range (TJ)
Storage Temperature Range,
Ambient (TA)
Rating
1.05 V
1.05 V
2.00 V
2.70 V
1.05 V
1.05 V
2.00 V
2.00 V
−0.3 V to +0.3 V
−0.3 V to +0.3 V
−0.3 V to +0.3 V
AGND − 0.3 V to AVDD3 + 0.3 V
AGND − 0.3 V to AVDD1 + 0.3 V
DGND − 0.3 V to SPIVDD + 0.3 V
DGND − 0.3 V to SPIVDD + 0.3 V
2.5 V
2.5 V
−40°C to +125°C
−65°C to +150°C
Typical θJA, θJB, and θJC are specified vs. the number of printed
circuit board (PCB) layers in different airflow velocities (in m/sec).
Airflow increases heat dissipation effectively reducing θJA and
θJB. In addition, metal in direct contact with the package leads
and exposed pad from metal traces, through holes, ground, and
power planes, reduces θJA. Thermal performance for actual
applications requires careful inspection of the conditions in
an application. The use of appropriate thermal management
techniques is recommended to ensure that the maximum
junction temperature does not exceed the limits shown in Table 7.
Table 8. Thermal Resistance
Package
Type
CP-64-17
Airflow
Velocity
(m/sec)
0
1.0
2.5
θJA1, 2 θJC_BOT1, 3
22.5 1.7
17.9
16.8
1
θJC_TOP1, 3
7.6
θJB1, 4
4.3
Per JEDEC 51-7, plus JEDEC 51-5 2S2P test board.
Per JEDEC JESD51-2 (still air) or JEDEC JESD51-6 (moving air).
Per MIL-Std 883, Method 1012.1.
4
Per JEDEC JESD51-8 (still air).
2
3
ESD CAUTION
Stresses at or above those listed under Absolute Maximum
Ratings may cause permanent damage to the product. This is a
stress rating only; functional operation of the product at these
or any other conditions above those indicated in the operational
section of this specification is not implied. Operation beyond
the maximum operating conditions for extended periods may
affect product reliability.
Rev. 0 | Page 13 of 135
θJT1,2
0.2
Unit
°C/W
°C/W
°C/W
AD9695
Data Sheet
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
AVDD1
AVDD2
AVDD2
AVDD1
AGND_SR
SYSREF–
SYSREF+
AVDD1_SR
AGND_SR
AVDD1
CLK–
CLK+
AVDD1
AVDD2
AVDD2
AVDD1
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
AD9695
TOP VIEW
(Not to Scale)
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
AVDD1
AVDD1
AVDD2
AVDD3
VIN–B
VIN+B
AVDD3
AVDD2
AVDD2
AVDD2
SPIVDD
CSB
SCLK
SDIO
DVDD
DGND
NOTES
1. ANALOG GROUND. CONNECT THE EXPOSED PAD TO THE
ANALOG GROUND PLANE.
15660-005
FD_A/GPIO_A0
DRGND
DRVDD1
SYNCINB–
SYNCINB+
SERDOUT0–
SERDOUT0+
SERDOUT1–
SERDOUT1+
SERDOUT2–
SERDOUT2+
SERDOUT3–
SERDOUT3+
DRVDD1
DRGND
FD_B/GPIO_B0
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
AVDD1
AVDD1
AVDD2
AVDD3
VIN–A
VIN+A
AVDD3
AVDD2
AVDD2
AVDD2
DRVDD2
VREF
SPIVDD
PDWN/STBY
DVDD
DGND
Figure 5. Pin Configuration (Top View)
Table 9. Pin Function Descriptions
Pin No.
1, 2, 47 to 49, 52,
55, 61, 64
3, 8 to 10, 39 to 41,
46, 50, 51, 62, 63
4, 7, 42, 45
5, 6
11
12
Mnemonic
AVDD1
Type
Power supply
Description
Analog Power Supply (0.95 V Nominal).
AVDD2
Power supply
Analog Power Supply (1.8 V Nominal).
AVDD3
VIN−A, VIN+A
DRVDD2
VREF
Power supply
Analog input
Power supply
Input/output
13, 38
14
SPIVDD
PDWN/STBY
Power supply
Digital control input
15, 34
16, 33
DVDD
DGND
17
FD_A/GPIO_A0
Power supply
Ground power
supply
CMOS output
Analog Power Supply (2.5 V Nominal).
ADC A Analog Input Complement/True.
Digital Driver Power Supply (1.8 V Nominal).
Reference Voltage Input (0.50 V)/Do Not Connect. This pin is configurable
through the SPI as a no connect pin or as an input. Do not connect this pin if
using the internal reference. This pin requires a 0.50 V reference voltage
input if using an external voltage reference source.
Digital Power Supply for SPI (1.8 V Nominal).
Power-Down Input (Active High). The operation of this pin depends on the
SPI mode and can be configured as power-down or standby.
Digital Power Supply (0.95 V Nominal).
Digital Control Ground Supply. These pins connect to the digital ground plane.
32
18, 31
FD_B/GPIO_B0
DRGND
19, 30
20
21
22, 23
DRVDD1
SYNCINB−
SYNCINB+
SERDOUT0−,
SERDOUT0+
SERDOUT1−,
SERDOUT1+
24, 25
CMOS output
Ground power
supply
Power supply
Digital input
Digital input
Data output
Fast Detect Output for Channel A (FD_A). General-purpose input/output
(GPIO) Pin A0 (GPIO_A0).
Fast Detect Output for Channel B (FD_B). GPIO Pin B0 (GPIO_B0).
Digital Driver Ground Supply. This pin connects to the digital driver ground
plane.
Digital Driver Power Supply (0.95 V Nominal).
Active Low JESD204B LVDS/CMOS Sync Input True.
Active Low JESD204B LVDS Sync Input Complement.
Lane 0 Output Data Complement/True.
Data output
Lane 1 Output Data Complement/True.
Rev. 0 | Page 14 of 135
Data Sheet
Pin No.
26, 27
AD9695
35
Mnemonic
SERDOUT2−
SERDOUT2+
SERDOUT3−,
SERDOUT3+
SDIO
36
37
43, 44
53, 54
56, 60
SCLK
CSB
VIN+B, VIN−B
CLK+, CLK−
AGND_SR
57
58, 59
AVDD1_SR
SYSREF+,
SYSREF−
EPAD
28, 29
Type
Data output
Description
Lane 2 Output Data Complement/True.
Data output
Lane 3 Output Data Complement/True.
Digital control
input/output
Digital control input
Digital control input
Analog input
Analog input
Ground power
supply
Power supply
Digital input
SPI Serial Data Input/Output.
Ground power
supply
SPI Serial Clock.
SPI Chip Select (Active Low).
ADC B Analog Input True/Complement.
Clock Input True/Complement.
Ground Reference for SYSREF±.
Analog Power Supply for SYSREF± (0.95 V Nominal).
Active High JESD204B LVDS System Reference Input Complement/True.
Analog Ground. Connect the exposed pad to the analog ground plane.
Rev. 0 | Page 15 of 135
AD9695
Data Sheet
TYPICAL PERFORMANCE CHARACTERISTICS
1300 MSPS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, sample rate = 625 MSPS
(AD9695-625 speed grade), sample rate = 1300 MSPS (AD9695-1300 speed grade), DCS on (AD9695-1300 speedgrade), DCS off (AD9695-625
speed grade), buffer current setting specified in Table 11, dc offset calibration enabled, unless otherwise noted. Minimum and maximum
specifications are guaranteed for the full operating junction temperature (TJ) range of −40°C to +105°C. Typical specifications represent
performance at TJ = 35°C (TA = 25°C for the AD9695-625 speed grade) and TJ = 40°C (TA = 25°C for the AD9695-1300 speed grade).
10
AIN = 10.3MHz
SNR = 65.7dBFS
SFDR = 79.0dBFS
BUFFER CURRENT = 300µA
–30
–30
AMPLITUDE (dBFS)
–50
–70
–90
–70
–90
200
400
600
FREQUENCY (MHz)
–130
15660-305
0
0
AIN = 172.3MHz
SNR = 65.6dBFS
SFDR = 78.0dBFS
BUFFER CURRENT = 300µA
AIN = 1002.3MHz
SNR = 64.9dBFS
SFDR = 81.0dBFS
BUFFER CURRENT = 300µA
–10
–30
AMPLITUDE (dBFS)
–50
–70
–90
–110
–50
–70
–90
0
200
400
600
FREQUENCY (MHz)
15660-306
–110
–130
0
200
400
600
FREQUENCY (MHz)
Figure 7. Single-Tone FFT with Analog Input Frequency (fIN) = 172.3 MHz
Figure 10. Single-Tone FFT with fIN = 1002.3 MHz
AIN = 342.3MHz
SNR = 65.6dBFS
SFDR = 77.0dBFS
BUFFER CURRENT = 300µA
–10
AMPLITUDE (dBFS)
–30
–50
–70
–90
–110
–50
–70
–90
–110
0
200
400
FREQUENCY (MHz)
600
15660-307
AMPLITUDE (dBFS)
AIN = 1402.3MHz
SNR = 64.2dBFS
SFDR = 76.0dBFS
BUFFER CURRENT = 300µA
–10
–30
–130
15660-309
AMPLITUDE (dBFS)
600
Figure 9. Single-Tone FFT with fIN =752.3 MHz
–30
–130
400
FREQUENCY (MHz)
Figure 6. Single-Tone FFT with Analog Input Frequency (fIN) = 10.3 MHz
–10
200
15660-308
–110
–110
–130
–50
–130
0
200
400
600
FREQUENCY (MHz)
Figure 8. Single-Tone FFT with fIN = 342.3 MHz
Figure 11. Single-Tone FFT with fIN = 1402.3 MHz
Rev. 0 | Page 16 of 135
15660-310
AMPLITUDE (dBFS)
–10
AIN = 752.3MHz
SNR = 65.2dBFS
SFDR = 80.0dBFS
BUFFER CURRENT = 300µA
–10
Data Sheet
AD9695
67
AIN = 1702.3MHz
SNR = 63.6dBFS
SFDR = 80.0dBFS
BUFFER CURRENT = 300µA
–10
TJ = +105°C
ROOM
TJ = –40°C
66
65
–70
64
–90
63
–110
62
–130
0
200
600
400
FREQUENCY (MHz)
61
Figure 12. Single-Tone FFT with fIN = 1702.3 MHz
500
1000
1500
2000
ANALOG INPUT FREQUENCY (MHz)
Figure 15. SNR vs. Analog Input Frequency (fIN) at Minimum, Room, and
Maximum Temperatures
90
AIN = 1980.3MHz
SNR = 63.0dBFS
SFDR = 79.0dBFS
BUFFER CURRENT = 300µA
–10
0
15660-314
SNR (dBFS)
–50
15660-311
AMPLITUDE (dBFS)
–30
TJ = +105°C
ROOM
TJ = –40°C
85
80
–70
75
–90
70
–110
65
–130
200
0
400
600
FREQUENCY (MHz)
60
0
500
1000
2000
Figure 16. SFDR vs. Analog Input Frequency (fIN) at Minimum, Room, and
Maximum Temperatures
Figure 13. Single-Tone FFT with fIN = 1980.3 MHz
85
fIN1 = 170.8MHz
fIN2 = 173.8MHz
IMD = –84dBFS
BUFFER CURRENT = 300µA
–10
SNR
SFDR
AMPLITUDE (dBFS)
80
75
70
–60
–110
60
500
700
900
1100
1300
SAMPLE RATE (MHz)
–160
0
200
400
600
FREQUENCY (MHz)
Figure 17. Two-Tone FFT; fIN1 = 170.8 MHz, fIN2 = 173.8 MHz
Figure 14. SNR/SFDR vs. Sample Rate, fIN =172.3 MHz
Rev. 0 | Page 17 of 135
15660-316
65
15660-313
SNR/SFDR (dBFS)
1500
ANALOG INPUT FREQUENCY (MHz)
15660-315
SDFR (dBFS)
–50
15660-312
AMPLITUDE (dBFS)
–30
AD9695
0
Data Sheet
85
fIN1 = 343.5MHz
fIN2 = 346.5MHz
IMD = –82dBFS
BUFFER CURRENT = 300µA
SNR
SFDR
80
SNR/SFDR (dBFS)
AMPLITUDE (dBFS)
–40
–90
75
70
0
200
400
60
–40
15660-317
–140
600
FREQUENCY (MHz)
10
60
110
JUNCTION TEMPERATURE (°C)
15660-320
65
Figure 21. SNR/SFDR vs. Junction Temperature, fIN = 172.3 MHz
Figure 18. Two-Tone FFT; fIN1 = 343.5 MHz, fIN2 = 346.5 MHz
2
120
100
1
0
60
INL (LSB)
SNR/SFDR (dB)
80
40
20
–1
–2
0
SFDR (dBFS)
SNRFS
SNR (dBc)
SFDR (dBc)
–60
–40
–20
0
ANALOG INPUT AMPLITUDE (dBFS)
–4
0
15000
Figure 22. INL, fIN = 10.3 MHz
10
0.3
SFDR (dBFS)
SFDR (dBc)
IMD3 (dBc)
IMD3 (dBFS)
–10
0.2
0.1
DNL (LSB)
–30
–50
–70
0
–0.1
–90
–0.2
–110
–0.3
–75
–55
–35
–15
ANALOG INPUT AMPLITUDE (dB)
15660-319
SFDR/IMD3 (dB)
10000
OUTPUT CODE
Figure 19. SNR/SFDR vs. Analog Input Amplitude, fIN = 172.3 MHz
–130
–95
5000
15660-321
–80
–0.4
0
5000
10000
OUTPUT CODE
Figure 23. DNL, fIN = 10.3 MHz
Figure 20. SFDR/IMD3 vs. Analog Input Amplitude, fIN = 172.3 MHz
Rev. 0 | Page 18 of 135
15000
15660-322
–40
–100
–3
15660-318
–20
Data Sheet
AD9695
1.8
NUMBER OF HITS
15000
10000
0
–16
–13
–10
–7
–4
–1
2
5
8
11
14
16
CODE
1.4
1.2
500
15660-323
5000
1.6
700
900
15660-326
POWER CONSUMPTION (W)
20000
1300
1100
SAMPLE RATE (MHz)
Figure 27. Total Power Dissipation vs. Sample Rate (fS)
Figure 24. Input Referred Noise Histogram
66
0
–2
65
–8
–10
64
CLOCK AMPLITUDE
400mV
600mV
800mV
1000mV
1200mV
1400mV
1600mV
1800mV
2000mV
63
–12
62
–14
0
1000
2000
3000
4000
AIN FREQUENCY (MHz)
61
15660-019
–16
Figure 25. Full Power Bandwidth
500
1000
1500
2000
ANALOG INPUT FREQUENCY (MHz)
Figure 28. SNR vs. Analog Input Frequency at Different Clock Amplitudes
2.0
86
84
1.8
SFDR (dBFS)
82
1.6
80
78
76
BUFFER CURRENT
BUFFER CURRENT
BUFFER CURRENT
BUFFER CURRENT
74
1.2
–40
10
60
110
JUNCTION TEMPERATURE (°C)
Figure 26. Total Power Dissipation vs. Junction Temperature
72
0
500
= 460µA
= 400µA
= 360µA
= 300µA
1000
1500
ANALOG INPUT FREQUENCY (MHz)
2000
15660-328
1.4
15660-325
TOTAL POWER CONSUMPTION (W)
0
15660-327
–6
SNR (dBFS)
AMPLITUDE (dBFS)
–4
Figure 29. SFDR vs. Analog Input Frequency with Different Buffer Current Settings
Rev. 0 | Page 19 of 135
AD9695
Data Sheet
90
68
67
80
IAVDD3 (mA)
65
64
63
60
0
500
1000
1500
2000
ANALOG INPUT FREQUENCY (MHz)
15660-329
2.04V
1.81V
1.59V
1.36V
62
61
Figure 30. SNR vs. Analog Input Frequency with Different Analog Input FullScale Values
90
85
SFDR (dBFS)
80
75
70
0
500
1000
1500
ANALOG INPUT FREQUENCY (MHz)
2000
15660-330
2.04V
1.81V
1.59V
1.36V
65
60
70
Figure 31. SFDR vs. Analog Input Frequency with Different Analog Input FullScale Values
Rev. 0 | Page 20 of 135
50
300
350
400
450
BUFFER CURRENT SETTING (µA)
Figure 32. IAVDD3 vs. Buffer Control 1 Setting in Register 0x1A4C
15660-331
SNR (dBFS)
66
Data Sheet
AD9695
625 MSPS
AVDD1 = 0.95 V, AVDD1_SR = 0.95 V, AVDD2 = 1.8 V, AVDD3 = 2.5 V, DVDD = 0.95 V, DRVDD1 = 0.95 V, DRVDD2 = 1.8 V, SPIVDD =
1.8 V, clock divider = 2, default input full scale, 0.5 V internal reference, AIN = −1.0 dBFS, default SPI settings, sample rate = 625 MSPS
(AD9695-625 speed grade), sample rate = 1300 MSPS (AD9695-1300 speed grade), DCS on (AD9695-1300 speedgrade), DCS off
(AD9695-625 speed grade), buffer current setting specified in Table 11, and dc offset calibration enabled, unless otherwise noted. Minimum
and maximum specifications are guaranteed for the full operating junction temperature (TJ) range of −40°C to +105°C. Typical specifications
represent performance at TJ = 35°C (TA = 25°C for the AD9695-625 speed grade) and TJ = 40°C (TA = 25°C for the AD9695-1300 speed grade).
0
0
AIN = –1dBFS
SNR = 67.2dBFS
SFDR = 89dBFS
BUFFER CURRENT = 160µA
–20
AMPLITUDE (dBFS)
–40
–60
–80
–80
–100
0
100
200
300
FREQUENCY (MHz)
–120
15660-006
–120
–60
0
250
300
80
–40
SNR/SFDR (dBFS)
–60
–80
60
40
20
–100
100
200
300
FREQUENCY (MHz)
15660-007
0
0
250
SNR
SFDR
350
450
550
650
750
850
SAMPLE RATE (MSPS)
Figure 34. Single-Tone FFT with fIN = 340 MHz
15660-010
AMPLITUDE (dBFS)
200
100
AIN = –1dBFS
SNR = 67.1dBFS
SFDR = 80dBFS
BUFFER CURRENT = 160µA
–20
Figure 37. SNR/SFDR vs. Sample Rate, fIN =172.3 MHz
0
100
AIN = –1dBFS
SNR = 66.6dBFS
SFDR = 84dBFS
BUFFER CURRENT = 300µA
–20
90
SNR/SFDR (dBFS)
–40
–60
–80
–100
80
70
60
SFDR (TJ = –40°C)
SNR (TJ = –40°C)
SFDR (ROOM)
SNR (ROOM)
SFDR (TJ = +105°C)
SNR (TJ = +105°C)
50
0
100
200
FREQUENCY (MHz)
Figure 35. Single-Tone FFT with fIN =750 MHz
300
40
15660-008
AMPLITUDE (dBFS)
150
Figure 36. Single-Tone FFT with fIN = 1000 MHz
0
–120
100
FREQUENCY (MHz)
Figure 33. Single-Tone FFT with Analog Input Frequency (fIN) = 172.3 MHz
–120
50
15660-009
–100
–40
0
250
500
750
ANALOG INPUT FREQUENCY (MHz)
1000
15660-011
AMPLITUDE (dBFS)
–20
AIN = –1dBFS
SNR = 66.3dBFS
SFDR = 83dBFS
BUFFER CURRENT = 300µA
Figure 38. SNR/SFDR vs. Analog Input Frequency (fIN) at Minimum, Room,
and Maximum Temperatures
Rev. 0 | Page 21 of 135
AD9695
Data Sheet
0
100
AIN1 AND A IN2 = –7dBFS
SFDR = 88dBFS
BUFFER CURRENT = 160µA
90
–60
–80
80
70
60
–100
50
–120
40
–40
0
100
200
300
FREQUENCY (MHz)
SNR
SFDR
10
15660-015
SNR/SFDR (dBFS)
–40
15660-012
AMPLITUDE (dBFS)
–20
60
JUNCTION TEMPERATURE (°C)
Figure 39. Two-Tone FFT; fIN1 = 170.8 MHz, fIN2 = 173.8 MHz
Figure 42. SNR/SFDR vs. Junction Temperature, fIN = 172.3 MHz
0
4
AIN1 AND A IN2 = –7dBFS
SFDR = 89dBFS
BUFFER CURRENT = 160µA
–20
3
1
INL (LSB)
AMPLITUDE (dBFS)
2
–40
–60
0
–1
–80
–2
–100
100
200
300
FREQUENCY (MHz)
–4
0
5000
10000
15000
OUTPUT CODE
Figure 40. Two-Tone FFT; fIN1 = 343.5 MHz, fIN2 = 346.5 MHz
15660-016
0
15660-013
–120
–3
Figure 43. INL, fIN = 10.3 MHz
120
0.25
0.20
100
0.15
0.10
DNL (LSB)
60
40
0
–0.05
–0.15
20
–0.20
SFDR (dBFS)
SFDR (dBc)
–80
–60
–40
–20
0
ANALOG INPUT AMPLITUDE (dBFS)
Figure 41. SNR/SFDR vs. Analog Input Amplitude, fIN = 172.3 MHz
–0.25
0
5000
10000
OUTPUT CODE
Figure 44. DNL, fIN = 10.3 MHz
Rev. 0 | Page 22 of 135
15000
15660-017
0
–100
0.05
–0.10
15660-014
SFDR (dB)
80
Data Sheet
AD9695
6000
1.4
4000
3000
2000
1000
N + 15
0.8
0.6
0.4
0.2
650
850
750
68
67
–4
66
–6
–8
65
CLOCK AMPLITUDE
400mV p-p
600mV p-p
800mV p-p
1000mV p-p
1200mV p-p
1400mV p-p
1600mV p-p
1800mV p-p
2000mV p-p
2200mV p-p
64
–10
63
–12
62
–14
1000
2000
3000
4000
AIN FREQUENCY (MHz)
15660-019
61
0
0
200
400
600
800
1000
1200
ANALOG INPUT FREQUENCY (MHz)
15660-022
–2
SNR (dBFS)
Figure 49. SNR vs. Analog Input Frequency at Different Clock Amplitudes
Figure 46. Full Power Bandwidth
1.4
100
1.2
90
1.0
SFDR (dBFS)
80
0.8
0.6
70
60
0.4
10
60
110
JUNCTION TEMPERATURE (°C)
Figure 47. Total Power Dissipation vs. Junction Temperature
40
15660-020
0
–40
BUFFER CURRENT
BUFFER CURRENT
BUFFER CURRENT
BUFFER CURRENT
50
0.2
0
200
400
600
800
= 160µA
= 200µA
= 240µA
= 300µA
1000
ANALOG INPUT FREQUENCY (MHz)
1200
15660-023
AMPLITUDE (dBFS)
550
Figure 48. Total Power Dissipation vs. Sample Rate (fS)
0
TOTAL POWER DISSIPATION (W)
450
SAMPLE RATE (MSPS)
Figure 45. Input Referred Noise Histogram
–16
350
15660-018
N + 12
N+9
N+6
N+3
0
N–3
N–6
N–9
N – 12
N – 15
1.0
0
250
0
CODE
1.2
15660-021
TOTAL POWER DISSIPATION (W)
NUMBER OF HITS
5000
Figure 50. SFDR vs. Analog Input Frequency with Different Buffer Current
Settings (AIN < 1250 MHz)
Rev. 0 | Page 23 of 135
AD9695
Data Sheet
80
69
BUFFER CURRENT = 400µA
BUFFER CURRENT = 300µA
BUFFER CURRENT = 240µA
68
67
75
SNR (dBFS)
SFDR (dBFS)
66
70
65
64
63
65
62
1450
1650
60
650
15660-348
60
1250
1850
ANALOG INPUT FREQUENCY (MHz)
Figure 51. SFDR vs. Analog Input Frequency with Different Buffer Current
Settings (AIN > 1250 MHz), Register 0x1B03 = 0x02, Register 0x1B08 = 0xC1,
Register 0x1B10 = 0x1C
750
850
950
1050
1150
1250
ANALOG INPUT FREQUENCY (MHz)
15660-026
INPUT FULL-SCALE = 1.36V p-p
INPUT FULL-SCALE = 1.7V p-p
INPUT FULL-SCALE = 2.04V p-p
61
Figure 54. SNR vs. Analog Input Frequency with Different Analog Input
Full-Scale Values (AIN > 650 MHz)
85
70
69
80
68
75
SFDR (dBFS)
66
65
64
63
65
60
62
INPUT FULL-SCALE = 1.36V p-p
INPUT FULL-SCALE = 1.7V p-p
INPUT FULL-SCALE = 2.04V p-p
0
200
400
600
ANALOG INPUT FREQUENCY (MHz)
INPUT FULL-SCALE = 1.36V p-p
INPUT FULL-SCALE = 1.7V p-p
INPUT FULL-SCALE = 2.04V p-p
55
50
650
15660-024
61
60
70
Figure 52. SNR vs. Analog Input Frequency with Different Analog Input
Full-Scale Values (AIN < 650 MHz)
750
850
950
1050
1150
1250
ANALOG INPUT FREQUENCY (MHz)
15660-027
SNR (dBFS)
67
Figure 55. SFDR vs. Analog Input Frequency with Different Analog Input
Full-Scale Values (AIN > 650 MHz)
80
100
95
70
90
60
IAVDD3 (mA)
80
75
70
40
65
60
30
0
200
400
ANALOG INPUT FREQUENCY (MHz)
600
15660-025
INPUT FULL-SCALE = 1.36V p-p
INPUT FULL-SCALE = 1.7V p-p
INPUT FULL-SCALE = 2.04V p-p
55
50
50
Figure 53. SFDR vs. Analog Input Frequency with Different Analog Input
Full-Scale Values (AIN < 650 MHz)
Rev. 0 | Page 24 of 135
20
160
210
260
310
360
BUFFER CURRENT (µA)
Figure 56. IAVDD3 vs. Buffer Control 1 Setting in Register 0x1A4C
15660-028
SFDR (dBFS)
85
Data Sheet
AD9695
EQUIVALENT CIRCUITS
AVDD3
AVDD3
VIN+x
3.5pF
AVDD3
400Ω
EMPHASIS/SWING
CONTROL (SPI)
VCM
BUFFER
10pF
AVDD3
DRVDD1
AVDD3
100Ω
DATA+
SERDOUTx+
x = 0, 1, 2, 3
VIN–x
DRGND
OUTPUT
DRIVER
AIN
CONTROL
(SPI)
DATA–
SERDOUTx–
x = 0, 1, 2, 3
15660-029
3.5pF
DRGND
Figure 57. Analog Inputs
Figure 60. Digital Outputs
DRVDD1
AVDD1
CLK+
DRVDD1
15660-032
100Ω
25Ω
DRGND
2.5kΩ
16kΩ
DRVDD1
100Ω
SYNCINB+
CMOS
PATH
SYNCINB PIN
CONTROL (SPI)
10kΩ
AVDD1
1.9pF
25Ω
16kΩ
LEVEL
TRANSLATOR
15660-030
DRGND
VCM = 0.65V
130kΩ
Figure 58. Clock Inputs
100Ω
SYNCINB–
100Ω
DRVDD1
10kΩ
1.9pF
AVDD1_SR
SYSREF+
DRGND
130kΩ
10kΩ
DRGND
15660-033
CLK–
DRGND
Figure 61. SYNCINB± Inputs
1.9pF
130kΩ
LEVEL
TRANSLATOR
100Ω
SPIVDD
10kΩ
1.9pF
ESD
PROTECTED
14808-026
SYSREF–
AVDD1_SR
SCLK
56kΩ
ESD
PROTECTED
Figure 59. SYSREF± Inputs
SPIVDD
DGND
Figure 62. SCLK Input
Rev. 0 | Page 25 of 135
DGND
15660-034
130kΩ
AD9695
Data Sheet
SPIVDD
SPIVDD
ESD
PROTECTED
ESD
PROTECTED
56kΩ
PDWN/
STBY
CSB
56kΩ
ESD
PROTECTED
DGND
PDWN
CONTROL (SPI)
15660-035
DGND
DGND
DGND
15660-037
ESD
PROTECTED
Figure 65. PDWN/STBY Input
Figure 63. CSB Input
SPIVDD
SPIVDD
SDI
DGND
56kΩ
DGND
DGND
TEMPERATURE DIODE
VOLTAGE OUTPUT
AVDD2
SDO
DGND
EXTERNAL REFERENCE
VOLTAGE INPUT
VREF
15660-036
ESD
PROTECTED
VCM OUTPUT
SPIVDD
VREF PIN
CONTROL (SPI)
AGND
Figure 66. VREF Input/Output
Figure 64. SDIO Input
SPIVDD
SPIVDD
ESD
PROTECTED
NCO BAND SELECT
DGND
FD_A/GPIO_A0,
FD_B/GPIO_B0
SPIVDD
FD
JESD204B LMFC
56kΩ
ESD
PROTECTED
JESD204B SYNC~
DGND
DGND
DGND
FD PIN CONTROL (SPI)
Figure 67. FD_A/GPIO_A0 and FD_B/GPIO_B0
Rev. 0 | Page 26 of 135
15660-039
SDIO
15660-038
ESD
PROTECTED
Data Sheet
AD9695
THEORY OF OPERATION
The AD9695 has two analog input channels and up to eight
JESD204B output lane pairs. The ADC samples wide bandwidth
analog signals of up to 2 GHz. The actual −3 dB roll-off of the
analog inputs is 2 GHz. The AD9695 is optimized for wide input
bandwidth, high sampling rate, excellent linearity, and low
power in a small package.
The dual ADC cores feature a multistage, differential pipelined
architecture with integrated output error correction logic. Each
ADC features wide bandwidth inputs supporting a variety of
user-selectable input ranges. An integrated voltage reference
eases design considerations.
low-pass filter that limits unwanted broadband noise. For more
information, refer to the Analog Dialogue article “TransformerCoupled Front-End for Wideband A/D Converters”
(Volume 39, April 2005). In general, the precise front-end
network component values depend on the application.
Figure 68 shows the differential input return loss curve for the
analog inputs across a frequency fange of 1 MHz to 10 GHz.
The reference impedence is 100 Ω.
The AD9695 has several functions that simplify the AGC
function in a communications receiver. The programmable
threshold detector allows monitoring of the incoming signal
power using the fast detect output bits of the ADC. If the input
signal level exceeds the programmable threshold, the fast detect
indicator goes high. Because this threshold indicator has low
latency, the user can quickly turn down the system gain to avoid
an overrange condition at the ADC input.
1: 1.000MHz
170.59nF
2: 100.000 MHz
45.72pF
3: 200.000MHz
12.07pF
4: 300.000MHz
6.49pF
5: 400.000MHz
4.70pF
6: 500.000MHz
4.00pF
182.88Ω
–932.98mΩ
177.37Ω
–34.81Ω
157.29Ω
–65.95Ω
128.82Ω
–81.70Ω
102.55Ω
–84.58Ω
82.01Ω
1
–79.60Ω
2
3
6
4
ADC ARCHITECTURE
The architecture of the AD9695 consists of an input buffered
pipelined ADC. The input buffer provides a termination
impedance to the analog input signal. This termination
impedance is set to 200 Ω. The equivalent circuit diagram of the
analog input termination is shown in Figure 29. The input
buffer is optimized for high linearity, low noise, and low power
across a wide bandwidth.
The input buffer provides a linear high input impedance (for
ease of drive) and reduces kickback from the ADC. The
quantized outputs from each stage are combined into a final
14-bit result in the digital correction logic. The pipelined
architecture permits the first stage to operate with a new input
sample; at the same time, the remaining stages operate with the
preceding samples. Sampling occurs on the rising edge of the clock.
ANALOG INPUT CONSIDERATIONS
The analog input to the AD9695 is a differential buffer. The
internal common-mode voltage of the buffer is 1.41 V. The
clock signal alternately switches the input circuit between
sample mode and hold mode.
Either a differential capacitor or two single-ended capacitors (or
a combination of both) can be placed on the inputs to provide a
matching passive network. These capacitors ultimately create a
CH1 AVG = 1
> CH1: START 1.0MHz
STOP 10.0000GHz
15660-200
5
The Subclass 1 JESD204B-based high speed, serialized output data
lanes can be configured in one-lane (L = 1), two-lane (L = 2),
and four-lane (L = 4) configurations, depending on the sample
rate and the decimation ratio. Multiple device synchronization
is supported through the SYSREF± and SYNCINB± input pins.
The SYSREF± pin in the AD9695 can also be used as a timestamp
of data as it passes through the ADC and out of the JESD204B
interface.
Figure 68. AD9695 Different Input Return Loss
For best dynamic performance, the source impedances driving
VIN+x and VIN−x must be matched such that any commonmode settling errors are symmetrical. These errors are reduced
by the common-mode rejection of the ADC. An internal reference
buffer creates a differential reference that defines the span of the
ADC core.
Maximum SNR performance is achieved by setting the ADC to
the largest span in a differential configuration. For the AD9695,
the available span is programmable through the SPI port from
1.36 V p-p to 2.04 V p-p differential, with 1.7 V p-p differential
being the default.
Differential Input Configurations
There are several ways to drive theAD9695, either actively or
passively. Optimum performance is achieved by driving the
analog input differentially.
For applications where SNR and SFDR are key parameters,
differential transformer coupling is the recommended input
configuration (see Figure 69 and Table 9) because the noise
performance of most amplifiers is not adequate to achieve the
true performance of the AD9695.
For low to midrange frequencies, a double balun or double
transformer network (see Figure 69 and Table 9) is recommended
for optimum performance of the AD9695. For higher frequencies
in the second or third Nyquist zones, it is recommended to
remove some of the front-end passive components to ensure
wideband operation (see Table 9).
Rev. 0 | Page 27 of 135
AD9695
Data Sheet
C2
R1
R3
C3
R2
MARKI
BAL-0006
C4
C2
ADC
R3
C3
15547-050
R1
200Ω
C1
R2
NOTES:
1. SEE TABLE 9 FOR COMPONENT VALUES
Figure 69. Differential Transformer-Coupled Configuration for the AD9695
Table 10. Differential Transformer-Coupled Input Configuration Component Values
Speed Grade
AD9695-625
AD9695-1300
Transformer
BAL-0006/BAL-0006SMG
BAL-0006/BAL-0006SMG
R1
25 Ω
25 Ω
R2
25 Ω
25 Ω
R3
10 Ω
10 Ω
C1
0.1 μF
0.1 μF
C2
0.1 μF
0.1 μF
C3
DNI1
DNI1
C4
DNI1
DNI1
DNI means do not insert.
The analog inputs of the AD9695 are internally biased to the
common mode, as shown in Figure 71.
For dc-coupled applications, the recommended operation
procedure is to export the common-mode voltage to the VREF pin
using the SPI writes listed in this section. The common-mode
voltage must be set by the exported value to ensure proper ADC
operation. Disconnect the internal common-mode buffer from
the analog input using Register 0x1908.
When performing SPI writes for dc coupling operation, use the
following register settings, in order:
1.
2.
3.
4.
5.
Set Register 0x1908, Bit 2 to 1 to disconnect the internal
common-mode buffer from the analog input.
Set Register 0x18A6 to 0x00 to turn off the voltage
reference.
Set Register 0x18E6 to 0x00 to turn off the temperature
diode export.
Set Register 0x18E3, Bit 6 to 0x01 to turn on the VCM export.
Set Register 0x18E3, Bits[5:0] to the buffer current setting
(copy the buffer current setting from Register 0x1A4C and
Register 0x1A4D to improve the accuracy of the commonmode export).
Figure 70 shows the block diagram representation of a dccoupled application.
ADC
ADC
AMP A
VOCM
VREF
VOCM
ADC
AMP B
VCM EXPORT SELECT
SPI REGISTERS 0x1908,
0x18A6, 0x18E3, 0x18E6)
15660-041
Input Common Mode
Figure 70. DC-Coupled Application Using the AD9695
Analog Input Buffer Controls and SFDR Optimization
The AD9695 input buffer offers flexible controls for the analog
inputs, such as, buffer current, and input full-scale adjustment.
All the available controls are shown in Figure 71.
AVDD3
AVDD3
VIN+
3.5pF
100Ω
AVDD3
100Ω
AVDD3
VIN–
REG
(0x0008,
0x1908)
AVDD3
3.5pF
REG (0x0008, 0x1A4C,
0x1A4D, 0x1910)
Figure 71. Analog Input Controls
Rev. 0 | Page 28 of 135
15660-042
1
Frequency Range
<2 GHz
<2 GHz
Data Sheet
AD9695
Using Register 0x1A4C and Register 0x1A4D, the buffer behavior
on each channel can be adjusted to optimize the SFDR over various
input frequencies and bandwidths of interest. Use Register 0x1910
to change the internal reference voltage. Changing the internal
reference voltage results in a change in the input full-scale voltage.
When the input buffer current in Register 0x1A4C and
Register 0x1A4D is set, the amount of current required by the
AVDD3 supply changes. This relationship is shown in Figure 72.
For a complete list of buffer current settings, see Table 11.
85
80
IAVDD3 (mA)
75
70
65
Table 11 shows the recommended values for the buffer current
for various Nyquist zones.
Absolute Maximum Input Swing
The absolute maximum input swing allowed at the inputs of the
AD9695 is 5.6 V p-p differential. Signals operating near or at
this level can cause permanent damage to the ADC.
Dither
The AD9695 has internal on-chip dither circuitry that improves
the ADC linearity and SFDR, particularly at smaller signal
levels. A known but random amount of white noise is injected into
the input of theAD9695. This dither improves the small signal
linearity within the ADC transfer function and is precisely
subtracted out digitally. The dither is turned on by default and
does not reduce the ADC input dynamic range. The data sheet
specifications and limits are obtained with the dither turned on.
The dither is on by default. It is not recommended to turn it off.
60
50
300
350
400
450
BUFFER CURRENT SETTING (µA)
15660-369
55
Figure 72. AVDD3 Current (IAVDD3) vs. Buffer Current Setting (Buffer Control 1
Setting in Register 0x1A4C and Buffer Control 2 Setting in Register 0x1A4D)
Table 11. SFDR Optimization for Input Frequencies
AD9695-1300
Frequency
DC to 650 MHz
650 MHz to 1250 MHz
>1250 MHz
All AIN frequencies
Register 0x1A4C and
Register 0x1A4D
160 μA
300 μA
400 μA
300 μA
Register 0x1B03
0x00
0x00
0x02
0x02
Register 0x1B08
0x01
0x01
0xC1
0xC1
VIN+A/VIN+B
VIN–A/VIN–B
INTERNAL
0.5V
REFERENCE
GENERATOR
VFS
ADJUST
ADC
CORE
INPUT FULL SCALE
RANGE ADJUST
SPI REGISTER
(0x1910)
VREF
VREF PIN
CONTROL SPI
REGISTER
(0x18A6)
Figure 73. Internal Reference Configuration and Controls
Rev. 0 | Page 29 of 135
15660-044
Speed Grade
AD9695-625
Register 0x1B10
0x00
0x00
0x1C
0x00
AD9695
Data Sheet
INTERNAL
0.5V
REFERENCE
GENERATOR
ADR130
NC
NC
ADC
GND SET
INPUT
VFS
ADJUST
VREF
VIN VOUT
0.1µF
0.1µF
15660-045
VREF PIN
AND VFS
CONTROL
Figure 74. External Reference Using the ADR130
VOLTAGE REFERENCE
A stable and accurate 0.5 V voltage reference is built into the
AD9695. This internal 0.5 V reference sets the full-scale input
range of the ADC. The full-scale input range can be adjusted via
the ADC input full-scale control register (Register 0x1910). For
more information on adjusting the input swing, see Table 47.
Figure 73 shows the block diagram of the internal 0.5 V reference
controls.
The SPI Register 0x18A6 enables the user to either use this
internal 0.5 V reference, or to provide an external 0.5 V
reference. When using an external voltage reference, provide a
0.5 V reference. The full-scale adjustment is made using the SPI,
irrespective of the reference voltage. For more information on
adjusting the full-scale level of the AD9695, refer to the
Memory Map section.
The SPI writes required to use the external voltage reference, in
order, are as follows:
1.
2.
Set Register 0x18E3 to 0x00 to turn off the VCM export.
Set Register 0x18E6 to 0x00 to turn off the temperature
diode export.
Set Register 0x18A6 to 0x01 to turn on the external voltage
reference.
3.
The use of an external reference may be necessary, in some
applications, to enhance the gain accuracy of the ADC or to
improve thermal drift characteristics. Figure 75 shows the
typical drift characteristics of the internal 0.5 V reference.
The external reference must be a stable 0.5 V reference. The
ADR130 is a sufficient option for providing the 0.5 V reference.
Figure 74 shows how the ADR130 can be used to provide the
external 0.5 V reference to the AD9695. The dashed lines show
unused blocks within the AD9695 while using the ADR130 to
provide the external reference.
DC OFFSET CALIBRATION
The AD9695 contains a digital filter to remove the dc offset
from the output of the ADC. For ac-coupled applications, this
filter can be enabled by setting Register 0x0701, Bit 7 to 0x1 and
setting Register 0x73B, Bit 7 to 0x0. The filter computes the
average dc signal and it is digitally subtracted from the ADC
output. As a result, the dc offset is improved to better than
70 dBFS at the output. Because the filter does not distinguish
between the source of dc signals, this feature can be used when
the signal content at dc is not of interest. The filter corrects dc
up to ±512 codes and saturates beyond that.
CLOCK INPUT CONSIDERATIONS
For optimum performance, drive the AD9695 sample clock
inputs (CLK+ and CLK−) with a differential signal. This
signal is typically ac-coupled to the CLK+ and CLK− pins via a
transformer or clock drivers. These pins are biased internally
and require no additional biasing.
Figure 76 shows a preferred method for clocking the AD9695.
The low jitter clock source is converted from a single-ended
signal to a differential signal using an RF transformer.
0.503
CLK+
0.501
100Ω
1:2Z
ADC
CLK–
0.500
Figure 76. Transformer Coupled Differential Clock
0.499
0.498
0.497
0.496
0.495
–40
–20
0
20
40
60
JUNCTION TEMPERATURE (°C)
80
100
Figure 75. Typical VREF Drift
Rev. 0 | Page 30 of 135
15660-047
CLOCK INPUT
15660-046
BANDGAP VOLTAGE (V)
0.502
Data Sheet
AD9695
Another option is to ac couple a differential CML or LVDS
signal to the sample clock input pins, as shown in Figure 77 and
Figure 78.
150Ω
ADC
CLOCK
INPUT
CLK–
15660-048
LVDS
DRIVER
150Ω
The AD9695 contains an input clock divider with the ability to
divide the input clock by 1, 2, or 4. Select the divider ratios
using Register 0x0108 (see Figure 80).
The maximum frequency at the CLK± inputs is 1.28 GHz,
which is the limit of the divider. In applications where the clock
input is a multiple of the sample clock, take care to program the
appropriate divider ratio into the clock divider before applying
the clock signal; this ensures that the current transients during
device startup are controlled.
CLK+
100Ω
DIFFERENTIAL 100Ω
TRACE
Input Clock Divider
Figure 77. Differential LVPECL Sample Clock
REG 0x011C,
0x011E
CLK+
CLK+
ADC
CLOCK
INPUT
CLK–
CLK–
DIFFERENTIAL
TRACE
REG 0x0108
Figure 78. Differential CML Sample Clock
Figure 80. Clock Divider Circuit
The AD9695 clock divider can be synchronized using the
external SYSREF± input. A valid SYSREF± signal causes the
clock divider to reset to a programmable state. This synchronization feature allows multiple devices to have their clock
dividers aligned to guarantee simultaneous input sampling.
See the Memory Map Register section for more information.
DAC
CLOCK
INPUT
100Ω
CLKOUT–
CLK+
ADC
CLOCK
INPUT
CLK–
15660-050
ADC
CLKOUT+
÷2
÷4
15660-051
100Ω
15660-049
CML
DRIVER
Figure 79. Clock Output Clocking the AD9695
Clock Duty Cycle Considerations
Input Clock Divider ½ Period Delay Adjust
Typical high speed ADCs use both clock edges to generate a
variety of internal timing signals. The AD9695 contains an
internal clock divider and a duty cycle stabilizer comprised of
Duty Cycle Stabilizer 1 (DCS1) and Duty Cycle Stabilizer 2
(DCS2).
The input clock divider inside the AD9695 provides phase delay
in increments of ½ the input clock cycle. Register 0x10C can be
programmed to enable this delay independently for each channel.
Changing this register does not affect the stability of the
JESD204B link.
For the AD9695 625 MSPS speed grade, the DCS is disabled by
default. In applications where the clock duty cycle cannot be
guaranteed to be 50%, a higher multiple frequency clock along
with the usage of the clock divider is recommended.
Clock Fine Delay and Superfine Delay Adjust
In the AD9695 625 MSPS speed grade, when it is not possible to
provide a higher frequency clock, it is recommended to turn on
DCSx using Register 0x011C and Register 0x011E. Figure 80
shows the different controls to the AD9695 clock inputs. The
output of the divider offers a 50% duty cycle, high slew rate (fast
edge) clock signal to the internal ADC.
In the AD9695 1300 MSPS speed grade, the DCS is enabled by
default. It is recommended to keep DCS on inrespective of clock
divide ratio in the AD9695.
See the Memory Map section for more details on using this feature.
Adjust the AD9695 sampling edge instant by writing to
Register 0x0110, Register 0x0111, and Register 0x0112. Bits[2:0]
of Register 0x0110 enable the selection of the fine delay, or the
fine delay with superfine delay. The fine delay allows the user to
delay the clock edges with 16 step or 192 step delay options. The
superfine delay is an unsigned control to adjust the clock delay
in superfine steps of 0.25 ps each.
Register 0x0112, Bits[7:0] offer the user the option to delay
the clock in 192 delay steps. Register 0x0111, Bits[7:0] offer the
user the option to delay the clock in 128 superfine steps. These
values can be programmed individually for each channel. To
use the superfine delay option, set the clock delay control in
Register 0x0110, Bits[2:0] to 0x2 or 0x6. Figure 81 shows the
controls available to the clock dividers within AD9695. It is
recommended to apply the same delay settings to the digital
delay circuits as are applied to the analog delay circuits to
maintain sample accuracy through the pipe.
Rev. 0 | Page 31 of 135
AD9695
Data Sheet
130
CHANNEL A
12.5fS
25fS
50fS
100fS
200fS
400fS
800fS
120
110
PHASE
CH. A
CLK_DIV
SNR (dB)
PHASE
CH. B
0x0108
FINE DELAY
0x0110,
0x0111,
0x0112
70
50
CHANNEL B
40
30
10
Figure 81. Clock Divider Phase and Delay Controls
The clock delay adjustment takes effect immediately when it is
enabled via SPI writes. Enabling the clock fine delay adjust in
Register 0x0110 causes a datapath reset. However, the contents
of Register 0x0111 and Register 0x0112 can be changed without
affecting the stability of the JESD204B link.
Clock Coupling Considerations
The AD9695 has many different domains within the analog
supply that control various aspects of the data conversion. The
clock domain is supplied by Pin 49 and Pin 64 on the analog
supply (AVDD1). To minimize coupling between the clock
supply domain and the other analog domains, it is recommended
to add a supply Q factor reduction circuitry (de-Q) for Pin 49
and Pin 64, as shown in Figure 82.
FERRITE BEAD
220Ω AT
100MHz
DCR ≤ 0.5Ω
80
60
15660-052
0x0109
90
100
1000
10000
ANALOG INPUT FREQUENCY (MHz)
15660-054
CLK INPUT
100
Figure 83. Ideal SNR vs. Input Frequency and Jitter
Treat the clock input as an analog signal when aperture jitter
may affect the dynamic range of the AD9695. Separate power
supplies for clock drivers from the ADC output driver supplies
to avoid modulating the clock signal with digital noise. If the
clock is generated from another type of source (by gating,
dividing, or other methods), retime the clock by the original clock
at the last step. Refer to the AN-501 Application Note and the
AN-756 Application Note for more in depth information about
jitter performance as it relates to ADCs.
Figure 84 shows the estimated SNR of the AD9695 across input
frequency for different clock induced jitter values. Estimate the
SNR by using the following equation:
  SNR JITTER
  SNR ADC 

  10 
10
 10 
SNR (dBFS) = −10log10 10


PIN 49
100nF
10Ω

 




70
AVDD1
PLANE
65
10Ω
60
55
Figure 82. De-Q Network Recommendation for the Clock Domain Supply
50
Clock Jitter Considerations
High speed, high resolution ADCs are sensitive to the quality of
the clock input. Calculate the degradation in SNR at a given
input frequency (fIN) due only to aperture jitter (tJ) by
45
10
25fS
50fS
75fS
100fS
125fS
150fS
175fS
100
INPUT FREQUENCY (MHz)
1000
15660-484
100nF
PIN 64
15660-053
220Ω AT
100MHz
DCR ≤ 0.5Ω
SNR (dBFS)
FERRITE BEAD
Figure 84. Estimated SNR Degradation for the AD9695 vs. Analog Input
Frequency and RMS Jitter
SNRJITTER = −20 × log10 (2 × π × fIN × tJ)
In this equation, the rms aperture jitter represents the root
mean square of all jitter sources, including the clock input,
analog input signal, and ADC aperture jitter specifications.
IF undersampling applications are particularly sensitive to jitter
(see Figure 83).
Rev. 0 | Page 32 of 135
Data Sheet
AD9695
The AD9695 has a PDWN/STBY pin that configure the device
in power-down or standby mode. The default operation is PDWN.
The PDWN/STBY pin is a logic high pin. When in power-down
mode, the JESD204B link is disrupted. The power-down option
can also be set via Register 0x003F and Register 0x0040.
In standby mode, the JESD204B link is not disrupted and transmits
zeros for all converter samples. Change this transmission using
Register 0x0571, Bit 7 to select /K/ characters.
The SPI writes required to export the central temperature diode
are as follows (see the Memory Map section for more information):
4.
5.
6.
Set Register 0x0008 to 0x03 to select both channels.
Set Register 0x18E3 to 0x00 to turn off VCM export.
Set Register 0x18A6 to 0x00 to turn off voltage reference
export.
Set Register 0x18E6 to 0x01 to turn on voltage export of
the central 1× temperature diode. The typical voltage
response of the temperature diode is shown in Figure 87.
Although this voltage represents the die temperature, it is
recommended to take measurements from a pair of diodes
for improved accuracy. The following step explains how to
enable the 20× diode.
Set Register 0x18E6 to 0x02 to turn on the second central
temperature diode of the pair, which is 20× the size of the
first. For the method using two diodes simultaneously to
achieve a more accurate result, see the AN-1432 Application
Note, Practical Thermal Modeling and Measurements in
High Power ICs.
7.
TEMPERATURE DIODE
ADC
ADC
B
DIGITAL
JESD204B DRIVER
TEMPERATURE DIODE
LOCATIONS
CHANNEL A, CENTRAL,
CHANNEL B
0.75
0.70
0.65
0.60
0.55
0.50
–40
–20
0
40
60
80
100
Figure 87. Typical Voltage Response of the 1× Temperature Diode
The relationship between the measured delta voltage (ΔV) and
the junction temperature in degrees Celcius is shown in Figure 88.
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
–10
–20
–30
–40
60
TJ (°C)
The temperature diode voltages can be exported to the VREF pin
using the SPI. Use Register 0x18E6 to enable or disable diodes.
It is important to note that other voltages may be exported to
the VREF pin at the same time, which can result in undefined
behavior. To ensure a proper readout, switch off all other voltage
exporting circuits as described in this section. Figure 86 shows
the block diagram of the controls that are required to enable the
diode voltage readout.
VREF PIN
CONTROL
SPI REGISTER
(0x18A6)
VREF
20
JUNCTION TEMPERATURE (°C)
Figure 85. Temperature Diode Locations in the Die
CHANNEL A
CENTRAL
TEMPERATURE DIODE
LOCATION SELECT
SPI REGISTER (0x18E6)
15660-057
CHANNEL B
65
70
75
80
85
90
95
100
105
110
DELTA VOLTAGE (mV)
Figure 88. Junction Temperature (TJ) vs. Delta Voltage (ΔV)
Figure 86. Register Controls to Output Temperature Diode Voltage on the
VREF Pin
Rev. 0 | Page 33 of 135
15660-059
VREF
0.80
15660-056
ADC
A
8.
TEMPERATURE DIODE VOLTAGE (V)
The AD9695 contains diode-based temperature sensors. The
diodes output voltages commensurate to the temperature of the
silicon. There are multiple diodes on the die, but the results
established using the temperature diode at the central location
of the die can be regarded as representative of the entire die.
However, in applications where only one channel is used (the
other channel being in a power-down state), it is recommended
to read the temperature diode corresponding to the channel
that is on. Figure 85 shows the locations of the diodes in the
AD9695 with voltages that can be output to the VREF pin. In
each location, there is a pair of diodes, one of which is 20× the
size of the other. It is recommended to use both diodes in a
location to obtain an accurate estimate of the die temperature.
For more information, see the AN-1432 Application Note, Practical
Thermal Modeling and Measurements in High Power ICs.
15660-058
POWER-DOWN/STANDBY MODE
AD9695
Data Sheet
ADC OVERRANGE AND FAST DETECT
The operation of the upper threshold and lower threshold
registers, along with the dwell time registers, is shown in
Figure 89.
In receiver applications, it is desirable to have a mechanism to
reliably determine when the converter is about to be clipped.
The standard overrange bit in the JESD204B outputs provides
information on the state of the analog input that is of limited
usefulness. Therefore, it is helpful to have a programmable
threshold below full scale that allows time to reduce the gain
before the clip actually occurs. In addition, because input
signals can have significant slew rates, the latency of this
function is of major concern. Highly pipelined converters can
have significant latency. The AD9695 contains fast detect
circuitry for individual channels to monitor the threshold and
assert the FD_A and FD_B pins.
The FD indicator is asserted if the input magnitude exceeds the
value programmed in the fast detect upper threshold registers,
located at Register 0x0247 and Register 0x0248. The selected
threshold register is compared with the signal magnitude at the
output of the ADC. The fast upper threshold detection has a
latency of 28 clock cycles (maximum). The approximate upper
threshold magnitude is defined by
Upper Threshold Magnitude (dBFS) = 20 log (Threshold
Magnitude/213)
ADC OVERRANGE
The FD indicators are not cleared until the signal drops below
the lower threshold for the programmed dwell time. The lower
threshold is programmed in the fast detect lower threshold
registers, located at Register 0x0249 and Register 0x024A. The
fast detect lower threshold register is a 13-bit register that is
compared with the signal magnitude at the output of the ADC.
This comparison is subject to the ADC pipeline latency, but is
accurate in terms of converter resolution. The lower threshold
magnitude is defined by
The ADC overrange indicator is asserted when an overrange is
detected on the input of the ADC. The overrange indicator can
be embedded within the JESD204B link as a control bit (when
CSB > 0). The latency of this overrange indicator matches the
sample latency.
The AD9695 also records any overrange condition in any of the
eight virtual converters. For more information on the virtual
converters, refer to Figure 90. The overrange status of each virtual
converter is registered as a sticky bit in Register 0x563. The
contents of Register 0x563 can be cleared using Register 0x562,
by toggling the bits corresponding to the virtual converter to set
and reset position.
Lower Threshold Magnitude (dBFS) = 20 log (Threshold
Magnitude/213)
For example, to set an upper threshold of −6 dBFS, write 0xFFF
to Register 0x0247 and Register 0x0248. To set a lower threshold of
−10 dBFS, write 0xA1D to Register 0x0249 and Register 0x024A.
FAST THRESHOLD DETECTION (FD_A AND FD_B)
The fast detect bit is immediately set whenever the absolute value of
the input signal exceeds the programmable upper threshold
level. The FD bit is only cleared when the absolute value of the
input signal drops below the lower threshold level for greater
than the programmable dwell time. This feature provides
hysteresis and prevents the FD bit from excessively toggling.
The dwell time can be programmed from 1 to 65,535 sample
clock cycles by placing the desired value in the fast detect dwell
time registers, located at Register 0x24B and Register 0x024C. See
the Memory Map section (Register 0x0040, and Register 0x0245 to
Register 0x024C in Table 47) for more details.
UPPER THRESHOLD
DWELL TIME
TIMER RESET BY
RISE ABOVE
LOWER
THRESHOLD
DWELL TIME
FD_A OR FD_B
Figure 89. Threshold Settings for FD_A and FD_B Signals
Rev. 0 | Page 34 of 135
TIMER COMPLETES BEFORE
SIGNAL RISES ABOVE
LOWER THRESHOLD
15660-060
MIDSCALE
LOWER THRESHOLD
Data Sheet
AD9695
ADC APPLICATION MODES AND JESD204B Tx CONVERTER MAPPING
Table 12 shows the number of virtual converters required and
the transport layer mapping when channel swapping is disabled.
Figure 62 shows the virtual converters and their relationship to
the DDC outputs when complex outputs are used.
The AD9695 contains a configurable signal path that allows
different features to be enabled for different applications. These
features are controlled using the chip application mode register,
Register 0x0200. The chip operating mode is controlled by Bits[3:0]
in this register, and the chip Q ignore is controlled by Bit 5.
Each DDC channel outputs either two sample streams (I/Q) for
the complex data components (real + imaginary), or one sample
stream for real (I) data. The AD9695 can be configured to use up
to eight virtual converters, depending on the DDC configuration.
The AD9695 contains the following modes:


Full bandwidth mode: two 14-bit ADC cores running at
the full sample rate.
DDC mode: up to four digital downconverter (DDC)
channels.
The I/Q samples are always mapped in pairs with the I samples
mapped to the first virtual converter and the Q samples mapped
to the second virtual converter. With this transport layer mapping,
the number of virtual converters are the same whether a single
real converter is used along with a digital downconverter block
producing I/Q outputs, or whether an analog downconversion
is used with two real converters producing I/Q outputs.
After the chip application mode is selected, the output
decimation ratio is set using the chip decimation ratio in
Register 0x0201, Bits[3:0]. The output sample rate = ADC
sample rate/the chip decimation ratio.
To support the different application layer modes, the AD9695
treats each sample stream (real, I, or Q) as originating from
separate virtual converters.
Figure 63 shows a block diagram of the two scenarios described
for I/Q transport layer mapping.
Table 12. Virtual Converter Mapping
1
2
2
4
4
8
Chip Operating
Mode
(Reg. 0x0200,
Bits[3:0])
Full bandwidth
mode (0x0)
One DDC mode
(0x1)
One DDC mode
(0x1)
Two DDC mode
(0x2)
Two DDC mode
(0x2)
Four DDC mode
(0x3)
Four DDC mode
(0x3)
Virtual Converter Mapping
Chip Q Ignore
(0x0200, Bit 5)
Real or
complex (0x0)
Real (I only)
(0x1)
Complex (I/Q)
(0x0)
Real (I only)
(0x1)
Complex (I/Q)
(0x0)
Real (I only)
(0x1)
Complex (I/Q)
(0x0)
REAL/I
0
ADC A
samples
DDC0 I
samples
DDC0 I
samples
DDC0 I
samples
DDC0 I
samples
DDC0 I
samples
DDC0 I
samples
1
ADC B
samples
Unused
2
Unused
3
Unused
4
Unused
5
Unused
6
Unused
7
Unused
Unused
Unused
Unused
Unused
Unused
Unused
DDC0 Q
samples
DDC1 I
samples
DDC0 Q
samples
DDC1 I
samples
DDC0 Q
samples
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
DDC1 I
samples
DDC2 I
samples
DDC1 I
samples
DDC1 Q
samples
DDC3 I
samples
DDC1 Q
samples
Unused
Unused
Unused
Unused
Unused
Unused
Unused
Unused
DDC2 I
samples
DDC2 Q
samples
DDC3 I
samples
DDC3 Q
samples
REAL/I
ADC A
SAMPLING
AT fS
REAL/Q
REAL/I
I/Q
CROSSBAR
MUX
REAL/Q
REAL/I
REAL/Q
REAL/Q
ADC B
SAMPLING
AT fS
REAL/I
REAL/Q
DDC 0
I
REAL/I
CONVERTER 0
Q
Q
CONVERTER 1
I
Q
DDC 1
I
REAL/I
CONVERTER 2
Q
Q
CONVERTER 3
I
Q
DDC 2
I
Q
DDC 3
I
Q
REAL/I
CONVERTER 6
Q
Q
CONVERTER 7
I
Figure 90. DDCs and Virtual Converter Mapping
Rev. 0 | Page 35 of 135
OUTPUT
INTERFACE
REAL/I
I
CONVERTER 4
Q
Q
CONVERTER 5
15660-061
Number of
Virtual
Converters
Supported
1 to 2
AD9695
Data Sheet
DIGITAL DOWNCONVERSION
M=2
I
CONVERTER 0
ADC
REAL
I
REAL
DIGITAL
DOWNCONVERSION
Q
CONVERTER 1
L LANES
JESD204B
Tx
L LANES
I/Q ANALOG MIXING
M=2
I
CONVERTER 0
ADC
90°
PHASE
Q
JESD204B
Tx
ADC
Q
CONVERTER 1
Figure 91. I/Q Transport Layer Mapping
Rev. 0 | Page 36 of 135
15660-062
REAL
Data Sheet
AD9695
PROGRAMMABLE FINITE IMPULSE RESPONSE (FIR) FILTERS

SUPPORTED MODES
The AD9695 supports the following modes of operation (the
asterisk symbol (*) denotes convolution):


PROGRAMMABLE FILTER (PFILT)
I (REAL)
ADC A
CORE
DINI [n]
48-TAP FIR
FILTER
xyI [n]
DOUTI [n]
I′ (REAL)
SIGNAL
PROCESSING
BLOCKS
Q (IMAG)
ADC B
CORE
DINQ [n]
48-TAP FIR
FILTER
xyQ [n]
DOUTQ [n]
JESD204B
INTERFACE
Q′ (IMAG)
15660-063


Real 48-tap filter for each I/Q channel (see Figure 92)
 DOUT_I[n] = DIN_I[n] * XY_I[n]
 DOUT_Q[n] = DIN_Q[n] * XY_Q[n]
Real 96-tap filter for on either I or Q channel (see Figure 93)
DOUT_I[n] = DIN_I[n] * XY_I_XY_Q[n]
 DOUT_Q[n] = DIN_Q[n]
Real set of two cascaded 24-tap filters for each I/Q channel
(see Figure 94)
 DOUT_I[n] = DIN_I[n] * X_I[n] * Y_I[n]
 DOUT_Q[n] = DIN_Q[n] * X_Q[n] * Y_Q[n]
Figure 92. Real 48-Tap Filter Configuration
PROGRAMMABLE FILTER (PFILT)
I (REAL)
ADC A
CORE
DINI [n]
96-TAP FIR
FILTER
xIyIxQyQ [n]
DOUTI [n]
I′ (REAL)
SIGNAL
PROCESSING
BLOCKS
Q (IMAG)
ADC B
CORE
DINQ [n]
DOUTQ [n]
Q′ (IMAG)
Figure 93. Real 96-Tap Filter Configuration
Rev. 0 | Page 37 of 135
JESD204B
INTERFACE
15660-064

Half complex filter using two real 48-tap filters for the I/Q
channels (see Figure 95)
 DOUT_I[n] = DIN_I[n]
 DOUT_Q[n] = DIN_Q[n] * XY_Q[n] + DIN_I[n] *
XY_I[n]
Full complex filter using four real 24-tap filters for the I/Q
channels (see Figure 96)
 DOUT_I[n] = DIN_I[n] * X_I[n] + DIN_Q[n] *
Y_Q[n]
 DOUT_Q[n] = DIN_Q[n] * X_Q[n] + DIN_I[n] *
Y_I[n]
AD9695
Data Sheet
PROGRAMMABLE FILTER (PFILT)
I (REAL)
ADC A
CORE
DINI [n]
24-TAP FIR
FILTER
xI [n]
DOUTI [n]
24-TAP FIR
FILTER
yI [n]
I′ (REAL)
SIGNAL
PROCESSING
BLOCKS
JESD204B
INTERFACE
24-TAP FIR
FILTER
yQ [n]
ADC B
CORE
DINQ [n]
24-TAP FIR
FILTER
xQ [n]
DOUTQ [n]
Q′ (IMAG)
15660-065
Q (IMAG)
Figure 94. Real, Two Cascaded, 24-Tap Filter Configuration
PROGRAMMABLE FILTER (PFILT)
I (REAL)
ADC A
CORE
DINI [n]
DOUTI [n]
0 TO 47
DELAY TAPS
48-TAP FIR
FILTER
xyI [n]
I′ (REAL)
SIGNAL
PROCESSING
BLOCKS
JESD204B
INTERFACE
+
ADC B
CORE
DINQ [n]
48-TAP FIR
FILTER
xyQ [n]
+
DOUTQ [n]
Q′ (IMAG)
15660-066
Q (IMAG)
Figure 95. 48-Tap Half Complex Filter Configuration
PROGRAMMABLE FILTER (PFILT)
I (REAL)
ADC A
CORE
DINI [n]
24-TAP FIR
FILTER
xI [n]
DOUTI [n]
+
I′ (REAL)
+
24-TAP FIR
FILTER
yI [n]
SIGNAL
PROCESSING
BLOCKS
JESD204B
INTERFACE
24-TAP FIR
FILTER
yQ [n]
ADC B
CORE
DINQ [n]
24-TAP FIR
FILTER
xQ [n]
+
DOUTQ [n]
Q′ (IMAG)
Figure 96. 24-Tap Full Complex Filter Configuration.
Rev. 0 | Page 38 of 135
15660-067
+
Q (IMAG)
Data Sheet
AD9695
PROGRAMMING INSTRUCTIONS
Table 12. Register 0x0DF8 Definition
Use the following procedure to set up the programmable FIR filter:
Bits
[7:3]
[2:0]
1.
2.
3.
4.
5.
6.
7.
Enable the sample clock to the device.
Configure the mode registers as follows:
a. Set the device index to Channel A (I path)
(Register 0x0008 = 0x01).
b. Set the I path mode (I mode) and gain in
Register 0x0DF8 and Register 0x0DF9 (see Table 12
and Table 13).
c. Set the device index to Channel B (Q path)
(Register 0x0008 = 0x02).
d. Set the Q path mode (Q mode) and gain in
Register 0x0DF8 and Register 0x0DF9.
Wait at least 5 μs to allow the programmable filter to power up.
Program the I path coefficients to the internal shadow
registers as follows:
a. Set the device index to Channel A (I path)
(Register 0x0008 = 0x01).
b. Program the XI coefficients in Register 0x0E00 to
Register 0x0E2F (see Table 14 and Table 15).
c. Program the YI coefficients in Register 0x0F00 to
Register 0x0F2F (see Table 14 and Table 15).
d. Program the tapped delay in Register 0x0F30 (note
that this step is optional).
Program the Q path coefficients to the internal shadow
registers as follows:
a. Set the device index to Channel B (Q path)
(Register 0x0008 = 0x02).
b. Set the Q path mode and gain in Register 0x0DF8 and
Register 0x0DF9 (see Table 12 and Table 13).
c. Program the XQ coefficients in Register 0x0E00 to
Register 0x0E2F (see Table 14 and Table 15).
d. Program the YQ coefficients in Register 0x0F00 to
Register 0x0F2F (see Table 14 and Table 15)
e. Program the tapped delay in Register 0x0F30 (note
that this step is optional).
Set the chip transfer bit using either of the following
methods (note that setting the chip transfer bit applies the
programmed shadow coefficients to the filter):
a. Via the register map by setting the chip transfer bit
(Register 0x000F = 0x01).
b. Via a GPIO pin, as follows:
i. Configure one of the GPIO pins as the chip
transfer bit in Register 0x0040 to Register 0x0042.
ii. Toggle the GPIO pin to initiate the chip transfer
(the rising edge is triggered).
When the I or Q path mode register changes in
Register 0x0DF8, all coefficients must be reprogrammed.
Description
Reserved
Filter mode (I mode or Q mode)
000: filters bypassed
001: real 24-tap filter (X only)
010: real 48-tap filter (X and Y together)
100: real set of two cascaded 24-tap filters (X then Y
cascaded)
101: full complex filter using four real 24-tap filters for the
A/B channels (opposite channel must also be set to 101)
110: half complex filter using two real 48-tap filters +
48-tap delay line (X and Y together) (opposite channel
must also be set to 010)
111: real 96-tap filter (XI, YI, XQ, and YQ together)
(opposite channel must be set to 000)
Table 13. Register 0x0DF9 Definition
Bits
7
[6:4]
3
[2:0]
Description
Reserved
Y filter gain
110: −12 dB loss
111: −6 dB loss
000: 0 dB gain
001: 6 dB gain
010: 12 dB gain
Reserved
X filter gain
110: −12 dB loss
111: −6 dB loss
000: 0 dB gain
001: 6 dB gain
010: 12 dB gain
Table 14 and Table 15 show the coefficient tables in
Register 0x0E00 to Register 0x0F30. All coefficients are
Q1.15 format (sign bit + 15 fractional bits).
Rev. 0 | Page 39 of 135
AD9695
Data Sheet
Table 14. I Coefficient Table (Device Selection = 0x1)1
Addr.
0x0E00
0x0E01
0x0E02
0x0E03
…
0x0E2E
0x0E2F
0x0F00
0x0F01
0x0F02
0x0F03
…
0x0F2E
0x0F2F
0x0F30
Single 24-Tap
Filter (I Mode
[2:0] = 0x1)
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
Unused
Unused
Unused
Unused
…
Unused
Unused
Unused
Single 48-Tap
Filter (I Mode
[2:0] = 0x2)
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
YI C24 [7:0]
YI C24 [15:8]
YI C25 [7:0]
YI C25 [15:8]
…
YI C47 [7:0]
YI C47 [15:0]
Unused
Two Cascaded
24-Tap Filters (I
Mode [2:0] = 0x4)
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
YI C0 [7:0]
YI C0 [15:8]
YI C1 [7:0]
YI C1 [15:8]
…
YI C23 [7:0]
YI C23 [15:0]
Unused
Full Complex
24-Tap Filters (I
Mode [2:0] = 0x5
and Q Mode
[2:0] = 0x5)
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
YI C0 [7:0]
YI C0 [15:8]
YI C1 [7:0]
YI C1 [15:8]
…
YI C23 [7:0]
YI C23 [15:0]
Unused
1
Half Complex
48-Tap Filters (I
Mode [2:0] = 0x6
and Q Mode
[2:0] = 0x2)2
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
YI C24 [7:0]
YI C24 [15:8]
YI C25 [7:0]
YI C25 [15:8]
…
YI C47 [7:0]
YI C47 [15:0]
I path tapped delay
0: 0 tapped delay
(matches C0 in the
filter)
1: 1 tapped delays
…
47: 47 tapped delays
XI Cn means I Path X Coefficient n. YI Cn means I Path Y Coefficient n.
When using the I path in half-complex 48-tap filter mode, the Q path must be in single 48-tap filter mode.
3
When using the I path in 96-tap filter mode, the Q path must be in bypass mode.
2
Rev. 0 | Page 40 of 135
I Path 96-Tap Filter
(I Mode[2:0] = 0x7
and Q Mode
[2:0] = 0x0)3
XI C0 [7:0]
XI C0 [15:8]
XI C1 [7:0]
XI C1 [15:8]
…
XI C23 [7:0]
XI C23 [15:0]
YI C24 [7:0]
YI C24 [15:8]
YI C25 [7:0]
YI C25 [15:8]
…
YI C47 [7:0]
YI C47 [15:0]
Unused
Q Path 96-Tap
Filter (I Mode
[2:0] = 0x0 and Q
Mode [2:0] = 0x7)3
XQ C48 [7:0]
XQ C48 [15:8]
XQ C49 [7:0]
XQ C49 [15:8]
…
XQ C71 [7:0]
XQ C71 [15:0]
YQ C72 [7:0]
YQ C72 [15:8]
YQ C73 [7:0]
YQ C73 [15:8]
…
YQ C95 [7:0]
YQ C95 [15:0]
Unused
Data Sheet
AD9695
Table 15. Q Coefficient Table (Device Selection = 0x2)1
Addr.
0x0E00
0x0E01
0x0E02
0x0E03
…
0x0E2E
0x0E2F
0x0F00
0x0F01
0x0F02
0x0F03
…
0x0F2E
0x0F2F
0x0F30
1
2
3
Single 24-Tap
Filter (Q Mode
[2:0] = 0x1)
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
Unused
Unused
Unused
Unused
…
Unused
Unused
Unused
Single 48-Tap
Filter (Q Mode
[2:0] = 0x2)
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
YQ C24 [7:0]
YQ C24 [15:8]
YQ C25 [7:0]
YQ C25 [15:8]
…
YQ C47 [7:0]
YQ C47 [15:0]
Unused
Two Cascaded
24-Tap Filters (Q
Mode [2:0] = 0x4)
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
YQ C0 [7:0]
YQ C0 [15:8]
YQ C1 [7:0]
YQ C1 [15:8]
…
YQ C23 [7:0]
YQ C23 [15:0]
Unused
Full Complex
24-Tap Filters (Q
Mode [2:0] = 0x5
and I Mode
[2:0] = 0x5)
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
YQ C0 [7:0]
YQ C0 [15:8]
YQ C1 [7:0]
YQ C1 [15:8]
…
YQ C23 [7:0]
YQ C23 [15:0]
Unused
Half Complex
48-Tap Filters (Q
Mode [2:0] = 0x6
and I Mode
[2:0] = 0x2)2
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
YQ C24 [7:0]
YQ C24 [15:8]
YQ C25 [7:0]
YQ C25 [15:8]
…
YQ C47 [7:0]
YQ C47 [15:0]
Q path tapped
delay
0: 0 tapped delay
(matches C0 in the
filter)
1: 1 tapped delays
…
47: 47 tapped
delays
XQ Cn means Q Path X Coefficient n. YQ Cn means Q Path Y Coefficient n.
When using the I path in half-complex 48-tap filter mode, the Q path must be in single 48-tap filter mode.
When using the I path in 96-tap filter mode, the Q path must be in bypass mode.
Rev. 0 | Page 41 of 135
I Path 96-Tap
Filter (Q Mode
[2:0] = 0x0 and I
Mode [2:0] = 0x7)3
XI C48 [7:0]
XI C48 [15:8]
XI C49 [7:0]
XI C49 [15:8]
…
XI C71 [7:0]
XI C71 [15:0]
YI C72 [7:0]
YI C72 [15:8]
YI C73 [7:0]
YI C73 [15:8]
…
YI C95 [7:0]
YI C95 [15:0]
Unused
Q Path 96-Tap
Filter (Q Mode
[2:0] = 0x7 and I
Mode [2:0] = 0x0)3
XQ C0 [7:0]
XQ C0 [15:8]
XQ C1 [7:0]
XQ C1 [15:8]
…
XQ C23 [7:0]
XQ C23 [15:0]
YQ C24 [7:0]
YQ C24 [15:8]
YQ C25 [7:0]
YQ C25 [15:8]
…
YQ C47 [7:0]
YQ C47 [15:0]
Unused
AD9695
Data Sheet
DIGITAL DOWNCONVERTER (DDC)
The AD9695 includes four digital downconverters (DDC 0 to
DDC 3) that provide filtering and reduce the output data rate.
This digital processing section includes an NCO, multiple
decimating FIR filters, a gain stage, and a complex to real
conversion stage. Each of these processing blocks has control lines
that allow it to be independently enabled and disabled to provide
the desired processing function. The digital downconverter can be
configured to output either real data or complex output data.
to use both DDC Output Port I and DDC Output Port Q. For
more information, see Figure 130.
The DDCs output a 16-bit stream. To enable this operation, the
converter number of bits, N, is set to a default value of 16, even
though the analog core only outputs 14 bits. In full bandwidth
operation, the ADC outputs are the 14-bit word followed by two
zeros, unless the tail bits are enabled.
Each DDC block contains the following signal processing stages:
DDC I/Q INPUT SELECTION
DDC GENERAL DESCRIPTION
The four DDC blocks extract a portion of the full digital
spectrum captured by the ADC(s). They are intended for IF
sampling or oversampled baseband radios requiring wide
bandwidth input signals.




Frequency translation stage (optional)
Filtering stage
Gain stage (optional)
Complex to real conversion stage (optional)
Frequency Translation Stage (Optional)
The AD9695 has two ADC channels and four DDC channels.
Each DDC channel has two input ports that can be paired to
support both real and complex inputs through the I/Q crossbar
mux. For real signals, both DDC input ports must select the
same ADC channel (that is, DDC Input Port I = ADC Channel A
and DDC Input Port Q = ADC Channel A). For complex
signals, each DDC input port must select different ADC
channels (that is, DDC Input Port I = ADC Channel A and
DDC Input Port Q = ADC Channel B).
The inputs to each DDC are controlled by the DDC input selection registers (Register 0x0311, Register 0x0331, Register 0x0351
and Register 0x0371). See Table 47 for information on how to
configure the DDCs.
DDC I/Q OUTPUT SELECTION
This stage consists of a phase coherent NCO and quadrature
mixers that can be used for frequency translation of both real or
complex input signals. The phase coherent NCO allows an
infinite number of frequency hops that are all referenced back
to a single synchronization event. It also includes 16 shadow
registers for fast switching applications. This stage shifts a
portion of the available digital spectrum down to baseband.
Filtering Stage
After shifting down to baseband, this stage decimates the
frequency spectrum using multiple low pass finite impulse
response (FIR) filters for rate conversion. The decimation
process lowers the output data rate, which in turn reduces the
output interface rate.
Gain Stage (Optional)
Each DDC channel has two output ports that can be paired to
support both real and complex outputs. For real output signals,
only the DDC Output Port I is used (the DDC Output Port Q is
invalid). For complex I/Q output signals, both DDC Output
Port I and DDC Output Port Q are used.
The I/Q outputs to each DDC channel are controlled by the
DDC complex to real enable bit, Bit 3, in the DDC control
registers (Register 0x0310, Register 0x0330, Register 0x0350 and
Register 0x370).
The chip Q ignore bit in the chip mode register (Register 0x0200,
Bit 5) controls the chip output muxing of all the DDC channels.
When all DDC channels use real outputs, set this bit high to
ignore all DDC Q output ports. When any of the DDC channels
are set to use complex I/Q outputs, the user must clear this bit
Due to losses associated with mixing a real input signal down to
baseband, this stage compensates by adding an additional 0 dB
or 6 dB of gain.
Complex to Real Conversion Stage (Optional)
When real outputs are necessary, this stage converts the complex
outputs back to real by performing an fS/4 mixing operation
plus a filter to remove the complex component of the signal.
Figure 97 shows the detailed block diagram of the DDCs
implemented in the AD9695.
Figure 98 shows an example usage of one of the four DDC
channels with a real input signal and four half-band filters
(HB4 + HB3 + HB2 + HB1) used. It shows both complex
(decimate by 16) and real (decimate by 8) output options.
Rev. 0 | Page 42 of 135
Data Sheet
AD9695
COMPLEX TO REAL
CONVERSION (OPTIONAL)
COMPLEX TO REAL
CONVERSION (OPTIONAL)
REAL/I
CONVERTER 2
Q CONVERTER 3
I
DECIMATION
FILTERS
Q
REAL/I
CONVERTER 4
JESD204B TRANSMIT INTERFACE
DECIMATION
FILTERS
Q
ADC B
SAMPLING
AT fS
L
JESD204B
LANES
AT UP TO
16Gbps
Q CONVERTER 5
DDC 3
I
NCO
+
MIXER
(OPTIONAL)
REAL/I
DECIMATION
FILTERS
Q
SYSREF±
REAL/I
CONVERTER 6
Q CONVERTER 7
SYSREF
DCM = DECIMATION
15660-068
NCO CHANNEL
SELECTION
CIRCUITS
COMPLEX TO REAL
CONVERSION (OPTIONAL)
REAL/I
REAL/I
GPIO PINS
Q CONVERTER 1
DDC 2
REAL/I
REGISTER MAP
CONTROLS
COMPLEX TO REAL
CONVERSION (OPTIONAL)
I/Q CROSSBAR MUX
REAL/I
NCO
+
MIXER
(OPTIONAL)
SYNCHRONIZATION
CONTROL CIRCUITS
GAIN = 0 OR +6dB
I
NCO
+
MIXER
(OPTIONAL)
SYSREF
PIN
REAL/I
CONVERTER 0
DDC 1
REAL/I
REAL/Q
DECIMATION
FILTERS
Q
ADC A
SAMPLING
AT fS
REAL/I
GAIN = 0 OR +6dB
I
NCO
+
MIXER
(OPTIONAL)
REAL/I
GAIN = 0 OR +6dB
REAL/I
GAIN = 0 OR +6dB
DDC 0
NCO CHANNEL SELECTION
Figure 97. DDC Detailed Block Diagram
Rev. 0 | Page 43 of 135
AD9695
Data Sheet
ADC
–fS/2
–fS/3
ADC
SAMPLING
AT fS
REAL
REAL INPUT—SAMPLED AT fS
BANDWIDTH OF
INTEREST IMAGE
–fS/4
REAL
BANDWIDTH OF
INTEREST
–fS/32
fS/32
DC
–fS/16
fS/16
–fS/8
fS/8
fS/4
fS/3
fS/2
FREQUENCY TRANSLATION STAGE (OPTIONAL)
I
DIGITAL MIXER + NCO
FOR fS/3 TUNING, THE FREQUENCY TUNING WORD = ROUND
((fS/3)/fS × 248) = +9.382513
(0x5555_5555_5555)
NCO TUNES CENTER OF
BANDWIDTH OF INTEREST
TO BASEBAND
cos(ωt)
REAL
48-BIT
NCO
90°
0°
–sin(ωt)
Q
DIGITAL FILTER
RESPONSE
–fS/3
–fS/4
–fS/32
fS/32
DC
–fS/16
fS/16
–fS/8
FILTERING STAGE
HB4 FIR
4 DIGITAL HALF-BAND FILTERS
(HB4 + HB3 + HB2 + HB1)
I
HALFBAND
FILTER
Q
HALFBAND
FILTER
HB3 FIR
2
HALFBAND
FILTER
2
HALFBAND
FILTER
HB4 FIR
fS/8
HB2 FIR
2
HALFBAND
FILTER
2
HALFBAND
FILTER
HB3 FIR
fS/4
fS/3
fS/2
HB1 FIR
2
HB2 FIR
HALFBAND
FILTER
I
HB1 FIR
2
HALFBAND
FILTER
Q
6dB GAIN TO
COMPENSATE FOR
NCO + MIXER LOSS
DIGITAL FILTER
RESPONSE
0dB OR 6dB GAIN
I
GAIN STAGE (OPTIONAL)
Q
0dB OR 6dB GAIN
COMPLEX TO REAL
CONVERSION STAGE (OPTIONAL)
–fS/32
fS/32
DC
–fS/16
fS/16
–fS/8
COMPLEX (I/Q) OUTPUTS
DECIMATE BY 16
GAIN STAGE (OPTIONAL)
fS/8
fS/4 MIXING + COMPLEX FILTER TO REMOVE Q
2
+6dB
2
+6dB
I
Q
–fS/32
fS/32
DC
–fS/16
fS/16
DOWNSAMPLE BY 2
I
REAL (I) OUTPUTS
+6dB
I
DECIMATE BY 8
Q
+6dB
Q
COMPLEX REAL/I
TO
REAL
6dB GAIN TO
COMPENSATE FOR
NCO + MIXER LOSS
–fS/8
–fS/32
fS/32
DC
–fS/16
fS/16
fS/8
Figure 98. DDC Theory of Operation Example (Real Input)
Rev. 0 | Page 44 of 135
15660-069
–fS/2
BANDWIDTH OF
INTEREST IMAGE
(–6dB LOSS DUE TO
NCO + MIXER)
BANDWIDTH OF INTEREST
(–6dB LOSS DUE TO
NCO + MIXER)
Data Sheet
AD9695
DDC FREQUENCY TRANSLATION
Variable IF Mode
DDC Frequency Translation General Description
NCO and mixers are enabled. NCO output frequency can be
used to digitally tune the IF frequency.
Frequency translation is accomplished by using a 48-bit
complex NCO with a digital quadrature mixer. This stage
translates either a real or complex input signal from an IF to a
baseband complex digital output (carrier frequency = 0 Hz).
0 Hz IF (ZIF) Mode
The mixers are bypassed, and the NCO is disabled.
fS/4 Hz IF Mode
The frequency translation stage of each DDC can be controlled
individually and supports four different IF modes using Bits[5:4]
of the DDC control registers (Register 0x0310, Register 0x0330,
Register 0x0350, and Register 0x0370). These IF modes are as
follows:
Test Mode
Input samples are forced to 0.999 to positive full scale. The
NCO is enabled. This test mode allows the NCOs to directly
drive the decimation filters.
Variable IF mode
0 Hz IF or zero IF (ZIF) mode
fS/4 Hz IF mode
Test mode
Figure 99 and Figure 100 show examples of the frequency
translation stage for both real and complex inputs.
NCO FREQUENCY TUNING WORD (FTW) SELECTION
48-BIT NCO FTW = MIXING FREQUENCY/ADC SAMPLE RATE × 4096
I
ADC + DIGITAL MIXER + NCO
REAL INPUT—SAMPLED AT fS
REAL
cos(ωt)
ADC
SAMPLING
AT fS
REAL
48-BIT
NCO
90°
0°
COMPLEX
–sin(ωt)
Q
BANDWIDTH OF
INTEREST
BANDWIDTH OF
INTEREST IMAGE
–fS/2
–fS/3
–fS/4
–fS/8
fS/32
–fS/32
DC
–fS/16
fS/16
fS/8
fS/4
fS/3
fS/2
–6dB LOSS DUE TO
NCO + MIXER
48-BIT NCO FTW =
ROUND ((fS/3)/fS × 248) = +9.382513
(0x5555_5555_5555)
POSITIVE FTW VALUES
–fS/32
DC
fS/32
48-BIT NCO FTW =
ROUND ((fS/3)/fS × 248) = –9.382513
(0xAAAA_AAAA_AAAA)
NEGATIVE FTW VALUES
–fS/32
DC
fS/32
Figure 99. DDC NCO Frequency Tuning Word Selection—Real Inputs
Rev. 0 | Page 45 of 135
15660-070




The mixers and the NCO are enabled in special downmixing by
fS/4 mode to save power.
AD9695
Data Sheet
NCO FREQUENCY TUNING WORD (FTW) SELECTION
48-BIT NCO FTW = MIXING FREQUENCY/ADC SAMPLE RATE × 248
QUADRATURE ANALOG MIXER +
2 ADCs + QUADRATURE DIGITAL
MIXER + NCO
QUADRATURE MIXER
ADC
SAMPLING
AT fS
I
+
I
I
Q
Q
REAL
90°
PHASE
48-BIT
NCO
90°
0°
COMPLEX INPUT—SAMPLED AT fS
Q
Q
ADC
SAMPLING
AT fS
Q
Q
I
I
–
–sin(ωt)
I
I
+
COMPLEX
Q
+
BANDWIDTH OF
INTEREST
IMAGE DUE TO
ANALOG I/Q
MISMATCH
–fS/3
–fS/4
–fS/8
fS/32
–fS/32
–fS/16
fS/16
DC
fS/8
fS/4
fS/3
fS/2
48-BIT NCO FTW =
ROUND ((fS/3)/fS × 248) = +9.382513
(0x5555_5555_5555)
POSITIVE FTW VALUES
–fS/32
fS/32
DC
Figure 100. DDC NCO Frequency Tuning Word Selection—Complex Inputs
Rev. 0 | Page 46 of 135
15660-071
–fS/2
Data Sheet
AD9695
DDC NCO Description
DDC NCO Coherent Mode
Each DDC contains one NCO. Each NCO enables the
frequency translation process by creating a complex exponential
frequency (e-jωct), which can be mixed with the input spectrum
to translate the desired frequency band of interest to dc, where
it can be filtered by the subsequent low-pass filter blocks to
prevent aliasing.
This mode allows an infinite number of frequency hops where the
phase is referenced to a single synchronization event at time 0.
This mode is useful when phase coherency must be maintained
when switching between different frequency bands. In this mode,
the user can switch to any tuning frequency without the need to
reset the NCO. Although only one FTW is required, the NCO
contains 16 shadow registers for fast-switching applications.
Selection of the shadow registers is controlled by the CMOS
GPIO pins or through the register map of the SPI. In this mode,
the NCO can be set up by providing the following:
When placed in variable IF mode, the NCO supports two
different additional modes.
DDC NCO Programmable Modulus Mode
This mode supports >48-bit frequency tuning accuracy for
applications that require exact rational (M/N) frequency
synthesis at a single carrier frequency. In this mode, the NCO is
set up by providing the following:







Figure 73 shows a block diagram of one NCO and its
connection to the rest of the design. The coherent phase
accumulator block contains the logic that allows an infinite
number of frequency hops.
48-bit frequency tuning word (FTW)
48-bit Modulus A word (MAW)
48-bit Modulus B word (MBW)
48-bit phase offset word (POW)
NCO
NCO CHANNEL
SELECTION
FTW/POW
WRITE INDEX
SYNCHRONIZATION
CONTROL CIRCUITS
I/O
CROSSBAR
MUX
0
48-BIT
FTW/POW
0
1
48-BIT
FTW/POW
1
48-BIT
FTW/POW
15
15
COHERENT
PHASE
ACCUMULATOR
BLOCK
COS/SIN
GENERATOR
SYSREF
I
I
Q
Q
DIGITAL
QUADRATURE
MIXER
FTW = FREQUENCY TUNING WORD
POW = PHASE OFFSET WORD
MAW = MODULUS A WORD (NUMERATOR)
MBW = MODULUS B WORD (DENOMINATOR)
Figure 101. NCO + Mixer Block Diagram
Rev. 0 | Page 47 of 135
DECIMATION
FILTERS
15660-072
FTW/POW
REGISTER
MAP
MODULUS
ERROR
48-BIT
MAW/MBW
–sin(x)
MAW/MBW
cos(x)
NCO
CHANNEL
SELECTION
CIRCUITS
Up to sixteen 48-bit FTWs.
Up to sixteen 48-bit POWs.
The 48-bit MAW must be set to zero in coherent mode.
AD9695
Data Sheet
NCO FTW/POW/MAW/MAB Description
The NCO frequency value is determined by the following settings:



48-bit twos complement number entered in the FTW
48-bit unsigned number entered in the MAW
48-bit unsigned number entered in the MBW
M and N are integers reduced to their lowest terms. MAW and
MBW are integers reduced to their lowest terms. When MAW is
set to zero, the programmable modulus logic is automatically
disabled.
Frequencies between −fS/2 and +fS/2 (fS/2 excluded) are
represented using the following values:



Equation 1 to Equation 4 apply to the aliasing of signals in the
digital domain (that is, aliasing introduced when digitizing
analog signals).
FTW = 0x8000_0000_0000 and MAW = 0x0000_0000_0000
represents a frequency of –fS/2.
FTW = 0x0000_0000_0000 and MAW = 0x0000_0000_0000
represents dc (frequency is 0 Hz).
FTW = 0x7FFF_FFFF_FFFF and MAW = 0x0000_0000_0000
represents a frequency of +fS/2.
For example, if the ADC sampling frequency (fS) is 625 MSPS
and the carrier frequency (fC) is 208.6 MHz, then,
2089
M
mod  417 . 8 ,1300 


1300
6250
N
mod2417.8,1300 

FTW  floor 248

1300


= 0x5590_C0AD_03D9
NCO FTW/POW/MAW/MAB Programmable
Modulus Mode
MAW = mod(248 × 2089, 6250) = 0x0000_0000_1117
For programmable modulus mode, the MAW must be set to a
nonzero value (not equal to 0x0000_0000_0000). This mode is
only needed when frequency accuracy of >48 bits is required.
One example of a rational frequency synthesis requirement that
requires >48 bits of accuracy is a carrier frequency of 1/3 the
sample rate. When frequency accuracy of ≤48 bits is required,
coherent mode must be used (see the NCO FTW/POW/MAW/
MAB Coherent Mode section).
MBW = 0x0000_0000_186A
The actual carrier frequency can be calculated based on the
following equation:
f C _ ACTUAL 
FTW  floor(248
FTW 
2
MAW
MBW
48
mod( fc , f s )
fs
)
(1)
(2)
MAW = mod(248 × M, N)
(3)
MBW = N
(4)
where:
fC is the desired carrier frequency.
fS is the ADC sampling frequency.
M is the integer representing the rational numerator of the
frequency ratio.
N is the integer representing the rational denominator of the
frequency ratio.
FTW is the 48-bit twos complement number representing the
NCO FTW.
MAW is the 48-bit unsigned number representing the NCO
MAW (must be <247).
MBW is the 48-bit unsigned number representing the NCO MBW.
mod(x) is a remainder function. For example mod(110,100) =
10 and for negative numbers, mod(–32,10)= –2.
floor(x) is defined as the largest integer less than or equal to x.
For example, floor(3.6) = 3.
MAW
f
MBW S
2 48
For the previous example, the actual carrier frequency (fC_ACTUAL) is
fC _ ACTUAL
In programmable modulus mode, the FTW, MAW, and MBW
must satisfy the following four equations (for a detailed
description of the programmable modulus feature, see the DDS
architecture described in the AN-953 Application Note):
mod( f c , f s ) M


fs
N
FTW 
0x5590_C0AD_03D9 

 417.8MHz
2 48
0x0000_0000_1117
0x0000_0000_186A
 1300 MHz
A 48-bit POW is available for each NCO to create a known
phase relationship between multiple chips or individual DDC
channels inside the chip.
While in programmable modulus mode, the FTW and POW
registers can be updated at any time while still maintaining
deterministic phase results in the NCO. However, the following
procedure must be followed to update the MAW and/or MBW
registers to ensure proper operation of the NCO:
1.
2.
Rev. 0 | Page 48 of 135
Write to the MAW and MBW registers for all the DDCs.
Synchronize the NCOs either through the DDC soft reset
bit accessible through the SPI or through the assertion of
the SYSREF± pin (see the Memory Map section).
Data Sheet
AD9695
For the previous example, the actual carrier frequency (fC_ACTUAL) is
NCO FTW/POW/MAW/MAB Coherent Mode
For coherent mode, the NCO MAW must be set to zero
(0x0000_0000_0000). In this mode, the NCO FTW can be
calculated by the following equation:
FTW  round(2 48
mod( f c , f s )
)
fs
fC_ACTUAL =
(5)
where:
FTW is the 48-bit twos complement number representing the
NCO FTW.
fS is the ADC sampling frequency.
fC is the desired carrier frequency.
mod() is a remainder function. For example mod(110,100) = 10
and for negative numbers, mod(–32,10) = –2.
round() is a rounding function. For example round(3.6) = 4 and
for negative numbers, round(–3.4)= –3.
Note that Equation 5 applies to the aliasing of signals in the
digital domain (that is, aliasing introduced when digitizing
analog signals). The MAW must be set to zero to use coherent
mode. When MAW is zero, the programmable modulus logic is
automatically disabled.
For example, if the ADC sampling frequency (fS) is 1300 MSPS
and the carrier frequency (fC) is 417.3333 MHz, then,
mod(417.33331300 

NCO_FTW = round  2 48
 =
1300


0x5578_49CE_E73F
0x5578_49C
E_E73F 1300
248
A 48-bit POW is available for each NCO to create a known
phase relationship between multiple chips or individual DDC
channels inside the chip.
While in coherent mode, the FTW and POW registers can be
updated at any time while still maintaining deterministic phase
results in the NCO.
NCO Channel Selection
When configured in coherent mode, only one FTW is required
in the NCO. In this mode, the user can switch to any tuning
frequency without the need to reset the NCO by writing to the
FTW directly. However, for fast switching applications, where
either all FTWs are known beforehand or it is possible to queue
up the next set of FTWs, the NCO contains 16 additional
shadow registers (see Figure 73). These shadow registers are
hereafter referred to as the NCO channels.
Figure 71 shows a simplified block diagram of the NCO channel
selection block.
Only one NCO channel is active at a time, and NCO channel
selection is controlled either by the CMOS GPIO pins or
through the register map.
Each NCO channel selector supports three different modes, as
described in the following sections.
The actual carrier frequency can be calculated based on the
following equation:

FTW  f S
2 48
NCO CHANNEL
SELECTION
IN
GPIO
CMOS
PINS
IN
[3:0]
IN
IN
MUX
REGISTER
MAP
[0]
GPIO
SELECTION
COUNTER
INC
NCO CHANNEL SELECTION
NCO
REGISTER MAP NCO
CHANNEL SELECTION
0x0314, 0x0334, 0x0354, 0x0374
NCO CHANNEL MODE
15660-073
f C _ ACTUAL
= 417.33 MHz
Figure 71. NCO Channel Selection Block
Rev. 0 | Page 49 of 135
AD9695
Data Sheet
The following procedure must be followed to use GPIO edge
control for NCO channel selection:
GPIO Level Control Mode
The GPIO pins determine the exact NCO channel selected.
The following procedure must be followed to use GPIO level
control for NCO channel selection:
1.
2.
3.
Configure one or more GPIO pins as NCO channel
selection inputs. GPIO pins not configured as NCO
channel selection are internally tied low.
a. To use GPIO_A0, write Bits[2:0] in Register 0x0040 to
0x6 and Bits[3:0] in Register 0x0041 to 0x0.
b. To use GPIO_B0, write Bits[5:3] in Register 0x0040 to
0x6 and Bits [7:4] in Register 0x0041 to 0x0.
Configure the NCO channel selector in GPIO level control
mode by setting Bits[7:4] in the NCO control registers
(Register 0x0314, Register 0x0334, Register 0x0354, and
Register 0x0374) to 0x1 through 0x6, depending on the
desired GPIO pin ordering.
Select the desired NCO channel through the GPIO pins.
1.
2.
3.
Configure one or more GPIO pins as NCO channel
selection inputs.
a. To use GPIO_A0, write Bits[2:0] in Register 0x0040 to
0x6 and Bits[3:0] in Register 0x0041 to 0x0.
b. To use GPIO_B0, write Bits[5:3] in Register 0x0040 to
0x6 and Bits[7:4] in Register 0x0041 to 0x0.
Configure the NCO channel selector in GPIO edge control
mode by setting Bits[7:4] in the NCO control registers
(Register 0x0314, Register 0x0334, Register 0x0354, and
Register 0x0374) to 0x8 through 0xB, depending on the
desired GPIO Pin.
Configure the wrap point for the NCO channel selection
by setting Bits[3:0] in the NCO control registers
(Register 0x0314, Register 0x0334, Register 0x0354, and
Register 0x0374). A value of 4 causes the channel selection
to wrap at Channel 4 (0, 1, 2, 3, 4, 0, 1, 2, 3, 4, and so on).
Transition the selected GPIO pin from low to high to
increment the NCO channel selection.
GPIO Edge Control Mode
4.
Low to high transition on a single GPIO pin determines the
exact NCO channel selected. The internal channel selection
counter is reset by either SYSREF± or the DDC soft reset.
Register Map Mode
NCO channel selection is controlled directly through the
register map.
Rev. 0 | Page 50 of 135
Data Sheet
AD9695
f0
f1
f2
ACTIVE
DDC
NCO CHANNEL 0
CARRIER FREQUENCY 0
(ACTIVE)
NCO CHANNEL 1
CARRIER FREQUENCY 1
(STANDBY)
NCO CHANNEL 2
CARRIER FREQUENCY 2
(STANDBY)
fS/2
15660-074
DC
B2
B1
B0
Figure 72. NCO Coherent Mode with Three NCO Channels (B0 Selected)
Figure 72 shows an example use case for coherent mode
utilizing three NCO channels. In this example, NCO Channel 0 is
actively downconverting bandwidth 0 (B0) while NCO Channel 1
and Channel 2 are in standby and tuned to Bandwidth 1 and
Bandwidth 2 (B1 and B2), respectively.
The phase coherent NCO switching feature allows an infinite
number of frequency hops that are all phase coherent. The initial
phase of the NCO is established at time t0 from SYSREF±
synchronization. Switching the NCO FTW does not affect the
phase. With this feature, only one FTW is required; however, the
user may want to use all 16 channels to queue up the next hop.
After SYSREF± synchronization at start-up, all NCOs across
multiple chips are inherently synchronized.
Setting Up the Multichannel NCO Feature
The first step to configure the multichannel NCO is to program
the FTWs. The AD9695 memory map has a FTW index register
for each DDC. This index determines which NCO channel
receives the FTW from the register map. The following
sequence describes the method for programming the FTWs.
1.
2.
3.
Write the FTW index register with the desired DDC channel.
Write the FTW with the desired value. This value is applied
to the NCO channel index mentioned in Step 1.
Repeat Step 1 and Step 2 for other NCO channels.
After setting the FTWs, the user must then select an active
NCO channel. This selection can be done either through the
SPI registers or through the external GPIO pins. The following
sequence describes the method for selecting the active NCO
channel using SPI.
1.
2.
The following sequence describes the method for selecting the
active NCO channel using GPIO CMOS pins.
1.
2.
3.
Set NCO channel selection mode (Bits[7:4]) in
Register 0x0314, Register 0x0334, Register 0x0354, and
Register 0x0374 to a nonzero value to enable GPIO pin
selection.
Configure the GPIO pins as NCO channel selection inputs
by writing to Register 0x0040, Register 0x0041, and
Register 0x0042.
NCO switching is done by externally controlling the GPIO
CMOS pins.
NCO Synchronization
Each NCO contains a separate phase accumulator word (PAW).
The initial reset value of each PAW is set to zero and increments
every clock cycle. The instantaneous phase of the NCO is
calculated using the PAW, FTW, MAW, MBW, and POW. Due to
this architecture, the FTW and POW registers can be updated at
any time while still maintaining deterministic phase results in
the PAW of the NCO.
Two methods can be used to synchronize multiple PAWs within
the chip:


Set the NCO channel selection mode (Bits[7:4]) in
Register 0x0314, Register 0x0334, Register 0x0354, and
Register 0x0374 to 0x0 to enable SPI selection.
Choose the active NCO channel (Bits[3:0]) in Register 0x0314,
Register 0x0334, Register 0x0354, and Register 0x0374.
Rev. 0 | Page 51 of 135
Using the SPI. Use the DDC soft reset bit in the DDC
synchronization control register (Register 0x0300, Bit 4) to
reset all the PAWs in the chip. This reset is accomplished
by setting the DDC soft reset bit high, and then setting this
bit low. Note that this method can only be used to
synchronize DDC channels within the same chip.
Using the SYSREF± pin. When the SYSREF± pin is enabled
in the SYSREF control registers (Register 0x0120 and
Register 0x0121), and the DDC synchronization is enabled
in the DDC synchronization control register (Register 0x0300,
Bits[1:0]), any subsequent SYSREF± event resets all the
PAWs in the chip. Note that this method can be used to
synchronize DDC channels within the same chip or DDC
channels within separate chips.
AD9695
Data Sheet
NCO Multichip Synchronization
MNTO
ADC DEVICE 0
(MASTER)
In some applications, it is necessary to synchronize all the NCOs
and local multiframe clocks (LMFCs) within multiple devices in
a system. For applications requiring multiple NCO tuning
frequencies in the system, a designer likely needs to generate a
single SYSREF± pulse at all devices simultaneously. For many
systems, generating or receiving a single-shot SYSREF± pulse at
all devices is challenging because of the following factors:

ADC DEVICE 1
(SLAVE)
SNTI
Enabling or disabling the SYSREF± pulse is often an
asynchronous event.
Not all clock generation chips support this feature.
SNTI
For these reasons, the AD9695 contains a synchronization
triggering mechanism that allows the following:


Multichip synchronization of all NCOs and LMFCs at
system startup.
Multichip synchronization of all NCOs after applying new
tuning frequencies during normal operation.
The synchronization triggering mechanism uses a master/slave
arrangement, as shown in Figure 104.
ADC DEVICE 2
(SLAVE)
ADC DEVICE 3
(SLAVE)
1 LINK,
L LANES
1 LINK,
L LANES
1 LINK,
L LANES
SYSREF±
DEVICE_CLOCK±
CLOCK
GENERATION
MNTO = MASTER NEXT TRIGGER OUTPUT (CMOS)
SNTI = SLAVE NEXT TRIGGER INPUT (CMOS)
15660-075

SNTI
1 LINK,
L LANES
Figure 104. System Using Master/Slave Synchronization Triggering
Each device has an internal next synchronization trigger enable
(NSTE) signal that controls whether the next SYSREF± signal
causes a synchronization event. Slave ADC devices must source
their NSTE from an external slave next trigger input (SNTI) pin.
Master devices can either use an external master next trigger
output (MNTO) pin (default setting), or use an external SNTI pin.
See Table 47 (Register 0x0041 and Register 0x0042) to configure
the FD/GPIO pins for this operation.
NCO Multichip Synchronization at Startup
Figure 105 shows a timing diagram along with the required
sequence of events for NCO multichip synchronization using
triggering and SYSREF± at startup. Using this startup sequence
synchronizes all the NCOs and LMFCs in the system at once.
NCO Multichip Synchronization During Normal Operation
See the NCO Multichip Synchronization section.
Rev. 0 | Page 52 of 135
Data Sheet
CONFIGURE MASTER
AND SLAVE DEVICES
AD9695
ENABLE TRIGGER IN
MASTER DEVICES
MNTO SET HIGH
SYSTEM
SYNCHRONIZATION
ACHIEVED
SNTI SET HIGH
SYSREF
IGNORED
DEVICE
CLOCK
SYSREF
MNTO
BOARD PROPAGATION
DELAY
SNTI
INPUT DELAY
NSTE
LMFCs
DON’T CARE
NCOs
DON’T CARE
LMFC
SYNCHRONIZED
NCO
SYNCHRONIZED
15660-076
MNTO = MASTER NEXT TRIGGER OUTPUT (CMOS)
SNTI = SLAVE NEXT TRIGGER INPUT (CMOS)
NSTE = NEXT SYNCHRONIZATION TRIGGER ENABLE
LMFC = LOCAL MULTIFRAME CLOCK
NCO = NUMERICALLY CONTROLLED OSCILLATOR
Figure 105. NCO Multichip Synchronization at Startup (Using Triggering and SYSREF)
DDC Mixer Description
When not bypassed (Register 0x0200 ≠ 0x00), the digital
quadrature mixer performs a similar operation to an analog
quadrature mixer. It performs the downconversion of input
signals (real or complex) by using the NCO frequency as a local
oscillator. For real input signals, a real mixer operation (with
two multipliers) is performed. For complex input signals, a
complex mixer operation (with four multipliers and two adders)
is performed. The selection of real or complex inputs can be
controlled individually for each DDC block using Bit 7 of the
DDC control registers (Register 0x0310, Register 0x0330,
Register 0x0350, and Register 0x0370).
DDC NCO + Mixer Loss and SFDR
When mixing a real input signal down to baseband, −6 dB of
loss is introduced in the signal due to filtering of the negative
image. An additional −0.05 dB of loss is introduced by the
NCO. The total loss of a real input signal mixed down to
baseband is −6.05 dB. For this reason, it is recommended that
the user compensate for this loss by enabling the 6 dB of gain in
the gain stage of the DDC to recenter the dynamic range of the
signal within the full scale of the output bits (see the DDC Gain
Stage section for more information).
When mixing a complex input signal (where I and Q DDC
inputs come from the different ADCs) down to baseband, the
maximum value each I/Q sample can reach is 1.414 × full-scale
after it passes through the complex mixer. To avoid overrange of
the I/Q samples and to keep the data bit widths aligned with
real mixing, −3.06 dB of loss is introduced in the mixer for
complex signals. An additional −0.05 dB of loss is introduced by
the NCO. The total loss of a complex input signal mixed down
to baseband is −3.11 dB.
The worst case spurious signal from the NCO is greater than
102 dBc SFDR for all output frequencies.
DDC DECIMATION FILTERS
After the frequency translation stage, there are multiple
decimation filter stages used to reduce the output data rate.
After the carrier of interest is tuned down to dc (carrier
frequency = 0 Hz), these filters efficiently lower the sample rate,
while providing sufficient alias rejection from unwanted
adjacent carriers around the bandwidth of interest.
Figure 73 shows a simplified block diagram of the decimation
filter stage, and Table 16 describes the filter characteristics of
the different FIR filter blocks.
Table 17 and Table 18 show the different filter configurations
selectable by including different filters. In all cases, the DDC
filtering stage provides 80% of the available output bandwidth,
<±0.005 dB of pass-band ripple, and >100 dB of stop band alias
rejection.
Rev. 0 | Page 53 of 135
AD9695
Data Sheet
DCM = 3
DECIMATION FILTERS
I
DCM = 2
DCM = 3
TB2
FIR
DCM = 2
HB3
FIR
HB2
FIR
I
DCM = 2
FB2
FIR
I
I
Q
Q
DCM = 5
Q
FB2
FIR
I
I
Q
TB2
FIR
DCM = 3
Q
Q
HB3
FIR
HB4
FIR
DCM = 2
HB2
FIR
DCM = 2
I
HB1
FIR
DCM = 5
I
NCO
AND
MIXERS
(OPTIONAL)
HB4
FIR
DCM = 2
DCM = 2
COMPLEX TO REAL CONVERSION
(OPTIONAL)
I
GAIN = 0dB OR +6dB
I
TB1
FIR
HB1
FIR
DCM = 2
Q
Q
Q
TB1
FIR
Q
DCM = 3
15660-077
FIR = FINITE IMPULSE RESPONSE FILTER
DCM = DECIMATION
NOTES
1. TB1 IS ONLY SUPPORTED IN DDC0 AND DDC1
Figure 73. DDC Decimation Filter Block Diagram
Table 16. DDC Decimation Filter Characteristics
Filter Name
HB4
HB3
HB2
HB1
TB2
TB11
FB2
1
Filter Type
FIR low-pass
FIR low-pass
FIR low-pass
FIR low-pass
FIR low-pass
FIR low-pass
FIR low-pass
Decimation
Ratio
2
2
2
2
3
3
5
Pass Band
(rad/sec)
0.1 x π/2
0.2 x π/2
0.4 x π/2
0.8 x π/2
0.4 x π/3
0.8 x π/3
0.4 x π/5
Stop Band
(rad/sec)
1.9 x π/2
1.8 x π/2
1.6 x π/2
1.2 x π/2
1.6 x π/3
1.2 x π/3
1.6 x π/5
TB1 is only supported in DDC0 and DDC1.
Rev. 0 | Page 54 of 135
Pass-Band
Ripple (dB)
<±0.001
<±0.001
<±0.001
<±0.001
<±0.002
<±0.005
<±0.001
Stop-Band
Attenuation (dB)
>100
>100
>100
>100
>100
>100
>100
Data Sheet
AD9695
Table 17. DDC Filter Configurations1
ADC
Sample
Rate
fS
fS
fS
fS
fS
fS
fS
fS
fS
fS
fS
fS
fS
fS
DDC Filter Configuration
HB1
TB13
HB2 + HB1
TB2 + HB1
HB3 + HB2 + HB1
FB2 + HB1
TB2 + HB2 + HB1
FB2 + TB13
HB4 + HB3 + HB2 + HB1
FB2 + HB2 + HB1
TB2 + HB3 + HB2 + HB1
HB2 + FB2 + TB13
FB2 + HB3 + HB2 + HB1
TB2 + HB4 + HB3 + HB2 + HB1
Real (I) Output
Decimation Sample
Ratio
Rate
1
fS
N/A
N/A
2
fS/2
3
fS/3
4
fS/4
5
fS/5
6
fS/6
N/A
N/A
8
fS/8
10
fS/10
12
fS/12
N/A
N/A
20
fS/20
24
fS/24
Complex (I/Q) Outputs
Decimation
Ratio
Sample Rate
2
fS/2 (I) + fS/2 (Q)
3
f fS/3 (I) + fS/3 (Q)
4
fS/4 (I) + fS/4 (Q)
6
fS/6 (I) + fS/6 (Q)
8
fS/8 (I) + fS/8 (Q)
10
fS/10 (I) + fS/10 (Q)
12
fS/12 (I) + fS/12 (Q)
15
fS/15 (I) + fS/15 (Q)
16
fS/16 (I) + fS/16 (Q)
20
fS/20 (I) + fS/20 (Q)
24
fS/24 (I) + fS/24 (Q)
30
fS/30 (I) + fS/30 (Q)
40
fS/40 (I) + fS/40 (Q)
48
fS/48 (I) + fS/48 (Q)
Alias
Protected
Bandwidth
fS/2 × 80%
fS/3 × 80%
fS/4 × 80%
fS/6 × 80%
fS/8 × 80%
fS/10 × 80%
fS/12 × 80%
fS/15 × 80%
fS/16 × 80%
fS/20 × 80%
fS/24 × 80%
fS/30 × 80%
fS/40 × 80%
fS/48 × 80%
Ideal SNR
Improvement
(dB)2
1
2.7
4
5.7
7
8
8.8
9.7
10
11
11.8
12.7
14
14.8
1
N/A means not applicable.
Ideal SNR improvement due to oversampling + filtering > 10log(bandwidth/fS/2).
3
TB1 is only supported in DDC0 and DDC1.
2
Table 18. DDC Filter Configurations (fS = 1300 MSPS)1
ADC Sample
Rate (MSPS)
1300
1300
DDC Filter Configuration
HB1
TB12
Real (I) Output
Decimation Sample Rate
Ratio
(MSPS)
1
1300
N/A
N/A
1300
1300
HB2 + HB1
TB2 + HB1
2
3
650
433.33
1300
1300
1300
HB3 + HB2 + HB1
FB2 + HB1
TB2 + HB2 + HB1
4
5
6
325
260
216.67
1300
1300
1300
1300
1300
1300
1300
FB2 + TB12
HB4 + HB3 + HB2 + HB1
FB2 + HB2 + HB1
TB2 + HB3 + HB2 + HB1
HB2 + FB2 + TB12
FB2 + HB3 + HB2 + HB1
TB2 + HB4 + HB3 + HB2 + HB1
N/A
8
10
12
N/A
20
24
N/A
162.5
130
108.33
N/A
65
54.16
1
2
N/A means not applicable.
TB1 is only supported in DDC0 and DDC1.
Rev. 0 | Page 55 of 135
Complex (I/Q) Outputs
Decimation Sample Rate
Ratio
(MSPS)
2
650 (I) + 650 (Q)
3
433.33 (I) + 433.33
(Q)
4
325 (I) + 325 (Q)
6
216.67 (I) + 216.67
(Q)
8
162.5 (I) + 162.5 (Q)
10
130 (I) + 130 (Q)
12
108.33 (I) + 108.33
(Q)
15
86.67 (I) + 86.67 (Q)
16
81.25 (I) + 81.25 (Q)
20
65 (I) + 65 (Q)
24
54.16 (I) + 54.16 (Q)
30
43.44 (I) + 43.44 (Q)
40
32.5 (I) + 32.5 (Q)
48
27.08 (I) + 27.08 (Q)
Alias-Protected
Bandwidth
(MHz)
520
346.67
260
173.33
130
104
86.67
69.33
65
52
43.33
34.67
26
21.67
AD9695
Data Sheet
Table 18. DDC Filter Configurations (fS = 625 MSPS)1
ADC Sample
Rate (MSPS)
625
625
DDC Filter Configuration
HB1
TB12
Real (I) Output
Decimation Sample Rate
Ratio
(MSPS)
1
625
N/A
N/A
625
HB2 + HB1
2
312.5
625
TB2 + HB1
3
208.33
625
HB3 + HB2 + HB1
4
156.25
625
625
625
625
625
625
625
625
FB2 + HB1
TB2 + HB2 + HB1
FB2 + TB12
HB4 + HB3 + HB2 + HB1
FB2 + HB2 + HB1
TB2 + HB3 + HB2 + HB1
HB2 + FB2 + TB12
FB2 + HB3 + HB2 + HB1
5
6
N/A
8
10
12
N/A
20
125
104.17
N/A
78.125
62.5
52.08
N/A
31.25
625
TB2 + HB4 + HB3 + HB2 + HB1
24
26.04
1
2
Complex (I/Q) Outputs
Decimation Sample Rate
Ratio
(MSPS)
2
312.5 (I) + 312.5 (Q)
3
208.33 (I) + 208.33
(Q)
4
156.25 (I) + 156.25
(Q)
6
104.17 (I) + 104.17
(Q)
8
78.125 (I) + 78.125
(Q)
10
62.5 (I) + 62.5 (Q)
12
52.08 (I) + 52.08 (Q)
15
41.67 (I) + 41.67 (Q)
16
39.06 (I) + 39.06 (Q)
20
31.25 (I) + 31.25 (Q)
24
26.04 (I) + 26.04 (Q)
30
20.83 (I) + 20.83 (Q)
40
15.625 (I) + 15.625
(Q)
48
13.02 (I) + 13.02 (Q)
Alias-Protected
Bandwidth
(MHz)
250
166.67
125
83.33
62.5
50
41.67
33.33
31.25
25
20.83
16.67
12.5
10.42
N/A means not applicable.
TB1 is only supported in DDC0 and DDC1.
20
HB4 Filter Description
0
–20
MAGNITUDE (dB)
The first decimate by 2, half-band, low-pass, FIR filter (HB4) uses
an 11-tap, symmetrical, fixed coefficient filter implementation
that is optimized for low power consumption. The HB4 filter is
only used when complex outputs (decimate by 16) or real outputs
(decimate by 8) are enabled; otherwise, it is bypassed. Table 21
and Figure 107 show the coefficients and response of the HB4 filter.
–80
–100
Decimal
Coefficient (15-Bit)
+99
0
−809
0
+4806
+8192
Rev. 0 | Page 56 of 135
–140
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
Figure 107. HB4 Filter Response
0.9
1.0
15660-078
Normalized
Coefficient
+0.006042
0
−0.049377
0
+0.293335
+0.5
–60
–120
Table 21. HB4 Filter Coefficients
HB4 Coefficient
Number
C1, C11
C2, C10
C3, C9
C4, C8
C5, C7
C6
–40
Data Sheet
AD9695
HB3 Filter Description
Table 23. HB2 Filter Coefficients
The second decimate by 2, half-band, low-pass, FIR filter (HB3)
uses an 11-tap, symmetrical, fixed coefficient filter implementation that is optimized for low power consumption. The HB3
filter is only used when complex outputs (decimate by 8 or 16)
or real outputs (decimate by 4 or 8) are enabled; otherwise, it is
bypassed. Table 22 and Figure 108 show the coefficients and
response of the HB3 filter.
HB2 Coefficient
Number
C1, C19
C2, C18
C3, C17
C4, C16
C5, C15
C6, C14
C7, C13
C8, C12
C9, C11
C10
HB3 Coefficient
Number
C1, C11
C2, C10
C3, C9
C4, C8
C5, C7
C6
Normalized
Coefficient
+0.006638
0
−0.051056
0
+0.294418
+0.500000
Decimal Coefficient
(17-Bit)
+435
0
−3346
0
+19,295
+32,768
20
0
0
–20
–40
–60
–80
–100
–40
–120
–60
–140
–80
–160
–100
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
–120
Figure 109. HB2 Filter Response
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
0.9
1.0
15660-079
–140
Figure 108. HB3 Filter Response
HB2 Filter Description
The third decimate by 2, half-band, low-pass, FIR filter (HB2)
uses a 19-tap, symmetrical, fixed coefficient filter implementation that is optimized for low power consumption.
The HB2 filter is only used when complex or real outputs
(decimate by 4, 8, or 16) is enabled; otherwise, it is bypassed.
Table 23 and Figure 109 show the coefficients and response of
the HB2 filter.
Rev. 0 | Page 57 of 135
0.9
1.0
15660-080
MAGNITUDE (dB)
–20
Decimal Coefficient
(18-Bit)
+88
0
−698
0
+2981
0
−9723
0
+40120
+65536
20
MAGNITUDE (dB)
Table 22. HB3 Filter Coefficients
Normalized
Coefficient
+0.000671
0
−0.005325
0
+0.022743
0
−0.074181
0
+0.306091
+0.5
AD9695
Data Sheet
20
HB1 Filter Description
0
–20
MAGNITUDE (dB)
Table 24. HB1 Filter Coefficients
Decimal Coefficient
(20-Bit)
−10
0
+38
0
−102
0
+232
0
−467
0
+862
0
−1489
0
+2440
0
−3833
0
+5831
0
−8679
0
12803
0
−19086
0
+29814
0
−53421
0
+166138
+262144
–60
–80
–100
–120
–140
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
NORMALIZED FREQUENCY (× Π RAD/s)
Figure 110. HB1 Filter Response
TB2 Filter Description
The TB2 uses a 26-tap, symmetrical, fixed coefficient filter
implementation that is optimized for low power consumption.
The TB2 filter is only used when decimation ratios of 6, 12, or
24 are required. Table 25 and Figure 111 show the coefficients
and response of the TB2 filter.
Table 25. TB2 Filter Coefficients
TB2 Coefficient
Number
C1, C26
C2, C25
C3, C24
C4, C23
C5, C22
C6, C21
C7, C20
C8, C19
C9, C18
C10, C17
C11, C16
C12, C15
C13, C14
Normalized
Coefficient
−0.000191
−0.000793
−0.001137
+0.000916
+0.006290
+0.009823
+0.000916
−0.023483
−0.043152
−0.019318
+0.071327
+0.201172
+0.297756
Decimal Coefficient
(19-Bit)
−50
+208
−298
+240
+1649
+2575
+240
−6156
−11312
−5064
+18698
+52736
+78055
20
0
–20
–40
–60
–80
–100
–120
–140
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
Figure 111. TB2 Filter Response
Rev. 0 | Page 58 of 135
0.9
1.0
15660-082
Normalized
Coefficient
−0.000019
0
+0.000072
0
−0.000195
0
+0.000443
0
−0.000891
0
+0.001644
0
−0.002840
0
+0.004654
0
−0.007311
0
+0.011122
0
−0.016554
0
0.024420
0
−0.036404
0
+0.056866
0
−0.101892
0
+0.316883
+0.5
MAGNITUDE (dB)
HB1 Coefficient
Number
C1, C63
C2, C62
C3, C61
C4, C60
C5, C59
C6, C58
C7, C57
C8, C56
C9, C55
C10, C54
C11, C53
C12, C52
C13, C51
C14, C50
C15, C49
C16, C48
C17, C47
C18, C46
C19, C45
C20, C44
C21, C43
C22, C42
C23, C41
C24, C40
C25, C39
C26, C38
C27, C37
C28, C36
C29, C35
C30, C34
C31, C33
C32
–40
15660-081
The fourth and final decimate by 2, half-band, low-pass, FIR
filter (HB1) uses a 63-tap, symmetrical, fixed coefficient filter
implementation that is optimized for low power consumption.
The HB1 filter is always enabled and cannot be bypassed.
Table 24 and Figure 110 show the coefficients and response of
the HB1 filter.
Data Sheet
AD9695
20
TB1 Filter Description
TB1 Coefficient
Number
1, 76
2, 75
3, 74
4, 73
5, 72
6, 71
7, 70
8, 69
9, 68
10, 67
11, 66
12, 65
13, 64
14, 63
15, 62
16, 61
17, 60
18, 59
19, 58
20, 57
21, 56
22, 55
23, 54
24, 53
25, 52
26, 51
27, 50
28, 49
29, 48
30, 47
31, 46
32, 45
33, 44
34, 43
35, 42
36, 41
37, 40
38, 39
Normalized
Coefficient
−0.000023
−0.000053
−0.000037
+0.000090
+0.000291
+0.000366
+0.000095
−0.000463
−0.000822
−0.000412
+0.000739
+0.001665
+0.001132
−0.000981
−0.002961
−0.002438
+0.001087
+0.004833
+0.004614
−0.000871
−0.007410
−0.008039
+0.000053
+0.010874
+0.013313
+0.001817
−0.015579
−0.021590
−0.005603
+0.022451
+0.035774
+0.013541
−0.034655
−0.066549
−0.035213
+0.071220
+0.210777
+0.309200
Decimal Coefficient
(22-Bit)
−96
−224
−156
+379
+1220
+1534
+398
−1940
−3448
−1729
+3100
+6984
+4748
−4114
−12418
−10226
+4560
+20272
+19352
−3652
−31080
−33718
+222
+45608
+55840
+7620
−65344
−90556
−23502
+94167
+150046
+56796
−145352
−279128
−147694
+298720
+884064
+1296880
–20
–40
–60
–80
–100
–120
–140
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
Figure 112. TB1 Filter Response
Rev. 0 | Page 59 of 135
0.9
1.0
15660-083
Table 26. TB1 Filter Coefficients
0
MAGNITUDE (dB)
The TB1 decimate by 3, low-pass, FIR filter uses a 76-tap,
symmetrical, fixed coefficient filter implementation. Table 26
shows the TB1 filter coefficients, and Figure 112 shows the TB1
filter response. TB1 is only supported in DDC0 and DDC1.
AD9695
Data Sheet
20
FB2 Filter Description
FB2 Coefficient
Number
1, 48
2, 47
3, 46
4, 45
5, 44
6, 43
7, 42
8, 41
9, 40
10, 39
11, 38
12, 37
13, 36
14, 35
15, 34
16, 33
17, 32
18, 31
19, 30
20, 29
21, 28
22, 27
23, 26
24, 25
Normalized
Coefficient
+0.000007
−0.000004
−0.000069
−0.000244
−0.000544
−0.000870
−0.000962
−0.000448
+0.000977
+0.003237
+0.005614
+0.006714
+0.004871
−0.001011
−0.010456
−0.020729
−0.026978
−0.023453
−0.005608
+0.027681
+0.072720
+0.121223
+0.162346
+0.185959
Decimal Coefficient
(21-Bit)
7
−4
−72
−256
−570
−912
−1009
−470
+1024
+3394
+5887
+7040
+5108
−1060
−10964
−21736
−28288
−24592
−5880
+29026
+76252
+127112
+170232
+194992
–40
–60
–80
–100
–120
–140
–160
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
NORMALIZED FREQUENCY (× Π RAD/s)
0.9
1.0
15660-084
Table 27. FB2 Filter Coefficients
0
–20
MAGNITUDE (dB)
The FB2 decimate by 5, low-pass, FIR filter uses a 48-tap,
symmetrical, fixed coefficient filter implementation. Table 27
shows the FB2 filter coefficients, and Figure 113 shows the FB2
filter response.
Figure 113. FB2 Filter Response
DDC GAIN STAGE
Each DDC contains an independently controlled gain stage.
The gain is selectable as either 0 dB or 6 dB. When mixing a real
input signal down to baseband, it is recommended that the user
enable the 6 dB of gain to recenter the dynamic range of the
signal within the full scale of the output bits.
When mixing a complex input signal down to baseband, the
mixer has already recentered the dynamic range of the signal
within the full scale of the output bits, and no additional gain is
necessary. However, the optional 6 dB gain compensates for low
signal strengths. The downsample by 2 portion of the HB1 FIR
filter is bypassed when using the complex to real conversion
stage. The TB1 filter does not have the 6 dB gain stage.
Rev. 0 | Page 60 of 135
Data Sheet
AD9695
the signal, the Q portion of the complex mixer is no longer
needed and is dropped. The TB1 filter does not support
complex to real conversion.
DDC COMPLEX TO REAL CONVERSION
Each DDC contains an independently controlled complex to
real conversion block. The complex to real conversion block
reuses the last filter (HB1 FIR) in the filtering stage along with
an fS/4 complex mixer to upconvert the signal. After upconverting
HB1 FIR
Figure 114 shows a simplified block diagram of the complex to
real conversion.
GAIN STAGE
COMPLEX TO
REAL ENABLE
LOW-PASS
FILTER
I
2
0dB
OR
6dB
I
0 I/REAL
1
COMPLEX TO REAL CONVERSION
0dB
OR
6dB
I
cos(wt)
+
REAL
90°
fS/4
0°
–
LOW-PASS
FILTER
2
0dB
OR
6dB
Q
0dB
OR
6dB
Q
Q
15660-085
Q
sin(wt)
HB1 FIR
Figure 114. Complex to Real Conversion Block
Rev. 0 | Page 61 of 135
AD9695
Data Sheet
DDC MIXED DECIMATION SETTINGS
The AD9695 also supports DDCs with different decimation
rates. In this scenario, the chip decimation ratio must be set to
the lowest decimation ratio of all the DDC channels. Samples of
higher decimation ratio DDCs are repeated to match the chip
decimation ratio sample rate. Only mixed decimation ratios that
are integer multiples of 2 are supported. For example, decimate
by 1, 2, 4, 8, or 16 can be mixed together, decimate by 3, 6, 12,
24, or 48 can be mixed together, or decimate by 5, 10, 20, or 40
can be mixed together.
Table 26 shows the DDC sample mapping when the chip
decimation ratio is different than the DDC decimation ratio.
For example, if the chip decimation ratio is set to decimate by 4,
DDC0 is set to use the HB2 + HB1 filters (complex outputs,
decimate by 4) and DDC1 is set to use the HB4 + HB3 + HB2 +
HB1 filters (real outputs, decimate by 8), then DDC1 repeats its
output data 2 times for every one DDC0 output. The resulting
output samples are shown in Table 27.
Table 26. Sample Mapping when Chip Decimation Ratio (DCM) Does Not Match DDC DCM
Sample Index
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
DDC DCM = Chip DCM
N
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+8
N+9
N + 10
N + 11
N + 12
N + 13
N + 14
N + 15
N + 16
N + 17
N + 18
N + 19
N + 20
N + 21
N + 22
N + 23
N + 24
N + 25
N + 26
N + 27
N + 28
N + 29
N + 30
N + 31
DDC DCM = 2 × Chip DCM
N
N
N+1
N+1
N+2
N+2
N+3
N+3
N+4
N+4
N+5
N+5
N+6
N+6
N+7
N+7
N+8
N+8
N+9
N+9
N + 10
N + 10
N + 11
N + 11
N + 12
N + 12
N + 13
N + 13
N + 14
N + 14
N + 15
N + 15
DDC DCM = 4 × Chip DCM
N
N
N
N
N+1
N+1
N+1
N+1
N+2
N+2
N+2
N+2
N+3
N+3
N+3
N+3
N+4
N+4
N+4
N+4
N+5
N+5
N+5
N+5
N+6
N+6
N+6
N+6
N+7
N+7
N+7
N+7
Rev. 0 | Page 62 of 135
DDC DCM = 8 × Chip DCM
N
N
N
N
N
N
N
N
N+1
N+1
N+1
N+1
N+1
N+1
N+1
N+1
N+2
N+2
N+2
N+2
N+2
N+2
N+2
N+2
N+3
N+3
N+3
N+3
N+3
N+3
N+3
N+3
Data Sheet
AD9695
Table 27. Chip DCM = 4, DDC0 DCM = 4 (Complex), and DDC1 DCM = 8 (Real)1
DDC Input Samples
N
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+8
N+9
N + 10
N + 11
N + 12
N + 13
N + 14
N + 15
1
Output Port I
I0[N]
I0[N]
I0[N]
I0[N]
I0[N + 1]
I0[N + 1]
I0[N + 1]
I0[N + 1]
I0[N + 2]
I0[N + 2]
I0[N + 2]
I0[N + 2]
I0[N + 3]
I0[N + 3]
I0[N + 3]
I0[N + 3]
DDC0
Output Port Q
Q0[N]
Q0[N]
Q0[N]
Q0[N]
Q0[N + 1]
Q0[N + 1]
Q0[N + 1]
Q0[N + 1]
Q0[N + 2]
Q0[N + 2]
Q0[N + 2]
Q0[N + 2]
Q0[N + 3]
Q0[N + 3]
Q0[N + 3]
Q0[N + 3]
DCM means decimation.
Rev. 0 | Page 63 of 135
Output Port I
I1[N]
I1[N]
I1[N]
I1[N]
I1[N]
I1[N]
I1[N]
I1[N]
I1[N + 1]
I1[N + 1]
I1[N + 1]
I1[N + 1]
I1[N + 1]
I1[N + 1]
I1[N + 1]
I1[N + 1]
DDC1
Output Port Q
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
Not applicable
AD9695
Data Sheet
DDC EXAMPLE CONFIGURATIONS
Table 28 describes the register settings for multiple DDC example configurations. Bandwidths listed are with <−0.005 dB of pass-band
ripple and >100 dB of stop band alias rejection.
Table 28. DDC Example Configurations (per ADC Channel Pair)
Chip
Application
Layer
One DDC
Chip
Decimation
Ratio
2
DDC
Input
Type
Complex
DDC
Output
Type
Complex
Bandwidth
Per DDC1
40% × fS
No. of Virtual
Converters
Required
2
Two DDCs
4
Complex
Complex
20% × fS
4
Two DDCs
4
Complex
Real
10% × fS
2
Two DDCs
4
Real
Real
10% × fS
2
Rev. 0 | Page 64 of 135
Register Settings
0x0200 = 0x01 (one DDC; I/Q selected)
0x0201 = 0x01 (chip decimate by 2)
0x0310 = 0x83 (complex mixer; 0 dB gain; variable IF;
complex outputs; HB1 filter)
0x0311 = 0x04 (DDC I Input = ADC Channel A; DDC Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0200 = 0x02 (two DDCs; I/Q selected)
0x0201 = 0x02 (chip decimate by 4)
0x0310, 0x0330 = 0x80 (complex mixer; 0 dB gain;
variable IF; complex outputs; HB2+HB1 filters)
0x0311, 0x0331 = 0x04 (DDC I input = ADC Channel A;
DDC Q input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0200 = 0x22 (two DDCs; I only selected)
0x0201 = 0x02 (chip decimate by 4)
0x0310, 0x0330 = 0x89 (complex mixer; 0 dB gain;
variable IF; real output; HB3 + HB2 + HB1 filters)
0x0311, 0x0331 = 0x04 (DDC I Input = ADC Channel A;
DDC Q input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0200 = 0x22 (two DDCs; I only selected)
0x0201 = 0x02 (chip decimate by 4)
0x0310, 0x0330 = 0x49 (real mixer; 6 dB gain; variable IF;
real output; HB3 + HB2 + HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x05 (DDC1 I input = ADC Channel B; DDC1 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
Data Sheet
AD9695
Chip
Application
Layer
Two DDCs
Chip
Decimation
Ratio
4
DDC
Input
Type
Real
DDC
Output
Type
Complex
Bandwidth
Per DDC1
20% × fS
No. of Virtual
Converters
Required
4
Two DDCs
8
Real
Real
5% × fS
2
Four DDCs
8
Real
Complex
10% × fS
8
Rev. 0 | Page 65 of 135
Register Settings
0x0200 = 0x02 (two DDCs; I/Q selected)
0x0201 = 0x02 (chip decimate by 4)
0x0310, 0x0330 = 0x40 (real mixer; 6 dB gain; variable IF;
complex output; HB2 + HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x05 (DDC1 I input = ADC Channel B; DDC1 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0200 = 0x22 (two DDCs; I only selected)
0x0201 = 0x03 (chip decimate by 8)
0x0310, 0x0330 = 0x4A (real mixer; 6 dB gain; variable
IF; real output; HB4 + HB3 + HB2 + HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x05 (DDC1 I input = ADC Channel B; DDC1 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0200 = 0x03 (four DDCs; I/Q selected)
0x0201 = 0x03 (chip decimate by 8)
0x0310, 0x0330, 0x0350, 0x0370 = 0x41 (real mixer; 6 dB
gain; variable IF; complex output; HB3 + HB2 + HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x00 (DDC1 I input = ADC Channel A; DDC1 Q
input = ADC Channel A)
0x0351 = 0x05 (DDC2 I input = ADC Channel B; DDC2 Q
input = ADC Channel B)
0x0371 = 0x05 (DDC3 I input = ADC Channel B; DDC3 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0356, 0x0357, 0x0358, 0x0359, 0x035A, 0x035B,
0x035D, 0x035E, 0x035F, 0x0360, 0x0361, 0x0362 = FTW
and POW set as required by application for DDC2
0x0376, 0x0377, 0x0378, 0x0379, 0x037A, 0x037B,
0x037D, 0x037E, 0x037F, 0x0380, 0x0381, 0x0382 = FTW
and POW set as required by application for DDC3
AD9695
Data Sheet
Chip
Application
Layer
Four DDCs
Chip
Decimation
Ratio
8
DDC
Input
Type
Real
DDC
Output
Type
Real
Bandwidth
Per DDC1
5% × fS
No. of Virtual
Converters
Required
4
Four DDCs
16
Real
Complex
5% × fS
8
1
fS is the ADC sample rate.
Rev. 0 | Page 66 of 135
Register Settings
0x0200 = 0x23 (four DDCs; I only selected)
0x0201 = 0x03 (chip decimate by 8)
0x0310, 0x0330, 0x0350, 0x0370 = 0x4A (real mixer; 6 dB
gain; variable IF; real output; HB4 + HB3 + HB2 + HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x00 (DDC1 I input = ADC Channel A; DDC1 Q
input = ADC Channel A)
0x0351 = 0x05 (DDC2 I input = ADC Channel B; DDC2 Q
input = ADC Channel B)
0x0371 = 0x05 (DDC3 I input = ADC Channel B; DDC3 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0356, 0x0357, 0x0358, 0x0359, 0x035A, 0x035B,
0x035D, 0x035E, 0x035F, 0x0360, 0x0361, 0x0362 = FTW
and POW set as required by application for DDC2
0x0376, 0x0377, 0x0378, 0x0379, 0x037A, 0x037B,
0x037D, 0x037E, 0x037F, 0x0380, 0x0381, 0x0382 = FTW
and POW set as required by application for DDC3
0x0200 = 0x03 (four DDCs; I/Q selected)
0x0201 = 0x04 (chip decimate by 16)
0x0310, 0x0330, 0x0350, 0x0370 = 0x42 (real mixer; 6 dB
gain; variable IF; complex output; HB4 + HB3 + HB2 +
HB1 filters)
0x0311 = 0x00 (DDC0 I input = ADC Channel A; DDC0 Q
input = ADC Channel A)
0x0331 = 0x00 (DDC1 I input = ADC Channel A; DDC1 Q
input = ADC Channel A)
0x0351 = 0x05 (DDC2 I input = ADC Channel B; DDC2 Q
input = ADC Channel B)
0x0371 = 0x05 (DDC3 I input = ADC Channel B; DDC3 Q
input = ADC Channel B)
0x0316, 0x0317, 0x0318, 0x0319, 0x031A, 0x031B,
0x031D, 0x031E, 0x031F, 0x0320, 0x0321, 0x0322 = FTW
and POW set as required by application for DDC0
0x0336, 0x0337, 0x0338, 0x0339, 0x033A, 0x033B,
0x033D, 0x033E, 0x033F, 0x0340, 0x0341, 0x0342 = FTW
and POW set as required by application for DDC1
0x0356, 0x0357, 0x0358, 0x0359, 0x035A, 0x035B,
0x035D, 0x035E, 0x035F, 0x0360, 0x0361, 0x0362 = FTW
and POW set as required by application for DDC2
0x0376, 0x0377, 0x0378, 0x0379, 0x037A, 0x037B,
0x037D, 0x037E, 0x037F, 0x0380, 0x0381, 0x0382 = FTW
and POW set as required by application for DDC3
Data Sheet
AD9695
SIGNAL MONITOR
The signal monitor block provides additional information about
the signal being digitized by the ADC. The signal monitor
computes the peak magnitude of the digitized signal. This
information can be used to drive an AGC loop to optimize the
range of the ADC in the presence of real-world signals.
The results of the signal monitor block can be obtained either
by reading back the internal values from the SPI port or by
embedding the signal monitoring information into the JESD204B
interface as special control bits. A global, 24-bit programmable
period controls the duration of the measurement. Figure 82
shows the simplified block diagram of the signal monitor block.
FROM
MEMORY
MAP
SIGNAL MONITOR
PERIOD REGISTER
(SMPR)
0x0271, 00x272, 0x0273
DOWN
COUNTER
IS
COUNT = 1?
LOAD
FROM
INPUT
LOAD
LOAD
SIGNAL
MONITOR
HOLDING
REGISTER
TO SPORT OVER
JESD204B AND
MEMORY MAP
15660-086
CLEAR
MAGNITUDE
STORAGE
REGISTER
COMPARE
A>B
Figure 82. Signal Monitor Block
The peak detector captures the largest signal within the
observation period. The detector only observes the magnitude
of the signal. The resolution of the peak detector is a 13-bit
value, and the observation period is 24 bits and represents
converter output samples. The peak magnitude can be derived
by using the following equation:
The magnitude of the input port signal is monitored over a
programmable time period, which is determined by the signal
monitor period register (SMPR). The peak detector function is
enabled by setting Bit 1 of Register 0x0270 in the signal monitor
control register. The 24-bit SMPR must be programmed before
activating this mode.
After enabling peak detection mode, the value in the SMPR is
loaded into a monitor period timer, which decrements at the
decimated clock rate. The magnitude of the input signal is
compared with the value in the internal magnitude storage
register (not accessible to the user), and the greater of the two
is updated as the current peak level. The initial value of the
magnitude storage register is set to the current ADC input signal
magnitude. This comparison continues until the monitor period
timer reaches a count of 1.
When the monitor period timer reaches a count of 1, the 13-bit
peak level value is transferred to the signal monitor holding
register, which can be read through the memory map or output
through the SPORT over the JESD204B interface. The monitor
period timer is reloaded with the value in the SMPR, and the
countdown restarts. In addition, the magnitude of the first
input sample is updated in the magnitude storage register, and
the comparison and update procedure, as explained previously,
continues.
Peak Magnitude (dBFS) = 20log(Peak Detector Value/213)
Rev. 0 | Page 67 of 135
AD9695
Data Sheet
significant control bit is used (see Example Configuration 1 and
Example Configuration 2 in Figure 83). To select the SPORT
over JESD204B option, program Register 0x0559, Register 0x055A,
and Register 0x058F. See Table 47 for more information on
setting these bits.
SPORT OVER JESD204B
The signal monitor data can also be serialized and sent over the
JESD204B interface as control bits. These control bits must be
deserialized from the samples to reconstruct the statistical data.
The signal control monitor function is enabled by setting Bits[1:0]
of Register 0x0279 and Bit 1 of Register 0x027A. Figure 83
shows two different example configurations for the signal
monitor control bit locations inside the JESD204B samples. A
maximum of three control bits can be inserted into the
JESD204B samples; however, only one control bit is required for
the signal monitor. Control bits are inserted from MSB to LSB. If
only one control bit is to be inserted (CS = 1), only the most
Figure 84 shows the 25-bit frame data that encapsulates the
peak detector value. The frame data is transmitted MSB first
with five 5-bit subframes. Each subframe contains a start bit
that can be used by a receiver to validate the deserialized data.
Figure 85 shows the SPORT over JESD204B signal monitor data
with a monitor period timer set to 80 samples.
16-BIT JESD204B SAMPLE SIZE (N' = 16)
EXAMPLE
CONFIGURATION 1
(N' = 16, N = 15, CS = 1)
1-BIT
CONTROL
BIT
(CS = 1)
15-BIT CONVERTER RESOLUTION (N = 15)
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
S[14]
X
S[13]
X
S[12]
X
S[11]
X
S[10]
X
S[9]
X
S[8]
X
S[7]
X
S[6]
X
S[5]
X
S[4]
X
S[3]
X
S[2]
X
S[1]
X
S[0]
X
CTRL
[BIT 2]
X
SERIALIZED SIGNAL MONITOR
FRAME DATA
16-BIT JESD204B SAMPLE SIZE (N' = 16)
15
S[13]
X
14
13
S[12]
X
S[11]
X
12
S[10]
X
11
10
S[9]
X
9
S[8]
X
8
S[7]
X
7
S[6]
X
6
S[5]
X
5
S[4]
X
4
S[3]
X
S[2]
X
3
S[1]
X
2
1
0
S[0]
X
CTRL
[BIT 2]
X
TAIL
X
SERIALIZED SIGNAL MONITOR
FRAME DATA
Figure 83. Signal Monitor Control Bit Locations
5-BIT SUBFRAMES
5-BIT IDLE
SUBFRAME
(OPTIONAL)
25-BIT
FRAME
IDLE
1
IDLE
1
IDLE
1
IDLE
1
IDLE
1
5-BIT IDENTIFIER START
0
SUBFRAME
ID[3]
0
ID[2]
0
ID[1]
0
ID[0]
1
5-BIT DATA
MSB
SUBFRAME
START
0
P[12]
P[11]
P[10]
P[9]
5-BIT DATA
SUBFRAME
START
0
P[8]
P[7]
P[6]
P5]
5-BIT DATA
SUBFRAME
START
0
P[4]
P[3]
P[2]
P1]
5-BIT DATA
LSB
SUBFRAME
START
0
P[0]
0
0
0
P[x] = PEAK MAGNITUDE VALUE
Figure 84. SPORT over JESD204B Signal Monitor Frame Data
Rev. 0 | Page 68 of 135
15660-088
EXAMPLE
CONFIGURATION 2
(N' = 16, N = 14, CS = 1)
15660-087
1
CONTROL
1 TAIL
BIT
BIT
(CS = 1)
14-BIT CONVERTER RESOLUTION (N = 14)
Data Sheet
AD9695
SMPR = 80 SAMPLES (0x0271 = 0x50; 0x0272 = 0x00; 0x0273 = 0x00)
80 SAMPLE PERIOD
PAYLOAD
25-BIT FRAME (N)
IDENT.
DATA
MSB
DATA
DATA
LSB
DATA
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
IDLE
80 SAMPLE PERIOD
PAYLOAD
25-BIT FRAME (N + 1)
IDENT.
DATA
MSB
DATA
DATA
LSB
DATA
IDLE
IDLE
IDLE
IDLE
IDLE
80 SAMPLE PERIOD
IDENT.
DATA
MSB
DATA
DATA
DATA
LSB
IDLE
IDLE
IDLE
IDLE
IDLE
Figure 85. SPORT over JESD204B Signal Monitor Example
Rev. 0 | Page 69 of 135
15660-089
PAYLOAD
25-BIT FRAME (N + 2)
AD9695
Data Sheet
DIGITAL OUTPUTS

INTRODUCTION TO THE JESD204B INTERFACE
The AD9695 digital outputs are designed to the JEDEC standard
JESD204B, serial interface for data converters. JESD204B is a
protocol to link the AD9695 to a digital processing device over
a serial interface with lane rates of up to 16 Gbps. The benefits
of the JESD204B interface over LVDS include a reduction in
required board area for data interface routing, and an ability to
enable smaller packages for converter and logic devices.
JESD204B OVERVIEW



K is the number of frames per multiframe
(AD9695 value = 4, 8, 12, 16, 20, 24, 28, or 32 )
S is the samples transmitted/single converter/frame cycle
(AD9695 value = set automatically based on L, M, F, and N΄)
HD is the high density mode (AD9695 = set automatically
based on L, M, F, and N΄)
CF is the number of control words/frame clock
cycle/converter device (AD9695 value = 0)
The JESD204B data transmit block assembles the parallel data
from the ADC into frames and uses 8-bit/10-bit encoding as
well as optional scrambling to form serial output data. Lane
synchronization is supported through the use of special control
characters during the initial establishment of the link. Additional
control characters are embedded in the data stream to maintain
synchronization thereafter. A JESD204B receiver is required to
complete the serial link. For additional details on the JESD204B
interface, refer to the JESD204B standard.
Figure 119 shows a simplified block diagram of the AD9695
JESD204B link. By default, the AD9695 is configured to use
two converters and four lanes. Converter A data is output to
SERDOUT0± and/or SERDOUT1±, and Converter B is output
to SERDOUT2± and/or SERDOUT3±. The AD9695 allows
other configurations, such as combining the outputs of both
converters onto a single lane, or changing the mapping of the
A and B digital output paths. These modes are customizable,
and can be set up via the SPI. Refer to the Memory Map section
for more details.
The AD9695 JESD204B data transmit block maps up to two
physical ADCs or up to eight virtual converters (when DDCs
are enabled) over a link. A link can be configured to use one,
two, or four JESD204B lanes. The JESD204B specification refers
to a number of parameters to define the link, and these parameters
must match between the JESD204B transmitter (the AD9695
output) and the JESD204B receiver (the logic device input).
By default in the AD9695, the 14-bit converter word from each
converter is broken into two octets (eight bits of data). Bit 13
(MSB) through Bit 6 are in the first octet. The second octet
contains Bit 5 through Bit 0 (LSB) and two tail bits. The tail bits
can be configured as zeros or a pseudorandom number
sequence. The tail bits can also be replaced with control bits
indicating overrange, SYSREF±, or fast detect output.
The JESD204B link is described according to the following
parameters:
The two resulting octets can be scrambled. Scrambling is
optional; however, it is recommended to avoid spectral peaks
when transmitting similar digital data patterns. The scrambler
uses a self-synchronizing, polynomial-based algorithm defined
by the equation 1 + x14 + x15. The descrambler in the receiver is
a self synchronizing version of the scrambler polynomial.





L is the number of lanes/converter device (lanes/link)
(AD9695 value = 1, 2, or 4)
M is the number of converters/converter device (virtual
converters/link) (AD9695 value = 1, 2, 4, or 8)
F is the octets/frame (AD9695 value = 1, 2, 4, 8, or 16)
N΄ is the number of bits per sample (JESD204B word size)
(AD9695 value = 8 or 16)
N is the converter resolution (AD9695 value = 7 to 16)
CS is the number of control bits/sample
(AD9695 value = 0, 1, 2, or 3)
The two octets are then encoded with an 8-bit/10-bit encoder.
The 8-bit/10-bit encoder works by taking eight bits of data (an
octet) and encoding them into a 10-bit symbol. Figure 119
shows how the 14-bit data is taken from the ADC, how the tail
bits are added, how the two octets are scrambled, and how the
octets are encoded into two 10-bit symbols. Figure 120 shows the
default data format.
CONVERTER 0
CONVERTER A
INPUT
ADC A
MUX/
FORMAT
(SPI
REGISTERS
0x0561,
0x0564)
CONVERTER B
INPUT
JESD204B LINK
CONTROL
(L, M, F)
(SPI REGISTER
0x058B,
0x058E, 0x058C)
ADC B
LANE MUX
AND
MAPPING
(SPI
REGISTERS
0x05B0,
0x05B2,
0x05B3,
0x05B5,
0x05B6)
SERDOUT0±
SERDOUT1±
SERDOUT2±
SERDOUT3±
CONVERTER 1
15660-090

SYSREF±
SYNCINB±
Figure 119. Transmit Link Simplified Block Diagram Showing Full Bandwidth Mode (Register 0x200 = 0x00)
Rev. 0 | Page 70 of 135
Data Sheet
AD9695
JESD204B DATA
LINK LAYER TEST
PATTERNS
0x0574[2:0]
JESD204B
INTERFACE TEST
PATTERN
(0x0573,
0x0551 TO 0x0558)
MSB A13
A12
A11
A10
A9
A8
ADC
A7
A6
A5
A4
A3
A2
A1
LSB A0
OCTET1
TAIL BITS
0x0571[6]
OCTET0
JESD204B SAMPLE
CONSTRUCTION
MSB A13
A12
A11
A10
A9
A8
A7
LSB A6
A5
A4
A3
A2
A1
A0
C2
T
SCRAMBLER
1 + x14 + x15
(OPTIONAL)
MSB S7
S6
S5
S4
S3
S2
S1
LSB S0
S7
S6
S5
S4
S3
S2
S1
S0
SERIALIZER
8-BIT/
10-BIT
ENCODER
a b
i j a b
SERDOUT0±
SERDOUT1±
SERDOUT2±
SERDOUT3±
i j
SYMBOL0 SYMBOL1
a b c d e f g h i j
a b c d e f g h i j
C2
C1
C0
15660-091
CONTROL BITS
FRAME
CONSTRUCTION
OCTET1
ADC TEST PATTERNS
(0x0550,
0x0551 TO 0x0558)
OCTET0
JESD204B LONG
TRANSPORT TEST
PATTERN
0x0571[5]
Figure 120. ADC Output Datapath Showing Data Framing
TRANSPORT
LAYER
SAMPLE
CONSTRUCTION
FRAME
CONSTRUCTION
SCRAMBLER
ALIGNMENT
CHARACTER
GENERATION
8-BIT/10-BIT
ENCODER
PHYSICAL
LAYER
CROSSBAR
MUX
SERIALIZER
Tx
OUTPUT
15660-092
PROCESSED
SAMPLES
FROM ADC
DATA LINK
LAYER
SYSREF±
SYNCINB±
Figure 121. Data Flow
FUNCTIONAL OVERVIEW
Data Link Layer
The block diagram in Figure 121 shows the flow of data through
the JESD204B hardware from the sample input to the physical
output. The processing can be divided into layers that are
derived from the open source initiative (OSI) model, widely
used to describe the abstraction layers of communications
systems. These layers are the transport layer, data link layer,
and physical layer (serializer and output driver).
The data link layer is responsible for the low level functions
of passing data across the link. These include optionally
scrambling the data, inserting control characters during the
initial lane alignment sequence (ILAS) and for frame and
multiframe synchronization monitoring, and encoding 8-bit
octets into 10-bit symbols. The data link layer is also responsible for
sending the ILAS, which contains the link configuration data
used by the receiver to verify the settings in the transport layer.
Transport Layer
The transport layer handles packing the data (consisting of
samples and optional control bits) into JESD204B frames that
are mapped to 8-bit octets. The packing of samples into frames
are determined by the JESD204B configuration parameters for
number of lanes (L), number of converters (M), the number of
octets per lane per frame (F), the number of samples per
converter per frame (S), and the number of bits in a nibble
group (sometimes called the JESD204 word size − N’).
Samples are mapped in order starting from Converter 0, then
Converter 1, and so on until Converter M − 1. If S > 1, each
sample from the converter is mapped before mapping the
samples from the next converter. Each sample is mapped into
words formed by appending converter control bits, if enabled,
to the LSBs of each sample. The words are then padded with tail
bits, if necessary, to form nibble groups (NGs) of the appropriate
size as determined by the N’ parameter. The following equation
can be used to determine the number of tail bits within a nibble
group (JESD204B word):
T = N΄ − N − CS
Physical Layer
The physical layer consists of the high speed circuitry clocked at
the serial clock rate. In this layer, parallel data is converted into
one, two, or four lanes of high speed differential serial data.
JESD204B LINK ESTABLISHMENT
The AD9695 JESD204B transmitter (Tx) interface operates in
Subclass 0 or Sunclass 1 as defined in the JEDEC Standard
JESD204B (July 2011 specification). The link establishment
process is divided into the following steps: code group synchronization, initial lane alignment sequence, and user data and error
correction.
Code Group Synchronization (CGS)
CGS is the process by which the JESD204B receiver finds the
boundaries between the 10-bit symbols in the stream of data.
During the CGS phase, the JESD204B transmit block transmits
/K/ characters (/K28.5/ symbols). The receiver must locate the
/K/ characters in its input data stream using clock and data
recovery (CDR) techniques.
Rev. 0 | Page 71 of 135
AD9695
Data Sheet

The receiver issues a synchronization request by asserting the
SYNCINB± pin of the AD9695 low. The JESD204B Tx then begins
sending /K/ characters. Once the receiver has synchronized, it waits
for the correct reception of at least four consecutive /K/ symbols. It
then deasserts SYNCINB±. The AD9695 then transmits an ILAS
on the following local multiframe clock (LMFC) boundary.
User Data and Error Detection
After the initial lane alignment sequence is complete, the user
data (ADC samples) is sent. During transmission of the user data, a
mechanism called character replacement monitors the frame
clock and multiframe clock alignment. This mechanism replaces
the last octet of a frame or multiframe with an /F/ or /A/
alignment characters when the data meets certain conditions.
These conditions are different for unscrambled and scrambled
data. The scrambling operation is enabled by default, but it can
be disabled using the SPI.
For more information on the code group synchronization
phase, refer to the JEDEC Standard JESD204B, July 2011,
Section 5.3.3.1.
The SYNCINB± pin operation can also be controlled by the
SPI. The SYNCINB± signal is a differential dc-coupled LVDS
mode signal by default, but it can also be driven single-ended.
For more information on configuring the SYNCINB± pin
operation, refer to Register 0x572.
For scrambled data, any 0xFC character at the end of a frame
is replaced by an /F/, and any 0x7C character at the end of a
multiframe is replaced with an /A/. The JESD204B receiver (Rx)
checks for /F/ and /A/ characters in the received data stream
and verifies that they only occur in the expected locations. If an
unexpected /F/ or /A/ character is found, the receiver handles
the situation by using dynamic realignment or asserting the
SYNCINB± signal for more than four frames to initiate a
resynchronization. For unscrambled data, if the final octet of
two subsequent frames are equal, the second octet is replaced
with an /F/ symbol if it is at the end of a frame, and an /A/
symbol if it is at the end of a multiframe.
The SYNCINB± pins can also be configured to run in CMOS
(single-ended) mode by setting Bit 4 in Register 0x572. When
running SYNCINB± in CMOS mode, connect the CMOS
SYNCINB signal to Pin 21 (SYNCINB+) and leave Pin 20
(SYNCINB−) disconnected.
Initial Lane Alignment Sequence (ILAS)
The ILAS phase follows the CGS phase and begins on the next
LMFC boundary after SYNCINB± deassertion. The ILAS
consists of four mulitframes, with an /R/ character marking the
beginning and an /A/ character marking the end. The ILAS
begins by sending an /R/ character followed by 0 to 255 ramp
data for one multiframe. On the second multiframe, the link
configuration data is sent, starting with the third character. The
second character is a /Q/ character to confirm that the link
configuration data follows. All undefined data slots are filled
with ramp data. The ILAS sequence is never scrambled.
Insertion of alignment characters can be modified using SPI.
The frame alignment character insertion (FACI) is enabled by
default. More information on the link controls is available in the
Memory Map section, Register 0x571.
8-Bit/10-Bit Encoder
The 8-bit/10-bit encoder converts 8-bit octets into 10-bit symbols
and inserts control characters into the stream when needed.
The control characters used in JESD204B are shown in Table 31.
The 8-bit/10-bit encoding ensures that the signal is dc balanced by
using the same number of ones and zeros across multiple symbols.
The ILAS sequence construction is shown in Figure 122. The
four multiframes include the following:
Multiframe 1 begins with an /R/ character (/K28.0/) and
ends with an /A/ character (/K28.3/).
Multiframe 2 begins with an /R/ character followed by a
/Q/ character (/K28.4/), followed by link configuration
parameters over 14 configuration octets (see Table 31) and
ends with an /A/ character. Many of the parameter values
are of the value – 1 notation.
Multiframe 3 begins with an /R/ character (/K28.0/) and
ends with an /A/ character (/K28.3/).


K
K
R
D
●●●
D
A
R
Q
C
●●●
C
D
●●●
The 8-bit/10-bit interface has options that can be controlled via
the SPI. These operations include bypass and invert. These options
are troubleshooting tools for the verification of the digital front
end (DFE). Refer to the Memory Map section, Register 0x572,
Bits[2:1] for information on configuring the 8-bit/10-bit encoder.
D
A
R
D
●●●
D
A
R
D
●●●
D
A
D
END OF
MULTIFRAME
●●●
START OF
ILAS
●●●
●●●
●●●
START OF LINK
CONFIGURATION DATA
●●●
START OF
USER DATA
Figure 122. Initial Lane Alignment Sequence
Rev. 0 | Page 72 of 135
15660-093

Multiframe 4 begins with an /R/ character (/K28.0/) and
ends with an /A/ character (/K28.3/).
Data Sheet
AD9695
Table 31. AD9695 Control Characters Used in JESD204B
Abbreviation
/R/
/A/
/Q/
/K/
/F/
1
Control Symbol
/K28.0/
/K28.3/
/K28.4/
/K28.5/
/K28.7/
8-Bit Value
000 11100
011 11100
100 11100
101 11100
111 11100
10-Bit Value,
RD1 = −1
001111 0100
001111 0011
001111 0100
001111 1010
001111 1000
10-Bit Value,
RD1 = +1
110000 1011
110000 1100
110000 1101
110000 0101
110000 0111
Description
Start of multiframe
Lane alignment
Start of link configuration data
Group synchronization
Frame alignment
RD means running disparity.
DRVDD1
SERDOUTx+
0.1µF
100Ω
DIFFERENTIAL
TRACE PAIR
100Ω
0.1µF
RECEIVER
SERDOUTx–
15660-094
OUTPUT SWING = 0.85 × DRVDD1 V p-p DIFFERENTIAL
ADJUSTABLE TO
1 × DRVDD1, 0.75 × DRVDD1
Figure 123. AC-Coupled Digital Output Termination Example
PHYSICAL LAYER (DRIVER) OUTPUTS
The AD9695 physical layer consists of drivers that are defined in
the JEDEC Standard JESD204B, July 2011. The differential digital
outputs are powered up by default. The drivers use a dynamic
100 Ω internal termination to reduce unwanted reflections.
15660-095
Digital Outputs, Timing, and Controls
Figure 124. Digital Outputs Data Eye, External 100 Ω Terminations at 16 Gbps
If there is no far end receiver termination, or if there is poor
differential trace routing, timing errors can result. To avoid such
timing errors, it is recommended that the trace length be less
than six inches, and that the differential output traces be close
together and at equal lengths.
Figure 125. Digital Outputs Jitter Histogram, External 100 Ω Terminations at
16 Gbps
15660-097
The AD9695 digital outputs can interface with custom ASICs
and field programmable gate array (FPGA) receivers, providing
superior switching performance in noisy environments. Single
point-to-point network topologies are recommended with a
single differential 100 Ω termination resistor placed as close to
the receiver inputs as possible.
15660-096
Place a 100 Ω differential termination resistor at each receiver
input to result in a nominal 0.85 × DRVDD1 V p-p swing at the
receiver (see Figure 123). The swing is adjustable through the
SPI registers. AC coupling is recommended to connect to the
receiver. See the Memory Map section (Register 0x05C0 to
Register 0x05C3 in Table 47) for more details.
Figure 126. Digital Outputs Bathtub Curve, External 100 Ω Terminations at
16 Gbps
Figure 124 to Figure 126 show an example of the digital output
data eye, jitter histogram, and bathtub curve for one AD9695
lane running at 16 Gbps. The format of the output data is twos
complement by default. To change the output data format, see
the Memory Map section (Register 0x0561).
Rev. 0 | Page 73 of 135
AD9695
Data Sheet
Deemphasis
Table 33. AD9695 JESD204B Initialization
Deemphasis enables the receiver eye diagram mask to be met in
conditions where the interconnect insertion loss does not meet
the JESD204B specification. Use the deemphasis feature only
when the receiver is unable to recover the clock due to excessive
insertion loss. Under normal conditions, it is disabled to
conserve power. Additionally, enabling and setting too high a
deemphasis value on a short link can cause the receiver eye
diagram to fail. Use the deemphasis setting with caution
because it can increase electromagnetic interference (EMI). See
the Memory Map section (Register 0x05C4 to Register 0x05CA
in Table 47) for more details.
Register
0x1228
0x1228
0x1222
0x1222
0x1222
0x1262
0x1262
Phase-Locked Loop (PLL)
The PLL generates the serializer clock, which operates at the
JESD204B lane rate. The status of the PLL lock can be checked
in the PLL locked status bit (Register 0x056F, Bit 7). This read
only bit notifies the user if the PLL achieved a lock for the
specific setup. Register 0x056F also has a loss of lock (LOL)
sticky bit (Bit 3) that notifies the user that a loss of lock is
detected. The sticky bit can be reset by issuing a JESD204B link
restart (Register 0x0571, Bit 0 = 0x1, followed by Register 0x0571,
Bit 0 = 0x0). Refer to Table 33 for the reinitialization of the link
following a link power cycle.
The JESD204B lane rate control, Bits[7:4] of Register 0x056E,
must be set to correspond with the lane rate. Table 32 shows the
lane rates supported by the AD9695 using Register 0x056E.
Table 32. AD9695 Register 0x056E Supported Lane Rates
Value
0x00
0x10
0x30
0x50
Lane Rate
Lane rate = 6.75 Gbps to 13.5
Lane rate = 3.375 Gbps to 6.75 Gbps (default)
Lane rate = 13.5 Gbps to 16 Gbps
Lane rate = 1.6875 Gbps to 3.375 Gbps



The initialization SPI writes are as shown in Table 33.
Number of lanes per link (L)
Number of converters per link (M)
Number of octets per frame (F)
If the internal DDCs are used for on-chip digital processing, M
represents the number of virtual converters. The virtual converter
mapping setup is shown in Table 11.
By default in the AD9695, the 14-bit converter word from each
converter is broken into two octets (eight bits of data). Bit 13
(MSB) through Bit 6 are in the first octet. The second octet
contains Bit 5 through Bit 0 (LSB) and two tail bits. The tail bits
can be configured as zeros or a pseudorandom number
sequence. The tail bits can also be replaced with control bits
indicating overrange, SYSREF±, or fast detect output. Control
bits are filled and inserted MSB first such that enabling CS = 1
activates Control Bit 2, enabling CS = 2 activates Control Bit 2
and Control Bit 1, and enabling CS = 3 activates Control Bit 2,
Control Bit 1, and Control Bit 0.
The maximum lane rate allowed by the AD9695 is 16 Gbps. The
lane rate is related to the JESD204B parameters using the
following equation:
To ensure proper operation of the AD9695 at startup, some SPI
writes are required to initialize the link. Additionally, these
registers must be written every time the ADC is reset. Any one
of the following resets warrants the initialization routine for the
digital interface:
Hard reset, as with power-up.
Power-up using the PDWN pin.
Power-up using the SPI via Register 0x0002, Bits[1:0].
SPI soft reset by setting Register 0x0000 = 0x81.
Datapath soft reset by setting Register 0x0001 = 0x02.
JESD204B link power cycle by setting Register 0x0571,
Bit 0 = 0x1, then 0x0.
Comment
Reset JESD204B start-up circuit
JESD204B start-up circuit in normal operation
JESD204B PLL force normal operation
Reset JESD204B PLL calibration
JESD204B PLL normal operation
Clear loss of lock bit
Loss of lock bit normal operation
The AD9695 has one JESD204B link. The serial outputs
(SERDOUT0± to SERDOUT3±) are considered to be part of
one JESD204B link. The basic parameters that determine the
link setup are
SETTING UP THE AD9695 DIGITAL INTERFACE






Value
0x4F
0x0F
0x00
0x04
0x00
0x08
0x00
10
M  N '   f OUT
 8 
Lane Rate =
L
where fOUT =
f ADC _ CLOCK
Decimation Ratio
The decimation ratio (DCM) is the parameter programmed in
Register 0x0201.
Use the following procedure to configure the output:
1.
2.
3.
4.
5.
6.
7.
Rev. 0 | Page 74 of 135
Power down the link.
Select the JESD204B link configuration options.
Configure the detailed options.
Set output lane mapping (optional).
Set additional driver configuration options (optional).
Power up the link.
Initialize the JESD204B link by issuing the commands
described in Table 33.
Data Sheet
AD9695
Register 0x056E must be programmed according to the lane
rate calculated. Refer to the Phase-Locked Loop (PLL) section
for more details.
Table 34 and Table 35 show the JESD204B output configurations
supported for both N΄ = 16, N’=12, and N΄ = 8 for a given
number of virtual converters. Take care to ensure that the serial
lane rate for a given configuration is within the supported range
of 1.6875 Gbps to 16 Gbps.
Table 34. JESD204B Output Configurations for N΄ = 161
Number
of Virtual
Converters
Supported
(Same as M)
1
2
4
8
Supported Decimation Rates
JESD204B
Serial
Lane
Rate2
20 × fOUT
Lane Rate =
1.6875 Gbps to
3.375 Gbps
2, 4, 5, 6
Lane Rate =
3.375 Gbps to
6.75 Gbps
1, 2, 3
Lane Rate =
6.75 Gbps to
13.5 Gbps
1
Lane Rate =
13.5 Gbps to
16 Gbps
N/A
L
1
M
1
F
2
S
1
HD
0
N
8 to 16
N'
16
CS
0 to 3
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
1
1
4
2
0
8 to 16
16
0 to 3
10 × fOUT
1, 2, 3
1,
N/A
N/A
2
1
1
1
1
8 to 16
16
0 to 3
10 × fOUT
1, 2, 3
1
N/A
N/A
2
1
2
2
0
8 to 16
16
0 to 3
5 × fOUT
1
N/A
N/A
N/A
4
1
1
2
1
8 to 16
16
0 to 3
5 × fOUT
1
N/A
N/A
N/A
4
1
2
4
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 5, 6
1, 2, 3
1
1
2
4
1
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 5, 6
1, 2, 3
1
1
2
8
2
0
8 to 16
16
0 to 3
20 × fOUT
4, 5, 6
1, 2, 3
1
N/A
2
2
2
1
0
8 to 16
16
0 to 3
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
2
2
4
2
0
8 to 16
16
0 to 3
10 × fOUT
1, 2, 3
1
N/A
N/A
4
2
1
1
1
8 to 16
16
0 to 3
10 × fOUT
1, 2, 3
1
N/A
N/A
4
2
2
2
0
8 to 16
16
0 to 3
80 × fOUT
8, 16, 20, 24
4, 8, 10, 12
2, 4, 6
2
1
4
8
1
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 5, 6
1, 2, 3
1
2
4
4
1
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 5, 6
1, 2, 3
1
2
4
8
2
0
8 to 16
16
0 to 3
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
4
4
2
1
0
8 to 16
16
0 to 3
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
4
4
4
2
0
8 to 16
16
0 to 3
160 × fOUT
16, 40, 48
8, 16, 20, 24
4, 8, 12
4
1
8
16
1
0
8 to 16
16
0 to 3
80 × fOUT
8, 16, 20, 24
4, 8, 10, 12
2, 4, 6
2
2
8
8
1
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 6
2
N/A
4
8
4
1
0
8 to 16
16
0 to 3
40 × fOUT
4, 8, 10, 12
2, 4, 6
2
N/A
4
8
8
2
0
8 to 16
16
0 to 3
1
JESD204B Transport Layer Settings3
K
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
Due to the internal clock requirements, only certain decimation rates are supported for certain link parameters.
JESD204B transport layer descriptions are as follows: L is the number of lanes per converter device (lanes per link); M is the number of virtual converters per converter
device (virtual converters per link); F is the octets per frame; S is the samples transmitted per virtual converter per frame cycle; HD is the high density mode; N is the virtual
converter resolution (in bits); N' is the total number of bits per sample (JESD204B word size); CS is the number of control bits per conversion sample; K is the number of frames per
multiframe.
3
fADC_CLK is the ADC sample rate; DCM = chip decimation ratio; fOUT is the output sample rate = fADC_CLK/DCM; SLR is the JESD204B serial lane rate. The following equations must be
met due to internal clock divider requirements: SLR ≥ 1.6875 Gbps and SLR ≤ 16 Gbps; SLR/40 ≤ fADC_CLK; least common multiple (20 × DCM × fOUT/SLR, DCM) ≤ 64. When the SLR
is ≤ 16,000 Mbps and > 13,500 Mbps, Register 0x056E must be set to 0x30. When the SLR is ≤ 13,500 Mbps and ≥ 6750 Mbps, Register 0x056E must be set to 0x00. When the SLR
is < 6750 Mbps and ≥ 3375 Mbps, Register 0x056E must be set to 0x10. When the SLR is < 3375 Mbps and ≥ 1687.5 Mbps, Register 0x056E must be set to 0x50.
4
Only valid K × F values that are divisible by 4 are supported: for F = 1, K = 20, 24, 28, 32; for F = 2, K = 12, 16, 20, 24, 28, 32; for F = 4, K = 8, 12, 16, 20, 24, 28, 32; for F = 8, K = 4,
8, 12, 16, 20, 24, 28, 32; and for F = 16, K = 4, 8, 12, 16, 20, 24, 28, 32.
2
Rev. 0 | Page 75 of 135
AD9695
Data Sheet
Table 35. JESD204B Output Configurations (N' = 12)1
No. of
Virtual
Converters
Supported
(Same Value
as M)
1
2
4
8
JESD204B Transport Layer Settings3
Supported Decimation Rates
Serial
Lane
Rate2
15 × fOUT
Lane Rate =
1.6875 Gbps
to
3.375 Gbps
3
Lane Rate =
3.375 Gbps to
6.75 Gbps
N/A
Lane Rate =
6.75 Gbps
to
13.5 Gbps
N/A
Lane Rate =
13.5 Gbps to
16 Gbps
N/A
L
1
M
1
F
3
S
2
HD
0
N
8 to 12
N'
12
L
0 to 3
7.5 × fOUT
N/A
N/A
N/A
N/A
2
1
3
4
1
8 to 12
12
0 to 3
7.5 × fOUT
N/A
N/A
N/A
N/A
2
1
6
8
0
8 to 12
12
0 to 3
5 × fOUT
1
N/A
N/A
N/A
3
1
1
2
1
8 to 12
12
0 to 3
30 × fOUT
3, 6
3
N/A
N/A
1
2
3
1
0
8 to 12
12
0 to 3
15 × fOUT
3
N/A
N/A
N/A
2
2
3
2
0
8 to 12
12
0 to 3
10 × fOUT
1, 2, 3
1
N/A
N/A
3
2
1
1
1
8 to 12
12
0 to 3
7.5 × fOUT
N/A
N/A
N/A
N/A
4
2
3
4
0
8 to 12
12
0 to 3
60 × fOUT
6, 12
3, 6
3
N/A
1
4
6
1
0
8 to 12
12
0 to 3
30 × fOUT
3, 6
3
N/A
N/A
2
4
3
1
0
8 to 12
12
0 to 3
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
3
4
2
1
1
8 to 12
12
0 to 3
15 × fOUT
3
N/A
N/A
N/A
4
4
3
2
0
8 to 12
12
0 to 3
60 × fOUT
6, 12
6
N/A
N/A
2
8
6
1
0
8 to 12
12
0 to 3
30 × fOUT
6
N/A
N/A
N/A
4
8
3
1
0
8 to 12
12
0 to 3
1
K
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
Due to the internal clock requirements, only certain decimation rates are supported for certain link parameters.
fADC_CLK is the ADC sample rate; DCM is the chip decimation ratio; fOUT is the output sample rate = fADC_CLK/DCM; SLR is the JESD204B serial lane rate. The following equations
must be met due to internal clock divider requirements: SLR ≥ 1.6875 Gbps and SLR ≤ 16 Gbps; SLR/40 ≤ fADC_CLK; least common multiple (20 × DCM × fOUT/SLR, DCM) ≤ 64.
When the SLR is ≤ 16,000 Mbps and > 13,500 Mbps, Register 0x056E must be set to 0x30. When the SLR is ≤ 13,500 Mbps and ≥ 6750 Mbps, Register 0x056E must be
set to 0x00. When the SLR is < 6750 Mbps and ≥ 3375 Mbps, Register 0x056E must be set to 0x10. When the SLR is < 3375 Mbps and ≥ 1687.5 Mbps, Register 0x056E
must be set to 0x50.
3
JESD204B transport layer descriptions are as follows: L is the number of lanes per converter device (lanes per link); M is the number of virtual converters per converter device
(virtual converters per link); F is the octets per frame; S is the samples transmitted per virtual converter per frame cycle; HD is the high density mode; N is the virtual converter
resolution (in bits); N' is the total number of bits per sample (JESD204B word size); CS is the number of control bits per conversion sample; K is the number of frames per
multiframe.
4
Only valid K × F values that are divisible by 4 are supported: for F = 1, K = 20, 24, 28, 32; for F = 2, K = 12, 16, 20, 24, 28, 32; for F = 4, K = 8, 12, 16, 20, 24, 28, 32; for F = 8, K = 4,
8, 12, 16, 20, 24, 28, 32; and for F = 16, K = 4, 8, 12, 16, 20, 24, 28, 32.
2
Rev. 0 | Page 76 of 135
Data Sheet
AD9695
Table 32. JESD204B Output Configurations for N΄ = 81
JESD204B Transport Layer Settings3
Supported decimation rates
No. of
Virtual
Converters
Supported
(Same Value
as M)
1
Serial
Lane
Rate2
10 × fOUT
Lane Rate =
1.6875 Gbps
to
3.375 Gbps
1, 2, 3
Lane Rate =
3.375 Gbps
to
6.75 Gbps
1
Lane Rate =
6.75 Gbps
to
13.5 Gbps
N/A
Lane Rate =
13.5 Gbps to
16 Gbps
N/A
L
1
M
1
F
1
S
1
HD
0
N
7 to 8
N'
8
CS
0 to 1
1
10 × fOUT
1, 2, 3
1
N/A
N/A
1
1
2
2
0
7 to 8
8
0 to 1
1
5 × fOUT
1
N/A
N/A
N/A
2
1
1
2
0
7 to 8
8
0 to 1
1
5 × fOUT
1
N/A
N/A
N/A
2
1
2
4
0
7 to 8
8
0 to 1
1
5 × fOUT
1
N/A
N/A
N/A
2
1
4
8
0
7 to 8
8
0 to 1
1
2.5 × fOUT
N/A
N/A
N/A
N/A
4
1
1
4
0
7 to 8
8
0 to 1
1
2.5 × fOUT
N/A
N/A
N/A
N/A
4
1
2
8
0
7 to 8
8
0 to 1
2
20 × fOUT
2, 4, 5, 6
1, 2, 3
1
N/A
1
2
2
1
0
7 to 8
8
0 to 1
2
10 × fOUT
1, 2, 3
1
N/A
N/A
2
2
1
1
0
7 to 8
8
0 to 1
2
10 × fOUT
1, 2, 3
1
N/A
N/A
2
2
2
2
0
7 to 8
8
0 to 1
2
5 × fOUT
1
N/A
N/A
N/A
4
2
1
2
0
7 to 8
8
0 to 1
2
5 × fOUT
1
N/A
N/A
N/A
4
2
2
4
0
7 to 8
8
0 to 1
2
5 × fOUT
1
N/A
N/A
N/A
4
2
4
8
0
7 to 8
8
0 to 1
1
K
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
See
Note 4
Due to the internal clock requirements, only certain decimation rates are supported for certain link parameters.
fADC_CLK is the ADC sample rate; DCM is the chip decimation ratio; fOUT is the output sample rate = fADC_CLK/DCM; SLR is the JESD204B serial lane rate. The following equations
must be met due to internal clock divider requirements: SLR ≥ 1.6875 Gbps and SLR ≤ 16 Gbps; SLR/40 ≤ fADC_CLK; least common multiple (20 × DCM × fOUT/SLR, DCM) ≤ 64.
When the SLR is ≤ 16,000 Mbps and > 13,500 Mbps, Register 0x056E must be set to 0x30. When the SLR is ≤ 13,500 Mbps and ≥ 6750 Mbps, Register 0x056E must be
set to 0x00. When the SLR is < 6750 Mbps and ≥ 3375 Mbps, Register 0x056E must be set to 0x10. When the SLR is < 3375 Mbps and ≥ 1687.5 Mbps, Register 0x056E
must be set to 0x50.
3
JESD204B transport layer descriptions are as follows: L is the number of lanes per converter device (lanes per link); M is the number of virtual converters per converter device
(virtual converters per link); F is the octets per frame; S is the samples transmitted per virtual converter per frame cycle; HD is the high density mode; N is the virtual converter
resolution (in bits); N' is the total number of bits per sample (JESD204B word size); CS is the number of control bits per conversion sample; K is the number of frames per
multiframe.
4
Only valid K × F values that are divisible by 4 are supported: for F = 1, K = 20, 24, 28, 32; for F = 2, K = 12, 16, 20, 24, 28, 32; for F = 4, K = 8, 12, 16, 20, 24, 28, 32; for F = 8, K = 4,
8, 12, 16, 20, 24, 28, 32; and for F = 16, K = 4, 8, 12, 16, 20, 24, 28, 32.
2
Rev. 0 | Page 77 of 135
AD9695
Data Sheet
Example Setup 1—Full Bandwidth Mode
1300MSPS
13GBit/sec
SYNC~
F=1
VIN_B
REAL
14-BIT
ADC
CORE
14-BIT
ADC
CORE
L0
M0
L1
M1
JESD204B
LINK
(L = 4,
M = 2,
F = 1,
S = 2,
N' = 16,
N = 16,
CS = 0,
HD = 1)
L2
L3
I = REAL COMPONENT
Q = QUADRATURE COMPONENT
DCM = DECIMATION
C2R = COMPLEX TO REAL
MX = VIRTUAL CONVERTER X
LY = LANE Y
SZ = SAMPLE Z INSIDE A JESD204B FRAME
C = CONTROL BIT (OVERRANGE, AMONG OTHERS)
T = TAIL BIT
M0S0[15:8]
M0S0[7:0]
M0S1[15:8]
M0S1[7:0]
15660-098
VIN_A
REAL
Figure 127. Full Bandwidth Mode
The AD9695 is set up as shown in Figure 127, with the
following configurations:
The JESD204B supported output configurations are as follows
(see Table 34):









Two 14-bit converters at 1300 MSPS.
Full bandwidth application layer mode.
Decimation filters bypassed.
The JESD204B output configuration is as follows:


Two virtual converters required (see Table 34).
Output sample rate (fOUT) = 1300/1 = 1300 MSPS.

Rev. 0 | Page 78 of 135
N΄ = 16 bits.
N = 14 bits.
L = 4, M = 2, and F = 1.
CS = 0.
K = 32.
Output serial lane rate:
 13 Gbps per lane (L = 4).
PLL control register:
 Register 0x056E is set to 0x00 (L = 4).
Data Sheet
AD9695
Example Setup 2—ADC with DDC Option (Two ADCs Plus Two DDCs)
Fs (MHz)
FOUT = Fs / 4 (MHz)
40 x FOUT MbPS
SYNCAB~
14-BIT ADC CORE
DDC0
(REAL INPUT,
DCM = 4
C2R = BYPASS)
JESD204B
LINK
(L = 2
M=4
F = 4,
S = 1,
N' = 16,
N = 16,
CS = 0,
HD =0)
M1(Q)
HB2_HB1 USED
VIN_B
REAL
14-BIT ADC CORE
DDC1
(REAL INPUT,
DCM = 4
C2R = BYPASS)
M2(I)
LAB0
LAB1
M3(Q)
M2(I)S0[15:8]
M2(I)S0[7:0]
M3(Q)S0[15:8]
M3(Q)S0[7:0]
VIN_A
REAL
M0(I)S0[15:8]
M0(I)S0[7:0]
M1(Q)S0[15:8]
M1(Q)S0[7:0]
F=4
M0(I)
LEGEND
15660-099
I = REAL COMPONENT
Q = QUADRATURE COMPONENT
DCM = DECIMATION
C2R = COMPLEX TO REAL
MX= VIRTUAL CONVERTER X
LY = LINK LANE Y
SZ = SAMPLE Z INSIDE A JESD204B FRAME
C = CONTROL BIT (OVER-RANGE, ETC.)
T = TAIL BIT
Figure 128. Two ADCs Plus Two DDCs Mode (L = 4, M = 4, F = 2, S = 1)
This example shows the flexibility in the digital and lane
configurations for the AD9695. The sample rate is 1300 MSPS;
whereas the outputs are all combined in a combination of either
two, four, or eight lanes, depending on the input/output speed
capability of the receiving device.
The AD9695 is set up as shown in Figure 128, with the
following configuration:




Two 14-bit converters at 1300 MSPS.
Two DDC application layer mode with complex outputs
(I/Q).
Chip decimation ratio = 4.
DDC decimation ratio = 4 (see the Memory Map section).
The JESD204B output configuration is as follows:


Four virtual converters required (see Table 34).
Output sample rate (fOUT) = 1300 MSPS/4 = 325 MSPS.
The JESD204B supported output configurations are as follows
(see Table 34):






N΄ = 16 bits.
N = 14 bits.
L = 2, M = 4, and F = 4, or L = 4, M = 4, and F = 4.
CS = 0.
K = 32.
Output serial lane rate = 6.5 Gbps per lane (L = 4), 13 Gbps
per lane (L = 2)
For L = 2, set the PLL control register, Register 0x056E, to 0x00.
For L = 4, set the PLL control register, Register 0x056E, to 0x10.
Rev. 0 | Page 79 of 135
AD9695
Data Sheet
DETERMINISTIC LATENCY
Both ends of the JESD204B link contain various clock domains
distributed throughout each system. Data traversing from one
clock domain to a different clock domain can lead to ambiguous
delays in the JESD204B link. These ambiguities lead to nonrepeatable latencies across the link from one power cycle or link
reset to the next. Section 6 of the JESD204B specification addresses
the issue of deterministic latency with mechanisms defined as
Subclass 1 and Subclass 2.
boundaries and buffering to achieve consistent latency across
lanes (or even multiple devices), and to achieve a fixed latency
between power cycles and link reset conditions.
Deterministic Latency Requirements
Several key factors are required for achieving deterministic
latency in a JESD204B Subclass 1 system:

The AD9695 supports JESD204B Subclass 0 and Subclass 1
operation. Register 0x0590, Bit 5 sets the subclass mode for the
AD9695; the default mode is the Subclass 1 operating mode
(Register 0x0590, Bit 5 = 1). If deterministic latency is not a
system requirement, Subclass 0 operation is recommended and
the SYSREF± signal may not be required. Even in Subclass 0
mode, the SYSREF± signal may be required in an application
where multiple AD9695 devices must be synchronized with each
other. This topic is addressed in the Timestamp Mode section.


SYSREF± signal distribution skew within the system must
be less than the desired uncertainty for the system.
SYSREF± setup and hold time requirements must be met
for each device in the system.
The total latency variation across all lanes, links, and
devices must be ≤1 LMFC period (see Figure 129). This
includes both variable delays and the variation in fixed
delays from lane to lane, link to link, and device to device
in the system.
Setting Deterministic Latency registers
SUBCLASS 0 OPERATION
The JESD204B receive buffer in the logic device buffers data
starting on the LMFC boundary. If the total link latency in the
system is near an integer multiple of the LMFC period, it is possible
that from one power cycle to the next, the data arrival time at
the receive buffer may straddle an LMFC boundary. To ensure
deterministic latency in this case, a phase adjustment of the
LMFC at either the transmitter or receiver must be performed.
Typically, adjustments to accommodate the receive buffer are
made to the LMFC of the receiver. In the AD9695, this adjustment
can be made using the LMFC offset bits (Register 0x578, Bits[4:0]).
These bits delay the LMFC in frame clock increments, depending
on the F parameter, which is the number of octets per lane per
frame). For F = 1, every fourth setting (0, 4, 8, …, and so on)
results in a one frame clock shift. For F = 2, every other setting
(0, 2, 4, …, and so on) results in a 1-frame clock shift. For all
other values of F, each setting results in a 1-frame clock shift.
If there is no requirement for multichip synchronization while
operating in Subclass 0 mode (Register 0x0590, Bit 5 = 0), the
SYSREF± input can be left disconnected. In this mode, the
relationship of the JESD204B clocks between the JESD204B transmitter and receiver are arbitrary but does not affect the ability of the
receiver to capture and align the lanes within the link.
SUBCLASS 1 OPERATION
The JESD204B protocol organizes data samples into octets, frames,
and multiframes as described in the Transport Layer section.
The LMFC is synchronous with the beginnings of these
multiframes. In Subclass 1 operation, the SYSREF± signal
synchronizes the LMFCs for each device in a link or across
multiple links (within the AD9695, SYSREF± also synchronizes
the internal sample dividers). This synchronization is shown in
Figure 129. The JESD204B receiver uses the multiframe
SYSREF
DEVICE CLOCK
SYSREF-ALIGNED
GLOBAL LMFC
SYSREF TO LMFC DELAY
ALL LMFCs
DATA
ILAS
Figure 129. SYSREF± and LMFC
Rev. 0 | Page 80 of 135
DATA
15660-100
POWER CYCLE VARIATION
(MUST BE < tLMFC)
Data Sheet
AD9695
Figure 130 shows that, in the case where the link latency is near
an LMFC boundary, the local LMFC of the AD9695 can be
delayed to in turn delay the data arrival time at the receiver.
Figure 131 shows how the LMFC of the receiver is delayed to
accommodate the receive buffer timing. Refer to the applicable
JESD204B receiver user guide for details on making this
adjustment. If the total latency in the system is not near an
integer multiple of the LMFC period, or if the appropriate
adjustments are made to the LMFC phase at the clock source, it
is still possible to have variable latency from one power cycle to
the next. In this case, check for the possibility that the setup and
hold time requirements for the SYSREF± signal are not being
met. Perform this check by reading the SYSREF± setup and
hold monitor register (Register 0x0128). This function is
described in the SYSREF± Setup/Hold Window Monitor
section.
If reading Register 0x0128 indicates a timing problem, there are
adjustments that can made in the AD9695. Changing the
SYSREF± level used for alignment is possible using the SYSREF±
transition select bit (Register 0x0120, Bit 4). Also, changing
which edge of the clock is used to capture SYSREF± can be
performed using the clock edge select bit (Register 0x0120,
Bit 3). Both of these options are described in the SYREF±
Control Features section. If neither of these measures help
achieve an acceptable setup and hold time, adjusting the phase
of SYSREF± and/or the device clock (CLK±) may be required.
POWER CYCLE VARIATION
LMFCTX DELAY TIME
SYSREF-ALIGNED
GLOBAL LMFC
Tx LOCAL LMFC
ILAS
DATA
(AT Rx INPUT)
DATA
ILAS
DATA
Tx LMFC MOVED (DELAYING THE ARRIVAL OF DATA RELATIVE
TO THE GLOBAL LMFC) SO THE RECIEVE BUFFER RELEASE TIME
IS ALWAYS REFERENCED TO THE SAME LMFC EDGE
15660-101
DATA
(AT Tx INPUT)
Figure 130. Adjusting the JESD204B Tx LMFC in the AD9695
LMFCRX DELAY TIME
POWER CYCLE VARIATION
SYSREF-ALIGNED
GLOBAL LMFC
DATA
(AT Tx INPUT)
DATA
(AT Rx INPUT)
DATA
ILAS
ILAS
ILAS
DATA
Rx LMFC MOVED SO THE RECEIVE BUFFER RELEASE TIME
IS ALWAYS REFERENCED TO THE SAME LMFC EDGE
Figure 131. Adjusting the JESD204B Rx LMFC in the Logic Device
Rev. 0 | Page 81 of 135
15660-102
Rx LOCAL LMFC
AD9695
Data Sheet
MULTICHIP SYNCHRONIZATION
The flowchart shown in Figure 133 describes the internal
mechanism for multichip synchronization in the AD9695.
There are two methods by which multichip synchronization can
take place, as determined by the chip synchronization mode bit
(Register 0x1FF , Bit 0). Each method involves different
applications of the SYSREF± signal.
timestamp method is used for synchronization of multiple
channels and/or devices. In timestamp mode, the clocks are not
reset but instead, the coinciding sample is time stamped using
the JESD204B control bits of that sample. To operate in timestamp
mode, the following additional settings are necessary:

Continuous or N-shot SYSREF enabled (Register 0x0120,
Bits[2:1] = 1 or 2).
At least one control bit must be enabled (CS > 0,
Register 0x058F, Bits[7:6] = 1, 2, or 3).
Set the function for one of the control bits to SYSREF:
 Register 0x0559, Bits[2:0] = 5 if using Control Bit 0.
 Register 0x0559, Bits[6:4] = 5 if using Control Bit 1.
 Register 0x055A, Bits[2:0] = 5 if using Control Bit 2.
NORMAL MODE
The default sate of the chip synchronization mode bit is 0,
which configures the AD9695 for normal chip synchronization.
The JESD204B standard specifies the use of SYSREF± to provide
deterministic latency within a single link. This same concept, when
applied to a system with multiple converters and logic devices, can
also provide multichip synchronization. In Figure 133, this is
referred to as normal mode. Following the process outlined in the
flowchart ensures that the AD9695 is configured appropriately.
Consult the logic devices user intellectual property (IP) guide to
ensure that the JESD204B receivers are configured appropriately.
TIMESTAMP MODE
For all AD9695 full bandwidth operating modes, the SYSREF±
input can also be used to timestamp samples. This is another
method by which multiple channels and multiple devices can
achieve synchronization. This is especially effective when
synchronizing multiple devices to one or more logic devices.
The logic devices buffer the data streams, identify the time
stamped samples, and align them. When the chip synchronization mode bit (Register 0x1FF, Bit 0) is set to 1, the


Enable control bits MSB first. In other words, if only using one
control bit (CS = 1), then Control Bit 2 must be enabled. If two
control bits are used, then Control Bits[2:1] must be enabled.
Figure 132 shows how the input sample coincident with
SYSREF± is time stamped and ultimately output of the ADC. In
this example, there are two control bits and Control Bit 1 is the
bit indicating which sample was coincident with the SYSREF
rising edge. Note that the pipeline latencies for each channel are
identical. If so desired, the SYSREF timestamp delay register
(Register 0x0123) can be used to adjust the timing of which
sample is time stamped.
Note that time stamping is not supported by any AD9695
operating modes that use decimation.
14-BIT SAMPLES OUT
N–1
N+1
N+2
N+3
N – 1 00
N
CONTROL BIT 0 USED TO
TIME STAMP SAMPLE N
ENCODE CLK
SYSREF
AINB
01 N + 1 00 N + 2 00 N + 3 00
CHANNEL A
N–1
N
CHANNEL B
N+1
N+2
N – 1 00
N
01 N + 1 00 N + 2 00 N + 3 00
N+3
2 CONTROL BITS
15660-103
AINA
N
Figure 132. AD9695 Timestamping—CS = 2 (Register 0x058F, Bits[7:6] = 2), Control Bit 1 is SYSREF± (Register 0x0559, Bits[6:4] = 5)
Rev. 0 | Page 82 of 135
Data Sheet
AD9695
INCREMENT
SYSREF IGNORE
COUNTER
START
NO
NO
RESET
SYSREF IGNORE
COUNTER
NO
SYSREF
ENABLED?
(0x0120)
YES
SYSREF
ASSERTED?
SYSREF
MODE
(0x0120)
YES
SYSREF
IGNORE
COUNTER
EXPIRED?
(0x0121)
N-SHOT
MODE
YES
CONTINUOUS
MODE
CLEAR SYSREF IGNORE COUNTER
AND DISABLE SYSREF
(CLEAR BIT 2 IN 0x0120)
UPDATE SETUP/HOLD
DETECTOR STATUS
(0x0128)
ALIGN CLOCK
DIVIDER PHASE
TO SYSREF
YES
INPUT
CLOCK DIVIDER
ALIGNMENT
REQUIRED?
NO
SYNCHRONIZATION
MODE?
(0x01FF)
TIMESTAMP
MODE
SYSREF
TIMESTAMP
DELAY
(0x0123)
YES
CLOCK
DIVIDER
AUTO ADJUST
ENABLED?
CLOCK
DIVIDER
>1?
(0x010B)
YES
NO
INCREMENT
SYSREF COUNTER
(0x012A)
NO
SYSREF
ENABLED
IN CONTROL BITS?
(0x0559, 0x055A,
0x058F)
SYSREF INSERTED
IN JESD204B
CONTROL BITS
YES
NO
RAMP
TEST MODE
ENABLED?
(0x0550)
NORMAL
MODE
SYSREF RESETS
RAMP TEST MODE
GENERATOR
YES
BACK TO START
NO
JESD204B
LMFC
ALIGNMENT
REQUIRED?
YES
ALIGN PHASE OF ALL
INTERNAL CLOCKS
(INCLUDING LMFC)
TO SYSREF
SEND INVALID 8-BIT/
10-BIT CHARACTERS
(ALL 0s)
SYNC~
ASSERTED
NO
NO
SEND K28.5
CHARACTERS
NORMAL
JESD204B
INITIALIZATION
NO
YES
ALIGN SIGNAL
MONITOR
COUNTERS
DDC NCO
ALIGNMENT
ENABLED?
(0x0300)
YES
ALIGN DDC NCO
PHASE
ACCUMULATOR
BACK TO START
NO
15660-104
SIGNAL
MONITOR
ALIGNMENT
ENABLED?
(0x026F)
YES
Figure 133. SYSREF± Capture Scenarios and Multichip Synchronization
Rev. 0 | Page 83 of 135
AD9695
Data Sheet
SETUP
REQUIREMENT
–70ps
HOLD
REQUIREMENT
120ps
SYSREF
KEEP OUT WINDOW
The input clock divider, DDCs, signal monitor block, and
JESD204B link are all synchronized using the SYSREF± input
when in normal synchronization mode (Register 0x01FF,
Bits[1:0] = 0). The SYSREF± input can also be used to time
stamp an ADC sample to provide a mechanism for synchronizing
multiple AD9695 devices in a system. For the highest level of
timing accuracy, SYSREF± must meet the setup and hold
requirements relative to the CLK± input. There are several
features in the AD9695 to ensure these requirements are met
(see the SYREF± Control Features section).
SYREF± Control Features
SYSREF± is used, along with the input clock (CLK±), as part of
a source synchronous timing interface and requires setup and
hold timing requirements of 117 ps and −96 ps, relative to the
input clock (see Figure 134). The AD9695 has several features
to meet these requirements. First, the SYSREF± sample event
can be defined as either a synchronous low to high transition
or synchronous high to low transition. Second, the AD9695
allows the SYSREF± signal to be sampled using either the rising
edge or falling edge of the input clock. Figure 134, Figure 135,
Figure 136, and Figure 137 show all four possible combinations.
15660-105
CLK
Figure 134. SYSREF± Setup and Hold Time Requirements; SYSREF± Low to
High Transition Using the Rising Edge Clock (Default)
SETUP
REQUIREMENT
–70ps
LMFC = ADC Clock/S × K
HOLD
REQUIREMENT
120ps
SYSREF
SAMPLE POINT
CLK
15660-106
where:
S is the the JESD204B parameter for number of samples per
converter.
K is JESD204B parameter for number of frames per multiframe.
SYSREF
SAMPLE POINT
SYSREF
Figure 135. SYSREF± Low to High Transition Using Falling Edge Clock
Capture (Register 0x0120, Bit 4 = 1’b0 and Register 0x0120, Bit 3 = 1’b1)
SETUP
REQUIREMENT
–70ps
HOLD
REQUIREMENT
120ps
SYSREF
SAMPLE POINT
CLK
15660-107
The SYSREF± input signal is used as a high accuracy system
reference for deterministic latency and multichip synchronization.
The AD9695 accepts a single-shot or periodic input signal. The
SYSREF± mode select bits (Register 0x0120, Bits[2:1]) select the
input signal type and also arm the SYSREF± state machine
when set. If in single- (or N) shot mode (Register 0x0120,
Bits[2:1] = 2), the SYSREF± mode select bit self clears after the
appropriate SYSREF± transition is detected. The pulse width
must have a minimum width of two CLK± periods. If the clock
divider (Register 0x010B, Bits[2:0]) is set to a value other than
divide by 1, then multiply this minimum pulse width requirement
by the divide ratio (for emample, if set to divide by 8, the
minimum pulse width is 16 CLK± cycles). When using a
continuous SYSREF± signal (Register 0x0120, Bits[2:1] = 1), the
period of the SYSREF± signal must be an integer multiple of the
LMFC. Derive the LMFC using the following formula:
labeled as N-shot mode. The AD9695 is able to ignore N
SYSREF± events, which is useful to handle periodic SYSREF±
signals that require time to settle after startup. Ignoring SYSREF±
until the clocks in the system have settled avoids an inaccurate
SYSREF± trigger. Figure 138 shows an example of the SYSREF±
ignore feature when ignoring three SYSREF± events.
SYSREF
Figure 136. SYSREF± High to Low Transition Using Rising Edge Clock Capture
(Register 0x0120, Bit 4 = 1’b1 and Register 0x0120, Bit 3 = 1’b0)
SETUP
REQUIREMENT
–70ps
HOLD
REQUIREMENT
120ps
SYSREF
SAMPLE POINT
CLK
SYSREF
Figure 137. SYSREF± High to Low Transition Using Falling Edge Clock
Capture (Register 0x0120, Bit 4= 1’b1 and Register 0x0120, Bit 3 = 1’b1)
The third SYSREF± related feature available is the ability to
ignore a programmable number (up to 16) of SYSREF± events.
The SYSREF± ignore feature is enabled by setting the SYSREF±
mode register (Register 0x0120, Bits[2:1]) to 2’b10, which is
Rev. 0 | Page 84 of 135
15660-108
SYSREF± INPUT
Data Sheet
AD9695
SYSREF SAMPLE PART 1
SYSREF SAMPLE PART 2
SYSREF SAMPLE PART 3
SYSREF SAMPLE PART 4
SYSREF SAMPLE PART 5
CLK
15660-109
SYSREF
SAMPLE THE FOURTH SYSREF
IGNORE FIRST THREE SYSREFs
Figure 138. SYSREF± Ignore Example; SYSREF± Ignore Count Bits (Register 0x0121, Bits[3:0]) = 3
SYSREF SKEW WINDOW = ±3
SYSREF SKEW WINDOW = ±2
SYSREF SKEW WINDOW = ±1
SYSREF SKEW WINDOW = 0
15660-110
SAMPLE CLOCK
SYSREF
Figure 139. SYSREF± Skew Window
When in continuous SYSREF± mode (Register 0x0120, Bits[2:1] =
1), the AD9695 monitors the placement of the SYSREF± leading
edge compared to the internal LMFC. If the SYSREF± edge is
captured with a clock edge other than the one that is aligned
with LMFC, the AD9695 initiates a resynchronization of the
link. Because the input clock rates for the AD9695 can be up to
4 GHz, the AD9695 provides another SYSREF± related feature
that makes it possible to accommodate periodic SYSREF±
signals where cycle accurate capture is not feasible or not
required. For these scenarios, the AD9695 has a programmable
SYSREF± skew window that allows the internal dividers to
remain undisturbed, unless SYSREF± occurs outside the skew
window. The resolution of the SYSREF± skew window is set in
sample clock cycles. If the SYSREF± negative skew window is 1
and the positive skew window is 1, then the total skew window
is ±1 sample clock cycles, meaning that, as long as SYSREF± is
captured within ±1 sample clock cycle of the clock that is aligned
with LMFC, the link continues to operate normally. If the SYSREF±
has jitter, which can cause a misalignment between SYSREF± and
the LMFC, the system continues to run without a resynchronization, while still allowing the device to monitor for larger
errors not caused by jitter. For the AD9695, the positive and
negative skew window is controlled by the SYSREF± window
negative bits (Register 0x0122, Bits[3:2]) and the SYSREF±
window positive bits (Register 0x0122, Bits[1:0]). Figure 139
shows information on the location of the skew window settings
relative to Phase 0 of the internal dividers. Negative skew is defined
as occurring before the internal dividers reach Phase 0 and positive
skew is defined after the internal dividers reach Phase 0.
Rev. 0 | Page 85 of 135
AD9695
Data Sheet
SYSREF± SETUP/HOLD WINDOW MONITOR
To ensure a valid SYSREF± signal capture, the AD9695 has a
SYSREF± setup/hold window monitor. This feature allows the
system designer to determine the location of the SYSREF±
signals relative to the CLK± signals by reading back the amount
of setup/hold margin on the interface through the memory
map. Figure 140 and Figure 141 show the setup and hold status
values for different phases of SYSREF±. The setup detector
returns the status of the SYSREF± signal before the CLK± edge,
and the hold detector returns the status of the SYSREF signal
after the CLK± edge. Register 0x0128 stores the status of
SYSREF± and lets the user know if the SYSREF± signal is
captured by the ADC.
Table 37 describes the contents of Register 0x0128 and how to
interpret them.
0xF
0xE
0xD
0xC
0xB
0xA
0x9
REG 0x0128[3:0] 0x8
0x7
0x6
0x5
0x4
0x3
0x2
0x1
0x0
CLK±
INPUT
VALID
SYSREF±
INPUT
FLIP-FLOP
HOLD (MIN)
FLIP-FLOP
HOLD (MIN)
Figure 140. SYSREF± Setup Detector
Rev. 0 | Page 86 of 135
15660-111
FLIP-FLOP
SETUP (MIN)
Data Sheet
AD9695
0xF
0xE
0xD
0xC
0xB
0xA
0x9
0x8
0x7
0x6
0x5
0x4
0x3
0x2
0x1
REG 0x0128[7:4] 0x0
CLK±
INPUT
SYSREF±
INPUT
FLIP-FLOP
SETUP (MIN)
FLIP-FLOP
HOLD (MIN)
FLIP-FLOP
HOLD (MIN)
15660-112
VALID
Figure 141. SYSREF± Hold Detector
Table 37. SYSREF± Setup/Hold Monitor, Register 0x0128
Register 0x0128, Bits[7:4]
Hold Status
0x0
0x0 to 0x8
0x8
0x8
0x9 to 0xF
0x0
Register 0x0128, Bits[3:0]
Setup Status
0x0 to 0x7
0x8
0x9 to 0xF
0x0
0x0
0x0
Description
Possible setup error. The smaller this number, the smaller the setup margin.
No setup or hold error (best hold margin).
No setup or hold error (best setup and hold margin).
No setup or hold error (best setup margin).
Possible hold error. The larger this number, the smaller the hold margin.
Possible setup or hold error.
Rev. 0 | Page 87 of 135
AD9695
Data Sheet
LATENCY
END TO END TOTAL LATENCY
EXAMPLE LATENCY CALCULATIONS
Total latency in the AD9695 is dependent on the chip
application mode and the JESD204B configuration. For any
given combination of these parameters, the latency is
deterministic, however, the value of this deterministic latency
must be calculated as described in the Example Latency
Calculations section.
Example Configuration 1 is as follows:
Table 38 shows the combined latency through the ADC and
DSP for the different chip application modes supported by the
AD9695. Table 37 shows the latency through the JESD204B
block for each application mode based on the M/L ratio. For
both tables, latency is typical and is in units of the encode clock.
The latency through the JESD204B block does not depend on
the output data type (real or complex). Therefore, data type is
not included in Table 37.
To determine the total latency, select the appropriate ADC + DSP
latency from Table 34 and add it to the appropriate JESD204B
latency from Table 35. Example calculations are provided in the
following section.





ADC application mode = full bandwidth
Real outputs
L = 4, M = 2, F = 1, S = 1 (JESD204B mode)
M/L = 0.5
Latency = 31 + 25 = 56 encode clocks
Example Configuration 2 is as follows:





ADC application mode = DCM4
Complex outputs
L = 2, M = 2, F = 2, S = 1 (JESD204B mode)
M/L = 1
Latency = 162 + 50 = 212 encode clocks
LMFC REFERENCED LATENCY
Some FPGA vendors may require the end user to know LMFC
referenced latency to make appropriate deterministic latency
adjustments. If they are required, the latency values in Table 38
and Table 39 can be used for the analog input to LMFC and
LMFC to data output latency values, respectively.
Rev. 0 | Page 88 of 135
Data Sheet
AD9695
Table 38. Latency Through the ADC + DSP Blocks (Number of Sample Clocks)1
Chip Application Mode
Full Bandwidth
DCM1 (Real)
DCM2 (Complex)
DCM3 (Complex)
DCM2 (Real)
DCM4 (Complex)
DCM3 (Real)
DCM6 (Complex)
DCM4 (Real)
DCM8 (Complex)
DCM5 (Real)
DCM10 (Complex)
DCM6 (Real)
DCM12 (Complex)
DCM15 (Real)
DCM8 (Real)
DCM16 (Complex)
DCM10 (Real)
DCM20 (Complex)
DCM12 (Real)
DCM24 (Complex)
DCM30 (Complex)
DCM20 (Real)
DCM40 (Complex)
DCM24 (Real)
DCM48 (Complex)
1
Enabled Filters
Not applicable
HB1
HB1
TB1
HB2 + HB1
HB2 + HB1
TB2 + HB1
TB2 + HB1
HB3 +HB2 + HB1
HB3 +HB2 + HB1
FB2 + HB1
FB2 + HB1
TB2 + HB2 + HB1
TB2 + HB2 + HB1
FB2 + TB1
HB4 + HB3 + HB2 + HB1
HB4 + HB3 + HB2 + HB1
FB2 + HB2 + HB1
FB2 + HB2 + HB1
TB2 + HB3 + HB2 + HB1
TB2 + HB3 + HB2 + HB1
HB2 + FB2 + TB1
FB2 + HB3 + HB2 + HB1
FB2 + HB3 + HB2 + HB1
TB2 + HB4 + HB3 + HB2 + HB1
TB2 + HB4 + HB3 + HB2 + HB1
ADC + DSP Latency
31
90
90
102
162
162
212
212
292
292
380
380
424
424
500
552
552
694
694
814
814
836
1420
1420
1594
1594
DCMx indicates the decimation ratio.
Table 39. Latency Through JESD204B Block (Number of Sample Clocks)1
Chip Application Mode
Full Bandwidth
DCM1
DCM2
DCM3
DCM4
DCM5
DCM6
DCM8
DCM10
DCM12
DCM15
DCM16
DCM20
DCM24
DCM30
DCM40
DCM48
0.125
82
82
160
237
315
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.25
44
44
84
124
164
2033
243
323
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.5
25
25
46
67
88
1093
130
172
213
255
3184
3394
N/A
N/A
N/A
N/A
N/A
1
M/L Ratio2
1
14
14
27
39
50
623
73
96
119
142
1764
1884
233
279
3484
N/A
N/A
N/A means not applicable and indicates that the application mode is not supported at the M/L ratio listed.
M/L ratio is the number of converters divided by the number of lanes for the configuration.
3
The application mode at the M/L ratio listed is only supported in real output mode.
4
The application mode at the M/L ratio listed is only supported in complex output mode.
2
Rev. 0 | Page 89 of 135
2
7
7
14
21
27
433
39
50
62
73
904
964
119
142
1764
2334
2794
4
9
N/A
7
11
14
N/A
21
27
33
39
474
504
62
73
904
1194
1424
8
3
N/A
N/A
N/A
9
N/A
14
18
22
27
334
354
43
51
624
824
974
AD9695
Data Sheet
TEST MODES
If the application mode is set to select a DDC mode of
operation, the test modes must be enabled for each DDC
enabled. The test patterns can be enabled via Bit 2 and Bit 0 of
Register 0x0327, Register 0x0347, and Register 0x0367,
depending on which DDC(s) are selected. The (I) data uses the
test patterns selected for Channel A, and the (Q) data uses the
test patterns selected for Channel B. For DDC3 only, the (I) data
uses the test patterns from Channel A, and the (Q) data does
not output test patterns. Bit 0 of Register 0x0387 selects the
Channel A test patterns to be used for the (I) data. For more
information, see the AN-877 Application Note, Interfacing to
High Speed ADCs via SPI.
ADC TEST MODES
The AD9695 has various test options that aid in the system level
implementation. The AD9695 has ADC test modes that are
available in Register 0x0550. These test modes are described in
Table 36. When an output test mode is enabled, the analog section
of the ADC is disconnected from the digital back end blocks,
and the test pattern is run through the output formatting block.
Some of the test patterns are subject to output formatting and
some are not. The pseudorandom number (PN) generators
from the PN sequence tests can be reset by setting Bit 4 or Bit 5
of Register 0x0550. These tests can be performed with or
without an analog signal (if present, the analog signal is
ignored); however, they do require an encode clock.
JESD204B
INTERFACE
TEST PATTERN
(REG 0x573,
REG 0x551 TO
REG 0x558)
JESD204B
LONG TRANSPORT
TEST PATTERN
REG 0x571[5]
SERIALIZER
MSB A13
A12
A11
A10
A9
A8
A7
LSB A6
A5
A4
A3
A2
A1
A0
C2
T
MSB S7
S6
S5
S4
S3
S2
S1
LSB S0
S7
S6
S5
S4
S3
S2
S1
S0
8-BIT/10-BIT
ENCODER
a b c d e f g h i j
a b
SERDOUT0±
SERDOUT1±
i j a b
SYMBOL0
i j
SYMBOL1
a b c d e f g h i j
15660-214
TAIL BITS
0x571[6]
SCRAMBLER
1 + x14 + x15
(OPTIONAL)
OCTET 1
JESD204B SAMPLE
CONSTRUCTION
MSB A13
A12
A11
A10
A9
A8
A7
A6
A5
A4
A3
A2
A1
LSB A0
OCTET 1
OCTET 0
FRAME
CONSTRUCTION
OCTET 0
ADC TEST PATTERNS
(RE0x550,
REG 0x551 TO
REG 0x558)
ADC
JESD204B DATA
LINK LAYER TEST
PATTERNS
REG 0x574[2:0]
C2
CONTROL BITS C1
C0
Figure 142. ADC Output Datapath Showing Data Framing
Table 36. ADC Test Modes1
Output Test Mode
Bit Sequence
0000
0001
0010
0011
0100
0101
0110
0111
1000
Pattern Name
Off (default)
Midscale short
Positive full-scale short
Negative full-scale short
Checkerboard
PN sequence long
PN sequence short
One-/zero-word toggle
User input
Expression
N/A
0000 0000 0000
01 1111 1111 1111
10 0000 0000 0000
10 1010 1010 1010
x23 + x18 + 1
x9 + x5 + 1
11 1111 1111 1111
Register 0x0551 to
Register 0x0558
1111
Ramp output
(x) % 214
1
Default/
Seed Value
N/A
N/A
N/A
N/A
N/A
0x3AFF
0x0092
N/A
N/A
N/A
N/A means not applicable.
Rev. 0 | Page 90 of 135
Sample (N, N + 1, N + 2, …)
N/A
N/A
N/A
N/A
0x1555, 0x2AAA, 0x1555, 0x2AAA, 0x1555
0x3FD7, 0x0002, 0x26E0, 0x0A3D, 0x1CA6
0x125B, 0x3C9A, 0x2660, 0x0C65, 0x0697
0x0000, 0x3FFF, 0x0000, 0x3FFF, 0x0000
User Pattern 1[15:2], User Pattern 2[15:2],
User Pattern 3[15:2], User Pattern 4[15:2],
User Pattern 1[15:2] … for repeat mode.
User Pattern 1[15:2], User Pattern 2[15:2],
User Pattern 3[15:2], User Pattern 4[15:2],
0x0000 … for single mode.
(x) % 214, (x +1) % 214, (x +2) % 214, (x +3) % 214
Data Sheet
AD9695
JESD204B BLOCK TEST MODES
In addition to the ADC pipeline test modes, the AD9695 also has
flexible test modes in the JESD204B block. These test modes are
listed in Register 0x0573 and Register 0x0574. These test patterns
can be injected at various points along the output datapath. These
test injection points are shown in Figure 142. Table 37 describes
the various test modes available in the JESD204B block. For the
AD9695, a transition from test modes (Register 0x0573 ≠ 0x00)
to normal mode (Register 0x0573 = 0x00) requires an SPI soft
reset. This is done by writing 0x81 to Register 0x0000 (self cleared).
Transport Layer Sample Test Mode
The transport layer samples are implemented in the AD9695 as
defined by Section 5.1.6.3 in the JEDEC JESD204B specification.
These tests are shown in Register 0x0571, Bit 5. The test pattern
is equivalent to the raw samples from the ADC.
Interface Test Modes
The interface test modes are described in Register 0x0573, Bits[3:0].
These test modes are also explained in Table 37. The interface tests
can be injected at various points along the data. See Figure 142
for more information on the test injection points. Register 0x0573,
Bits[5:4] show where these tests are injected.
Table 38, Table 39, and Table 40 show examples of some of the
test modes when injected at the JESD sample input, PHY 10-bit
input, and scrambler 8-bit input. UPx in the tables represent the
user pattern control bits from the customer register map.
Table 37. JESD204B Interface Test Modes
Output Test Mode
Bit Sequence
0000
0001
0010
0011
0100
0101
0110
0111
1000
1110
1111
Pattern Name
Off (default)
Alternating checker board
1/0 word toggle
31-bit PN sequence
23-bit PN sequence
15-bit PN sequence
9-bit PN sequence
7-bit PN sequence
Ramp output
Continuous/repeat user test
Single user test
Expression
Not applicable
0x5555, 0xAAAA, 0x5555, …
0x0000, 0xFFFF, 0x0000, …
x31 + x28 + 1
x23 + x18 + 1
x15 + x14 + 1
x9 + x5 + 1
x7 + x6 + 1
(x) % 216
Register 0x0551 to Register 0x0558
Register 0x0551 to Register 0x0558
Default
Not applicable
Not applicable
Not applicable
0x0003AFFF
0x003AFF
0x03AF
0x092
0x07
Ramp size depends on test injection point
User Pattern 1 to User Pattern 4, then repeat
User Pattern 1 to User Pattern 4, then zeros
Table 38. JESD204B Sample Input for M = 2, S = 2, N' = 16 (Register 0x0573, Bits[5:4] = 'b00)
Frame
Number
0
0
0
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
Converter
Number
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
Sample
Number
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Alternating
Checkerboard
0x5555
0x5555
0x5555
0x5555
0xAAAA
0xAAAA
0xAAAA
0xAAAA
0x5555
0x5555
0x5555
0x5555
0xAAAA
0xAAAA
0xAAAA
0xAAAA
0x5555
0x5555
0x5555
0x5555
1/0 Word
Toggle
0x0000
0x0000
0x0000
0x0000
0xFFFF
0xFFFF
0xFFFF
0xFFFF
0x0000
0x0000
0x0000
0x0000
0xFFFF
0xFFFF
0xFFFF
0xFFFF
0x0000
0x0000
0x0000
0x0000
Ramp
(x) % 216
(x) % 216
(x) % 216
(x) % 216
(x +1) % 216
(x +1) % 216
(x +1) % 216
(x +1) % 216
(x +2) % 216
(x +2) % 216
(x +2) % 216
(x +2) % 216
(x +3) % 216
(x +3) % 216
(x +3) % 216
(x +3) % 216
(x +4) % 216
(x +4) % 216
(x +4) % 216
(x +4) % 216
Rev. 0 | Page 91 of 135
PN9
0x496F
0x496F
0x496F
0x496F
0xC9A9
0xC9A9
0xC9A9
0xC9A9
0x980C
0x980C
0x980C
0x980C
0x651A
0x651A
0x651A
0x651A
0x5FD1
0x5FD1
0x5FD1
0x5FD1
PN23
0xFF5C
0xFF5C
0xFF5C
0xFF5C
0x0029
0x0029
0x0029
0x0029
0xB80A
0xB80A
0xB80A
0xB80A
0x3D72
0x3D72
0x3D72
0x3D72
0x9B26
0x9B26
0x9B26
0x9B26
User Repeat
UP1[15:0]
UP1[15:0]
UP1[15:0]
UP1[15:0]
UP2[15:0]
UP2[15:0]
UP2[15:0]
UP2[15:0]
UP3[15:0]
UP3[15:0]
UP3[15:0]
UP3[15:0]
UP4[15:0]
UP4[15:0]
UP4[15:0]
UP4[15:0]
UP1[15:0]
UP1[15:0]
UP1[15:0]
UP1[15:0]
User Single
UP1[15:0]
UP1[15:0]
UP1[15:0]
UP1[15:0]
UP2[15:0]
UP2[15:0]
UP2[15:0]
UP2[15:0]
UP3[15:0]
UP3[15:0]
UP3[15:0]
UP3[15:0]
UP4[15:0]
UP4[15:0]
UP4[15:0]
UP4[15:0]
0x0000
0x0000
0x0000
0x0000
AD9695
Data Sheet
Table 39. Physical Layer 10-Bit Input (Register 0x0573, Bits[5:4] = 'b01)
10-Bit Symbol
Number
0
1
2
3
4
5
6
7
8
9
10
11
Alternating
Checkerboard
0x155
0x2AA
0x155
0x2AA
0x155
0x2AA
0x155
0x2AA
0x155
0x2AA
0x155
0x2AA
1/0 Word
Toggle
0x000
0x3FF
0x000
0x3FF
0x000
0x3FF
0x000
0x3FF
0x000
0x3FF
0x000
0x3FF
Ramp
(x) % 210
(x + 1) % 210
(x + 2) % 210
(x + 3) % 210
(x + 4) % 210
(x + 5) % 210
(x + 6) % 210
(x + 7) % 210
(x + 8) % 210
(x + 9) % 210
(x + 10) % 210
(x + 11) % 210
PN9
0x125
0x2FC
0x26A
0x198
0x031
0x251
0x297
0x3D1
0x18E
0x2CB
0x0F1
0x3DD
PN23
0x3FD
0x1C0
0x00A
0x1B8
0x028
0x3D7
0x0A6
0x326
0x10F
0x3FD
0x31E
0x008
User Repeat
UP1[15:6]
UP2[15:6]
UP3[15:6]
UP4[15:6]
UP1[15:6]
UP2[15:6]
UP3[15:6]
UP4[15:6]
UP1[15:6]
UP2[15:6]
UP3[15:6]
UP4[15:6]
User Single
UP1[15:6]
UP2[15:6]
UP3[15:6]
UP4[15:6]
0x000
0x000
0x000
0x000
0x000
0x000
0x000
0x000
Table 40. Scrambler 8-Bit Input (Register 0x0573, Bits[5:4] = 'b10)
8-Bit Octet
Number
0
1
2
3
4
5
6
7
8
9
10
11
Alternating
Checkerboard
0x55
0xAA
0x55
0xAA
0x55
0xAA
0x55
0xAA
0x55
0xAA
0x55
0xAA
1/0 Word
Toggle
0x00
0xFF
0x00
0xFF
0x00
0xFF
0x00
0xFF
0x00
0xFF
0x00
0xFF
Ramp
(x) % 28
(x + 1) % 28
(x + 2) % 28
(x + 3) % 28
(x + 4) % 28
(x + 5) % 28
(x + 6) % 28
(x + 7) % 28
(x + 8) % 28
(x + 9) % 28
(x + 10) % 28
(x + 11) % 28
Data Link Layer Test Modes
The data link layer test modes are implemented in the AD9695
as defined by Section 5.3.3.8.2 in the JEDEC JESD204B
specification. These tests are shown in Register 0x0574,
PN9
0x49
0x6F
0xC9
0xA9
0x98
0x0C
0x65
0x1A
0x5F
0xD1
0x63
0xAC
PN23
0xFF
0x5C
0x00
0x29
0xB8
0x0A
0x3D
0x72
0x9B
0x26
0x43
0xFF
User Repeat
UP1[15:9]
UP2[15:9]
UP3[15:9]
UP4[15:9]
UP1[15:9]
UP2[15:9]
UP3[15:9]
UP4[15:9]
UP1[15:9]
UP2[15:9]
UP3[15:9]
UP4[15:9]
User Single
UP1[15:9]
UP2[15:9]
UP3[15:9]
UP4[15:9]
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
Bits[2:0]. Test patterns inserted at this point are useful for
verifying the functionality of the data link layer. When the data
link layer test modes are enabled, disable SYNCINB± by writing
0xC0 to Register 0x0572.
Rev. 0 | Page 92 of 135
Data Sheet
AD9695
SERIAL PORT INTERFACE (SPI)
The AD9695 SPI allows the user to configure the converter for
specific functions or operations through a structured register
space provided inside the ADC. The SPI gives the user added
flexibility and customization, depending on the application.
Addresses are accessed via the serial port and can be written to
or read from via the port. Memory is organized into bytes that
can be further divided into fields. These fields are documented
in the Memory Map section. For detailed operational information,
see the Serial Control Interface Standard (Rev. 1.0).
CONFIGURATION USING THE SPI
Three pins define the SPI of the AD9695 ADC: the SCLK pin,
the SDIO pin, and the CSB pin (see Table 41). The SCLK (serial
clock) pin is used to synchronize the read and write data
presented to and from the ADC. The SDIO (serial data input/
output) pin is a dual-purpose pin that allows data to be sent and
read from the internal ADC memory map registers. The CSB
(chip select bar) pin is an active low control that enables or
disables the read and write cycles.
Table 41. Serial Port Interface Pins
Pin
SCLK
SDIO
CSB
Function
Serial clock. The serial shift clock input that is used to
synchronize serial interface, reads, and writes.
Serial data input/output. A dual-purpose pin that
typically serves as an input or an output, depending on
the instruction being sent and the relative position in the
timing frame.
Chip select bar. An active low control that gates the read
and write cycles.
The falling edge of CSB, in conjunction with the rising edge of
SCLK, determines the start of the framing. An example of the
serial timing and its definitions can be found in Figure 4 and
Table 5.
Other modes involving the CSB pin are available. The CSB pin
can be held low indefinitely, which permanently enables the
device; this is called streaming. The CSB can stall high between
bytes to allow additional external timing. When CSB is tied
high, SPI functions are placed in a high impedance mode. This
mode turns on any SPI pin secondary functions.
All data is composed of 8-bit words. The first bit of each
individual byte of serial data indicates whether a read or write
command is issued, which allows the SDIO pin to change
direction from an input to an output.
In addition to word length, the instruction phase determines
whether the serial frame is a read or write operation, allowing
the serial port to be used both to program the chip and to read
the contents of the on-chip memory. If the instruction is a
readback operation, performing a readback causes the SDIO pin
to change direction from an input to an output at the appropriate
point in the serial frame.
Data can be sent in MSB first mode or in LSB first mode. MSB first
is the default on power-up and can be changed via the SPI port
configuration register. For more information about this and other
features, see the Serial Control Interface Standard (Rev. 1.0).
HARDWARE INTERFACE
The pins described in Table 41 comprise the physical interface
between the user programming device and the serial port of the
AD9695. The SCLK pin and the CSB pin function as inputs
when using the SPI interface. The SDIO pin is bidirectional,
functioning as an input during write phases and as an output
during readback.
The SPI interface is flexible enough to be controlled by either
FPGAs or microcontrollers. One method for SPI configuration
is described in detail in the AN-812 Application Note,
Microcontroller-Based Serial Port Interface (SPI) Boot Circuit.
Do not activate the SPI port during periods when the full
dynamic performance of the converter is required. Because the
SCLK signal, the CSB signal, and the SDIO signal are typically
asynchronous to the ADC clock, noise from these signals can
degrade converter performance. If the on-board SPI bus is
used for other devices, it may be necessary to provide buffers
between this bus and the AD9695 to prevent these signals from
transitioning at the converter inputs during critical sampling
periods.
SPI ACCESSIBLE FEATURES
Table 42 provides a brief description of the general features that
are accessible via the SPI. These features are described in detail
in the Serial Control Interface Standard (Rev. 1.0). The AD9695
device specific features are described in the Memory Map section.
Table 42. Features Accessible Using the SPI
Feature Name
Mode
Clock
DDC
Test Input/Output
Output Mode
SERDES Output Setup
Description
Allows the user to set either power-down mode or standby mode.
Allows the user to access the clock divider via the SPI.
Allows the user to set up decimation filters for different applications.
Allows the user to set test modes to have known data on output bits.
Allows the user to set up outputs.
Allows the user to vary SERDES settings such as swing and emphasis.
Rev. 0 | Page 93 of 135
AD9695
Data Sheet
MEMORY MAP
READING THE MEMORY MAP REGISTER TABLE
Default Values
Each row in the memory map register table has eight bit
locations. The memory map is divided into the following
sections:
After the AD9695 is reset, critical registers are loaded with
default values. The default values for the registers are given in
the memory map register table, Table 47.

Logic Levels






Analog Devices SPI registers (Register 0x0000 to
Register 0x000F)
Clock/SYSREF/chip power-down pin control registers
(Register 0x003F to Register 0x0201)
Fast detect and signal monitor control registers
(Register 0x0245 to Register 0x027A)
DDC function registers (Register 0x0300 to
Register 0x03CD)
Digital outputs and test modes registers (Register 0x0550 to
Register 0x05CB)
Programmable filter control and coefficients registers
(Register 0x0DF8 to Register 0x0F7F)
VREF/analog input control registers (Register 0x18A6 to
Register 0x1A4D)
Table 47 (see the Memory Map Register Details section)
documents the default hexadecimal value for each hexadecimal
address shown. The column with the heading Bit 7 (MSB) is the
start of the default hexadecimal value given. For example,
Address 0x0561, the output sample mode register, has a
hexadecimal default value of 0x01, which means that Bit 0 = 1,
and the remaining bits are 0s. This setting is the default output
format value, which is twos complement. For more information
on this function and others, see Table 47.
Open and Reserved Locations
An explanation of logic level terminology follows:



“Bit is set” is synonymous with “bit is set to Logic 1” or
“writing Logic 1 for the bit.”
“Clear a bit” is synonymous with “bit is set to Logic 0” or
“writing Logic 0 for the bit.”
X denotes a don’t care bit.
Channel Specific Registers
Some channel setup functions, such as the input buffer control
register (Register 0x1A4C), can be programmed to a different
value for each channel. In these cases, channel address locations
are internally duplicated for each channel. These registers and
bits are designated in Table 47 as local. These local registers and
bits can be accessed by setting the appropriate Channel A or
Channel B bits in Register 0x0008. If both bits are set, the
subsequent write affects the registers of both channels. In a read
cycle, set only Channel A or Channel B to read one of the two
registers. If both bits are set during an SPI read cycle, the device
returns the value for Channel A. Registers and bits designated as
global in Table 47 affect the entire device and the channel
features for which independent settings are not allowed
between channels. The settings in Register 0x0005 do not affect
the global registers and bits.
SPI Soft Reset
All address and bit locations that are not included in Table 47
are not currently supported for this device. Write unused bits
of a valid address location with 0s unless the default value is
set otherwise. Writing to these locations is required only when
part of an address location is unassigned (for example,
Address 0x0561). If the entire address location is open (for
example, Address 0x0013), do not write to this address location.
After issuing a soft reset by programming 0x81 to Register 0x0000,
the AD9695 requires 5 ms to recover. When programming the
AD9695 for application setup, ensure that an adequate delay is
programmed into the firmware after asserting the soft reset and
before starting the device setup.
Rev. 0 | Page 94 of 135
Data Sheet
AD9695
MEMORY MAP REGISTERS
All address locations that are not included in Table 47 are not currently supported for this device and must not be written.
Table 47. Memory Map Registers
Address
Name
Analog Devices SPI Registers
0x0000
SPI
Configuration A
Bits
Bit Name
Settings
7
Soft reset mirror (self
clearing)
0
1
6
LSB first mirror
1
0
5
Address ascension
mirror
0
1
[4:3]
2
Reserved
Address ascension
0
1
1
LSB first
1
0
0
Soft reset (self clearing)
0
1
0x0001
0x0002
SPI
Configuration B
Chip
configuration
(local)
[7:2]
1
Reserved
Datapath soft reset
(self clearing)
0
[7:2]
[1:0]
Reserved
Reserved
Channel power mode
0
1
00
10
11
0x0003
Chip type
[7:0]
Chip type
0x3
0x0004
Chip ID LSB
[7:0]
Chip ID LSB [7:0]
0xDE
0x0005
Chip ID MSB
[7:0]
Chip ID MSB [15:8]
Rev. 0 | Page 95 of 135
Description
Reset
Access
Whenever a soft reset is issued, the
user must wait 5 ms before writing to
any other register; this provides
sufficient time for the boot loader to
complete.
Do nothing.
Reset the SPI and registers (self
clearing).
Least significant bit shifted first for all
SPI operations.
Most significant bit shifted first for all
SPI operations.
Multibyte SPI operations cause
addresses to auto decrement.
Multibyte SPI operations cause
addresses to auto increment.
Reserved.
Multibyte SPI operations cause
addresses to auto decrement.
Multibyte SPI operations cause
addresses to auto increment.
Least significant bit shifted first for all
SPI operations.
Most significant bit shifted first for all
SPI operations.
Whenever a soft reset is issued, the
user must wait 5 ms before writing to
any other register; this provides
sufficient time for the boot loader to
complete.
Do nothing.
Reset the SPI and registers (self
clearing).
Reserved.
Normal operation.
0x0
R/WC
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/WC
0x0
0x0
R
R/WC
0x0
0x0
0x0
R
R
R/W
0x3
R
Datapth soft reset (self clearing).
Reserved.
Reserved.
Channel power modes.
Normal mode (power-up).
Standby mode. Digital datapath
clocks disabled; JESD204B interface
enabled.
Power-down mode. Digital datapath
clocks disabled; digital datapath held
in reset; JESD204B interface disabled.
Chip type.
High speed ADC.
Chip ID.
AD9695.
Chip ID.
R
0x0
R
AD9695
Address
0x0006
Data Sheet
Name
Chip grade
Bits
[7:4]
Bit Name
Chip speed grade
Settings
0x00
0x01
0x0008
Device index
[3:0]
[7:2]
1
Reserved
Reserved
Channel B
0
1
0
Channel A
0
1
0x000A
Scratch pad
[7:0]
Scratch pad
0x000B
SPI revision
[7:0]
SPI revision
0x01
00000001
0x000C
0x000D
0x000F
Vendor ID LSB
Vendor ID MSB
Transfer
[7:0]
[7:0]
[7:1]
0
Reserved
Chip transfer
0
1
Clock/SYSREF/Chip Power-Down Pin Control Registers
0x003F
7
Chip powerLocal chip powerdown pin (local)
down pin disable
0
1
[6:0]
Reserved
Rev. 0 | Page 96 of 135
Description
Chip speed grade.
625 MSPS.
1300 MSPS.
Reserved.
Reserved.
ADC Core B does not receive the next
SPI command.
ADC Core B receives the next SPI
command.
ADC Core A does not receive the next
SPI command.
ADC Core A receives the next SPI
command.
Chip scratch pad register. This register
provides a consistent memory
location for software debug.
SPI revision register.
Revision 1.0.
Revision 1.0.
Vendor ID [7:0].
Vendor ID [15:8].
Reserved.
Self clearing chip transfer bit. This bit
is used to update the DDC phase
increment and phase offset registers
when the DDC phase update mode bit
(Register 0x0300, Bit 7) = 1. This setting
makes it possible to synchronously
update the DDC mixer frequencies. It
is also used to update the coefficients
for the programmable filter (PFILT).
Do nothing. This bit it is only cleared
after the transfer completes.
Self clearing bit used to synchronize
the transfer of data from master to
slave registers.
Function is determined by
Register 0x0040, Bits[7:6].
Power-down pin (PDWN/STBY)
enabled (default).
Power-down pin (PDWN/STBY)
disabled/ignored.
Reserved.
Reset
0x0
Access
R
DNC
0x0
0x1
R
R
R/W
0x1
R/W
0x0
R/W
0x1
R
0x56
0x04
0x0
0x0
R
R
R
R/W
0x0
R/W
0x0
R
Data Sheet
Address
0x0040
Name
Chip Pin Control 1
AD9695
Bits
[7:6]
Bit Name
Global chip powerdown pin functionality
Settings
00
01
10
[5:3]
Chip FD_B/GPIO_B0
pin functionality
000
001
110
111
[2:0]
Chip FD_A/GPIO_A0
pin functionality
000
001
110
111
0x0041
Chip Pin Control 2
[7:4]
Chip FD_B/GPIO_B0
pin secondary
functionality
0000
0001
1000
1001
[3:0]
Chip FD_A/GPIO_A0
pin secondary
functionality
0000
0001
1000
1001
Rev. 0 | Page 97 of 135
Description
External power-down pin functionality.
Assertion of the external power-down
pin (PDWN/STBY) has higher priority
than the channel power mode control
bits (Register 0x0002, Bits[1:0]). The
PDWN/STBY pin is only used when
Register 0x0040, Bits[7:6] = 00 or 01.
Power-down pin (default). Assertion
of the external power-down pin
(PDWN/STBY) causes the chip to
enter full power-down mode.
Standby-pin. Assertion of the external
power-down pin (PDWN/STBY) causes
the chip to enter standby mode.
Pin disabled. Power-down pin
(PDWN/STBY) is ignored.
Fast Detect B/GPIO B0 pin functionality.
Fast Detect B output.
JESD204B LMFC output.
Pin functionality determined by
Register 0x0041, Bits[7:4].
Disabled. Configured as input with
weak pull-down (default).
Fast Detect A/GPIO A0 pin functionality.
Fast Detect A output.
JESD204B LMFC output.
Pin functionality determined by
Regiser 0x0041, Bits[3:0].
Disabled. Configured as input with
weak pull-down. (default)
Fast Detect B/GPIO B0 pin secondary
functionality. Only used when
Register 0x0040, Bits[5:3] = 110.
Chip GPIO B0 input (NCO channel
selection).
Chip transfer input.
Master next trigger output (MNTO).
Slave next trigger input (SNTI).
Fast Detect A/GPIO B0 pin secondary
functionality. Only used when
Register 0x0040, Bits[2:0] = 110.
Chip GPIO A0 input (NCO channel
selection).
Chip transfer input.
Master next trigger output (MNTO).
Slave next trigger input (SNTI).
Reset
0x0
Access
R/W
0x7
R/W
0x7
R/W
0x0
R/W
0x0
R/W
AD9695
Address
0x0042
Data Sheet
Name
Chip Pin Control 3
Bits
[7:4]
Bit Name
Chip GPIO_B1 pin
functionality
Settings
0000
1000
1001
1111
[3:0]
Chip GPIO_A1 pin
functionality
0000
1000
1001
1111
0x0108
0x0109
Clock divider
control
[7:3]
[2:0]
Reserved
Input clock divider
(CLK± pins)
Clock divider
phase (local)
[7:4]
[3:0]
Reserved
Clock divider phase
offset
00
01
11
0000
0001
0010
1110
1111
0x010A
Clock divider and
SYSREF control
7
Clock divider
autophase adjust
enable
0
1
[6:4]
[3:2]
Reserved
Clock divider negative
skew window
0
1
10
11
Rev. 0 | Page 98 of 135
Description
GPIO_B1 pin functionality.
Chip GPIO B1 input (NCO channel
selection).
Master next trigger output (MNTO).
Slave next trigger input (SNTI).
Disabled (configured as an input with
a weak pull down).
GPIO_A1 pin functionality.
Chip GPIO A1 input (NCO channel
selection).
Master next trigger output (MNTO).
Slave next trigger input (SNTI).
Disabled (configured as an input with
a weak pull down).
Reserved.
Divide by 1.
Divide by 2.
Divide by 4.
Reserved.
0 input clock cycles delayed.
1/2 input clock cycles delayed (invert
clock).
1 input clock cycles delayed.
…
7 input clock cycles delayed.
7 1/2 input clock cycles delayed.
Clock divider autophase adjust enable.
When enabled, Register 0x0129,
Bits[3:0] contain the phase of the
divider when SYSREF occurred. The
actual divider phase offset =
Register 0x0129, Bits[3:0] +
Register 0x0109, Bits[3:0].
Clock divider phase is not changed by
SYSREF± (disabled).
Clock divider phase is automatically
adjusted by SYSREF± (enabled).
Reserved.
Clock divider negative skew window
(measured in 1/2 input device clocks).
Number of 1/2 clock cycles before the
input device clock by which captured
SYSREF± transitions are ignored. Only
used when Register 0x010A, Bit 7 = 1.
Register 0x010A, Bits[3:2] +
Register 0x010A, Bits[1:0] <
Register 0x0108, Bits[2:0]; this allows
some uncertainty in the sampling of
SYSREF± without disturbing the input
clock divider. Also, SYSREF± must be
disabled (Register 0x0120, Bits[2:1] =
0x0) when changing this control field.
No negative skew. SYSREF± must be
captured accurately.
1/2 device clock of negative skew.
1 device clocks of negative skew.
1 1/2 device clocks of negative skew.
Reset
0x0
Access
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
0x0
R
R/W
Data Sheet
Address
Name
AD9695
Bits
[1:0]
Bit Name
Clock divider positive
skew window
Settings
0
1
10
11
0x010B
Clock divider
SYSREF status
[7:4]
[3:0]
Reserved
Clock divider SYSREF±
offset
0x0110
Clock delay
control
[7:3]
[2:0]
Reserved
Clock delay mode
select
000
010
011
100
110
0x0111
Clock super fine
delay (local)
[7:0]
Clock super fine delay
adjust
Rev. 0 | Page 99 of 135
Description
Clock divider positive skew window
(measured in 1/2 input device clocks).
Number of clock cycles after the
input device clock by which captured
SYSREF± transitions aree ignored.
Only used when Register 0x010A,
Bit 7 = 1. Register 0x010A, Bits[3:2] +
Register 0x010A, Bits[1:0] <
Register 0x0108, Bits[2:0]; this allows
some uncertainty in the sampling of
SYSREF± without disturbing the input
clock divider. Also, SYSREF± must be
disabled (Register 0x0120, Bits[2:1] =
0x0) when changing this control field.
No positive skew. SYSREF± must be
captured accurately.
1/2 device clock of positive skew.
1 device clocks of positive skew.
1 1/2 device clocks of positive skew.
Reserved
Clock divider phase status (measured in
1/2 clock cycles). Internal clock divider
phase of the captured SYSREF± signal
applied to the phase offset. Only used
when Register 0x010A, Bit 7 = 1.
When Register 0x010A, Bit 7 = 1,
Register 0x010A, Bits[3:2] = 0, and
Register 0x010A, Bits[1:0] = 0, clock
divider SYSREF± offset =
Register 0x0129, Bits[3:0].
Reserved.
Clock delay mode select. Used in
conjunction with Register 0x0111 and
Register 0x0112.
No clock delay.
Fine delay. Only Delay Step 0 to Delay
Step 16 are valid.
Fine delay (lowest jitter). Only Delay
Step 0 to Delay Step 16 are valid.
Fine delay. All 192 delay steps valid.
Fine delay enabled (all 192 delay
steps valid). Super fine delay enabled
(all 128 delay steps valid).
Clock super fine delay adjust. This is
an unsigned control to adjust the
super fine sample clock delay in
0.25 ps steps. These bits are only used
when Register 0x0110, Bits[2:0] = 010
or 110.
0x00 = 0 delay steps.
…
0x08 = 8 delay steps.
…
0x80 = 128 delay steps.
Reset
0x0
Access
R/W
0x0
0x0
R
R
0x0
0x0
R
R/W
0x0
R/W
AD9695
Data Sheet
Address
0x0112
Name
Clock fine delay
(local)
Bits
[7:0]
Bit Name
Set clock fine delay
0x0113
Digital clock
super fine delay
[7:0]
Digital clock super fine
delay adjust
0x0114
Digital clock fine
delay
[7:0]
Set digital clock fine
delay
0x011A
Clock detection
control
[7:5]
[4:3]
Reserved
Clock detection
threshold
0x011B
Clock status
[2:0]
[7:1]
0
Settings
01
11
Reserved
Reserved
Input clock detect
0
1
Rev. 0 | Page 100 of 135
Description
Clock fine delay adjust. This is an
unsigned control to adjust the fine
sample clock skew in 1.725 ps steps.
These bits are only used when
Register 0x0110, Bits[2:0] = 0x2, 0x3,
0x4, or 0x6.
0x00 = 0 delay steps.
…
0x08 = 8 delay steps.
…
0xC0 = 192 delay steps.
Minimum = 0.
Maximum = 192.
Increment = 1.
Unit = delay steps.
Digital clock super fine delay adjust.
This is an unsigned control to adjust
the super fine sample clock delay in
0.25ps steps. These bits are only used
when Register 0x0110, Bits[2:0] = 010
or 110.
0x00 = 0 delay steps.
…
0x08 = 8 delay steps.
…
0x80 = 128 delay steps.
Digital clock fine delay adjust. This is
an unsigned control to adjust the fine
sample clock skew in 1.725 ps steps.
These bits are only used when
Register 0x0110, Bits[2:0] = 0x2, 0x3,
0x4 or 0x6.
0x00 = 0 delay steps.
…
0x08 = 8 delay steps.
…
0xC0 = 192 delay steps.
Minimum = 0.
Maximum = 192.
Increment = 1.
Unit = delay steps.
Reserved.
Clock detection threshold.
Threshold 1 for sample rate ≥ 300 MSPS.
Threshold 2 for sample rate <300 MSPS.
Reserved
Reserved.
Clock detection status.
Input clock not detected.
Input clock detected/locked.
Reset
0xC0
Access
R/W
0x0
R/W
0xC0
R/W
0x0
0x1
R/W
R/W
0x6
0x0
0x0
R/W
R
R
Data Sheet
Address
0x011C
Name
Clock Duty Cycle
Stabilizer 1
(DCS1) control
(local)
AD9695
Bits
[7:2]
1
Bit Name
Reserved
DCS1 enable
Settings
0
1
0
DCS1 power-up
0
1
0x011E
Clock Duty Cycle
Stabilizer 2
(DCS2) control
[7:2]
1
Reserved
DCS2 enable
0
1
0
DCS2 power-up
0
1
0x0120
SYSREF±
Control 1
7
6
Reserved
SYSREF± flag reset
5
4
Reserved
SYSREF± transition
select
0
1
0
1
0x0121
SYSREF±
Control 2
3
CLK± edge select
[2:1]
SYSREF± mode select
0
[7:4]
[3:0]
Reserved
Reserved
SYSREF± N-shot ignore
counter select
0
1
0
1
10
0000
0001
0010
0011
1110
1111
Rev. 0 | Page 101 of 135
Description
Reserved
Clock DCS1 enable.
DCS1 bypassed.
DCS1 enabled.
Clock DCS1 power-up.
DCS1 powered down.
DCS1 powered up.
Reserved.
Clock DCS2 enable.
DCS2 bypassed.
DCS2 enabled.
Clock DCS2 power-up.
DCS2 powered down.
DCS2 powered up.
Reserved.
Normal flag operation.
SYSREF flags held in reset (setup/hold
error flags cleared).
Reserved.
SYSREF is valid on low to high transitions using the selected CLK± edge.
When changing this setting, SYSREF±
mode select must be set to disabled.
SYSREF is valid on high to low
transitions using the selected CLK±
edge. When changing this setting,
SYSREF± mode select must be set to
disabled.
Captured on rising edge of CLK± input.
Captured on falling edge of CLK± input.
Disabled.
Continuous.
N-shot.
Reserved.
Reserved.
Next SYSREF only (do not ignore).
Ignore the first SYSREF± transition.
Ignore the first two SYSREF± transitions.
Ignore the first three SYSREF±
transitions.
…
Ignore the first 14 SYSREF± transitions.
Ignore the first 15 SYSREF± transitions.
Reset
0x0
0x1
Access
R/W
R/W
0x1
R/W
0x0
0x1
R/W
R/W
0x1
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
0x0
R
R
R/W
AD9695
Address
0x0122
Data Sheet
Name
SYSREF±
Control 3
Bits
[7:4]
[3:2]
Bit Name
Reserved
SYSREF± window
negative
Settings
00
01
10
11
[1:0]
SYSREF± window
negative
00
01
10
11
0x0123
SYSREF Control 4
7
[6:0]
Reserved
SYSREF± timestamp
delay
0x0128
SYSREF Status 1
0x0129
SYSREF Status 2
[7:4]
[3:0]
[7:4]
[3:0]
SYSREF± hold status
SYSREF± setup status
Reserved
Clock divider phase
when SYSREF± is
captured
0x012A
SYSREF Status 3
[7:0]
SYSREF± counter,
Bits[7:0] increments
when a SYSREF±is
captured
Rev. 0 | Page 102 of 135
Description
Reserved
Negative skew window (measured in
sample clocks). Number of clock cycles
before the sample clock by which
captured SYSREF± transitions are
ignored.
No negative skew. SYSREF± must be
captured accurately.
One sample clock of negative skew.
Two sample clocks of negative skew.
Three sample clocks of negative skew.
Positive skew window (measured in
sample clocks). Number of clock
cycles before the sample clock by
which captured SYSREF± transitions
are ignored.
No positive skew. SYSREF± must be
captured accurately.
One sample clock of positive skew.
Two sample clocks of positive skew.
Three sample clocks of positive skew.
Reserved.
SYSREF± timestamp delay (in
converter sample clock cycles).
0: 0 sample clock cycle delay.
1: 1 sample clock cycle delay.
…
127: 127 sample clock cycle delay.
SYSREF± hold status.
SYSREF± setup status.
Reserved.
SYSREF divider phase. These bits
represent the phase of the divider
when SYSREF± is captured.
0000 = in phase.
0001 = SYSREF± is ½ cycle delayed
from clock.
0010 = SYSREF± is 1 cycle delayed
from clock.
0011 = 1½ input clock cycles delayed.
0100 = 2 input clock cycles delayed.
0101 = 2½ input clock cycles delayed.
…
1111 = 7½ input clock cycles delayed.
SYSREF± count. Running counter that
increments whenever a SYSREF event
is captured. Reset by Register 0x0120,
Bit 6. Wraps around at 255. Only read
these bits while Register 0x0120,
Bits[2:1] are set to disabled.
Reset
0x0
0x0
Access
R
R/W
0x0
R/W
0x0
0x40
R
R/W
0x0
0x0
0x0
0x0
R
R
R
R
0x0
R
Data Sheet
Address
0x01FF
Name
Chip sync mode
AD9695
Bits
[7:1]
0
Bit Name
Reserved
Synchronization mode
Settings
0x0
0x1
Chip Operating Mode Control Registers
0x0200
Chip mode
[7:6]
5
Reserved
Chip Q ignore
0
1
4
[3:0]
Reserved
Chip application mode
0000
0001
0010
0011
0x0201
Chip decimation
ratio
[7:4]
[3:0]
Reserved
Chip decimation ratio
0000
0001
1000
0010
0101
1001
0011
0110
1010
0111
0100
1101
1011
1110
1111
1100
Rev. 0 | Page 103 of 135
Description
Reserved.
JESD204B synchronization mode. The
SYSREF± signal resets all internal
clock dividers. Use this mode when
synchro-nizing multiple chips as
specified in the JESD204B standard. If
the phase of any of the dividers must
change, the JESD204B link goes down.
Timestamp mode. The SYSREF± signal
does not reset the internal clock
dividers. In this mode, the JESD204B
link and the signal monitor are not
affected by the SYSREF± signal. The
SYSREF± signal simply time stamps a
sample as it passes through the ADC
and is used as a control bit in the
JESD204B output word.
Reset
0x0
0x0
Access
R
R/W
Reserved.
Chip real (I) only selection.
Both real (I) and complex (Q) selected.
Only real (I) selected. Complex (Q) is
ignored.
Reserved.
Full bandwidth mode. (default).
One DDC mode (DDC 0 only).
Two DDC mode (DDC 0 and DDC 1
only).
Four DDC mode (DDC 0, DDC 1, DDC 2,
and DDC 3).
Reserved.
Chip decimation ratio.
Full sample rate (decimate by 1, DDCs
are bypassed).
Decimate by 2.
Decimate by 3.
Decimate by 4.
Decimate by 5.
Decimate by 6.
Decimate by 8.
Decimate by 10.
Decimate by 12.
Decimate by 15.
Decimate by 16.
Decimate by 20.
Decimate by 24.
Decimate by 30.
Decimate by 40.
Decimate by 48.
0x0
0x0
R/W
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
AD9695
Data Sheet
Address
Name
Bits Bit Name
Fast Detect and Signal Monitor Control Registers
0x0245
[7:4] Reserved
Fast detect
control (local)
3
Force FD A/FD B pins
2
Force value of
FD_A/FD_B pins
1
0
Reserved
Enable fast detect
output
0x0247
Fast detect
uppper LSB
(local)
[7:0]
Fast detect upper
threshold
0x0248
Fast detect upper
MSB (local)
[7:5]
[4:0]
Reserved
Fast detect upper
threshold
0x0249
Fast detect low
LSB (local)
[7:0]
Fast detect lower
threshold
0x024A
Fast detect low
MSB (local)
[7:5]
[4:0]
Reserved
Fast detect lower
threshold
0x024B
Fast detect dwell
LSB (local)
[7:0]
Fast detect dwell time
0x024C
FAST detect
dwell MSB (local)
[7:0]
Fast detect dwell time
0x026F
SIgnal monitor
sync control
[7:2]
1
Reserved
Signal monitor next
synchronization mode
Settings
0
1
0
1
0
1
Rev. 0 | Page 104 of 135
Description
Reset
Access
Reserved.
Normal operation of fast detect pin.
Force a value on fast detect pin (see
Bit 2 in this register).
The fast detect output pin for this
channel is set to this value when the
output is forced.
Reserved.
Fast detect disabled.
Fast detect enabled.
LSBs of fast detect upper threshold.
Eight LSBs of the programmable
13-bit upper threshold compared to
the fine ADC magnitude
Reserved.
LSBs of fast detect upper threshold.
Eight LSBs of the programmable 13-bit
upper threshold compared to the fine
ADC magnitude.
LSBs of the fast detect lower threshold.
Eight LSBs of the programmable 13-bit
lower threshold compared to the fine
ADC magnitude.
Reserved.
LSBs of fast detect lower threshold.
Eight LSBs of the programmable 13-bit
lower threshold compared to the fine
ADC magnitude.
LSBs of fast detect dwell time counter
target. This is a load value for a 16-bit
counter that determines how long
the ADC data must remain below the
lower threshold before the FD_x pins
are reset to 0.
LSBs of fast detect dwell time counter
target. This is a load value for a 16-bit
counter that determines how long
the ADC data must remain below the
lower threshold before the FD_x pins
are reset to 0.
Reserved.
Signal monitor next synchronization
mode.
Continuous mode.
Next synchronization mode. Only the
next valid edge of SYSREF± pin
synchronizes the signal monitor
block. Subsequent edges of the
SYSREF± pin are ignored. After the
next SYSREF is found, Register 0x026F,
Bit 0 is cleared. The SYSREF± pin must
be an integer multiple of the signal
monitor period for this function to
operate correctly in continuous mode.
0x0
0x0
R
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
Data Sheet
Address
0x0270
Name
Signal monitor
control (local)
0x0271
Signal Monitor
Period 0 (local)
0x0272
Signal Monitor
Period 1 (local)
Signal Monitor
Period 2 (local)
Signal monitor
status control
(local)
0x0273
0x0274
AD9695
Bits
0
Bit Name
Signal monitor
synchronization mode
Reserved
Peak detector
0
[7:0]
Reserved
Signal monitor period
[7:0]
[7:0]
[7:5]
4
Signal monitor period
[15:8]
Signal monitor period
[23:16]
Reserved
Result update
1
3
[2:0]
Reserved
Result selection
001
[7:0]
0
1
Signal Monitor
Status 0 (local)
[7:0]
Signal monitor result
[7:0]
0x0276
Signal Monitor
Status 1 (local)
Signal Monitor
Status 2 (local)
[7:0]
Signal monitor
status frame
counter (local)
Signal monitor
serial framer
control (local)
[7:0]
Signal monitor result
[15:8]
Reserved
Signal monitor result
[19:16]
Period count result
[7:2]
[1:0]
Reserved
Signal monitor SPORT
over JESD204B enable
SPORT over
JESD204B input
selection (local)
[7:6]
1
Reserved
SPORT over JESD204B
input selection
0x0278
0x0279
0x027A
0
1
[7:2]
1
0x0275
0x0277
Settings
[7:4]
[3:0]
00
11
0
1
0
Reserved
Rev. 0 | Page 105 of 135
Description
Signal monitor synchronization enable.
Synchronization disabled.
If Register 0x026F, Bit 1 = 1, only the
next valid edge of the SYSREF± pin is
used to synchronize the signal
monitor block. Subsequent edges of
the SYSREF± pin are ignored. After
the next SYSREF± isreceived, this bit
is cleared. The SYSREF± input pin
must be enabled to synchronize the
signal monitor blocks.
Reserved.
Peak detector disabled.
Peak detector enabled.
Reserved.
This 24-bit value sets the number of
output clock cycles over which the
signal monitor performs its operation.
Only even values are supported.
Reset
0x0
Access
R/W
0x0
0x0
R
R/W
0x0
0x80
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/WC
0x0
0x1
R
R/W
0x0
R
0x0
R
Reserved.
Signal monitor status result.
0x0
0x0
R
R
Signal monitor frame counter status
bits. The frame counter increments
whenever the period counter expires.
Reserved.
Disabled.
Enabled.
Reserved.
Signal monitor serial framer input
selection. When each individual bit is
a 1, the corresponding signal statistics
information is sent within the frame.
Disabled.
Peak detector data inserted in serial
frame.
Reserved
0x0
R
0x0
0x0
R
R/W
0x0
0x1
R
R/W
0x0
R
Reserved.
Update signal monitor status,
Register 0x0275 to Register 0x0278.
Self clearing.
Reserved.
Peak detector placed on status
readback signals.
Signal monitor status result. This 20-bit
value contains the status result
calculated by the signal monitor block.
Signal monitor status result.
AD9695
Data Sheet
Address
Name
Bits Bit Name
Settings
DDC Function Registers (See the Digital Downconverter (DDC) Section)
0x0300
DDC sync control 7
DDC FTW/POW/MAW/
MBW update mode
0
1
[6:5]
4
Reserved
DDC NCO soft reset
0
1
[3:2]
1
Reserved
DDC next sync
0
1
0
DDC synchronization
mode
0
1
0x0310
DDC 0 control
7
DDC 0 mixer select
0
1
6
DDC 0 gain select
0
1
Rev. 0 | Page 106 of 135
Description
Reset
Access
Select DDC FTW/POW/MAW/MBW
update mode.
Instantaneous/continuous update.
The FTW/POW/MAW/MBW values are
updated immediately.
The FTW/POW/MAW/MBW values are
updated synchronously when the
chip transfer bit (Register 0x000F,
Bit 0) is set.
Reserved.
This bit can be used to synchronize all
the NCOs inside the DDC blocks.
Normal operation.
DDC held in reset.
Reserved.
Continuous mode. The SYSREF±
frequency must be an integer
multiple of the NCO frequencyr for
this function to operate correctly in
continuous mode.
Only the next valid edge of the
SYSREF± pin is used to synchronize the
NCO in the DDC block. Subsequent
edges of the SYSREF± pin are ignored.
After the next SYSREF is found, the
DDC synchronization enable bit
(Register 0x0300, Bit 0) is cleared.
The SYSREF± input pin must be
enabled to synchronize the DDCs.
Synchronization disabled.
If the DDC next sync bit
(Register 0x0300, Bit 1) = 1, only the
next valid edge of the SYSREF± pin is
used to synchronize the NCO in the
DDC block. Subsequent edges of the
SYSREF± pin are ignored. After the next
SYSREF± is received, this bit is cleared.
Real mixer (I and Q inputs must be
from the same real channel).
Complex mixer (I and Q must be from
separate, real and imaginary
quadrature ADC receive channels;
analog demodulator).
Gain can be used to compensate for the
6 dB loss associated with mixing an
input signal down to baseband and
filtering out its negative component.
0 dB gain.
6 dB gain (multiply by 2).
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
Name
AD9695
Bits
[5:4]
Bit Name
DDC 0 intermediate
frequency (IF) mode
3
DDC 0 complex to real
enable
Settings
00
01
10
11
0
1
[2:0]
DDC 0 decimation rate
select
000
001
010
011
100
101
110
111
0x0311
DDC 0 input
select
[7:4]
DDC 0 decimation rate
select
0
10
11
100
Rev. 0 | Page 107 of 135
Description
Variable IF mode.
0 Hz IF mode.
fS/4 Hz IF mode.
Test mode.
Complex (I and Q) outputs contain
valid data.
Real (I) output only. Complex to real
enabled. Uses extra fS /4 mixing to
convert to real.
Decimation filter selection.
HB1 + HB2 filter selection: decimate
by 2 (complex to real enabled), or
decimate by 4 (complex to real
disabled).
HB1 + HB2 + HB3 filter selection:
decimate by 4 (complex to real
enabled), or decimate by 8 (complex
to real disabled).
HB1 + HB2 + HB3 + HB4 filter
selection: decimate by 8 (complex to
real enabled), or decimate by 16
(complex to real disabled).
HB1 filter selection: decimate by 1
(complex to real enabled), or
decimate by 2 (complex to real
disabled).
HB1 + TB2 filter selection: decimate
by 3 (complex to real enabled), or
decimate by 6 (complex to real
disabled).
HB1 + HB2 + TB2 filter selection:
decimate by 6 (complex to real
enabled), or decimate by 12 (complex
to real disabled).
HB1 + HB2 + HB3 + TB2 filter
selection: decimate by 12 (complex to
real enabled), or decimate by 24
(complex to real disabled).
Decimation determined by
Register 0x0311, Bits[7:4].
Only valid when Register 0x0310,
Bits[2:0] = 3'b111.
TB2 + HB4 + HB3 + HB2 + HB1 filter
selection: decimate by 48 (complex to
real disabled), or decimate by 24
(complex to real enabled).
FB2 + HB1 filter selection: decimate by
10 (complex to real disabled) or
decimate by 5 (complex to real
enabled).
FB2 + HB2 + HB1 filter selection:
decimate by 20 (complex to real
disabled), or decimate by 10
(complex to real enabled).
FB2 + HB3 + HB2 + HB1 filter
selection: decimate by 40 (complex to
real disabled), or decimate by 20
(complex to real enabled).
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
R/W
AD9695
Address
Data Sheet
Name
Bits
Bit Name
Settings
111
1000
1001
0x0314
DDC 0 NCO
control
3
2
Reserved
DDC 0 Q input select
1
0
Reserved
DDC 0 I input select
[7:4]
DDC 0 NCO channel
select mode
0
1
0
1
0
1
1000
1010
[3:0]
DDC 0 NCO register
map channel select
0
1
10
11
100
101
110
111
1000
1001
1010
1011
1100
1101
1110
1111
0x0315
DDC 0 phase
control
[7:4]
[3:0]
Reserved
DDC 0 phase update
index
0000
0001
0010
0011
0x0316
DDC 0 Phase
Increment 0
[7:0]
DDC 0 phase
increment [7:0]
Rev. 0 | Page 108 of 135
Description
TB1 filter selection: decimate by 3
(decimate by 1.5 not supported).
FB2 + TB1 filter selection: decimate by
15 (decimate by 7.5 not supported).
HB2 + FB2 + TB1 filter selection:
decimate by 30 (decimate by 15 not
supported).
Reserved.
Channel A.
Channel B.
Reserved.
Channel A.
Channel B.
For edge control, the internal counter
wraps after the Register 0x0314,
Bits[3:0] value is reached.
Use Register 0x0314, Bits[3:0].
2'b0, GPIO B0, GPIO A0.
Increment internal counter on rising
edge of the GPIO A0 pin.
Increment internal counter on rising
edge of the GPIO B0 pin.
NCO channel select register map
control.
Select NCO Channel 0.
Select NCO Channel 1.
Select NCO Channel 2.
Select NCO Channel 3.
Select NCO Channel 4.
Select NCO Channel 5.
Select NCO Channel 6.
Select NCO Channel 7.
Select NCO Channel 8.
Select NCO Channel 9.
Select NCO Channel 10.
Select NCO Channel 11.
Select NCO Channel 12.
Select NCO Channel 13.
Select NCO Channel 14.
Select NCO Channel 15.
Reserved.
Indexes the NCO channel whose
phase and offset gets updated. The
update method is based on the DDC
pyhase update mode, which may be
continuous or require chip transfer.
Update NCO Channel 0.
Update NCO Channel 1.
Update NCO Channel 2.
Update NCO Channel 3.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
Reset
Access
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
Data Sheet
AD9695
Address
0x0317
Name
DDC 0 Phase
Increment 1
Bits
[7:0]
Bit Name
DDC 0 phase
increment [15:8]
0x0318
DDC 0 Phase
Increment 2
[7:0]
DDC 0 phase
increment [23:16]
0x0319
DDC 0 Phase
Increment 3
[7:0]
DDC 0 phase
increment [31:24]
0x031A
DDC 0 Phase
Increment 4
[7:0]
DDC 0 phase
increment [39:32]
0x031B
DDC 0 Phase
Increment 5
[7:0]
DDC 0 phase
increment [47:40]
0x031D
DDC 0 Phase
Offset 0
DDC 0 Phase
Offset 1
DDC 0 Phase
Offset 2
DDC 0 Phase
Offset 3
DDC 0 Phase
Offset 4
DDC 0 Phase
Offset 5
DDC 0 test enable
[7:0]
DDC 0 phase offset
[7:0]
DDC 0 phase offset
[15:8]
DDC 0 phase offset
[23:16]
DDC 0 phase offset
[31:24]
DDC 0 phase offset
[39:32]
DDC 0 phase offset
[47:40]
Reserved
DDC 0 Q output test
mode enable
0x031E
0x031F
0x0320
0x0321
0x0322
0x0327
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:3]
2
Settings
0
1
1
0
0x0330
DDC 1 control
7
Reserved
DDC 0 I output test
mode enable
DDC 1 mixer select
0
1
0
1
6
DDC 1 gain select
0
1
Rev. 0 | Page 109 of 135
Description
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Reserved.
Q samples always use the Test Mode B
block. The test mode is selected using
the channel dependent bits
(Register 0x0550, Bits[3:0]).
Test mode disabled.
Test mode enabled.
Reserved.
I samples always use the Test Mode A
block. The test mode is selected using
the channel dependent bits
(Register 0x0550, Bits[3:0]).
Test mode disabled.
Test mode enabled.
Real mixer (I and Q inputs must be
from the same real channel).
Complex mixer (I and Q must be from
separate, real and imaginary
quadrature ADC receive channels;
analog demodulator).
Gain can be used to compensate for the
6 dB loss associated with mixing an
input signal down to baseband and
filtering out its negative component.
0 dB gain.
6 dB gain (multiply by 2).
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
AD9695
Address
Data Sheet
Name
Bits
[5:4]
Bit Name
DDC 1 intermediate
frequency (IF) mode
3
DDC 1 Complex to real
enable
Settings
00
01
10
11
0
1
[2:0]
DDC 1 decimation rate
select
000
001
010
011
100
101
110
111
Rev. 0 | Page 110 of 135
Description
Variable IF mode.
0 Hz IF mode.
fS/4 Hz IF mode.
Test mode.
Complex (I and Q) outputs contain
valid data.
Real (I) output only. Complex to real
enabled. Uses extra fS/4 mixing to
convert to real.
Decimation filter selection.
HB1 + HB2 filter selection: decimate
by 2 (complex to real enabled), or
decimate by 4 (complex to real
disabled).
HB1 + HB2 + HB3 filter selection:
decimate by 4 (complex to real
enabled), or decimate by 8 (complex
to real disabled).
HB1 + HB2 + HB3 + HB4 filter
selection: decimate by 8 (complex to
real enabled), or decimate by 16
(complex to real disabled).
HB1 filter selection: decimate by 1
(complex to real enabled), or
decimate by 2 (complex to real
disabled).
HB1 + TB2 filter selection: decimate
by 3 (complex to real enabled), or
decimate by 6 (complex to real
disabled).
HB1 + HB2 + TB2 filter selection:
decimate by 6 (complex to real
enabled), or decimate by 12 (complex
to real disabled).
HB1 + HB2 + HB3 + TB2 filter
selection: decimate by 12 (complex to
real enabled), or decimate by 24
(complex to real disabled).
Decimation determined by
Register 0x0331, Bits[7:4].
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x0331
Name
DDC 1 input
select
AD9695
Bits
[7:4]
Bit Name
DDC 1 decimation rate
select
Settings
0
10
11
100
111
1000
1001
0x0334
DDC 1 NCO
control
3
2
Reserved
DDC 1 Q input select
1
0
Reserved
DDC 1 I input select
[7:4]
DDC 1 NCO channel
select mode
0
1
0
1
0
1
1000
1010
[3:0]
DDC 1 NCO register
map channel select
0
1
10
11
100
101
110
111
1000
1001
1010
1011
1100
1101
1110
1111
Rev. 0 | Page 111 of 135
Description
Only valid when Register 0x0310,
Bits[2:0] = 3'b111.
TB2 + HB4 + HB3 + HB2 + HB1 filter
selection: decimate by 48 (complex to
real disabled), or decimate by 24
(complex to real enabled).
FB2 + HB1 filter selection: decimate by
10 (complex to real disabled), or decimate by 5 (complex to real enabled).
FB2 + HB2 + HB1 filter selection:
decimate by 20 (complex to real
disabled), or decimate by 10
(complex to real enabled).
FB2 + HB3 + HB2 + HB1 filter
selection: decimate by 40 (complex to
real disabled), or decimate by 20
(complex to real enabled).
TB1 filter selection: decimate by 3
(decimate by 1.5 not supported).
FB2 + TB1 filter selection: decimate by
15 (decimate by 7.5 not supported).
HB2 + FB2 + TB1 filter select: decim-ate
by 30 (decimate by 15 not supported).
Reserved.
Channel A.
Channel B.
Reserved.
Channel A.
Channel B.
For edge control, the internal counter
wraps after the Register 0x0334,
Bits[3:0] value is reached.
Use Register 0x0314, Bits[3:0].
2'b0, GPIO B0, GPIO A0.
Increment internal counter when
rising edge of the GPIO A0 pin.
Increment internal counter when
rising edge of the GPIO B0 pin.
NCO channel select register map
control.
Select NCO Channel 0.
Select NCO Channel 1.
Select NCO Channel 2.
Select NCO Channel 3.
Select NCO Channel 4.
Select NCO Channel 5.
Select NCO Channel 6.
Select NCO Channel 7.
Select NCO Channel 8.
Select NCO Channel 9.
Select NCO Channel 10.
Select NCO Channel 11.
Select NCO Channel 12.
Select NCO Channel 13.
Select NCO Channel 14.
Select NCO Channel 15.
Reset
0x0
Access
R/W
0x0
0x1
R
R/W
0x0
0x1
R
R/W
0x0
R/W
0x0
R/W
AD9695
Address
0x0335
Data Sheet
Name
DDC 1 phase
control
Bits
[7:4]
[3:0]
Bit Name
Reserved
DDC 1 phase update
index
Settings
0000
0001
0010
0011
0x0336
DDC 1 Phase
Increment 0
[7:0]
DDC 1 phase
increment [7:0]
0x0337
DDC 1 Phase
Increment 1
[7:0]
DDC 1 phase
increment [15:8]
0x0338
DDC 1 Phase
Increment 2
[7:0]
DDC 1 phase
increment [23:16]
0x0339
DDC 1 Phase
Increment 3
[7:0]
DDC 1 phase
increment [31:24]
0x033A
DDC 1 Phase
Increment 4
[7:0]
DDC 1 phase
increment [39:32]
0x033B
DDC 1 Phase
Increment 5
[7:0]
DDC 1 phase
increment [47:40]
0x033D
DDC 1 Phase
Offset 0
DDC 1 Phase
Offset 1
DDC 1 Phase
Offset 2
DDC 1 Phase
Offset 3
DDC 1 Phase
Offset 4
DDC 1 Phase
Offset 5
[7:0]
DDC 1 phase offset
[7:0]
DDC 1 phase offset
[15:8]
DDC 1 phase offset
[23:16]
DDC 1 phase offset
[31:24]
DDC 1 phase offset
[39:32]
DDC 1 phase offset
[47:40]
0x033E
0x033F
0x0340
0x0341
0x0342
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
Rev. 0 | Page 112 of 135
Description
Reserved.
Indexes the NCO channel whose
phase and offset gets updated. The
update method is based on the DDC
pyhase update mode, which may be
continuous or require chip transfer.
Update NCO Channel 0.
Update NCO Channel 1.
Update NCO Channel 2.
Update NCO Channel 3.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Reset
0x0
0x0
Access
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x0347
Name
DDC 1 test enable
AD9695
Bits
[7:3]
2
Bit Name
Reserved
DDC 1 Q output test
mode enable
Settings
0
1
1
0
0x0350
DDC 2 control
7
Reserved
DDC 1 I output test
mode enable
DDC 2 mixer select
0
1
0
1
6
DDC 2 gain select
[5:4]
DDC 2 intermediate
frequency (IF) mode
3
DDC 2 complex to real
enable
0
1
00
01
10
11
0
1
[2:0]
DDC 2 decimation rate
select
000
001
010
011
100
Rev. 0 | Page 113 of 135
Description
Reserved.
Q Samples always use the Test Mode B
block. The test mode is selected using
the channel dependent bits,
Register 0x0550, Bits[3:0].
Test mode disabled.
Test mode enabled.
Reserved.
I samples always use the Test Mode A
block. The test mode is selected using
the channel dependent bits,
Register 0x0550, Bits[3:0] bits.
Test mode disabled.
Test mode enabled.
Real mixer (I and Q inputs must be
from the same real channel).
Complex mixer (I and Q must be from
separate, real and imaginary
quadrature ADC receive channels;
analog demodulator).
Gain can be used to compensate for the
6 dB loss associated with mixing an
input signal down to baseband and
filtering out its negative component.
0 dB gain.
6 dB gain (multiply by 2).
Variable IF mode.
0 Hz IF mode.
fS/4 Hz IF mode.
Test mode.
Complex (I and Q) outputs contain
valid data.
Real (I) output only. Complex to real
enabled. Uses extra fS/4 mixing to
convert to real.
Decimation filter selection.
HB1 + HB2 filter selection: decimate
by 2 (complex to real enabled), or
decimate by 4 (complex to real
disabled).
HB1 + HB2 + HB3 filter selection:
decimate by 4 (complex to real
enabled), or decimate by 8 (complex
to real disabled).
HB1 + HB2 + HB3 + HB4 filter
selection: decimate by 8 (complex to
real enabled), or decimate by 16
(complex to real disabled).
HB1 filter selection: decimate by 1
(complex to real enabled), or decimate by 2 (complex to real disabled).
HB1 + TB2 filter selection: decimate
by 3 (complex to real enabled), or
decimate by 6 (complex to real
disabled).
Reset
0x0
0x0
Access
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
AD9695
Address
Data Sheet
Name
Bits
Bit Name
Settings
101
110
111
0x0351
DDC 2 input
select
[7:4]
DDC 2 decimation rate
select
0
10
11
100
0x0354
DDC 2 NCO
control
3
2
Reserved
DDC 2 Q input select
1
0
Reserved
DDC 2 I input select
[7:4]
DDC 2 NCO channel
select mode
0
1
0
1
0
1
1000
1010
[3:0]
DDC 2 NCO register
map channel select
0
1
10
11
100
101
110
111
1000
1001
1010
1011
Rev. 0 | Page 114 of 135
Description
HB1 + HB2 + TB2 filter selection:
decimate by 6 (complex to real
enabled), or decimate by 12 (complex
to real disabled).
HB1 + HB2 + HB3 + TB2 filter
selection: decimate by 12 (complex to
real enabled), or decimate by 24
(complex to real disabled).
Decimation determined by
Register 0x0351, Bits[7:4].
Only valid when Register 0x0310,
Bits[2:0] = 3'b111.
TB2 + HB4 + HB3 + HB2 + HB1 filter
selection: decimate by 48 (complex to
real disabled), or decimate by 24
(complex to real enabled).
FB2 + HB1 filter selection: decimate
by 10 (complex to real disabled), or
decimate by 5 (complex to real
enabled).
FB2 + HB2 + HB1 filter selection:
decimate by 20 (complex to real
disabled), or decimate by 10
(complex to real enabled).
FB2 + HB3 + HB2 + HB1 filter
selection: decimate by 40 (complex to
real disabled), or decimate by 20
(complex to real enabled).
Reserved.
Channel A.
Channel B.
Reserved.
Channel A.
Channel B.
For edge control, the internal counter
wraps after the Register 0x0354,
Bits[3:0] value is reached.
Use Register 0x0314, Bits[3:0].
2'b0, GPIO B0, GPIO A0.
Increment internal counter when
rising edge of the GPIO A0 pin.
Increment internal counter when
rising edge of the GPIO B0 pin.
NCO channel select register map
control.
Select NCO Channel 0.
Select NCO Channel 1.
Select NCO Channel 2.
Select NCO Channel 3.
Select NCO Channel 4.
Select NCO Channel 5.
Select NCO Channel 6.
Select NCO Channel 7.
Select NCO Channel 8.
Select NCO Channel 9.
Select NCO Channel 10.
Select NCO Channel 11.
Reset
Access
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
Data Sheet
AD9695
Address
Name
Bits
Bit Name
Settings
1100
1101
1110
1111
0x0355
DDC 2 phase
control
[7:4]
[3:0]
Reserved
DDC 2 phase update
index
0000
0001
0010
0011
0x0356
DDC 2 Phase
Increment 0
[7:0]
DDC 2 phase
increment [7:0]
0x0357
DDC 2 Phase
Increment 1
[7:0]
DDC 2 phase
increment [15:8]
0x0358
DDC 2 Phase
Increment 2
[7:0]
DDC 2 phase
increment [23:16]
0x0359
DDC 2 Phase
Increment 3
[7:0]
DDC 2 phase
increment [31:24]
0x035A
DDC 2 Phase
Increment 4
[7:0]
DDC 2 phase
increment [39:32]
0x035B
DDC 2 Phase
Increment 5
[7:0]
DDC 2 phase
increment [47:40]
0x035D
DDC 2 Phase
Offset 0
DDC 2 Phase
Offset 1
DDC 2 Phase
Offset 2
DDC 2 Phase
Offset 3
DDC 2 Phase
Offset 4
DDC 2 Phase
Offset 5
[7:0]
DDC 2 phase offset
[7:0]
DDC 2 phase offset
[15:8]
DDC 2 phase offset
[23:16]
DDC 2 phase offset
[31:24]
DDC 2 phase offset
[39:32]
DDC 2 phase offset
[47:40]
0x035E
0x035F
0x0360
0x0361
0x0362
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
Rev. 0 | Page 115 of 135
Description
Select NCO Channel 12.
Select NCO Channel 13.
Select NCO Channel 14.
Select NCO Channel 15.
Reserved.
Indexes the NCO channel whose
phase and offset gets updated. The
update method is based on the DDC
pyhase update mode, which may be
continuous or require chip transfer.
Update NCO Channel 0.
Update NCO Channel 1.
Update NCO Channel 2.
Update NCO Channel 3.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Reset
Access
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
AD9695
Address
0x0367
Data Sheet
Name
DDC 2 test enable
Bits
[7:3]
2
Bit Name
Reserved
DDC 2 Q output test
mode enable
Settings
0
1
1
0
0x0370
DDC 3 control
7
Reserved
DDC 2 I output test
mode enable
DDC 3 mixer select
0
1
0
1
6
DDC 3 gain select
[5:4]
DDC 3 intermediate
frequency (IF) mode
3
DDC 3 complex to real
enable
0
1
00
01
10
11
0
1
[2:0]
DDC 3 decimation rate
select
000
001
010
011
100
Rev. 0 | Page 116 of 135
Description
Reserved.
Q samples always use the Test Mode B
block. The test mode is selected using
the channel dependent bits,
Register 0x0550, Bits[3:0].
Test mode disabled.
Test mode enabled.
Reserved.
I samples always use the Test Mode A
block. The test mode is selected using
the channel dependent bits,
Register 0x0550, Bits[3:0].
Test mode disabled.
Test mode enabled.
Real mixer (I and Q inputs must be
from the same real channel).
Complex mixer (I and Q must be from
separate, real and imaginary
quadrature ADC receive channels;
analog demodulator).
Gain can be used to compensate for
the 6 dB loss associated with mixing
an input signal down to baseband
and filtering out its negative
component.
0 dB gain.
6 dB gain (multiply by 2).
Variable IF mode.
0 Hz IF mode.
fS/4 Hz IF mode.
Test mode.
Complex (I and Q) outputs contain
valid data.
Real (I) output only. Complex to real
enabled. Uses extra fS/4 mixing to
convert to real.
Decimation filter selection.
HB1 + HB2 filter selection: decimate
by 2 (complex to real enabled), or
decimate by 4 (complex to real
disabled).
HB1 + HB2 + HB3 filter selection:
decimate by 4 (complex to real
enabled), or decimate by 8 (complex
to real disabled).
HB1 + HB2 + HB3 + HB4 filter
selection: decimate by 8 (complex to
real enabled), or decimate by 16
(complex to real disabled).
HB1 filter selection: decimate by 1
(complex to real enabled), or
decimate by 2 (complex to real
disabled).
HB1 + TB2 filter selection: decimate
by 3 (complex to real enabled), or
decimate by 6 (complex to real
disabled).
Reset
0x0
0x0
Access
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
Name
AD9695
Bits
Bit Name
Settings
101
110
111
0x0371
DDC 3 input
select
[7:4]
DDC 3 decimation rate
select
0
10
11
100
0x0374
DDC 3 NCO
control
3
2
Reserved
DDC 3 Q input select
1
0
Reserved
DDC 3 I input select
[7:4]
DDC 3 NCO channel
select mode
0
1
0
1
0
1
1000
1010
[3:0]
DDC 3 NCO register
map channel select
0
1
10
11
100
101
110
111
1000
1001
1010
1011
Rev. 0 | Page 117 of 135
Description
HB1 + HB2 + TB2 filter selection:
decimate by 6 (complex to real
enabled), or decimate by 12 (complex
to real disabled).
HB1 + HB2 + HB3 + TB2 filter
selection: decimate by 12 (complex to
real enabled), or decimate by 24
(complex to real disabled).
Decimation determined by
Register 0x0371, Bits[7:4].
Only valid when Register 0x0310,
Bits[2:0] = 3'b111.
TB2 + HB4 + HB3 + HB2 + HB1 filter
selection: decimate by 48 (complex to
real disabled), or decimate by 24
(complex to real enabled).
FB2 + HB1 filter selection: decimate
by 10 (complex to real disabled), or
decimate by 5 (complex to real
enabled)
FB2 + HB2 + HB1 filter selection:
decimate by 20 (complex to real
disabled), or decimate by 10
(complex to real enabled)
FB2 + HB3 + HB2 + HB1 filter
selection: decimate by 40 (complex to
real disabled), or decimate by 20
(complex to real enabled)
Reserved.
Channel A.
Channel B.
Reserved.
Channel A.
Channel B.
For edge control, the internal counter
wraps after the Register 0x0374,
Bits[3:0] value is reached.
Use Register 0x0314, Bits[3:0].
2'b0, GPIO B0, GPIO A0.
Increment internal counter when
rising edge of the GPIO A0 pin.
Increment internal counter when
rising edge of the GPIO B0 pin.
NCO channel select register map
control.
Select NCO Channel 0.
Select NCO Channel 1.
Select NCO Channel 2.
Select NCO Channel 3.
Select NCO Channel 4.
Select NCO Channel 5.
Select NCO Channel 6.
Select NCO Channel 7.
Select NCO Channel 8.
Select NCO Channel 9.
Select NCO Channel 10.
Select NCO Channel 11.
Reset
Access
0x0
R/W
0x0
0x1
R
R/W
0x0
0x1
R
R/W
0x0
R/W
0x0
R/W
AD9695
Data Sheet
Address
Name
Bits
Bit Name
Settings
1100
1101
1110
1111
0x0375
DDC 3 phase
control
[7:4]
[3:0]
Reserved
DDC 3 phase update
index
0000
0001
0010
0011
0x0376
DDC 3 Phase
Increment 0
[7:0]
DDC 3 phase
increment [7:0]
0x0377
DDC 3 Phase
Increment 1
[7:0]
DDC 3 phase
increment [15:8]
0x0378
DDC 3 Phase
Increment 2
[7:0]
DDC 3 phase
increment [23:16]
0x0379
DDC 3 Phase
Increment 3
[7:0]
DDC 3 phase
increment [31:24]
0x037A
DDC 3 Phase
Increment 4
[7:0]
DDC 3 phase
increment [39:32]
0x037B
DDC 3 Phase
Increment 5
[7:0]
DDC 3 phase
increment [47:40]
0x037D
DDC 3 Phase
Offset 0
DDC 3 Phase
Offset 1
DDC 3 Phase
Offset 2
DDC 3 Phase
Offset 3
DDC 3 Phase
Offset 4
DDC 3 Phase
Offset 5
[7:0]
DDC 3 phase offset [7:0]
[7:0]
DDC 3 phase offset
[15:8]
DDC 3 phase offset
[23:16]
DDC 3 phase offset
[31:24]
DDC 3 phase offset
[39:32]
DDC 3 phase offset
[47:40]
0x037E
0x037F
0x0380
0x0381
0x0382
[7:0]
[7:0]
[7:0]
[7:0]
Rev. 0 | Page 118 of 135
Description
Select NCO Channel 12.
Select NCO Channel 13.
Select NCO Channel 14.
Select NCO Channel 15.
Reserved.
Indexes the NCO channel whose
phase and offset gets updated. The
update method is based on the DDC
phase update mode, which may be
continuous or require chip transfer.
Update NCO Channel 0.
Update NCO Channel 1.
Update NCO Channel 2.
Update NCO Channel 3.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
FTW. Twos complement phase
increment value for the NCO.
Complex mixing frequency =
(DDC_PHASE_INC × fS)/248.
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Twos complement phase offset value
for the NCO (POW).
Reset
Access
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x0387
Name
DDC 3 test enable
AD9695
Bits
[7:3]
2
Bit Name
Reserved
DDC 3 Q output test
mode enable
Settings
0
1
1
0
Reserved
DDC 3 I output test
mode enable
0
1
0x0390
0x0391
0x0392
0x0393
0x0394
0x0395
0x0398
0x0399
0x039A
0x039B
0x039C
0x039D
0x03A0
0x5C00x03A1
DDC 0 Phase
Increment
Fractional A0
DDC 0 Phase
Increment
Fractional A1
DDC 0 Phase
Increment
Fractional A2
DDC 0 Phase
Increment
Fractional A3
DDC 0 Phase
Increment
Fractional A4
DDC 0 Phase
Increment
Fractional A5
DDC 0 Phase
Increment
Fractional B0
DDC 0 Phase
Increment
Fractional B1
DDC 0 Phase
Increment
Fractional B2
DDC 0 Phase
Increment
Fractional B3
DDC 0 Phase
Increment
Fractional B4
DDC 0 Phase
Increment
Fractional B5
DDC 1 Phase
Increment
Fractional A0
DDC 1 Phase
Increment
Fractional A1
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
DDC 0 Phase
Increment Fractional A
[7:0]
DDC 0 Phase
Increment Fractional A
[15:8]
DDC 0 Phase
Increment Fractional A
[23:16]
DDC 0 Phase
Increment Fractional A
[31:24]
DDC 0 Phase
Increment Fractional A
[39:32]
DDC 0 Phase
Increment Fractional A
[47:40]
DDC 0 Phase
Increment Fractional B
[7:0]
DDC 0 Phase
Increment Fractional B
[15:8]
DDC 0 Phase
Increment Fractional B
[23:16]
DDC 0 Phase
Increment Fractional B
[31:24]
DDC 0 Phase
Increment Fractional B
[39:32]
DDC 0 Phase
Increment Fractional B
[47:40]
DDC 1 Phase
Increment Fractional A
[7:0]
DDC 1 Phase
Increment Fractional A
[15:8]
Rev. 0 | Page 119 of 135
Description
Reserved.
Q samples always use the Test Mode B
block. The test mode is selected using
the channel dependent bit,
Register 0x0550, Bits[3:0].
Test mode disabled.
Test mode enabled.
Reserved.
I samples always use the Test Mode A
block. The test mode is selected using
the channel dependent bits,
Register 0x0550, Bits[3:0].
Test mode disabled.
Test mode enabled.
Numerator correction term for the
modulus phase accumulator (MAW).
Reset
0x0
0x0
Access
R
R/W
0x0
0x0
R
R/W
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Denominator correction term for the
modulus phase accumulator (MBW).
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
Numerator correction term for the
modulus phase accumulator (MAW).
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
AD9695
Address
0x03A2
0x03A3
0x03A4
0x03A5
0x03A8
0x03A9
0x03AA
0x03AB
0x03AC
0x03AD
0x03B0
0x03B1
0x03B2
0x03B3
0x03B4
0x03B5
0x03B8
0x03B9
0x03BA
Data Sheet
Name
DDC 1 Phase
Increment
Fractional A2
DDC 1 Phase
Increment
Fractional A3
DDC 1 Phase
Increment
Fractional A4
DDC 1 Phase
Increment
Fractional A5
DDC 1 Phase
Increment
Fractional B0
DDC 1 Phase
Increment
Fractional B1
DDC 1 Phase
Increment
Fractional B2
DDC 1 Phase
Increment
Fractional B3
DDC 1 Phase
Increment
Fractional B4
DDC 1 Phase
Increment
Fractional B5
DDC 2 Phase
Increment
Fractional A0
DDC 2 Phase
Increment
Fractional A1
DDC 2 Phase
Increment
Fractional A2
DDC 2 Phase
Increment
Fractional A3
DDC 2 Phase
Increment
Fractional A4
DDC 2 Phase
Increment
Fractional A5
DDC 2 Phase
Increment
Fractional B0
DDC 2 Phase
Increment
Fractional B1
DDC 2 Phase
Increment
Fractional B2
Bits
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
Bit Name
DDC 1 Phase
Increment Fractional A
[23:16]
DDC 1 Phase
Increment Fractional A
[31:24]
DDC 1 Phase
Increment Fractional A
[39:32]
DDC 1 Phase
Increment Fractional A
[47:40]
DDC 1 Phase Increment
Fractional B [7:0]
Settings
Description
Numerator correction term for the
MAW.
Reset
0x0
Access
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Numerator correction term for the
MAW.
0x0
R/W
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 1 Phase Increment
Fractional B [15:8]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 1 Phase Increment
Fractional B [23:16]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 1 Phase Increment
Fractional B [31:24]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 1 Phase Increment
Fractional B [39:32]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 1 Phase Increment
Fractional B [47:40]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [7:0]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [15:8]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [23:16]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [31:24]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [39:32]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional A [47:40]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional B [7:0]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional B [15:8]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional B [23:16]
Denominator correction term for the
MBW.
0x0
R/W
Rev. 0 | Page 120 of 135
Data Sheet
Address
0x03BB
0x03BC
0x03BD
0x03C0
0x03C1
0x03C2
0x03C3
0x03C4
0x03C5
0x03C8
0x03C9
0x03CA
0x03CB
0x03CC
Name
DDC 2 Phase
Increment
Fractional B3
DDC 2 Phase
Increment
Fractional B4
DDC 2 Phase
Increment
Fractional B5
DDC 3 Phase
Increment
Fractional A0
DDC 3 Phase
Increment
Fractional A1
DDC 3 Phase
Increment
Fractional A2
DDC 3 Phase
Increment
Fractional A3
DDC 3 Phase
Increment
Fractional A4
DDC 3 Phase
Increment
Fractional A5
DDC 3 Phase
Increment
Fractional B0
DDC 3 Phase
Increment
Fractional B1
DDC 3 Phase
Increment
Fractional B2
DDC 3 Phase
Increment
Fractional B3
DDC 3 Phase
Increment
Fractional B4
AD9695
Bits
[7:0]
Bit Name
DDC 2 Phase Increment
Fractional B [31:24]
[7:0]
Settings
Description
Denominator correction term for the
MBW.
Reset
0x0
Access
R/W
DDC 2 Phase Increment
Fractional B [39:32]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 2 Phase Increment
Fractional B [47:40]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [7:0]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [15:8]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [23:16]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [31:24]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [39:32]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional A [47:40]
Numerator correction term for the
MAW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional B [7:0]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional B [15:8]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional B [23:16]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional B [31:24]
Denominator correction term for the
MBW.
0x0
R/W
[7:0]
DDC 3 Phase Increment
Fractional B [39:32]
Denominator correction term for the
MBW.
0x0
R/W
Rev. 0 | Page 121 of 135
AD9695
Data Sheet
Address
0x03CD
Name
Bits
[7:0]
DDC 3 Phase
Increment
Fractional B5
Digital Outputs and Test Mode Registers
0x0550
7
ADC test mode
control (local)
Bit Name
DDC 3 Phase Increment
Fractional B [47:40]
Settings
User pattern selection
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R/W
R/W
0x0
R/W
User Pattern 1 [7:0]
User Pattern 1 [15:8]
[7:0]
User Pattern 2 [7:0]
User Test Pattern 2 LSB.
6
5
Reserved
Reset psuedorandom
long generator
0
1
4
Reset psuedorandom
short generator
0
1
[3:0]
Test mode selection
0000
0001
0010
0011
0100
0101
0110
0111
1000
0x0553
Access
R/W
[7:0]
[7:0]
1
User Pattern 1 LSB
User Pattern 1
MSB
User Pattern 2 LSB
Reset
0x0
Test mode user pattern selection.
These bits are only used when
TMODE_GEN_SEL is in user input
mode (TMODE_GEN_SEL = 1000).
Otherwise, they are ignored. User
Pattern 1 is found in the USR_PAT_
1_MSB and USR_PAT_1_LSB registers.
User Pattern 2 is found in the USR_
PAT_2_MSB and USR_PAT_2_LSB
registers, and so on.
Continuous/repeat pattern. Place
each user pattern (User Pattern 1
through User Pattern 4) on the output
for 1 clock cycle and then repeat.
(Output User Pattern 1, User Pattern 2,
User Pattern 3, User Pattern 4, User
Pattern 1, User Pattern 2, User Pattern 3,
User Pattern 4, User Pattern 1, User
Pattern 2, User Pattern 3, User Pattern 4,
and so on).
Single Pattern. Place each User
Pattern (User Pattern 1 through User
Pattern 4) on the output for 1 clock
cycle and then output all zeros.
(Output User Pattern 1 through User
Pattern 4, then output all zeros).
Reserved.
Test mode long psuedorandom
number test generator reset.
Long psuedorandom enabled.
Long psuedorandom held in reset.
Test mode short psuedorandom
number Test generator reset.
Short psuedorandom enabled.
Short psuedorandom held in reset.
Test mode generation selection.
Off; normal operation.
Midscale short.
Positive full scale.
Negative full scale.
Alternating checker board.
Psuedorandom sequence, long.
Psuedorandom sequence, short.
1/0 word toggle.
User pattern test mode (used with
TMODE_USR_PAT_SEL and the User
Pattern 1 through User Pattern 4
registers).
1111: ramp output.
User Test Pattern 1 LSB.
User Test Pattern 1 LSB.
0
0x0551
0x0552
Description
Denominator correction term for the
MBW.
Rev. 0 | Page 122 of 135
Data Sheet
Address
0x0554
0x0555
0x0556
0x0557
0x0558
0x0559
0x055A
0x0561
0x0562
Name
User Pattern 2
MSB
User Pattern 3 LSB
User Pattern 3
MSB
User Pattern 4 LSB
User Pattern 4
MSB
Output Mode
Control 1
AD9695
Description
User Test Pattern 2 LSB.
Reset
0x0
Access
R/W
User Pattern 3 [7:0]
User Pattern 3 [15:8]
User Test Pattern 3 LSB.
User Test Pattern 3 LSB.
0x0
0x0
R/W
R/W
[7:0]
[7:0]
User Pattern 4 [7:0]
User Pattern 4 [15:8]
User Test Pattern 4 LSB.
User Test Pattern 4 LSB.
0x0
0x0
R/W
R/W
[7:4]
Converter control Bit 1
selection
0x0
R/W
[3:0]
Converter control Bit 0
selection
0x0
R/W
Output Mode
Control 2
[7:4]
[3:0]
Reserved
Converter control Bit 2
selection
0x0
0x1
R
R/W
Output sample
mode
[7:3]
2
Reserved
Sample invert
0x0
0x0
R/W
R/W
[1:0]
Data format select
0x1
R/W
[7:0]
Data format overrange
clear
Tie low (1'b0).
Overrange bit.
Signal monitor bit.
Fast detect (FD) bit.
SYSREF±.
Tie low (1'b0).
Overrange bit.
Signal monitor bit.
Fast detect (FD) bit.
SYSREF±.
Reserved.
Tie low (1'b0).
Overrange bit.
Signal monitor bit.
Fast detect (FD) bit.
SYSREF±.
Reserved.
ADC sample data is not inverted.
ADC sample data is inverted.
Offset binary.
Twos complement (default).
Overrange clear bits (one bit for each
virtual converter).
Overrange bit enabled.
Overrange bit cleared. Writing a 1 to
the overrange clear bit clears the
corresponding overrange sticky bit.
Overrange sticky bit status (one bit
for each virtual converter).
No overrange occurred
Overrange occurred. Writing a 1 to
the overrange clear bit clears the
corresponding overrange sticky bit.
Reserved.
Normal channel ordering.
Channel swap enabled. Depending
on the application mode selected in
Register 0x0200, enabling the channel
swap bit (Register 0x0564, Bit 0)
swaps the A/B or I/Q converters.
Lane rate = 13.5 Gbps to 16 Gbps.
Lane rate = 6.75 Gbps to 13.5 Gbps.
Lane rate = 3.375 Gbps to 6.75 Gbps.
Lane rate = 1.6875 Gbps to 3.375 Gbps.
Reserved.
0x0
R/W
0x0
R
0x0
0x0
R
R/W
0x3
R/W
0x0
R
Output
overrange clear
Bits
[7:0]
Bit Name
User Pattern 2 [15:8]
[7:0]
[7:0]
Settings
0000
0001
0010
0011
0101
0000
0001
0010
0011
0101
0000
0001
0010
0011
0101
0
1
00
01
0
1
0x0563
Output
overrange status
[7:0]
Data format overrange
0
1
0x0564
0x056E
Output channel
select
[7:1]
0
Reserved
Converter channel
swap control
PLL control
[7:4]
JESD204B lane rate
control
[3:0]
Reserved
0
1
0011
0000
0001
0101
Rev. 0 | Page 123 of 135
AD9695
Address
0x056F
Data Sheet
Name
PLL status
Bits
7
Bit Name
PLL lock status
[6:4]
3
Reserved
PLL loss of lock
Settings
0
1
1
0x0571
JESD204B Link
Control 1
[2:0]
7
Reserved
Standby mode
0
1
6
Tail bit (t) PN
5
Long transport layer
test
4
Lane synchronization
[3:2]
ILAS sequence mode
0
1
0
1
0
1
00
01
11
1
FACI
0
1
0
Link control
0
1
0x0572
JESD204B Link
Control 2
[7:6]
SYNCINB± pin control
5
SYNCINB± pin invert
4
SYNCINB± pin type
3
2
Reserved
8-bit/10-bit bypass
1
8-bit/10-bit bit invert
0
Reserved
00
10
11
0
1
0
1
0
1
0
1
Rev. 0 | Page 124 of 135
Description
Not locked.
Locked.
Reserved.
Loss of lock sticky bit.
Indicates a loss of lock occurred at
some time; cleared by setting
Register 0x0571, Bit 0.
Reserved
Standby mode forces zeros for all
converter samples.
Standby mode forces code group
synchronization (K28.5 characters).
Disable.
Enable.
JESD204B test samples disabled.
JESD204B test samples enabled; long
transport layer test sample sequence
(as specified in JESD204B Section
5.1.6.3) sent on all link lanes.
Disable FACI uses /K28.7/.
Enable FACI uses /K28.3/ and /K28.7/.
Initial lane alignment sequence
disabled (JESD204B, Section 5.3.3.5).
Initial lane alignment sequence
enabled (JESD204B, Section 5.3.3.5).
Initial lane alignment sequence always
on test mode (JESD204B data link layer
test mode) where repeated lane
alignment sequence (as specified in
JESD204B, Section 5.3.3.8.2) sent on
all lanes.
Frame alignment character insertion
enabled (JESD204B, Section 5.3.3.4).
Frame alignment character insertion
disabled; for debug only (JESD204B,
Section 5.3.3.4).
JESD204B serial transmit link enabled.
Transmission of the /K28.5/ characters
for code group synchronization is
controlled by the SYNCINB± pin.
JESD204B serial transmit link powered
down (held in reset and clock gated).
Normal mode.
Ignore SYNCINB± (force CGS).
Ignore SYNCINB± (force ILAS/user data).
SYNCINB± pin not inverted.
SYNCINB± pin inverted.
LVDS differential pair SYNC input.
CMOS single-ended SYNC input.
Reserved.
8-bit/10-bit enabled.
8-bit/10-bit bypassed (the most
significant 2 bits are 0).
Normal.
Invert a, b, c, d, e, f, g, h, I, and j symbols.
Reserved.
Reset
0x0
Access
R
0x0
R
0x0
R/W
0x0
R/W
0x0
R/W
0x1
R/W
0x1
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x0573
Name
JESD204B Link
Contro 3
AD9695
Bits
[7:6]
Bit Name
Checksum mode
Settings
00
01
10
[5:4]
0x0574
JESD204B Link
Control 4
Test injection point
[3:0]
JESD204B test mode
patterns
[7:4]
ILAS delay
11
0
1
10
0
1
10
11
100
101
110
111
1000
1110
1111
0
1
10
11
100
101
110
111
1000
1001
1010
1011
1100
1101
1110
1111
3
Reserved
Rev. 0 | Page 125 of 135
Description
Checksum is the sum of all 8-bit
registers in the link configuration table.
Checksum is the sum of all individual
link configuration fields (LSB aligned).
Checksum is disabled (set to zero).
For test purposes only.
Unused.
N' sample input.
10 bit data at 8-bit/10-bit output (for
PHY testing)
8-bit data at scrambler input.
Normal operation (test mode disabled).
Alternating checkerboard.
1/0 word toggle.
31-bit PN sequence (x31 + x28 + 1).
23-bit PN sequence (x23 + x18 + 1).
15-bit PN sequence (x15 + x14 + 1).
9-bit PN sequence (x9 + x5 + 1).
7-bit PN sequence (x7 + x6 + 1).
Ramp output.
Continuous/repeat user test.
Single user test.
Transmit ILAS on first LMFC after
SYNCINB± deasserted.
Transmit ILAS on second LMFC after
SYNCINB± deasserted.
Transmit ILAS on third LMFC after
SYNCINB± deasserted.
Transmit ILAS on fourth LMFC after
SYNCINB± deasserted.
Transmit ILAS on fifth LMFC after
SYNCINB± deasserted.
Transmit ILAS on sixth LMFC after
SYNCINB± deasserted.
Transmit ILAS on seventh LMFC after
SYNCINB± deasserted.
Transmit ILAS on eighth LMFC after
SYNCINB± deasserted.
Transmit ILAS on ninth LMFC after
SYNCINB± deasserted.
Transmit ILAS on tenth LMFC after
SYNCINB± deasserted.
Transmit ILAS on eleventh LMFC after
SYNCINB± deasserted.
Transmit ILAS on twelfth LMFC after
SYNCINB± deasserted.
Transmit ILAS on thirteenth LMFC
after SYNCINB± deasserted.
Transmit ILAS on fourteenth LMFC
after SYNCINB± deasserted.
Transmit ILAS on fifteenth LMFC after
SYNCINB± deasserted.
Transmit ILAS on sixteenth LMFC after
SYNCINB± deasserted.
Reserved.
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R
AD9695
Address
Data Sheet
Name
Bits
[2:0]
Bit Name
Link layer test mode
Settings
000
001
010
011
100
101
110
111
Description
Normal operation (link layer test
mode disabled).
Continuous sequence of /D21.5/
characters.
Reserved.
Reserved.
Modified RPAT test sequence.
JSPAT test sequence.
JTSPAT test sequence.
Reserved.
Reserved.
Local multiframe clock (LMFC) phase
offset value. Reset value for the LMFC
phase counter when SYSREF± is
asserted. Used for deterministic delay
applications.
JESD204B serial DID number.
Reset
0x0
Access
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0578
JESD204B LMFC
offset
[7:5]
[4:0]
Reserved
LMFC phase offset
value
0x0580
JESD204B device
identification
(DID) configuration
JESD204B bank
identification
(BID) configuration
JESD204B Lane
Identification 0
(LID0) configuration
JESD204B LID1
configuration
[7:0]
JESD204B Tx DID value
[7:4]
[3:0]
Reserved
JESD204B Tx BID value
Reserved.
JESD204B serial BID number
(extension to DID).
0x0
0x0
R
R/W
[7:5]
[4:0]
Reserved
Lane 0 LID value
Reserved.
JESD204B serial LID number for Lane 0.
0x0
0x0
R
R/W
[7:5]
[4:0]
[7:5]
[4:0]
[7:5]
[4:0]
7
Reserved
Lane 1 LID value
Reserved
Lane 2 LID value
Reserved
Lane 3 LID value
JESD204B scrambling
(SCR)
0x0
0x1
0x0
0x2
0x0
0x3
0x1
R
R/W
R
R/W
R
R/W
R/W
[6:5]
[4:0]
Reserved
JESD204B lanes (L)
0x0
0x3
R
R/W
JESD204B link
number of octets
per frames (F)
[7:0]
JESD204B F
configuration
0x0
R/W
JESD204B link
number of
frames per
multiframe (K)
[7:5]
[4:0]
Reserved.
JESD204B serial LID number for Lane 1.
Reserved.
JESD204B serial LID number for Lane 2.
Reserved.
JESD204B serial LID number for Lane 3.
JESD204B scrambler disabled (SCR = 0).
JESD204B scrambler enabled (SCR = 1).
Reserved.
One lane per link (L = 1).
Two lanes per link (L = 2).
Four lanes per link (L = 4).
JESD204B number of octets per frame
(F = JESD204B_F_CONFIG + 1).
F = 1.
F = 2.
F = 3.
F = 4.
F = 6.
F = 8.
F = 16.
Reserved.
JESD204B number of frames per multiframe (K = JESD204B_K_CONFIG + 1).
Only values where F × K which are
divisible by 4 can be used.
0x0
0x1F
R
R/W
0x0581
0x0583
0x0584
0x0585
JESD204B LID2
configuration
0x0586
JESD204B LID3
configuration
0x058B
JESD204B
scrambling and
number of lanes
(L) configuration
0x058C
0x058D
0
1
0x0
0x1
0x3
0
1
10
11
101
111
1111
Reserved
JESD204B K
configuration
Rev. 0 | Page 126 of 135
Data Sheet
Address
0x058E
Name
JESD204B link
number of
converters (M)
AD9695
Bits
[7:0]
Bit Name
JESD204B M
configuration
Settings
0
1
11
111
0x058F
0x0590
JESD204B
number Of
control bits (CS)
and ADC
resolution (N)
JESD204B SCV
NP configuration
[7:6]
Number of control bits
(CS) per sample
0
1
10
11
5
[4:0]
Reserved
ADC converter
resolution (N)
[7:5]
Subclass support
[4:0]
ADC number of bits
per sample (N')
Reserved
Samples per converter
frame cycle (S)
HD value
0x0591
JESD204B JV S
configuration
[7:5]
[4:0]
0x0592
JESD204B HD CF
configuration
7
[6:5]
[4:0]
00110
00111
01000
01001
01010
01011
01100
01101
01110
01111
000
001
00111
01111
0
1
Reserved
Control words per
frame clock cycle per
link (CF)
Checksum 0 value for
SERDOUT0±
0x05A0
JESD204B
Checksum 0
configuration
[7:0]
0x05A1
JESD204B
Checksum 1
configuration
[7:0]
Checksum 1 value for
SERDOUT1±
0x05A2
JESD204B
Checksum 2
configuration
[7:0]
Checksum 2 value for
SERDOUT2±
0x05A3
JESD204B
Checksum 3
configuration
[7:0]
Checksum 3 value for
SERDOUT3±
Rev. 0 | Page 127 of 135
Description
JESD204B number of converters per
link/device (M = JESD204B_M_CFG).
Link connected to one virtual
converter (M = 1).
Link connected to two virtual
converters (M = 2).
Link connected to four virtual
converters (M = 4).
Link connected to eight virtual
converters (M = 8).
No control bits (CS = 0)
1 control bit (CS = 1), Control Bit 2 only.
2 control bits (CS = 2), Control Bit 2
and Control Bit 1only.
3 control bits (CS = 3), all control bits
(2, 1,and 0).
Reserved.
N = 7-bit resolution.
N = 8-bit resolution.
N = 9-bit resolution.
N = 10-bit resolution.
N = 11-bit resolution.
N = 12-bit resolution.
N = 13-bit resolution.
N = 14-bit resolution.
N = 15-bit resolution.
N = 16-bit resolution.
Subclass 0.
Subclass 1.
N' = 8.
N' = 16.
Reserved.
Samples per converter frame cycle (S =
Register 0x0591, Bits[4:0]+1)
High density format disabled.
High density format enabled.
Reserved.
Number of control words per frame
clock cycle per link (CF =
Register 0x0592, Bits[4:0]).
Serial Checksum Value for Lane 0.
Automatically calculated for each
lane. Sum(all link configuration
parameters for Lane 0) mod 256.
Serial Checksum Value for Lane 1.
Automatically calculated for each
lane. Sum(all link configuration
parameters for Lane 1) mod 256.
Serial Checksum Value for Lane 2.
Automatically calculated for each
lane. Sum(all link configuration
parameters for each lane) mod 256.
Serial Checksum Value for Lane 3.
Automatically calculated for each
lane. Sum(all link configuration
parameters for Lane 3) mod 256.
Reset
0x1
Access
R/W
0x0
R/W
0x0
0xF
R
R/W
0x1
R/W
0xF
R/W
0x1
0x0
R
R
0x0
R
0x0
0x0
R
R
0xC3
R
0xC4
R
0xC5
R
0xC6
R
AD9695
Address
0x05B0
Data Sheet
Name
JESD204B lane
power-down
Bits
7
6
5
4
3
2
1
0
0x05B2
0x05B3
0x05B5
0x05B6
0x05BF
JESD204B Lane
Assignment 1
JESD204B Lane
Assignment 2
JESD204B Lane
Assignment 3
JESD204B Lane
Assignment 4
SERDOUTx± data
invert
[7:3]
[2:0]
[7:3]
[2:0]
[7:3]
[2:0]
[7:3]
[2:0]
7
6
Bit Name
Reserved
JESD204B Lane 3
power-down
Reserved
JESD204B Lane 2
power-down
Reserved
JESD204B Lane 1
power-down
Reserved
JESD204B Lane 0
power-down
Reserved
SERDOUT0± lane
assignment
Reserved
SERDOUT1± lane
assignment
Reserved
SERDOUT2± lane
assignment
Reserved
SERDOUT3± lane
assignment
Settings
0
1
10
11
0
1
10
11
0
1
10
11
0
1
10
11
Reserved
Invert SERDOUT3± data
0
1
5
4
Reserved
Invert SERDOUT2± data
0
1
3
2
Reserved
Invert SERDOUT1± data
0
1
1
0
Reserved
Invert SERDOUT0± data
0
1
Rev. 0 | Page 128 of 135
Description
Reserved.
Physical Lane 3 force power-down.
Reset
0x0
0x0
Access
R
R/W
Reserved.
Physical Lane 2 force power-down.
0x0
0x0
R
R/W
Reserved.
Physical Lane 1 force power-down.
0x0
0x0
R
R/W
Reserved.
Physical Lane 0 force power-down.
0x0
0x0
R
R/W
Reserved.
Physical Lane 0 assignment.
Logical Lane 0.
Logical Lane 1.
Logical Lane 2.
Logical Lane 3.
Reserved.
Physical Lane 1 assignment.
Logical Lane 0.
Logical Lane 1
Logical Lane 2
Logical Lane 3.
Reserved.
Physical Lane 2 assignment.
Logical Lane 0.
Logical Lane 1.
Logical Lane 2.
Logical Lane 3.
Reserved.
Physical Lane 3 assignment.
Logical Lane 0.
Logical Lane 1.
Logical Lane 2.
Logical Lane 3.
Reserved.
Invert SERDOUT3± data.
Normal.
Invert.
Reserved.
Invert SERDOUT2± data.
Normal.
Invert.
Reserved.
Invert SERDOUT1± data
Normal.
Invert.
Reserved.
Invert SERDOUT0± data.
Normal.
Invert.
0x0
0x0
R
R/W
0x0
0x1
R
R/W
0x0
0x2
R
R/W
0x0
0x3
R
R/W
0x0
0x0
R/W
R/W
0x0
0x0
R/W
R/W
0x0
0x0
R/W
R/W
0x0
0x0
R/W
R/W
Data Sheet
Address
0x05C0
0x05C1
0x05C2
0x05C3
0x05C4
Name
JESD204B Swing
Adjust 1
JESD204B Swing
Adjust 2
JESD204B Swing
Adjust 3
JESD204B Swing
Adjust 4
SERDOUT0± preemphasis select
AD9695
Bits
[7:3]
[2:0]
[7:3]
[2:0]
[7:3]
[2:0]
[7:3]
[2:0]
7
SERDOUT1± preemphasis select
[3:0]
7
SERDOUT2± preemphasis select
Reserved
SERDOUT2± voltage
swing adjust
Reserved
SERDOUT3± voltage
swing adjust
0
1
2
0
1
2
0
1
2
0
1
2
Post tap enable
Set post tap level for
SERDOUT0±
000
001
010
011
100
Reserved
Post tap enable
0
1
[6:4]
0x05C8
Reserved
SERDOUT1± voltage
swing adjust
Settings
0
1
[6:4]
0x05C6
Bit Name
Reserved
SERDOUT0± voltage
swing adjust
[3:0]
7
Set post tap level for
SERDOUT1
000
001
010
011
100
Reserved
Post tap enable
0
1
[6:4]
[3:0]
Set post tap level for
SERDOUT2
000
001
010
011
100
Reserved
Rev. 0 | Page 129 of 135
Description
Reserved.
Output swing level for SERDOUT0±.
1.0 × DRVDD1.
0.850 × DRVDD1.
0.750 × DRVDD1.
Reserved.
Output swing level for SERDOUT1±.
1.0 × DRVDD1.
0.850 × DRVDD1.
0.750 × DRVDD1.
Reserved.
Output swing level for SERDOUT2±.
1.0 × DRVDD1.
0.850 × DRVDD1.
0.750 × DRVDD1.
Reserved.
Output swing level for SERDOUT3±.
1.0 × DRVDD1.
0.850 × DRVDD1.
0.750 × DRVDD1.
Post tab enable.
Disable.
Enable.
Set post tap level.
0 dB.
3 dB.
6 dB.
9 dB.
12 dB.
Reserved.
Post tab enable.
Disable.
Enable.
Set post tap level.
0 dB.
3 dB.
6 dB.
9 dB.
12 dB.
Reserved.
Post tab enable.
Disable.
Enable.
Set post tap level.
0 dB.
3 dB.
6 dB.
9 dB.
12 dB.
Reserved.
Reset
0x0
0x1
Access
R
R/W
0x0
0x1
R
R/W
0x0
0x1
R
R/W
0x0
0x1
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R/W
R/W
0x0
R/W
0x0
0x0
R/W
R/W
0x0
R/W
0x0
R/W
AD9695
Address
0x05CA
Data Sheet
Name
SERDOUT3± preemphasis select
Bits
7
Bit Name
Post tap enable
Settings
0
1
[6:4]
Set post tap level for
SERDOUT3±
[3:0] Reserved
Programmable Filter Control and Coefficients Registers
0x0DF8
[7:3] Reserved
Programmable
filter control
[2:0] Programmable filter
mode
000
001
010
011
100
000
001
010
100
101
0x0DF9
PFILT gain
7
[6:4]
Reserved
PFILT Y gain
3
[2:0]
Reserved
PFILT X gain
Rev. 0 | Page 130 of 135
Description
Post tab enable.
Disabled
Enabled.
Set post tap level.
0 dB.
3 dB.
6 dB.
9 dB.
12 dB.
Reserved.
Reserved.
Programmable filter (PFILT) mode.
Disabled (filters bypassed).
Single filter (Filter X only).
DOUT_I[n] = DIN_I[n] × X_I[n].
DOUT_Q[n] = DIN_Q[n] × X_Q[n].
Single filter (Filter X and Filter Y
together).
DOUT_I[n] = DIN_I[n] × XY_I[n].
DOUT_Q[n] = DIN_Q[n] × XY_Q[n].
Cascaded filters (Filter X to Filter Y).
DOUT_I[n] = DIN_I[n] × X_I[n] × Y_I[n].
DOUT_Q[n] = DIN_Q[n] × X_Q[n] ×
Y_Q[n].
Complex filters.
DOUT_I[n] = DIN_I[n] × X_I[n] +
DIN_Q[n] × Y_Q[n].
DOUT_Q[n] = DIN_Q[n] × X_Q[n] +
DIN_I[n] × Y_I[n].
Reserved.
Programmable filter (PFILT) Y gain
100 = reserved.
101 = reserved.
110 = −12 dB loss.
111 = −6 dB loss.
000: 0 dB gain.
001: +6 dB gain.
010: +12 dB gain.
011: reserved.
Reserved.
Programmable filter (PFILT) X gain.
100 = reserved.
101 = reserved.
110 = −12 dB loss.
111 = −6 dB loss.
000: 0 dB gain.
001: +6 dB gain.
010: +12 dB gain.
011: reserved.
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
Data Sheet
AD9695
Address
0x0E00 to
0x0E7F
Name
Programmable
Filter X
Coefficient x
Bits
[7:0]
Bit Name
Programmable Filter X
Coefficient 0 to
Programmable Filter X
Coefficient 127
0x0F00 to
0x0F7F
Programmable
Filter Y
Coefficient x
[7:0]
Programmable Filter Y
Coefficient 0 to
Programmable Filter Y
Coefficient 127
VREF/Analog Input Control Registers
0x0701
[7]
DC Offset
Calibration
Control 1 (local)
0x073B
0x18A6
0x18E3
0x18E6
DC Offset
Calibration
Control 2 (local)
VREF control
External VCM
buffer control
Temperature
diode export
[6:0]
[7]
DC Offset Calibration
Enable 1
Settings
0
1
Reserved
DC Offset Calibration
Enable 2
110
0
1
[6:0]
[7:1]
0
Reserved
Reserved
VREF control
7
6
Reserved
External VCM buffer
[5:0]
External VCM buffer
[5:0]
Temperature diode
location select
[7:0]
111111
0
1
0
1
0x00
0x01
0x02
0x03
0x40
0x41
0x42
0x43
0x50
0x51
0x52
0x53
0x1908
Analog input
control (local)
[7:3]
2
Reserved
Enable dc coupling
[1:0]
Reserved
0
1
Rev. 0 | Page 131 of 135
Description
Refer to the I coefficient table (Table 14)
and the Q coefficient table (Table 15) in
the Programmable Finite Impulse
Response (FIR) Filters section for
details. Coefficients are only applied
after the chip transfer bit
(Register 0x000F, Bit 0) is set.
Refer to the I coefficient table (Table 14)
and the Q coefficient table (Table 15) in
the Programmable Finite Impulse
Response (FIR) Filters section for
details. Coefficients are only applied
after the chip transfer bit
(Register 0x000F, Bit 0) is set.
Reset
0x0
Access
R/W
0x0
R/W
Disabled (must set to 0 when
Register 0x073B, Bit 7 = 1).
Enabled (must set to 1 when
Register 0x073B, Bit 7 = 0).
Reserved.
Enabled (must set to 0 when
Register 0x0701, Bit 7 = 1).
Disabled (must set to 1 when
Register 0x0701, Bit 7 = 0).
Reserved.
Reserved.
Internal reference.
External reference.
Reserved.
Disable.
Enable.
See the Input Common Mode section.
0x0
R/W
0x6
0x1
R
R/W
0x3F
0x0
0x0
R
R
R/W
0x0
0x0
R
R/W
0x0
R/W
See the Temperature Diode section.
0x0
Central diode. VREF pin = high-Z.
Central diode. VREF pin = 1× diode
voltage output.
Central diode. VREF pin = 20× diode
voltage output.
Central diode. VREF pin = GND.
Channel A diode. VREF pin = high-Z.
Channel A diode. VREF pin = 1× diode
voltage output.
Channel A diode. VREF pin = 20× diode
voltage output.
Channel A diode. VREF pin = GND.
Channel B diode. VREF pin = high-Z.
Channel B diode. VREF pin = 1× diode
voltage output.
Channel B diode. VREF pin = 20× diode
voltage output.
Channel B diode. VREF pin = GND.
Reserved.
0x0
Analog input optimized for ac coupling. 0x0
Analog input optimized for dc coupling.
Reserved.
0x0
R/W
R
R/W
R
AD9695
Address
0x1910
Data Sheet
Name
Input full-scale
control (local)
Bits
[7:4]
[3:0]
Bit Name
Reserved
TRM VREF 1.8 V
Settings
0
1010
1011
1100
1101
1110
1111
0x1A4C
Buffer Control 1
(local)
[7:6]
[5:0]
Reserved
Buffer Control P
00110
01000
01010
01100
01110
10000
10010
10100
0x1A4D
Buffer Control 2
(local)
[7:6]
[5:0]
Reserved
Buffer Control N
00110
01000
01010
01100
01110
10000
10010
10100
0x1B03
Buffer Control 3
(local)
[7:0]
Buffer Control 3
0x00
0x02
0x1B08
0x1B10
Buffer Control 4
(local)
[7:0]
Buffer Control 5
(local)
[7:0]
Buffer Control 4
0x01
0xC1
Buffer Control 5
0x00
0x1C
Rev. 0 | Page 132 of 135
Description
Reserved.
Full-scale voltage setting.
2.04 V p-p differential.
1.36 V p-p differential.
1.47 V p-p differential.
1.59 V p-p differential.
1.70 V p-p differential.
1.81 V p-p differential.
1.93 V p-p differential.
Reserved.
Input buffer main current (P).
Buffer current set to 120 μA.
Buffer current set to 160 μA.
Buffer current set to 200 μA.
Buffer current set to 240 μA.
Buffer current set to 280 μA.
Buffer current set to 320 μA.
Buffer current set to 360 μA.
Buffer current set to 400 μA.
Reserved.
Input buffer main current (N).
Buffer current set to 120 μA.
Buffer current set to 160 μA.
Buffer current set to 200 μA.
Buffer current set to 240 μA.
Buffer current set to 280 μA.
Buffer current set to 320 μA.
Buffer current set to 360 μA.
Buffer current set to 400 μA.
Buffer Control 3.
Setting 1
Setting 2
Buffer Control 4.
Setting 1
Setting 2
Buffer Control 5.
Setting 1
Setting 2
Reset
0x0
0xD
Access
R
R/W
0x0
0x1E
R
R/W
0x0
0x1E
R
R/W
0x00
R/W
0x01
R/W
0x00
R/W
Data Sheet
AD9695
APPLICATIONS INFORMATION
POWER SUPPLY RECOMMENDATIONS
The power supplies required to power the AD9695 are shown in
Table 44.
Table 44. Typical Power Supplies for AD9695
Voltage (V)
0.95
0.95
0.95
0.95
1.8
1.8
1.8
2.5
Tolerance (%)
±2.5
±2.5
±2.5
±2.5
±5
±5
±5
±2.5
For applications requiring an optimal high power efficiency and
low noise performance, it is recommended that the ADP5054
quad switching regulator be used to convert the 6.0 V or 12 V
input rails to intermediate rails (1.3 V, 2.4 V, and 3.0 V). These
intermediate rails are then postregulated by very low noise, low
dropout (LDO) regulators (ADP1763, ADP7159, and ADP151).
Figure 143 shows the recommended power supply scheme for
the AD9695.
6.0V
TO
15.0V
1.3V
ANALOG
ADP5054 1.3V
DIGITAL
ADP1763
ADP1763
ADP7159
ADP151
ADP5054
ADP1763
1.3V
DIGITAL
AVDD1
0.95V
AVDD1_SR
0.95V
DVDD
0.95V
SW3
DRVDD1
0.95V
SW4
2.4V
ADP7159
AVDD1_SR
0.95V
OPTIONAL
DVDD
0.95V
AVDD2
1.8V
SPIVDD
1.8V
FERRITE BEAD
LDO
SWITCHER
OPTIONAL PATH
3.0V
ADP7159
AVDD3
2.5V
NOTES
1. ALL VOLTAGES REFERENCED TO AGND.
AVDD2
1.8V
Figure 144. Simplified Power Solution for the AD9695
OPTIONAL
DRVDD2
1.8V
The user can employ several different decoupling capacitors to
cover both high and low frequencies. These capacitors must be
located close to the point of entry at the PCB level and close to
the devices, with minimal trace lengths.
AVDD3
2.5V
15660-113
LDO
SWITCHER
OPTIONAL PATH
REFERENCED TO AGND
ADP7159
SW2
1.3V
ANALOG
DRVDD2
1.8V
SPIVDD
1.8V
3.0V
SW1
AVDD1
0.95V
DRVDD1
0.95V
2.4V
12V FROM FMC OR
6.0V FROM WALL
SUPPLY
Figure 143. High Efficiency, Low Noise Power Solution for the AD9695
Rev. 0 | Page 133 of 135
15660-114
Domain
AVDD1
AVDD1_SR
DVDD
DRVDD1
AVDD2
DRVDD2
SPIVDD
AVDD3
It is not necessary to split all of these power domains in all cases.
The recommended solution shown in Figure 143 provides the
lowest noise, highest efficiency power delivery system for the
AD9695. If only one 0.975 V supply is available, route to AVDD1
first and then tap it off and isolate it with a ferrite bead or a
filter choke, preceded by decoupling capacitors for AVDD1_SR,
DVDD, and DRVDD1, in that order. Figure 144 shows the
simplified schematic. Alternatively, the LDOs can be bypassed
altogether and the AD9695 can be driven directly from the dcto-dc converter. Note that this approach has risks in that there
may be more power supply noise injected into the power supply
domains of the ADC. To minimize noise, follow the layout
guidelines of the dc-to-dc converter.
AD9695
Data Sheet
LAYOUT GUIDELINES
The ADC evaluation board can be used as a guide to follow
good layout practices. The evaluation board layout is set up in
such a way as to




Minimize coupling between the analog inputs (Channel A
to Channel B and Channel B to Channel A).
Minimize clock coupling to the analog inputs.
Provide enough power and ground planes for the various
supply domains while reducing cross coupling.
Provide adequate thermal relief to the ADC.
AVDD1_SR (PIN 57) AND AGND_SR (PIN 56 AND
PIN 60)
AVDD1_SR (Pin 57) and AGND_SR (Pin 56 and Pin 60) can be
used to provide a separate power supply node to the SYSREF±
circuits of AD9695. If running in Subclass 1, the AD9695 can
support periodic one-shot or gapped signals. To minimize the
coupling of this supply into the AVDD1 supply node, adequate
supply bypassing is needed.
Figure 145 shows the overall layout scheme used for the
AD9695 evaluation board.
CH.A
EF
SR
SY
ADC
CH.B
JESD204B LANES
POWER
Figure 145. Recommended PCB Layout for the AD9695
.
Rev. 0 | Page 134 of 135
15660-115
CLK
Data Sheet
AD9695
OUTLINE DIMENSIONS
9.10
9.00 SQ
8.90
0.30
0.25
0.18
PIN 1
INDICATOR
49
PIN 1
INDICATOR
64
1
48
0.50
BSC
EXPOSED
PAD
6.30
6.20 SQ
6.10
33
PKG-004559
0.80
0.75
0.70
32
17
16
BOTTOM VIEW
0.05 MAX
0.02 NOM
COPLANARITY
0.08
0.203 REF
SEATING
PLANE
0.20 MIN
7.50 REF
COMPLIANT TO JEDEC STANDARDS MO-220-WMMD
06-06-2014-A
TOP VIEW
0.45
0.40
0.35
Figure 146. 64-Lead Lead Frame Chip Scale Package [LFCSP]
9 mm × 9 mm Body and 0.75 mm Package Height
(CP-64-17)
Dimensions shown in millimeters
ORDERING GUIDE
Model1
AD9695BCPZ-625
AD9695BCPZRL7-625
AD9695-625EBZ
AD9695BCPZ-1300
AD9695BCPZRL7-1300
AD9695-1300EBZ
1
Temperature Range
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
Package Description
64-Lead Lead Frame Chip Scale Package [LFCSP]
64-Lead Lead Frame Chip Scale Package [LFCSP]
Evaluation Board
64-Lead Lead Frame Chip Scale Package [LFCSP]
64-Lead Lead Frame Chip Scale Package [LFCSP]
Evaluation Board
Z = RoHS Compliant Part.
©2017 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D15660-0-9/17(0)
Rev. 0 | Page 135 of 135
Package Option
CP-64-17
CP-64-17
CP-64-17
CP-64-17
Similar pages