AD AD9144BCPZRL Multiple chip synchronization Datasheet

Quad, 16-Bit, 2.8 GSPS, TxDAC+®
Digital-to-Analog Converter
AD9144
Data Sheet
FEATURES
TYPICAL APPLICATION CIRCUIT
Supports input data rate >1 GSPS
Proprietary low spurious and distortion design
6-carrier GSM IMD = 77 dBc at 75 MHz IF
SFDR = 82 dBc at dc IF, −9 dBFS
Flexible 8-lane JESD204B interface
Support quad or dual DAC mode at 2.8 GSPS
Multiple chip synchronization
Fixed latency
Data generator latency compensation
Selectable 1×, 2×, 4×, 8× interpolation filter
Low power architecture
Input signal power detection
Emergency stop for downstream analog circuitry protection
Transmit enable function allows extra power saving
High performance, low noise phase-locked loop (PLL) clock
multiplier
Digital inverse sinc filter
Low power: 1.6 W at 1.6 GSPS, 1.7 W at 2.0 GSPS,
full operating conditions
88-lead LFCSP with exposed pad
QUAD MOD
ADRF6720
LPF
DAC
0°/90° PHASE
SHIFTER
JESD204B
SYNCOUTx±
SYSREF±
DAC
AD9144
LO_IN
MOD_SPI
QUAD MOD
ADRF6720
QUAD
DAC
LPF
DAC
0°/90° PHASE
SHIFTER
JESD204B
SYNCOUTx±
LO_IN
MOD_SPI
CLK±
DAC
SPI
11675-001
DAC
Figure 1.
APPLICATIONS
Wireless communications
3G/4G W-CDMA base stations
Wideband repeaters
Software defined radios
Wideband communications
Point-to-point
Local multipoint distribution service (LMDS) and
multichannel multipoint distribution service (MMDS)
Transmit diversity, multiple input/multiple output (MIMO)
Instrumentation
Automated test equipment
PRODUCT HIGHLIGHTS
1.
GENERAL DESCRIPTION
The AD9144 is a quad, 16-bit, high dynamic range digital-toanalog converter (DAC) that provides a maximum sample rate
of 2.8 GSPS, permitting a multicarrier generation up to the
Nyquist frequency. The DAC outputs are optimized to interface
seamlessly with the ADRF6720 analog quadrature modulator
(AQM) from Analog Devices, Inc. An optional 3-wire or 4-wire
serial port interface (SPI) provides for programming/readback
of many internal parameters. Full-scale output current can be
programmed over a typical range of 13.9 mA to 27.0 mA. The
AD9144 is available in an 88-lead LFCSP.
Rev. A
2.
3.
4.
5.
6.
Greater than 1 GHz, ultrawide complex signal bandwidth
enables emerging wideband and multiband wireless
applications.
Advanced low spurious and distortion design techniques
provide high quality synthesis of wideband signals from
baseband to high intermediate frequencies.
JESD204B Subclass 1 support simplifies multichip
synchronization in software and hardware design.
Fewer pins for data interface width with a serializer/
deserializer (SERDES) JESD204B eight-lane interface.
Programmable transmit enable function allows easy design
balance between power consumption and wake-up time.
Small package size with 12 mm × 12 mm footprint.
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
©2014 Analog Devices, Inc. All rights reserved.
Technical Support
www.analog.com
AD9144* Product Page Quick Links
Last Content Update: 11/01/2016
Comparable Parts
Reference Materials
View a parametric search of comparable parts
Informational
• JESD204 Serial Interface
Technical Articles
• MS-2773: Advancement in High Speed Converter
Technology Enables Next-Generation Wireless
Communications Systems Designs
Evaluation Kits
• AD-FMCDAQ2-EBZ Evaluation Board
• AD9144 Evaluation Board
Documentation
Data Sheet
• AD9144: Quad, 16-Bit, 2.8 GSPS, TxDAC+® Digital-toAnalog Converter Data Sheet
Tools and Simulations
• AD9144 IBIS Model
• AD9144/AD9152/AD9154/AD9135/AD9136 AMI Model
Download
Design Resources
•
•
•
•
AD9144 Material Declaration
PCN-PDN Information
Quality And Reliability
Symbols and Footprints
Discussions
View all AD9144 EngineerZone Discussions
Sample and Buy
Visit the product page to see pricing options
Technical Support
Submit a technical question or find your regional support
number
* This page was dynamically generated by Analog Devices, Inc. and inserted into this data sheet. Note: Dynamic changes to
the content on this page does not constitute a change to the revision number of the product data sheet. This content may be
frequently modified.
AD9144
Data Sheet
TABLE OF CONTENTS
Features .............................................................................................. 1
JESD204B Setup ......................................................................... 29
Applications ....................................................................................... 1
SERDES Clocks Setup ................................................................ 31
General Description ......................................................................... 1
Equalization Mode Setup .......................................................... 31
Typical Application Circuit ............................................................. 1
Link Latency Setup ..................................................................... 31
Product Highlights ........................................................................... 1
Crossbar Setup ............................................................................ 33
Revision History ............................................................................... 3
JESD204B Serial Data Interface .................................................... 34
Functional Block Diagram .............................................................. 4
JESD204B Overview .................................................................. 34
Specifications..................................................................................... 5
Physical Layer ............................................................................. 35
DC Specifications ......................................................................... 5
Data Link Layer .......................................................................... 38
Digital Specifications ................................................................... 6
Transport Layer .......................................................................... 47
Maximum DAC Update Rate Speed Specifications by Supply ..... 7
JESD204B Test Modes ............................................................... 60
JESD204B Serial Interface Speed Specifications ...................... 7
JESD204B Error Monitoring..................................................... 61
SYSREF to DAC Clock Timing Specifications ......................... 8
Hardware Considerations ......................................................... 63
Digital Input Data Timing Specifications ................................. 8
Digital Datapath ............................................................................. 67
Latency Variation Specifications ................................................ 9
Dual Paging ................................................................................. 67
JESD204B Interface Electrical Specifications ........................... 9
Data Format ................................................................................ 67
AC Specifications........................................................................ 10
Interpolation Filters ................................................................... 67
Absolute Maximum Ratings.......................................................... 11
Digital Modulation ..................................................................... 68
Thermal Resistance .................................................................... 11
Inverse Sinc ................................................................................. 69
ESD Caution ................................................................................ 11
Digital Gain, Phase Adjust, DC Offset, and Group Delay .... 69
Pin Configuration and Function Descriptions ........................... 12
I to Q Swap .................................................................................. 70
Terminology .................................................................................... 15
NCO Alignment ......................................................................... 70
Typical Performance Characteristics ........................................... 16
Downstream Protection ............................................................ 72
Theory of Operation ...................................................................... 21
Datapath PRBS ........................................................................... 74
Serial Port Operation ..................................................................... 22
DC Test Mode ............................................................................. 74
Data Format ................................................................................ 22
Interrupt Request Operation ........................................................ 75
Serial Port Pin Descriptions ...................................................... 22
Interrupt Service Routine .......................................................... 75
Serial Port Options ..................................................................... 22
DAC Input Clock Configurations ................................................ 76
Chip Information ............................................................................ 24
Driving the CLK± Inputs .......................................................... 76
Device Setup Guide ........................................................................ 25
DAC PLL Fixed Register Writes ............................................... 76
Overview...................................................................................... 25
Clock Multiplication .................................................................. 76
Step 1: Start Up the DAC ........................................................... 25
Starting the PLL .......................................................................... 78
Step 2: Digital Datapath ............................................................. 26
Analog Outputs............................................................................... 79
Step 3: Transport Layer .............................................................. 26
Transmit DAC Operation.......................................................... 79
Step 4: Physical Layer ................................................................. 27
Device Power Dissipation.............................................................. 82
Step 5: Data Link Layer .............................................................. 27
Temperature Sensor ................................................................... 82
Step 6: Optional Error Monitoring .......................................... 28
Start-Up Sequence .......................................................................... 83
Step 7: Optional Features ........................................................... 28
Step 1: Start Up the DAC ........................................................... 83
DAC PLL Setup ........................................................................... 29
Step 2: Digital Datapath............................................................. 83
Interpolation ............................................................................... 29
Step 3: Transport Layer .............................................................. 84
Rev. A | Page 2 of 125
Data Sheet
AD9144
Step 4: Physical Layer..................................................................84
Device Configuration Register Map ......................................... 86
Step 5: Data Link Layer ..............................................................85
Device Configuration Register Descriptions .......................... 94
Step 6: Error Monitoring ............................................................85
Outline Dimensions ......................................................................124
Register Maps and Descriptions ....................................................86
Ordering Guide .........................................................................125
REVISION HISTORY
6/15—Rev. 0 to Rev. A
Changed Functional Block Diagram Section to Typical
Application Circuit Section.............................................................. 1
Changes to Figure 1........................................................................... 1
Changed Detailed Functional Block Diagram Section to
Functional Block Diagram Section ................................................. 4
Deleted Reference Voltage Parameter, Table 1 .............................. 5
Changes to Output Voltage (VOUT) Logic High Parameter,
Output Voltage (VOUT) Logic Low Parameter, and SYSREF±
Frequency Parameter, Table 2 .......................................................... 6
Changes to Table 4 ............................................................................ 7
Changes to Interpolation Parameter, Table 6 ................................ 8
Deleted Sync Off, Subclass Mode 0 Parameter, Table 7 ............... 9
Changed Junction Temperature Parameter to Operating
Junction Temperature, Table 10 ....................................................11
Changes to Terminology Section ..................................................15
Changes to Figure 26 Caption .......................................................19
Changes to Figure 29 Caption .......................................................20
Change to Device Revision Parameter, Table 14 .........................24
Changes to Step 1: Start Up the DAC Section, Table 16, and
Table 17 .............................................................................................25
Changes to Step 3: Transport Layer Section and Table 19 .........26
Changes to Table 20 and Table 21 .................................................27
Changes to Step 7: Optional Features Section .............................28
Added Table 25; Renumbered Sequentially .................................29
Changes to DAC PLL Setup Section and Table 26 ......................29
Changes to Lane0Checksum Section ...........................................30
Changes to Table 30 and Subclass 0 Section ................................31
Changes to Table 33 ........................................................................32
Changes to Table 37 ........................................................................35
Changes to Table 38 ........................................................................36
Added SERDES PLL Fixed Register Writes Section and
Table 39 .............................................................................................36
Changes to Figure 38 and Table 40 ...............................................37
Changes to Figure 29 and Data Link Layer Section ...................38
Added Figure 42; Renumbered Sequentially ...............................39
Changes to Figure 44 ......................................................................40
Changes to Continuous Sync Mode (SYNCMOD = 0x2)
Section ..............................................................................................42
Changes to Subclass 0 Section .......................................................43
Changes to Figure 53 ......................................................................50
Changes to Table 49 and Figure 54 ............................................... 51
Changes to Table 50 and Figure 55 ............................................... 52
Changes to Table 51 and Figure 56 ............................................... 53
Changes to Table 52 and Figure 57 ............................................... 54
Changes to Table 53, Table 54, and Figure 58 ............................. 55
Changes to Table 55 and Figure 59 ............................................... 56
Changes to Table 56 and Figure 60 ............................................... 57
Changes to Table 57 and Figure 61 ............................................... 58
Changes to Table 58 and Figure 62 ............................................... 59
Changes to Power Supply Recommendations Section ............... 63
Added Figure 64 .............................................................................. 64
Changes to Figure 68 ...................................................................... 66
Changes to Table 66 ........................................................................ 67
Changes to Table 70, Table 71, Table 72, and I to Q Swap
Section .............................................................................................. 70
Changes to Power Detection and Protection Section ................ 72
Changes to DC Test Mode Section ............................................... 73
Moved Figure 75 and Table 78 ...................................................... 75
Deleted Table 80; Renumbered Sequentially ............................... 76
Added DAC PLL Fixed Register Writes Section and
Table 79 ............................................................................................. 76
Changes to Clock Multiplication Section .................................... 76
Added Loop Filter Section and Charge Pump Section .............. 77
Added Temperature Tracking Section and Table 83 .................. 78
Changes to Starting the PLL Section and Figure 79 ................... 78
Changes to Transmit DAC Operation Section ............................ 79
Changes to Self Calibration Section ............................................. 81
Added Figure 86 and Figure 87 ..................................................... 81
Changes to Device Power Dissipation Section............................ 82
Changes to Table 88 and Table 89 ................................................. 83
Changes to Table 93 ........................................................................ 84
Changes to Table 94, Table 95, and Table 96 ............................... 85
Changes to Table 97 ........................................................................ 86
Changes to Table 98 ........................................................................ 94
Deleted Lookup Tables for Three Different DAC PLL Reference
Frequencies Section and Table 96 to Table 98 ........................... 122
Added Figure 89 ............................................................................ 124
Updated Outline Dimensions...................................................... 124
Changes to Ordering Guide ......................................................... 125
7/14—Revision 0: Initial Version
Rev. A | Page 3 of 125
AD9144
Data Sheet
FUNCTIONAL BLOCK DIAGRAM
DACCLK
SERDES
PLL
INV SINC
FSC
NCO
MODE CONTROL
HB1
HB2
HB3
HB1
HB2
HB3
Q-GAIN
I-GAIN
OUT3–
Q-OFFSET
PHASE
ADJUST
DACCLK
I-OFFSET
OUT2+
÷4, ÷8
FSC
OUT2–
COMPLEX
MODULATION
OUT1+
INV SINC
FSC
PDP0
SERDIN0±
OUT3+
COMPLEX
MODULATION
HB3
fDAC
CLOCK DATA RECOVERY
AND CLOCK FORMATTER
VTT
SERDIN7±
HB2
PDP1
HB1
NCO
MODE CONTROL
Q-GAIN
I-GAIN
OUT1–
Q-OFFSET
PHASE
ADJUST
DACCLK
I-OFFSET
fDAC
HB1
PROTECT_OUT0
PROTECT_OUT1
FSC
PLL_CTRL
SERIAL
I/O PORT
POWER-ON
RESET
DACCLK
OUT0–
REF
AND
BIAS
SYSREF
Rx
CLK
Rx
I120
SYSREF+
SYSREF–
CLK+
CLK–
11675-002
DAC PLL
TXEN1
TXEN0
PLL_LOCK
IRQ
DAC
ALIGN
DETECT
CLK_SEL
CONFIG
REGISTERS
CLOCK DISTRIBUTION
AND
CONTROL LOGIC
RESET
SYNCOUT1+
SYNCOUT1–
OUT0+
÷4, ÷8
HB3
SYNCHRONIZATION
LOGIC
SDO
SDIO
SCLK
CS
SYNCOUT0+
SYNCOUT0–
HB2
Figure 2.
Rev. A | Page 4 of 125
Data Sheet
AD9144
SPECIFICATIONS
DC SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted.
Table 1.
Parameter
RESOLUTION
ACCURACY
Differential Nonlinearity (DNL)
Integral Nonlinearity (INL)
MAIN DAC OUTPUTS
Gain Error
I/Q Gain Mismatch
Full-Scale Output Current
Maximum Setting
Minimum Setting
Output Compliance Range
Output Resistance
Output Capacitance
Gain DAC Monotonicity
Settling Time
MAIN DAC TEMPERATURE DRIFT
Offset
Gain
REFERENCE
Internal Reference Voltage
ANALOG SUPPLY VOLTAGES
AVDD33
PVDD12
CVDD12
DIGITAL SUPPLY VOLTAGES
SIOVDD33
VTT
DVDD12
Test Conditions/Comments
Min
Typ
16
Max
Unit
Bits
With calibration
±1.0
±2.0
With internal reference
LSB
LSB
−2.5
−0.6
+2
+5.5
+0.6
% FSR
% FSR
25.5
13.1
−250
27.0
13.9
28.6
14.8
+750
mA
mA
mV
MΩ
pF
Based on a 4 kΩ external resistor between I120 and GND
0.2
3.0
Guaranteed
20
To within ±0.5 LSB
ns
0.04
32
ppm
ppm/°C
1.2
V
3.13
1.14
1.14
3.3
1.2
1.2
3.47
1.26
1.26
V
V
V
3.13
1.1
3.3
1.2
3.47
1.37
V
V
1.14
1.274
1.2
1.3
1.26
1.326
V
V
1.14
1.274
1.71
1.2
1.3
1.8
1.26
1.326
3.47
V
V
V
1.59
1.84
W
126
95.3
101
518.2
234
11
36
134
112.4
111
654
255
12
50
mA
mA
mA
mA
mA
mA
μA
SVDD12
IOVDD
POWER CONSUMPTION
4× Interpolation Mode,
JESD Mode 4, 8 SERDES Lanes
AVDD33
PVDD12
CVDD12
SVDD12
DVDD12
SIOVDD33
IOVDD
fDAC = 1.6 GSPS, IF = 40 MHz, NCO off, PLL on, digital gain
on, inverse sinc on, DAC FSC = 20 mA
Includes VTT
Rev. A | Page 5 of 125
AD9144
Data Sheet
DIGITAL SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted.
Table 2.
Parameter
CMOS INPUT LOGIC LEVEL
Input Voltage (VIN) Logic
High
Low
Symbol
CMOS OUTPUT LOGIC LEVEL
Output Voltage (VOUT) Logic
High
Low
Test Conditions/Comments
Min
Typ
Max
Unit
1.8 V  IOVDD  3.3 V
1.8 V  IOVDD  3.3 V
0.7 × IOVDD
0.3 × IOVDD
V
V
1.8 V  IOVDD  3.3 V
1.8 V  IOVDD  3.3 V
0.75 × IOVDD
0.25 × IOVDD
V
V
1× interpolation2 (see Table 4)
2× interpolation3
4× interpolation
8× interpolation
1060
2120
2800
2800
MSPS
MSPS
MSPS
MSPS
1× interpolation
2× interpolation
4× interpolation
8× interpolation
1060
1060
700
350
MSPS
MSPS
MSPS
MSPS
MAXIMUM DAC UPDATE RATE1
ADJUSTED DAC UPDATE RATE
INTERFACE4
Number of JESD204B Lanes
JESD204B Serial Interface Speed
Minimum
Maximum
DAC CLOCK INPUT (CLK+, CLK−)
Differential Peak-to-Peak Voltage
Common-Mode Voltage
Maximum Clock Rate
REFCLK Frequency (PLL Mode)
SYSTEM REFERENCE INPUT
(SYSREF+, SYSREF−)
Differential Peak-to-Peak
Voltage
Common-Mode Voltage
SYSREF± Frequency5
SYSREF TO DAC CLOCK6
Setup Time
Hold Time
Keep Out Window
SPI
Maximum Clock Rate
Minimum SCLK Pulse Width
High
Low
SDIO to SCLK
Setup Time
Hold Time
8
Per lane
Per lane, SVDD12 = 1.3 V ± 2%
1.44
Gbps
Gbps
2000
1000
mV
mV
MHz
MHz
2000
mV
2000
fDATA/(K × S)
mV
Hz
10.6
400
Self biased input, ac-coupled
6.0 GHz ≤ fVCO ≤ 12.0 GHz
Lanes
1000
600
2800
35
400
1000
0
SYSREF differential swing = 0.4 V, slew
rate = 1.3 V/ns, common modes tested:
ac-coupled, 0 V, 0.6 V, 1.25 V, 2.0 V
tSSD
tHSD
KOW
SCLK
131
119
ps
ps
ps
20
IOVDD = 1.8 V
10
tPWH
tPWL
MHz
8
12
tDS
tDH
5
2
Rev. A | Page 6 of 125
ns
ns
ns
ns
Data Sheet
Parameter
SDO to SCLK
Data Valid Window
CS to SCLK
Setup Time
Hold Time
AD9144
Symbol
Test Conditions/Comments
Min
Typ
Max
Unit
tDV
25
ns
tSCS
5
ns
tHCS
2
ns
1
See Table 3 for detailed specifications for DAC update rate conditions.
Maximum speed for 1× interpolation is limited by the JESD interface. See Table 4 for details.
Maximum speed for 2× interpolation is limited by the JESD interface. See Table 4 for details.
4
See Table 4 for detailed specifications for JESD speed conditions.
5
K, F, and S are JESD204B transport layer parameters. See Table 44 for the full definitions.
6
See Table 5 for detailed specifications for SYSREF to DAC clock timing conditions.
2
3
MAXIMUM DAC UPDATE RATE SPEED SPECIFICATIONS BY SUPPLY
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted.
Table 3.
Parameter
MAXIMUM DAC UPDATE RATE
Test Conditions/Comments
DVDD12, CVDD12 = 1.2 V ± 5%
DVDD12, CVDD12 = 1.2 V ± 2%
DVDD12, CVDD12 = 1.3 V ± 2%
Min
2.23
2.41
2.80
Typ
Max
Unit
GSPS
GSPS
GSPS
JESD204B SERIAL INTERFACE SPEED SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted.
Table 4.
Parameter
HALF RATE
FULL RATE
OVERSAMPLING
Test Conditions/Comments
SVDD12 = 1.2 V ± 5%
SVDD12 = 1.2 V ±2%
SVDD12 = 1.3 V ± 2%
SVDD12 = 1.2 V ± 5%
SVDD12 = 1.2 V ± 2%
SVDD12 = 1.3 V ± 2%
SVDD12 = 1.2 V ± 5%
SVDD12 = 1.2 V ± 2%
SVDD12 = 1.3 V ± 2%
Rev. A | Page 7 of 125
Min
5.75
5.75
5.75
2.88
2.88
2.88
1.44
1.44
1.44
Typ
Max
8.92
9.42
10.64
4.63
4.93
5.52
2.31
2.46
2.76
Unit
Gbps
Gbps
Gbps
Gbps
Gbps
Gbps
Gbps
Gbps
Gbps
AD9144
Data Sheet
SYSREF TO DAC CLOCK TIMING SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, SYSREF± common-mode voltages = 0.0 V, 0.6 V, 1.25 V, and 2.0 V, unless otherwise noted.
Table 5.
Parameter
SYSREF DIFFERENTIAL SWING = 0.4 V, SLEW RATE = 1.3 V/ns
Setup Time
Hold Time
SYSREF DIFFERENTIAL SWING = 0.7 V, SLEW RATE = 2.28 V/ns
Setup Time
Hold Time
SYSREF SWING = 1.0 V, SLEW RATE = 3.26 V/ns
Setup Time
Hold Time
Test Conditions/Comments
Min
Typ
Max
Unit
AC-coupled
DC-coupled
AC-coupled
DC-coupled
126
131
92
119
ps
ps
ps
ps
AC-coupled
DC-coupled
AC-coupled
DC-coupled
96
104
77
95
ps
ps
ps
ps
AC-coupled
DC-coupled
AC-coupled
DC-coupled
83
90
68
84
ps
ps
ps
ps
DIGITAL INPUT DATA TIMING SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = 25°C, IOUTFS = 20 mA, unless otherwise noted.
Table 6.
Parameter
LATENCY
Interface
Interpolation
1×
2×
4×
8×
Inverse Sinc
Fine Modulation
Coarse Modulation
fS/8
fS/4
Digital Phase Adjust
Digital Gain Adjust
Power-Up Time
Dual A Only
Dual B Only
All DACs
1
Test Conditions/Comments
Min
Register 0x011 from 0x60 to 0x00
Register 0x011 from 0x18 to 0x00
Register 0x011 from 0x7C to 0x00
PClock is the AD9144 internal processing clock and equals the lane rate ÷ 40.
Rev. A | Page 8 of 125
Typ
Max
Unit
17
PClock1 cycles
58
137
251
484
17
20
DAC clock cycles
DAC clock cycles
DAC clock cycles
DAC clock cycles
DAC clock cycles
DAC clock cycles
8
4
12
12
DAC clock cycles
DAC clock cycles
DAC clock cycles
DAC clock cycles
60
60
60
μs
μs
μs
Data Sheet
AD9144
LATENCY VARIATION SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = 25°C, IOUTFS = 20 mA, unless otherwise noted.
Table 7.
Parameter
DAC LATENCY VARIATION
SYNC On
PLL Off
PLL On
Min
Typ
Max
Unit
0
1
+1
DACCLK cycles
DACCLK cycles
−1
JESD204B INTERFACE ELECTRICAL SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V, VTT = 1.2 V,
TA = −40°C to +85°C, IOUTFS = 20 mA, unless otherwise noted.
Table 8.
Parameter
JESD204B DATA INPUTS
Input Leakage Current
Logic High
Logic Low
Unit Interval
Common-Mode Voltage
Differential Voltage
VTT Source Impedance
Differential Impedance
Differential Return Loss
Common-Mode Return Loss
DIFFERENTIAL OUTPUTS (SYNCOUT±)2
Output Differential Voltage
Output Offset Voltage
Output Differential Voltage
DETERMINISTIC LATENCY
Fixed
Variable
SYSREF±-to-LMFC DELAY
1
2
3
Symbol
Test Conditions/Comments
Min
25°C
Input level = 1.2 V ± 0.25 V, VTT = 1.2 V
Input level = 0 V
UI
VRCM
R_VDIFF
ZTT
ZRDIFF
RLRDIF
RLRCM
VOD
VOS
VOD
AC-coupled, VTT = SVDD121
At dc
At dc
High swing mode: Register 0x2A5[0] = 1
94
−0.05
110
100
8
6
192
1.19
341
4
As measured on the input side of the ac coupling capacitor.
IEEE Standard 1596.3 LVDS compatible.
PClock is the AD9144 internal processing clock and equals the lane rate ÷ 40.
Rev. A | Page 9 of 125
Max
Unit
714
+1.85
1050
30
120
μA
μA
ps
V
mV
Ω
Ω
dB
dB
235
1.27
394
mV
V
mV
17
2
PClock3 cycles
PClock3 cycles
DAC clock cycles
10
−4
80
Normal swing mode: Register 0x2A5[0] = 0
Typ
AD9144
Data Sheet
AC SPECIFICATIONS
AVDD33 = 3.3 V, SIOVDD33 = 3.3 V, IOVDD = 1.8 V, DVDD12 = 1.2 V, CVDD12 = 1.2 V, PVDD12 = 1.2 V, SVDD12 = 1.2 V,1 VTT = 1.2 V,
TA = 25°C, IOUTFS = 20 mA, unless otherwise noted.
Table 9.
Parameter
SPURIOUS-FREE DYNAMIC RANGE (SFDR)
fDAC = 983.04 MSPS
fDAC = 983.04 MSPS
fDAC = 1966.08 MSPS
fDAC = 1966.08 MSPS
TWO-TONE INTERMODULATION DISTORTION (IMD)
fDAC =983.04 MSPS
fDAC = 983.04 MSPS
fDAC = 1966.08 MSPS
fDAC = 1966.08 MSPS
NOISE SPECTRAL DENSITY (NSD), SINGLE-TONE
fDAC = 983.04 MSPS
fDAC = 1966.08 MSPS
W-CDMA FIRST ADJACENT CHANNEL LEAKAGE
RATIO (ACLR), SINGLE CARRIER
fDAC = 983.04 MSPS
fDAC = 983.04 MSPS
fDAC = 1966.08 MSPS
W-CDMA SECOND ACLR, SINGLE CARRIER
fDAC = 983.04 MSPS
fDAC = 983.04 MSPS
fDAC = 1966.08 MSPS
1
Test Conditions/Comments
−9 dBFS single-tone
fOUT = 20 MHz
fOUT = 150 MHz
fOUT = 20 MHz
fOUT = 170 MHz
−9 dBFS
fOUT = 20 MHz
fOUT = 150 MHz
fOUT = 20 MHz
fOUT = 170 MHz
0 dBFS
fOUT = 150 MHz
fOUT = 150 MHz
0 dBFS
fOUT = 30 MHz
fOUT = 150 MHz
fOUT = 150 MHz
0 dBFS
fOUT = 30 MHz
fOUT = 150 MHz
fOUT = 150 MHz
SVDD12 = 1.3 V for all fDAC = 1966.08 MSPS conditions in Table 9.
Rev. A | Page 10 of 125
Min
Typ
Max
Unit
82
76
81
69
dBc
dBc
dBc
dBc
90
82
90
81
dBc
dBc
dBc
dBc
−162
−163
dBm/Hz
dBm/Hz
82
80
80
dBc
dBc
dBc
84
85
85
dBc
dBc
dBc
Data Sheet
AD9144
ABSOLUTE MAXIMUM RATINGS
THERMAL RESISTANCE
Table 10.
Parameter
I120 to Ground
SERDINx±, VTT, SYNCOUT1±/
SYNCOUT0±, TXENx
OUTx±
SYSREF±
CLK± to Ground
RESET, IRQ, CS, SCLK, SDIO, SDO,
PROTECT_OUTx to Ground
LDO_BYP1
LDO_BYP2
LDO24
Ambient Operating Temperature (TA)
Operating Junction Temperature
Storage Temperature
Rating
−0.3 V to AVDD33 + 0.3 V
−0.3 V to SIOVDD33 + 0.3 V
−0.3 V to AVDD33 + 0.3 V
GND − 0.5 V to +2.5 V
−0.3 V to PVDD12 + 0.3 V
−0.3 V to IOVDD + 0.3 V
−0.3 V to SVDD12 + 0.3 V
−0.3 V to PVDD12 + 0.3 V
−0.3 V to AVDD33 + 0.3 V
−40°C to +85°C
125°C
−65°C to +150°C
The exposed pad (EPAD) must be soldered to the ground plane
for the 88-lead LFCSP. The EPAD provides an electrical,
thermal, and mechanical connection to the board.
Typical θJA, θJB, and θJC values are specified for a 4-layer JESD51-7
high effective thermal conductivity test board for leaded
surface-mount packages. θJA is obtained in still air conditions
(JESD51-2). Airflow increases heat dissipation, effectively reducing
θJA. θJB is obtained following double-ring cold plate test conditions
(JESD51-8). θJC is obtained with the test case temperature monitored at the bottom of the exposed pad.
ΨJT and ΨJB are thermal characteristic parameters obtained with
θJA in still air test conditions.
Junction temperature (TJ) can be estimated using the following
equations:
TJ = TT + (ΨJT × P), or
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.
TJ = TB + (ΨJB × P)
where:
TT is the temperature measured at the top of the package.
P is the total device power dissipation.
TB is the temperature measured at the board.
Table 11. Thermal Resistance
Package
88-Lead LFCSP1
1
θJA
22.6
θJB
5.59
θJC
1.17
ΨJT
0.1
ΨJB
5.22
The exposed pad must be securely connected to the ground plane.
ESD CAUTION
Rev. A | Page 11 of 125
Unit
°C/W
AD9144
Data Sheet
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
LDO_BYP2
CVDD12
I120
AVDD33
OUT0+
OUT0–
LDO24
CVDD12
LDO24
OUT1–
OUT1+
AVDD33
CVDD12
AVDD33
OUT2+
OUT2–
LDO24
CVDD12
LDO24
OUT3–
OUT3+
AVDD33
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
AD9144
TOP VIEW
(Not to Scale)
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
IOVDD
CS
SCLK
SDIO
SDO
RESET
IRQ
PROTECT_OUT0
PROTECT_OUT1
PVDD12
PVDD12
GND
GND
DVDD12
SERDIN7+
SERDIN7–
SVDD12
SERDIN6+
SERDIN6–
SVDD12
VTT
SVDD12
NOTES
1. THE EXPOSED PAD MUST BE SECURELY CONNECTED TO THE GROUND PLANE.
11675-003
SYNCOUT0+
SYNCOUT0–
VTT
SERDIN2+
SERDIN2–
SVDD12
SERDIN3+
SERDIN3–
SVDD12
SVDD12
SVDD12
LDO_BYP1
SIOVDD33
SVDD12
SERDIN4–
SERDIN4+
SVDD12
SERDIN5–
SERDIN5+
VTT
SYNCOUT1–
SYNCOUT1+
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
PVDD12
CLK+
CLK–
PVDD12
SYSREF+
SYSREF–
PVDD12
PVDD12
PVDD12
PVDD12
TXEN0
TXEN1
DVDD12
DVDD12
SERDIN0+
SERDIN0–
SVDD12
SERDIN1+
SERDIN1–
SVDD12
VTT
SVDD12
Figure 3. Pin Configuration
Table 12. Pin Function Descriptions
Pin No.
1
2
Mnemonic
PVDD12
CLK+
3
CLK−
4
5
PVDD12
SYSREF+
6
SYSREF−
7
8
9
10
11
12
13
14
15
PVDD12
PVDD12
PVDD12
PVDD12
TXEN0
TXEN1
DVDD12
DVDD12
SERDIN0+
16
SERDIN0−
17
18
SVDD12
SERDIN1+
19
SERDIN1−
Description
1.2 V Supply. PVDD12 provides a clean supply.
PLL Reference/Clock Input, Positive. When the PLL is used, this pin is the positive reference clock input. When
the PLL is not used, this pin is the positive device clock input. This pin is self biased and must be ac-coupled.
PLL Reference/Clock Input, Negative. When the PLL is used, this pin is the negative reference clock input. When
the PLL is not used, this pin is the negative device clock input. This pin is self biased and must be ac-coupled.
1.2 V Supply. PVDD12 provides a clean supply.
Positive Reference Clock for Deterministic Latency. This pin is self biased for ac coupling. It can be ac-coupled or
dc-coupled.
Negative Reference Clock for Deterministic Latency. This pin is self biased for ac coupling. It can be ac-coupled or
dc-coupled.
1.2 V Supply. PVDD12 provides a clean supply.
1.2 V Supply. PVDD12 provides a clean supply.
1.2 V Supply. PVDD12 provides a clean supply.
1.2 V Supply. PVDD12 provides a clean supply.
Transmit Enable for DAC0 and DAC1. The CMOS levels are determined with respect to IOVDD.
Transmit Enable for DAC2 and DAC3. The CMOS levels are determined with respect to IOVDD.
1.2 V Digital Supply.
1.2 V Digital Supply.
Serial Channel Input 0, Positive. CML compliant. SERDIN0+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 0, Negative. CML compliant. SERDIN0− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 1, Positive. CML compliant. SERDIN1+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 1, Negative. CML compliant. SERDIN1− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Rev. A | Page 12 of 125
Data Sheet
Pin No.
20
21
22
23
24
25
26
Mnemonic
SVDD12
VTT
SVDD12
SYNCOUT0+
SYNCOUT0−
VTT
SERDIN2+
27
SERDIN2−
28
29
SVDD12
SERDIN3+
30
SERDIN3−
31
32
33
34
35
36
37
SVDD12
SVDD12
SVDD12
LDO_BYP1
SIOVDD33
SVDD12
SERDIN4−
38
SERDIN4+
39
40
SVDD12
SERDIN5−
41
SERDIN5+
42
43
44
45
46
47
48
VTT
SYNCOUT1−
SYNCOUT1+
SVDD12
VTT
SVDD12
SERDIN6−
49
SERDIN6+
50
51
SVDD12
SERDIN7−
52
SERDIN7+
53
54
55
56
57
58
59
60
61
DVDD12
GND
GND
PVDD12
PVDD12
PROTECT_OUT1
PROTECT_OUT0
IRQ
RESET
AD9144
Description
1.2 V JESD204B Receiver Supply.
1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins.
1.2 V JESD204B Receiver Supply.
Positive LVDS Sync (Active Low) Output Signal Channel Link 0.
Negative LVDS Sync (Active Low) Output Signal Channel Link 0.
1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins.
Serial Channel Input 2, Positive. CML compliant. SERDIN2+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 2, Negative. CML compliant. SERDIN2− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 3, Positive. CML compliant. SERDIN3+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 3, Negative. CML compliant. SERDIN3− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V JESD204B Receiver Supply.
1.2 V JESD204B Receiver Supply.
1.2 V JESD204B Receiver Supply.
LDO SERDES Bypass. This pin requires a 1 Ω resistor in series with a 1 μF capacitor to ground.
3.3 V Supply for SERDES.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 4, Negative. CML compliant. SERDIN4− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 4, Positive. CML compliant. SERDIN4+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 5, Negative. CML compliant. SERDIN5− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 5, Positive. CML compliant. SERDIN5+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins.
Negative LVDS Sync (Active Low) Output Signal Channel Link 1.
Positive LVDS Sync (Active Low) Output Signal Channel Link 1.
1.2 V JESD204B Receiver Supply.
1.2 V Termination Voltage. Connect VTT to the SVDD12 supply pins.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 6, Negative. CML compliant. SERDIN6− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 6, Positive. CML compliant. SERDIN6+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V JESD204B Receiver Supply.
Serial Channel Input 7, Negative. CML compliant. SERDIN7− is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
Serial Channel Input 7, Positive. CML compliant. SERDIN7+ is internally terminated to the VTT pin voltage
using a calibrated 50 Ω resistor. This pin is ac-coupled only.
1.2 V Digital Supply.
Ground. Connect GND to the ground plane.
Ground. Connect GND to the ground plane.
1.2 V Supply. PVDD12 provides a clean supply.
1.2 V Supply. PVDD12 provides a clean supply.
Power Detection Protection Pin Output for DAC2 and DAC3. Pin 58 is high when power protection is in process.
Power Detection Protection Pin Output for DAC0 and DAC1. Pin 59 is high when power protection is in process.
Interrupt Request (Active Low, Open Drain).
Reset. This pin is active low. CMOS levels are determined with respect to IOVDD.
Rev. A | Page 13 of 125
AD9144
Pin No.
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
Mnemonic
SDO
SDIO
SCLK
CS
IOVDD
AVDD33
OUT3+
OUT3−
LDO24
CVDD12
LDO24
OUT2−
OUT2+
AVDD33
CVDD12
AVDD33
OUT1+
OUT1−
LDO24
CVDD12
LDO24
OUT0−
OUT0+
AVDD33
I120
CVDD12
LDO_BYP2
EPAD
Data Sheet
Description
Serial Port Data Output. CMOS levels are determined with respect to IOVDD.
Serial Port Data Input/Output. CMOS levels are determined with respect to IOVDD.
Serial Port Clock Input. CMOS levels are determined with respect to IOVDD.
Serial Port Chip Select. This pin is active low; CMOS levels are determined with respect to IOVDD.
IOVDD Supply for CMOS Input/Output and SPI. Operational for 1.8 V  IOVDD  3.3 V.
3.3 V Analog Supply for DAC Cores.
DAC3 Positive Current Output.
DAC3 Negative Current Output.
2.4 V LDO. Requires a 1 μF capacitor to ground.
1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 71.
2.4 V LDO. Requires a 1 μF capacitor to ground.
DAC2 Negative Current Output.
DAC2 Positive Current Output.
3.3 V Analog Supply for DAC Cores.
1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 76.
3.3 V Analog Supply for DAC Cores.
DAC1 Positive Current Output.
DAC1 Negative Current Output.
2.4 V LDO. Requires a 1 μF capacitor to ground.
1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 81.
2.4 V LDO. Requires a 1 μF capacitor to ground.
DAC0 Negative Current Output.
DAC0 Positive Current Output.
3.3 V Analog Supply for DAC Cores.
Output Current Generation Pin for DAC Full-Scale Current. Tie a 4 kΩ resistor from the I120 pin to ground.
1.2 V Clock Supply. Place bypass capacitors as near as possible to Pin 87.
LDO Clock Bypass for DAC PLL. This pin requires a 1 Ω resistor in series with a 1 μF capacitor to ground.
Exposed Pad. The exposed pad must be securely connected to the ground plane.
Rev. A | Page 14 of 125
Data Sheet
AD9144
TERMINOLOGY
Integral Nonlinearity (INL)
INL is the maximum deviation of the actual analog output from
the ideal output, determined by a straight line drawn from zero
scale to full scale.
Differential Nonlinearity (DNL)
DNL is the measure of the variation in analog value, normalized
to full scale, associated with a 1 LSB change in digital input code.
Offset Error
Offset error is the deviation of the output current from the ideal
of 0 mA. For OUTx+, 0 mA output is expected when all inputs
are set to 0. For OUTx−, 0 mA output is expected when all
inputs are set to 1.
Gain Error
Gain error is the difference between the actual and ideal output
span. The actual span is determined by the difference between
the output when the input is at its minimum code and the
output when the input is at its maximum code.
Output Compliance Range
The output compliance range is the range of allowable voltages
at the output of a current output DAC. Operation beyond the
maximum compliance limits can cause either output stage
saturation or breakdown, resulting in nonlinear performance.
Temperature Drift
Offset drift is a measure of how far from full-scale range (FSR)
the DAC output current is at 25°C (in ppm). Gain drift is a
measure of the slope of the DAC output current across its full
ambient operating temperature range, TA, (in ppm/°C).
Power Supply Rejection (PSR)
PSR is the maximum change in the full-scale output as the supplies
are varied from minimum to maximum specified voltages.
Settling Time
Settling time is the time required for the output to reach and
remain within a specified error band around its final value,
measured from the start of the output transition.
Spurious-Free Dynamic Range (SFDR)
SFDR is the difference, in decibels, between the peak amplitude
of the output signal and the peak spurious signal within the dc
to Nyquist frequency of the DAC. Typically, energy in this band
is rejected by the interpolation filters. This specification,
therefore, defines how well the interpolation filters work and
the effect of other parasitic coupling paths on the DAC output.
Signal-to-Noise Ratio (SNR)
SNR is the ratio of the rms value of the measured output signal
to the rms sum of all other spectral components below the Nyquist
frequency, excluding the first six harmonics and dc. The value
for SNR is expressed in decibels.
Interpolation Filter
If the digital inputs to the DAC are sampled at a multiple rate of
fDATA (interpolation rate), a digital filter can be constructed that
has a sharp transition band near fDATA/2. Images that typically
appear around fDAC (output data rate) can be greatly suppressed.
Adjacent Channel Leakage Ratio (ACLR)
ACLR is the ratio in decibels relative to the carrier (dBc)
between the measured power within a channel relative to
its adjacent channel.
Complex Image Rejection
In a traditional two part upconversion, two images are created
around the second IF frequency. These images have the effect
of wasting transmitter power and system bandwidth. By placing
the real part of a second complex modulator in series with the
first complex modulator, either the upper or lower frequency
image near the second IF can be rejected.
Adjusted DAC Update Rate
The adjusted DAC update rate is defined as the DAC update
rate divided by the smallest interpolating factor. For clarity on
DACs with multiple interpolating factors, the adjusted DAC
update rate for each interpolating factor may be given.
Physical Lane
Physical Lane x refers to SERDINx±.
Logical Lane
Logical Lane x refers to physical lanes after optionally being
remapped by the crossbar block (Register 0x308 to
Register 0x30B).
Link Lane
Link Lane x refers to logical lanes considered per link. When
paging Link 0 (Register 0x300[2] = 0), Link Lane x = Logical
Lane x. When paging Link 1 (Register 0x300[2] = 1, dual-link
only), Link Lane x = Logical Lane x + 4.
Rev. A | Page 15 of 125
AD9144
Data Sheet
TYPICAL PERFORMANCE CHARACTERISTICS
0
–20
–40
–60
–80
–60
200
300
400
500
fOUT (MHz)
–100
200
300
400
500
Figure 7. Single-Tone SFDR vs. fOUT in the First Nyquist Zone
over Digital Back Off, fDAC = 983 MHz
0
fDAC = 1966MHz
fDAC = 2456MHz
MEDIAN
–20
100
fOUT (MHz)
Figure 4. Single-Tone SFDR vs. fOUT in the First Nyquist Zone,
fDAC = 983 MHz, 1228 MHz, and 1474 MHz
0
0
11675-107
100
11675-104
0
0dBFS
–6dBFS
–9dBFS
–12dBFS
–20
–40
SFDR (dBc)
–60
–80
–40
–60
–100
100
200
300
400
500
fOUT (MHz)
–100
11675-305
0
0
100
200
300
400
500
fOUT (MHz)
Figure 5. Single-Tone SFDR vs. fOUT in the First Nyquist Zone,
fDAC = 1966 MHz and 2456 MHz
11675-108
–80
Figure 8. Single-Tone SFDR vs. fOUT in the First Nyquist Zone
over Digital Back Off, fDAC = 1966 MHz
0
0
IN-BAND SECOND HARMONIC
IN-BAND THIRD HARMONIC
MAX DIGITAL SPUR
–20
fDAC = 983MHz
fDAC = 1228MHz
fDAC = 1474MHz
–20
IMD3 (dBc)
–40
–60
–80
–40
–60
–80
0
100
200
300
fOUT (MHz)
400
500
11675-106
–100
Figure 6. Single-Tone Second and Third Harmonics and Maximum Digital Spur
in the First Nyquist Zone, fDAC = 1966 MHz, 0 dB Back Off
Rev. A | Page 16 of 125
–100
0
100
200
300
400
fOUT (MHz)
Figure 9. Two-Tone Third IMD (IMD3) vs. fOUT,
fDAC = 983 MHz, 1228 MHz, and 1474 MHz
500
11675-109
SFDR (dBc)
–40
–80
–100
SFDR (dBc)
0dBFS
–6dBFS
–9dBFS
–12dBFS
–20
SFDR (dBc)
SFDR (dBc)
0
fDAC = 983MHz
fDAC = 1228MHz
fDAC = 1474MHz
Data Sheet
0
fDAC = 1966MHz
fDAC = 2456MHz
–20
–20
–40
–40
IMD3 (dBc)
–60
–80
1MHz TONE SPACING
16MHz TONE SPACING
35MHz TONE SPACING
–60
–80
0
100
200
300
400
500
fOUT (MHz)
Figure 10. Two-Tone Third IMD (IMD3) vs. fOUT,
fDAC = 1966 MHz and 2456 MHz
0
–100
11675-110
–100
fDAC = 983MHz
fDAC = 1966MHz
100
200
300
400
500
fOUT (MHz)
Figure 13. Two-Tone Third IMD (IMD3) vs. fOUT over Tone Spacing at
0 dB Back Off, fDAC = 983 MHz and 1966 MHz
–130
0dBFS
–6dBFS
–9dBFS
–12dBFS
–20
0
11675-113
IMD3 (dBc)
0
AD9144
fDAC = 983MHz
fDAC = 1228MHz
fDAC = 1474MHz
–135
NSD (dBm/Hz)
IMD3 (dBc)
–140
–40
–60
–145
–150
–155
–160
–80
100
200
300
400
500
fOUT (MHz)
–170
0
300
400
500
500
Figure 14. Single-Tone (0 dBFS) NSD vs. fOUT,
fDAC = 983 MHz, 1228 MHz, and 1474 MHz
–130
0dBFS
–6dBFS
–9dBFS
–12dBFS
–20
200
fOUT (MHz)
Figure 11. Two-Tone Third IMD (IMD3) vs. fOUT over Digital Back Off,
fDAC = 983 MHz, Each Tone Is at −6 dBFS
0
100
11675-114
0
11675-111
–100
11675-115
–165
fDAC = 1966MHz
fDAC = 2456MHz
–135
NSD (dBm/Hz)
IMD3 (dBc)
–140
–40
–60
–145
–150
–155
–160
–80
–100
0
100
200
300
fOUT (MHz)
400
500
11675-112
–165
Figure 12. Two-Tone Third IMD (IMD3) vs. fOUT over Digital Back Off,
fDAC = 1966 MHz, Each Tone Is at −6 dBFS
Rev. A | Page 17 of 125
–170
0
100
200
300
400
fOUT (MHz)
Figure 15. Single-Tone (0 dBFS) NSD vs. fOUT,
fDAC = 1966 MHz and 2456 MHz
AD9144
–130
Data Sheet
–60
0dBFS
–6dBFS
–9dBFS
–12dBFS
–135
–80
PHASE NOISE (dBc/Hz)
–140
NSD (dBm/Hz)
fOUT = 30MHz
fOUT = 200MHz
fOUT = 400MHz
–145
–150
–155
PLL: OFF
PLL: ON
–100
–120
–140
–160
–160
–180
0
100
200
300
400
500
fOUT (MHz)
11675-116
–170
100
1k
10k
100k
1M
10M
OFFSET FREQUENCY (Hz)
Figure 19. Single-Tone Phase Noise vs. Offset Frequency over fOUT,
fDAC = 2.0 GHz, PLL On and Off
Figure 16. Single-Tone NSD vs. fOUT over Digital Back Off,
fDAC = 983 MHz
–130
10
0dBFS
–6dBFS
–9dBFS
–12dBFS
–135
NSD (dBm/Hz)
–140
–145
–150
–155
–160
0
100
200
300
400
500
fOUT (MHz)
11675-117
–170
Figure 20. 1C WCDMA ACLR, fOUT = 30 MHz,
fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz
Figure 17. Single-Tone NSD vs. fOUT over Digital Back Off,
fDAC = 1966 MHz
–130
PLL OFF
PLL ON
–135
fDAC = 983MHz
fDAC = 1966MHz
–140
–145
–150
–155
–160
–170
0
100
200
300
400
fOUT (MHz)
500
Figure 18. Single-Tone NSD (0 dBFS) vs. fOUT,
fDAC = 983 MHz and 1966 MHz, PLL On and Off
Rev. A | Page 18 of 125
11675-316
–165
11675-118
NSD (dBm/Hz)
11675-315
–165
Figure 21. 1C WCDMA ACLR, fOUT = 122 MHz,
fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz
11675-119
–165
AD9144
11675-317
11675-320
Data Sheet
Figure 22. 4C WCDMA ACLR, fOUT = 30 MHz,
fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz
Figure 25. 4C WCDMA ACLR, fOUT = 245 MHz,
fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 245 MHz
1800
1×
2×
4×
8×
POWER CONSUMPTION (mW)
1700
1600
1500
1400
1300
1200
1000
0
500
1000
1500
2000
2500
fDAC (MHz)
Figure 23. 4C WCDMA ACLR, fOUT = 122 MHz,
fDAC = 983 MHz, 2× Interpolation, PLL Frequency = 122 MHz
11675-326
11675-318
1100
Figure 26. Total Power Consumption vs. fDAC over Interpolation, 8 SERDES Lanes
Enabled, 4 DACs Enabled, NCO, Digital Gain, Inverse Sinc and DAC PLL Disabled
120
POWER CONSUMPTION (mW)
100
NCO
PLL (fDAC /fREF RATIO:4)
DIGITAL GAIN
INVERSE SINC
80
60
40
0
200
400
600
800
1000
1200
1400
1600
fDAC (MHz)
Figure 24. 4C WCDMA ACLR, fOUT = 30 MHz,
fDAC = 1966 MHz, 4× Interpolation, PLL Frequency = 245 MHz
Figure 27. Power Consumption vs. fDAC over Digital Functions
Rev. A | Page 19 of 125
11675-327
11675-319
20
AD9144
700
Data Sheet
2 LANES
4 LANES
8 LANES
350
1.2V SVDD12 SUPPLY
1.3V SVDD12 SUPPLY
300
SUPPLY CURRENT (mA)
500
400
300
200
250
200
150
100
50
1
2
3
4
5
LANE RATE (Gbps)
6
7
8
0
200
11675-328
100
1.2V SUPPLY
1.3V SUPPLY
3.3V SUPPLY
Figure 28. SVDD12 Current vs. Lane Rate over Number of SERDES Lanes and
Supply Voltage Setting
400
600
800
100
fDAC (MHz)
1200
1400
1600
11675-329
SVDD12 CURRENT (mA)
600
DVDD12
CVDD12
PVDD12
AVDD33
Figure 29. DVDD12, CVDD12, PVDD12, and AVDD33 Supply Current vs. fDAC
over Supply Voltage Setting, 4 DACs Enabled
Rev. A | Page 20 of 125
Data Sheet
AD9144
THEORY OF OPERATION
The AD9144 is a 16-bit, quad DAC with a SERDES interface.
Figure 2 shows a detailed functional block diagram of the
AD9144. Eight high speed serial lanes carry data at a maximum
speed of 10.6 Gbps, and a 1.06 GSPS input data rate to the DACs.
Compared to either LVDS or CMOS interfaces, the SERDES
interface simplifies pin count, board layout, and input clock
requirements to the device.
The clock for the input data is derived from the device clock
(required by the JESD204B specification). This device clock can
be sourced with a PLL reference clock used by the on-chip PLL
to generate a DAC clock or a high fidelity direct external DAC
sampling clock. The device can be configured to operate in one-,
two-, four-, or eight-lane modes, depending on the required
input data rate. To add application flexibility, the quad DAC can
be configured as a dual-link device with each JESD204B link
providing data for a dual DAC pair.
The digital datapath of the AD9144 offers four interpolation
modes (1×, 2×, 4×, and 8×) through three half-band filters with
a maximum DAC sample rate of 2.8 GSPS. An inverse sinc filter
is provided to compensate for sinc related roll-off.
The AD9144 DAC cores provide a fully differential current
output with a nominal full-scale current of 20 mA. The full-scale
current, IOUTFS, is user adjustable to between 13.9 mA and 27.0 mA,
typically. The differential current outputs are complementary and
are optimized for easy integration with the Analog Devices
ADRF6720 AQM. The AD9144 is capable of multichip
synchronization that can both synchronize multiple DACs and
establish a constant and deterministic latency (latency locking)
path for the DACs. The latency for each of the DACs remains
constant from link establishment to link establishment. An
external alignment (SYSREF±) signal makes the AD9144
Subclass 1 compliant. Several modes of SYSREF± signal
handling are available for use in the system.
An SPI configures the various functional blocks and monitors
their statuses. The various functional blocks and the data
interface must be set up in a specific sequence for proper operation
(see the Device Setup Guide section). Simple SPI initialization
routines set up the JESD204B link and are included in the
evaluation board package. The following sections describe the
various blocks of the AD9144 in greater detail. Descriptions of
the JESD204B interface, control parameters, and various registers
to set up and monitor the device are provided. The recommended
start-up routine reliably sets up the data link.
Rev. A | Page 21 of 125
AD9144
Data Sheet
SERIAL PORT OPERATION
The serial port is a flexible, synchronous serial communications
port that allows easy interfacing with many industry-standard
microcontrollers and microprocessors. The serial input/output
(I/O) is compatible with most synchronous transfer formats,
including both the Motorola SPI and Intel® SSR protocols. The
interface allows read/write access to all registers that configure
the AD9144. MSB first or LSB first transfer formats are supported.
The serial port interface can be configured as a 4-wire interface
or a 3-wire interface in which the input and output share a singlepin I/O (SDIO).
SDO 62
SCLK 64
CS 65
SPI
PORT
Serial Clock (SCLK)
Figure 30. Serial Port Interface Pins
There are two phases to a communication cycle with the AD9144.
Phase 1 is the instruction cycle (the writing of an instruction
byte into the device), coincident with the first 16 SCLK rising
edges. The instruction word provides the serial port controller
with information regarding the data transfer cycle, Phase 2 of
the communication cycle. The Phase 1 instruction word defines
whether the upcoming data transfer is a read or write, along with
the starting register address for the following data transfer.
A logic high on the CS pin followed by a logic low resets the
serial port timing to the initial state of the instruction cycle.
From this state, the next 16 rising SCLK edges represent the
instruction bits of the current I/O operation.
The remaining SCLK edges are for Phase 2 of the communication
cycle. Phase 2 is the actual data transfer between the device and
the system controller. Phase 2 of the communication cycle is a
transfer of one or more data bytes. Eight × N SCLK cycles are
needed to transfer N bytes during the transfer cycle. Registers
change immediately upon writing to the last bit of each transfer
byte, except for the frequency tuning word (FTW) and numerically
controlled oscillator (NCO) phase offsets, which change only when
the frequency tuning word FTW_UPDATE_REQ bit is set.
DATA FORMAT
The instruction byte contains the information shown in Table 13.
Table 13. Serial Port Instruction Word
I[15] (MSB)
R/W
I[14:0]
A[14:0]
A14 to A0, Bit 14 to Bit 0 of the instruction word, determine the
register that is accessed during the data transfer portion of the
communication cycle. For multibyte transfers, A[14:0] is the
starting address. The remaining register addresses are generated
by the device based on the ADDRINC bit. If ADDRINC is set
high (Register 0x000, Bit 5 and Bit 2), multibyte SPI writes start
on A[14:0] and increment by 1 every 8 bits sent/received. If
ADDRINC is set to 0, the address decrements by 1 every 8 bits.
SERIAL PORT PIN DESCRIPTIONS
11675-044
SDIO 63
R/W, Bit 15 of the instruction word, determines whether a read
or a write data transfer occurs after the instruction word write.
Logic 1 indicates a read operation, and Logic 0 indicates a write
operation.
The serial clock pin synchronizes data to and from the device
and runs the internal state machines. The maximum frequency
of SCLK is 10 MHz. All data input is registered on the rising edge
of SCLK. All data is driven out on the falling edge of SCLK.
Chip Select (CS)
An active low input starts and gates a communication cycle.
CS allows more than one device to be used on the same serial
communications lines. The SDIO pin goes to a high impedance
state when this input is high. During the communication cycle,
chip select must stay low.
Serial Data I/O (SDIO)
This pin is a bidirectional data line. In 4-wire mode, this pin
acts as the data input, and SDO acts as the data output.
SERIAL PORT OPTIONS
The serial port can support both MSB first and LSB first data
formats. This functionality is controlled by the LSBFIRST bit
(Register 0x000, Bit 6 and Bit 1). The default is MSB first
(LSBFIRST = 0).
When LSBFIRST = 0 (MSB first), the instruction and data bits
must be written from MSB to LSB. R/W is followed by A[14:0]
as the instruction word, and D[7:0] is the data-word. When
LSBFIRST = 1 (LSB first), the opposite is true. A[0:14] is
followed by R/W, which is subsequently followed by D[0:7].
The serial port supports a 3-wire or 4-wire interface. When
SDOACTIVE = 1 (Register 0x000, Bit 4 and Bit 3), a 4-wire
interface with a separate input pin (SDIO) and output pin
(SDO) is used. When SDOACTIVE = 0, the SDO pin is unused
and the SDIO pin is used for both input and output.
Rev. A | Page 22 of 125
Data Sheet
AD9144
DATA TRANSFER CYCLE
SDIO
R/W A14 A13
A3
A2 A1
A0 D7N D6 N D5N
D30 D20 D1 0 D00
Figure 31. Serial Register Interface Timing, MSB First, ADDRINC = 0
INSTRUCTION CYCLE
DATA TRANSFER CYCLE
CS
SCLK
SDIO
A0
A1 A2
A12 A13 A14 R/W D00 D10 D20
D4 N D5N D6N D7N
Figure 32. Serial Register Interface Timing, LSB First, ADDRINC = 1
CS
SCLK
tDV
SDIO
DATA BIT n
DATA BIT n – 1
Figure 33. Timing Diagram for Serial Port Register Read
tSCS
tHCS
CS
tPWH
tPWL
SDIO
tDH
INSTRUCTION BIT 15
INSTRUCTION BIT 14
INSTRUCTION BIT 0
Figure 34. Timing Diagram for Serial Port Register Write
Rev. A | Page 23 of 125
11675-047
SCLK
tDS
11675-045
SCLK
11675-046
To prevent confusion and to ensure consistency between devices,
the chip tests the first nibble following the address phase, ignoring
the second nibble. This is completed independently from the LSB
first bit and ensures that there are extra clock cycles following
the soft reset bits (Register 0x000, Bit 0 and Bit 7). This only
applies when writing to Register 0x000.
INSTRUCTION CYCLE
CS
11675-048
Multibyte data transfers can be performed as well. This is done
by holding the CS pin low for multiple data transfer cycles
(eight SCLKs) after the first data transfer word following the
instruction cycle. The first eight SCLKs following the
instruction cycle read from or write to the register provided in
the instruction cycle. For each additional eight SCLK cycles, the
address is either incremented or decremented and the read/write
occurs on the new register. The direction of the address can be set
using ADDRINC (Register 0x000, Bit 5 and Bit 2). When
ADDRINC is 1, the multicycle addresses are incremented.
When ADDRINC is 0, the addresses are decremented. A new
write cycle can always be initiated by bringing CS high and then
low again.
AD9144
Data Sheet
CHIP INFORMATION
Register 0x003 to Register 0x006 contain chip information, as shown in Table 14.
Table 14. Chip Information
Information
Chip Type
Product ID
Product Grade
Device Revision
Description
The product type is high speed DAC, which is represented by a code of 0x04 in Register 0x003.
8 MSBs in Register 0x005 and 8 LSBs in Register 0x004. The product ID is 0x9144.
Register 0x006[7:4]. The product grade is 0x00.
Register 0x006[3:0]. The device revision is 0x06.
Rev. A | Page 24 of 125
Data Sheet
AD9144
DEVICE SETUP GUIDE
OVERVIEW
The sequence of steps to properly set up the AD9144 is as follows:
1.
2.
3.
4.
5.
6.
Set up the SPI interface, power up necessary circuit blocks,
make required writes to the configuration registers, and set
up the DAC clocks (see the Step 1: Start Up the DAC section).
Set the digital features of the AD9144 (see the Step 2:
Digital Datapath section).
Set up the JESD204B links (see the Step 3: Transport Layer
section).
Set up the physical layer of the SERDES interface (see the
Step 4: Physical Layer section).
Set up the data link layer of the SERDES interface (see the
Step 5: Data Link Layer section).
Check for errors (see the Step 6: Optional Error
Monitoring section).
Optionally, enable any needed features as described in the
Step 7: Optional Features section.
The registers in Table 16 must be written from their default
values to be the values listed in the table for the device to work
correctly. These registers must be written after any soft reset,
hard reset, or power-up occurs.
Table 16. Required Device Configurations
Addr.
0x12D
0x146
0x2A4
0x232
0x333
Value
0x8B
0x01
0xFF
0xFF
0x01
If using the optional DAC PLL, also set the registers in Table 17.
Table 17. Optional DAC PLL Configuration Procedure
Addr.
0x087
Value1
0x62
0x088
0xC9
0x089
0x0E
0x08A
0x12
0x08D
0x7B
0x1B0
0x00
0x1B9
0x24
STEP 1: START UP THE DAC
0x1BC
0x0D
This section describes how to set up the SPI interface, power up
necessary circuit blocks, write required configuration registers,
and set up the DAC clocks, as listed in Table 15.
0x1BE
0x02
0x1BF
0x8E
0x1C0
0x2A
0x1C1
0x2A
0x1C4
0x7E
0x08B
0x08C
0x085
Various
0x
0x
0x
0x
0x083
0x10
7.
The register writes listed in Table 15 to Table 21 give the register
writes necessary to set up the AD9144. Consider printing out
this setup guide and filling in the Value column with appropriate
variable values for the conditions of the desired application.
The notation 0x, shaded in gray, indicates register settings that
must be filled in by the user. To fill in the unknown register
values, select the correct settings for each variable listed in the
Variable column of Table 15 to Table 21. The Description
column describes how to set variables or provides a link to a
section where this is described.
Table 15. Power-Up and DAC Initialization Settings
Addr.
0x000
0x000
0x011
Bit
No.
7
[6:3]
2
0x080
0x081
1
1
Value
0xBD
0x3C
0x
0
Variable
PdDACs
0
0x
0x
PdClocks
PdSysref
Description
Soft reset.
Deassert reset, set 4-wire SPI.
Power up band gap.
PdDACs = 0 if all 4 DACs are
being used. If not, see the DAC
Power-Down Setup section.
Power up master DAC.
PdClocks = 0 if all 4 DACs are
being used. If not, see the DAC
Power-Down Setup section.
PdSysref = 0x00 for Subclass 1.
PdSysref = 0x10 for Subclass 0.
See the Subclass Setup section
for details on subclass.
Description
Digital datapath configuration
Digital datapath configuration
Clock configuration
SERDES interface configuration
SERDES interface configuration
1
Variable
LODivMode
RefDivMode
BCount
LookUpVals
Description
Optimal DAC PLL loop filter
settings
Optimal DAC PLL loop filter
settings
Optimal DAC PLL loop filter
settings
Optimal DAC PLL charge pump
settings
Optimal DAC LDO settings for
DAC PLL
Power DAC PLL blocks when
power machine is disabled
Optimal DAC PLL charge pump
settings
Optimal DAC PLL VCO control
settings
Optimal DAC PLL VCO power
control settings
Optimal DAC PLL VCO calibration
settings
Optimal DAC PLL lock counter
length setting
Optimal DAC PLL charge pump
setting
Optimal DAC PLL varactor
settings
See the DAC PLL Setup section
See the DAC PLL Setup section
See the DAC PLL Setup section
See Table 25 in the DAC PLL Setup
section for the list of register
addresses and values for each.
Enable DAC PLL2
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the appropriate register
value.
2
Verify that Register 0x084[1] reads back 1 after enabling the DAC PLL to
indicate that the DAC PLL has locked.
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the appropriate register value.
Rev. A | Page 25 of 125
AD9144
Data Sheet
STEP 2: DIGITAL DATAPATH
Table 19. Transport Layer Settings
This section describes which interpolation filters to use and
how to set the data format being used. Additional digital
features are available including fine and coarse modulation,
digital gain scaling, and an inverse sinc filter used to improve
pass-band flatness. Table 22 provides further details on the
feature blocks available.
Addr.
0x200
0x201
Addr.
0x112
Bit No.
0x110
Variable
InterpMode
Description
Select interpolation
mode; see the
Interpolation section.
DataFmt
DataFmt = 0 if twos
complement;
DataFmt = 1 if unsigned
binary.
0x
7
1
Value
0x
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the appropriate register value.
0x
0x300
Table 18. Digital Datapath Settings
1
Bit No. Value1 Variable
0x00
UnusedLanes
0x
6
CheckSumMode
3
DualLink
2
CurrentLink
0x450
0x
DID
0x451
0x
BID
0x452
0x
LID
0x453
This section describes how to set up the JESD204B links. The
parameters are determined by the desired JESD204B operating
mode. See the JESD204B Setup section for details.
0x454
0x
Scrambling
L − 12
F − 12
0x455
0x
K − 12
Table 19 shows the register settings for the transport layer.
If using dual-link mode, perform writes from Register 0x300 to
Register 0x47D with CurrentLink = 0 and then repeat the same
set of register writes with CurrentLink = 1 (Register 0x200 and
Register 0x201 need only be written once).
0x456
0x
M − 12
0x457
0x458
0x
0x
N − 12
5
Subclass
[4:0]
NP − 12
0x459
Set DID to match
the device ID sent
by the transmitter.
Set BID to match
the bank ID sent
by the transmitter.
Set LID to match
the lane ID sent
by the transmitter.
See the JESD204B
Setup section.
See the JESD204B
Setup section.
See the JESD204B
Setup section.
See the JESD204B
Setup section.
N = 16.
See the JESD204B
Setup section.
NP = 16.
0x
5
JESDVer
[4:0]
S − 12
0x45A
JESDVer = 1 for
JESD204B,
JESDVer = 0 for
JESD204A.
See the JESD204B
Setup section.
0x
7
[4:0]
HD
0x45D
0
0x
CF
Lane0Checksum
0x46C
0x
Lanes
0x476
0x
F
0x47D
0x
Lanes
1
See the JESD204B
Setup section for
details on these
variables.
0x
7
[4:0]
STEP 3: TRANSPORT LAYER
Description
Power up the
interface.
See the JESD204B
Setup section.
See the JESD204B
Setup section.
CF must equal 0.
See the JESD204B
Setup section.
Deskew lanes. See
the JESD204B
Setup section.
See the JESD204B
Setup section.
Enable lanes.
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the correct register value.
2
This JESD204B link parameter is programmed in n − 1 notation as noted. For
example, if the setup requires L = 8 (8 lanes per link), program L − 1 or 7 into
Register 0x453[4:0].
Rev. A | Page 26 of 125
Data Sheet
AD9144
STEP 4: PHYSICAL LAYER
STEP 5: DATA LINK LAYER
This section describes how to set up the physical layer of the
SERDES interface. In this section, the input termination settings
are configured along with the CDR sampling and SERDES PLL.
This section describes how to set up the data link layer of the
SERDES interface. This section deals with SYSREF processing,
setting deterministic latency, and establishing the link.
Table 20. Device Configurations and Physical Layer Settings
Table 21. Data Link Layer Settings
Addr.
0x2AA
0x2AB
0x2B1
0x2B2
0x2A7
0x2AE
0x314
0x230
Bit
No.
Value1
0xB7
0x87
0xB7
0x87
0x01
0x01
0x01
0x
5
[4:2]
Halfrate
0x2
1
0x206
0x206
0x289
2
[1:0]
OvSmp
0x00
0x01
0x
1
PLLDiv
0x284
0x285
0x286
0x287
0x62
0xC9
0x0E
0x12
0x28A
0x28B
0x7B
0x00
0x290
0x89
0x294
0x24
0x296
0x297
0x299
0x03
0x0D
0x02
0x29A
0x8E
0x29C
0x2A
0x29F
0x78
0x2A0
0x06
0x280
0x268
0x01
0x
[7:6]
[5:0]
1
Variable
EqMode
0x22
Description
SERDES interface termination
setting
SERDES interface termination
setting
Autotune PHY setting
Autotune PHY setting
SERDES SPI configuration
Set up CDR; see the SERDES
Clocks Setup section
SERDES PLL default
configuration
Set up CDR; see the SERDES
Clocks Setup section
Reset CDR
Release CDR reset
SERDES PLL configuration
Set CDR oversampling for
PLL; see the SERDES Clocks
Setup section
Optimal SERDES PLL loop filter
Optimal SERDES PLL loop filter
Optimal SERDES PLL loop filter
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO LDO
Optimal SERDES PLL
configuration
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO
Optimal SERDES PLL VCO
Optimal SERDES PLL
configuration
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL VCO
varactor
Enable SERDES PLL2
See the Equalization Mode
Setup section
Required value (default)
Bit
No.
Value1
0x
Variable
Subclass
0x304
0x
LMFCDel
0x305
0x
LMFCDel
0x306
0x
LMFCVar
0x307
0x
LMFCVar
0x03A
0x01
0x03A
0x81
0x03A
0xC1
Addr.
0x301
SYSREF±
Signal
0x308
to
0x30B
0x
XBarVals
0x334
0x
InvLanes
0x300
1
0x
6
3
2
CheckSumMode
DualLink
CurrentLink
[1:0]
EnLinks
Description
See the JESD204B
Setup section.
See the Link Latency
Setup section.
See the Link Latency
section.
See the Link Latency
Setup section.
See the Link Latency
Setup section.
Set sync mode =
one-shot sync; see
the Syncing LMFC
Signals section for
other sync options.
Enable the sync
machine.
Arm the sync
machine.
If Subclass = 1, ensure
that at least one
SYSREF± edge is sent
to the device.2
If remapping lanes,
set up crossbar; see
the Crossbar Setup
section.
Invert polarity of
desired logical lanes.
Bit x of InvLanes
must be a 1 for each
Logical Lane x to
invert.
Enable the links.
See the JESD204B
Setup section.
Set to 0 to access
Link 0 status or 1 for
Link 1 status
readbacks. See the
JESD204B Setup
section.
EnLinks = 3 if
DualLink = 1
(enables Link 0 and
Link 1);
EnLinks = 1 if
DualLink = 0 (enables
Link 0 only).
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the correct register value.
2
Verify that Register 0x03B[3] reads back 1 after sending at least one SYSREF±
edge to the device to indicate that the LMFC sync machine has properly locked.
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the correct register value.
2
Verify that Register 0x281[0] reads back 1 after enabling the SERDES PLL to
indicate that the SERDES PLL has locked.
Rev. A | Page 27 of 125
AD9144
Data Sheet
Table 22. Optional Features
STEP 6: OPTIONAL ERROR MONITORING
Feature
Digital
Modulation
Default
Off
STEP 7: OPTIONAL FEATURES
Inverse Sinc
On
There are a number of optional features that can be enabled.
Table 22 provides links to the sections describing each feature.
These features can be enabled during the digital datapath
configuration step or after the link is set up, because it is not
required to configure them for the link to be established, unlike
interpolation. Unless otherwise noted, these features are paged
as described in the Dual Paging section. Paging is particularly
important for dual specific settings like digital gain, phase adjust,
and dc offset.
Digital Gain
2.7 dB
Phase Adjust
Off
DC Offset
Off
Group Delay
0
Downstream
Protection
Off
Self Calibration
Off
For JESD204B error monitoring, see the JESD204B Error
Monitoring section. For other error checks, see the Interrupt
Request Operation section.
Rev. A | Page 28 of 125
Description
Modulates the data with a desired
carrier. See the Digital Modulation
section.
Improves pass-band flatness. See
the Inverse Sinc section.
Multiplies data by a factor. Can
compensate inverse sinc usage or
balance I/Q amplitude. See the
Digital Gain section.
Used to balance I/Q phase. See the
Phase Adjust section.
Used to cancel LO leakage. See the
DC Offset section.
Used to control overall latency. See
the Group Delay section.
Used to protect downstream
components. See the Downstream
Protection section.
Used to improve DAC linearity. Not
paged by the dual paging register.
See the Self Calibration section.
Data Sheet
AD9144
DAC PLL SETUP
INTERPOLATION
This section explains how to select the appropriate LODivMode,
RefDivMode, and BCount in the Step 1: Start Up the DAC
section. These parameters depend on the desired DAC clock
frequency (fDACCLK) and DAC reference clock frequency (fREF).
When using the DAC PLL, the reference clock signal is applied
to the CLK± differential pins (Pin 2 and Pin 3).
The transmit path can use zero to three cascaded interpolation
filters, which each provides a 2× increase in output data rate and
a low-pass function. Table 26 shows the different interpolation
modes and the respective usable bandwidth along with the
maximum fDATA rate attainable.
Table 23. DAC PLL LODivMode Settings
DAC Frequency Range (MHz)
1500 to 2800
750 to 1500
420 to 750
LODivMode,
Register 0x08B[1:0]
1
2
3
Table 24. DAC PLL RefDivMode Settings
DAC PLL Reference
Frequency (fREF) (MHz)
35 to 80
80 to 160
160 to 320
320 to 640
640 to 1000
Divide by
(RefDivFactor)
1
2
4
8
16
RefDivMode,
Register 0x08C[2:0]
0
1
2
3
4
The VCO frequency (fVCO) is related to the DAC clock frequency
according to the following equation:
fVCO = fDACCLK × 2LODivMode + 1
where 6 GHz  fVCO  12 GHz.
BCount must be between 6 and 127 and is calculated based on
fDACCLK and fREF as follows:
BCount = floor((fDACCLK)/(2 × fREF/RefDivFactor))
where RefDivFactor = 2RefDivMode (see Table 24).
Table 25. VCO Control Lookup Table Reference
Register
0x1B5
Setting
0x08
0x09
0x09
Register
0x1BB
Setting
0x03
0x03
0x13
Interpolation
Mode
1× (bypass)
InterpMode
0x00
Usable
Bandwidth
0.5 × fDATA
2×
0x01
0.4 × fDATA
4×
8×
0x03
0x04
0.4 × fDATA
0.4 × fDATA
Max fDATA (MHz)
1060 (SERDES
limited)
1060 (SERDES
limited)
700
350
The usable bandwidth is defined for 1×, 2×, 4×, and 8× modes
as the frequency band over which the filters have a pass-band
ripple of less than ±0.001 dB and an image rejection of greater
than 85 dB. For more information, see the Interpolation Filters
section.
JESD204B SETUP
This section explains how to select a JESD204B operating mode
for a desired application. This section defines appropriate values
for CheckSumMode, UnusedLanes, DualLink, CurrentLink,
Scrambling, L, F, K, M, N, NP, Subclass, S, HD, Lane0Checksum,
and Lanes needed for the Step 3: Transport Layer section.
Note that DualLink, Scrambling, L, F, K, M, N, NP, S, HD, and
Subclass must be set the same on the transmit side.
For a summary of how a JESD204B system works and what each
parameter means, see the JESD204B Serial Data Interface section.
Available Operating Modes
Finally, to finish configuring the DAC PLL, set the VCO control
registers up as described in Table 25 based on the VCO
frequency (fVCO). Write the registers listed in the table with the
corresponding LookUpVals.
VCO Frequency
Range (GHz)
fVCO < 6.3
6.3 ≤ fVCO < 7.25
fVCO ≥ 7.25
Table 26. Interpolation Modes and Their Usable Bandwidth
Register
0x1C5
Setting
0x07
0x06
0x06
For more information on the DAC PLL, see the DAC Input
Clock Configurations section.
Table 27. JESD204B Operating Modes (Single-Link Only)
Parameter
M (Converter Count)
L (Lane Count)
S ((Samples per Converter) per Frame)
F ((Octets per Frame) per Lane)
0
4
8
1
1
1
4
8
2
2
Mode
2
4
4
1
2
3
4
2
1
4
Table 28. JESD204B Operating Modes (Single- or Dual-Link)
Parameter
M (Converter Count)
L (Lane Count)
S ((Samples per Converter) per Frame)
F ((Octets per Frame) per Lane)
Rev. A | Page 29 of 125
4
2
4
1
1
5
2
4
2
2
Mode
6 7
2 2
2 1
1 1
2 4
9
1
2
1
1
10
1
1
1
2
AD9144
Data Sheet
For a particular application, the number of converters to use (M)
and the fDATA (DataRate) are known. The LaneRate and number
of lanes (L) can be traded off as follows:
DataRate = (DACRate)/(InterpolationFactor)
LaneRate = (20 × DataRate × M)/L
where LaneRate is between 1.44 Gbps and 10.64 Gbps.
Octets per frame per lane (F) and samples per convertor per
frame (S) define how the data is packed. If F = 1, the high density
setting must be set to one (HD = 1). Otherwise, set HD = 0.
Converter resolution and bits per sample (N and NP) must both be
set to 16. Frames per multiframe (K) must be set to 32 for Mode 0,
Mode 4 and Mode 9. Other modes can use either K = 16 or K = 32.
DualLink
DualLink sets up two independent JESD204B links, which allows
each link to be reset independently. If this functionality is desired,
set DualLink to 1; if a single link is desired, set DualLink to 0.
Note that Link 0 and Link 1 must have identical parameters.
The operating modes available when using dual-link mode are
shown in Table 28. In addition to these operating modes, the
modes in Table 28 can also be used when using single-link mode.
Scrambling
Scrambling is a feature that makes the spectrum of the link data
independent. This avoids spectral peaking and provides some
protection against data dependent errors caused by frequency
selective effects in the electrical interface. Set to 1 if scrambling
is being used, or to 0 if it is not.
Subclass
Subclass determines whether the latency of the device is
deterministic, meaning it requires an external synchronization
signal. See the Subclass Setup section for more information.
CurrentLink
Set CurrentLink to either 0 or 1 depending on whether Link 0
or Link 1, respectively, needs to be configured.
Lanes
Lanes is used to enable and deskew particular lanes in two
thermometer coded registers.
Lanes = (2L) − 1.
CheckSumMode
CheckSumMode must match the checksum mode used on the
transmit side. If the checksum used is the sum of fields in the
link configuration table, CheckSumMode = 0. If summing the
registers containing the packed link configuration fields,
CheckSumMode = 1. For more information on the how to
calculate the two checksum modes, see the Lane0Checksum
section.
Lane0Checksum
Lane0Checksum can be used for error checking purposes
to ensure that the transmitter is set up as expected. Both
CheckSumMode calculations use the fields contained in
Register 0x450 to Register 0x45A. Select whether to sum by
fields or by registers, matching the setting on the transmitter.
If CheckSumMode = 0, the summation is computed by fields.
The checksum is the lower 8 bits of the sum of the DID,
ADJCNT, BID, ADJDIR, PHADJ, LID, Scrambling, L – 1, F − 1,
K − 1, M − 1, CS, N − 1, Subclass, NP − 1, JESDVer, S − 1, HD,
and CF variables.
If CheckSumMode = 1, the summation is computed by registers.
The checksum is the sum of Register 0x450 to Register 0x45A,
Modulo 256.
DAC Power-Down Setup
As described in the Step 1: Start Up the DAC section, PdDACs
must be set to 0 if all 4 converters are being used. If fewer than
four converters are being used, the unused converters must be
powered down. Table 29 can be used to determine which DACs
are powered down based on the number of converters per link
(M) and whether the device is in DualLink mode.
Table 29. DAC Power-Down Configuration Settings
M (Converters
per link)
1
1
2
2
4
DualLink
0
1
0
1
0
DACs to Power Down
0
1
2
3
0
1
1
1
0
1
0
1
0
0
1
1
0
0
0
0
0
0
0
0
PdDACs
0b0111
0b0101
0b0011
0b0000
0b0000
PdClocks
UnusedLanes
UnusedLanes is used to turn off unused circuit blocks to save
power. Each physical lane that is not being used (SERDINx±)
must be powered off by writing a 1 to the corresponding bit of
Register 0x201.
If both DACs in DAC Dual B (DAC2 and DAC3) are powered
down, the clock for DAC Dual B can be powered down. In this
case, PdClocks = 0x40; if not, PdClocks = 0x00.
For example, if using Mode 6 in dual-link mode and sending
data on SERDIN0±, SERDIN1±, SERDIN4±, and SERDIN5±,
set UnusedLanes = 0xCC to power off Physical Lane 2, Lane 3,
Lane 6, and Lane 7.
Rev. A | Page 30 of 125
Data Sheet
AD9144
SERDES CLOCKS SETUP
This section describes how to select the appropriate Halfrate,
OvSmp, and PLLDiv settings in the Step 4: Physical Layer
section. These parameters depend solely on the lane rate (the
lane rate is established in the JESD204B Setup section).
Table 30. SERDES Lane Rate Configuration Settings
Lane Rate (Gbps)
1.44 to 2.76
2.88 to 5.52
5.75 to 10.64
Halfrate
0
0
1
OvSmp
1
0
0
PLLDiv
2
1
0
Delays that are not in PClock cycles must be converted before
they are used.
Some useful internal relationships are defined as follows:
PClockPeriod = 40/LaneRate
The PClockPeriod can be used to convert from time to PClock
cycles when needed.
PClockFactor = 4/F (frames per PClock)
The PClockFactor is used to convert from units of PClock
cycles to frame clock cycles, which is needed to set LMFCDel in
Subclass 1.
Halfrate and OvSmp set how the clock detect and recover
(CDR) circuit sample. See the SERDES PLL section for an
explanation of how that circuit blocks works and the role of
PLLDiv in the block.
PClocksPerMF = K/PClockFactor (PClocks per LMFC cycle)
where PClocksPerMF is the number or PClock cycles in a
multiframe cycle.
EQUALIZATION MODE SETUP
Set EqMode = 1 for a low power setting. Select this mode if the
insertion loss in the printed circuit board (PCB) is less than
12 dB. For insertion losses greater than 12 dB, but less than
17.5 dB, set EqMode = 0. More details can be found in the
Equalization section.
LINK LATENCY SETUP
The values for PClockFactor and PClockPerMF are given per
JESD mode in Table 31 and Table 32.
Table 31. PClockFactor and PClockPerMF per LMFC
JESD Mode ID
PClockFactor
PClockPerMF (K = 32)
PClockPerMF (K = 16)
1
This section describes the steps necessary to guarantee
multichip deterministic latency in Subclass 1 and to guarantee
synchronization of links within a device in Subclass 0. Use this
section to fill in LMFCDel, LMFCVar, and Subclass in the Step
5: Data Link Layer section. For more information, see the
Syncing LMFC Signals section.
0
4
8
N/A1
1
2
16
8
2
2
16
8
3
1
32
16
N/A means not applicable.
Table 32. PClockFactor and PClockPerMF per LMFC
JESD Mode ID
PClockFactor
PClockPerMF (K = 32)
PClockPerMF (K = 16)
Subclass Setup
1
The AD9144 supports JESD204B Subclass 0 and Subclass 1
operation.
With Known Delays
4
4
8
N/A1
5
2
16
8
6
2
16
8
7
1
32
16
9
4
8
N/A1
10
2
16
8
N/A means not applicable.
With information about all the system delays, LMFCVar and
LMFCDel can be calculated directly.
Subclass 1
This mode gives deterministic latency and allows links to be
synced to within ½ DAC clock periods. It requires an external
SYSREF± signal that is accurately phase aligned to the DAC clock.
Subclass 0
This mode does not require any signal on the SYSREF± pins
(the pins can be left disconnected).
Subclass 0 still requires that all lanes arrive within the same
LMFC cycle and that the dual DACs must be synchronized to
each other (they are synchronized to an internal clock instead of
to the SYSREF± signal).
RxFixed (the fixed receiver delay in PClock cycles) and RxVar
(the variable receiver delay in PClock cycles) can be found in
Table 8. TxFixed (the fixed transmitter delay in PClock cycles)
and TxVar (the variable receiver delay in PClock cycles) can be
found in the data sheet of the transmitter used. PCBFixed (the
fixed PCB trace delay in PClock cycles) can be extracted from
software; because this is generally much smaller than a PClock
cycle, it can also be omitted. For both the PCB and transmitter
delays, convert the delays into PClock cycles.
For each lane
MinDelayLane = floor(RxFixed + TxFixed + PCBFixed)
Set Subclass to 0 or 1 as desired.
MaxDelayLane = ceiling(RxFixed + RxVar + TxFixed +
TxVar + PCBFixed))
Link Delay Setup
LMFCVar and LMFCDel are used to impose delays such that all
lanes in a system arrive in the same LMFC cycle.
The unit used internally for delays is the period of the internal
processing clock (PClock), whose rate is 1/40th the lane rate.
where:
MinDelay is the minimum of all MinDelayLane values across
lanes, links, and devices.
MaxDelay is the maximum of all MaxDelayLane values across
lanes, links, and devices.
Rev. A | Page 31 of 125
AD9144
Data Sheet
For safety, add a guard band of 1 PClock cycle to each end of
the link delay as in the following equations:
LMFCVar = (MaxDelay + 1) − (MinDelay − 1)
Note that if LMFCVar must be more than 10, the AD9144 is
unable to tolerate the variable delay in the system.
For Subclass 1
LMFCDel = ((MinDelay − 1) × PClockFactor) % K
For Subclass 0
LMFCDel = (MinDelay − 1) % PClockPerMF
Program the same LMFCDel and LMFCVar across all links and
devices.
See the Link Delay Setup Example, with Known Delays section
for an example calculation.
Table 33. Register Configuration and Procedure for OneShot Sync
Addr.
0x301
0x03A
Bit.
No.
Value1
0x
0x01
0x03A
0x81
0x03A
0xC1
Variable
Subclass
SYSREF±
Signal
0x300
Without Known Delays
If comprehensive delay information is not available or known,
the AD9144 can read back the link latency between the LMFCRX
and the last arriving LMFC boundary in PClock cycles. This
information is then used to calculate LMFCVar and LMFCDel.
0x
6
CheckSumMode
3
DualLink
2
CurrentLink
[1:0]
EnLinks
For each link (on each device)
1.
2.
3.
4.
5.
Power up the board.
Follow the steps in Table 15 through Table 21 of the Device
Setup Guide.
Set the subclass and perform a sync. For one-shot sync,
perform the writes in Table 33. See the Syncing LMFC
Signals section for alternate sync modes.
Record DYN_LINK_LATENCY_0 (Register 0x302) as a
value of Delay for that link and power cycle.
Record DYN_LINK_LATENCY_1 (Register 0x303) as a
value of Delay for that link and power cycle the system.
Repeat Step 1 to Step 5 twenty times for each device in the
system. Keep a single list of the Delay values across all runs and
devices.
1
Description
Set subclass
Set sync mode to
one-shot sync
Enable the sync
machine
Arm the sync
machine
If Subclass = 1, ensure
that at least one
SYSREF± edge is sent
to the device.
Enable the links
See the JESD204B
Setup section
See the JESD204B
Setup section
Set to 0 to access
Link 0 status or 1
for Link 1 status
readbacks. See the
JESD204B Setup
section.
EnLinks = 3 if in
DualLink mode to
enable Link 0 and
Link 1; EnLinks = 1 if
not in DualLink mode
to enable Link 0
0x denotes a register value that the user must fill in. See the Variable and
Description columns for information on selecting the appropriate register value.
The list of delay values is used to calculate LMFCDel and
LMFCVar; however, first some of the delay values may need to
be remapped.
The maximum possible value for DYN_LINK_LATENCY_x
is one less than the number of PClocks in a multiframe
(PClocksPerMF). It is possible that a rollover condition may be
encountered, meaning the set of recorded Delay values might roll
over the edge of a multiframe. If so, Delay values may be near
both 0 and PClocksPerMF. If this occurs, add PClocksPerMF to
the set of values near 0.
For example, for Delay value readbacks of 6, 7, 0, and 1, the 0
and 1 Delay values must be remapped to 8 and 9, making the
new set of Delay values 6, 7, 8, and 9.
Rev. A | Page 32 of 125
Data Sheet
AD9144
Across power cycles, links, and devices
CROSSBAR SETUP


Register 0x308 to Register 0x30B allow arbitrary mapping of
physical lanes (SERDINx±) to logical lanes used by the SERDES
deframers.
MinDelay is the minimum of all Delay measurements
MaxDelay is the maximum of all Delay measurements
For safety, a guard band of 1 PClock cycle is added to each end
of the link delay and calculate LMFCVar and LMFCDel with
the following equation:
LMFCVar = (MaxDelay + 1) − (MinDelay − 1)
Note that if LMFCVar must be more than 10, the AD9144 is
unable to tolerate the variable delay in the system.
For Subclass 1
LMFCDel = ((MinDelay − 1) × PClockFactor) % K
For Subclass 0
LMFCDel = (MinDelay − 1) % PClockPerMF
Program the same LMFCDel and LMFCVar across all links and
devices.
See the Link Delay Setup Example, Without Known Delay
section for an example calculation.
Table 34. Crossbar Registers
Address
0x308
0x308
0x309
0x309
0x30A
0x30A
0x30B
0x30B
Bits
[2:0]
[5:3]
[2:0]
[5:3]
[2:0]
[5:3]
[2:0]
[5:3]
Logical Lane
LOGICAL_LANE0_SRC
LOGICAL_LANE1_SRC
LOGICAL_LANE2_SRC
LOGICAL_LANE3_SRC
LOGICAL_LANE4_SRC
LOGICAL_LANE5_SRC
LOGICAL_LANE6_SRC
LOGICAL_LANE7_SRC
Write each LOGICAL_LANEy_SRC with the number (x) of the
desired physical lane (SERDINx±) from which to obtain data.
By default, all logical lanes use the corresponding physical lane as
their data source. For example, by default LOGICAL_LANE0_
SRC = 0, meaning that Logical Lane 0 receives data from
Physical Lane 0 (SERDIN0±). If instead the user wants to use
SERDIN4± as the source for Logical Lane 0, the user must write
LOGICAL_LANE0_SRC = 4.
Rev. A | Page 33 of 125
AD9144
Data Sheet
JESD204B SERIAL DATA INTERFACE
Only certain combinations of parameters are supported. Each
supported combination is called a mode. In total, there are 10
single-link modes supported by the AD9144, as described in
Table 35. In dual-link mode, there are six supported modes, as
described in Table 36. Each of these tables shows the associated
clock rates when the lane rate is 10 Gbps.
JESD204B OVERVIEW
The AD9144 has eight JESD204B data ports that receive data.
The eight JESD204B ports can be configured as part of a single
JESD204B link or as part of two separate JESD204B links (duallink mode) that share a single system reference (SYSREF±) and
device clock (CLK±).
For a particular application, the number of converters to use
(M) and the DataRate are known. The LaneRate and number of
lanes (L) can be traded off as follows:
The JESD204B serial interface hardware consists of three layers:
the physical layer, the data link layer, and the transport layer.
These sections of the hardware are described in subsequent
sections, including information for configuring every aspect of
the interface. Figure 35 shows the communication layers
implemented in the AD9144 serial data interface to recover the
clock and deserialize, descramble, and deframe the data before it
is sent to the digital signal processing section of the device.
DataRate = (DACRate)/(InterpolationFactor)
LaneRate = (20 × DataRate × M)/L
where LaneRate must be between 1.44 Gbps and 10.64 Gbps.
Achieving and recovering synchronization of the lanes is very
important. To simplify the interface to the transmitter, the
AD9144 designates a master synchronization signal for each
JESD204B link. In single-link mode, SYNCOUT0± is used as the
master signal for all lanes; in dual-link mode, SYNCOUT0± is used
as the master signal for Link 0, and SYNCOUT1± is used as the
master signal for Link 1. If any lane in a link loses synchronization,
a resynchronization request is sent to the transmitter via the
synchronization signal of the link. The transmitter stops sending
data and instead sends synchronization characters to all lanes in
that link until resynchronization is achieved.
The physical layer is responsible for establishing a reliable channel
between the transmitter and the receiver, the data link layer is
responsible for unpacking the data into octets and descrambling
the data, and the transport layer receives the descrambled
JESD204B frames and converts them to DAC samples.
There are a number of JESD204B parameters (L, F, K, M, N, NP,
S, HD, and Scrambling) that define how the data is packed and
tell the device how to turn the serial data into samples. These
parameters are defined in detail in the Transport Layer section.
SYNCOUT0±
SYNCOUT1±
DATA LINK
LAYER
PHYSICAL
LAYER
SERDIN0±
TRANSPORT
LAYER
DUAL A Q DATA[15:0]
DESERIALIZER
TO
DAC
FRAME TO
SAMPLES
QBD/
DESCRAMBLER
SERDIN7±
DUAL A I DATA[15:0]
DUAL B I DATA[15:0]
DESERIALIZER
11675-004
DUAL B Q DATA[15:0]
SYSREF±
Figure 35. Functional Block Diagram of Serial Link Receiver
Table 35. Single-Link JESD204B Operating Modes
Parameter
M (Converter Counts)
L (Lane Counts)
S (Samples per Converter per Frame)
F (Octets per Frame per Lane)
Example Clocks for 10 Gbps Lane Rate
PClock (MHz)
Frame Clock (MHz)
Sample Clock (MHz)
0
4
8
1
1
1
4
8
2
2
2
4
4
1
2
3
4
2
1
4
4
2
4
1
1
250
1000
1000
250
500
1000
250
500
500
250
250
250
250
1000
1000
Rev. A | Page 34 of 125
Mode
5
2
4
2
2
250
500
1000
6
2
2
1
2
7
2
1
1
4
9
1
2
1
1
10
1
1
1
2
250
500
500
250
250
250
250
1000
1000
250
500
500
Data Sheet
AD9144
Table 36. Dual-Link JESD204B Operating Modes for Link 0 and Link 1
Mode
Parameter
M (Converter Counts)
L (Lane Counts)
S (Samples per Converter per Frame)
F (Octets/Frame per Lane)
Example Clock for 10 Gbps Lane Rate
PClock (MHz)
Frame Clock (MHz)
Sample Clock (MHz)
4
2
4
1
1
5
2
4
2
2
6
2
2
1
2
7
2
1
1
4
9
1
2
1
1
10
1
1
1
2
250
1000
1000
250
500
1000
250
500
500
250
250
250
250
1000
1000
250
500
500
PHYSICAL LAYER
Table 37. PHY Termination Autocalibration Routine
The physical layer of the JESD204B interface, hereafter referred
to as the deserializer, has eight identical channels. Each channel
consists of the terminators, an equalizer, a clock and data recovery
(CDR) circuit, and the 1:40 demux function (see Figure 36).
Address
0x2AA
0x2AB
0x2B1
0x2B2
0x2A7
0x2AE
TERMINATION
EQUALIZER
CDR
1:40
11675-006
SPI
CONTROL
FROM SERDES PLL
Figure 36. Deserializer Block Diagram
JESD204B data is input to the AD9144 via the SERDINx± 1.2 V
differential input pins as per the JESD204B specification.
Interface Power-Up and Input Termination
Before using the JESD204B interface, it must be powered up by
setting Register 0x200[0] = 0. In addition, each physical lane that is
not being used (SERDINx±) must be powered down. To do so,
set the corresponding Bit x for Physical Lane x in Register 0x201 to
0 if the physical lane is being used, and to 1 if it is not being used.
Description
SERDES interface termination configuration
SERDES interface termination configuration
SERDES interface termination configuration
SERDES interface termination configuration
Autotune PHY terminations
Autotune PHY terminations
The input termination voltage of the DAC is sourced externally
via the VTT pins (Pin 21, Pin 23, Pin 40, and Pin 43). Set VTT by
connecting it to SVDD12. It is recommended that the JESD204B
inputs be ac-coupled to the JESD204B transmit device using
100 nF capacitors.
Receiver Eye Mask
The AD9144 complies with the JESD204B specification
regarding the receiver eye mask and is capable of capturing data
that complies with this mask. Figure 37 shows the receiver eye
mask normalized to the data rate interval with a 600 mV VTT
swing. See the JESD204B specification for more information
regarding the eye mask and permitted receiver eye opening.
The AD9144 autocalibrates the input termination to 50 Ω.
Before running the termination calibration, Register 0x2AA,
Register 0x2AB, Register 0x2B1, and Register 0x2B2 must be
written as described in Table 37 to guarantee proper calibration.
The termination calibration begins when Register 0x2A7[0] and
Register 0x2AE[0] transition from low to high. Register 0x2A7
controls autocalibration for PHY 0, PHY 1, PHY 6, and PHY 7.
Register 0x2AE controls autocalibration for PHY 2, PHY 3,
PHY 4, and PHY 5.
The PHY termination autocalibration routine is as shown in
Table 37.
Rev. A | Page 35 of 125
LV-OIF-11G-SR RECEIVER EYE MASK
525
55
0
–55
–525
0
0.35
0.5
0.65
TIME (UI)
Figure 37. Receiver Eye Mask
1.00
11675-007
SERDINx±
AMPLITUDE (mV)
DESERIALIZER
Value
0xB7
0x87
0xB7
0x87
0x01
0x01
AD9144
Data Sheet
Clock Relationships
Register 0x280 controls the synthesizer enable and recalibration.
The following clocks rates are used throughout the rest of the
JESD204B section. The relationship between any of the clocks
can be derived from the following equations:
To enable the SERDES PLL, first set the PLL divider register
according to Table 38, then enable the SERDES PLL by writing
Register 0x280[0] to 1.
DataRate = (DACRate)/(InterpolationFactor)
Confirm that the SERDES PLL is working by reading
Register 0x281. If Register 0x281[0] = 1, the SERDES PLL has
locked. If Register 0x281[3] = 1, the SERDES PLL was successfully
calibrated. If Register 0x281[4] or Register 0x281[5] are high, the
PLL hit the upper or lower end of its calibration band and must be
recalibrated by writing 0 and then 1 to Register 0x280[2].
LaneRate = (20 × DataRate × M)/L
ByteRate = LaneRate/10
This comes from 8-bit/10-bit encoding, where each byte is
represented by 10 bits.
PClockRate = ByteRate/4
SERDES PLL Fixed Register Writes
The processing clock is used for a quad-byte decoder.
To optimize the SERDES PLL across all operating conditions,
the register writes in Table 39 are recommended.
FrameRate = ByteRate/F
where F is defined as (bytes per frame) per lane.
Table 39. SERDES PLL Fixed Register Writes
PClockFactor = FrameRate/PClockRate = 4/F
where:
M is the JESD204B parameter for converters per link.
L is the JESD204B parameter for lanes per link.
F is the JESD204B parameter for octets per frame per lane.
SERDES PLL
Functional Overview of the SERDES PLL
The independent SERDES PLL uses integer-N techniques to
achieve clock synthesis. The entire SERDES PLL is integrated
on-chip, including the VCO and the loop filter. The SERDES
PLL VCO operates over the range of 5.65 GHz to 11.04 GHz.
In the SERDES PLL, a VCO divider block divides the VCO
clock by 2 to generate a 2.825 GHz to 5.52 GHz quadrature
clock for the deserializer cores. This clock is the input to the
clock and data recovery block that is described in the Clock and
Data Recovery section.
The reference clock to the SERDES PLL is always running at a
frequency, fREF = 1/40 of the lane rate = PClockRate. This clock
is divided by a DivFactor to deliver a clock to the PFD block
that is between 35 MHz and 80 MHz. Table 38 includes the
respective SERDES_PLL_DIV_MODE register settings for each
of the desired DivFactor options available.
Register
Address
0x284
0x285
0x286
0x287
0x28A
0x28B
0x290
0x294
0x296
0x297
0x299
0x29A
0x29C
0x29F
0x2A0
Divide by
(DivFactor)
1
2
4
Description
Optimal SERDES PLL loop filter
Optimal SERDES PLL loop filter
Optimal SERDES PLL loop filter
Optimal SERDES PLL charge pump
Optimal SERDES PLL VCO LDO
Optimal SERDES PLL configuration
Optimal SERDES PLL VCO varactor
Optimal SERDES PLL charge pump
Optimal SERDES PLL VCO
Optimal SERDES PLL VCO
Optimal SERDES PLL configuration
Optimal SERDES PLL VCO varactor
Optimal SERDES PLL charge pump
Optimal SERDES PLL VCO varactor
Optimal SERDES PLL VCO varactor
SERDES PLL IRQ
SERDES PLL lock and lost signals are available as IRQ events.
Use Register 0x01F[3:2] to enable these signals, and then use
Register 0x023[3:2] to read back their statuses and reset the IRQ
signals. See the Interrupt Request Operation section for more
information.
Table 38. SERDES PLL Divider Settings
LaneRate (Gbps)
1.44 to 2.76
2.88 to 5.52
5.75 to 10.64
Register
Value
0x62
0xC9
0x0E
0x12
0x7B
0x00
0x89
0x24
0x03
0x0D
0x02
0x8E
0x2A
0x78
0x06
SERDES_PLL_DIV_MODE,
Register 0x289[1:0]
2
1
0
Rev. A | Page 36 of 125
Data Sheet
AD9144
2.825GHz TO 5.52GHz
OUTPUT
VCO
LDO
CHARGE
PUMP
I Q
PFD
80MHz
MAX
fREF
BIT RATE ÷ 40
DivFactor
(1, 2, 4)
C1
R1
UP
C2
C3
LC VCO
5.65GHz TO 11.04GHz
÷2
DOWN
÷80
R3
ALC CAL
CAL CONTROL BITS
11675-011
FO CAL
3.2mA
Figure 38. SERDES PLL Synthesizer Block Diagram Including VCO Divider Block
Clock and Data Recovery
Equalization
The deserializer is equipped with a CDR circuit. Instead of
recovering the clock from the JESD204B serial lanes, the CDR
recovers the clocks from the SERDES PLL. The 2.825 GHz to
5.52 GHz output from the SERDES PLL, shown in Figure 38,
is the input to the CDR.
To compensate for signal integrity distortions for each PHY
channel due to PCB trace length and impedance, the AD9144
employs an easy to use, low power equalizer on each JESD204B
channel. The AD9144 equalizers can compensate for insertion
losses far greater than required by the JESD204B specification.
The equalizers have two modes of operation that are determined
by the EQ_POWER_MODE register setting in Register 0x268[7:6].
In low power mode (Register 0x268[7:6] = 2b’01) and operating
at the maximum lane rate of 10 Gbps, the equalizer can
compensate for up to 12 dB of insertion loss. In normal mode
(Register 0x268[7:6] = 2b’00), the equalizer can compensate for
up to 17.5 dB of insertion loss. This performance is shown in
Figure 39 as an overlay to the JESD204B specification for
insertion loss. Figure 39 shows the equalization performance at
10.0 Gbps, near the maximum baud rate for the AD9144.
A CDR sampling mode must be selected to generate the lane
rate clock inside the device. If the desired lane rate is greater
than 5.65 GHz, half rate CDR operation must be used. If the
desired lane rate is less than 5.65 GHz, disable half rate operation.
If the lane rate is less than 2.825 GHz, disable half rate and
enable 2× oversampling to recover the appropriate lane rate clock.
Table 40 gives a breakdown of CDR sampling settings that must
be set dependent on the LaneRate.
Table 40. CDR Operating Modes
LaneRate (Gbps)
1.44 to 2.76
2.88 to 5.52
5.75 to 10.64
ENHALFRATE,
Register 0x230[5]
0
0
1
CDR_OVERSAMP,
Register 0x230[1]
1
0
0
The CDR circuit synchronizes the phase used to sample the data on
each serial lane independently. This independent phase adjustment
per serial interface ensures accurate data sampling and eases the
implementation of multiple serial interfaces on a PCB.
After configuring the CDR circuit, reset it and then release the
reset by writing 1 and then 0 to Register 0x206[0].
Power-Down Unused PHYs
Note that any unused and enabled lanes consume extra power
unnecessarily. Each lane that is not being used (SERDINx±)
must be powered off by writing a 1 to the corresponding bit of
PHY_PD (Register 0x201).
Figure 40 and Figure 41 are provided as points of reference for
hardware designers and show the insertion loss for various
lengths of well laid out stripline and microstrip transmission
lines. See the Hardware Considerations section for specific layout
recommendations for the JESD204B channel.
Low power mode is recommended if the insertion loss of the
JESD204B PCB channels is less than that of the most lossy
supported channel for lower power mode (shown in Figure 39).
If the insertion loss is greater than that, but still less than that of
the most lossy supported channel for normal mode (shown in
Figure 39), use normal mode. At 10 Gbps operation, the EQ in
normal mode consumes about 4 mW more power per lane used
than in low power EQ mode. Note that either mode can be used
in conjunction with transmitter preemphasis to ensure
functionality and/or to optimize for power.
Rev. A | Page 37 of 125
AD9144
Data Sheet
0
JESD204B SPEC ALLOWED
CHANNEL LOSS
2
INSERTION LOSS (dB)
4
6
AD9144 ALLOWED
CHANNEL LOSS
(LOW POWER MODE)
8
EXAMPLE OF
AD9144
COMPATIBLE
CHANNEL (LOW
POWER MODE)
10
12
14
EXAMPLE OF
AD9144
COMPATIBLE
CHANNEL
(NORMAL MODE)
AD9144 ALLOWED
CHANNEL LOSS
(NORMAL MODE)
16
DATA LINK LAYER
EXAMPLE OF
JESD204B
COMPLIANT
CHANNE L
18
20
22
2.5
5.0
11675-339
24
7.5
FREQUENCY (GHz)
Figure 39. Insertion Loss Allowed
0
The AD9144 can operate as a single-link or dual-link high
speed JESD204B serial data interface. When operating in duallink mode, configure both links with the same JESD204B
parameters because they share a common device clock and system
reference. All eight lanes of the JESD204B interface handle link
layer communications such as code group synchronization,
frame alignment, and frame synchronization.
–5
ATTENUATION (dB)
–10
–15
–20
STRIPLINE = 6”
STRIPLINE = 10”
STRIPLINE = 15”
STRIPLINE = 20”
STRIPLINE = 25”
STRIPLINE = 30”
–25
–30
–40
1
2
3
4
5
6
7
8
9
10
FREQUENCY (GHz)
11675-010
–35
0
Figure 40. Insertion Loss of 50 Ω Striplines on FR4
0
–5
–15
–20
6” MICROSTRIP
10” MICROSTRIP
15” MICROSTRIP
20” MICROSTRIP
25” MICROSTRIP
30” MICROSTRIP
–30
–35
–40
0
1
2
3
4
5
6
7
8
9
FREQUENCY (GHz)
10
The AD9144 decodes 8-bit/10-bit control characters, allowing
marking of the start and end of the frame and alignment between
serial lanes. Each AD9144 serial interface link can issue a
synchronization request by setting its SYNCOUT0±/SYNCOUT1±
signal low. The synchronization protocol follows Section 4.9 of the
JESD204B standard. When a stream of four consecutive /K/
symbols is received, the AD9144 deactivates the synchronization
request by setting the SYNCOUT0±/SYNCOUT1± signal high
at the next internal LMFC rising edge. Then, it waits for the
transmitter to issue an ILAS. During the ILAS sequence, all
lanes are aligned using the /A/ to /R/ character transition as
described in the JESD204B Serial Link Establishment section.
Elastic buffers hold early arriving lane data until the alignment
character of the latest lane arrives. At this point, the buffers for
all lanes are released and all lanes are aligned (see Figure 44).
11675-011
ATTENUATION (dB)
–10
–25
The data link layer of the AD9144 JESD204B interface accepts
the deserialized data from the PHYs and deframes and descrambles
them so that data octets are presented to the transport layer to
be put into DAC samples. Figure 42 shows the link mode block
diagrams for single-link and dual-link configurations and the
interaction between the physical layer and logical layer. The
logical lanes and DACs can only be configured in sequential
order; for example in Mode 10, when in single-link mode, the
AD9144 only uses Logical Lane 0 and DAC0, and in dual-link
mode, only uses Logical Lane 0, Logical Lane 1 and DAC0,
DAC1. See the Mode Configuration Maps section for further
details on each of the mode configurations supported. The
architecture of the data link layer is shown in Figure 43. The
data link layer consists of a synchronization FIFO for each lane,
a crossbar switch, a deframer, and descrambler.
Figure 41. Insertion Loss of 50 Ω Microstrips on FR4
Rev. A | Page 38 of 125
Data Sheet
AD9144
PHYSICAL
LAYER (PHY)
LOGICAL
LAYER
CROSSBAR
DAC CORE
QBD
SERDIN0±/PHYSICAL LANE 0
LOGICAL LANE 0/LINK 0 LANE 0
SERDIN1±/PHYSICAL LANE 1
LOGICAL LANE 1/LINK 0 LANE 1
SINGLE-LINK MODE
DAC0
SERDIN2±/PHYSICAL LANE 2
SERDIN3±/PHYSICAL LANE 3
SERDIN4±/PHYSICAL LANE 4
SERDIN5±/PHYSICAL LANE 5
LOGICAL LANE 2/LINK 0 LANE 2
CROSSBAR
REGISTER
0x308
TO
REGISTER
0x30B
DAC1
LOGICAL LANE 3/LINK 0 LANE 3
QUAD-BYTE
DEFRAMER
(QBD0)
LOGICAL LANE 4/LINK 0 LANE 4
DAC2
LOGICAL LANE 5/LINK 0 LANE 5
SERDIN6±/PHYSICAL LANE 6
LOGICAL LANE 6/LINK 0 LANE 6
SERDIN7±/PHYSICAL LANE 7
LOGICAL LANE 7/LINK 0 LANE 7
PHYSICAL LANES
DAC3
LOGICAL LANES
SERDIN0±/PHYSICAL LANE 0
LOGICAL LANE 0/LINK 0 LANE 0
SERDIN1±/PHYSICAL LANE 1
LOGICAL LANE 1/LINK 0 LANE 1
SERDIN2±/PHYSICAL LANE 2
LOGICAL LANE 2/LINK 0 LANE 2
SERDIN3±/PHYSICAL LANE 3
SERDIN4±/PHYSICAL LANE 4
SERDIN5±/PHYSICAL LANE 5
CROSSBAR
REGISTER
0x308
TO
REGISTER
0x30B
QUAD-BYTE
DEFRAMER 0
(QBD0)
DAC1
LOGICAL LANE 3/LINK 0 LANE 3
LOGICAL LANE 4/LINK 1 LANE 0
DAC2
LOGICAL LANE 5/LINK 1 LANE 1
SERDIN6±/PHYSICAL LANE 6
LOGICAL LANE 6/LINK 1 LANE 2
SERDIN7±/PHYSICAL LANE 7
LOGICAL LANE 7/LINK 1 LANE 3
QUAD-BYTE
DEFRAMER 1
(QBD1)
DAC3
LOGICAL LANES
11675-442
PHYSICAL LANES
Figure 42. Link Mode Functional Diagram
DATA LINK LAYER
SYNCOUTx±
LANE 7 DESERIALIZED
AND DESCRAMBLED DATA
LANE 7 DATA CLOCK
SYSREF
SERDIN0
FIFO
CROSS
BAR
SWITCH
SERDIN7
FIFO
LANE0 OCTETS
LANE7 OCTETS
SYSTEM CLOCK
PHASE DETECT
11675-012
LANE 0 DATA CLOCK
QUAD-BYTE
DEFRAMER
QBD
DESCRAMBLE
LANE 0 DESERIALIZED
AND DESCRAMBLED DATA
10-BIT/8-BIT DECODE
DUAL-LINK MODE
DAC0
PCLK
SPI CONTROL
Figure 43. Data Link Layer Block Diagram
Rev. A | Page 39 of 125
AD9144
Data Sheet
L RECEIVE LANES
(EARLIEST ARRIVAL) K K K R D D
D D A R Q C
L RECEIVE LANES
(LATEST ARRIVAL) K K K K K K K R D D
C
D D A R Q C
D D A R D D
C
D D A R D D
0 CHARACTER ELASTIC BUFFER DELAY OF LATEST ARRIVAL
4 CHARACTER ELASTIC BUFFER DELAY OF EARLIEST ARRIVAL
L ALIGNED
RECEIVE LANES K K K K K K K R D D
D D A R Q C
C
11675-013
K = K28.5 CODE GROUP SYNCHRONIZATION COMMA CHARACTER
A = K28.3 LANE ALIGNMENT SYMBOL
F = K28.7 FRAME ALIGNMENT SYMBOL
R = K28.0 START OF MULTIFRAME
Q = K28.4 START OF LINK CONFIGURATION DATA
C = JESD204B LINK CONFIGURATION PARAMETERS
D = Dx.y DATA SYMBOL
D D A R D D
Figure 44. Lane Alignment During ILAS
JESD204B Serial Link Establishment
A brief summary of the high speed serial link establishment
process for Subclass 1 is provided. See Section 5.3.3 of the
JESD204B specifications document for complete details.
After the last /A/ character of the last ILAS, multiframe data
begins streaming. The receiver adjusts the position of the /A/
character such that it aligns with the internal LMFC of the
receiver at this point.
Step 1: Code Group Synchronization
Step 3: Data Streaming
Each receiver must locate K (K28.5) characters in its input data
stream. After four consecutive K characters are detected on all
link lanes, the receiver block deasserts the SYNCOUTx± signal
to the transmitter block at the receiver local multiframe clock
(LMFC) edge.
In this phase, data is streamed from the transmitter block to the
receiver block.
The transmitter captures the change in the SYNCOUTx± signal,
and at a future transmitter LMFC rising edge, starts the initial
lane alignment sequence (ILAS).
Optionally, data can be scrambled. Scrambling does not start
until the very first octet following the ILAS.
The receiver block processes and monitors the data it receives
for errors, including:
The main purposes of this phase are to align all the lanes of the
link and to verify the parameters of the link.





Before the link is established, write each of the link parameters
to the receiver device to designate how data is sent to the
receiver block.
If any of these errors exist, they are reported back to the
transmitter in one of a few ways (see the JESD204B Error
Monitoring section for details).
The ILAS consists of four or more multiframes. The last character
of each multiframe is a multiframe alignment character, /A/.
The first, third, and fourth multiframes are populated with
predetermined data values. Note that Section 8.2 of the JESD204B
specifications document describes the data ramp that is expected
during ILAS. By default, the AD9144 does not require this ramp.
Register 0x47E[0] can be set high to require the data ramp. The
deframer uses the final /A/ of each lane to align the ends of the
multiframes within the receiver. The second multiframe contains
an R (K28.0), Q (K28.4), and then data corresponding to the
link parameters. Additional multiframes can be added to the
ILAS if needed by the receiver. By default, the AD9144 uses four
multiframes in the ILAS (this can be changed in Register 0x478).
If using Subclass 1, exactly four multiframes must be used.

Step 2: Initial Lane Alignment Sequence


Bad running disparity (8-bit/10-bit error)
Not in table (8-bit/10-bit error)
Unexpected control character
Bad ILAS
Interlane skew error (through character replacement)
SYNCOUTx± signal assertion: resynchronization
(SYNCOUTx± signal pulled low) is requested at each error
for the last two errors. For the first three errors, an optional
resynchronization request can be asserted when the error
counter reaches a set error threshold.
For the first three errors, each multiframe with an error in
it causes a small pulse on SYNCOUTx±.
Errors can optionally trigger an IRQ event, which can be
sent to the transmitter.
Various test modes for verifying the link integrity can be found
in the JESD204B Test Modes section.
Rev. A | Page 40 of 125
Data Sheet
AD9144
Lane FIFO
The FIFOs in front of the crossbar switch and deframer
synchronize the samples sent on the high speed serial data
interface with the deframer clock by adjusting the phase of the
incoming data. The FIFO absorbs timing variations between the
data source and the deframer; this allows up to two PClock cycles
of drift from the transmitter. The FIFO_STATUS_REG_0 register
and FIFO_STATUS_REG_1 register (Register 0x30C and
Register 0x30D, respectively) can be monitored to identify
whether the FIFOs are full or empty.
Lane FIFO IRQ
An aggregate lane FIFO error bit is also available as an IRQ
event. Use Register 0x01F[1] to enable the FIFO error bit, and
then use Register 0x023[1] to read back its status and reset the
IRQ signal. See the Interrupt Request Operation section for
more information.
Crossbar Switch
Register 0x308 to Register 0x30B allow arbitrary mapping of
physical lanes (SERDINx±) to logical lanes used by the SERDES
deframers.
Table 41. Crossbar Registers
Address
0x308
0x308
0x309
0x309
0x30A
0x30A
0x30B
0x30B
Bits
[2:0]
[5:3]
[2:0]
[5:3]
[2:0]
[5:3]
[2:0]
[5:3]
In single-link mode, Deframer 0 is used exclusively and Deframer 1
remains inactive. In dual-link mode, both QBDs are active and
must be configured separately using the LINK_PAGE bit
(Register 0x300[2]) to select which link is being configured. The
LINK_MODE bit (Register 0x300[3]) is 1 for dual-link, or 0 for
single-link.
Each deframer uses the JESD204B parameters that the user has
programmed into the register map to identify how the data has
been packed and how to unpack it. The JESD204B parameters
are discussed in detail in the Transport Layer section; many of
the parameters are also needed in the transport layer to convert
JESD204B frames into samples.
Descrambler
The AD9144 provides an optional descrambler block using a
self synchronous descrambler with a polynomial: 1 + x14 + x15.
Enabling data scrambling reduces spectral peaks that are
produced when the same data octets repeat from frame to
frame. It also makes the spectrum data independent so that
possible frequency-selective effects on the electrical interface do
not cause data-dependent errors. Descrambling of the data is
enabled by setting the SCR bit (Register 0x453[7]) to 1.
Syncing LMFC Signals
Logical Lane
LOGICAL_LANE0_SRC
LOGICAL_LANE1_SRC
LOGICAL_LANE2_SRC
LOGICAL_LANE3_SRC
LOGICAL_LANE4_SRC
LOGICAL_LANE5_SRC
LOGICAL_LANE6_SRC
LOGICAL_LANE7_SRC
The first step in guaranteeing synchronization across links and
devices begins with syncing the LMFC signals. Each DAC dual
(DAC Dual A: DAC0/DAC1 and DAC Dual B: DAC2/DAC3)
has its own LMFC signal. In Subclass 0, the LMFC signals for
each of the two links are synchronized to an internal processing
clock. In Subclass 1, all LMFC signals (for all duals and devices)
are synchronized to an external SYSREF signal. All LMFC sync
registers are paged as described in the Dual Paging section.
SYSREF Signal
Write each LOGICAL_LANEy_SRC with the number (x) of the
desired physical lane (SERDINx±) from which to obtain data.
By default, all logical lanes use the corresponding physical lane
as their data source. For example, by default LOGICAL_
LANE0_SRC = 0; therefore, Logical Lane 0 obtains data from
Physical Lane 0 (SERDIN0±). If instead the user wants to use
SERDIN4± as the source for Logical Lane 0, the user must write
LOGICAL_LANE0_SRC = 4.
Lane Inversion
Register 0x334 allows inversion of desired logical lanes, which
can be used to ease routing of the SERDINx± signals. For each
Logical Lane x, set Bit x of Register 0x334 to 1 to invert it.
Deframers
The AD9144 consists of two quad-byte deframers (QBDs). Each
deframer takes in the 8-bit/10-bit encoded data from the
deserializer (via the crossbar switch), decodes it, and descrambles it
into JESD204B frames before passing it to the transport layer to be
converted to DAC samples. The deframer processes four symbols
(or octets) per processing clock (PClock) cycle.
The SYSREF signal is a differential source synchronous input
that synchronizes the LMFC signals in both the transmitter
and receiver in a JESD204B Subclass 1 system to achieve
deterministic latency.
The SYSREF signal is an active high signal that is sampled by
the device clock rising edge. It is best practice that the device
clock and SYSREF signals be generated by the same source,
such as the AD9516-1 clock generator, so that the phase
alignment between the signals is fixed. When designing for
optimum deterministic latency operation, consider the timing
distribution skew of the SYSREF signal in a multipoint link
system (multichip).
The AD9144 supports a single pulse or step, or a periodic
SYSREF± signal. The periodicity can be continuous, strobed, or
gapped periodic. The SYSREF± signal can always be dc-coupled
(with a common-mode voltage of 0 V to 2 V). When dc-coupled,
a small amount of common-mode current (<500 μA) is drawn
from the SYSREF± pins. See Figure 45 for the SYSREF± internal
circuit.
Rev. A | Page 41 of 125
AD9144
Data Sheet
To avoid this common-mode current draw, a 50% duty-cycle
periodic SYSREF± signal can be used with ac coupling capacitors.
If ac-coupled, the ac coupling capacitors combine with the
resistors shown in Figure 45 to make a high-pass filter with RC
time constant τ = RC. Select C such that τ > 4/SYSREF Freq.
In addition, the edge rate must be sufficiently fast—at least
1.3 V/ns is recommended per Table 5—to meet the SYSREF vs.
DAC clock keep out window (KOW) requirements.
It is possible to use ac-coupled mode without meeting the
frequency to time-constant constraint mentioned by using
SYSREF hysteresis (Register 0x081 and Register 0x082).
However, this increases the DAC clock KOW (Table 5 does not
apply) by an amount depending on SYSREF frequency, level of
hysteresis, capacitor choice, and edge rate.
1.2V
3kΩ
SYSREF–
3kΩ
The maximum acceptable phase error (in DAC clock cycles)
between the alignment edge and the LMFC edge is set in the
error window tolerance register. If continuous sync mode is
used with a nonzero error window tolerance, a phase check
occurs on every SYSREF pulse, but an alignment occurs only if
the phase error is greater than the specified error window
tolerance. If the jitter of the SYSREF± signal violates the KOW
specification given in Table 5 and therefore causes phase error
uncertainty, the error tolerance can be increased to avoid
constant clock rotations. Note that this means the latency is less
deterministic by the size of the window. If the error window
tolerance must be set above 3, Subclass 0 with a one-shot sync is
recommended.
For debug purposes, SYNCARM (Register 0x03A[6]) can be
used to inform the user that alignment edges are being received
in continuous mode. Because the SYNCARM bit is self cleared
after an alignment edge is received, the user can arm the sync
(SYNCARM (Register 0x03A[6]) = 1), and then read back
SYNCARM. If SYNCARM = 0, the alignment edges are being
received and phase checks are occurring. Arming the sync
machine in this mode does not affect the operation of the
device.
~600mV
11675-015
SYSREF+
occurs on every alignment edge in continuous mode. The one
caveat to the previous statement is that when a phase rotation cycle
is underway, subsequent alignment edges are ignored until the
logic lane is ready again.
Figure 45. SYSREF± Input Circuit
Sync Processing Modes Overview
The AD9144 supports various LMFC sync processing modes.
These modes are one-shot, continuous, windowed continuous,
and monitor modes. All sync processing modes perform a
phase check to see that the LMFC is phase aligned to an
alignment edge. In Subclass 1, the SYSREF pulse acts as the
alignment edge; in Subclass 0, an internal processing clock acts
as the alignment edge. If the signals are not in phase, a clock
rotation occurs to align the signals. The sync modes are
described in the following sections. See the Sync Procedure
section for details on the procedure for syncing the LMFC
signals.
One-Shot Sync Mode (SYNCMODE = 0x1)
In one-shot sync mode, a phase check occurs on only the first
alignment edge that is received after the sync machine is armed.
If the phase error is larger than a specified window error
tolerance, a phase adjustment occurs. Though an LMFC
synchronization occurs only once, the SYSREF signal can
still be continuous.
Continuous Sync Mode (SYNCMODE = 0x2)
Continuous mode must only be used in Subclass 1 with a periodic
SYSREF± signal. In continuous mode, a phase check/alignment
occurs on every alignment edge.
One-Shot then Monitor Sync Mode (SYNCMODE = 0x9)
In one-shot then monitor mode, the user can monitor the phase
error in real time. Use this sync mode with a periodic SYSREF±
signal. A phase check and alignment occurs on the first alignment
edge received after the sync machine is armed. On all subsequent
alignment edges the phase is monitored and reported, but no clock
phase adjustment occurs.
The phase error can be monitored on the SYNC_CURRERR_L
register (Register 0x03C[3:0]). Immediately after an alignment
occurs, CURRERR = 0 indicates that there is no difference
between the alignment edge and the LMFC edge. On every
subsequent alignment edge, the phase is checked. If the
alignment is lost, the phase error is reported in the SYNC_
CURRERR_L register in DAC clock cycles. If the phase error is
beyond the selected window tolerance (Register 0x034[2:0]),
one bit of Register 0x03D[7:6] is set high depending on whether
the phase error is on the low or high side.
When an alignment occurs, snapshots of the last phase error
(Register 0x03C[3:0]) and the corresponding error flags
(Register 0x03D[7:6]) are placed into readable registers for
reference (Register 0x038 and Register 0x039, respectively).
Continuous mode differs from one-shot mode in two ways.
First, no SPI cycle is required to arm the device; the alignment
edge seen after continuous mode is enabled results in a phase
check. Second, a phase check (and when necessary, clock rotation)
Rev. A | Page 42 of 125
Data Sheet
AD9144
Sync Procedure
LMFC Sync IRQ
The procedure for enabling the sync is as follows:
The sync status bits (SYNCLOCK, SYNCROTATE, SYNCTRIP,
and SYNCWLIM) are available as IRQ events.
1.
2.
3.
4.
5.
6.
7.
Set Register 0x008 to 0x03 to sync the LMFC for both
duals (DAC0/DAC1 and DAC2/DAC3).
Set the desired sync processing mode. The sync processing
mode settings are listed in Table 42.
For Subclass 1, set the error window according to the
uncertainty of the SYSREF± signal relative to the DAC
clock and the tolerance of the application for deterministic
latency uncertainty. Sync window tolerance settings are
given in Table 43.
Enable sync by writing SYNCENABLE
(Register 0x03A[7] = 1).
If in one-shot mode, arm the sync machine by writing
SYNCARM (Register 0x03A[6] = 1).
If in Subclass 1, ensure that at least one SYSREF pulse is
sent to the device.
Check the status by reading the following bit fields:
a) SYNC_BUSY (Register 0x03B[7]) = 0 to indicate that
the sync logic is no longer busy.
b) SYNC_LOCK (Register 0x03B[3]) = 1 to indicate that
the signals are aligned. This bit updates on every
phase check.
c) SYNC_WLIM (Register 0x03B[1]) = 0 to indicate that
the phase error is not beyond the specified error
window. This bit updates on every phase check.
d) SYNC_ROTATE (Register 0x03B[2]) = 1 if the phases
were not aligned before the sync and an alignment
occurred; this indicates that a clock alignment occurred.
This bit is sticky and can be cleared only by writing to
the SYNCCLRSTKY control bit (Register 0x03A[5]).
e) SYNC_TRIP (Register 0x03B[0]) = 1 to indicate
alignment edge received and phase check occurred.
This bit is sticky and can be cleared only by writing to
the SYNCCLRSTKY control bit (Register 0x03A[5]).
Table 42. Sync Processing Modes
Sync Processing Mode
One-shot
Continuous
One-shot then monitor
SYNCMODE (Register 0x03A[3:0])
0x01
0x02
0x09
Table 43. Sync Window Tolerance
Sync Error Window
Tolerance
±½ DAC clock cycles
±1 DAC clock cycles
±2 DAC clock cycles
±3 DAC clock cycles
Use Register 0x021[3:0] to enable the sync status bits for DAC
Dual A (DAC0 and DAC1), and then use Register 0x025[3:0] to
read back their statuses and reset the IRQ signals.
Use Register 0x022[3:0] to enable the sync status bits for DAC
Dual B (DAC2 and DAC3), and then use Register 0x026[3:0]
read back their statuses and reset the IRQ signals.
See the Interrupt Request Operation section for more information.
Deterministic Latency
JESD204B systems 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 power cycle to
power cycle with each new link establishment. Section 6 of the
JESD204B specification addresses the issue of deterministic
latency with mechanisms defined as Subclass 1 and Subclass 2.
The AD9144 supports JESD204B Subclass 0 and Subclass 1 operation, but not Subclass 2. Write the subclass to Register 0x301[2:0]
and once per link to Register 0x458[7:5].
Subclass 0
This mode does not require any signal on the SYSREF± pins,
which can be left disconnected.
Subclass 0 still requires that all lanes arrive within the same
LMFC cycle, and the dual DACs must be synchronized to each
other.
Minor Subclass 0 Caveats
Because the AD9144 requires an ILAS, the nonmultiple
converter single lane (NMCDA-SL) case from the JESD204A
specification is only supported when using the optional ILAS.
Error reporting using SYNCOUTx± is not supported when
using Subclass 0 with F = 1.
Subclass 1
This mode gives deterministic latency and allows links to be
synced to within ½ of a DAC clock period. It requires an
external SYSREF± signal that is accurately phase aligned to the
DAC clock.
ERRWINDOW (Register 0x034[2:0])
0x00
0x01
0x02
0x03
Rev. A | Page 43 of 125
AD9144
Data Sheet
Because the LMFC is periodic, this can account for any amount
of fixed delay. As a result, the LMFC period must only be larger
than the variation in the link delays, and the AD9144 can achieve
proper performance with a smaller total latency. Figure 46 and
Figure 47 show a case where the link delay is larger than an
LMFC period. Note that it can be accommodated by delaying
LMFCRx.
Deterministic Latency Requirements
Several key factors are required for achieving deterministic
latency in a JESD204B Subclass 1 system.
POWER CYCLE
VARIANCE
LMFC
ILAS
ALIGNED DATA
DATA
LATE ARRIVING
LMFC REFERENCE
EARLY ARRIVING
LMFC REFERENCE
Link Delay
Figure 46. Link Delay > LMFC Period Example
The link delay of a JESD204B system is the sum of fixed and
variable delays from the transmitter, channel, and receiver as
shown in Figure 48.
POWER CYCLE
VARIANCE
LMFC
For proper functioning, all lanes on a link must be read during
the same LMFC period. Section 6.1 of the JESD204B
specification states that the LMFC period must be larger than
the maximum link delay. For the AD9144, this is not necessarily
the case; instead, the AD9144 uses a local LMFC for each link
(LMFCRx) that can be delayed from the SYSREF aligned LMFC.
ILAS
ALIGNED DATA
DATA
LMFCRX
LMFC REFERENCE FOR ALL POWER CYCLES
LMFC_DELAY
FRAME CLOCK
Figure 47. LMFC_DELAY to Compensate for Link Delay > LMFC
LINK DELAY = DELAYFIXED + DELAYVARIABLE
LOGIC DEVICE
(JESD204B Tx)
CHANNEL
JESD204B Rx
DSP
DAC
POWER CYCLE
VARIANCE
LMFC
DATA AT
Tx INPUT
ALIGNED DATA
AT Rx OUTPUT
11675-018

ILAS
DATA
ILAS
DATA
FIXED DELAY
VARIABLE
DELAY
Figure 48. JESD204B Link Delay = Fixed Delay + Variable Delay
Rev. A | Page 44 of 125
11675-019

SYSREF± signal distribution skew within the system must
be less than the desired uncertainty.
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 ≤10 PClock periods. 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.
11675-017

Data Sheet
AD9144
1.
The method for setting the LMFCDel and LMFCVar is
described in the Link Delay Setup section.
Setting LMFCDel appropriately ensures that all the corresponding
data samples arrive in the same LMFC period. Then LMFCVar
is written into the receive buffer delay (RBD) to absorb all link
delay variation. This ensures that all data samples have arrived
before reading. By setting these to fixed values across runs and
devices, deterministic latency is achieved.
2.
The RBD described in the JESD204B specification takes values
from 1 to K frame clock cycles, while the RBD of the AD9144
takes values from 0 to 10 PClock cycles. As a result, up to
10 PClock cycles of total delay variation can be absorbed.
Because LMFCVar is in PClock cycles and LMFCDel is in frame
clock cycles, a conversion between these two units is needed.
The PClockFactor, or number of frame clock cycles per PClock
cycle, is equal to 4/F. For more information on this relationship,
see the Clock Relationships section.
3.
4.
Two examples follow that show how to determine LMFCVar
and LMFCDel. After they are calculated, write LMFCDel into
both Register 0x304 and Register 0x305 for all devices in the
system, and write LMFCVar to both Register 0x306 and
Register 0x307 for all devices in the system.
5.
Link Delay Setup Example, with Known Delays
All the known system delays can be used to calculate LMFCVar
and LMFCDel, as described in the Link Delay Setup section.
The example shown in Figure 49 is demonstrated in the
following steps according to the procedure outlined in the Link
Delay Setup section. Note that this example is in Subclass 1 to
achieve deterministic latency, which has a PClockFactor (4/F)
of 2 frameclock cycles per PClock cycle, and uses K = 32
(frames/multiframe). Because PCBFixed << PClockPeriod,
PCBFixed is negligible in this example and not included in the
calculations.
6.
7.
Find the receiver delays using Table 8.
RxFixed = 17 PClock cycles
RxVar = 2 PClock cycles
Find the transmitter delays. The equivalent table in the
example JESD204B core (implemented on a GTH or GTX
transceiver on a Virtex-6 FPGA) states that the delay is
56 ± 2 byte clock cycles.
Because the PClockRate = ByteRate/4, as described in the
Clock Relationships section, the transmitter delays in
PClock cycles are
TxFixed = 54/4 = 13.5 PClock cycles
TxVar = 4/4 = 1 PClock cycle
Calculate MinDelayLane as follows:
MinDelayLane = floor(RxFixed + TxFixed + PCBFixed)
= floor(17 + 13.5 + 0)
= floor(30.5)
MinDelayLane = 30
Calculate MaxDelayLane as follows:
MaxDelayLane = ceiling(RxFixed + RxVar + TxFixed +
TxVar + PCBFixed))
= ceiling(17 + 2 + 13.5 + 1 + 0)
= ceiling(33.5)
MaxDelayLane = 34
Calculate LMFCVar as follows:
LMFCVar = (MaxDelay + 1) − (MinDelay − 1)
= (34 + 1) − (30 − 1) = 35 − 29
LMFCVar = 6 PClock cycles
Calculate LMFCDel as follows:
LMFCDel = ((MinDelay − 1) × PClockFactor) % K
= ((30 − 1) × 2) % 32 = (29 × 2) % 32
= 58 % 32
LMFCDel = 26 frame clock cycles
Write LMFCDel to both Register 0x304 and Register 0x305
for all devices in the system. Write LMFCVar to both
Register 0x306 and Register 0x307 for all devices in the
system.
LMFC
PCLOCK
FRAME CLOCK
ALIGNED LANE DATA
AT Rx DEFRAMER OUTPUT
ILAS
DATA
ILAS
Tx VAR
DELAY
Rx VAR
DELAY
DATA
PCB FIXED
DELAY
LMFCRX
LMFC DELAY = 26 FRAME CLOCK CYCLES
TOTAL FIXED LATENCY = 30 PCLOCK CYCLES
Figure 49. LMFC_DELAY Calculation Example
Rev. A | Page 45 of 125
TOTAL VARIABLE
LATENCY = 4
PCLOCK CYCLES
11675-024
DATA AT Tx FRAMER
AD9144
Data Sheet

Link Delay Setup Example, Without Known Delay
If the system delays are not known, the AD9144 can read back
the link latency between LMFCRX for each link and the SYSREF
aligned LMFC. This information is then used to calculate
LMFCVar and LMFCDel, as shown in the Without Known
Delays section.


Figure 51 shows how DYN_LINK_LATENCY_x (Register 0x302
and Register 0x303) provides a readback showing the delay (in
PClock cycles) between LMFCRX and the transition from ILAS
to the first data sample. By repeatedly power-cycling and taking
this measurement, the minimum and maximum delays across
power cycles can be determined and used to calculate LMFCVar
and LMFCDel.
2.
3.
4.
The example shown in Figure 51 is demonstrated in the following
steps according to the procedure outlined in the Without Known
Delays section. Note that this example is in Subclass 1 to achieve
deterministic latency, which has a PClockFactor (FrameClockRate/
PClockRate) of 2 and uses K = 16; therefore PClocksPerMF = 8.
5.
In Figure 51, for Link A, Link B, and Link C, the system
containing the AD9144 (including the transmitter) is
power cycled and configured 20 times. The AD9144 is
configured as described in the Device Setup Guide.
Because the point of this exercise is to determine LMFCDel
and LMFCVar, the LMFCDel is programmed to 0 and the
DYN_ LINK_LATENCY_x is read from Register 0x302
and Register 0x303 for Link 0 and Link 1, respectively. The
variation in the link latency over the 20 runs is shown in
Figure 51 in grey.
6.
Calculate the minimum of all Delay measurements across
all power cycles, links, and devices:
MinDelay = min(all Delay values) = 4
Calculate the maximum of all Delay measurements across
all power cycles, links, and devices:
MaxDelay = max(all Delay values) = 9
Calculate the total Delay variation (with guard band)
across all power cycles, links, and devices:
LMFCVar = (MaxDelay + 1) − (MinDelay − 1)
= (9 + 1) − (4 − 1) = 10 − 3 = 7 PClock cycles
Calculate the minimum delay in frame clock cycles (with
guard band) across all power cycles, links, and devices:
LMFCDel = ((MinDelay − 1) × PClockFactor) % K
= ((4 − 1) × 2) % 16 = (3 × 2) % 16
= 6 % 16 = 6 frame clock cycles
Write LMFCDel to both Register 0x304 and Register 0x305
for all devices in the system. Write LMFCVar to both
Register 0x306 and Register 0x307 for all devices in the
system.
SYSREF
LMFCRX
ILAS
ALIGNED DATA
DATA
11675-022
DYN_LINK_LATENCY
Figure 50. DYN_LINK_LATENCY Illustration
LMFC
PCLOCK
FRAME CLOCK
DYN_LINK_LATENCY_CNT
0
1
2
ALIGNED DATA (LINK A)
ALIGNED DATA (LINK B)
ALIGNED DATA (LINK C)
3
4
5
6
7
0
1
2
3
ILAS
4
5
6
7
DATA
ILAS
DATA
ILAS
DATA
LMFCRX
DETERMINISTICALLY
DELAYED DATA
ILAS
LMFC_DELAY = 6
(FRAME CLOCK CYCLES)
DATA
LMFC_VAR = 7
(PCLOCK CYCLES)
Figure 51. Multilink Synchronization Settings, Derived Method Example
Rev. A | Page 46 of 125
11675-025
1.
Link A gives readbacks of 6, 7, 0, and 1. Note that the
set of recorded delay values rolls over the edge of a
multiframe at the boundary K/PClockFactor = 8. Add
PClocksPerMF = 8 to low set. Delay values range from
6 to 9.
Link B gives Delay values from 5 to 7.
Link C gives Delay values from 4 to 7.
Data Sheet
AD9144
TRANSPORT LAYER
TRANSPORT LAYER
(QBD)
LANE 0 OCTETS
DAC A_I0[15:0]
DELAY
BUFFER 0
F2S_0
DAC A_Q0[15:0]
LANE 3 OCTETS
PCLK_0
SPI CONTROL
LANE 4 OCTETS
DAC B_I0[15:0]
DELAY
BUFFER 1
F2S_1
DAC B_Q0[15:0]
LANE 7 OCTETS
11675-026
PCLK_1
SPI CONTROL
Figure 52. Transport Layer Block Diagram
The transport layer receives the descrambled JESD204B frames
and converts them to DAC samples based on the programmed
JESD204B parameters shown in Table 44. A number of device
parameters are defined in Table 45.
Table 45. JESD204B Device Parameters
Parameter
CF
Table 44. JESD204B Transport Layer Parameters
CS
Parameter
F
K
HD
L
M
S
Description
Number of octets per frame per lane: 1, 2, or 4.
Number of frames per multiframe.
K = 32 if F = 1, K = 16 or 32 otherwise.
Number of lanes per converter device (per link), as
follows:
1, 2, 4, or 8 (single-link mode).
1, 2, or 4 (dual-link mode).
Number of converters per device (per link), as follows:
1, 2, or 4 (single-link mode).
1 or 2 (dual-link mode).
Number of samples per converter, per frame: 1 or 2.
N
Nʹ (NP)
Description
Number of control words per device clock per link.
Not supported, must be 0.
Number of control bits per conversion sample.
Not supported, must be 0.
High density user data format. Used when samples
must be split across lanes.
Set to 1 when F = 1, otherwise 0.
Converter resolution = 16.
Total number of bits per sample = 16.
Certain combinations of these parameters, called JESD204B
operating modes, are supported by the AD9144. See Table 46
and Table 47 for a list of supported modes, along with their
associated clock relationships.
Rev. A | Page 47 of 125
AD9144
Data Sheet
Table 46. Single-Link JESD204B Operating Modes
Parameter
M (Converter Count)
L (Lane Count)
S (Samples per Converter per Frame)
F (Octets per Frame, per Lane)
K1 (Frames per Multiframe)
HD (High Density)
N (Converter Resolution)
NP (Bits per Sample)
Example Clocks for 10 Gbps Lane Rate
PClock Rate (MHz)
Frame Clock Rate (MHz)
Data Rate (MHz)
1
0
4
8
1
1
32
1
16
16
1
4
8
2
2
16/32
0
16
16
2
4
4
1
2
16/32
0
16
16
3
4
2
1
4
16/32
0
16
16
4
2
4
1
1
32
1
16
16
250
1000
1000
250
500
1000
250
500
500
250
250
250
250
1000
1000
Mode
5
2
4
2
2
16/32
0
16
16
250
500
1000
6
2
2
1
2
16/32
0
16
16
7
2
1
1
4
16/32
0
16
16
9
1
2
1
1
32
1
16
16
10
1
1
1
2
16/32
0
16
16
250
500
500
250
250
250
250
1000
1000
250
500
500
K must be 32 in Mode 0, Mode 4, and Mode 9. K can be 16 or 32 in all other modes.
Table 47. Dual-Link JESD204B Operating Modes for Link 0 and Link 1
Parameter
M (Converter Count)
L (Lane Count)
S (Samples per Converter per Frame)
F (Octets per Frame per Lane)
K1 (Frames per Multiframe)
HD (High Density)
N (Converter Resolution)
NP (Bits per Sample)
Example Clocks for 10 Gbps Lane Rate
PClock Rate (MHz)
Frame Clock Rate (MHz)
Data Rate (MHz)
1
4
2
4
1
1
32
1
16
16
5
2
4
2
2
16/32
0
16
16
Mode
6
2
2
1
2
16/32
0
16
16
7
2
1
1
4
16/32
0
16
16
9
1
2
1
1
32
1
16
16
10
1
1
1
2
16/32
0
16
16
250
1000
1000
250
500
1000
250
500
500
250
250
250
250
1000
1000
250
500
500
K must be 32 in Mode 4 and Mode 9. K can be 16 or 32 in all other modes.
Rev. A | Page 48 of 125
Data Sheet
AD9144
Configuration Parameters
Data Flow Through the JESD204B Receiver
The AD9144 modes refer to the link configuration parameters
for L, K, M, N, NP, S, and F. Table 48 provides the description
and addresses for these settings.
The link configuration parameters determine how the serial bits
on the JESD204B receiver interface are deframed and passed on
to the DACs as data samples. Figure 53 shows a detailed flow of
the data through the various hardware blocks for Mode 4 (L = 4,
M = 2, S = 1, F = 1). Simplified flow diagrams for all other modes
are provided in Figure 54 through Figure 62.
Table 48. Configuration Parameters
JESD204B
Setting
L−1
F−1
K−1
M−1
N−1
NP − 1
S−1
HD
F1
DID
BID
LID0
JESDV
1
Description
Number of lanes − 1.
Number of ((octets per frame) per
lane) − 1.
Number of frames per multiframe − 1.
Number of converters − 1.
Converter bit resolution − 1.
Bit packing per sample − 1.
Number of ((samples per converter)
per frame) − 1.
High density format. Set to 1 if F = 1.
Leave at 0 if F ≠ 1.
F parameter, in ((octets per frame) per
lane).
Device ID. Match the Device ID sent
by the transmitter.
Bank ID. Match the Bank ID sent by
the transmitter.
Lane ID for lane 0. Match the Lane ID
sent by the transmitter on Logical
Lane 0.
JESD204x Version. Match the version
sent by the transmitter (0x0 =
JESD204A, 0x1 = JESD204B).
F must be programmed in two places.
Address
0x453[4:0]
0x454[7:0]
0x455[4:0]
0x456[7:0]
0x457[4:0]
0x458[4:0]
0x459[4:0]
Single-Link and Dual-Link Configuration
The AD9144 uses the settings contained in Table 46 and Table 47.
Mode 0 to Mode 10 can be used for single-link operation.
Mode 4 to Mode 10 can also be used for dual-link operation.
To use dual-link mode, set LINK_MODE (Register 0x300[3]) to 1.
In dual-link mode, Link 1 must be programmed with identical
parameters to Link 0. To write to Link 1, set LINK_PAGE
(Register 0x300[2]) to 1.
0x476[7:0]
If single-link mode is being used, a small amount of power can
be saved by powering down the output buffer for SYNCOUT1±,
which can be done by setting Register 0x203[0] = 1.
0x450[7:0]
Checking Proper Configuration
0x45A[7]
0x451[3:0]
0x452[4:0]
0x459[7:5]
As a convenience, the AD9144 provides some quick configuration
checks. Register 0x030[5] is high if an illegal LMFC_DELAY is
used. Register 0x030[3] is high if an unsupported combination
of L, M, F, and S is used. Register 0x030[2] is high if an illegal K
is used. Register 0x030[1] is high if an illegal SUBCLASSV is used.
Deskewing and Enabling Logical Lanes
After proper configuration, the logical lanes must be deskewed and
enabled to capture data.
Set Bit x in Register 0x46C to 1 to deskew Logical Lane x and to 0 if
that logical lane is not being used. Then, set Bit x in Register 0x47D
to 1 to enable Logical Lane x and to 0 if that logical lane is not being
used.
Rev. A | Page 49 of 125
AD9144
Data Sheet
J0
SERDINx±
J19 J18
DESERIALIZER
DESERIALIZER
J11 J10
SERDINx±
J9
J8
J1
J0
DESERIALIZER
SERIAL JESD204B DATA (L = 4)
SAMPLES SPLIT ACROSS LANES
(HD = 1)
S19
S18
S17
S16
S15
S14
S13
S12
S11
S10
S9
S8
S7
S6
S5
S4
S3
S2
S1
S0
LANE 0, OCTET 0
10-BIT/8-BIT
DECODE
DAC0
D15
D14
D13
D12
D11
D10
D9
D8
D7
D6
D5
D4
D3
D2
D1
D0
DAC1
DESCRAMBLER
40 BITS PARALLEL DATA
(ENCODED AND SCRAMBLED)
2 CONVERTERS
(M = 2)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
16-BIT NIBBLE GROUP
(N = 16)
1 OCTET PER LANE
(F = 1)
Figure 53. JESD204B Mode 4 Data Deframing
Rev. A | Page 50 of 125
11675-027
J1
CONVERTER 0, SAMPLE 0
J8
D15
D14
D13
D12
D11
D10
D9
D8
D7
D6
D5
D4
D3
D2
D1
D0
CONVERTER 1, SAMPLE 0
J9
NIBBLE GROUP 0
SERDINx±
LANE 1, OCTET 0
DESERIALIZER
LANE 2, OCTET 0
J11 J10
TRANSPORT
LAYER
NIBBLE GROUP 1
J19 J18
DATA LINK LAYER
LANE 3, OCTET 0
PHYSICAL
LAYER
SERDINx±
S19
S18
S17
S16
S15
S14
S13
S12
S11
S10
S9
S8
S7
S6
S5
S4
S3
S2
S1
S0
Data Sheet
AD9144
Mode Configuration Maps
also apply to dual-link operation. Register 0x300 must be set
accordingly for single- or dual-link operation, as previously
discussed.
Table 49 to Table 58 contain the SPI configuration map for each
mode shown in Figure 54 through Figure 62. Figure 54 through
Figure 62 show the associated data flow through the deframing
process of the JESD204B receiver for each of the modes. Mode 0
to Mode 10 apply to single-link operation. Mode 4 to Mode 10
Additional details regarding all the SPI registers can be found in
the Register Maps and Descriptions section.
Table 49. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 0
LANE 3,
OCTET 0
J0
J1
SERDINx±
J7
LANE 6,
OCTET 0
J6
J9
SERDINx±
J8
J0
J1
SERDINx±
LANE 5,
OCTET 0
J6
J9
LANE 4,
OCTET 0
J15 J14
LANE 2,
OCTET 0
J15 J14
SERDINx±
J8
J0
J1
SERDINx±
J6
J8
SERDINx±
J9
J7
J7
J15 J14
J9
SERDINx±
SERDINx±
LANE 1,
OCTET 0
LANE 7,
OCTET 0
NIBBLE GROUP 0
NIBBLE GROUP 1
NIBBLE GROUP 2
NIBBLE GROUP 3
CONVERTER 0, SAMPLE 0
CONVERTER 1, SAMPLE 0
CONVERTER 2, SAMPLE 0
CONVERTER 3, SAMPLE 0
D15 ... D0
D15 ... D0
D15 ... D0
D15 ... D0
DAC1
DAC2
DAC0
DAC3
11675-028
4 CONVERTERS
(M = 4)
LANE 0,
OCTET 0
J7
1 OCTET PER LANE
(F = 1)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J15 J14
SERIAL JESD204B DATA (L = 8)
SAMPLES SPLIT ACROSS LANES
(HD = 1)
J1
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x7: L = 8 lanes per link
Register 0x454[7:0] = 0x00: F = 1 octet per frame
Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe
Register 0x456[7:0] = 0x03: M = 4 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0xF: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0
Register 0x46C[7:0] = 0xFF: deskew Link Lane 0 to Link Lane 7
Register 0x476[7:0] = 0x01: F = 1 octet per frame
Register 0x47D[7:0] = 0xFF: enable Link Lane 0 to Link Lane 7
J6
Setting
0x07 or 0x87
0x00
0x1F
0x03
0x0F
0x0F or 0x2F
0x20
0x80
0xFF
0x01
0xFF
J8
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 54. JESD204B Mode 0 Data Deframing
Rev. A | Page 51 of 125
AD9144
Data Sheet
Table 50. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 1
16-BIT NIBBLE GROUP
(N = 16)
2 SAMPLES PER
CONVERTER PER FRAME
(S = 2)
4 CONVERTERS
(M = 4)
L 0,
O1
L 1,
O0
NIBBLE
GROUP 0
CONV 0,
SMPL 0
L 3,
O0
NIBBLE
GROUP 2
CONV 1,
SMPL 0
L 3,
O1
NIBBLE
GROUP 3
CONV 1,
SMPL 1
NIBBLE
GROUP 4
CONV 2,
SMPL 0
NIBBLE
GROUP 5
CONV 2,
SMPL 1
L 6,
O0
J0
L 7,
O0
NIBBLE
GROUP 6
CONV 3,
SMPL 0
J1
SERDINx±
L 6,
O1
J15 J14
J15 J14
SERDINx±
L 5,
O1
J1
J0
J0
L 5,
O0
J1
SERDINx±
L 4,
O1
J15 J14
J1
J0
L 4,
O0
J15 J14
SERDINx±
J1
SERDINx±
L 2,
O1
J15 J14
J0
NIBBLE
GROUP 1
CONV 0,
SMPL 1
L 2,
O0
J1
SERDINx±
L 1,
O1
J15 J14
J1
J15 J14
SERDINx±
J1
SERDINx±
L 0,
O0
L 7,
O1
NIBBLE
GROUP 7
CONV 3,
SMPL 1
D15 ... D0 (0) D15 ... D0 (1)
D15 ... D0 (0) D15 ... D0 (1)
D15 ... D0 (0) D15 ... D0 (1)
D15 ... D0 (0) D15 ... D0 (1)
DAC0
DAC1
DAC2
DAC3
11675-029
2 OCTETS PER LANE
(F = 2)
J15 J14
SERIAL JESD204B DATA (L = 8)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x7: L = 8 lanes per link
Register 0x454[7:0] = 0x01: F = 2 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x03: M = 4 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x1: S = 2 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0xFF: deskew Link Lane 0 to Link Lane 7
Register 0x476[7:0] = 0x02: F = 2 octets per frame
Register 0x47D[7:0] = 0xFF: enable Link Lane 0 to Link Lane 7
J0
Setting
0x07 or 0x87
0x01
0x0F or 0x1F
0x03
0x0F
0x0F or 0x2F
0x21
0x00
0xFF
0x02
0xFF
J0
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 55. JESD204B Mode 1 Data Deframing
Rev. A | Page 52 of 125
Data Sheet
AD9144
Table 51. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 2
4 CONVERTERS
(M = 4)
LANE 0,
OCTET 1
LANE 1,
OCTET 0
LANE 2,
OCTET 0
J0
J1
SERDINx±
LANE 2,
OCTET 1
J15 J14
J1
J0
LANE 1,
OCTET 1
J15 J14
SERDINx±
J1
SERDINx±
J1
SERDINx±
LANE 0,
OCTET 0
LANE 3,
OCTET 0
LANE 3,
OCTET 1
NIBBLE GROUP 0
NIBBLE GROUP 1
NIBBLE GROUP 2
NIBBLE GROUP 3
CONVERTER 0, SAMPLE 0
CONVERTER 1, SAMPLE 0
CONVERTER 2, SAMPLE 0
CONVERTER 3, SAMPLE 0
D15 ... D0 (0)
D15 ... D0 (0)
D15 ... D0 (0)
D15 ... D0 (0)
DAC1
DAC2
DAC0
DAC3
11675-030
2 OCTETS PER LANE
(F = 2)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J15 J14
SERIAL JESD204B DATA (L = 4)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x3: L = 4 lanes per link
Register 0x454[7:0] = 0x01: F = 2 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x03: M = 4 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x0F: deskew Link Lane 0 to Link Lane 3
Register 0x476[7:0] = 0x02: F = 2 octets per frame
Register 0x47D[7:0] = 0x0F: enable Link Lane 0 to Link Lane 3
J15 J14
Setting
0x03 or 0x83
0x01
0x0F or 0x1F
0x03
0x0F
0x0F or 0x2F
0x20
0x00
0x0F
0x02
0x0F
J0
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 56. JESD204B Mode 2 Data Deframing
Rev. A | Page 53 of 125
AD9144
Data Sheet
Table 52. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 3
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x1: L = 2 lanes per link
Register 0x454[7:0] = 0x03: F = 4 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x03: M = 4 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] =0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x03: deskew Link Lane 0 and Link Lane 1
Register 0x476[7:0] = 0x04: F = 4 octets per frame
Register 0x47D[7:0] = 0x03: enable Link Lane 0 and Link Lane 1
J1
J1
SERDINx±
J31 J30
LANE 0,
LANE 0,
OCTET 0
OCTET 1
NIBBLE GROUP 0
LANE 0,
LANE 0,
OCTET 2
OCTET 3
NIBBLE GROUP
GROUP 11
LANE 1,
LANE 1,
OCTET 0
OCTET 1
NIBBLE
NIBBLE GROUP
GROUP 22
LANE 1,
LANE 1,
OCTET 2
OCTET 3
NIBBLE
NIBBLE GROUP
GROUP 33
CONVERTER 0, SAMPLE 0
CONVERTER 1, SAMPLE 0
CONVERTER 2, SAMPLE 0
CONVERTER 3, SAMPLE 0
D15 ... D0 (0)
D15 ... D0 (0)
D15 ... D0 (0)
D15 ... D0 (0)
DAC1
DAC2
DAC0
DAC3
11675-031
4 CONVERTERS
(M = 4)
J17 J16 J15 J14
SERDINx±
4 OCTETS PER LANE
(F = 4)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J31 J30
SERIAL JESD204B DATA (L = 2)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J17 J16 J15 J14
Setting
0x01 or 0x81
0x03
0x0F or 0x1F
0x03
0x0F
0x0F or 0x2F
0x20
0x00
0x03
0x04
0x03
J0
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 57. JESD204B Mode 3 Data Deframing
Rev. A | Page 54 of 125
Data Sheet
AD9144
Table 53. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 4
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Setting
0x03 or 0x83
0x00
0x1F
0x01
0x0F
0x0F or 0x2F
0x20
0x01
0x0F
0x01
0x0F
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x3: L = 4 lanes per link
Register 0x454[7:0] = 0x00: F = 1 octet per frame
Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe
Register 0x456[7:0] = 0x01: M = 2 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x0F: deskew Link Lane 0 to Link Lane 3
Register 0x476[7:0] = 0x01: F = 1 octet per frame
Register 0x47D[7:0] = 0x0F: enable Link Lane 0 to Link Lane 3
See Figure 53 for an illustration of the AD9144 JESD204B Mode 4 data deframing process.
Table 54. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 5
2 CONVERTERS
(M = 2)
LANE 0,
OCTET 0
LANE 1,
OCTET 0
LANE 2,
OCTET 0
J0
LANE 3,
OCTET 0
J1
SERDINx±
LANE 2,
OCTET 1
J15 J14
J1
J0
LANE 1,
OCTET 1
J15 J14
SERDINx±
J1
J15 J14
J1 J0
LANE 0,
OCTET 1
LANE 3,
OCTET 1
NIBBLE GROUP 0
NIBBLE GROUP 1
NIBBLE GROUP 2
NIBBLE GROUP 3
CONVERTER 0, SAMPLE 0
CONVERTER 0, SAMPLE 1
CONVERTER 1, SAMPLE 0
CONVERTER 1, SAMPLE 1
D15 ... D0 (0)
D15 ... D0 (1)
D15 ... D0 (0)
D15 ... D0 (1)
DAC0
DAC1
11675-032
2 OCTETS PER LANE
(F = 2)
16-BIT NIBBLE GROUP
(N = 16)
2 SAMPLES PER
CONVERTER PER FRAME
(S = 2)
SERDINx±
SERIAL JESD204B DATA (L = 4)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled; Register 0x453[4:0] = 0x3: L = 4 lanes per link
Register 0x454[7:0] = 0x01: F = 2 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x01: M = 2 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x1: S = 2 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x0F: deskew Link Lane 0 to Link Lane 3
Register 0x476[7:0] = 0x02: F = 2 octets per frame
Register 0x47D[7:0] = 0x0F: enable Link Lane 0 to Link Lane 3
SERDINx±
Setting
0x03 or 0x83
0x01
0x0F or 0x1F
0x01
0x0F
0x0F or 0x2F
0x21
0x00
0x0F
0x02
0x0F
J15 J14
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 58. JESD204B Mode 5 Data Deframing
Rev. A | Page 55 of 125
AD9144
Data Sheet
Table 55. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 6
J1
DAC0
D0
D1
D4
D5
D6
D7
D9
D10
D11
D12
D13
D14
D15
D0
D1
D2
D3
D4
D5
D6
D7
D8
D9
CONVERTER 1, SAMPLE 0
D10
CONVERTER 0, SAMPLE 0
D2
LANE 1, OCTET 1
D3
SERDINx±
J15 J14
SERDINx±
J1
LANE 1, OCTET 0
NIBBLE GROUP 1
D11
D12
D13
D14
LANE 0, OCTET 1
NIBBLE GROUP 0
DAC1
11675-033
2 CONVERTERS
(M = 2)
LANE 0, OCTET 0
D15
2 OCTETS PER LANE
(F = 2)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J15 J14
SERIAL JESD204B DATA (L = 2)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x1: L = 2 lanes per link
Register 0x454[7:0] = 0x01: F = 2 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x01: M = 2 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x03: deskew Link Lane 0 and Link Lane 1
Register 0x476[7:0] = 0x02: F = 2 octets per frame
Register 0x47D[7:0] = 0x03: enable Link Lane 0 and Link Lane 1
D8
Setting
0x01 or 0x81
0x01
0x0F or 0x1F
0x01
0x0F
0x0F or 0x2F
0x20
0x00
0x03
0x02
0x03
J0
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 59. JESD204B Mode 6 Data Deframing
Rev. A | Page 56 of 125
Data Sheet
AD9144
Table 56. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 7
SERDINx±
J1
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x0: L = 1 lane per link
Register 0x454[7:0] = 0x03: F = 4 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x01: M = 2 converters per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x01: deskew Link Lane 0
Register 0x476[7:0] = 0x04: F = 4 octets per frame
Register 0x47D[7:0] = 0x01: enable Link Lane 0
SERIAL JESD204B DATA (L = 1)
SAMPLES NOT SPLIT
ACROSS LANES
(HD = 0)
J17 J16 J15 J14
4 OCTETS PER LANE
(F = 4)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
2 CONVERTERS
(M = 2)
LANE 0,
LANE 0,
OCTET 0
OCTET 1
NIBBLE GROUP 0
LANE 0,
LANE 0,
OCTET 2
OCTET 3
NIBBLE
NIBBLE GROUP
GROUP 12
CONVERTER 0, SAMPLE 0
CONVERTER 1, SAMPLE 0
D15 ... D0
D15 ... D0
DAC0
DAC1
11675-034
Setting
0x00 or 0x80
0x03
0x0F or 0x1F
0x01
0x0F
0x0F or 0x2F
0x20
0x00
0x01
0x04
0x01
J31 J30
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 60. JESD204B Mode 7 Data Deframing
Rev. A | Page 57 of 125
AD9144
Data Sheet
Table 57. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 9
J1
J6
SERDINx±
LANE 0,
OCTET 0
LANE 1,
OCTET 0
NIBBLE GROUP 0
CONVERTER 0, SAMPLE 0
1 CONVERTER
(M = 1)
D15 ... D0
DAC0
Figure 61. JESD204B Mode 9 Data Deframing
Rev. A | Page 58 of 125
11675-035
1 OCTET PER LANE
(F = 1)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J7
SERIAL JESD204B DATA (L = 2)
SAMPLES SPLIT ACROSS LANES
(HD = 1)
J9
J8
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x1: L = 2 lanes per link
Register 0x454[7:0] = 0x00: F = 1 octet per frame
Register 0x455[4:0] = 0x1F: K = 32 frames per multiframe
Register 0x456[7:0] = 0x00: M = 1 converter per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: Set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 1: HD = 1; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x03: deskew Link Lane 0 and Link Lane 1
Register 0x476[7:0] = 0x01: F = 1 octet per frame
Register 0x47D[7:0] = 0x03: enable Link Lane 0 and Link Lane 1
SERDINx±
Setting
0x01 or 0x81
0x00
0x1F
0x00
0x0F
0x0F or 0x2F
0x20
0x01
0x03
0x01
0x03
J15 J14
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Data Sheet
AD9144
Table 58. SPI Configuration Map—Register Settings for JESD204B Parameters for Mode 10
SERIAL JESD204BB DATA (L = 1)
SAMPLES SPLIT ACROSS LANES
(HD = 0)
2 OCTETS PER LANE
(F = 2)
16-BIT NIBBLE GROUP
(N = 16)
1 SAMPLE PER
CONVERTER PER FRAME
(S = 1)
J1
J0
Description
Register 0x453[7] = 0 or 1: scrambling disabled or enabled, Register 0x453[4:0] = 0x0: L = 1 lane per link
Register 0x454[7:0] = 0x01: F = 2 octets per frame
Register 0x455[4:0] = 0x0F or 0x1F: K = 16 or 32 frames per multiframe
Register 0x456[7:0] = 0x00: M = 1 converter per link
Register 0x457[7:6] = 0x0: always set CS = 0; Register 0x457[4:0] = 0x0F: N = 16, always set to 16-bit resolution
Register 0x458[7:5] = 0x0 or 0x1: Subclass 0 or Subclass 1; Register 0x458[4:0] = 0xF: NP = 16 bits per sample
Register 0x459[7:5] = 0x1: set to JESD204B version; Register 0x459[4:0] = 0x0: S = 1 (sample/converter)/frame
Register 0x45A[7] = 0: HD = 0; Register 0x45A[4:0] = 0x00: always set CF = 0
Register0x46C[7:0] = 0x01: deskew Link Lane 0
Register 0x476[7:0] = 0x02: F = 2 octets per frame
Register 0x47D[7:0] = 0x01: enable Link Lane 0
SERDINx±
LANE 0,
OCTET 0
LANE 1,
OCTET 0
NIBBLE GROUP 0
CONVERTER 0, SAMPLE 0
1 CONVERTER
(M = 1)
D15 ... D0
DAC0
11675-036
Setting
0x00 or 0x80
0x01
0x0F or 0x1F
0x00
0x0F
0x0F or 0x2F
0x20
0x00
0x01
0x02
0x01
J15 J14
Address
0x453
0x454
0x455
0x456
0x457
0x458
0x459
0x45A
0x46C
0x476
0x47D
Figure 62. JESD204B Mode 10 Data Deframing
Rev. A | Page 59 of 125
AD9144
Data Sheet
JESD204B TEST MODES
Transport Layer Testing
PHY PRBS Testing
The JESD204B receiver in the AD9144 supports the short
transport layer (STPL) test as described in the JESD204B
standard. This test can be used to verify the data mapping
between the JESD204B transmitter and receiver. To perform
this test, this function must be implemented in the logic device
and enabled there. Before running the test on the receiver side,
the link must be established and running without errors (see the
Device Setup Guide section).
The JESD204B receiver on the AD9144 includes a PRBS pattern
checker on the back end of its physical layer. This functionality
enables bit error rate (BER) testing of each physical lane of the
JESD204B link. The PHY PRBS pattern checker does not
require that the JESD204B link be established. The pattern checker
can synchronize with a PRBS7, PRBS15, or PRBS31 data pattern.
PRBS pattern verification can be done on multiple lanes
simultaneously. The error counts for failing lanes are reported
for one JESD204B lane at a time. The process for performing
PRBS testing on the AD9144 is as follows:
1.
2.
3.
4.
5.
6.
7.
8.
9.
Start sending a PRBS7, PRBS15, or PRBS31 pattern from
the JESD204B transmitter.
Select and write the appropriate PRBS pattern to
Register 0x316[3:2], as shown in Table 59.
Enable the PHY test for all lanes being tested by writing to
PHY_TEST_EN (Register 0x315). Each bit of Register 0x315
enables the PRBS test for the corresponding lane. For example,
writing a 1 to Bit 0 enables the PRBS test for Physical Lane 0.
Toggle PHY_TEST_RESET (Register 0x316[0]) from 0 to 1
then back to 0.
Set PHY_PRBS_ERROR_THRESHOLD (Register 0x319 to
Register 0x317) as desired.
Write a 0 and then a 1 to PHY_TEST_START
(Register 0x316[1]). The rising edge of PHY_TEST_START
starts the test.
Wait 500 ms.
Stop the test by writing PHY_TEST_START
(Register 0x316[1]) = 0.
Read the PRBS test results.
a. Each bit of PHY_PRBS_PASS (Register 0x31D)
corresponds to one SERDES lane: 0 is fail, 1 is pass.
b. The number of PRBS errors seen on each failing lane can
be read by writing the lane number to check (0 to 7) in
the PHY_SRC_ERR_CNT (Register 0x316[6:4]) and
reading the PHY_PRBS_ERR_COUNT (Register 0x31C
to Register 0x31A). The maximum error count is 224 − 1.
If all bits of Register 0x31C to Register 0x31A are high,
the maximum error count on the selected lane has
been exceeded.
Table 59. PHY PRBS Pattern Selection
PHY_PRBS_PAT_SEL Setting
(Register 0x316[3:2])
0b00 (default)
0b01
0b10
PRBS Pattern
PRBS7
PRBS15
PRBS31
The STPL test ensures that each sample from each converter is
mapped appropriately according to the number of converters (M)
and the number of samples per converter (S). As specified in
the JESD204B standard, the converter manufacturer specifies
what test samples are transmitted. Each sample must have a
unique value. For example, if M = 2 and S = 2, there are 4
unique samples transmitted repeatedly until the test is stopped.
The expected sample must be programmed into the device, and
the expected sample is compared to the received sample one
sample at a time until all have been tested. The process for
performing this test on the AD9144 is as follows:
1.
2.
3.
4.
5.
6.
7.
8.
Synchronize the JESD204B link.
Enable the STPL test at the JESD204B Tx.
Select Converter 0 Sample 0 for testing. Write
SHORT_TPL_DAC_SEL (Register 0x32C[3:2]) = 0 and
SHORT_TPL_SP_SEL (Register 0x32C[5:4]) = 0.
Set the expected test sample for Converter 0, Sample 0.
Program the expected 16-bit test sample into the
SHORT_TPL_REF_SP registers (Register 0x32E and
Register 0x32D).
Enable the STPL test. Write SHORT_TPL_TEST_EN
(Register 0x32C[0]) = 1.
Toggle the STPL reset. SHORT_TPL_TEST_RESET
(Register 0x32C[1]) from 0 to 1 then back to 0.
Check for failures. Read SHORT_TPL_FAIL
(Register 0x32F[0]): 0 is pass, 1 is fail.
Repeat Step 3 to Step 7 for each sample of each converter,
Conv0Sample0 through ConvM − 1SampleS − 1.
Repeated CGS and ILAS Test
As per Section 5.3.3.8.2 of the JESD204B specification, the
AD9144 can check that a constant stream of /K28.5/ characters
is being received, or that CGS followed by a constant stream of
ILAS is being received.
To run a repeated CGS test, send a constant stream of /K28.5/
characters to the AD9144 SERDES inputs. Next, set up the
device and enable the links as described in the Device Setup
Guide section. Ensure that the /K28.5/ characters are being
received by verifying that the SYNCOUTx± has been deasserted and that CGS has passed for all enabled link lanes by
reading Register 0x470. Program Register 0x300[2] = 0 to
monitor the status of lanes on Link 0, and Register 0x300[2] = 1
to monitor the status of lanes on Link 1 for dual-link mode.
Rev. A | Page 60 of 125
Data Sheet
AD9144
To run the CGS followed by a repeated ILAS sequence test,
follow the Device Setup Guide section; however, before
performing the last write (enabling the links), enable the ILAS
test mode by writing a 1 to Register 0x477[7]. Then, enable the
links. When the device recognizes four CGS characters on each
lane, it de-asserts the SYNCOUTx±. At this point, the
transmitter starts sending a repeated ILAS sequence.
Read Register 0x473 to verify that initial lane synchronization has
passed for all enabled link lanes. Program Register 0x300[2] = 0
to monitor the status of lanes on Link 0, and Register 0x300[2] = 1
to monitor the status of lanes on Link 1 for dual-link mode.
JESD204B ERROR MONITORING
Check for Error Count Over Threshold
In addition to reading the error count per lane and error type as
described in the Checking Error Counts section, the user can
check a register to see if the error count for a given error type
has reached a programmable threshold.
The same error threshold is used for the three error types
(disparity, not in table, and unexpected control character). The
error counters are on a per error type basis. To use this feature,
complete the following steps:
1.
2.
Disparity, Not in Table, and Unexpected Control
Character Errors
As per Section 7.6 of the JESD204B specification, the AD9144
can detect disparity errors, not in table errors, and unexpected
control character errors, and can optionally issue a sync request
and reinitialize the link when errors occur.
Note that the disparity error counter counts all characters with
invalid disparity, regardless of whether they are in the 8-bit/10-bit
decoding table. This is a minor deviation from the JESD204B
specification, which only counts disparity errors when they are
in the 8-bit/10-bit decoding table.
Checking Error Counts
The error count can be checked for disparity errors, not in table
errors, and unexpected control character errors. The error counts
are on a per lane and per error type basis. Note that the lane select
and counter select are programmed into Register 0x46B, and the
error count is read back from the same address. To check the
error count, complete the following steps:
1.
2.
Select the desired link lane and error type of the counter to
view. Write these to Register 0x46B according to Table 60.
To select a link lane, first select a link (Register 0x300[2] =
0 to select Link 0 or Register 0x300[2] = 1 to select Link 1
(dual-link only)).
Note that when using Link 1, Link Lane x refers to Logical
Lane x + 4.
Read the error count from Register 0x46B. Note that the
maximum error count is equal to the error threshold set in
Register 0x47C.
Table 60. Error Counters
Addr.
0x46B
Bits
[6:4]
[1:0]
Variable
LaneSel
CntrSel
Description
LaneSel = x to monitor the error
count of Link Lane x. See the notes
on link lane in Step 1 of the Checking
Error Counts section.
CntrSel = 0b00 for bad running
disparity counter.
CntrSel = 0b01 for not in table error
counter.
CntrSel = 0b10 for unexpected control
character counter.
Program the desired error count threshold into
ERRORTHRES (Register 0x47C).
Read back the error status for each error type to see if the
error count has reached the error threshold.
 Disparity errors are reported in Register 0x46D.
 Not in table errors are reported in Register 0x46E.
 Unexpected control characters are reported in
Register 0x46F.
Error Counter and IRQ Control
The user can write to Register 0x46D and Register 0x46F to
reset or disable the error counts and to reset the IRQ for a given
lane. Note that these are the same registers that are used to
report error count over threshold (see the Check for Error
Count Over Threshold section); therefore, the readback is not
the value that was written. For each error type
1.
2.
3.
Select the link lane to access. To select a link lane, first
select a link (Register 0x300[2] = 0 to select Link 0,
Register 0x300[2] = 1 to select Link 1 (dual-link only)).
Note that when using Link 1, Link Lane x refers to Logical
Lane x + 4.
Decide whether to reset the IRQ, disable the error count,
and/or reset the error count for the given lane and error
type.
Write the link lane and desired reset or disable action to
Register 0x46D to Register 0x46F according to Table 61.
Table 61. Error Counter and IRQ Control: Disparity
(Register 0x46D), Not In Table (Register 0x46E), Unexpected
Control Character (Register 0x46F)
Bits
7
Variable
RstIRQ
6
Disable_ErrCnt
5
RstErrCntr
[2:0]
LaneAddr
Rev. A | Page 61 of 125
Description
RstIRQ = 1 to reset IRQ for the lane
selected in Bits[2:0].
Disable_ErrCnt = 1 to disable the error
count for the lane selected in Bits[2:0].
RsteErrCntr = 1 to reset the error
count for the lane selected in Bits[2:0].
LaneAddr = x to monitor the error
count of Link Lane x. See the notes on
link lane in Step 1 of the Checking
Error Counts section.
AD9144
Data Sheet
Table 63. Sync Assertion Mask
Monitoring Errors via SYNCOUTx±
When one or more disparity, not in table, or unexpected control
character error occurs, the error is reported on the SYNCOUTx±
pins as per Section 7.6 of the JESD204B specification. The
JESD204B specification states that the SYNCOUTx± signal is
asserted for exactly 2 frame periods when an error occurs. For the
AD9144, the width of the SYNCOUTx± pulse can be programmed
to ½, 1, or 2 PClock cycles. The settings to achieve a SYNCOUTx±
pulse of 2 frame clock cycles are given in Table 62.
Addr.
0x47B
Table 62. Setting SYNCOUTx± Error Pulse Duration
JESD Mode
IDs
0, 4, 9
1, 2, 5, 6, 10
3, 7
1
PClockFactor
(Frames/PClock)
4
2
1
SYNCB_ERR_DUR
(Register 0x312[5:4]) Setting1
0 (default)
1
2
These register settings assert the SYNCOUTx± signal for 2 frame clock cycles
pulse widths.
Disparity, NIT, Unexpected Control Character IRQs
For disparity, not in table, and unexpected control character
errors, error count over the threshold events are available as
IRQ events. Enable these events by writing to Register 0x47A[7:5].
The IRQ event status can be read at the same address
(Register 0x47A[7:5]) after the IRQs are enabled.
See the Error Counter and IRQ Control section for information
on resetting the IRQ. See the Interrupt Request Operation
section for more information on IRQs.
Errors Requiring Reinitializing
A link reinitialization automatically occurs when four invalid
disparity characters are received, as per Section 7.1 of the
JESD204B specification. When a link reinitialization occurs, the
resync request is 5 frames and 9 octets long.
The user can optionally reinitialize the link when the error
count for disparity errors, not in table errors, or unexpected
control characters reaches a programmable error threshold. The
process to enable the reinitialization feature for certain error
types is as follows:
1.
2.
3.
4.
Set THRESHOLD_MASK_EN (Register 0x477[3]) = 1.
Note that when this bit is set, unmasked errors do not
saturate at either the threshold or maximum value.
Enable the sync assertion mask for each type of error by
writing to SYNC_ASSERTION_MASK (Register 0x47B[7:5])
according to Table 63.
Program the desired error counter threshold into
ERRORTHRES (Register 0x47C).
For each error type enabled in the SYNC_ASSERTION_
MASK register, if the error counter on any lane reaches the
programmed threshold, SYNCOUTx± falls, issuing a sync
request. Note that all error counts are reset when a link
reinitialization occurs. The IRQ does not reset and must be
reset manually.
Bit No.
7
Bit Name
BADDIS_S
6
NIT_S
5
UCC_S
Description
Set to 1 to assert SYNCOUTx±
if the disparity error count
reaches the threshold
Set to 1 to assert SYNCOUTx±
if the not in table error count
reaches the threshold
Set to 1 to assert SYNCOUTx±
if the unexpected control
character count reaches the
threshold
CGS, Frame Sync, Checksum, and ILAS Monitoring
Register 0x470 to Register 0x473 can be monitored to verify
that each stage of JESD204B link establishment has occurred.
Program Register 0x300[2] = 0 to monitor the status of the
lanes on Link 0, and Register 0x300[2] = 1 to monitor the status
of the lanes on Link 1.
Bit x of CODEGRPSYNCFLAG (Register 0x470) is high if Link
Lane x received at least four K28.5 characters and passed code
group synchronization.
Bit x of FRAMESYNCFLAG (Register 0x471) is high if Link
Lane x completed initial frame synchronization.
Bit x of GOODCHKSUMFLG (Register 0x472) is high if the
checksum sent over the lane matches the sum of the JESD204B
parameters sent over the lane during ILAS for Link Lane x. The
parameters can be added either by summing the individual fields
in registers or summing the packed register. If Register 0x300[6] =
0 (default), the calculated checksums are the lower 8 bits of the
sum of the following fields: DID, BID, LID, SCR, L − 1, F − 1, K − 1,
M − 1, N − 1, SUBCLASSV, NP − 1, JESDV, S − 1, and HD. If
Register 0x300[6] = 1, the calculated checksums are the lower
8 bits of the sum of Register 0x400 to Register 0x40C and LID.
Bit x of INITIALLANESYNC (Register 0x473) is high if Link
Lane x passed the initial lane alignment sequence.
CGS, FrameSync, Checksum, and ILAS IRQs
Fail signals for CGS, FrameSync, CheckSum, and ILAS are
available as IRQ events. Enable them by writing to
Register 0x47A[3:0]. The IRQ event status can be read at the
same address (Register 0x47A[3:0]) after the IRQs are enabled.
Write a 1 to Register 0x470[7] to reset the CGS IRQ. Write a 1
to Register 0x471 to reset the FrameSync IRQ. Write a 1 to
Register 0x472 to reset the CheckSum IRQ. Write a 1 to
Register 0x473 to reset the ILAS IRQ.
See the Interrupt Request Operation section for more
information.
Rev. A | Page 62 of 125
Data Sheet
AD9144
Configuration Mismatch IRQ
Power and Ground Planes
The AD9144 has a configuration mismatch flag that is available
as an IRQ event. Use Register 0x47B[3] to enable the mismatch
flag (it is enabled by default), and then use Register 0x47B[4] to
read back its status and reset the IRQ signal. See the Interrupt
Request Operation section for more information.
Solid ground planes are recommended to avoid ground loops
and to provide a solid, uninterrupted ground reference for the
high speed transmission lines that require controlled impedances.
Do not use segmented power planes as a reference for controlled
impedances unless the entire length of the controlled impedance
trace traverses across only a single segmented plane. These and
additional guidelines for the topology of high speed transmission
lines are described in the JESD204B Serial Interface Inputs
(SERDIN0± to SERDIN7±) section.
The configuration mismatch event flag is high when the link
configuration settings (in Register 0x450 to Register 0x45D) do
not match the JESD204B transmitted settings (Register 0x400 to
Register 0x40D). All these registers are paged per link (in
Register 0x300).
Note that this function is different from the good checksum
flags in Register 0x472. The good checksum flags ensure that
the transmitted checksum matches a calculated checksum based
on the transmitted settings. The configuration mismatch event
ensures that the transmitted settings match the configured settings.
HARDWARE CONSIDERATIONS
Power Supply Recommendations
The power supply domains are described in Table 64. The
power supplies can be grouped into separate PCB domains,
as show in Figure 63. All the AD9144 supply domains must
remain as noise free as possible for the best operation. Power
supply noise has a frequency component that affects performance,
and is specified in terms of V rms. Figure 64 shows the
recommended power supply components.
An LC filter on the output of the power supply is recommended
to attenuate the noise, and must be placed as close to the AD9144
as possible. An effective filter is shown in Figure 63. This filter
scheme reduces high frequency noise components. Each of the
power supply pins of the AD9144 must also have a 0.1 μF capacitor
connected to the ground plane, as shown in Figure 63. Place the
capacitor as close to the supply pin as possible. Adjacent power
pins can share a bypass capacitor. Connect the ground pins of
the AD9144 to the ground plane using vias.
Table 64. Power Supplies
Supply Domain
DVDD121
PVDD122
SVDD123
CVDD121
IOVDD
VTT4
SIOVDD33
AVDD33
1
Voltage (V)
1.2
1.2
1.2
1.2
1.8
1.2
3.3
3.3
Circuitry
Digital core
DAC PLL
JESD204B receiver interface
DAC clocking
SPI interface
VTT
Sync LVDS transmit
DAC
This supply requires a 1.3 V supply when operating at maximum DAC sample
rates. See Table 3 for details.
2
This supply can be combined with CVDD12 on the same regulator with a
separate supply filter network and sufficient bypass capacitors near the pins.
3
This supply requires a 1.3 V supply when operating at maximum interface
rates. See Table 4 for details.
4
This supply can be connected to SVDD12 and does not need separate
circuitry.
Rev. A | Page 63 of 125
AD9144
Data Sheet
1.8V FPGA VCCIO
OR
OTHER SYSTEM SUPPLY
10µH
IOVDD
66
10µF
1.2V
LINEAR
REGULATOR
DVDD12
13
10µF
14
53
3.3V
LINEAR
REGULATOR
SIOVDD33
35
10µF
AVDD33
67
75
1.2V
LINEAR
REGULATOR
SVDD12
17
10µF
77
20
85
22
28
1.2V
LINEAR
REGULATOR
10µH
31
CVDD12
10µF
71
32
76
33
81
36
87
39
1
45
10µH
PVDD12
10µF
47
4
50
7
VTT
8
21
10µF
9
25
10
42
56
46
Figure 63. JESD204B Interface PCB Power Domain Recommendation
AD9144
1.8V
POWER
INPUT
+12V
ADP1741
ADP2119
ADP1741
ADP1753
BUCK
1.2MHz/600kHz
800mA
3.8V
ADM7154-3.3
1.2V
1.2V
SVDD12 + V TT
DVDD12
CVDD12 + PVDD12
3.3V
AVDD33
3.3V
IOVDD + SIOVDD33
ADP2370
ADM7160-3.3
Figure 64. Power Supply Connections
Rev. A | Page 64 of 125
11675-464
+3.3V
STEP-DOWN DDC
1.2MHz, 2A
1.2V
11675-039
57
NOTES
1. UNLABELED CAPACITORS ARE 0.1µF, AS CLOSE AS POSSIBLE TO DEVICE PIN(S), WITH MINIMUM DISTANCE
AND VIAS BETWEEN CAPACITORS AND PIN(S).
Data Sheet
AD9144
JESD204B Serial Interface Inputs (SERDIN0± to SERDIN7±)
Return Loss
When considering the layout of the JESD204B serial interface
transmission lines, there are many factors to consider to
maintain optimal link performance. Among these factors are
insertion loss, return loss, signal skew, and the topology of the
differential traces.
The JESD204B specification limits the amount of return loss
allowed in a converter device and a logic device, but does not
specify return loss for the channel. However, every effort must
be made to maintain a continuous impedance on the transmission
line between the transmitting logic device and the AD9144. As
mentioned in the Insertion Loss section, minimizing the use of
vias, or eliminating them altogether, reduces one of the primary
sources for impedance mismatches on a transmission line.
Maintain a solid reference beneath (for microstrip) or above and
below (for stripline) the differential traces to ensure continuity in
the impedance of the transmission line. If the stripline technique
is used, follow the guidelines listed in the Insertion Loss section
to minimize impedance mismatches and stub effects.
The JESD204B specification limits the amount of insertion loss
allowed in the transmission channel (see Figure 39). The AD9144
equalization circuitry allows significantly more loss in the channel
than is required by the JESD204B specification. It is still important
that the designer of the PCB minimize the amount of insertion
loss by adhering to the following guidelines:



Keep the differential traces short by placing the AD9144 as
near to the transmitting logic device as possible and routing
the trace as directly as possible between the devices.
Route the differential pairs on a single plane using a solid
ground plane as a reference.
Use a PCB material with a low dielectric constant (<4) to
minimize loss, if possible.
When choosing between the stripline and microstrip
techniques, keep in mind the following considerations: stripline
has less loss (see Figure 40 and Figure 41) and emits less EMI,
but requires the use of vias that can add complexity to the task
of controlling the impedance; whereas microstrip is easier to
implement if the component placement and density allow
routing on the top layer, and eases the task of controlling the
impedance.
If using the top layer of the PCB is problematic or the advantages
of stripline are desirable, follow these recommendations:


Minimize the number of vias.
If possible, use blind vias to eliminate via stub effects and
use micro vias to minimize via inductance.
If using standard vias, use the maximum via length to
minimize the stub size. For example, on an 8-layer board,
use Layer 7 for the stripline pair (see Figure 65).
For each via pair, place a pair of ground vias adjacent to them
to minimize the impedance discontinuity (see Figure 65).
LAYER 1
LAYER 2
ADD GROUND VIAS
y
DIFF–
LAYER 3
LAYER 4
GND
y
STANDARD VIA
LAYER 5
y
DIFF+
LAYER 6
GND
LAYER 7
LAYER 8
MINIMIZE STUB EFFECT
Figure 65. Minimizing Stub Effect and Adding Ground Vias for
Differential Stripline Traces
11675-040


Another primary source for impedance mismatch is at either
end of the transmission line, where care must be taken to match
the impedance of the termination to that of the transmission
line. The AD9144 handles this internally with a calibrated
termination scheme for the receiving end of the line. See the
Interface Power-Up and Input Termination section for details
on this circuit and the calibration routine.
Signal Skew
There are many sources for signal skew, but the two sources to
consider when laying out a PCB are interconnect skew within a
single JESD204B link and skew between multiple JESD204B
links. In each case, keeping the channel lengths matched to
within 15 mm is adequate for operating the JESD204B link at
speeds of up to 10.6 Gbps. Managing the interconnect skew
within a single link is fairly straightforward. Managing multiple
links across multiple devices is more complex. However, follow
the 15 mm guideline for length matching.
Topology
Structure the differential SERDINx± pairs to achieve 50 Ω to
ground for each half of the pair. Stripline vs. microstrip tradeoffs are described in the Insertion Loss section. In either case,
it is important to keep these transmission lines separated from
potential noise sources such as high speed digital signals and
noisy supplies. If using stripline differential traces, route them
using a coplanar method, with both traces on the same layer.
Although this does not offer more noise immunity than the
broadside routing method (traces routed on adjacent layers),
it is easier to route and manufacture so that the impedance
continuity is maintained. An illustration of broadside vs.
coplanar differential routing techniques is shown in Figure 66.
Tx DIFF A
Tx DIFF B
Tx ACTIVE
BROADSIDE DIFFERENTIAL Tx LINES
Tx
DIFF A
Tx
DIFF B
Tx
ACTIVE
COPLANAR DIFFERENTIAL Tx LINES
Figure 66. Broadside vs. Coplanar Differential Stripline Routing Techniques
Rev. A | Page 65 of 125
11675-041
Insertion Loss
AD9144
Data Sheet
SYNCOUTx±, SYSREF±, and CLK± Signals
When considering the trace width vs. copper weight and
thickness, the speed of the interface must be considered. At
multigigabit speeds, the skin effect of the conducting material
confines the current flow to the surface. Maximize the surface
area of the conductor by making the trace width wider to
reduce the losses. Additionally, loosely couple differential traces
to accommodate the wider trace widths. This helps to reduce
the crosstalk and minimize the impedance mismatch when the
traces must separate to accommodate components, vias,
connectors, or other routing obstacles. Tightly coupled vs.
loosely coupled differential traces are shown in Figure 67.
TIGHTLY COUPLED
DIFFERENTIAL Tx LINES
Tx
DIFF A
Tx
DIFF B
LOOSELY COUPLED
DIFFERENTIAL Tx LINES
It is important to keep similar trace lengths for the CLK± and
SYSREF± signals from the clock source to each of the devices
on either end of the JESD204B links (see Figure 68). If using a
clock chip that can tightly control the phase of CLK± and
SYSREF±, the trace length matching requirements are greatly
reduced.
Figure 67. Tightly Coupled vs. Loosely Coupled Differential Traces
AC Coupling Capacitors
The AD9144 requires that the JESD204B input signals be
ac-coupled to the source. These capacitors must be 100 nF and
placed as close as possible to the transmitting logic device. To
minimize the impedance mismatch at the pads, select the
package size of the capacitor so that the pad size on the PCB
matches the trace width as closely as possible.
LANE 0
LANE 1
Tx
DEVICE
Rx
DEVICE
LANE N – 1
LANE N
SYSREF
DEVICE CLOCK
SYSREF
CLOCK SOURCE
(AD9516-1, ADCLK925)
SYSREF TRACE LENGTH
DEVICE CLOCK TRACE LENGTH
DEVICE CLOCK
SYSREF TRACE LENGTH
DEVICE CLOCK TRACE LENGTH
Figure 68. SYSREF Signal and Device Clock Trace Length
Rev. A | Page 66 of 125
11675-043
Tx
DIFF B
Separate the SYNCOUTx± signal from other noisy signals,
because noise on the SYNCOUTx± might be interpreted as a
request for K characters. The SYNCOUTx± signal has two
modes of operation available for use. Register 0x2A5[0] defaults
to 0, which sets the SYNCOUTx± swing to normal swing mode.
When this bit is set to 1, the SYNCOUTx± swing is configured
for high swing mode. For more details, see Table 8.
11675-042
Tx
DIFF A
The SYNCOUTx± and SYSREF± signals on the AD9144 are
low speed LVDS differential signals. Use controlled impedance
traces routed with 100 Ω differential impedance and 50 Ω to
ground when routing these signals. As with the SERDIN0± to
SERDIN7± data pairs, it is important to keep these signals
separated from potential noise sources such as high speed
digital signals and noisy supplies.
Data Sheet
AD9144
INPUT
POWER
DETECTION
AND
PROTECTION
INTERPOLATION
MODES
1×, 2×, 4×, 8×
COARSE
AND
FINE
MODULATION
INV
SINC
DIGITAL GAIN,
PHASE OFFSET,
DC OFFSET,
AND
GROUP DELAY
ADJUSTMENT
11675-049
DIGITAL DATAPATH
Figure 69. Block Diagram of Digital Datapath
The block diagram in Figure 69 shows the functionality of the
digital datapath (all blocks can be bypassed). The digital
processing includes an input power detection block; three halfband interpolation filters; a quadrature modulator consisting of a
fine resolution NCO and fDAC/4 and fDAC/8 coarse modulation
block; an inverse sinc filter; and gain, phase, offset, and group
delay adjustment blocks.
The interpolation filters take independent I and Q data streams.
If using the modulation function, I and Q must be quadrature
data to function properly.
Note that the pipeline delay changes when digital datapath
functions are enabled/disabled. If fixed DAC pipeline latency
is desired, do not reconfigure these functions after initial
configuration.
DUAL PAGING
Digital datapath registers are paged to allow configuration of
either DAC dual independently or both simultaneously. Table 65
shows how to use the dual paging register.
Table 65. Paging Modes
Duals
Paged
A
B
A and B
The transmit path contains three half-band interpolation filters,
which each provides a 2× increase in output data rate and a lowpass function. The filters can be cascaded to provide a 4× or 8×
interpolation ratio. Table 66 shows how to select each available
interpolation mode, their usable bandwidths, and their
maximum data rates. Note that fDATA = fDAC/InterpolationFactor.
Interpolation mode is paged as described in the Dual Paging
section. Register 0x030[0] is high if an unsupported
interpolation mode is selected.
Table 66. Interpolation Modes and Usable Bandwidth
Interpolation
Mode
1× (Bypass)
2×
4×
8×
1
INTERP_MODE,
Reg 0x112[2:0]
0x00
0x01
0x03
0x04
Usable
Bandwidth
0.5 × fDATA
0.4 × fDATA
0.4 × fDATA
0.4 × fDATA
Maximum
fDATA (MHz)
10601
10601
700
350
The maximum speed for 1× and 2× interpolation is limited by the JESD204B
interface.
Filter Performance
DACs Updated
DAC0 and DAC1
DAC2 and DAC3
DAC0, DAC1, DAC2, and DAC3
Several functions are paged by DAC dual, such as input data
format, downstream protection, interpolation, modulation,
inverse sinc, digital gain, phase offset, dc offset, group delay, IQ
swap, datapath PRBS, LMFC sync, and NCO alignment.
The interpolation filters interpolate between existing data in
such a way that they minimize changes in the incoming data
while suppressing the creation of interpolation images. This is
shown for each filter in Figure 70.
The usable bandwidth (as shown in Table 66) is defined as the
frequency band over which the filters have a pass-band ripple of
less than ±0.001 dB and an image rejection of greater than 85 dB.
0
DATA FORMAT
2×
4×
8×
MAGNITUDE (dB)
–20
BINARY_FORMAT (Register 0x110[7], paged as described in
the Dual Paging section) controls the expected input data
format. By default it is 0, which means that the input data must
be in twos complement. It can also be set to 1, which means
input data is in offset binary (0x0000 is negative full scale and
0xFFFF is positive full scale).
–40
–60
–80
–100
0
0.2
0.4
0.6
0.8
FREQUENCY (×fDAC )
Figure 70. All Band Responses of Interpolation Filters
Rev. A | Page 67 of 125
1.0
11675-368
DUAL_PAGE,
Reg. 0x008[1:0]
1
2
3 (default)
INTERPOLATION FILTERS
AD9144
Data Sheet
The NCO produces a quadrature carrier to translate the input
signal to a new center frequency. A quadrature carrier is a pair of
sinusoidal waveforms of the same frequency, offset 90° from
each other. The frequency of the quadrature carrier is set via an
FTW. The quadrature carrier is mixed with the I and Q data and
then summed into the I and Q datapaths, as shown in Figure 72.
Filter Performance Beyond Specified Bandwidth
0
80
–0.1
70
–0.2
60
–0.3
50
–0.4
40
30
20
40
–0.5
IMAGE REJECTION
PASS-BAND RIPPLE
41
42
43
44
BANDWIDTH (% fDATA )
45
–0.6
MAXIMUM PASS-BAND RIPPLE (dB)
90
11675-369
MINIMUM INTERPOLATION IMAGE REJECTION (dB)
The interpolation filters are specified to 0.4 × fDATA (with pass
band). The filters can be used slightly beyond this ratio at the
expense of increased pass-band ripple and decreased
interpolation image rejection.
Figure 71. Interpolation Filter Performance Beyond Specified Bandwidth
−fDAC/2 ≤ fCARRIER < +fDAC/2
FTW = (fCARRIER/fDAC) × 248
where FTW is a 48-bit twos complement number.
The frequency tuning word is set as shown in Table 68 and
paged as described in the Dual Paging section.
Table 68. NCO FTW Registers
Address
0x114
0x115
0x116
0x117
0x118
0x119
Value
FTW[7:0]
FTW[15:8]
FTW[23:16]
FTW[31:24]
FTW[39:32]
FTW[47:40]
Description
8 LSBs of FTW
Next 8 bits of FTW
Next 8 bits of FTW
Next 8 bits of FTW
Next 8 bits of FTW
8 MSBs of FTW
Figure 71 shows the performance of the interpolation filters
beyond 0.4 × fDATA. Note that the ripple increases much slower
than the image rejection decreases. This means that if the
application can tolerate degraded image rejection from the
interpolation filters, more bandwidth can be used.
Unlike other registers, the FTW registers are not updated
immediately upon writing. Instead, the FTW registers update on
the rising edge of FTW_UPDATE_REQ (Register 0x113[0]).
After an update request, FTW_UPDATE_ACK (Register 0x113[1])
must be high to acknowledge that the FTW has updated.
DIGITAL MODULATION
SEL_SIDEBAND (Register 0x111[1], paged as described in the
Dual Paging section) is a convenience bit that can be set to use
the negative modulation result. This is equivalent to flipping the
sign of FTW.
The AD9144 has digital modulation features to modulate the
baseband quadrature signal to the desired DAC output frequency.
The coarse modulation modes (fDAC/4 and fDAC/8) allow
modulation by those particular frequencies. The NCO fine
modulation mode allows modulating by a programmable
frequency at the cost of 30 mW to 120 mW, depending on the
DAC rate. Modulation mode is selected as shown in Table 67
and paged as described in the Dual Paging section.
I DATA
COS(ωn + θ)
ω
π
NCO
θ
SIN(ωn + θ)
FTW[47:0]
NCO_PHASE_OFFSET
[15:0]
Table 67. Modulation Mode Selection
MODULATION_TYPE,
Register 0x111[3:2]
0b00
0b01
0b10
0b11
OUT_I
–
OUT_Q
+
–1
SEL_SIDEBAND
Q DATA
0
1
INTERPOLATION
NCO Fine Modulation
This modulation mode uses an NCO, a phase shifter, and a
complex modulator to modulate the signal by a programmable
carrier signal as shown in Figure 72. This allows output signals to
be placed anywhere in the output spectrum with very fine
frequency resolution.
Rev. A | Page 68 of 125
Figure 72. NCO Modulator Block Diagram
11675-056
Modulation Mode
None
NCO Fine Modulation
Coarse − fDAC/4
Coarse − fDAC/8
INTERPOLATION
Data Sheet
AD9144
NCO Phase Offset
The phase offset feature allows rotation of the I and Q phases.
Unlike phase adjust, this feature moves the phases of both I and
Q channels together. Phase offset can be used only when using
NCO fine modulation.
−180° ≤ DegreesOffset < +180°
PhaseOffset = (DegreesOffset/180°) × 215
where PhaseOffset is a 16-bit twos complement number.
The NCO phase offset is set as shown in Table 69 and paged as
described in the Dual Paging section. Because this function is
part of the fine modulation block, phase offset is not updated
immediately upon writing. Instead, it updates on the rising edge
of FTW_ UPDATE_REQ (Register 0x113[0]) along with the FTW.
Table 69. NCO Phase Offset Registers
Address
0x11A
0x11B
DIGITAL GAIN, PHASE ADJUST, DC OFFSET, AND
GROUP DELAY
Digital gain, phase adjust, and dc offset (as described in the
Digital Gain section, Phase Adjust section, and DC Offset
section) allow compensation of imbalances in the I and Q paths
due to analog mismatches between DAC I/Q outputs, quadrature
modulator I/Q baseband inputs, and DAC/modulator interface
I/Q paths. These imbalances can cause the two following issues:


Value
PhaseOffsetI[7:0]
PhaseOffset[15:8]
An unwanted sideband signal to appear at the quadrature
modulator output with significant energy. This can be
tuned out using digital gain and phase adjust. Tuning the
quadrature gain and phase adjust values can optimize
complex image rejection in single sideband radios or can
optimize the error vector magnitude (EVM) in zero IF
(ZIF) architectures.
The I/Q mismatch can cause LO leakage through a
modulator, which can be tuned out using dc offset.
Group delay allows adjustment of the delay through the DAC,
which can be used to adjust digital predistortion (DPD) loop delay.
INVERSE SINC
Digital Gain
The AD9144 provides a digital inverse sinc filter to compensate
the DAC roll-off over frequency. The filter is enabled by setting
the INVSINC_ENABLE bit (Register 0x111[7], paged as
described in the Dual Paging section) and is enabled by default.
The inverse sinc (sinc−1) filter is a seven-tap FIR filter. Figure 73
shows the frequency response of sin(x)/x roll-off, the inverse
sinc filter, and the composite response. The composite response
has less than ±0.05 dB pass-band ripple up to a frequency of
0.4 × fDACCLK. To provide the necessary peaking at the upper end
of the pass band, the inverse sinc filter shown has an intrinsic
insertion loss of approximately 3.8 dB; in many cases, this can be
partially compensated as described in the Digital Gain section.
1
Digital gain can be used to independently adjust the digital
signal magnitude being fed into each DAC. This is useful to
balance the gain between I and Q channels of a dual or to cancel
out the insertion loss of the inverse sinc filter. Digital gain must
be enabled when using the blanking state machine (see the
Downstream Protection section). If digital gain is disabled,
TXENx must be tied high.
Digital gain is enabled by setting the DIG_GAIN_ENABLE bit
(Register 0x111[5], paged as described in the Dual Paging
section). In addition to enabling the function, the amount of
digital gain (GainCode) desired must be programmed. By
default, digital gain is enabled and GainCode is 0xAEA.
SIN(x)/x ROLL-OFF
SINC–1 FILTER RESPONSE
COMPOSITE RESPONSE
0
0 ≤ Gain ≤ 4095/2048
−∞ dB ≤ dBGain ≤ 6.018 dB
MAGNITUDE (dB)
Gain = GainCode × (1/2048)
–1
dBGain = 20 × log10(Gain)
GainCode = 2048 × Gain = 2048 × 10dBGain/20
–2
where GainCode is a 12-bit unsigned binary number.
–3
The I/Q digital gain is set as shown in Table 70 and paged as
described in the Dual Paging section.
–5
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
FREQUENCY (× fDAC )
0.40
0.45
0.50
11675-058
–4
Figure 73. Responses of sin(x)/x Roll-Off, the Sinc−1 Filter, and the Composite
of the Two Input Signal Power Detection and Protection
The default GainCode (0xAEA = 2.7 dB), is appropriate to
counteract the insertion loss of the inverse sinc filter without
causing digital clipping when using 2× interpolation. This value
can be read off of Figure 73 at 0.25 × fDAC, as that is the Nyquist
rate when using a 2× interpolation. Recommended GainCode
values for 4× and 8× interpolation are 0xBB3 (3.3 dB) and
0xBF8 (3.5 dB), respectively.
Rev. A | Page 69 of 125
AD9144
Data Sheet
Group Delay
Table 70. Digital Gain Registers
Address
0x111[5]
0x13C
0x13D[3:0]
0x13E
0x13F[3:0]
Value
DIG_GAIN_ENABLE
GainCodeI[7:0]
GainCodeI[11:8]
GainCodeQ[7:0]
GainCodeQ[11:8]
Description
Set to 1 to enable digital gain
I DAC LSB gain code
I DAC MSB gain code
Q DAC LSB gain code
Q DAC MSB gain code
Group delay can be used to delay both I and Q channels together.
This can be useful, for example, for DPD loop delay adjust.
−4 ≤ DACClockCycles ≤ 3.5
GroupDelay = (DACClockCycles × 2) + 8
where GroupDelay is a 4-bit twos complement number.
Write GroupDelay to GROUP_DELAY (Register 0x014). This
feature is paged as described in the Dual Paging section.
Phase Adjust
Ordinarily, the I and Q channels of each DAC pair have an
angle of 90° between them. The phase adjust feature changes the
angle between the I and Q channels, which can help balance the
phase into a modulator.
−14 ≤ DegreesAdjust < 14
PhaseAdj = (DegreesAdjust/14) × 212
I TO Q SWAP
I_TO_Q (Register 0x111[0], paged as described in the Dual
Paging section) is a convenience bit that can be set to send the
I datapath to the Q DAC. Note that this operation occurs at the
end of the datapath (after any modulation, digital gain, phase
adjust, and phase offset).
where PhaseAdj is a 13-bit twos complement number.
NCO ALIGNMENT
The phase adjust is set as shown in Table 71 and paged as
described in the Dual Paging section.
The NCO alignment block is used to phase align the NCO
output from multiple converters. Two NCO alignment modes
are supported by the AD9144. The first is a SYSREF± alignment
mode that phase aligns the NCO outputs to the rising edge of a
SYSREF± pulse. The second alignment mode is a data key
alignment; when this mode is enabled, the AD9144 aligns the
NCO outputs when a user specified data pattern arrives at the
DAC input. Note that the NCO alignment is per dual, and is
paged as described in the Dual Paging section.
Table 71. I/Q Phase Adjustment Registers
Address
0x111[4]
Value
PHASE_ADJ_ENABLE
0x11C
0x11D[4:0]
PhaseAdj[7:0]
PhaseAdj[12:8]
Description
Set to 1 to enable phase
adjust
LSB phase adjust code
MSB phase adjust code
DC Offset
SYSREF± NCO Alignment
The dc offset feature is used to individually offset the data into
the I or Q DACs. This feature can be used to cancel LO leakage.
As with the LMFC alignment, in Subclass 1, a SYSREF± pulse
can be used to phase align the NCO outputs of multiple devices
in a system and multiple channels on the same device. Note that
in Subclass 0, this alignment mode can be used to align the
NCO outputs within a device to an internal processing clock
edge. No SYSREF± edge is needed in Subclass 0, but multichip
alignment cannot be achieved. The steps to achieve a SYSREF
NCO alignment are as follows:
The offset is programmed individually for I and Q as a 16-bit
twos complement number in LSBs, plus a 5-bit twos complement
number in sixteenths of an LSB, as shown in Table 72. DC offset
is paged as described in the Dual Paging section.
−215 ≤ LSBsOffset < 215
−16 ≤ SixteenthsOffset  15
1.
Table 72. DC Offset Registers
Address
0x135[0]
0x136
0x137
0x138
0x139
0x13A[4:0]
Value
DC_OFFSET_ON
LSBsOffsetI[7:0]
LSBsOffsetI[15:8]
LSBsOffsetQ[7:0]
LSBsOffsetQ[15:8]
SixteenthsOffsetI
0x13B[4:0]
SixteenthsOffsetQ
Description
Set to 1 to enable dc offset
I DAC LSB dc offset code
I DAC MSB dc offset code
Q DAC LSB dc offset code
Q DAC MSB dc offset code
I DAC sub-LSB dc offset
code
Q DAC sub-LSB dc offset
code
2.
3.
4.
Rev. A | Page 70 of 125
Set NCO_ALIGN_MODE (Register 0x050[1:0] = 0b01) for
SYSREF NCO alignment mode.
Set NCO_ALIGN_ARM (Register 0x050[7] = 1).
Perform an LMFC alignment to force the NCO phase align
(see the Syncing LMFC Signals section). The phase
alignment occurs on the next SYSREF edge.
Note that if in one-shot sync mode, the LMFC alignment
block must be armed by setting Register 0x03A[6] = 1. If in
continuous mode or one-shot then monitor mode, the
LMFC align block does not need to be armed; the NCO
align automatically trips on the next SYSREF± edge.
Check the alignment status. If NCO phase alignment was
successful, NCO_ALIGN_PASS (Register 0x050[4]) = 1.
If phase alignment failed, NCO_ALIGN_FAIL
(Register 0x050[3]) = 1.
Data Sheet
AD9144
Data Key NCO Alignment
In addition to supporting the SYSREF± alignment mode, the
AD9144 supports a mode where the NCO phase alignment
occurs when a user-specified pattern is seen at the DAC input.
The steps to achieve a data key NCO alignment are as follows:
1.
2.
3.
4.
5.
Set NCO_ALIGN_MODE (Register 0x050[1:0]) = 0b10.
Write the expected 16-bit data key for the I and Q datapath
into NCOKEYI (Register 0x051 to Register 0x052) and
NCOKEYQ (Register 0x053 to Register 0x054), respectively.
Set NCO_ALIGN_ARM (Register 0x050[7]) = 1.
Send the expected 16-bit I and Q data keys to the device to
achieve NCO alignment.
Check the alignment status. If the expected data key was
seen at the DAC input, NCO_ALIGN_MTCH
(Register 0x050[5]) = 1. If NCO phase alignment was
successful, NCO_ALIGN_PASS (Register 0x050[4]) = 1.
If phase alignment failed, NCO_ALIGN_FAIL
(Register 0x050[3]) = 1.
Multiple device NCO alignment can be achieved with the data
key alignment mode. To achieve multichip NCO alignment,
program the same expected data key on all devices, arm all
devices, and then send the data key to all devices/channels at
the same time.
NCO Alignment IRQ
An IRQ event showing whether the NCO align was tripped is
available.
Use Register 0x021[4] to enable DAC Dual A (DAC0 and
DAC1), and then use Register 0x025[4] to read back its status
and reset the IRQ signal.
Use Register 0x022[4] to enable DAC Dual B (DAC2 and
DAC3), and then use Register 0x026[4] to read back its status
and reset the IRQ signal.
See the Interrupt Request Operation section for more information.
Rev. A | Page 71 of 125
AD9144
Data Sheet
DOWNSTREAM PROTECTION
FILTER
AND
MODULATION
DATA
PDP
DIGITAL
GAIN
PDP_PROTECT
FROM LMFC
SYNC LOGIC
BSM
DATA TO DACs
1
BSM_PROTECT
0
PDP_PROTECT_OUT
TXENx
TxEnSm
Tx_PROTECT
1
1
0
0
Tx_PROTECT_OUT
SPI_PROTECT
PROTECT_OUTx
PROTECT_OUT_INVERT
1
SPI_PROTECT_OUT
PROTECT OUTx GENERATION
11675-372
0
Figure 74. Downstream Protection Block Diagram
The AD9144 has several blocks designed to protect the power
amplifier (PA) of the system, as well as other downstream
blocks. It consists of a power detection and protection (PDP)
block, a blanking state machine (BSM), and a transmit enable
state machine (TxEnSM).
The PDP block can be used to monitor incoming data. If a
moving average of the data power goes above a threshold, the
PDP block provides a signal (PDP_PROTECT) that can be
routed externally.
The TxEnSM is a simple block that controls delay between
TXENx and the Tx_PROTECT signal. The Tx_PROTECT
signal is used as an input to the BSM, and its inverse can
optionally be routed externally. Optionally, the TxEnSM can
also power down its associated DAC dual.
The BSM gently ramps data entering the DAC and flushes the
datapath. The BSM is activated by the Tx_PROTECT signal or
automatically by the LMFC sync logic during a rotation. For
proper function, digital gain must be enabled; tie TXEN high if
disabling digital gain.
Finally, some simple logic takes the outputs from each of those
blocks and uses them to generate a desired PROTECT_OUTx
signal on an external pin. This signal can be used to
enable/disable downstream components, such as a PA.
Power Detection and Protection
The input signal PDP block is designed to detect the average
power of the DAC input signal and to prevent overrange signals
from being passed to the next stage, which may potentially
cause destructive breakdown on power sensitive devices, such as
PAs. The protection function provides a signal (PDP_PROTECT)
that can be routed externally to shut down a PA.
The PDP block uses a separate path with a shorter latency than
the datapath to ensure that PDP_PROTECT is triggered before
the overrange signal reaches the analog DAC cores. The sum of
the I2 and Q2 are calculated as a representation of the input signal
power (only the top six MSBs of data samples are used). The
calculated sample power numbers are accumulated through a
moving average filter whose output is the average of the input
signal power in a certain number of samples. When the output of
the averaging filter exceeds the threshold, the internal signal
PDP_PROTECT goes high, which can optionally be configured to
trigger a signal on the PROTECT_OUTx. The PDP block is
configured as shown in Table 73 and paged as described in the
Dual Paging section.
The choice of PDP_AVG_TIME (Register 0x062) and
PDP_THRESHOLD (Register 0x060 to Register 0x061) for
effective protection are application dependent. Experiment with
real-world vectors to ensure proper configuration. The PDP_
POWER readback (Register 0x063 to Register 0x064) can help
by storing the maximum power when a set threshold was passed.
Rev. A | Page 72 of 125
Data Sheet
AD9144
Table 74. TxEnSM Registers
Table 73. PDP Registers
Addr.
0x060
0x061
0x062
0x063
0x064
Bit
No.
[7:0]
Value
PDP_THRESHOLD[7:0]
[4:0]
7
[3:0]
PDP_THRESHOLD[12:8]
PDP_ENABLE
PDP_AVG_TIME
[7:0]
[4:0]
PDP_POWER[7:0]
PDP_POWER[12:8]
Description
Power that triggers
PDP_PROTECT. 8 LSBs.
5 MSBs.
Set to 1 to enable PDP.
Can be set from 0 to
10. Averages across
2(9 + PDP_AVG_TIME), IQ
sample pairs.
If PDP_THRESHOLD
is crossed, this reads
back the maximum
power seen. If not,
this reads back the
instantaneous power.
8 LSBs.
5 MSBs.
Addr.
0x11F
Bit No.
[7:6]
Value
FALL_COUNTERS
[5:4]
RISE_COUNTERS
0x121
[7:0]
RISE_COUNT_0
0x122
[7:0]
RISE_COUNT_1
0x123
[7:0]
FALL_COUNT_0
0x124
[7:0]
FALL_COUNT_1
Power Detection and Protection IRQ
The PDP_PROTECT signal is available as an IRQ event.
Use Register 0x021[7] to enable PDP_PROTECT for Dual A
(DAC0 and DAC1), and then use Register 0x025[7] to read back
its status and reset the IRQ signal.
Use Register 0x022[7] to enable PDP_PROTECT for Dual B
(DAC2 and DAC3), and then use Register 0x026[7] to read back
its status and reset the IRQ signal.
See the Interrupt Request Operation section for more information.
Transmit Enable State Machine
The TxEnSM is a simple block that controls the delay between
the TXENx signal and the TX_PROTECT signal. This signal is
used as an input to the BSM and its inverse can be routed to an
external pin (PROTECT_OUTx) to turn downstream components
on or off as desired.
The TXENx signal can be used to power down their associated
DAC duals. If DUALA_MASK (Register 0x012[0]) = 1, a falling
edge of TXENx causes DAC Dual A (DAC0 and DAC1) to
power down. If DUALB_MASK (Register 0x012[1]) = 1, a
falling edge of TXENx causes DAC Dual B (DAC2 and DAC3)
to power down. On a rising edge of TXENx, without
DUALA_MASK and DUALB_MASK enabled, the output is
valid after the BSM settles (see the Blanking State Machine
(BSM) section). If the masks are enabled, an additional delay is
imposed; the output is not valid until the BSM settles and the
DACs fully power on (nominally an additional ~35 μs).
Description
Number of fall counters
to use (1 to 2).
Number of rise
counters to use (0 to 2).
Delay TX_PROTECT rise
from TXEN rising edge
by 32 × RISE_COUNT_0
DAC clock cycles.
Delay TX_PROTECT rise
from TXEN rising edge
by 32 × RISE_COUNT_1
DAC clock cycles.
Delay TX_PROTECT rise
from TXEN rising edge
by 32 × FALL_COUNT_0
DAC clock cycles. Must
be at least 0x12.
Delay TX_PROTECT rise
from TXEN rising edge
by 32 × FALL_COUNT_1
DAC clock cycles.
Blanking State Machine (BSM)
The BSM gently ramps data entering the DAC and flushes the
datapath.
On a falling edge of TX_PROTECT (the TXENx signal delayed
by the TxEnSM), the datapath holds the latest data value and
the digital gain gently ramps from its set value to 0. At the same
time, the datapath is flushed with zeroes.
On a rising edge of TX_PROTECT, the TXENx signal is delayed
by the TxEnSM; data is allowed to flow through the datapath
again, and the digital gain gently ramps the data from 0 up to
the set digital gain.
Both of these functions are also triggered automatically by the
LMFC sync logic during a rotation to prevent glitching on the
output.
Ramping
For proper ramping, digital gain must be enabled; tie TXEN
high if disabling digital gain.
The step size to use when ramping gain to 0 or its assigned value
can be controlled via the GAIN_RAMP_DOWN_STEP registers
(Register 0x142 and Register 0x143) and the GAIN_RAMP_
UP_STEP registers (Register 0x140 and Register 0x141). These
registers are paged as described in the Dual Paging section.
The current BSM state can be read back as shown in Table 75.
The TxEnSM is configured as shown in Table 74 and is paged as
described in the Dual Paging section.
Rev. A | Page 73 of 125
AD9144
Data Sheet
Table 75. Blanking State Machine Ramping Readbacks
Address
0x147[7:6]
Value
0b00
0b01
0b10
0b11
Description
Data is being held at midscale.
Ramping gain to 0. Data ramping to
midscale.
Ramping gain to assigned value. Data
ramping to normal amplitude.
Data at normal amplitude.
Blanking State Machine IRQ
Blanking completion is available as an IRQ event.
Use Register 0x021[5] to enable blanking completion for DAC
Dual A (DAC0 and DAC1), and then use Register 0x025[5] to
read back its status and reset the IRQ signal.
Use Register 0x022[5] to enable blanking completion for DAC
Dual B (DAC2 and DAC3), and then use Register 0x026[5] to
read back its status and reset the IRQ signal.
See the Interrupt Request Operation section for more information.
PROTECT_OUTx Generation
Register 0x013 controls which signals are OR’ed into the external
PROTECT_OUTx signal. Register 0x11F[2] can be used to
invert the PROTECT_OUTx signal, By default, PROTECT_OUTx
is high when the output is valid. Both of these registers are
paged as described in the Dual Paging section.
Table 76. PROTECT_OUTx Registers
Addr.
0x013
0x11F
Bit No.
6
Value
PDP_PROTECT_OUT
5
TX_PROTECT_OUT
3
SPI_PROTECT_OUT
2
2
SPI_PROTECT
PROTECT_OUT_
INVERT
Description
1: PDP block triggers
PROTECT_OUT
1: TxEnSM triggers
PROTECT_OUT
1: SPI_PROTECT
triggers PROTECT_OUT
Sets SPI_PROTECT
Inverts
PROTECT_OUTx
DATAPATH PRBS
The datapath PRBS can be used to verify that the AD9144
datapath is receiving and correctly decoding data. The datapath
PRBS verifies that the JESD204B parameters of the transmitter
and receiver match, that the lanes of the receiver are mapped
appropriately, that the lanes have been appropriately inverted,
if necessary, and in general that the start-up routine has been
implemented correctly.
The datapath PRBS is paged as described in the Dual Paging
section. To run the datapath PRBS test, complete the following
steps:
1.
Set up the device in the desired operating mode. See the
Device Setup Guide section for details on setting up the
device.
2. Send PRBS7 or PRBS15 data.
3. Write Register 0x14B[2] = 0 for PRBS7 or 1 for PRBS15.
4. Write Register 0x14B[1:0] = 0b11 to enable and reset the
PRBS test.
5. Write Register 0x14B[1:0] = 0b01 to enable the PRBS test
and release reset.
6. Wait 500 ms.
7. Check the status by checking the IRQ for DAC0 to DAC3
PRBS as described in the Datapath PRBS IRQ section.
8. If there are failures, set Register 0x008 = 0x01 to view the
status of Dual A (DAC0/DAC1). Set Register 0x08 = 0x02
to view the status of Dual B (DAC2/DAC3).
9. Read Register 0x14B[7:6]. Bit 6 is 0 if the I DAC of the
selected dual has any errors. Bit 7 is 0 if the Q DAC of the
selected dual has any errors. This must match the IRQ.
10. Read Register 0x14C to read the error count for the I DAC
of the selected dual. Read Register 0x14D to read the error
count for the Q DAC of the selected dual.
Note that the PRBS processes 32 bits at a time, and compares
the 32 new bits to the previous set of 32 bits. It detects (and
reports) only 1 error in every group of 32 bits; therefore, the
error count partly depends on when the errors are seen. For
example



Bits: 32 good, 31 good, 1 bad; 32 good [2 errors]
Bits: 32 good, 22 good, 10 bad; 32 good [2 errors]
Bits: 32 good, 31 good, 1 bad; 31 good, 1 bad; 32 good
[3 errors]
Datapath PRBS IRQ
The PRBS fail signals for each DAC are available as IRQ events.
Use Register 0x020[3:0] to enable the fail signals, and then use
Register 0x024[3:0] to read back their statuses and reset the IRQ
signals. See the Interrupt Request Operation section for more
information.
DC TEST MODE
As a convenience, the AD9144 provides a dc test mode, which
is enabled by setting Register 0x520[1] to 1 and clearing
Register 0x146[0] to 0. When this mode is enabled, the datapath
is given 0 (midscale) for its data. Register 0x146[0] must be set
to 1 for all other modes of operation.
In conjunction with dc offset, this test mode can provide desired dc
data to the DACs. This test mode can also provide sinusoidal data
to the DACs by combining digital modulation (to set frequency)
and dc offset (to set amplitude). See the DC Offset section.
Rev. A | Page 74 of 125
Data Sheet
AD9144
INTERRUPT REQUEST OPERATION
The AD9144 provides an interrupt request output signal on
Pin 60 (IRQ) that can be used to notify an external host processor
of significant device events. On assertion of the interrupt, query
the device to determine the precise event that occurred. The
IRQ pin is an open-drain, active low output. Pull the IRQ pin
high external to the device. This pin can be tied to the interrupt
pins of other devices with open-drain outputs to wire; OR these
pins together.
Figure 75 shows a simplified block diagram of how the IRQ
blocks works. If IRQ_EN is low, the INTERRUPT_SOURCE
signal is set to 0. If IRQ_EN is high, any rising edge of EVENT
causes the INTERRUPT_SOURCE signal to be set high. If any
INTERRUPT_SOURCE signal is high, the IRQ pin is pulled
low. INTERRUPT_SOURCE can be reset to 0 by either an
IRQ_RESET signal or a DEVICE_RESET.
Depending on STATUS_MODE, the EVENT_STATUS bit reads
back EVENT or INTERRUPT_SOURCE. The AD9144 has
several IRQ register blocks, which can monitor up to 75 events
(depending on device configuration). Certain details vary by
IRQ register block as described in Table 77. Table 78 shows
which registers the IRQ_EN, IRQ_RESET, and STATUS_MODE
signals in Figure 75 are coming from, as well as the address
where EVENT_STATUS is read back.
Table 77. IRQ Register Block Details
Register Block
0x01F to 0x026
EVENT
Reported
Per chip
0x46D to 0x46F; 0x470
to 0x473; 0x47A
0x47B[4]
Per link and
lane
Per link
EVENT_STATUS
INTERRUPT_SOURCE if
IRQ is enabled; if not, it
is EVENT
INTERRUPT_SOURCE if
IRQ is enabled; if not, 0
INTERRUPT_SOURCE if
IRQ is enabled; if not, 0
INTERRUPT SERVICE ROUTINE
Interrupt request management starts by selecting the set of event
flags that require host intervention or monitoring. Enable the
events that require host action so that the host is notified when
they occur. For events requiring host intervention upon IRQ
activation, run the following routine to clear an interrupt request:
1.
2.
3.
Read the status of the event flag bits that are being monitored.
Disable the interrupt by writing 0 to IRQ_EN.
Read the EVENT source. For Register 0x01F to
Register 0x026, EVENT_STATUS has a live readback.
For other events, see their registers.
Perform any actions required to clear the cause of the
EVENT. In many cases, no specific actions are required.
Verify that the EVENT source is functioning as expected.
Clear the interrupt by writing 1 to IRQ_RESET.
Enable the interrupt by writing 1 to IRQ_EN.
4.
5.
6.
7.
0
1
EVENT_STATUS
STATUS_MODE
IRQ
IRQ_EN
INTERRUPT_SOURCE
0
1
EVENT
IRQ_EN
OTHER
INTERRUPT
SOURCES
IRQ_RESET
11675-060
DEVICE_RESET
Figure 75. Simplified Schematic of IRQ Circuitry
Table 78. IRQ Register Block Address of IRQ Signal Details
Register Block
0x01F to 0x026
0x46D to 0x46F
IRQ_EN
0x01F to 0x022; R/W per chip
0x47A; W per link
0x470 to 0x473
0x47A; W per link
0x47B[4]
0x47B[3]; R/W per link;
1 by default
Address of IRQ Signals
IRQ_RESET
STATUS_MODE
0x023 to 0x026; W per chip
STATUS_MODE = IRQ_EN
0x46D to 0x46F; W per link
Not applicable,
and lane
STATUS_MODE = 1
0x470 to 0x473; W per link
Not applicable,
STATUS_MODE = 1
0x47B[4]; W per link
Not applicable,
STATUS_MODE = 1
Rev. A | Page 75 of 125
EVENT_STATUS
0x023 to 0x26; R per chip
0x47A; R per link
0x47A; R per link
0x47B[4]; R per link
AD9144
Data Sheet
DAC INPUT CLOCK CONFIGURATIONS
The AD9144 DAC sample clock (DACCLK) can be sourced
directly through CLK± (Pin 2 and Pin 3) or by clock
multiplication through the CLK± differential input. Clock
multiplication employs the on-chip PLL that accepts a reference
clock operating at a submultiple of the desired DACCLK rate.
The PLL then multiplies the reference clock up to the desired
DACCLK frequency, which is used to generate all the internal
clocks required by the DAC. The clock multiplier provides a
high quality clock that meets the performance requirements of
most applications. Using the on-chip clock multiplier removes
the burden of generating and distributing the high speed
DACCLK.
The second mode bypasses the clock multiplier circuitry and
allows DACCLK to be sourced directly to the DAC core. This
mode enables the user to source a very high quality clock directly
to the DAC core.
DRIVING THE CLK± INPUTS
The CLK± differential input circuitry is shown in Figure 76 as a
simplified circuit diagram of the input. The on-chip clock receiver
has a differential input impedance of 10 kΩ. It is self biased to a
common-mode voltage of approximately 600 mV. The inputs
can be driven by differential PECL or LVDS drivers with ac
coupling between the clock source and the receiver.
To optimize the PLL across all operating conditions, the register
writes in Table 79 are recommended. These writes properly set
up the DAC PLL, including the loop filter and the charge pump.
Table 79. DAC PLL Fixed Register Writes
Register
Address
0x087
0x088
0x089
0x08A
0x08D
0x1B0
Register
Value
0x62
0xC9
0x0E
0x12
0x7B
0x00
0x1B9
0x1BC
0x1BE
0x24
0x0D
0x02
0x1BF
0x1C0
0x8E
0x2A
0x1C1
0x1C4
0x2A
0x7E
Description
Optimal DAC PLL loop filter settings
Optimal DAC PLL loop filter settings
Optimal DAC PLL loop filter settings
Optimal DAC PLL charge pump settings
Optimal DAC LDO settings for DAC PLL
Power DAC PLL blocks when power
machine disabled
Optimal DAC PLL charge pump settings
Optimal DAC PLL VCO control settings
Optimal DAC PLL VCO power control
settings
Optimal DAC PLL VCO calibration settings
Optimal DAC PLL lock counter length
setting
Optimal DAC PLL charge pump setting
Optimal DAC PLL varactor settings
CLOCK MULTIPLICATION
CLK+
The on-chip PLL clock multiplier circuit can be used to generate
the DAC sample rate clock from a lower frequency reference clock.
The PLL is integrated on-chip, including the VCO and the loop
filter. The VCO operates over the frequency range of 6 GHz to
12 GHz.
5kΩ
600mV
11675-061
5kΩ
CLK–
DAC PLL FIXED REGISTER WRITES
Figure 76. Clock Receiver Input Simplified Equivalent Circuit
The minimum input drive level to the differential clock input is
400 mV p-p differential. The optimal performance is achieved
when the clock input signal is between 800 mV p-p differential
and 1000 mV p-p differential. Whether using the on-chip clock
multiplier or sourcing the DACCLK directly (the CLK± pins are
used in both cases), it is necessary that the input clock signal to
the device has low jitter and fast edge rates to optimize the DAC
noise performance. Direct clocking with a low noise clock
produces the lowest noise spectral density at the DAC outputs.
The clocks and clock receiver are powered down by default. The
clocks must be enabled by writing to Register 0x080. To enable
all clocks on the device, write Register 0x080 = 0x00.
Register 0x080, Bit 7 powers up the clocks for DAC0 and DAC1.
Bit 6 powers up the clocks for DAC2 and DAC3, Bit 5 powers
up the digital clocks, Bit 4 powers up the SERDES clocks, and
Bit 3 powers up the clock receiver.
The PLL configuration parameters must be programmed before
the PLL is enabled. Step by step instructions on how to program the
PLL can be found in the Starting the PLL section. The functional
block diagram of the clock multiplier is shown in Figure 79.
The clock multiplication circuit generates the DAC sampling
clock from the REFCLK input, which is fed in on the CLK±
differential pins (Pin 2 and Pin 3). The frequency of the
REFCLK input is referred to as fREF.
The REFCLK input is divided by the variable RefDivFactor. Select
the RefDivFactor variable to ensure that the frequency into the
phase frequency detector (PFD) block is between 35 MHz and
80 MHz. The valid values for RefDivFactor are 1, 2, 4, 8, 16, or 32.
Each RefDivFactor maps to the appropriate REF_DIV_MODE
register control according to Table 80. The REF_DIV_MODE
register is programmed through Register 0x08C[2:0].
Rev. A | Page 76 of 125
Data Sheet
AD9144
Table 80. Mapping of RefDivFactor to REF_DIV_MODE
Divide by
(RefDivFactor)
1
2
4
8
16
REF_DIV_MODE,
Reg. 0x08C[2:0]
0
1
2
3
4
The range of fREF is 35 MHz to 1 GHz, and the output frequency
of the PLL is 420 MHz to 2.8 GHz. Use the following equations to
determine the RefDivFactor:
35 MHz 
f REF
 80 MHz
RefDivFactor
(1)
where:
RefDivFactor is the reference divider division ratio.
fREF is the reference frequency on the CLK± input pins.
Table 82. Common Frequency Examples
fDACCLK
(MHz)
1474.56
1474.56
1228.88
983.04
983.04
1966.08
1966.08
Frequency
(MHz)
368.64
184.32
307.2
122.88
61.44
491.52
245.76
fVCO
(MHz)
11796.48
11796.48
9831.04
7864.35
7864.35
7864.35
7864.35
RefDivFactor
8
4
8
2
1
8
4
f DACCLK
f REF

RefDivFactor
2  BCount
(2)
The RF PLL filter is fully integrated on-chip and is a standard
passive third-order filter with five 4-bit programmable
components (see Figure 77). The C1, C2, C3, R1, and R3 filter
components are programmed with Register 0x087 through
Register 0x089, as described in the DAC PLL Fixed Register
Writes section.
R3
FROM CHARGE PUMP
where:
BCount is the feedback loop divider ratio.
fDACCLK is the DAC sample clock.
TO VCO
R1
C2
C3
C1
The BCount value is programmed with Bits[7:0] of
Register 0x085. It is programmable from 6 to 127.
TO VCO LDO
Figure 77. Loop Filter
The PFD compares fREF/RefDivRate to fDAC/(2 × BCount) and
pulses the charge pump up or down to control the frequency of
the VCO. A low noise VCO is tunable over an octave with an
oscillation range of 6 GHz to 12 GHz.
The clock multiplication circuit operates such that the VCO
outputs a frequency, fVCO.
(3)
Charge Pump
The charge pump current is 6-bit programmable and varies from
0.1 mA to 6.4 mA in 0.1 mA steps. The charge pump current is
programmed into Register 0x08A for the DAC PLL, as shown in
the DAC PLL Fixed Register Writes section. The charge pump
calibration must be run one time during chip initialization to
reduce reference spurs. This calibration is on by default.
And from Equation 2, the DAC sample clock frequency, fDACCLK,
is equal to
f REF
RefDivFactor
(4)
The LODivFactor is chosen to keep fVCO in the operating range
between 6 GHz and 12 GHz. The valid values for LODivFactor
are 4, 8, and 16. Each LODivFactor maps to a LO_DIV_MODE
value. The LO_DIV_MODE (Register 0x08B[1:0]) value is
programmed as described in Table 81.
Divide by
(LODivFactor)
4
8
16
TO LOOP FILTER
DOWN
CHARGE PUMP CURRENT = 0.1mA TO 6.4mA
Table 81. DAC VCO Divider Selection
DAC Frequency
Range (MHz)
>1500
750 to 1500
420 to 750
UP
LO_DIV_MODE,
Register 0x08B[1:0]
1
2
3
Rev. A | Page 77 of 125
Figure 78. Charge Pump
11675-063
f DACCLK  2  BCount 
BCount
16
16
16
8
8
16
16
Loop Filter
The BCount value is the divide ratio of the loop divider. It is set
to divide the fDACCLK to frequency match the fREF/RefDivFactor.
Select BCount so that the following equation is true:
fVCO  f DACCLK  LODivFactor
LODivFactor
8
8
8
8
8
4
4
11675-062
DAC Reference
Frequency Range (MHz)
35 to 80
80 to 160
160 to 320
320 to 640
640 to 1000
Table 82 lists some common frequency examples for the
RefDivFactor, LODivFactor, and BCount values that are needed
to configure the PLL properly.
AD9144
Data Sheet
Charge pump calibration is run during the first power-up of
the PLL, and the coefficient of the calibration is held for all
subsequent starts. The PLL is enabled by writing 0x10 into
Register 0x083; however, the configuration registers must be
programmed before the PLL is enabled. The calibration tries to
match the up and down current, which minimizes the spurs at
the reference frequency that appears at the DAC output. The
charge pump calibration takes 64 reference clock cycles. Bit 5
in Register 0x084 notifies the user that the charge pump
calibration is completed and is valid.
STARTING THE PLL
The programming sequence for the DAC PLL is as follows:
1.
2.
3.
4.
Temperature Tracking
When properly configured, the device automatically selects one
of the 512 VCO bands. The PLL settings selected by the device
ensure that the PLL remains locked over the full −40°C to +85°C
operating temperature range of the device without further
adjustment. The PLL remains locked over the full temperature
range even if the temperature during initialization is at one of
the temperature extremes. Check the PLL lock bit to make sure
that the calibration completed properly. The PLL lock bit is Bit 1
of Register 0x084.
5.
6.
7.
Register 0x084[5] notifies the user that the DAC PLL calibration is
completed and is valid.
To properly configure temperature tracking, follow the settings
in the DAC PLL Fixed Register Writes section and the fVCO
dependent SPI writes shown in Table 83.
Register 0x084[1] notifies the user that the PLL has locked.
Register 0x084[7] and Register 0x084[6] notify the user that the
DAC PLL hit the upper or lower edge of its operating band,
respectively. If either of these bits are high, recalibrate the DAC
PLL by setting Register 0x083[7] to 0 and then 1.
Table 83. VCO Control Lookup Table Reference
Register
0x1C5
Setting
0x07
0x06
0x06
CHARGE
PUMP
fREF
÷1
35MHz
TO 1GHz
÷2
÷4
÷8
÷16
The DAC PLL lock and lost signals are available as IRQ events.
Use Register 0x01F[5:4] to enable these signals, and then use
Register 0x023[5:4] to read back their statuses and reset the IRQ
signals. See the Interrupt Request Operation section for more
information.
4-BIT
PROGRAMMABLE,
INTEGRATED
LOOP FILTER
VCO
LDO
PFD
80MHz
MAX
C1
RETIMER
UP
C2
LC VCO
6GHz
TO
12GHz
C3
R1
DOWN
÷2
R3
÷2
÷2
I
Q
÷2
I
Q
I
Q
ALC CAL
FO CAL
CAL CONTROL BITS
0.1mA TO 6.4mA
MUX/SELECTABLE BUFFERS
LODivFactor =
4, 8, 16
÷2
B COUNTER
BCount (INTEGER FEEDBACK DIVIDER)
RANGE = 6 TO 127
Figure 79. Device Clock PLL Block Diagram
Rev. A | Page 78 of 125
DAC CLOCK
420MHz TO 2.8GHz
11675-064
RefDivFactor
DAC PLL IRQ
<750MHz
Register
0x1BB
Setting
0x03
0x03
0x13
750MHz TO 1.5GHz
Register
0x1B5
Setting
0x08
0x09
0x09
>1.5GHz
VCO Frequency
Range (GHz)
fVCO < 6.3
6.3 ≤ fVCO < 7.25
fVCO ≥ 7.25
Program the registers in the DAC PLL Fixed Register
Writes section.
Determine the VCO frequency based on the DAC
frequency requirements.
Determine the VCO divider ratio to achieve the desired
DAC frequency. Program the VCO divider ratio in
Register 0x08B[1:0].
Determine the BCount ratio to achieve the desired PLL
reference frequency (35 MHz to 80 MHz). Program the
BCount ratio in Register 0x085[7:0].
Determine the reference divider ratio to achieve the
desired PLL reference frequency. Program the reference
divider ratio in Register 0x08C[2:0].
Based on the fVCO found in Step 2, write the temperature
tracking registers as shown in Table 83.
Enable the DAC PLL synthesizer by setting Register 0x083[4]
to 1.
Data Sheet
AD9144
ANALOG OUTPUTS
TRANSMIT DAC OPERATION
Table 84. DAC Full-Scale Current Registers
Figure 80 shows a simplified block diagram of the transmit path
DACs. The DAC core consists of a current source array, a switch
core, digital control logic, and full-scale output current control.
The DAC full-scale output current (IOUTFS) is nominally 20.48 mA.
The output currents from the OUTx± pins are complementary,
meaning that the sum of the two currents always equals the fullscale current of the DAC. The digital input code to the DAC
determines the effective differential current delivered to the load.
Address
0x040[1:0]
0x041[7:0]
0x042[1:0]
0x043[7:0]
0x044[1:0]
0x045[7:0]
0x046[1:0]
0x047[7:0]
1.2V
Q DACs
FULL-SCALE
ADJUST
DAC3
Value
DACFSC_0[9:8]
DACFSC_0[7:0]
DACFSC_1[9:8]
DACFSC_1[7:0]
DACFSC_2[9:8]
DACFSC_2[7:0]
DACFSC_3[9:8]
DACFSC_3[7:0]
Description
Dual A I DAC MSB gain code
Dual A I DAC LSB gain code
Dual A Q DAC MSB gain code
Dual A Q DAC LSB gain code
Dual B I DAC MSB gain code
Dual B I DAC LSB gain code
Dual B Q DAC MSB gain code
Dual B Q DAC LSB gain code
OUT3+
28
OUT3–
26
OUT2+
4kΩ
I DACs
FULL-SCALE
ADJUST
IOUTFS (mA)
DAC1
OUT1+
OUT1–
OUT0+
DAC0
OUT0–
11675-065
I120
24
OUT2–
CURRENT
SCALING
22
20
18
16
14
Figure 80. Simplified Block Diagram of DAC Core
The DAC has a 1.2 V band gap reference. A 4 kΩ external
resistor, RSET, must be connected from the I120 pin to the
ground plane. This resistor, along with the reference control
amplifier, sets up the correct internal bias currents for the DAC.
Because the full-scale current is inversely proportional to this
resistor, the tolerance of RSET is reflected in the full-scale output
amplitude.
DACFSC_x (where x is a number from 0 to 3 that corresponds
to DAC0 through DAC3) is a 10-bit twos complement value
that controls the full-scale current of each of the four DAC outputs.
These values are stored in Register 0x040 to Register 0x047, as
shown in Table 84.
The typical full-scale current for each DAC is given by:
12
–512
–256
–128
0
128
256
384
512
GAIN DAC CODE
Figure 81. DAC Full-Scale Current (IOUTFS) vs. DAC Gain Code
Transmit DAC Transfer Function
The output currents from the OUTx+ and OUTx− pins are
complementary, meaning that the sum of the positive and negative
currents always equals the full-scale current of the DAC. The
digital input code to the DAC determines the effective differential
current delivered to the load. OUTx± provides the maximum
output current when all bits are high for binary data. The
output currents vs. DACCODE for the DAC outputs using
binary format are expressed as
IOUTP 
IOUTFS = 20.45 + (DACFSC_x × 6.55 mA)/2(10 − 1)
For nominal values of VREF (1.2 V), RSET (4 kΩ), and DACFSC_x
(0, which is midscale in twos complement), the full-scale current
of the DAC is typically 20.48 mA. The DAC full-scale current
can be adjusted from 13.9 mA to 27.0 mA, by programming the
appropriate DACFSC_x values in Register 0x040 to Register 0x047.
Analog output full-scale current vs. DAC gain code is plotted in
Figure 81.
–384
11675-066
DAC2
DACCODEBIN
 IOUTFS
2N  1
IOUTN  IOUTFS  IOUTP
(5)
(6)
where DACCODEBIN is the 16-bit input to the DAC in unsigned
binary. DACCODEBIN has a range of 0 to 2N − 1.
If the data format is twos complement, the output currents are
expressed as
IOUTP 
DACCODETWOS  2N 1
 IOUTFS
2N  1
I OUTN  I OUTFS  I OUTP
(7)
(8)
where DACCODETWOS is the 16-bit input to the DAC in twos
complement. DACCODETWOS has a range of −2N − 1 to 2N − 1 − 1.
Rev. A | Page 79 of 125
AD9144
Data Sheet
4
Powering Down Unused DACs
Register 0x011, Bit 7 and Bit 2, must stay low to enable the band
gap and DAC master bias, respectively.
CALIBRATION OFF
CALIBRATION ON
3
2
DNL (LSB)
Power down any unused DAC outputs to avoid burning excess
power. The DAC power downs are located in Register 0x011.
Register 0x011, Bit 6 corresponds to DAC0, Bit 5 corresponds to
DAC1, Bit 4 corresponds to DAC2, and Bit 3 corresponds to DAC3.
Write a 1 to each bit to power down the appropriate DACs.
1
0
For more information on which DACs to power down, see the
DAC Power-Down Setup section.
–1
0
30k
40k
50k
60k
70k
Figure 83. Pre-Calibration and Post-Calibration, DNL
–40
CALIBRATION OFF
CALIBRATION ON
SECOND HARMONIC
FOURTH HARMONIC
–50
–60
–70
–80
–90
–100
0
50
100
150
200
250
300
fOUT (MHz)
Figure 84. Pre-Calibration and Post-Calibration, HD2 and HD4
2
–60
CALIBRATION OFF
CALIBRATION ON
1
–65
0
–70
IMD2 (dBc)
–1
–2
–3
0
10k
20k
30k
40k
50k
60k
DAC GAIN CODE
70k
–75
–80
–85
–90
–95
Figure 82. Pre-Calibration and Post-Calibration, INL
–100
0
50
100
150
200
250
fOUT (MHz)
Figure 85. Pre-Calibration and Post-Calibration, IMD2
Rev. A | Page 80 of 125
300
11675-096
–4
11675-088
INL (LSB)
20k
DAC GAIN CODE
CALIBRATION OFF
CALIBRATION ON
3
10k
11675-095
4
–2
SFDR (dBc)
The AD9144 has a self calibration feature that improves the
DAC dc and ac linearity in zero or low IF applications. The
performance improvement includes the INL/DNL, second and
fourth harmonic distortions (HD2 and HD4), and second-order
intermodulation distortion (IMD2) of the device. Figure 82 and
Figure 83 show the typical DAC INL and DNL before and after
the calibration. Figure 84 and Figure 85 show the calibration effect
on the HD2, HD4, and IMD2 performance. The improvement
from calibration decreases with the DAC output frequency. For
improvement in HD2 and HD4, it is recommended to run the
calibration routine when the desired output frequency is below
100 MHz. For improvement in IMD2, it is recommended to run
the routine when the desired output frequency is below 200 MHz.
A single run of the routine is sufficient to obtain the desired
performance for both ac and dc performance.
11675-089
Self Calibration
Data Sheet
AD9144
Table 85. Device Self Calibration Procedure for 4-Converter
Setup
Addr.
0x0E7
0x0E8
0x0ED
0x0E2
0x0E2
Read
0x023[7:6]
0x0E7
SPI Data
Byte
0x38
0x0F
0xA2
0x01
0x03
0b10
0x30
Description
Enable calibration clock.
Calibrate all DACs.
Configure initial value.
Enable averaged calibration.
Start averaged calibration.
CAL_PASS (Register 0x023[7]) = 1 to
indicate that the calibration passed.
If CAL_PASS = 0, check CAL_FAIL
(Register 0x023[6]). If both CAL_PASS =
0 and CAL_FAIL = 0, calibration is
either still running or it never ran.
Try waiting ~100 ms and reread
CAL_PASS and CAL_FAIL, or rerun the
calibration routine.
Disable calibration clock.
The post-calibration result is a function of operating temperature.
A set of calibration coefficients obtained at one temperature
may not be the optimal setting for a different temperature.
Figure 86 and Figure 87 show the typical temperature drift effect
after a single run calibration.
For optimal performance, run the calibration again when the
operating temperature changes significantly. Note that it is
recommended to power down the DAC outputs when running
the calibration routine. If continuous transmission is required
in the system, running the calibration again during the operation
may not be an option. In this case, it is recommended to perform a
calibration at the average temperature of the operating temperature
range and to use the same set of coefficients during the operation.
This results in the best overall performance over temperature.
–40
–40°C
+25°C
+85°C
–50
SECOND HARMONIC
FOURTH HARMONIC
–60
SFDR (dBc)
When using all four DACs, follow the procedure in Table 85
to perform a device self calibration. However, when using
fewer than four DACs, follow the calibration routine shown in
Table 86.
–70
–80
If using fewer than four converters, use the calibration routine
in Table 86. See DAC Power-Down Setup for notes on which
DACs to power down when using fewer than four converters.
Bit
0x0E8
3
2
1
0
0x0ED
0x0E9
0x0E9
0x0E7
0b0 or 0b1
0b0 or 0b1
0b0 or 0b1
0b0 or 0b1
0xA2
0x01
0x03
0x30
Description
Use highest comparator speed
and set calibration clock divider
Select DACs to calibrate
1 if DAC3 is enabled
1 if DAC2 is enabled
1 if DAC1 is enabled
1 if DAC0 is enabled
Configure initial value
Enable calibration
Start calibration
Disable calibration clock
50
100
150
250
300
Figure 86. Post-Calibration HD2 and HD4 over Temperature, Calibrated at 25°C
–60
–40°C
+25°C
+85°C
–65
–70
–75
–80
–85
–90
–95
–100
0
50
100
150
fOUT (MHz)
For each DAC calibrated, check the calibration status by writing
a 1 in the corresponding bit of CAL_PAGE (Register 0x0E8)
and reading Register 0x0E9. If the calibration completed
correctly, CAL_FIN (Register 0x0E9[7]) = 1 to indicate that
calibration is complete, and Register 0x0E9[6:4] = 0 to indicate
that no errors occurred.
200
fOUT (MHz)
11675-486
0
200
250
300
11675-487
Addr.
0x0E7
SPI Data
Byte
0x38
–100
IMD2 (dBc)
Table 86. Device Self Calibration Procedure with Fewer than
Four Converters Enabled
–90
Figure 87. Post-Calibration IMD2 over Temperature, Calibrated at 25°C
Self Calibration IRQ
Self calibration pass and fail signals are available as IRQ events.
Use Register 0x01F[7:6] to enable these signals, and then use
Register 0x023[7:6] to read back their statuses and reset the IRQ
signals. See the Interrupt Request Operation section for more
information.
Rev. A | Page 81 of 125
AD9144
Data Sheet
DEVICE POWER DISSIPATION
The AD9144 has eight supply rails, AVDD33, DVDD12, SVDD12,
SIOVDD33, CVDD12, IOVDD, VTT, and PVDD12, which can
be driven from five regulators to achieve optimum performance, as
shown in Figure 63.
The AVDD33 supply powers the DAC core circuitry. The power
dissipation of the AVDD33 supply rail is independent of the
digital operating mode and sample rate. The current drawn
from the AVDD33 supply rail is typically 126 mA (416 mW)
when the full-scale current of DAC0 to DAC3 are set to the
nominal value of 20.48 mA.
PVDD12 powers the DAC PLLs and varies depending on the
DAC sample rate. CVDD12 can be combined with the PVDD12
regulator but requires proper bypass capacitor networks near
the pins. CVDD12 powers the clock tree, and the current varies
directly with the DAC sample rate. DVDD12 powers the DSP
core, and the current draw depends on the number of DSP
functions and the DAC sample rate used. SVDD12 supplies the
SERDES lanes and associated circuitry including the equalizers,
SERDES PLL, PHY, and up to the input of the DSP. The current
depends on the number lanes and the lane bit rate. IOVDD
powers the SPI circuit and draws very small current.
TEMPERATURE SENSOR
The AD9144 has a band gap temperature sensor for monitoring
the temperature changes of the AD9144. The temperature must
be calibrated against a known temperature to remove the
device-to-device variation on the band gap circuit used to sense
the temperature.
To monitor temperature change, the user must take a reading at
a known ambient temperature for a single-point calibration of
each AD9144 device.
Tx = TREF + 7.3 × (CODE_X − CODE_REF)/1000
where:
CODE_X is the readback code at the unknown temperature, Tx.
CODE_REF is the readback code at the calibrated temperature,
TREF.
To use the temperature sensor, it must be enabled by setting
Register 0x12F[0] to 1. The user must write a 1 to Register 0x134[0]
before reading back the die temperature from Register 0x132
and Register 0x133.
SIOVDD33 powers the equalizers for the SERDES lanes. The
VTT termination voltage draws a very small current of <5 mA.
Rev. A | Page 82 of 125
Data Sheet
AD9144
START-UP SEQUENCE
Table 87 through Table 96 show the register writes needed to set
up the AD9144 with fDAC = 1474.56 MHz, 2× interpolation, and
the DAC PLL enabled with a 368.64 MHz reference clock. The
JESD204B interface is configured in Mode 4, dual-link mode,
Subclass 1, and scrambling is enabled with all eight SERDES
lanes running at 7.3728 Gbps, inputting twos complement
formatted data. No remapping of lanes with the crossbar is
done in this example.
Step 1A: Configure the DAC PLL
Table 89. Configure DAC PLL
Command
W
Address
0x087
Value
0x62
W
0x088
0xC9
W
0x089
0x0E
The sequence of steps to properly start up the AD9144 are as
follows:
W
0x08A
0x12
1.
W
0x08D
0x7B
W
0x1B0
0x00
W
0x1B9
0x24
W
0x1BC
0x0D
W
0x1BE
0x02
W
0x1BF
0x8E
W
0x1C0
0x2A
W
0x1C1
0x2A
W
0x1C4
0x7E
W
0x08B
0x02
W
0x08C
0x03
W
0x085
0x10
W
0x1B5
0x09
W
0x1BB
0x13
W
0x1C5
0x06
W
R
0x083
0x084
0x10
0x01
2.
3.
4.
5.
6.
Set up the SPI interface, power up necessary circuit blocks,
make required writes to the configuration register, and set up
the DAC clocks (see the Step 1: Start Up the DAC section).
Set the digital features of the AD9144 (see the Step 2:
Digital Datapath section).
Set up the JESD204B links (see the Step 3: Transport Layer
section).
Set up the physical layer of the SERDES interface (see the
Step 4: Physical Layer section).
Set up the data link layer of the SERDES interface. This
procedure is for quick startup or debug only and does not
guarantee deterministic latency (see the Step 5: Data Link
Layer section).
Check for errors on Link 0 and Link 1 (see the Step 6:
Error Monitoring section).
These steps are outlined in detail in the following sections in
tables that list the required register write and read commands.
STEP 1: START UP THE DAC
Power-Up and DAC Initialization
Table 87. Power-Up and DAC Initialization
Command
W
W
W
Address
0x000
0x000
0x011
Value
0xBD
0x3C
0x00
W
W
0x080
0x081
0x00
0x00
Description
Soft reset
Deassert reset, set 4-wire SPI
Enable reference, DAC
channels, and master DAC
Power up all clocks
Power up SYSREF receiver,
disable hysteresis
Required Device Configurations
Table 88. Required Device Configuration
Command
W
Address
0x12D
Value
0x8B
W
0x146
0x01
W
W
0x2A4
0x232
0xFF
0xFF
W
0x333
0x01
Description
Digital datapath
configuration
Digital datapath
configuration
Clock configuration
SERDES interface
configuration
SERDES interface
configuration
Description
Optimal DAC PLL loop filter
settings
Optimal DAC PLL loop filter
settings
Optimal DAC PLL loop filter
settings
Optimal DAC PLL charge
pump settings
Optimal DAC LDO settings
for DAC PLL
Power DAC PLL blocks when
power machine is disabled
Optimal DAC PLL charge
pump settings
Optimal DAC PLL VCO
control settings
Optimal DAC PLL VCO
power control settings
Optimal DAC PLL VCO
calibration settings
Optimal DAC PLL lock
counter length setting
Optimal DAC PLL charge
pump setting
Optimal DAC PLL varactor
settings
Set the VCO LO divider to 8
so that 6 GHz ≤ fVCO = fDACCLK ×
2(LODivMode + 1) ≤ 12 GHz
Set the reference clock divider
to 8 so that the reference
clock into the PLL is less
than 80 MHz
Set the B counter to 16 to
divide the DAC clock down
to 2× the reference clock
PLL lookup value from
Table 25 for fVCO  7.25GHz
PLL lookup value from
Table 25 for fVCO  7.25GHz
PLL lookup value from
Table 25 for fVCO  7.25GHz
Enable DAC PLL
Verify that Bit 1 reads back
high for PLL locked
STEP 2: DIGITAL DATAPATH
Table 90. Digital Datapath
Command
W
W
Rev. A | Page 83 of 125
Address
0x112
0x110
Value
0x01
0x00
Description
Set the interpolation to 2×
Set twos complement data
format
AD9144
Data Sheet
STEP 3: TRANSPORT LAYER
STEP 4: PHYSICAL LAYER
Table 91. Link 0 Transport Layer
Command
W
W
W
Address
0x200
0x201
0x300
Value
0x00
0x00
0x08
W
0x450
0x00
W
0x451
0x00
W
0x452
0x00
W
0x453
0x83
W
W
W
W
W
0x454
0x455
0x456
0x457
0x458
0x00
0x1F
0x01
0x0F
0x2F
W
0x459
0x20
W
W
W
W
W
0x45A
0x45D
0x46C
0x476
0x47D
0x80
0x45
0x0F
0x01
0x0F
Table 93. Physical Layer
Description
Power up the interface
Enable all lanes
Bit 3 = 1 for dual-link, Bit 2 =
0 to access Link 0 registers
Set the device ID to match
Tx (0x00 in this example)
Set the bank ID to match Tx
(0x00 in this example)
Set the lane ID to match Tx
(0x00 in this example)
Set descrambling and L = 4
(in n − 1 notation)
Set F = 1 (in n − 1 notation)
Set K = 32 (in n − 1 notation)
Set M = 2 (in n − 1 notation)
Set N = 16 (in n − 1 notation)
Set Subclass 1 and NP = 16
(in n − 1 notation)
Set JESD204B Version and
S = 1 (in n − 1 notation)
Set HD = 1
Set checksum for Lane 0
Deskew Lane 0 to Lane 3
Set F (not in n − 1 notation)
Enable Lane 0 to Lane 3
Command
W
Address
0x2AA
Value
0xB7
W
0x2AB
0x87
W
0x2B1
0xB7
W
0x2B2
0x87
W
W
W
W
0x2A7
0x2AE
0x314
0x230
0x01
0x01
0x01
0x28
W
W
W
0x206
0x206
0x289
0x00
0x01
0x04
W
0x284
0x62
W
0x285
0xC9
W
0x286
0x0E
W
0x287
0x12
W
0x28A
0x7B
W
0x28B
0x00
W
0x290
0x89
W
0x294
0x24
W
W
W
0x296
0x297
0x299
0x03
0x0D
0x02
W
0x29A
0x8E
W
0x29C
0x2A
W
0x29F
0x78
W
0x2A0
0x06
W
R
0x280
0x281
0x01
0x01
W
0x268
0x62
Table 92. Link 1 Transport Layer
Command
W
Address
0x300
Value
0x0C
W
0x450
0x00
W
0x451
0x00
W
0x452
0x04
W
0x453
0x83
W
W
W
W
W
0x454
0x455
0x456
0x457
0x458
0x00
0x1F
0x01
0x0F
0x2F
W
0x459
0x20
W
W
W
W
W
0x45A
0x45D
0x46C
0x476
0x47D
0x80
0x45
0x0F
0x01
0x0F
Description
Bit 3 = 1 for dual-link, Bit 2 = 1
to access registers for Link 1
Set the device ID to match
Tx (0x00 in this example)
Set the bank ID to match Tx
(0x00 in this example)
Set the lane ID to match Tx
(0x04 in this example)
Set descrambling and L = 4
(in n − 1 notation)
Set F = 1 (in n − 1 notation)
Set K = 32 (in n − 1 notation)
Set M = 2 (in n − 1 notation)
Set N = 16 (in n − 1 notation)
Set Subclass 1 and NP = 16
(in n − 1 notation)
Set JESD204B and S = 1
(in n − 1 notation)
Set HD
Set checksum for Lane 0
Deskew Lane 4 to Lane 7
Set F (not in n − 1 notation)
Enable Lane 4 to Lane 7
Rev. A | Page 84 of 125
Description
SERDES interface
termination setting
SERDES interface
termination setting
SERDES interface
termination setting
SERDES interface
termination setting
Autotune PHY setting
Autotune PHY setting
SERDES SPI configuration
Configure CDRs in half rate
mode
Resets CDR logic
Release CDR logic reset
Configure PLL divider to 1
along with PLL required
configuration
Optimal SERDES PLL loop
filter
Optimal SERDES PLL loop
filter
Optimal SERDES PLL loop
filter
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO
LDO
Optimal SERDES PLL
configuration
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO
Optimal SERDES PLL VCO
Optimal SERDES PLL
configuration
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL charge
pump
Optimal SERDES PLL VCO
varactor
Optimal SERDES PLL VCO
varactor
Enable SERDES PLL
Verify that Bit 0 reads back
high for SERDES PLL lock
Set EQ mode to low power
Data Sheet
AD9144
STEP 5: DATA LINK LAYER
STEP 6: ERROR MONITORING
Note that this procedure does not guarantee deterministic latency.
Link 0 Checks
Confirm that the registers in Table 95 read back as noted and
that system tasks are completed as described.
Table 94. Data Link Layer—Does Not Guarantee
Deterministic Latency
Command
W
W
Address
0x301
0x304
Value
0x01
0x00
W
0x305
0x00
W
0x306
0x0A
W
0x307
0x0A
W
0x03A
0x01
W
W
SYSREF±
Signal
0x03A
0x03A
0x81
0xC1
W
0x300
0x0B
Description
Set subclass to 1
Set the LMFC delay setting
to 0
Set the LMFC delay setting
to 0
Set the LMFC receive buffer
delay to 10
Set the LMFC receive buffer
delay to 10
Set sync mode to one-shot
sync
Enable the sync machine
Arm the sync machine
Ensure that at least one
SYSREF± edge is sent to the
device
Bit 1 and Bit 0 = 1 to enable
Link 0 and Link 1, Bit 2 = 0
to access Link 0
Table 95. Link 0 Checks
Command
R
Address
0x470
Value
0x0F
SYNCOUT0±
Signal
SERDINx±
Signals
R
0x471
0x0F
R
R
0x472
0x473
0x0F
0x0F
Description
Acknowledge that four
consecutive K28.5 characters
have been detected on Lane 0
to Lane 3.
Confirm that SYNCOUT0± is
high.
Apply ILAS and data to
SERDES input pins.
Check for frame sync on all
lanes.
Check for good checksum.
Check for ILAS.
Link 1 Checks
Confirm that the registers in Table 96 read back as noted and
that system tasks are completed as described.
Table 96. Link 1 Checks
Command
W
R
Address
0x300
0x470
Value
0x0F
0x0F
SYNCOUT0±
Signal
SERDINx±
Signals
R
0x471
0x0F
R
R
0x472
0x473
0x0F
0x0F
Rev. A | Page 85 of 125
Description
Bit 2 = 1 to access Link 1.
Acknowledge that four
consecutive K28.5 characters
have been detected on Lane 4
to Lane 7.
Confirm that SYNCOUT0± is
high.
Apply ILAS and data to
SERDES input pins.
Check for frame sync on all
lanes.
Check for good checksum.
Check for ILAS.
AD9144
Data Sheet
REGISTER MAPS AND DESCRIPTIONS
In the following tables, register addresses (Reg. column) and reset (Reset column) values are hexadecimal, and in the read/write (R/W)
column, R means read only, W means write only, R/W means read/write, and N/A means not applicable. All values in the register address
and reset columns are hexadecimal numbers.
DEVICE CONFIGURATION REGISTER MAP
Table 97. Device Configuration Register Map
Reg.
Name
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
Bit 2
Bit 1
Bit 0
Reset
R/W
0x000
SPI_INTFCONFA
SOFT
RESET_M
LSBFIRST_
M
ADDRINC_M
SDOACTIVE_M
SDOACTIVE
ADDRINC
LSBFIRST
SOFTRESET
0x00
R/W
0x003
CHIPTYPE
CHIPTYPE
0x04
R
0x004
PRODIDL
PRODIDL
0x44
R
0x005
PRODIDH
0x006
CHIPGRADE
0x008
SPI_PAGEINDX
0x00A
SCRATCH_PAD
0x011
PWRCNTRL0
0x012
TXENMASK
0x013
PWRCNTRL3
PRODIDH
PROD_GRADE
DEV_REVISION
RESERVED
DUAL_PAGE
SCRATCHPAD
PD_BG
PD_DAC_0 PD_DAC_1
PD_DAC_2
PD_DAC_3
PD_DACM
RESERVED
RESERVED
PDP_
PROTECT_
OUT
IRQEN_
SMODE_
CALPASS
IRQEN_
SMODE_
CALFAIL
TX_PROTECT_ RESERVED
OUT
0x014
GROUP_DLY
0x01F
IRQEN_
STATUSMODE0
0x020
IRQEN_
STATUSMODE1
0x021
IRQEN_
STATUSMODE2
IRQEN_
SMODE_
PDPERR0
RESERVED
IRQEN_
SMODE_
BLNKDONE0
0x022
IRQEN_
STATUSMODE3
IRQEN_
SMODE_
PDPERR1
RESERVED
0x023
IRQ_STATUS0
CALPASS
CALFAIL
0x03
R/W
R/W
0x7C
R/W
0x00
R/W
RESERVED
0x20
R/W
DUALB_
MASK
RESERVED
R
R
0x00
RESERVED
SPI_PROTECT_OUT SPI_PROTECT
0x91
0x06
DUALA_
MASK
0x88
R/W
IRQEN_SMODE_
SERPLLLOST
IRQEN_SMODE_
SERPLLLOCK
GROUP_DLY
RESERVED
IRQEN_
SMODE_
LANEFIFOERR
0x00
R/W
IRQEN_SMODE_
PRBS3
IRQEN_SMODE_
PRBS2
IRQEN_
SMODE_
PRBS1
IRQEN_
SMODE_
PRBS0
0x00
R/W
IRQEN_SMODE
_NCO_ALIGN0
IRQEN_SMODE_
SYNC_LOCK0
IRQEN_SMODE_
SYNC_ROTATE0
IRQEN_
SMODE_
SYNC_
WLIM0
0x00
IRQEN_
SMODE_
SYNC_TRIP0
R/W
IRQEN_
SMODE_
BLNKDONE1
IRQEN_SMODE
_NCO_ ALIGN1
IRQEN_SMODE_
SYNC_LOCK1
IRQEN_
SMODE_SYNC_
ROTATE1
IRQEN_
SMODE_
SYNC_
WLIM1
0x00
IRQEN_
SMODE_
SYNC_TRIP1
R/W
DACPLLLOST
DACPLLLOCK
SERPLLLOST
SERPLLLOCK
LANEFIFOERR
RESERVED
0x00
R
IRQEN_
IRQEN_SMODE
SMODE_
_ DACPLLLOCK
DACPLLLOST
RESERVED
0x024
IRQ_STATUS1
PRBS3
PRBS2
PRBS1
PRBS0
0x00
R
0x025
IRQ_STATUS2
PDPERR0
RESERVED
RESERVED
BLNKDONE0
NCO_
ALIGN0
SYNC_
LOCK0
SYNC_
ROTATE0
SYNC_
WLIM0
SYNC_
TRIP0
0x00
R
0x026
IRQ_STATUS3
PDPERR1
RESERVED
BLNKDONE1
NCO_
ALIGN1
SYNC_
LOCK1
SYNC_
ROTATE1
SYNC_
WLIM1
SYNC_
TRIP1
0x00
R
0x030
JESD_CHECKS
RESERVED
ERR_JESDBAD
ERR_KUNSUPP
ERR_
SUBCLASS
ERR_
INTSUPP
0x00
R
0x034
SYNC_
ERRWINDOW
0x00
R/W
0x038
SYNC_LASTERR_L
0x00
R
0x039
SYNC_LASTERR_H LASTUNDER
LASTOVER
0x00
R
0x03A
SYNC_CONTROL
SYNCENABLE
SYNCARM
0x00
R/W
0x03B
SYNC_STATUS
SYNC_
BUSY
0x00
R
0x03C
SYNC_CURRERR_L
0x00
R
ERR_DLYOVER ERR_WINLIMIT
RESERVED
ERRWINDOW
RESERVED
LASTERROR
RESERVED
SYNCCLRSTKY
RESERVED
SYNCCLRLAST
SYNCMODE
SYNC_LOCK
RESERVED
SYNC_
ROTATE
SYNC_WLIM
CURRERROR
Rev. A | Page 86 of 125
SYNC_
TRIP
Data Sheet
AD9144
Reg.
Name
Bit 7
Bit 6
0x03D
SYNC_
CURRERR_H
CURRUNDER
CURROVER
0x040
DACGAIN0_1
0x041
DACGAIN0_0
0x042
DACGAIN1_1
Bit 5
Bit 4
Bit 3
Bit 2
Bit 1
Bit 0
Reset
R/W
0x00
R
DACFSC_0[9:8]
0x00
R/W
0x00
R/W
DACFSC_1[9:8]
0x00
R/W
0x00
R/W
DACFSC_2[9:8]
0x00
R/W
0x00
R/W
0x00
R/W
0x00
R/W
0x00
R/W
RESERVED
RESERVED
DACFSC_0[7:0]
RESERVED
0x043
DACGAIN1_0
0x044
DACGAIN2_1
DACFSC_1[7:0]
0x045
DACGAIN2_0
0x046
DACGAIN3_1
0x047
DACGAIN3_0
0x050
NCOALIGN_
MODE
0x051
NCOKEY_ILSB
NCOKEYI[7:0]
0x00
R/W
0x052
NCOKEY_IMSB
NCOKEYI[15:8]
0x00
R/W
RESERVED
DACFSC_2[7:0]
RESERVED
DACFSC_3[9:8]
DACFSC_3[7:0]
NCO_
ALIGN_
ARM
RESERVED
NCO_ALIGN_FAIL
NCO_ALIGN_ NCO_ALIGN_
MTCH
PASS
RESERVED
NCO_ALIGN_MODE
0x053
NCOKEY_QLSB
NCOKEYQ[7:0]
0x00
R/W
0x054
NCOKEY_QMSB
NCOKEYQ[15:8]
0x00
R/W
0x00
R/W
0x060
PDP_THRES0
0x061
PDP_THRES1
PDP_THRESHOLD[7:0]
0x062
PDP_AVG_TIME
0x063
PDP_POWER0
0x064
PDP_POWER1
0x080
CLKCFG0
PD_CLK01 PD_CLK23
0x081
SYSREF_ACTRL0
RESERVED
0x082
SYSREF_ACTRL1
0x083
DACPLLCNTRL
RECAL_
DACPLL
RESERVED
0x084
DACPLLSTATUS
DACPLL_
OVERRANGE_H
DACPLL_
OVERRANGE_L
0x085
DACINTEGERWORD0
0x087
DACLOOPFILT1
RESERVED
PDP_THRESHOLD[12:8]
RESERVED
PDP_
ENABLE
PDP_AVG_TIME
PDP_POWER[7:0]
RESERVED
PDP_POWER[12:8]
PD_CLK_DIG PD_SERDES_
PCLK
PD_SYSREF
PD_CLK_REC
RESERVED
HYS_ON
SYSREF_RISE
HYS_CNTRL1
DACLOOPFILT2
0x089
DACLOOPFILT3
LF_
BYPASS_
R3
0x08A
DACCPCNTRL
RESERVED
0x08B
DACLOGENCNTRL
0x08C
DACLDOCNTRL1
0x00
R
0x08
R/W
LF_C1_WORD
0x88
R/W
LF_C3_WORD
0x88
R/W
LF_R3_WORD
0x08
R/W
DACPLL_
LOCK
RESERVED
CP_CURRENT
LO_DIV_MODE
RESERVED
DACLDOCNTRL2
0x0E2
CAL_CTRL_
GLOBAL
0x0E7
CAL_CLKDIV
RESERVED
0x0E8
CAL_PAGE
RESERVED
0x0E9
CAL_CTRL
REF_DIV_MODE
DAC_LDO
RESERVED
CAL_FIN
0x0ED
CAL_INIT
0x110
DATA_FORMAT
BINARY_
FORMAT
0x111
DATAPATH_CTRL
INVSINC_
ENABLE
CAL_
ACTIVE
CAL_ERRHI
CAL_START_
AVG
CAL_CLK_EN
CAL_EN_
AVG
RESERVED
CAL_PAGE
CAL_ERRLO
RESERVED
CAL_START
CAL_EN
CAL_INIT
RESERVED
RESERVED
DIG_GAIN_
ENABLE
PHASE_ADJ_
ENABLE
MODULATION_TYPE
Rev. A | Page 87 of 125
R/W
R/W
RESERVED
0x08D
R
0xF8
0x00
RESERVED
LF_BYPASS_C1
R
0x00
R/W
RESERVED
LF_R1_WORD
LF_
LF_BYPASS_
BYPASS_R1 C2
0x00
R/W
B_COUNT
0x088
R/W
0x10
ENABLE_
DACPLL
LF_C2_WORD
R/W
0x00
0x00
HYS_CNTRL0
DACPLL_
CAL_VALID
0x00
SEL_
SIDEBAND
I_TO_Q
0x20
R/W
0x02
R/W
0x01
R/W
0x2B
R/W
0x00
R/W
0x30
R/W
0x0F
R/W
0x00
R/W
A6
R/W
00
R/W
0xA0
R/W
AD9144
Reg.
Name
0x112
INTERP_MODE
0x113
NCO_FTW_
UPDATE
Data Sheet
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
Bit 2
RESERVED
Bit 1
Bit 0
INTERP_MODE
RESERVED
FTW_UPDATE_ FTW_
ACK
UPDATE_
REQ
Reset
R/W
0x01
R/W
0x00
R/W
0x114
FTW0
FTW[7:0]
0x00
R/W
0x115
FTW1
FTW[15:8]
0x00
R/W
0x116
FTW2
FTW[23:16]
0x00
R/W
0x117
FTW3
FTW[31:24]
0x00
R/W
0x118
FTW4
FTW[39:32]
0x00
R/W
0x119
FTW5
FTW[47:40]
0x10
R/W
0x11A
NCO_PHASE_
OFFSET0
NCO_PHASE_OFFSET[7:0]
0x00
R/W
0x11B
NCO_PHASE_
OFFSET1
NCO_PHASE_OFFSET[15:8]
0x00
R/W
0x11C
PHASE_ADJ0
PHASE_ADJ[7:0]
0x00
R/W
0x11D
PHASE_ADJ1
0x11F
TXEN_SM_0
0x121
TXEN_RISE_
COUNT_0
0x122
RESERVED
0x00
R/W
0x83
R/W
RISE_COUNT_0
0x0F
R/W
TXEN_RISE_
COUNT_1
RISE_COUNT_1
0x00
R/W
0x123
TXEN_FALL_
COUNT_0
FALL_COUNT_0
0xFF
R/W
0x124
TXEN_FALL_
COUNT_1
FALL_COUNT_1
0xFF
R/W
0x12D
DEVICE_CONFIG_
REG_0
DEVICE_CONFIG_0
0x46
R/W
0x12F
DIE_TEMP_CTRL0
0x20
R/W
0x132
DIE_TEMP0
0x00
R
FALL_COUNTERS
PHASE_ADJ[12:8]
RISE_COUNTERS
RESERVED
PROTECT_OUT_
INVERT
RESERVED
RESERVED
AUXADC_
ENABLE
DIE_TEMP[7:0]
0x133
DIE_TEMP1
0x134
DIE_TEMP_
UPDATE
RESERVED
DIE_TEMP[15:8]
DIE_TEMP_
UPDATE
0x00
R
0x00
R/W
0x135
DC_OFFSET_CTRL
RESERVED
DC_OFFSET_ 0x00
ON
R/W
0x136
IPATH_DC_
OFFSET_1PART0
LSB_OFFSET_I[7:0]
0x00
R/W
0x137
IPATH_DC_
OFFSET_1PART1
LSB_OFFSET_I[15:8]
0x00
R/W
0x138
QPATH_DC_
OFFSET_1PART0
LSB_OFFSET_Q[7:0]
0x00
R/W
0x139
QPATH_DC_
OFFSET_1PART1
LSB_OFFSET_Q[15:8]
0x00
R/W
0x13A
IPATH_DC_
OFFSET_2PART
RESERVED
SIXTEENTH_OFFSET_I
0x00
R/W
0x13B
QPATH_DC_
OFFSET_2PART
RESERVED
SIXTEENTH_OFFSET_Q
0x00
R/W
0x13C
IDAC_DIG_GAIN0
0x13D
IDAC_DIG_GAIN1
0x13E
QDAC_DIG_
GAIN0
0x13F
QDAC_DIG_GAIN1
0x140
GAIN_RAMP_UP_
STEP0
IDAC_DIG_GAIN[7:0]
RESERVED
IDAC_DIG_GAIN[11:8]
QDAC_DIG_GAIN[7:0]
RESERVED
QDAC_DIG_GAIN[11:8]
GAIN_RAMP_UP_STEP[7:0]
Rev. A | Page 88 of 125
0xEA
R/W
0x0A
R/W
0xEA
R/W
0x0A
R/W
0x04
R/W
Data Sheet
Reg.
Name
0x141
GAIN_RAMP_
UP_STEP1
0x142
GAIN_RAMP_
DOWN_STEP0
0x143
GAIN_RAMP_
DOWN_STEP1
0x146
DEVICE_CONFIG_
REG_1
AD9144
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
Bit 2
RESERVED
Bit 1
Bit 0
GAIN_RAMP_UP_STEP[11:8]
GAIN_RAMP_DOWN_STEP[7:0]
RESERVED
GAIN_RAMP_DOWN_STEP[11:8]
DEVICE_CONFIG_1
0x09
R/W
0x00
R/W
0x00
R/W
0x00
R
0x10
R/W
BSM_STAT
0x14B
PRBS
0x14C
PRBS_ERROR_I
PRBS_COUNT_I
0x00
R
0x14D
PRBS_ERROR_Q
PRBS_COUNT_Q
0x00
R
DACPLLT0
0x1B5
DACPLLT5
0x1B9
DACPLLT9
PRBS_
GOOD_Q
RESERVED
R/W
R/W
0x147
0x1B0
SOFTBLANKRB
Reset
0x00
RESERVED
PRBS_
GOOD_I
PRBS_MODE
PRBS_RESET
PRBS_EN
DAC_PLL_PWR
RESERVED
VCO_VAR
DAC_PLL_CP1
RESERVED
VCO_BIAS_TCF
VCO_BIAS_REF
0xFA
R/W
0x83
R/W
0x34
R/W
0x1BB
DACPLLTB
0x0C
R/W
0x1BC
DACPLLTC
DAC_PLL_VCO_CTRL
0x00
R/W
0x1BE
DACPLLTE
DAC_PLL_VCO_PWR
0x00
R/W
0x1BF
DACPLLTF
DAC_PLL_VCOCAL
0x8D
R/W
0x1C0
DACPLLT10
DAC_PLL_LOCK_CNTR
0x2E
R/W
0x1C1
DACPLLT11
DAC_PLL_CP2
0x24
R/W
0x1C4
DACPLLT17
DAC_PLL_VAR1
0x33
R/W
0x1C5
DACPLLT18
DAC_PLL_VAR2
0x08
R/W
0x200
MASTER_PD
0x01
R/W
0x201
PHY_PD
0x203
GENERIC_PD
0x206
CDR_RESET
0x230
CDR_OPERATING_
MODE_REG_0
0x232
DEVICE_CONFIG_
REG_3
0x268
EQ_BIAS_REG
0x280
SERDESPLL_
ENABLE_CNTRL
0x281
PLL_STATUS
0x284
LOOP_FILTER_1
LOOP_FILTER_1
0x77
R/W
0x285
LOOP_FILTER_2
LOOP_FILTER_2
0x87
R/W
0x286
LOOP_FILTER_3
LOOP_FILTER_3
0x08
R/W
0x287
SERDES_PLL_CP1
SERDES_PLL_CP1
0x3F
R/W
0x289
REF_CLK_
DIVIDER_LDO
0x00
R/W
0x28A
VCO_LDO
SERDES_PLL_VCO_LDO
0x2B
R/W
0x28B
SERDES_PLL_PD1
SERDES_PLL_PD1
0x7F
R/W
0x290
SERDESPLL_VAR1
SERDES_PLL_VAR1
0x83
R/W
0x294
SERDES_PLL_CP2
SERDES_PLL_CP2
0xB0
R/W
0x296
SERDESPLL_VCO1
SERDES_PLL_VCO1
0x0C
R/W
0x297
SERDESPLL_VCO2
SERDES_PLL_VCO2
0x00
R/W
RESERVED
SPI_PD_
MASTER
SPI_PD_PHY
RESERVED
SPI_
SYNC1_PD
RESERVED
RESERVED
ENHALFRATE
RESERVED
CDR_
OVERSAMP
0x00
R/W
SPI_
SYNC2_PD
0x00
R/W
SPI_CDR_
RESETN
0x01
R/W
RESERVED
0x28
R/W
0x0
R/W
0x62
R/W
0x00
R/W
DEVICE_CONFIG_3
EQ_POWER_MODE
RESERVED
RESERVED
RESERVED
RECAL_
SERDESPLL
SERDES_PLL_ SERDES_PLL_
OVERRANGE_ OVERRANGE_L
H
RESERVED
SERDES_PLL_CAL_
VALID_RB
RESERVED
DEVICE_
CONFIG_4
Rev. A | Page 89 of 125
RESERVED
ENABLE_
SERDESPLL
SERDES_PLL_ 0x00
LOCK_RB
SERDES_PLL_DIV_MODE
R
AD9144
Data Sheet
Reg.
Name
Reset
R/W
0x299
SERDES_PLL_PD2
Bit 7
Bit 6
Bit 5
Bit 4
SERDES_PLL_PD2
Bit 3
Bit 2
Bit 1
Bit 0
0x00
R/W
0x29A
SERDESPLL_VAR2
SERDES_PLL_VAR2
0xFE
R/W
R/W
0x29C
SERDES_PLL_CP3
SERDES_PLL_CP3
0x17
0x29F
SERDESPLL_VAR3
SERDES_PLL_VAR3
0x33
R/W
0x2A0
SERDESPLL_VAR4
SERDES_PLL_VAR4
0x08
R/W
0x2A4
DEVICE_CONFIG_
REG_8
DEVICE_CONFIG_8
0x4B
R/W
0x2A5
SYNCOUTB_
SWING
RESERVED
SYNCOUTB_
SWING_MD
0x00
R/W
0x2A7
TERM_BLK1_
CTRLREG0
RESERVED
RCAL_
TERMBLK1
0x00
R/W
0x2AA DEVICE_CONFIG_
REG_9
DEVICE_CONFIG_9
0xC3
R/W
0x2AB
DEVICE_CONFIG_
REG_10
DEVICE_CONFIG_10
0x93
R/W
0x2AE
TERM_BLK2_
CTRLREG0
0x00
R/W
0x2B1
DEVICE_CONFIG_
REG_11
DEVICE_CONFIG_11
0xC3
R/W
0x2B2
DEVICE_CONFIG_
REG_12
DEVICE_CONFIG_12
0x93
R/W
0x300
GENERAL_JRX_
CTRL_0
0x00
R/W
0x301
GENERAL_JRX_
CTRL_1
0x01
R/W
0x302
DYN_LINK_
LATENCY_0
RESERVED
DYN_LINK_LATENCY_0
0x00
R
0x303
DYN_LINK_
LATENCY_1
RESERVED
DYN_LINK_LATENCY_1
0x00
R
0x304
LMFC_DELAY_0
RESERVED
LMFC_DELAY_0
0x00
R/W
0x305
LMFC_DELAY_1
RESERVED
LMFC_DELAY_1
0x00
R/W
0x306
LMFC_VAR_0
RESERVED
LMFC_VAR_0
0x06
R/W
RESERVED
RESERVED CHECKSUM
_MODE
RESERVED
RCAL_
TERMBLK2
LINK_MODE
LINK_PAGE
RESERVED
SUBCLASSV_LOCAL
0x307
LMFC_VAR_1
0x308
XBAR_LN_0_1
RESERVED
RESERVED
0x309
XBAR_LN_2_3
RESERVED
0x30A
XBAR_LN_4_5
RESERVED
0x30B
XBAR_LN_6_7
RESERVED
LOGICAL_LANE7_SRC
0x30C
FIFO_STATUS_
REG_0
0x30D
FIFO_STATUS_
REG_1
RESERVED
LINK_EN
LMFC_VAR_1
0x06
R/W
LOGICAL_LANE0_SRC
0x08
R/W
LOGICAL_LANE3_SRC
LOGICAL_LANE2_SRC
0x1A
R/W
LOGICAL_LANE5_SRC
LOGICAL_LANE4_SRC
0x2C
R/W
LOGICAL_LANE6_SRC
0x3E
R/W
LANE_FIFO_FULL
0x00
R
LANE_FIFO_EMPTY
0x00
R
LOGICAL_LANE1_SRC
0x312
SYNCB_GEN_1
0x00
R/W
0x314
SERDES_SPI_REG
SYNCB_ERR_DUR
SERDES_SPI_CONFIG
0x00
R/W
0x315
PHY_PRBS_TEST_
EN
PHY_TEST_EN
0x00
R/W
0x316
PHY_PRBS_TEST_ RESERVED
CTRL
0x00
R/W
0x317
PHY_PRBS_TEST_
THRESHOLD_
LOBITS
PHY_PRBS_THRESHOLD[7:0]
0x00
R/W
0x318
PHY_PRBS_TEST_
THRESHOLD_
MIDBITS
PHY_PRBS_THRESHOLD[15:8]
0x00
R/W
0x319
PHY_PRBS_TEST_
THRESHOLD_
HIBITS
PHY_PRBS_THRESHOLD[23:16]
0x00
R/W
PHY_SRC_ERR_CNT
RESERVED
PHY_PRBS_PAT_SEL
Rev. A | Page 90 of 125
PHY_TEST_
START
PHY_TEST_
RESET
Data Sheet
AD9144
Reg.
Name
Reset
R/W
0x31A
PHY_PRBS_TEST_
ERRCNT_LOBITS
Bit 7
Bit 6
Bit 5
PHY_PRBS_ERR_CNT[7:0]
0x00
R
0x31B
PHY_PRBS_TEST_
ERRCNT_MIDBITS
PHY_PRBS_ERR_CNT[15:8]
0x00
R
0x31C
PHY_PRBS_TEST_
ERRCNT_HIBITS
PHY_PRBS_ERR_CNT[23:16]
0x00
R
0x31D
PHY_PRBS_TEST_
STATUS
PHY_PRBS_PASS
0xFF
R
0x32C
SHORT_TPL_
TEST_0
0x32D
SHORT_TPL_
TEST_1
0x32E
Bit 3
Bit 1
SHORT_TPL_REF_SP_LSB
0x00
R/W
SHORT_TPL_
TEST_2
SHORT_TPL_REF_SP_MSB
0x00
R/W
0x32F
SHORT_TPL_
TEST_3
RESERVED
0x00
R
0x333
DEVICE_CONFIG_
REG_13
DEVICE_CONFIG_13
0x00
R/W
0x334
JESD_BIT_
INVERSE_CTRL
JESD_BIT_INVERSE
0x00
R/W
0x400
DID_REG
0x401
BID_REG
0x402
LID0_REG
RESERVED ADJDIR_RD PHADJ_RD
0x403
SCR_L_REG
SCR_RD
0x404
F_REG
K_REG
M_REG
0x407
CS_N_REG
0x408
NP_REG
0x409
S_REG
SHORT_TPL_DAC_SEL
Bit 0
R/W
0x406
SHORT_TPL_SP_SEL
Bit 2
SHORT_TPL_ 0x00
TEST_EN
0x405
RESERVED
Bit 4
SHORT_TPL_
TEST_RESET
SHORT_
TPL_FAIL
DID_RD
0x00
R
0x00
R
LID0_RD
0x00
R
L-1_RD
0x00
R
0x00
R
ADJCNT_RD
BID_RD
RESERVED
F-1_RD
RESERVED
K-1_RD
0x00
R
0x00
R
N-1_RD
0x00
R
SUBCLASSV_RD
NP-1_RD
0x00
R
JESDV_RD
S-1_RD
0x00
R
CF_RD
M-1_RD
CS_RD
HD_RD
RESERVED
0x40A
HD_CF_REG
0x00
R
0x40B
RES1_REG
RESERVED
RES1_RD
0x00
R
0x40C
RES2_REG
RES2_RD
0x00
R
R
0x40D
CHECKSUM_REG
FCHK0_RD
0x00
0x40E
COMPSUM0_REG
FCMP0_RD
0x00
R
0x412
LID1_REG
0x00
R
0x415
CHECKSUM1_REG
FCHK1_RD
0x00
R
0x416
COMPSUM1_REG
FCMP1_RD
0x00
R
0x41A
LID2_REG
0x00
R
0x41D
CHECKSUM2_REG
FCHK2_RD
0x00
R
0x41E
COMPSUM2_REG
FCMP2_RD
0x00
R
0x422
LID3_REG
0x00
R
0x425
CHECKSUM3_REG
FCHK3_RD
0x00
R
0x426
COMPSUM3_REG
FCMP3_RD
0x00
R
0x42A
LID4_REG
0x00
R
0x42D
CHECKSUM4_REG
0x00
R
0x42E
COMPSUM4_REG
0x432
LID5_REG
0x435
CHECKSUM5_REG
0x436
COMPSUM5_REG
0x43A
LID6_REG
0x43D
CHECKSUM6_REG
RESERVED
LID1_RD
RESERVED
LID2_RD
RESERVED
LID3_RD
RESERVED
LID4_RD
FCHK4_RD
FCMP4_RD
RESERVED
LID5_RD
FCHK5_RD
FCMP5_RD
RESERVED
LID6_RD
FCHK6_RD
Rev. A | Page 91 of 125
0x00
R
0x00
R
0x00
R
0x00
R
0x00
R
0x00
R
AD9144
Data Sheet
Reg.
Name
0x43E
COMPSUM6_REG
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
0x442
LID7_REG
0x445
CHECKSUM7_REG
0x446
COMPSUM7_REG
0x450
ILS_DID
DID
Bit 2
Bit 1
Bit 0
Reset
R/W
0x00
R
0x00
R
FCHK7_RD
0x00
R
FCMP7_RD
0x00
R
0x00
R/W
0x00
R/W
0x00
R/W
FCMP6_RD
RESERVED
0x451
ILS_BID
0x452
ILS_LID0
RESERVED ADJDIR
0x453
ILS_SCR_L
SCR
0x454
ILS_F
0x455
ILS_K
0x456
ILS_M
0x457
ILS_CS_N
0x458
ILS_NP
0x459
ILS_S
0x45A
ILS_HD_CF
LID7_RD
ADJCNT
BID
PHADJ
LID0
RESERVED
L-1
F-1
RESERVED
K-1
M-1
CS
RESERVED
0x83
R/W
0x00
R/W
0x1F
R/W
0x01
R/W
N-1
0x0F
R/W
SUBCLASSV
NP-1
0x2F
R/W
JESDV
S-1
0x20
R/W
CF
0x80
R/W
HD
RESERVED
0x45B
ILS_RES1
RES1
0x00
R/W
0x45C
ILS_RES2
RES2
0x00
R/W
0x45D
ILS_CHECKSUM
FCHK0
0x45
R/W
0x46B
ERRCNTRMON_RB
0x00
R
0x46B
ERRCNTRMON
0x46C
LANEDESKEW
0x46D
BADDISPARITY_RB
0x46D
BADDISPARITY
0x46E
NIT_RB
0x46E
NIT_W
0x46F
UNEXPECTEDCONTROL_RB
0x46F
UNEXPECTEDCONTROL_W
READERRORCNTR
RESERVED
LANESEL
RESERVED
CNTRSEL
LANEDESKEW
BADDIS
RST_IRQ_ DISABLE_ RST_ERR_
DIS
ERR_CNTR_ CNTR_DIS
DIS
RESERVED
RST_IRQ_ DISABLE_ RST_ERR_
NIT
ERR_CNTR_ CNTR_NIT
NIT
RESERVED
LANE_ADDR_DIS
NIT
LANE_ADDR_NIT
UCC
RESERVED
RST_IRQ_ DISABLE_ RST_ERR_
UCC
ERR_CNTR_ CNTR_UCC
UCC
LANE_ADDR_UCC
0x00
R/W
0x0F
R/W
0x00
R
0x00
R/W
0x00
R
0x00
R/W
0x00
R
0x00
R/W
0x470
CODEGRPSYNCFLG
0x471
FRAMESYNCFLG
CODEGRPSYNC
0x00
R/W
FRAMESYNC
0x00
R/W
0x472
0x473
GOODCHKSUMFLG
GOODCHECKSUM
0x00
R/W
INITLANESYNCFLG
INITIALLANESYNC
0x00
0x476
CTRLREG1
R/W
F
0x01
R/W
0x477
CTRLREG2
0x00
R/W
0x478
KVAL
0x01
R/W
0x47A
IRQVECTOR_MASK BADDIS_
MASK
NIT_MASK
UCC_
MASK
RESERVED
INITIALLANESYNC_
MASK
BADCHECK
SUM_MASK
0x00
CODEGRP
SYNC_MASK
R/W
0x47A
IRQVECTOR_FLAG BADDIS_
FLAG
NIT_FLAG
UCC_FLAG
RESERVED
INITIALLANESYNC_
FLAG
BADCHECKSUM FRAMESYNC_
_FLAG
FLAG
0x47B
SYNCASSERTIONMASK
UCC_S
CMM
CMM_ENABLE
RESERVED
0x47C
ERRORTHRES
ETH
0xFF
R/W
0x47D
LANEENABLE
LANE_ENA
0x0F
R/W
0x47E
RAMP_ENA
ENA_RAMP_ 0x00
CHECK
R/W
RESERVED
ILAS_
MODE
RESERVED
THRESHOLD_
MASK_EN
KSYNC
BADDIS_S NIT_S
RESERVED
Rev. A | Page 92 of 125
FRAMESYNC_
MASK
CODEGRP
SYNC_FLAG
0x00
R
0x008
R/W
Data Sheet
AD9144
Reg.
Name
0x520
DIG_TEST0
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
0x521
DC_TEST_VALUEI0
DC_TEST_VALUEI[7:0]
0x522
DC_TEST_VALUEI1
DC_TEST_VALUEI[15:8]
0x00
R/W
0x523
DC_TEST_
VALUEQ0
DC_TEST_VALUEQ[7:0]
0x00
R/W
0x524
DC_TEST_
VALUEQ1
DC_TEST_VALUEQ[15:8]
0x00
R/W
RESERVED
Rev. A | Page 93 of 125
Bit 2
Bit 1
Bit 0
Reset
R/W
DC_TEST_
MODE
RESERVED
0x1C
R/W
0x00
R/W
AD9144
Data Sheet
DEVICE CONFIGURATION REGISTER DESCRIPTIONS
Table 98. Device Configuration Register Descriptions
Address
0x000
Name
SPI_INTFCONFA
Bit No.
7
6
5
4
3
2
Bit Name
SOFTRESET_M
LSBFIRST_M
ADDRINC_M
SDOACTIVE_M
SDOACTIVE
ADDRINC
Settings
1
0
1
LSBFIRST
1
0
0
SOFTRESET
1
0x003
CHIPTYPE
[7:0]
CHIPTYPE
0x004
0x005
0x006
PRODIDL
PRODIDH
CHIPGRADE
0x008
SPI_PAGEINDX
[7:0]
[7:0]
[7:4]
[3:0]
[7:2]
[1:0]
PRODIDL
PRODIDH
PROD_GRADE
DEV_REVISION
RESERVED
DUAL_PAGE
0b01
0b10
0b11
0x00A
SCRATCH_PAD
[7:0]
SCRATCHPAD
0x011
PWRCNTRL0
7
PD_BG
6
PD_DAC_0
5
PD_DAC_1
1
1
1
Description
Soft Reset (Mirror).
LSB First (Mirror).
Address Increment (Mirror).
SDO Active (Mirror).
SDO Active.
Address Increment. Controls whether
addresses are incremented or decremented
during multibyte data transfers.
Addresses are incremented during multibyte
data transfers
Addresses are decremented during
multibyte data transfers
LSB First. Controls whether input and output
data are oriented as LSB first or MSB first.
Shift LSB in first
Shift MSB in first
Soft Reset. Setting this bit initiates a reset.
This bit is autoclearing after the soft reset is
complete.
Assert soft reset
The product type is “High Speed DAC”, which
is represented by a code of 0x04.
Product Identification Low.
Product Identification High.
Product Grade.
Device Revision.
Reserved.
Dual Paging. Selects which dual DAC pair is
accessed and written to when changing
digital features, such as digital gain, dc offset,
NCO FTW, and others. This paging affects
Register 0x013 to Register 0x014,
Register 0x034 to Register 0x03d,
Register 0x050 to Register 0x064,
Register 0x110 to Register 0x124, and
Register 0x135 to Register 0x14D.
Read and write Dual A
Read and write Dual B
Write both duals; read Dual A
This register does not affect any functions in
the device and can be used for testing SPI
communication with the part. Any value
written to this register will be read back to
reflect the change unless a reset or powercycle occurs.
Reference Power-Down. Powers down the
band gap reference for the entire chip.
Circuits will not be provided with bias
currents.
Power down reference
Powers Down DAC0. Powers down the Ichannel DAC of Dual A.
Powers down DAC0
Powers Down DAC1. Powers down the Qchannel DAC of Dual A.
Powers down DAC 1
Rev. A | Page 94 of 125
Reset
0x0
0x0
0x0
0x0
0x0
0x0
Access
R
R
R
R
R/W
R/W
0x0
R/W
0x0
R/W
0x4
R
0x44
0x91
0x0
0x6
0x0
0x3
R
R
R
R
R
R/W
0x00
R/W
0x0
R/W
0x1
R/W
0x1
R/W
Data Sheet
Address
Name
AD9144
Bit No.
4
Bit Name
PD_DAC_2
Settings
3
PD_DAC_3
2
PD_DACM
[1:0]
[7:2]
1
RESERVED
RESERVED
DUALB_MASK
0
DUALA_MASK
7
6
RESERVED
PDP_PROTECT_
OUT
5
4
3
TX_PROTECT_OUT
RESERVED
SPI_PROTECT_
OUT
SPI_PROTECT
RESERVED
RESERVED
GROUP_DLY
1
1
1
0x012
TXENMASK
1
1
0x013
PWRCNTRL3
0x014
GROUP_DLY
0x01F
IRQEN_
STATUSMODE0
2
[1:0]
[7:4]
[3:0]
7
IRQEN_SMODE_
CALPASS
1
1
1
1
0
6
IRQEN_SMODE_
CALFAIL
1
0
5
IRQEN_SMODE_
DACPLLLOST
1
0
4
IRQEN_SMODE_
DACPLLLOCK
1
0
3
IRQEN_SMODE_
SERPLLLOST
1
0
Description
Powers Down DAC2. Powers down the Ichannel DAC of Dual B.
Powers down DAC 2
Powers Down DAC3. Powers down the Qchannel DAC of Dual B.
Powers down DAC 3
Powers Down the DAC Master Bias. The
master bias cell provides currents and DAC
full-scale adjustments to the four DACs. With
the DAC master bias powered down, the
DACs are inoperative.
Powers down the DAC master bias
Reserved.
Reserved.
Dual B TXEN1 Mask. Power down Dual B on a
falling edge of TXEN1.
If TXEN1 is low, power down DAC2 and DAC3
Dual A TXEN0 Mask. Power down Dual A on a
falling edge of TXEN0.
If TXEN0 is low, power down DAC0 and DAC1
Reserved.
PDP_PROTECT triggers PROTECT_OUTx.
Reset
0x1
Access
R/W
0x1
R/W
0x1
R/W
0x0
0x0
0x0
R
R
R/W
0x0
R/W
0x0
0x0
R
R/W
TX_PROTECT triggers PROTECT_OUTx.
Reserved.
SPI_PROTECT triggers PROTECT_OUTx.
0x1
0x0
0x0
R/W
R
R/W
SPI_PROTECT
Reserved.
Reserved.
Group Delay Control. Delays the I and Q
channel outputs together. 0 = minimum
delay. 15 = maximum delay. The range of the
delay is −4 to +3.5 DAC clock periods, and
the resolution is 1/2 DAC clock period.
Calibration Pass Detection Status Mode.
If CALPASS goes high, it latches and pulls IRQ
low
CALPASS shows current status
Calibration Fail Detection Status Mode.
If CALFAIL goes high, it latches and pulls IRQ
low
CALFAIL shows current status
DAC PLL Lost Detection Status Mode.
If DACPLLLOST goes high, it latches and pulls
IRQ low
DACPLLLOST shows current status
DAC PLL Lock Detection Status Mode.
If DACPLLLOCK goes high, it latches and
pulls IRQ low
DACPLLLOCK shows current status
SERDES PLL Lost Detection Status Mode.
If SERPLLLOST goes high, it latches and pulls
IRQ low
SERPLLLOST shows current status
0x0
0x0
0x8
0x8
R/W
R
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Rev. A | Page 95 of 125
AD9144
Address
Name
Data Sheet
Bit No.
2
Bit Name
IRQEN_SMODE_
SERPLLLOCK
Settings
1
0
1
IRQEN_SMODE_
LANEFIFOERR
1
0
0x020
IRQEN_
STATUSMODE1
0
[7:4]
RESERVED
RESERVED
3
IRQEN_SMODE_
PRBS3
1
0
2
IRQEN_SMODE_
PRBS2
1
0
1
IRQEN_SMODE_
PRBS1
1
0
0
IRQEN_SMODE_
PRBS0
1
0
0x021
IRQEN_
STATUSMODE2
7
IRQEN_SMODE_
PDPERR0
1
0
6
5
RESERVED
IRQEN_SMODE_
BLNKDONE0
1
0
4
IRQEN_SMODE_
NCO_ALIGN0
1
0
3
IRQEN_SMODE_
SYNC_LOCK0
1
0
2
IRQEN_SMODE_
SYNC_ROTATE0
1
0
1
IRQEN_SMODE_
SYNC_WLIM0
1
0
0
IRQEN_SMODE_
SYNC_TRIP0
1
0
Description
SERDES PLL Lock Detection Status Mode.
If SERPLLLOCK goes high, it latches and pulls
IRQ low
SERPLLLOCK shows current status
Lane FIFO Error Detection Status Mode.
If LANEFIFOERR goes high, latches and pulls
IRQ low
LANEFIFOERR shows current status
Reserved.
Reserved.
DAC3 PRBS Error Status Mode.
If PRBS3 goes high, it latches and pulls IRQ
low
PRBS3 shows current status
DAC2 PRBS Error Status Mode.
If PRBS2 goes high, it latches and pulls IRQ
low
PRBS2 shows current status
DAC1 PRBS Error Status Mode.
If PRBS1 goes high, it latches and pulls IRQ
low
PRBS1 shows current status
DAC0 PRBS Error Status Mode.
If PRBS0 goes high, it latches and pulls IRQ
low
PRBS0 shows current status
Dual A PDP Error.
If PDPERR0 goes high, it latches and pulls IRQ
low
PDPERR0 shows current status
Reserved.
Dual A Blanking Done Status Mode.
If BLNKDONE0 goes high, it latches and pulls
IRQ low
BLNKDONE0 shows current status
Dual A NCO Align Tripped Status Mode
If NCO_ALIGN0 goes high, it latches and pulls
IRQ low
NCO_ALIGN0 shows current status
Dual A Alignment Locked Status Mode.
If SYNC_LOCK0 goes high, it latches and pulls
IRQ low
SYNC_LOCK0 shows current status
Dual A Alignment Rotate Status Mode.
If SYNC_ROTATE0 goes high, it latches and
pulls IRQ low
SYNC_ROTATE0 shows current status
Dual A Outside Window Status Mode.
If SYNC_WLIM0 goes high, it latches and
pulls IRQ low
SYNC_WLIM0 shows current status
Dual A Alignment Tripped Status Mode.
If SYNC_TRIP0 goes high, it latches and pulls
IRQ low
SYNC_TRIP0 shows current status
Rev. A | Page 96 of 125
Reset
0x0
Access
R/W
0x0
R/W
0x0
0x0
R
R
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
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x022
Name
IRQEN_
STATUSMODE3
AD9144
Bit No.
7
Bit Name
IRQEN_SMODE_
PDPERR1
Settings
1
0
6
5
RESERVED
IRQEN_SMODE_
BLNKDONE1
1
0
4
IRQEN_SMODE_
NCO_ALIGN1
1
0
3
IRQEN_SMODE_
SYNC_LOCK1
1
0
2
IRQEN_SMODE_
SYNC_ROTATE1
1
0
1
IRQEN_SMODE_
SYNC_WLIM1
1
0
0
IRQEN_SMODE_
SYNC_TRIP1
1
0
0x023
IRQ_STATUS0
7
CALPASS
6
CALFAIL
5
DACPLLLOST
4
DACPLLLOCK
3
SERPLLLOST
1
1
1
1
1
Description
Dual B PDP Error.
If PDPERR1 goes high, it latches and pulls IRQ
low
PDPERR1 shows current status
Reserved.
Dual B Blanking Done Status Mode.
If BLNKDONE1 goes high, it latches and pulls
IRQ low
BLNKDONE1 shows current status
Dual B NCO Align Tripped Status Mode
If NCO_ALIGN1 goes high, it latches and pulls
IRQ low
NCO_ALIGN1 shows current status
Dual B Alignment Locked Status Mode.
If SYNC_LOCK1 goes high, it latches and pulls
IRQ low
SYNC_LOCK1 shows current status
Dual B Alignment Rotate Status Mode.
If SYNC_ROTATE1 goes high, it latches and
pulls IRQ low
SYNC_ROTATE1 shows current status
Dual B Outside Window Status Mode.
If SYNC_WLIM1 goes high, it latches and
pulls IRQ low
SYNC_WLIM1 shows current status
Dual B Alignment Tripped Status Mode.
If SYNC_TRIP1 goes high, it latches and pulls
IRQ low
SYNC_TRIP1 shows current status
Calibration Pass Status. If
IRQEN_SMODE_CALPASS is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Calibration passed
Calibration Fail Detection Status. If
IRQEN_SMODE_CALFAIL is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Calibration failed
DAC PLL Lost Status. If
IRQEN_SMODE_DACPLLLOST is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
DAC PLL lock was lost
DAC PLL Lock Status. If
IRQEN_SMODE_DACPLLLOCK is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
DAC PLL locked
SERDES PLL Lost Status. If
IRQEN_SMODE_SERPLLLOST is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
SERDES PLL lock was lost
Rev. A | Page 97 of 125
Reset
0x0
Access
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R
0x0
R
0x0
R
0x0
R
0x0
R
AD9144
Address
Name
Data Sheet
Bit No.
2
Bit Name
SERPLLLOCK
1
LANEFIFOERR
0
[7:4]
3
RESERVED
RESERVED
PRBS3
2
PRBS2
1
PRBS1
0
PRBS0
7
PDPERR0
6
5
RESERVED
BLNKDONE0
4
NCO_ALIGN0
Settings
1
1
0x024
IRQ_STATUS1
1
1
1
1
0x025
IRQ_STATUS2
1
1
1
Description
SERDES PLL Lock Status. If
IRQEN_SMODE_SERPLLLOCK is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
SERDES PLL locked
Lane FIFO Error Status. If
IRQEN_SMODE_LANEFIFOERR is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low.
A lane FIFO error occurs when there is a full
or empty condition on any of the FIFOs
between the deserializer block and the core
digital. This error requires a link disable and
reenable to remove it. The status of the lane
FIFOs can be found in Register 0x30C (FIFO
full), and Register 0x30D (FIFO empty).
Lane FIFO error
Reserved.
Reserved.
DAC3 PRBS Error Status. If
IRQEN_SMODE_PRBS3 is low, this bit shows
current status. If not, this bit latches on a
rising edge and pull IRQ low. When latched,
write a 1 to clear this bit.
DAC3 failed PRBS
DAC2 PRBS Error Status. If
IRQEN_SMODE_PRBS2 is low, this bit shows
current status. If not, this bit latches on a
rising edge and pull IRQ low. When latched,
write a 1 to clear this bit.
DAC2 failed PRBS
DAC1 PRBS Error Status. If
IRQEN_SMODE_PRBS1 is low, this bit shows
current status. If not, this bit latches on a
rising edge and pull IRQ low. When latched,
write a 1 to clear this bit.
DAC1 failed PRBS
DAC0 PRBS Error Status. If
IRQEN_SMODE_PRBS0 is low, this bit shows
current status. If not, this bit latches on a
rising edge and pull IRQ low. When latched,
write a 1 to clear this bit.
DAC0 failed PRBS
Dual A PDP Error. If IRQEN_SMODE_PAERR0
is low, this bit shows current status. If not,
this bit latches on a rising edge and pull IRQ
low. When latched, write a 1 to clear this bit.
Data into Dual A over power threshold
Reserved.
Dual A Blanking Done Status. If
IRQEN_SMODE_BLNKDONE0 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual A blanking done
Dual A NCO Align Tripped Status. If
IRQEN_SMODE_NCO_ALIGN0 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual A NCO align tripped
Rev. A | Page 98 of 125
Reset
0x0
Access
R
0x0
R
0x0
0x0
0x0
R
R
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
Data Sheet
Address
Name
AD9144
Bit No.
3
Bit Name
SYNC_LOCK0
2
SYNC_ROTATE0
1
SYNC_WLIM0
0
SYNC_TRIP0
7
PDPERR1
6
5
RESERVED
BLNKDONE1
4
NCO_ALIGN1
3
SYNC_LOCK1
Settings
1
1
1
1
0x026
IRQ_STATUS3
1
1
1
1
2
SYNC_ROTATE1
1
SYNC_WLIM1
1
1
Description
Dual A LMFC Alignment Locked Status. If
IRQEN_SMODE_SYNC_LOCK0 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual A LMFC alignment locked
Dual A LMFC Alignment Rotate Status. If
IRQEN_SMODE_SYNC_ROTATE0 is low, this
bit shows current status. If not, this bit
latches on a rising edge and pull IRQ low.
When latched, write a 1 to clear this bit.
Dual A LMFC alignment rotated
Dual A Outside Window Status. If
IRQEN_SMODE_SYNC_WLIM0 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual A LMFC phase outside of window
Dual A LMFC Alignment Tripped Status. If
IRQEN_SMODE_SYNC_TRIP0 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual A LMFC alignment tripped
Dual B PDP Error. If IRQ_SMODE_PDPERR1 is
low, this bit shows current status. If not, this
bit latches on a rising edge and pull IRQ low.
When latched, write a 1 to clear this bit.
Data into Dual B over power threshold
Reserved.
Dual B Blanking Done Status. If
IRQEN_SMODE_BLNKDONE1 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual B blanking done
Dual B NCO Align Tripped Status. If
IRQEN_SMODE_NCO_ALIGN1 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual B NCO align tripped
Dual B LMFC Alignment Locked Status. If
IRQEN_SMODE_SYNC_LOCK1 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual B LMFC alignment locked
Dual B LMFC Alignment Rotate Status. If
IRQEN_SMODE_SYNC_ROTATE1 is low, this
bit shows current status. If not, this bit
latches on a rising edge and pull IRQ low.
When latched, write a 1 to clear this bit.
Dual B LMFC alignment rotated
Dual B Outside Window Status. If
IRQEN_SMODE_SYNC_WLIM1 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual B LMFC phase outside of window
Rev. A | Page 99 of 125
Reset
0x0
Access
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
0x0
R
AD9144
Data Sheet
Address
Name
Bit No.
0
Bit Name
SYNC_TRIP1
0x030
JESD_CHECKS
[7:6]
5
RESERVED
ERR_DLYOVER
4
ERR_WINLIMIT
Settings
1
1
1
3
ERR_JESDBAD
1
2
ERR_KUNSUPP
1
ERR_SUBCLASS
0
ERR_INTSUPP
1
1
1
0x034
SYNC_ERRWINDOW
[7:2]
[1:0]
RESERVED
ERRWINDOW
0x038
SYNC_LASTERR_L
[7:4]
[3:0]
RESERVED
LASTERROR
0x039
SYNC_LASTERR_H
7
LASTUNDER
1
6
LASTOVER
1
0x03A
SYNC_CONTROL
[5:0]
7
RESERVED
SYNCENABLE
1
0
6
SYNCARM
1
5
SYNCCLRSTKY
4
SYNCCLRLAST
[3:0]
SYNCMODE
0b0001
0b0010
0b1000
0b1001
Description
Dual B LMFC Alignment Tripped Status. If
IRQEN_SMODE_SYNC_TRIP1 is low, this bit
shows current status. If not, this bit latches
on a rising edge and pull IRQ low. When
latched, write a 1 to clear this bit.
Dual B LMFC alignment tripped
Reserved.
Error: LMFC_Delay > JESD_K Parameter.
LMFC_Delay > JESD_K
Unsupported Window Limit.
Unsupported SYSREF window limit
Unsupported M/L/S/F Selection.
This JESD combination is not supported
Unsupported K Values. 16 and 32 are
supported.
K value unsupported
Unsupported Subclass Value. 0 and 1 are
supported.
Unsupported subclass value
Unsupported Interpolation Rate Factor. 1, 2,
4, 8 are supported.
Unsupported interpolation rate factor
Reserved.
LMFC Sync Error Window. The error window
allows the SYSREF sample phase to vary
within the confines of the window without
triggering a clock adjustment. This is useful if
SYSREF cannot be guaranteed to always
arrive in the same period of the device clock
associated with the target phase.
Error window tolerance = ± ERRWINDOW
Reserved.
LMFC Sync Last Alignment Error. 4-bit twos
complement value that represents the phase
error (in number of DAC clock cycles) when the
clocks were last adjusted.
LMFC Sync Last Error Under Flag.
Last phase error was beyond lower window
tolerance boundary
LMFC Sync Last Error Over Flag.
Last phase error was beyond upper window
tolerance boundary
Reserved.
LMFC Sync Logic Enable.
Enable sync logic
Disable sync logic
LMFC Sync Arming Strobe.
Sync one-shot armed
LMFC Sync Sticky Bit Clear. On a rising edge,
this bit clears SYNC_ROTATE and SYNC_TRIP.
LMFC Sync Clear Last Error. On a rising edge,
this bit clears LASTERROR, LASTUNDER,
LASTOVER.
LMFC Sync Mode.
Sync one-shot mode
Sync continuous mode
Sync monitor only mode
Sync one-shot, then monitor
Rev. A | Page 100 of 125
Reset
0x0
Access
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R/W
0x0
R
R
0x0
R
0x0
R
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x03B
Name
SYNC_STATUS
AD9144
Bit No.
7
Bit Name
SYNC_BUSY
[6:4]
3
RESERVED
SYNC_LOCK
2
SYNC_ROTATE
Settings
1
1
1
1
SYNC_WLIM
1
0
SYNC_TRIP
1
0x03C
SYNC_CURRERR_L
[7:4]
[3:0]
RESERVED
CURRERROR
0x03D
SYNC_CURRERR_H
7
CURRUNDER
1
6
CURROVER
1
[5:0]
[7:2]
[1:0]
RESERVED
RESERVED
DACFSC_0[9:8]
DACGAIN0_0
DACGAIN1_1
[7:0]
[7:2]
[1:0]
DACFSC_0[7:0]
RESERVED
DACFSC_1[9:8]
0x043
0x044
DACGAIN1_0
DACGAIN2_1
[7:0]
[7:2]
[1:0]
DACFSC_1[7:0]
RESERVED
DACFSC_2[9:8]
0x045
DACGAIN2_0
[7:0]
DACFSC_2[7:0]
0x040
DACGAIN0_1
0x041
0x042
Description
LMFC Sync Machine Busy.
Sync logic SM is busy
Reserved.
LMFC Sync Alignment Locked.
Sync logic aligned within window
LMFC Sync Rotated.
Sync logic rotated with SYSREF (sticky)
LMFC Sync Alignment Limit Range.
Phase error outside window threshold
LMFC Sync Tripped After Arming.
Sync received SYSREF pulse (sticky)
Reserved.
LMFC Sync Alignment Error. 4-bit twos
complement value that represents the phase
error in number of DAC clock cycles (that is,
number of DAC clocks between LMFC edge and
SYSREF edge).
When an adjustment of the clocks is made
on any given SYSREF, the value of the phase
error is placed into SYNC_ LASTERR, and
SYNC_CURRERR is forced to 0.
LMFC Sync Current Error Under Flag.
Current phase error is beyond lower window
tolerance boundary
LMFC Sync Current Error Over Flag.
Current phase error is beyond upper window
tolerance boundary
Reserved.
Reserved.
2 MSBs of I-Channel DAC Gain Dual A. A 10bit twos complement value that is mapped
to analog full-scale current for DAC 0 as
shown:
01111111111 = 27.0 mA
0000000000 = 20.48 mA
1000000000 = 13.9 mA
8 LSBs of I-Channel DAC Gain Dual A.
Reserved.
2 MSBs of Q-Channel DAC Gain Dual A. A 10bit twos complement value that is mapped
to analog full-scale current for DAC 1 as
shown in Register 0x040.
01111111111 = 27.0 mA
0000000000 = 20.48 mA
1000000000 = 13.9 mA
8 LSBs of Q-Channel DAC Gain Dual A.
Reserved.
2 MSBs of I-Channel DAC Gain Dual B. A 10bit twos complement value that is mapped
to analog full-scale current for DAC as shown
in Register 0x040.
01111111111 = 27.0 mA
0000000000 = 20.48 mA
1000000000 = 13.9 mA
8 LSBs of I-Channel DAC Gain Dual B.
Rev. A | Page 101 of 125
Reset
0x0
Access
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
0x0
0x0
R
R
R/W
0x0
0x0
0x0
R/W
R
R/W
0x0
0x0
0x0
R/W
R
R/W
0x0
R/W
AD9144
Data Sheet
Address
0x046
Name
DACGAIN3_1
Bit No.
[7:2]
[1:0]
Bit Name
RESERVED
DACFSC_3[9:8]
Settings
0x047
0x050
DACGAIN3_0
NCOALIGN_MODE
[7:0]
7
DACFSC_3[7:0]
NCO_ALIGN_ARM
6
5
RESERVED
NCO_ALIGN_
MTCH
Description
Reserved.
2 MSBs of Q-Channel DAC Gain Dual B. A 10bit twos complement value that is mapped
to analog full-scale current for DAC 3 as
shown in Register 0x40.
01111111111 = 27.0 mA
0000000000 = 20.48 mA
1000000000 = 13.9 mA
8 LSBs of Q-Channel DAC Gain Dual B.
Arm NCO Align. On a rising edge, arms the
NCO align operation.
Reserved.
NCO Align Data Match.
1
0
4
NCO_ALIGN_PASS
1
0
3
NCO_ALIGN_FAIL
1
0
2
[1:0]
RESERVED
NCO_ALIGN_
MODE
00
10
01
0x051
0x052
0x053
0x054
0x060
NCOKEY_ILSB
NCOKEY_IMSB
NCOKEY_QLSB
NCOKEY_QMSB
PDP_THRES0
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
NCOKEYI[7:0]
NCOKEYI[15:8]
NCOKEYQ[7:0]
NCOKEYQ[15:8]
PDP_THRESHOLD[7:0]
0x061
PDP_THRES1
[7:5]
[4:0]
0x062
PDP_AVG_TIME
7
[6:4]
[3:0]
RESERVED
PDP_
THRESHOLD[12:8]
PDP_ENABLE
RESERVED
PDP_AVG_TIME
0x063
PDP_POWER0
[7:0]
PDP_POWER[7:0]
0x064
PDP_POWER1
[7:5]
[4:0]
RESERVED
PDP_POWER[12:8]
1
Key NCO align data match
If finished, NCO not aligned on data match
NCO Align Pass.
NCO align takes effect
Clear not taken effect yet
NCO Align Fail.
NCO reset during rotate
Not finished yet
Reserved.
NCO Align Mode.
NCO align disabled
NCO align on data key
NCO align on SYSREF
NCO Data Key for I Channel.
NCO Data Key for I Channel.
NCO Data Key for Q Channel.
NCO Data Key for Q Channel.
PDP_THRESHOLD is the average power
threshold for comparison. If the moving
average of signal power crosses this
threshold, PDP_PROTECT is set high.
Reserved.
See Register 0x60.
Enable average power calculation.
Reserved.
Can be set from 0-10. Averages across
2(9 + PDP_AVG_TIME) IQ sample pairs.
If PDP_POWER has not gone over PDP_
THRESHOLD, PDP_POWER reads back the
moving average of the signal power (I2 + Q2).
If PDP_THRESHOLD is crossed, PDP_POWER
will hold the max value until its corresponding
IRQ is cleared (0x025[7 or 0x026[7]). Only 6
data MSBs are used in calculating power.
Reserved.
See Register 0x063.
Rev. A | Page 102 of 125
Reset
0x0
0x0
Access
R
R/W
0x0
0x0
R/W
R/W
0x0
0x0
R
R
0x0
R
0x0
R
0x0
0x0
R
R/W
0x0
0x0
0x0
0x0
0x0
R/W
R/W
R/W
R/W
R/W
0x0
0x0
R
R/W
0x0
0x0
0x0
R/W
R
R/W
0x0
R
0x0
0x0
R
R
Data Sheet
Address
0x080
0x081
Name
CLKCFG0
SYSREF_ACTRL0
AD9144
Bit No.
7
Bit Name
PD_CLK01
Settings
6
PD_CLK23
5
PD_CLK_DIG
4
PD_SERDES_PCLK
3
PD_CLK_REC
[2:0]
[7:5]
4
RESERVED
RESERVED
PD_SYSREF
3
HYS_ON
2
SYSREF_RISE
0
1
0x082
0x083
0x084
SYSREF_ACTRL1
DACPLLCNTRL
DACPLLSTATUS
[1:0]
HYS_CNTRL1
[7:0]
7
HYS_CNTRL0
RECAL_DACPLL
[6:5]
4
RESERVED
ENABLE_DACPLL
[3:0]
7
RESERVED
DACPLL_
OVERRANGE_H
6
DACPLL_
OVERRANGE_L
5
DACPLL_CAL_
VALID
[4:2]
1
RESERVED
DACPLL_LOCK
0
RESERVED
Description
Power-Down Clock for Dual A. This bit
disables the digital and analog clocks for
Dual A.
Power-Down Clock for Dual B. This bit
disables the digital and analog clocks for
Dual B.
Power-Down Clocks to all DACs. This bit
disables the digital and analog clocks for
both duals. This includes all reference clocks,
PCLK, DAC clocks, and digital clocks.
Serdes PLL Clock Power-Down. This bit
disables the reference clock to the SERDES
PLL, which is needed to have an operational
serial interface.
Clock Receiver Power-Down. This bit powers
down the analog DAC clock receiver block.
With this bit set, clocks are not passed to
internal nets.
Reserved.
Reserved.
Power-Down SYSREF Buffer. This bit powers
down the SYSREF receiver. For Subclass 1
operation to work, this buffer must be enabled.
Hysteresis Enabled. This bit enables the
programmable hysteresis control for the
SYSREF receiver. Using hysteresis gives some
noise resistance, but delays the SYSREF±
edge an amount depending on HYS_CNTRL
and the SYSREF± edge rate. The SYSREF±
KOW is not guaranteed when using hysteresis.
Select DAC Clock Edge to Sample SYSREF.
Use falling edge of DAC clock to sample
SYSREF for alignment
Use rising edge of DAC clock to sample
SYSREF for alignment
Hysteresis Control Bits[9:8]. HYS_CNTRL is a
10-bit thermometer-coded number. Each bit
set adds 10 mV of differential hysteresis to
the SYSREF receiver.
Hysteresis Control Bits[7:0].
Recalibrate DAC PLL. On a rising edge of this bit,
recalibrate the DAC PLL.
Reserved.
Synthesizer Enable. This bit enables and
calibrates the DAC PLL.
Reserved.
DAC PLL High Overrange. This bit indicates
that the DAC PLL hit the upper edge of its
operating band. Recalibrate.
DAC PLL Low Overrange. This bit indicates
that the DAC PLL hit the lower edge of its
operating band. Recalibrate.
DAC PLL Calibration Valid. This bit indicates
that the DAC PLL has been successfully
calibrated.
Reserved.
DAC PLL Lock Bit. This bit is set high by the PLL
when it has achieved lock.
Reserved.
Rev. A | Page 103 of 125
Reset
0x1
Access
R/W
0x1
R/W
0x1
R/W
0x1
R/W
0x1
R/W
0x0
0x0
0x1
R
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
0x0
R/W
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
AD9144
Data Sheet
Address
0x085
Name
DACINTEGERWORD0
Bit No.
[7:0]
Bit Name
B_COUNT
0x087
DACLOOPFILT1
[7:4]
LF_C2_WORD
[3:0]
LF_C1_WORD
[7:4]
LF_R1_WORD
[3:0]
LF_C3_WORD
7
LF_BYPASS_R3
6
LF_BYPASS_R1
5
LF_BYPASS_C2
4
LF_BYPASS_C1
[3:0]
LF_R3_WORD
0x088
0x089
DACLOOPFILT2
DACLOOPFILT3
0x08A
DACCPCNTRL
[7:6]
[5:0]
RESERVED
CP_CURRENT
0x08B
DACLOGENCNTRL
[7:2]
[1:0]
RESERVED
LO_DIV_MODE
Settings
01
10
11
Description
Integer Division Word. This bit controls the
integer feedback divider for the DAC PLL.
Determine the frequency of the DAC clock by
the following equations (see the DAC PLL
Fixed Register Writes section for more
details):
fDAC = fREF/(REF_DIVRATE) × 2 × B_COUNT
fVCO = fREF/(REF_DIVRATE) × 2 × B_COUNT ×
LO_DIV_MODE
Minimum value is 6.
C2 Control Word. Set this control to 0x6 for
optimal performance.
C1 Control Word. Set this control to 0x2 for
optimal performance.
R1 Control Word. Set this control to 0xC for
optimal performance.
C3 Control Word. Set this control to 0x9 for
optimal performance.
Bypass R3 Resistor. When this bit is set,
bypass the R3 capacitor (set to 0 pF) when
R3_WORD is set to 0. Set this control to 0x0 for
optimal performance.
Bypass R1 Resistor. When this bit is set,
bypass the R1 capacitor (set to 0 pF) when
R1_WORD is set to 0. Set this control to 0x0 for
optimal performance.
Bypass C2 Capacitor. When this bit is set,
bypass the C2 capacitor (set to 0 pF) when
C2_WORD is set to 0. Set this control to 0x0 for
optimal performance.
Bypass C1 Capacitor. When this bit is set,
bypass the C1 capacitor (set to 0 pF) when
C1_WORD is set to 0. Set this control to 0x0 for
optimal performance.
R3 Control Word. Set this control to 0xE for
optimal performance.
Reserved.
Charge Pump Current Control. Set this control
to 0x12 for optimal performance.
Reserved.
This range controls the RF clock divider
between the VCO and DAC clock rates. The
options are 4×, 8×, or 16× division. Choose
the LO_DIV_MODE so that 6 GHz < fVCO < 12
GHz (see the DAC PLL Fixed Register Writes
section for more details):
DAC clock = VCO/4
DAC clock = VCO/8
DAC clock = VCO/16
Rev. A | Page 104 of 125
Reset
0x8
Access
R/W
0x8
R/W
0x8
R/W
0x8
R/W
0x8
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x8
R/W
0x0
0x20
R
R/W
0x0
0x2
R
R/W
Data Sheet
Address
0x08C
Name
DACLDOCNTRL1
AD9144
Bit No.
[7:3]
[2:0]
Bit Name
RESERVED
REF_DIV_MODE
Settings
000
001
010
011
100
0x08D
DACLDOCNTRL2
[7:0]
DAC_LDO
0x0E2
CAL_CTRL_GLOBAL
[7:2]
1
RESERVED
CAL_START_AVG
0
CAL_EN_AVG
[7:4]
RESERVED
3
CAL_CLK_EN
1
0x0E7
CAL_CLKDIV
1
0
0x0E8
CAL_PAGE
[2:0]
[7:4]
[3:0]
RESERVED
RESERVED
CAL_PAGE
0x0E9
CAL_CTRL
7
CAL_FIN
6
CAL_ACTIVE
1
1
Description
Reserved.
Reference Clock Division Ratio. This field
controls the amount of division that is done
to the input clock at the CLK+/CLK− pins
before it is presented to the PLL as a reference
clock. The reference clock frequency must be
between 35 MHz and 80 MHz, but the
CLK+/CLK− input frequency can range from
35 MHz to 1 GHz. The user sets this division
to achieve a 35 MHz to 80 MHz PLL reference
frequency. For more details see the DAC PLL
Fixed Register Writes section.
1
2
4
8
16
DAC PLL LDO setting. This register must be
written to 0x7B for optimal performance.
Reserved.
Averaged Calibration Start. On rising edge,
calibrate the DACs. Only use if calibrating all
DACs.
Averaged Calibration Enable. Set prior to
starting calibration with CAL_START_AVG.
While this bit is set, calibration can be
performed, and the results are applied.
Enable averaged calibration
Must write the default value for proper
operation.
Enable Self Calibration Clock.
Enable calibration clock
Disable calibration clock
Reserved.
Reserved.
DAC Calibration Paging. Selects which of the
DACs are being accessed for calibration or
calibration readback. This paging affects
Register 0x0E9 and Register 0x0ED.
Calibration: any number of DACs can be
accessed simultaneously to write and
calibrate. Write a 1 to Bit x to include DAC x.
Readback: only one DAC at a time can be
accessed when reading back CAL_CTRL
(Register 0x0E9). Write a 1 to Bit x to read
from DAC x (the other bits must be 0).
Calibration finished. This bit is high when the
calibration has completed. If the calibration
completes and either CAL_ERRHI or CAL_
ERRLO is high, then the calibration cannot be
considered valid and are considered a timeout
event.
Calibration ran and is finished
Calibration Active. This bit is high while the
calibration is in progress.
Calibration is running
Rev. A | Page 105 of 125
Reset
0x0
0x1
Access
R
R/W
0x2B
R/W
0x0
0x0
R
R/W
0x0
R/W
0x3
R/W
0x0
R/W
0x0
0x0
0xF
R
R
R/W
0x0
R
0x0
R
AD9144
Address
Name
Data Sheet
Bit No.
5
Bit Name
CAL_ERRHI
4
CAL_ERRLO
[3:2]
1
RESERVED
CAL_START
Settings
1
1
0
1
0
CAL_EN
0
1
0x0ED
CAL_INIT
[7:0]
CAL_INIT
0x110
DATA_FORMAT
7
BINARY_FORMAT
0
1
0x111
DATAPATH_CTRL
[6:0]
7
RESERVED
INVSINC_ENABLE
1
0
6
5
RESERVED
DIG_GAIN_ENABLE
1
0
4
PHASE_ADJ_
ENABLE
1
0
[3:2]
MODULATION_TYPE
00
01
10
11
1
SEL_SIDEBAND
0
I_TO_Q
Description
SAR Data Error: Too High. This bit is set at the
end of a calibration cycle if any of the calibration DACs has overranged to the high side.
This typically means that the algorithm adjusts
the calibration preset of the calibration DACs
and runs another cycle.
Data saturated high
SAR Data Error: Too Low. This bit is set at the
end of a calibration cycle if any of the calibration DACs has overranged to the low side.
This typically means that the algorithm adjusts
the calibration preset of the calibration DACs
and runs another cycle.
Data saturated low
Reserved.
Calibration Start. The rising edge of this bit
kicks off a calibration sequence for the DACs
that have been selected in the CAL_INDX
register.
Normal operation
Start calibration state machine
Calibration Enable. Enable the calibration
DAC of the converter. Enable to calibration
engine and machines. Prepare for a calibration
start. For calibration coefficients to be applied
to the calibrated DACs, this bit must be high.
Do not use calibration DACs
Use calibration DACs
Initialize Calibration. Must be written to 0xA2
before starting calibration or averaged
calibration.
Binary or Twos Complementary Format on
the Data Bus.
Input data is twos complement
Input data is offset binary
Reserved.
Enable Inverse Sinc Filter.
Enable inverse sinc filter
Disable inverse sinc filter
Reserved.
Enable Digital Gain.
Enable digital gain function
Disable digital gain function
Enable Phase Compensation.
Enable phase adjust compensation
Disable phase adjust compensation
Selects Type Of Modulation Operation.
No modulation
Fine modulation (uses FTW)
fS/4 coarse modulation
fS/8 coarse modulation
Spectrum Inversion Control. Can only be
used with fine modulation. This causes the
negative sideband to be selected and is
equivalent to changing the sign of FTW.
Send I Data into Q DAC datapath. Occurs at
the end of the digital datapath prior to
entering DACs.
Rev. A | Page 106 of 125
Reset
0x0
Access
R
0x0
R
0x0
0x0
R
R/W
0x0
R/W
0xA6
R/W
0x0
R/W
0x0
0x1
R
R/W
0x0
0x1
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Data Sheet
Address
0x112
Name
INTERP_MODE
AD9144
Bit No.
[7:3]
[2:0]
Bit Name
RESERVED
INTERP_MODE
Settings
000
001
011
100
0x113
NCO_FTW_UPDATE
[7:2]
1
RESERVED
FTW_UPDATE_ACK
0
FTW_UPDATE_REQ
0x114
0x115
0x116
0x117
0x118
0x119
0x11A
FTW0
FTW1
FTW2
FTW3
FTW4
FTW5
NCO_PHASE_
OFFSET0
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
FTW[7:0]
FTW[15:8]
FTW[23:16]
FTW[31:24]
FTW[39:32]
FTW[47:40]
NCO_PHASE_
OFFSET[7:0]
0x11B
[7:0]
0x11C
NCO_PHASE_
OFFSET1
PHASE_ADJ0
[7:0]
NCO_PHASE_
OFFSET[15:8]
PHASE_ADJ[7:0]
0x11D
PHASE_ADJ1
0x11F
TXEN_SM_0
[7:5]
[4:0]
[7:6]
RESERVED
PHASE_ADJ[12:8]
FALL_COUNTERS
[5:4]
RISE_COUNTERS
3
2
RESERVED
PROTECT_OUT_
INVERT
0
1
[1:0]
RESERVED
Description
Reserved.
Interpolation Mode.
1× mode
2× mode
4× mode
8× mode
Reserved.
Frequency tuning word update acknowledge.
This readback is high when an FTW has been
updated.
Frequency tuning word update request from
SPI. Unlike most registers, those relating to
fine NCO modulation (Register 0x114 to
Register 0x11B) are not updated immediately
upon writing to them. Once the desired FTW
and phase offset values are written, set this
bit. These registers update on the rising edge
of this bit. It is only after this update that the
internal state matches Register 0x114 to
Register 0x11B. Confirmation that this
update has occurred can be made by reading
back bit 1 of this register and ensuring it is
set high for the update acknowledge.
NCO Frequency Tuning Word.
NCO Frequency Tuning Word.
NCO Frequency Tuning Word.
NCO Frequency Tuning Word.
NCO Frequency Tuning Word.
NCO Frequency Tuning Word.
8 LSBs of NCO Phase Offset.
NCO_PHASE_OFFSET changes the phase of
both I and Q data, and is only functional
when using NCO fine modulation. It is a 16bit twos complement number ranging from
−180 to+180 degrees in steps of .0055°.
8 MSBs of NCO Phase Offset.
Reset
0x0
0x1
Access
R
R/W
0x0
0x0
R
R
0x0
R/W
0x0
0x0
0x0
0x0
0x0
0x10
0x0
R/W
R/W
R/W
R/W
R/W
R/W
R/W
0x0
R/W
8 LSBs of Phase Compensation Word. Phase
compensation changes the phase between
the I and Q data. PHASE_ADJ is a 13-bit twos
complement value. The control ranges from
−14° to +14° with 0.0035° resolution steps.
Reserved.
5 MSBs of Phase Compensation Word.
Fall Counters. The number of counters to use
to delay TX_PROTECT fall from TXENx falling
edge. Must be set to 1 or 2.
Rise Counters. The number of counters to
use to delay TX_PROTECT rise from TXENx
rising edge.
Reserved.
PROTECT_OUTx Invert.
PROTECT_OUTx is high when output is valid.
Suitable for enabling downstream
components during transmission
PROTECT_OUTx is high when output is
invalid. Suitable for disabling downstream
components when not transmitting
Must write the default value for proper
operation.
0x0
R/W
0x0
0x0
0x2
R
R/W
R/W
0x0
R/W
0x0
0x0
R
R/W
0x3
R/W
Rev. A | Page 107 of 125
AD9144
Data Sheet
Address
0x121
Name
TXEN_RISE_
COUNT_0
Bit No.
[7:0]
Bit Name
RISE_COUNT_0
0x122
TXEN_RISE_
COUNT_1
[7:0]
RISE_COUNT_1
0x123
TXEN_FALL_
COUNT_0
[7:0]
FALL_COUNT_0
0x124
TXEN_FALL_
COUNT_1
[7:0]
FALL_COUNT_1
0x12D
DEVICE_CONFIG_
REG_0
DIE_TEMP_CTRL0
[7:0]
DEVICE_CONFIG_0
[7:1]
RESERVED
0
AUXADC_ENABLE
0x12F
Settings
0
1
0x132
0x133
0x134
DIE_TEMP0
DIE_TEMP1
DIE_TEMP_UPDATE
[7:0]
[7:0]
[7:1]
0
0x135
DC_OFFSET_CTRL
[7:1]
0
DIE_TEMP[7:0]
DIE_TEMP[15:8]
RESERVED
DIE_TEMP_
UPDATE
RESERVED
DC_OFFSET_ON
0x136
IPATH_DC_OFFSET_
1PART0
[7:0]
LSB_OFFSET_I[7:0]
0x137
IPATH_DC_OFFSET_
1PART1
[7:0]
LSB_OFFSET_I[15:8]
0x138
QPATH_DC_OFFSET_
1PART0
[7:0]
LSB_OFFSET_
Q[7:0]
0x139
QPATH_DC_OFFSET_
1PART1
[7:0]
LSB_OFFSET_
Q[15:8]
0x13A
IPATH_DC_OFFSET_
2PART
[7:5]
RESERVED
[4:0]
SIXTEENTH_
OFFSET_I
[7:5]
RESERVED
[4:0]
SIXTEENTH_
OFFSET_Q
1
x
0x13B
QPATH_DC_OFFSET_
2PART
x
0x13C
IDAC_DIG_GAIN0
[7:0]
IDAC_DIG_
GAIN[7:0]
0x13D
IDAC_DIG_GAIN1
[7:4]
[3:0]
RESERVED
IDAC_DIG_
GAIN[11:8]
Description
First counter used to delay TX_PROTECT rise
from TXENx rising edge. Delays by 32 ×
RISE_COUNT_0 DAC clock cycles.
Second counter used to delay TX_PROTECT
rise from TXENx rising edge. Delays by 32 ×
RISE_COUNT_1 DAC clock cycles.
First counter used to delay TX_PROTECT fall
from TXENx falling edge. Delays by 32 ×
FALL_COUNT_0 DAC clock cycles. Must be
set to a minimum of 0x12.
Second counter used to delay TX_PROTECT
fall from TXENx falling edge. Delays by 32 ×
FALL_COUNT_1 DAC clock cycles.
Must be set to 0x8B for proper digital
datapath configuration.
Must write the default value for proper
operation.
Enables the AUX ADC Block.
AUX ADC disable
AUX ADC enable
Aux ADC Readback Value.
Aux ADC Readback Value.
Reserved.
Die Temperature Update. On a rising edge, a
new temperature code is generated.
Reserved.
DC Offset On.
Enables dc offset module
8 LSBs of IPath DC Offset. LSB_OFFSET_I is a
16-bit twos complement number that is
added to incoming data.
8 MSBs of IPath DC Offset. LSB_OFFSET_I is a
16-bit twos complement number that is
added to incoming I data.
8 LSBs of QPath DC Offset. LSB_OFFSET_Q is a
16-bit twos complement number that is
added to incoming Q data.
8 MSBs of QPath DC Offset. LSB_OFFSET_Q is
a 16-bit twos complement number that is
added to incoming Q data.
Reserved.
SIXTEENTH_OFFSET_I is a 5-bit twos
complement number in 16ths of an LSB that
is added to incoming I data.
x/16 LSB DC offset
Reserved.
SIXTEENTH_OFFSET_Q is a 5-bit twos
complement number in 16ths of an LSB that
is added to incoming Q data.
x/16 LSB DC offset
8 LSBs of I DAC Digital Gain. IDAC_DIG_GAIN
is the digital gain of the IDAC. The digital
gain is a multiplier from 0 to 4095/2048 in
steps of 1/2048.
Reserved.
4 MSBs of I DAC Digital Gain
Rev. A | Page 108 of 125
Reset
0xF
Access
R/W
0x0
R/W
0xFF
R/W
0xFF
R/W
0x46
R/W
0x10
R/W
0x0
R/W
0x0
0x0
0x0
0x0
R
R
R
R/W
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R
0x0
R/W
0x0
R
0x0
R/W
0xEA
R/W
0x0
0xA
R
R/W
Data Sheet
AD9144
Address
0x13E
Name
QDAC_DIG_GAIN0
Bit No.
[7:0]
Bit Name
QDAC_DIG_
GAIN[7:0]
Settings
0x13F
QDAC_DIG_GAIN1
[7:4]
[3:0]
0x140
GAIN_RAMP_UP_
STEP0
[7:0]
RESERVED
QDAC_DIG_
GAIN[11:8]
GAIN_RAMP_UP_
STEP[7:0]
0x0
0xFFF
0x141
0x142
GAIN_RAMP_UP_
STEP1
GAIN_RAMP_DOWN_
STEP0
[7:4]
RESERVED
[3:0]
GAIN_RAMP_UP_
STEP[11:8]
GAIN_RAMP_
DOWN_STEP[7:0]
[7:0]
0
0xFFF
0x143
0x146
0x147
GAIN_RAMP_
DOWN_STEP1
DEVICE_CONFIG_
REG_1
BSM_STAT
[7:4]
RESERVED
[3:0]
[7:0]
GAIN_RAMP_
DOWN_STEP[11:8]
DEVICE_CONFIG_1
[7:6]
SOFTBLANKRB
00
01
10
11
0x14B
PRBS
[5:0]
7
RESERVED
PRBS_GOOD_Q
0
1
6
PRBS_GOOD_I
0
1
[5:3]
2
RESERVED
PRBS_MODE
0
1
1
PRBS_RESET
0
1
0
PRBS_EN
0
1
0x14C
0x14D
0x1B0
PRBS_ERROR_I
PRBS_ERROR_Q
DACPLLT0
[7:0]
[7:0]
[7:0]
PRBS_COUNT_I
PRBS_COUNT_Q
DAC_PLL_PWR
Description
8 LSBs of Q DAC Digital Gain.
QDAC_DIG_GAIN is the digital gain of the
QDAC. The digital gain is a multiplier from 0
to 4095/2048 in steps of 1/2048.
Reserved.
4 MSBs of Q DAC Digital Gain.
Reset
0xEA
Access
R/W
0x0
0xA
R
R/W
8 LSBs of Gain Ramp Up Step.
GAIN_RAMP_UP_STEP controls the amplitude
step size of the BSM’s ramping feature when
the gain is being ramped to its assigned value.
Smallest ramp up step size
Largest ramp up step size
Reserved.
0x4
R/W
0x0
R
0x0
R/W
0x9
R/W
0x0
R
0x0
R/W
0x0
R/W
0x0
R
0x0
0x0
R
R
0x0
R
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
0x0
0xFA
R
R
R/W
4 MSBs of Gain Ramp Up Step. See
Register 0x140 for description.
8 LSBs of Gain Ramp Down Step.
GAIN_RAMP_DOWN_STEP controls the
amplitude step size of the BSM’s ramping
feature when the gain is being ramped to zero.
Smallest ramp down step size
Largest ramp down step size
Reserved.
4 MSBs of Gain Ramp Down Step. See
Register 0x142 for description.
Must be set to 0x01 for proper digital
datapath configuration.
Blanking State.
Data is fully blanked
Ramping from data process to full blanking
Ramping from fully blanked to data process
Data is being processed
Reserved.
Good Data Indicator Imaginary Channel.
Incorrect sequence detected
Correct PRBS sequence detected
Good Data Indicator Real Channel.
Incorrect sequence detected
Correct PRBS sequence detected
Reserved.
Polynomial Select
7-bit: x7 + x6 + 1
15-bit: x15 + x14 + 1
Reset Error Counters.
Normal operation
Reset counters
Enable PRBS Checker.
Disable
Enable
Error Count Value Real Channel.
Error Count Value Imaginary Channel.
DAC PLL PD settings. This register must be
written to 0x00 for optimal performance.
Rev. A | Page 109 of 125
AD9144
Address
0x1B5
Name
DACPLLT5
Data Sheet
Bit No.
[7:4]
Bit Name
RESERVED
[3:0]
VCO_VAR
0x1B9
DACPLLT9
[7:0]
DAC_PLL_CP1
0x1BB
DACPLLTB
[7:5]
[4:3]
RESERVED
VCO_BIAS_TCF
[2:0]
VCO_BIAS_REF
Settings
0x1BC
DACPLLTC
[7:0]
DAC_PLL_VCO_
CTRL
0x1BE
DACPLLTE
[7:0]
DAC_PLL_VCO_
PWR
0x1BF
DACPLLTF
[7:0]
DAC_PLL_VCOCAL
0x1C0
DACPLLT10
[7:0]
0x1C1
DACPLLT11
[7:0]
DAC_PLL_LOCK_
CNTR
DAC_PLL_CP2
0x1C4
DACPLLT17
[7:0]
DAC_PLL_VAR1
0x1C5
DACPLLT18
[7:0]
DAC_PLL_VAR2
0x200
MASTER_PD
[7:1]
0
RESERVED
SPI_PD_MASTER
0x201
PHY_PD
[7:0]
SPI_PD_PHY
0x203
GENERIC_PD
[7:2]
1
RESERVED
SPI_SYNC1_PD
0x206
CDR_RESET
Description
Must write the default value for proper
operation.
Varactor KVO Setting. See Table 83 for
optimal settings based on the fVCO being used.
DAC PLL Charge Pump settings. This register
must be written to 0x24 for optimal
performance.
Reserved.
Temperature Coefficient for VCO Bias. See
Table 83 for optimal settings based on the
fVCO being used.
VCO Bias Control. See Table 83 for optimal
settings based on the fVCO being used.
DAC PLL VCO control settings. This register
must be written to 0x0D for optimal
performance.
DAC PLL VCO power control settings. This
register must be written to 0x02 for optimal
performance.
DAC PLL VCO calibration settings. This
register must be written to 0x8E for optimal
performance.
This register must be written to 0x2A for
optimal performance.
This register must be written to0x2A for
optimal performance.
DAC PLL Varactor setting. Must be set to
0x7E for proper DAC PLL configuration.
DAC PLL Varactor setting. See Table 83 for
optimal settings based on the fVCO being used.
Reserved.
Power Down the Entire JESD Receiver Analog
(All Eight Channels Plus Bias).
SPI Override to Power Down the Individual
PHYs.
Set Bit x to power down the corresponding
SERDINx± PHY
Reserved.
Power down LVDS buffer for SYNCOUT0±.
0x232
CDR_OPERATING_
MODE_REG_0
DEVICE_CONFIG_
REG_3
Access
R/W
0x3
R/W
0x34
R/W
0x0
0x1
R
R/W
0x4
R/W
0x00
R/W
0x00
R/W
0x8D
R/W
0x2E
R/W
0x24
R/W
0x33
R/W
0x08
R/W
0x0
0x1
R
R/W
0x0
R/W
0x0
0x0
R
R/W
0
SPI_SYNC2_PD
Power down LVDS buffer for SYNCOUT1±.
0x0
R/W
[7:1]
0
RESERVED
SPI_CDR_RESETN
Reserved.
Resets the Digital Control Logic for All PHYs.
Hold CDR in reset
Enable CDR
Reserved.
0x0
0x1
R
R/W
0x0
R
0x1
R/W
0x2
R/W
0x0
R/W
0x0
0x0
R
R/W
0
1
0x230
Reset
0x8
[7:6]
RESERVED
5
ENHALFRATE
[4:2]
RESERVED
1
CDR_OVERSAMP
0
[7:0]
RESERVED
DEVICE_CONFIG_3
Enables Half-Rate CDR Operation. Set to 1
when 5.75 Gbps ≤ lane rate ≤ 10.64.
Must write the default value for proper
operation.
Enables Oversampling of the Input Data. Set
to 1 when 1.44 Gbps ≤ lane rate ≤ 2.76 Gbps.
Reserved.
Must be set to 0xFF for proper JESD interface
configuration.
Rev. A | Page 110 of 125
Data Sheet
Address
0x268
Name
EQ_BIAS_REG
AD9144
Bit No.
[7:6]
Bit Name
EQ_POWER_
MODE
Settings
00
01
0x280
0x281
SERDESPLL_
ENABLE_CNTRL
PLL_STATUS
[5:0]
RESERVED
[7:3]
RESERVED
2
RECAL_SERDESPLL
1
0
RESERVED
ENABLE_
SERDESPLL
RESERVED
SERDES_PLL_
OVERRANGE_H
[7:6]
5
4
SERDES_PLL_
OVERRANGE_L
3
SERDES_PLL_CAL_
VALID_RB
[2:1]
0
0x284
LOOP_FILTER_1
[7:0]
RESERVED
SERDES_PLL_
LOCK_RB
LOOP_FILTER_1
0x285
LOOP_FILTER_2
[7:0]
LOOP_FILTER_2
0x286
LOOP_FILTER_3
[7:0]
LOOP_FILTER_3
0x287
SERDES_PLL_CP1
[7:0]
SERDES_PLL_CP1
0x289
REF_CLK_DIVIDER_
LDO
[7:3]
RESERVED
2
DEVICE_CONFIG_4
[1:0]
SERDES_PLL_DIV_
MODE
Recalibrate SERDES PLL. On a rising edge,
recalibrate the SERDES PLL.
Reserved.
Enable the SERDES PLL. Setting this bit
enables and calibrates the SERDES PLL.
Reserved.
SERDES PLL High Overrange. This bit
indicates that the SERDES PLL hit the lower
edge of its operating band. Recalibrate.
SERDES PLL Low Overrange. This bit
indicates that the SERDES PLL hit the lower
edge of its operating band. Recalibrate.
SERDES PLL Calibration Valid. This bit
indicates that the SERDES PLL has been
successfully calibrated.
Reserved.
SERDES PLL Lock. This bit is set high by the PLL
when it has achieved lock.
SERDES PLL loop filter setting. This register
must be written to 0x62 for optimal
performance.
SERDES PLL loop filter setting. This register
must be written to 0xC9 for optimal
performance.
SERDES PLL loop filter setting. This register
must be written to 0x0E for optimal
performance.
SERDES PLL charge pump setting. This
register must be written to 0x12 for optimal
performance.
Reserved.
00
01
10
0x28A
VCO_LDO
[7:0]
SERDES_PLL_
VCO_LDO
0x28B
SERDES_PLL_PD1
[7:0]
SERDES_PLL_PD1
0x290
SERDESPLL_VAR1
[7:0]
SERDES_PLL_VAR1
Description
Control the Equalizer Power/Insertion Loss
Capability.
Normal mode
Low power mode
Must write the default value for proper
operation.
Reserved.
Must be set to 1 for proper SERDES PLL
configuration.
SERDES PLL Reference Clock Division Factor.
This field controls the division of the SERDES
PLL reference clock before it is fed into the
SERDES PLL Phase Frequency Detector (PFD).
It must be set so fREF/DivFactor is between
35 MHz and 80 MHz.
Divide by 4 for 5.75 Gbps to 10.64 Gbps lane
rate
Divide by 2 for 2.88 Gbps to 5.52 Gbps lane rate
Divide by 1 for 1.44 Gbps to 2.76 Gbps lane rate
SERDES PLL VCO LDO setting. This register
must be written to 0x7B for optimal
performance.
SERDES PLL PD setting. This register must be
written to 0x00 for optimal performance.
SERDES PLL Varactor setting. This register
must be written to 0x89 for optimal
performance.
Rev. A | Page 111 of 125
Reset
0x1
Access
R/W
0x22
R/W
0x0
R
0x0
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R
0x0
R
0x0
R
0x0
0x0
R
R
0x77
R/W
0x87
R/W
0x08
R/W
0x3F
R/W
0x0
R
0x0
R/W
0x0
R/W
0x2B
R/W
0x7F
R/W
0x83
R/W
AD9144
Data Sheet
Address
0x294
Name
SERDES_PLL_CP2
Bit No.
[7:0]
Bit Name
SERDES_PLL_CP2
0x296
SERDESPLL_VCO1
[7:0]
0x297
SERDESPLL_VCO2
[7:0]
0x299
SERDES_PLL_PD2
[7:0]
SERDES_PLL_
VCO1
SERDES_PLL_
VCO2
SERDES_PLL_PD2
0x29A
SERDESPLL_VAR2
[7:0]
SERDES_PLL_VAR2
0x29C
SERDES_PLL_CP3
[7:0]
SERDES_PLL_CP3
0x29F
SERDESPLL_VAR3
[7:0]
SERDES_PLL_VAR3
0x2A0
SERDESPLL_VAR4
[7:0]
SERDES_PLL_VAR4
0x2A4
DEVICE_CONFIG_
REG_8
SYNCOUTB_SWING
[7:0]
DEVICE_CONFIG_8
[7:1]
0
RESERVED
SYNCOUTB_
SWING_MD
0x2A5
Settings
0
1
0x2A7
TERM_BLK1_
CTRLREG0
[7:1]
0
RESERVED
RCAL_TERMBLK1
0x2AA
DEVICE_CONFIG_
REG_9
DEVICE_CONFIG_
REG_10
TERM_BLK2_
CTRLREG0
[7:0]
[7:1]
DEVICE_CONFIG_
9
DEVICE_CONFIG_
10
RESERVED
0
RCAL_TERMBLK2
[7:0]
DEVICE_CONFIG_
11
DEVICE_CONFIG_
12
RESERVED
CHECKSUM_MODE
0x2AB
0x2AE
0x2B1
0x2B2
0x300
DEVICE_CONFIG_
REG_11
DEVICE_CONFIG_
REG_12
GENERAL_JRX_
CTRL_0
[7:0]
[7:0]
7
6
0
1
[5:4]
RESERVED
Description
SERDES PLL Charge Pump setting. This
register must be set to 0x24 for optimal
performance.
SERDES PLL VCO setting. This register must
be set to 0x03 for optimal performance.
SERDES PLL VCO setting. This register must
be set to 0x0D for optimal performance.
SERDES PLL PD setting. This register must be
set to 0x02 for optimal performance.
SERDES PLL Varactor setting. This register
must be set to 0x8E for optimal performance.
SERDES PLL Charge Pump setting. Must be
set to 0x2A for proper SERDES PLL
configuration.
SERDES PLL varactor setting. Must be set to
0x78 for proper SERDES PLL configuration.
SERDES PLL varactor setting. This register
must be set to 0x06 for optimal performance.
Must be set to 0xFF for proper clock
configuration.
Reserved.
SYNCOUTx± Swing Mode. Sets the output
differential swing mode for the SYNCOUTx±
pins. See Table 8 for details.
Normal Swing Mode
High Swing Mode
Reserved.
Termination Calibration. The rising edge of
this bit calibrates PHY0, PHY1, PHY6, and PHY7
terminations to 50 Ω.
Must be set to 0xB7 for proper JESD interface
termination configuration.
Must be set to 0x87 for proper JESD interface
termination configuration.
Reserved.
Terminal Calibration. The rising edge of this
bit calibrates PHY2, PHY3, PHY4 and PHY5
terminations to 50 Ω.
Must be set to 0xB7 for proper JESD interface
termination configuration.
Must be set to 0x87 for proper JESD interface
termination configuration.
Reserved.
Checksum Mode. This bit controls the locally
generated JESD204B link parameter checksum
method. The value is stored in the FCMP
registers (Register 0x40E, Register 0x416,
Register 0x41E, Register 0x426, Register
0x42E, Register 0x436, Register 0x43E, and
Register 0x446).
Checksum is calculated by summing the
individual fields in the link configuration
table as defined in Section 8.3, Table 20 of
the JESD204B standard
Checksum is calculated by summing the registers containing the packed link configuration
fields (Σ[0x450:0x45A] modulo 256).
Reserved.
Rev. A | Page 112 of 125
Reset
0xB0
Access
R/W
0x0C
R/W
0x00
R/W
0x00
R/W
0xFE
R/W
0x17
R/W
0x33
R/W
0x08
R/W
0x4B
R/W
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0xC3
R/W
0x93
R/W
0x0
R
0x0
R/W
0xC3
R/W
0x93
R/W
0x0
0x0
R
R/W
0x0
R
Data Sheet
Address
Name
AD9144
Bit No.
3
Bit Name
LINK_MODE
Settings
0
1
2
LINK_PAGE
0
1
[1:0]
LINK_EN
0b00
0b01
0b10
0b11
0x301
GENERAL_JRX_CTRL_1
[7:3]
[2:0]
RESERVED
SUBCLASSV_
LOCAL
000
001
0x302
DYN_LINK_LATENCY_0
[7:5]
[4:0]
RESERVED
DYN_LINK_
LATENCY_0
0x303
DYN_LINK_LATENCY_1
[7:5]
[4:0]
RESERVED
DYN_LINK_
LATENCY_1
0x304
LMFC_DELAY_0
[7:5]
[4:0]
RESERVED
LMFC_DELAY_0
0x305
LMFC_DELAY_1
[7:5]
[4:0]
RESERVED
LMFC_DELAY_1
0x306
LMFC_VAR_0
[7:5]
[4:0]
RESERVED
LMFC_VAR_0
0x307
LMFC_VAR_1
[7:5]
[4:0]
RESERVED
LMFC_VAR_1
Description
Link Mode. This register selects either singlelink or dual-link mode.
Single-link mode
Dual-link mode
Link Paging. Selects which link’s register map
is used. This paging affects Registers 0x401
to 0x47E.
Use Link 0 register map
Use Link 1 register map
Link Enable. These bits bring up the JESD204B
receiver digital circuitry: Bit 0 for Link 0 and
Bit 1 for Link 1. Enable the link only after the
following has occurred: all JESD204B parameters are set, the DAC PLL is enabled and
locked (Register 0x084[1] = 1), and the
JESD204B PHY is enabled (Register 0x200 =
0x00) and calibrated (Register 0x281[2] = 0).
Disable both JESD Link 1 and JESD Link 0
Disable JESD Link 1, enable JESD Link 0
Enable JESD Link 1, disable JESD Link 0
Enable both JESD Link 1 and JESD Link 0
Reserved.
JESD204B Subclass.
Subclass 0
Subclass 1
Reserved.
Dynamic Link Latency: Link 0. Latency
between the LMFCRx for link 0 and the last
arriving LMFC boundary in units of PCLK
cycles. See the Deterministic Latency section.
Reserved.
Dynamic Link Latency: Link 1. Latency
between the LMFCRx for link 1 and the last
arriving LMFC boundary in units of PCLK
cycles. See the Deterministic Latency section.
Reserved.
LMFC Delay: Link 0 Delay from the LMFC to
LMFCRx for Link 0. In units of frame clock
cycles for subclass 1 and PCLK cycles for
subclass 0. See the Deterministic Latency
section.
Reserved.
LMFC Delay: Link 1. Delay from the LMFC to
LMFCRx for Link 1. In units of frame clock
cycles for subclass 1 and PCLK cycles for
subclass 0. See the Deterministic Latency
section.
Reserved.
Variable Delay Buffer: Link 0. Sets when data is
read from a buffer to be consistent across links
and power cycles. In units of PCLK cycles. See
the Deterministic Latency section.
This setting must not be more than 10.
Reserved.
Variable Delay Buffer: Link 1. Sets when data is
read from a buffer to be consistent across links
and power cycles. In units of PCLK cycles. See
the Deterministic Latency section.
This setting must not be more than 10.
Rev. A | Page 113 of 125
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
0x1
R
R/W
0x0
0x0
R
R
0x0
0x0
R
R
0x0
0x0
R
R/W
0x0
0x0
R
R/W
0x0
0x6
R
R/W
0x0
0x6
R
R/W
AD9144
Address
0x308
Name
XBAR_LN_0_1
Data Sheet
Bit No.
[7:6]
[5:3]
Bit Name
RESERVED
LOGICAL_LANE1_
SRC
[2:0]
LOGICAL_LANE0_
SRC
[7:6]
[5:3]
RESERVED
LOGICAL_LANE3_
SRC
[2:0]
LOGICAL_LANE2_
SRC
[7:6]
[5:3]
RESERVED
LOGICAL_LANE5_
SRC
[2:0]
LOGICAL_LANE4_
SRC
[7:6]
[5:3]
RESERVED
LOGICAL_LANE7_
SRC
[2:0]
LOGICAL_LANE6_
SRC
Settings
x
x
0x309
XBAR_LN_2_3
x
x
0x30A
XBAR_LN_4_5
x
x
0x30B
XBAR_LN_6_7
x
x
0x30C
FIFO_STATUS_REG_0
[7:0]
LANE_FIFO_FULL
0x30D
FIFO_STATUS_REG_1
[7:0]
LANE_FIFO_EMPTY
0x312
SYNCB_GEN_1
[7:6]
[5:4]
RESERVED
SYNCB_ERR _DUR
0
1
2
0x314
SERDES_SPI_REG
[3:0]
[7:0]
0x315
PHY_PRBS_TEST_EN
[7:0]
RESERVED
SERDES_SPI_
CONFIG
PHY_TEST_EN
Description
Reserved.
Logical Lane 1 Source. Selects a physical lane
to be mapped onto Logical Lane 1.
Data is from SERDINx
Logical Lane 0 Source. Selects a physical lane
to be mapped onto Logical Lane 0.
Data is from SERDINx
Reserved.
Logical Lane 3 Source. Selects a physical lane
to be mapped onto Logical Lane 3.
Data is from SERDINx
Logical Lane 2 source. Selects a physical lane
to be mapped onto Logical Lane 2.
Data is from SERDINx
Reserved.
Logical Lane 5 Source. Selects a physical lane
to be mapped onto Logical Lane 5.
Data is from SERDINx
Logical Lane 4 Source. Selects a physical lane
to be mapped onto Logical Lane 4.
Data is from SERDINx
Reserved.
Logical Lane 7 Source. Selects a physical lane
to be mapped onto Logical Lane 7.
Data is from SERDINx
Logical Lane 6 Source. Selects a physical lane
to be mapped onto Logical Lane 6.
Data is from SERDINx
FIFO Full Flags for Each Logical Lane. A full
FIFO indicates an error in the JESD204B
configuration or with a system clock.
If the FIFO for Lane x is full, Bit x in this
register will be high.
FIFO Empty Flags for Each Logical Lane. An
empty FIFO indicates an error in the JESD204B
configuration or with a system clock.
If the FIFO for Logical Lane x is empty, Bit x in
this register will be high.
Reserved.
Duration of SYNCOUTx± Low for Error. The
duration applies to both SYNCOUT0 and
SYNCOUT1. A sync error is asserted at the
end of a multiframe whenever one or more
disparity, not in table or unexpected control
character errors are encountered.
½ PCLK cycle
1 PCLK cycle
2 PCLK cycles
Reserved.
SERDES SPI Configuration. Must be written to
0x01 as part of the Physical Layer setup step.
PHY Test Enable. Enables the PHY BER test.
Set Bit x to enable the PHY test for Lane x.
Rev. A | Page 114 of 125
Reset
0x0
0x1
Access
R
R/W
0x0
R/W
0x0
0x3
R
R/W
0x2
R/W
0x0
0x5
R
R/W
0x4
R/W
0x0
0x7
R
R/W
0x6
R/W
0x0
R
0x0
R
0x0
R/W
0x0
0x0
R/W
R/W
0x0
R/W
Data Sheet
Address
0x316
Name
PHY_PRBS_TEST_CTRL
AD9144
Bit No.
7
[6:4]
Bit Name
RESERVED
PHY_SRC_ERR_CNT
[3:2]
PHY_PRBS_PAT_SEL
Settings
x
00
01
10
1
PHY_TEST_START
0
1
0
PHY_TEST_RESET
0
1
0x317
0x318
0x319
0x31A
0x31B
0x31C
0x31D
0x32C
PHY_PRBS_TEST_
THRESHOLD_LOBITS
PHY_PRBS_TEST_
THRESHOLD_
MIDBITS
PHY_PRBS_TEST_
THRESHOLD_HIBITS
PHY_PRBS_TEST_
ERRCNT_LOBITS
[7:0]
PHY_PRBS_TEST_
ERRCNT_MIDBITS
PHY_PRBS_TEST_
ERRCNT_HIBITS
PHY_PRBS_TEST_
STATUS
[7:0]
SHORT_TPL_TEST_0
[7:6]
[5:4]
RESERVED
SHORT_TPL_SP_
SEL
[3:2]
SHORT_TPL_DAC_
SEL
1
SHORT_TPL_TEST_
RESET
[7:0]
[7:0]
[7:0]
[7:0]
[7:0]
PHY_PRBS_
THRESHOLD[7:0]
PHY_PRBS_
THRESHOLD[15:8]
PHY_PRBS_
THRESHOLD[23:16]
PHY_PRBS_ERR_
CNT[7:0]
PHY_PRBS_ERR_
CNT[15:8]
PHY_PRBS_ERR_
CNT[23:16]
PHY_PRBS_PASS
x
x
0
1
0
SHORT_TPL_TEST_
EN
0
1
0x32D
SHORT_TPL_TEST_1
[7:0]
SHORT_TPL_REF_
SP_LSB
Description
Reserved.
PHY Error Count Source. Selects which PHY
errors are being reported in Register 0x31A
to Register 0x31C.
Report Lane x error count
PHY PRBS Pattern Select. Selects the PRBS
pattern for PHY BER test.
PRBS7
PRBS15
PRBS31
PHY PRBS Test Start. Starts and stops the PHY
PRBS test.
Test stopped
Test in progress
PHY PRBS Test Reset. Resets the PHY PRBS
test state machine and error counters.
Enable PHY PRBS test state machine
Hold PHY PRBS test state machine in reset
8 LSBs of PHY PRBS Error Threshold.
Reset
0x0
0x0
Access
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
8 ISBs of PHY PRBS Error Threshold.
0x0
R/W
8 MSBs of PHY PRBS Error Threshold.
0x0
R/W
8 LSBs of PHY PRBS Error Count.
Reported PHY BERT error count from lane
selected using Register 0x316[6:4].
8 ISBs of PHY PRBS Error Count.
0x0
R
0x0
R
8 MSBs of PHY PRBS Error Count.
0x0
R
PHY PRBS Test Pass/Fail.
Bit x corresponds to PHY PRBS pass/fail for
Physical Lane x.
The bit is set to 1 while the error count for
Physical Lane x is less than
PHY_PRBS_THRESHOLD.
Reserved.
Short Transport Layer Sample Select. Selects
which sample to check from the DAC
selected via Bits[3:2].
Sample x
Short Transport Layer Test DAC Select.
Selects which DAC to sample.
Sample from DAC x
Short Transport Layer Test Reset. Resets the
result of short transport layer test.
Not reset
Reset
Short Transport Layer Test Enable. See the
Subclass 0 section for details on how to
perform this test.
Disable
Enable
Short Transport Layer Test Reference, Sample
LSB. This is the lower eight bits of the
expected DAC sample. It is used to compare
with the received DAC sample at the output
of the JESD204B receiver.
0xFF
R
0x0
0x0
R
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
Rev. A | Page 115 of 125
AD9144
Data Sheet
Address
0x32E
Name
SHORT_TPL_TEST_2
Bit No.
[7:0]
Bit Name
SHORT_TPL_REF_
SP_MSB
0x32F
SHORT_TPL_TEST_3
[7:1]
0
RESERVED
SHORT_TPL_FAIL
Settings
0
1
0x333
[7:0]
DEVICE_CONFIG_
13
JESD_BIT_INVERSE
0x400
DEVICE_CONFIG_
REG_13
JESD_BIT_INVERSE_
CTRL
DID_REG
[7:0]
DID_RD
0x401
BID_REG
[7:4]
ADJCNT_RD
[3:0]
BID_RD
7
6
RESERVED
ADJDIR_RD
5
PHADJ_RD
[4:0]
LID0_RD
7
SCR_RD
0x334
0x402
0x403
LID0_REG
SCR_L_REG
[7:0]
0
1
[6:5]
[4:0]
RESERVED
L-1_RD
0
1
3
0x404
F_REG
[7:0]
F-1_RD
0
1
3
Description
Short Transport Layer Test Reference, Sample
MSB. This is the upper eight bits of the
expected DAC sample. It is used to compare
with the received DAC sample at the output
of the JESD204B receiver.
Reserved.
Short Transport Layer Test Fail. This bit shows
whether the selected DAC sample matches
the reference sample. If they match, it is a
test pass, otherwise it is a test fail.
Test pass
Test fail
Must be set to 0x01 for proper JESD interface
configuration.
Logical Lane Invert. Set Bit x high to invert
the JESD deserialized data on Logical Lane x.
Device Identification Number. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
Adjustment Resolution to DAC LMFC. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
Must be 0.
Bank Identification: Extension to DID. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
Reserved.
Direction to Adjust DAC LMFC. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B. Must be 0.
Phase Adjustment Request to DAC Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B. Must be 0.
Lane Identification for Lane 0. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
Transmit Scrambling Status.
Link information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
Scrambling is disabled
Scrambling is enabled
Reserved.
Number of Lanes per Converter Device. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
One lane per converter
Two lanes per converter
Four lanes per converter
Number of Octets per Frame. Settings of 1, 2
and 4 octets per frame are valid. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B.
(One octet per frame) per lane
(Two octets per frame) per lane
(Four octets per frame) per lane
Rev. A | Page 116 of 125
Reset
0x0
Access
R/W
0x0
0x0
R
R
00
R/W
0x0
R/W
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
Data Sheet
Address
0x405
Name
K_REG
AD9144
Bit No.
[7:5]
[4:0]
Bit Name
RESERVED
K-1_RD
Settings
0x0F
0x1F
0x406
M_REG
[7:0]
M-1_RD
0
1
3
0x407
CS_N_REG
[7:6]
CS_RD
5
[4:0]
RESERVED
N-1_RD
[7:5]
SUBCLASSV_RD
[4:0]
NP-1_RD
[7:5]
JESDV_RD
0x0F
0x408
NP_REG
0x0F
0x409
S_REG
000
001
[4:0]
S-1_RD
0
1
0x40A
HD_CF_REG
7
HD_RD
0
1
0x40B
RES1_REG
[6:5]
[4:0]
RESERVED
CF_RD
[7:0]
RES1_RD
Description
Reserved.
Number of Frames per Multiframe. Settings
of 16 or 32 are valid. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B.
16 frames per multiframe
32 frames per multiframe
Number of converters per device. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B. Must be
0, 1, or 3.
One converter per device
Two converters per device
Four converters per device
Number of Control Bits per Sample. Link
information received on Link Lane 0 as specified
in Section 8.3 of JESD204B. CS must be 0.
Reserved.
Converter Resolution. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B. Converter
resolution must be 16.
Converter resolution of 16
Device Subclass Version. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B.
Total Number of Bits per Sample. Link
information received on Link Lane 0 as
specified in Section 8.3 of JESD204B. Must be
16 bits per sample.
16 bits per sample.
JESD204 Version. Link information received
on Link Lane 0 as specified in Section 8.3 of
JESD204B.
JESD204A
JESD204B
Number of Samples per Converter per Frame
Cycle. Settings of one and two are valid. See
Table 35 and Table 36. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B.
One sample per converter per frame
Two samples per converter per frame
High Density Format. See Section 5.1.3 of the
JESD294B standard. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B.
Low density mode
High density mode: link information received
on Lane 0 as specified in Section 8.3 of
JESD204B
Reserved.
Number of Control Words per Frame Clock
Period per Link. Link information received on
Link Lane 0 as specified in Section 8.3 of
JESD204B. Bits[4:0] must be 0.
Reserved Field 1. Link information received on
Link Lane 0 as specified in Section 8.3 of
JESD204B.
Rev. A | Page 117 of 125
Reset
0x0
0x0
Access
R
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
AD9144
Data Sheet
Address
0x40C
Name
RES2_REG
Bit No.
[7:0]
Bit Name
RES2_RD
0x40D
CHECKSUM_REG
[7:0]
FCHK0_RD
0x40E
COMPSUM0_REG
[7:0]
FCMP0_RD
0x412
LID1_REG
[7:5]
[4:0]
RESERVED
LID1_RD
0x415
CHECKSUM1_REG
[7:0]
FCHK1_RD
0x416
COMPSUM1_REG
[7:0]
FCMP1_RD
0x41A
LID2_REG
0x41D
0x41E
CHECKSUM2_REG
COMPSUM2_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID2_RD
FCHK2_RD
FCMP2_RD
0x422
LID3_REG
0x425
0x426
CHECKSUM3_REG
COMPSUM3_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID3_RD
FCHK3_RD
FCMP3_RD
0x42A
LID4_REG
0x42D
0x42E
CHECKSUM4_REG
COMPSUM4_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID4_RD
FCHK4_RD
FCMP4_RD
0x432
LID5_REG
0x435
0x436
CHECKSUM5_REG
COMPSUM5_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID5_RD
FCHK5_RD
FCMP5_RD
0x43A
LID6_REG
0x43D
0x43E
CHECKSUM6_REG
COMPSUM6_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID6_RD
FCHK6_RD
FCMP6_RD
0x442
LID7_REG
0x445
0x446
CHECKSUM7_REG
COMPSUM7_REG
[7:5]
[4:0]
[7:0]
[7:0]
RESERVED
LID7_RD
FCHK7_RD
FCMP7_RD
0x450
ILS_DID
[7:0]
DID
Settings
Description
Reserved Field 2. Link information received on
Link Lane 0 as specified in Section 8.3 of
JESD204B.
Checksum for Link Lane 0. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B.
Computed Checksum for Link Lane 0. The
JESD204B receiver computes the checksum
of the link information received on Lane 0 as
specified in Section 8.3 of JESD204B. The
computation method is set by the
CHECKSUM_MODE bit (Address 0x300[6])
and must match the likewise calculated
checksum in Register 0x40D.
Reserved.
Lane Identification for Link Lane 1.Link
information received on Lane 0 as specified
in section 8.3 of JESD204B.
Checksum for Link Lane 1. Link information
received on Lane 0 as specified in Section 8.3
of JESD204B.
Computed Checksum for Link Lane 1. See the
description for Register 0x40E.
Reserved.
Lane Identification for Link Lane 2.
Checksum for Link Lane 2.
Computed Checksum for Link Lane 2 (see the
description for Register 0x40E).
Reserved.
Lane Identification for Link Lane 3.
Checksum for Link Lane 3.
Computed Checksum for Link Lane 3 (see the
description for Register 0x40E).
Reserved.
Lane Identification for Link Lane 4.
Checksum for Link Lane 4.
Computed Checksum for Link Lane 4 (see the
description for Register 0x40E).
Reserved.
Lane Identification for Link Lane 5.
Checksum for Link Lane 5.
Computed Checksum for Link Lane 5 (see the
description for Register 0x40E).
Reserved.
Lane Identification for Link Lane 6.
Checksum for Link Lane 6.
Computed Checksum for Link Lane 6 (see the
description for Register 0x40E).
Reserved.
Lane Identification for Link Lane 7.
Checksum for Link Lane 7.
Computed Checksum for Link Lane 7 (see the
description for Register 0x40E).
Device Identification Number. Link information
received on Link Lane 0 as specified in
Section 8.3 of JESD204B. Must be set to value
read in Register 0x400.
Rev. A | Page 118 of 125
Reset
0x0
Access
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
0x0
0x0
0x0
R
R
R
R
0x0
R/W
Data Sheet
Address
0x451
0x452
0x453
Name
ILS_BID
ILS_LID0
ILS_SCR_L
AD9144
Bit No.
[7:4]
Bit Name
ADJCNT
[3:0]
BID
7
6
5
RESERVED
ADJDIR
PHADJ
[4:0]
LID0
7
SCR
Settings
0
1
[6:5]
[4:0]
RESERVED
L-1
0
1
3
7
0x454
ILS_F
[7:0]
F-1
0
1
3
0x455
ILS_K
[7:5]
[4:0]
RESERVED
K-1
0x0F
0x1F
0x456
ILS_M
[7:0]
M-1
0
1
3
0x457
ILS_CS_N
[7:6]
CS
5
[4:0]
RESERVED
N-1
[7:5]
SUBCLASSV
0
0xF
0x458
ILS_NP
0
1
[4:0]
NP-1
[7:5]
JESDV
0xF
0x459
ILS_S
000
001
[4:0]
S-1
0
1
Description
Adjustment Resolution to DAC LMFC Must
be set to 0.
Bank Identification: Extension to DID Must be
set to value read in Register 0x401[3:0].
Reserved.
Direction to Adjust DAC LMFC. Must be set to 0.
Phase Adjustment Request to DAC. Must be
set to 0.
Lane Identification for Link Lane 0. Must be set
to the value read in Register 0x402[4:0].
Receiver Descrambling Enable.
Descrambling is disabled
Descrambling is enabled
Reserved.
Number of Lanes per Converter Device. See
Table 35 and Table 36.
One lane per converter
Two lanes per converter
Four lanes per converter
Eight lanes per converter (single link only)
Number of Octets per Lane per Frame. Settings
of 1, 2, and 4 (octets per lane) per frame are
valid. See Table 35 and Table 36.
(One octet per lane) per frame
(Two octets per lane) per frame
(Four octets per lane) per frame
Reserved.
Number of Frames per Multiframe. Settings
of 16 or 32 are valid. Must be set to 32 when
F = 1 (Register 0x476).
16 frames per multiframe
32 frames per multiframe
Number of Converters per Device. See Table 35
and Table 36.
One converter per link
Two converters per link
Four converters per link (single link only)
Number of Control Bits per Sample. Must be
set to 0. Control bits are not supported.
Zero control bits per sample
Reserved.
Converter Resolution. Must be set to 16 bits
of resolution.
Converter resolution of 16.
Device Subclass Version.
Subclass 0
Subclass 1
Total Number of Bits per Sample. Must be set
to 16 bits per sample.
16 bits per sample.
JESD204 Version.
JESD204A
JESD204B
Number of Samples per Converter per Frame
Cycle. Settings of one and two are valid. See
Table 35 and Table 36.
One sample per converter per frame
Two samples per converter per frame
Rev. A | Page 119 of 125
Reset
0x0
Access
R/W
0x0
R/W
0x0
0x0
0x0
R
R/W
R/W
0x0
R/W
0x1
R/W
0x0
0x3
R
R/W
0x0
R/W
0x0
0x1F
R
R/W
0x1
R/W
0x0
R/W
0x0
0xF
R
R/W
0x1
R/W
0xF
R/W
0x1
R/W
0x0
R/W
AD9144
Address
0x45A
Name
ILS_HD_CF
Data Sheet
Bit No.
7
Bit Name
HD
Settings
0
1
[6:5]
[4:0]
RESERVED
CF
0x45B
0x45C
0x45D
ILS_RES1
ILS_RES2
ILS_CHECKSUM
[7:0]
[7:0]
[7:0]
RES1
RES2
FCHK0
0x46B
ERRCNTRMON_RB
[7:0]
READERRORCNTR
0x46B
ERRCNTRMON
7
[6:4]
RESERVED
LANESEL
[3:2]
[1:0]
RESERVED
CNTRSEL
x
00
01
10
0x46C
0x46D
LANEDESKEW
BADDISPARITY_RB
[7:0]
[7:0]
LANEDESKEW
BADDIS
0x46D
BADDISPARITY
7
RST_IRQ_DIS
6
DISABLE_ERR_
CNTR_DIS
5
RST_ERR_CNTR_DIS
[4:3]
[2:0]
RESERVED
LANE_ADDR_DIS
0x46E
NIT_RB
[7:0]
NIT
0x46E
NIT_W
7
RST_IRQ_NIT
6
DISABLE_ERR_
CNTR_NIT
5
RST_ERR_CNTR_NIT
[4:3]
[2:0]
RESERVED
LANE_ADDR_NIT
Description
High Density Format. If F = 1, HD must be set
to 1. Otherwise, HD must be set to 0. See
Section 5.1.3 of JESD204B standard.
Low density mode
High density mode
Reserved.
Number of Control Words per Frame Clock
Period per Link. Must be set to 0. Control bits
are not supported.
Reserved Field 1.
Reserved Field 2.
Checksum for Link Lane 0. Calculated
checksum. Calculation depends on 0x300[6].
Read JESD204B Error Counter. After selecting
the lane and error counter by writing to
LANESEL and CNTRSEL (both in this same
register), the selected error counter is read
back here.
Reserved.
Link Lane select for JESD204B error counter.
Selects the lane whose errors are read back
in this register.
Selects Link Lane x
Reserved.
JESD204B Error Counter Select. Selects the
type of error that are read back in this
register.
BADDISCNTR: bad running disparity counter
NITCNTR: not in table error counter
UCCCNTR: Unexpected control character counter
Lane Deskew. Setting Bit x deskews Link Lane x
Bad Disparity Character Error (BADDIS). Bit x
is set when the bad disparity error count for
Link Lane x reaches the threshold in
Register 0x47C.
BADDIS IRQ Reset. Reset BADDIS IRQ for lane
selected via Bits[2:0] by writing 1 to this bit.
BADDIS Error Counter Disable. Disable the
BADDIS error counter for lane selected via
Bits[2:0] by writing 1 to this bit.
BADDIS Error Counter Reset. Reset BADDIS
error counter for lane selected via Bits[2:0] by
writing 1 to this bit.
Reserved.
Link Lane Address for Functions Described in
Bits[7:5].
Not in table Character Error (NIT). Bit x is set
when the NIT error count for Link Lane x
reaches the threshold in Register 0x47C.
IRQ Reset. Reset IRQ for lane selected via
Bits[2:0] by writing 1 to this bit.
Disable Error Counter. Disable the error
counter for lane selected via Bits[2:0] by
writing 1 to this bit.
Reset Error Counter. Reset error counter for lane
selected via Bits[2:0] by writing 1 to this bit.
Reserved.
Link Lane Address for Functions Described in
Bits[7:5].
Rev. A | Page 120 of 125
Reset
0x1
Access
R/W
0x0
0x0
R
R/W
0x0
0x0
0x45
R/W
R/W
R/W
0x0
R
0x0
0x0
R
W
0x0
0x0
R
W
0xF
0x0
R/W
R
0x0
W
0x0
W
0x0
W
0x0
0x0
R
W
0x0
R
0x0
W
0x0
W
0x0
W
0x0
0x0
R
W
Data Sheet
AD9144
Address
0x46F
Name
UNEXPECTEDCONTROL_RB
Bit No.
[7:0]
Bit Name
UCC
0x46F
UNEXPECTEDCONTROL_W
7
RST_IRQ_UCC
6
DISABLE_ERR_
CNTR_UCC
5
[4:3]
[2:0]
RST_ERR_CNTR_
UCC
RESERVED
LANE_ADDR_UCC
[7:0]
CODEGRPSYNC
0x470
CODEGRPSYNCFLG
Settings
0
1
0x471
FRAMESYNCFLG
[7:0]
FRAMESYNC
0
1
0x472
GOODCHKSUMFLG
[7:0]
GOODCHECKSUM
0
1
0x473
INITLANESYNCFLG
[7:0]
INITIALLANESYNC
0x476
CTRLREG1
[7:0]
F
1
2
4
0x477
CTRLREG2
7
ILAS_MODE
1
0
[6:4]
3
0x478
KVAL
[2:0]
[7:0]
RESERVED
THRESHOLD_
MASK_EN
RESERVED
KSYNC
x
Description
Unexpected Control Character Error (UCC).
Bit x is set when the UCC error count for Link
Lane x reaches the threshold in Register 0x47C.
IRQ Reset. Reset IRQ for lane selected via
Bits[2:0] by writing 1 to this bit.
Disable Error Counter. Disable the error
counter for lane selected via Bits[2:0] by
writing 1 to this bit.
Reset Error Counter. Reset error counter for
lane selected via Bits[2:0] by writing 1 to this bit.
Reserved.
Link Lane Address for Functions Described in
Bits[7:5].
Code Group Sync Flag (from Each
Instantiated Lane). Writing 1 to Bit 7 resets
the IRQ. The associated IRQ flag is located in
Register 0x47A[0]. A loss of CODEGRPSYNC
triggers sync request assertion. See the
SYNCOUT and SYSREF Signals section and
the Deterministic Latency section.
Synchronization is lost
Synchronization is achieved
Frame Sync Flag (from Each Instantiated
Lane). This register indicates the live status
for each lane. Writing 1 to Bit 7 resets the
IRQ. A loss of frame sync automatically
initiates a synchronization sequence.
Synchronization is lost
Synchronization is achieved
Good Checksum Flag (from Each Instantiated
Lane). Writing 1 to Bit 7 resets the IRQ. The
associated IRQ flag is located in
Register 0x47A[2].
Last computed checksum is not correct
Last computed checksum is correct
Initial Lane Sync Flag (from Each Instantiated
Lane). Writing 1 to Bit 7 resets the IRQ. The
associated IRQ flag is located in Register
0x47A[3]. Loss of synchronization is also
reported on SYNCOUT1± or SYNCOUT0±. See
the SYNCOUT and SYSREF± Signal section
and the Deterministic Latency section.
Number of Octets per Frame. Settings of 1, 2,
and 4 are valid. See Table 35 and Table 36.
One octet per frame
Two octets per frame
Four octets per frame
ILAS Test Mode. Defined in Section 5.3.3.8 of
JESD204B specification.
JESD204B receiver is constantly receiving
ILAS frames
Normal link operation
Reserved.
Threshold Mask Enable. Set this bit if using
SYNC_ASSERTION_MASK (Register 0x47B[7:5]).
Reserved.
Number of K Multiframes During ILAS
(Divided by Four). Sets the number of
multiframes to send initial lane alignment
sequence. Cannot be set to 0.
4x multiframes during ILAS
Rev. A | Page 121 of 125
Reset
0x0
Access
R
0x0
W
0x0
W
0x0
W
0x0
0x0
R
W
0x0
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x1
R/W
0x0
R/W
0x0
0x0
R
R/W
0x0
0x1
R
R/W
AD9144
Address
0x47A
Name
IRQVECTOR_MASK
Data Sheet
Bit No.
7
Bit Name
BADDIS_MASK
Settings
1
6
NIT_MASK
1
5
UCC_MASK
1
4
3
2
1
0
0x47A
IRQVECTOR_FLAG
7
RESERVED
INITIALLANESYNC_
MASK
1
BADCHECKSUM_
MASK
1
FRAMESYNC_
MASK
1
CODEGRPSYNC_
MASK
1
BADDIS_FLAG
1
6
NIT_FLAG
1
5
UCC_FLAG
1
4
3
2
1
0
RESERVED
INITIALLANESYNC_
FLAG
1
BADCHECKSUM_
FLAG
1
FRAMESYNC_
FLAG
1
CODEGRPSYNC_
FLAG
1
Description
Bad Disparity Mask.
If the bad disparity count reaches
ERRORTHRESH on any lane, IRQ is pulled low.
Not in table Mask.
If the not in table character count reaches
ERRORTHRESH on any lane, IRQ is pulled low.
Unexpected Control Character Mask.
If the unexpected control character count
reaches ERRORTHRESH on any lane, IRQ is
pulled low.
Reserved.
Initial Lane Sync Mask.
If initial lane sync (0x473) fails on any lane,
IRQ is pulled low.
Bad Checksum Mask.
If there is a bad checksum (0x472) on any
lane, IRQ is pulled low.
Frame Sync Mask
If frame sync (0x471) fails on any lane, IRQ is
pulled low.
Code Group Sync Machine Mask.
If code group sync (0x470) fails on any lane,
IRQ is pulled low.
Bad Disparity Error Count.
Bad disparity character count reached
ERRORTHRESH (0x47C) on at least one lane.
Read Register 0x46D to determine which
lanes are in error.
Not in table Error Count
Not in table character count reached
ERRORTHRESH (0x47C) on at least one lane.
Read Register 0x46E to determine which
lanes are in error.
Unexpected Control Character Error Count
Unexpected control character count reached
ERRORTHRESH (0x47C) on at least one lane.
Read Register 0x46F to determine which
lanes are in error.
Reserved.
Initial Lane Sync Flag.
Initial lane sync failed on at least one lane.
Read Register 0x473 to determine which
lanes are in error
Bad Checksum Flag.
Bad checksum on at least one lane. Read
Register 0x472 to determine which lanes are in
error.
Frame Sync Flag.
Frame sync failed on at least one lane. Read
Register 0x471 to determine which lanes are
in error.
Code Group Sync Flag.
Code group sync failed on at least one lane.
Read Register 0x470 to determine which
lanes are in error
Rev. A | Page 122 of 125
Reset
0x0
Access
W
0x0
W
0x0
W
0x0
0x0
R
W
0x0
W
0x0
W
0x0
W
0x0
R
0x0
R
0x0
R
0x0
0x0
R
R
0x0
R
0x0
R
0x0
R
Data Sheet
Address
0x47B
Name
SYNCASSERTIONMASK
AD9144
Bit No.
7
Bit Name
BADDIS_S
Settings
1
6
NIT_S
1
5
UCC_S
1
4
CMM
1
3
CMM_ENABLE
1
0
0x47C
ERRORTHRES
[2:0]
[7:0]
RESERVED
ETH
0x47D
LANEENABLE
[7:0]
LANE_ENA
0x47E
RAMP_ENA
[7:1]
0
RESERVED
ENA_RAMP_
CHECK
0
1
0x520
DIG_TEST0
[7:2]
RESERVED
0x521
DC_TEST_VALUEI0
1
0
[7:0]
0x522
DC_TEST_VALUEI1
[7:0]
0x523
DC_TEST_VALUEQ0
[7:0]
0x524
DC_TEST_VALUEQ1
[7:0]
DC_TEST_MODE
RESERVED
DC_TEST_
VALUEI[7:0]
DC_TEST_
VALUEI [15:8]
DC_TEST_
VALUEQ[7:0]
DC_TEST_
VALUEQ[15:8]
Description
Bad Disparity Error on Sync.
Asserts a sync request on SYNCOUTx± when
the bad disparity character count reaches the
threshold in Register 0x47C
Not in table Error on Sync.
Asserts a sync request on SYNCOUTx± when
the not in table character count reaches the
threshold in Register 0x47C
Unexpected Control Character Error on Sync.
Asserts a sync request on SYNCOUTx± when
the unexpected control character count
reaches the threshold in Register 0x47C
Configuration Mismatch IRQ. If
CMM_ENABLE is high, this bit latches on a
rising edge and pull IRQ low. When latched,
write a 1 to clear this bit. If CMM_ENABLE is
low, this bit is non-functional.
Link Lane 0 configuration registers (Register
0x450 to Register 0x45D) do not match the
JESD204B transmit settings (Register 0x400
to Register 0x40D)
Configuration Mismatch IRQ Enable.
Enables IRQ generation if a configuration
mismatch is detected
Configuration mismatch IRQ disabled
Reserved.
Error Threshold. Bad disparity, not in table,
and unexpected control character errors are
counted and compared to the error
threshold value. When the count reaches the
threshold, either an IRQ is generated or the
SYNCOUTx± signal is asserted per the mask
register settings, or both. Function is
performed in all lanes.
Lane Enable. Setting Bit x enables Link Lane x.
This register must be programmed before
receiving the code group pattern for proper
operation.
Reserved.
Enable Ramp Checking at the Beginning of
ILAS.
Disable ramp checking at beginning of ILAS;
ILAS data need not be a ramp
Enable ramp checking; ILAS data needs to be
a ramp starting at 00-01-02; otherwise, the
ramp ILAS fails and the device does not start up
Must write default value for proper
operation.
DC Test Mode
Reserved.
DC Value LSB of DC Test Mode for I DAC.
Reset
0x0
Access
R/W
0x0
R/W
0x0
R/W
0x0
R/W
0x1
R/W
0x0
0xFF
R
R/W
0xF
R/W
0x0
0x0
R
W
0x7
R/W
0x0
0x0
0x0
R/W
R/W
R/W
DC value MSB of DC Test Mode for I DAC.
0x0
R/W
DC value LSB of DC Test Mode for Q DAC.
0x0
R/W
DC value MSB of DC Test Mode for Q DAC.
0x0
R/W
Rev. A | Page 123 of 125
AD9144
Data Sheet
OUTLINE DIMENSIONS
12.10
12.00 SQ
11.90
0.28
0.23
0.18
0.60 MAX
0.60
MAX
67
66
88
PIN 1
INDICATOR
1
PIN 1
INDICATOR
11.85
11.75 SQ
11.65
0.50
BSC
0.50
0.40
0.30
22
23
45
44
TOP VIEW
BOTTOM VIEW
10.50
REF
0.70
0.65
0.60
FOR PROPER CONNECTION OF
THE EXPOSED PAD, REFER TO
THE PIN CONFIGURATION AND
FUNCTION DESCRIPTIONS
SECTION OF THIS DATA SHEET.
0.045
0.025
0.005
COPLANARITY
0.08
0.20 REF
SEATING
PLANE
08-10-2012-A
12° MAX
0.90
0.85
0.80
7.55
7.40 SQ
7.25
EXPOSED PAD
COMPLIANT TO JEDEC STANDARDS MO-220-VRRD
Figure 88. 88-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
12 mm × 12 mm Body, Very Thin Quad
(CP-88-6)
Dimensions shown in millimeters
12.10
12.00 SQ
11.90
0.30
0.25
0.20
0.60 MAX
0.60
MAX
67
88
66
1
PIN 1
INDICATOR
PIN 1
INDICATOR
0.50
BSC
7.55
7.40 SQ
5.25
EXPOSED
PAD
0.65
0.55
0.45
22
44
TOP VIEW
0.90
0.85
0.80
PKG-004598
SEATING
PLANE
0.70
0.65
0.60
12° MAX
0.190~0.245 REF
0.50
0.40
0.30
45
0.045
0.025
0.005
COPLANARITY
0.08
23
BOTTOM VIEW
10.50
REF
FOR PROPER CONNECTION OF
THE EXPOSED PAD, REFER TO
THE PIN CONFIGURATION AND
FUNCTION DESCRIPTIONS
SECTION OF THIS DATA SHEET.
COMPLIANT TO JEDEC STANDARDS MO-220
Figure 89. 88-Lead Lead Frame Chip Scale Package [LFCSP_VQ] (Variable Lead Length)
12 mm × 12 mm Body, Very Thin Quad
(CP-88-9)
Dimensions shown in millimeters
Rev. A | Page 124 of 125
1.00
0.90
0.80
0.80
0.70
0.60
08-04-2014-A
11.85
11.75 SQ
11.65
Data Sheet
AD9144
ORDERING GUIDE
Model1
AD9144BCPZ
AD9144BCPZRL
AD9144BCPAZ
AD9144BCPAZRL
AD9144-EBZ
AD9144-FMC-EBZ
AD9144-M6720-EBZ
1
Temperature Range
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
Package Description
88-Lead LFCSP_VQ
88-Lead LFCSP_VQ
88-Lead LFCSP_VQ (Variable Lead Length)
88-Lead LFCSP_VQ (Variable Lead Length)
DPG3 Evaluation Board
FMC Evaluation Board
DPG3 Evaluation Board with ADRF6720 Modulator
Z = RoHS Compliant Part.
©2014–2015 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D11675-0-6/15(A)
Rev. A | Page 125 of 125
Package Option
CP-88-6
CP-88-6
CP-88-9
CP-88-9
Similar pages