AD AD9866BCPRL

Broadband Modem Mixed Signal Front End
AD9866
Powerline networking
VDSL and HPNA
AD9866
PWR DWN
MODE
TXEN/SYNC
TXCLK
2-4X
IAMP
TxDAC
0 TO –7.5dB
IOUT_G+
IOUT_N+
IOUT_N–
IOUT_G–
0 TO –12dB
12
CLKOUT_1
CLKOUT_2
CLK
SYN.
ADIO[11:6]/
Tx[5:0]
2M
OSCIN
XTAL
CLK
MULTIPLIER
ADIO[5:0]/
Rx[5:0]
12
RXE/SYNC
RXCLK
AGC[5:0]
SPI
RX+
ADC
80MSPS
2-POLE
LPF
1-POLE
LPF
RX–
6
4
REGISTER
CONTROL
0 TO 6dB
∆ = 1dB
– 6 TO 18dB –6 TO 24dB
∆ = 6dB
∆ = 6dB
04560-0-001
APPLICATIONS
FUNCTIONAL BLOCK DIAGRAM
IOUT_P–
Low cost 3.3 V CMOS MxFETM for broadband modems
12-bit D/A converter
2×/4× interpolation filter
200 MSPS DAC update rate
Integrated 23 dBm line driver with 19.5 dB gain control
12-bit, 80 MSPS A/D converter
−12 dB to +48 dB low noise RxPGA (< 2.5 nV/rtHz)
Third order programmable low-pass filter
Flexible digital data path interface
Half- and full-duplex operation
Backward compatible with AD9975 and AD9876
Various power-down/reduction modes
Internal clock multiplier (PLL)
2 auxiliary programmable clock outputs
Available in 64-lead chip scale package or bare die
IOUT_P+
FEATURES
Figure 1.
GENERAL DESCRIPTION
The AD9866 is a mixed-signal front end (MxFE) IC for
transceiver applications requiring Tx and Rx path functionality
with data rates up to 80 MSPS. Its flexible digital interface,
power saving modes, and high Tx-to-Rx isolation make it well
suited for half- and full-duplex applications. The digital interface is extremely flexible allowing simple interfaces to digital
back ends that support half- or full-duplex data transfers, thus
often allowing the AD9866 to replace discrete ADC and DAC
solutions. Power saving modes include the ability to reduce
power consumption of individual functional blocks or to power
down unused blocks in half-duplex applications. A serial port
interface (SPI®) allows software programming of the various
functional blocks. An on-chip PLL clock multiplier and
synthesizer provide all the required internal clocks, as well as
two external clocks from a single crystal or clock source.
or to an internal low distortion current amplifier. The current
amplifier (IAMP) can be configured as a current or voltage
mode line driver (with two external npn transistors) capable of
delivering in excess of 23 dBm peak signal power. Tx power can
be digitally controlled over a 19.5 dB range in 0.5 dB steps.
The Tx signal path consists of a bypassable 2×/4× low-pass
interpolation filter, a 12-bit TxDAC, and a line driver. The
transmit path signal bandwidth can be as high as 34 MHz at an
input data rate of 80 MSPS. The TxDAC provides differential
current outputs that can be steered directly to an external load
The AD9866 provides a highly integrated solution for many
broadband modems. It is available in a space-saving 64-lead
chip scale package and is specified over the commercial (−40°C
to +85°C) temperature range.
The receive path consists of a programmable amplifier
(RxPGA), a tunable low pass filter (LPF), and a 12-bit ADC. The
low noise RxPGA has a programmable gain range of −12 dB to
+48 dB in 1 dB steps. Its input referred noise is less than
3.3 nV/rtHz for gain settings beyond 30 dB. The receive path
LPF cutoff frequency can either be set over a 15 MHz to
35 MHz range or simply bypassed. The 12-bit ADC achieves
excellent dynamic performance over a 5 MSPS to 80 MSPS
span. Both the RxPGA and the ADC offer scalable power
consumption allowing power/performance optimization.
Rev. 0
Information furnished by Analog Devices is believed to be accurate and reliable.
However, no responsibility is assumed by Analog Devices for its use, nor for any
infringements of patents or other rights of third parties that may result from its use.
Specifications subject to change without notice. No license is granted by implication
or otherwise under any patent or patent rights of Analog Devices. Trademarks and
registered trademarks are the property of their respective owners.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700
www.analog.com
Fax: 781.326.8703
© 2003 Analog Devices, Inc. All rights reserved.
AD9866
TABLE OF CONTENTS
Specifications..................................................................................... 3
Tx Programmable Gain Control .............................................. 30
Tx Path Specifications.................................................................. 3
TxDAC Output Operation........................................................ 30
Rx Path Specifications.................................................................. 4
IAMP Current Mode Operation .............................................. 30
Power Supply Specifications ....................................................... 5
IAMP Voltage Mode Operation ............................................... 31
Digital Specifications ................................................................... 6
IAMP Current Consumption Considerations........................ 32
Serial Port Timing Specifications............................................... 7
Receive Path .................................................................................... 33
Half-Duplex Data Interface (ADIO Port) Timing
Specifications ................................................................................ 7
Rx Programmable Gain Amplifier........................................... 33
Full-Duplex Data Interface (Tx and Rx PORT) Timing
Specifications ................................................................................ 8
Absolute Maximum Ratings............................................................ 9
Thermal Characteristics .............................................................. 9
ESD Caution.................................................................................. 9
Pin Configuration and Function Descriptions........................... 10
Typical Performance Characteristics ........................................... 12
Rx Path Typical Performance Characteristics ........................ 12
TxDAC Path Typical Performance Characteristics ............... 16
IAMP Path Typical Performance Characteristics .................. 18
Serial Port ........................................................................................ 19
Register Map Description ......................................................... 21
Serial Port Interface (SPI) ......................................................... 21
Digital Interface .............................................................................. 23
Half-Duplex Mode ..................................................................... 23
Full-Duplex Mode ...................................................................... 24
RxPGA Control .......................................................................... 25
TxPGA Control .......................................................................... 27
Transmit Path .................................................................................. 28
Low-Pass Filter ........................................................................... 34
Analog to Digital Converter (ADC)........................................ 35
AGC Timing Considerations.................................................... 36
Clock Synthesizer ........................................................................... 37
Power Control and Dissipation .................................................... 39
Power-Down ............................................................................... 39
Half-Duplex Power Savings ...................................................... 39
Power Reduction Options......................................................... 40
Power Dissipation ...................................................................... 42
Mode Select upon Power-Up and Reset.................................. 42
Analog and Digital Loop-back Test Modes ............................ 43
PCB Design Considerations.......................................................... 44
Component Placement.............................................................. 44
Power Planes and Decoupling .................................................. 44
Ground Planes ............................................................................ 44
Signal Routing ............................................................................ 44
Evaluation Board ............................................................................ 46
Outline Dimensions ....................................................................... 47
Ordering Guide .......................................................................... 47
Digital Interpolation Filters ...................................................... 28
TxDAC and IAMP Architecture............................................... 28
REVISION HISTORY
Revision 0: Initial Version
Rev. 0 | Page 2 of 48
AD9866
SPECIFICATIONS
TX PATH SPECIFICATIONS
Table 1. AVDD = 3.3 V ± 5%, DVDD = CLKVDD = DRVDD = 3.3 V ± 10%; fOSCIN = 50 MHz, fDAC = 200 MHz, RSET = 2.0 kΩ, unless
otherwise noted
Parameter
TxDAC DC CHARACTERISTICS
Resolution
Update Rate
Full-Scale Output Current (IOUTP_FS)
Gain Error1
Offset Error
Voltage Compliance Range
TxDAC GAIN CONTROL CHARACTERISTICS
Minimum Gain
Maximum Gain
Gain Step Size
Gain Step Accuracy
Gain Range Error
TxDAC AC CHARACTERISTICS2
Fundamental
Signal-to-Noise and Distortion
Signal-to-Noise Ratio
THD
SFDR
IAMP DC CHARACTERISTICS
IOUTN Full-Scale Current = IOUTN+ + IOUTN−
IOUTG Full-Scale Current = IOUTG+ + IOUTG−
AC Voltage Compliance Range
IAMPN AC CHARACTERISTICS3
Fundamental
IOUTN SFDR (Third Harmonic)
IAMP GAIN CONTROL CHARACTERISTICS
Minimum Gain
Maximum Gain
Gain Step Size
Gain Step Accuracy
IOUTN Gain Range Error
REFERENCE
Internal Reference Voltage4
Reference Error
Reference Drift
Tx DIGITAL FILTER CHARACTERISTICS (2× INTERPOLATION)
Latency (Relative to 1/ FDAC)
−0.2 dB Bandwidth
−3 dB Bandwidth
Stop-Band Rejection (0.289 FDAC to 0.711 FDAC)
Tx DIGITAL FILTER CHARACTERISTICS (4× Interpolation)
Latency (Relative to 1/ FDAC)
−0.2 dB Bandwidth
Temp
Test Level
Full
Full
Full
25°C
25°C
Full
II
IV
I
V
25°C
25°C
25°C
25°C
25°C
V
V
V
IV
V
−7.5
0
0.5
Monotonic
±2
Full
Full
Full
Full
IV
IV
IV
IV
68.5
0.5
69.2
69.8
−79
81
Full
Full
Full
IV
IV
IV
2
2
1
25°C
Full
IV
43.3
25°C
25°C
25°C
25°C
25°C
V
V
V
IV
V
−19.5
0
0.5
Monotonic
0.5
dB
dB
dB
dB
dB
25°C
I
1.23
V
Full
Full
V
V
0.7
30
Full
Full
V
V
43
0.2187
Cycles
fOUT/fDAC
Full
Full
V
V
0.2405
50
fOUT /fDAC
dB
Full
Full
V
V
Rev. 0 | Page 3 of 48
Min
Typ
Max
12
200
25
2
±2
2
−1
66.6
68.4
+1.5
Bits
MSPS
mA
% FS
uA
V
dB
dB
dB
dB
−68.7
105
150
7
13
45.2
96
0.1095
Unit
dBm
dBc
dBc
dBc
dBc
mA
mA
V
dBm
dBc
3.4
%
ppm/oC
Cycles
fOUT /fDAC
AD9866
Parameter
−3 dB Bandwidth
Stop Band Rejection (0.289 fOSCIN to 0.711 fOSCIN)
PLL CLK MULTIPLIER
OSCIN Frequency Range
Internal VCO Frequency Range
Duty Cycle
OSCIN Impedance
CLKOUT1 Jitter5
CLKOUT2 Jitter6
CLKOUT1 and CLKOUT2 Duty Cycle7
Temp
Full
Full
Test Level
V
V
Min
Full
Full
Full
25°C
25°C
IV
IV
II
V
III
5
20
40
25°C
Full
III
III
Typ
0.1202
50
Max
Unit
fOUT /fDAC
dB
80
200
60
MHz
MHz
%
ΜΩ//pF
ps rms
100//3
12
6
45
55
ps rms
%
1
Gain error and gain temperature coefficients are based on the ADC only (with a fixed 1.23 V external reference and a 1 V p-p differential analog input).
TxDAC IOUTFS = 20 mA, differential output with 1:1 transformer with source and load termination of 50 Ω, FOUT = 5 MHz, 4× interpolation.
IOUN full-scale current = 80 mA, fOSCIN= 80 MHz, fDAC=160 MHz, 2× interpolation.
4
Use external amplifier to drive additional load.
5
Internal VCO operates at 200 MHz , set to divide-by-1.
6
Because CLKOUT2 is a divided down version of OSCIN, its jitter is typically equal to OSCIN.
7
CLKOUT2 is an inverted replica of OSCIN, if set to divide-by-1.
2
3
RX PATH SPECIFICATIONS
Table 2. AVDD = 3.3 V ± 5%, DVDD = CLKVDD = DRVDD = 3.3 V ± 10%; half- or full-duplex operation with CONFIG = 0 default
power bias settings, unless otherwise noted
Parameter
Rx INPUT CHARACTERISTICS
Input Voltage Span (RxPGA gain = −10 dB)
Input Voltage Span (RxPGA gain = +48 dB)
Input Common-Mode Voltage
Differential Input Impedance
Temp
Test Level
Full
Full
25°C
25°C
III
III
III
III
Input Bandwidth (with RxLPF Disabled, RxPGA = 0 dB)
Input Voltage Noise Density (RxPGA Gain = 36 dB, f−3 dBF = 26 MHz)
Input Voltage Noise Density (RxPGA Gain = 48 dB, f−3 dBF = 26 MHz)
RxPGA CHARACTERISTICS
Minimum Gain
Maximum Gain
Gain Step Size
Gain Step Accuracy
Gain Range Error
RxLPF CHARACTERISTICS
Cutoff Frequency (f−3 dBF ) range
Attenuation at 55.2 MHz with f−3 dBF = 21 MHz
Pass-Band Ripple
Settling Time to 5 dB RxPGA Gain Step @ fADC = 50 MSPS
Settling Time to 60 dB RxPGA Gain Step @ fADC = 50 MSPS
ADC DC CHARACTERISTICS
Resolution
Conversion Rate
RX PATH LATENCY1
Full-Duplex Interface
Half-Duplex Interface
25°C
25°C
25°C
Min
Typ
Max
Unit
III
III
III
6.33
8
1.3
400
4.0
53
2.7
2.4
Ω
pF
MHz
nV/rtHz
nV/rtHz
25°C
25°C
25°C
25°C
25°C
III
III
III
III
III
−12
48
1
Monotonic
0.5
dB
dB
dB
dB
dB
Full
25°C
25°C
25°C
25°C
III
III
III
III
III
NA
FULL
NA
II
Full
Full
V
V
Rev. 0 | Page 4 of 48
15
V p-p
mV p-p
V
35
20
±1
20
100
12
5
80
10.5
10.0
MHz
dB
dB
ns
ns
Bits
MSPS
Cycles
Cycles
AD9866
Parameter
Rx PATH COMPOSITE AC PERFORMANCE @ fADC = 50 MSPS2
RxPGA Gain = 48 dB (Full-Scale = 8.0 mV p-p)
Signal-to-Noise and Distortion (SNR)
Total Harmonic Distortion (THD)
RxPGA Gain = 24 dB (Full-Scale = 126 mV p-p)
Signal-to-Noise (SNR)
Total Harmonic Distortion (THD)
RxPGA Gain = 0 dB (Full-Scale = 2.0 V p-p)
Signal-to-Noise and Distortion (SINAD)
Total Harmonic Distortion (THD)
Rx PATH COMPOSITE AC PERFORMANCE @ fADC = 80 MSPS3
RxPGA Gain = 48 dB (Full-Scale = 8.0 m V p-p)
Signal-to-Noise (SNR)
Total Harmonic Distortion (THD)
RxPGA Gain = 24 dB (Full-Scale = 126 m V p-p)
Signal-to-Noise (SNR)
Total Harmonic Distortion (THD)
RxPGA Gain = 0 dB (Full-Scale = 2.0 V p-p)
Signal-to-Noise (SNR)
Total Harmonic Distortion (THD)
Rx-to-Tx PATH FULL-DUPLEX ISOLATION
(1 V p-p, 10 MHz Sine Wave Tx Output)
RxPGA Gain = 40 dB
IOUTP± Pins to RX± Pins
IOUTG± Pins to RX± Pins
RxPGA Gain = 0 dB
IOUTP± Pins to RX± Pins
IOUTG± Pins to RX± Pins
Temp
Test Level
Min
25°C
25°C
III
III
43.7
−71
dBc
dBc
25°C
25°C
III
III
63.1
−67.2
dBc
dBc
Full
Full
IV
IV
64.3
−67.3
dBc
dBc
25°C
25°C
III
III
41.8
−67
dBc
dBc
25°C
25°C
III
III
58.6
−62.9
dBc
dBc
25°C
25°C
II
II
25°C
25°C
III
III
83
37
dBc
dBc
25°C
25°C
III
III
123
77
dBc
dBc
61.1
Typ
Max
62.9
−70.8
−60.8
Unit
dBc
dBc
1
Includes RxPGA, ADC pipeline, and ADIO bus delay relative to fADC.
fIN = 5 MHz, AIN = −1.0 dBFS , LPF cut-off frequency set to 15.5 MHz with Reg. 0x08 = 0x80.
3
fIN = 5 MHz, AIN = −1.0 dBFS , LPF cut-off frequency set to 26 MHz with Reg. 0x08 = 0x80.
2
POWER SUPPLY SPECIFICATIONS
Table 3. AVDD = 3.3 V, DVDD = CLKVDD = DRVDD = 3.3 V; RSET = 2 kΩ, full-duplex operation with fDATA = 80 MSPS,1 unless
otherwise noted
Parameter
SUPPLY VOLTAGES
AVDD
CLKVDD
DVDD
DRVDD
IS_TOTAL (Total Supply Current)
POWER CONSUMPTION
IAVDD + ICLKVDD (Analog Supply Current)
IDVDD + IDRVDD (Digital Supply Current)
POWER CONSUMPTION (Half-Duplex Operation with fDATA = 50 MSPS)2
Tx Mode
IAVDD + ICLKVDD
IDVDD + IDRVDD
Temp
Test Level
Min
Typ
Max
Unit
Full
Full
Full
Full
Full
V
V
V
V
II
3.135
3.0
3.0
3.0
3.3
3.3
3.3
3.3
406
3.465
3.6
3.6
3.6
475
V
V
V
V
mA
Full
IV
IV
311
95
342
133
mA
mA
25°C
25°C
IV
IV
112
46
130
49.5
mA
mA
Rev. 0 | Page 5 of 48
AD9866
Parameter
Rx Mode
IAVDD + ICLKVDD
IDVDD + IDRVDD
POWER CONSUMPTION OF FUNCTIONAL BLOCKS1 (IAVDD + ICLKVDD)
RxPGA and LPF
ADC
TxDAC
IAMP (Programmable)
Reference
CLK PLL and Synthesizer
MAXIMUM ALLOWABLE POWER DISSIPATION
STANDBY POWER CONSUMPTION
IS_TOTAL (Total Supply Current)
POWER DOWN DELAY (USING PWR_DWN PIN)
RxPGA and LPF
ADC
TxDAC
IAMP
CLK PLL and Synthesizer
POWER UP DELAY (USING PWR_DWN PIN)
RxPGA and LPF
ADC
TxDAC
IAMP
CLK PLL and Synthesizer
Temp
Test Level
Min
25°C
25°C
25°C
25°C
25°C
25°C
25°C
25°C
Full
III
III
III
III
III
III
IV
Typ
Max
Unit
225
36.5
253
39
mA
mA
120
mA
mA
mA
mA
mA
mA
W
87
108
38
10
170
107
1.66
Full
13
mA
25°C
25°C
25°C
25°C
25°C
III
III
III
III
III
440
12
20
20
27
ns
ns
ns
ns
ns
25°C
25°C
25°C
25°C
25°C
III
III
III
III
III
7.8
88
13
20
20
µs
ns
µs
ns
µs
1
Default power-up settings for MODE = HIGH and CONFIG = LOW, IOUTP_FS = 20 mA, does not include IAMP’s current consumption, which is application dependent.
Default power-up settings for MODE = LOW and CONFIG = LOW .
2
DIGITAL SPECIFICATIONS
Table 4. AVDD = 3.3 V ± 5%, DVDD = CLKVDD = DRVDD = 3.3 V ± 10%; RSET = 2 kΩ, unless otherwise noted
Parameter
CMOS LOGIC INPUTS
High Level Input Voltage
Low Level Input Voltage
Input Leakage Current
Input Capacitance
CMOS LOGIC OUTPUTS (CLOAD = 5 pF)
High Level Output Voltage (IOH = 1 mA)
Low Level Output Voltage (IOH = 1 mA)
Output Rise/Fall Time (High Strength Mode and CLOAD = 15 pF)
Output Rise/Fall Time (Low Strength Mode and CLOAD = 15 pF)
Output Rise/Fall Time (High Strength Mode and CLOAD = 5 pF)
Output Rise/Fall Time (Low Strength Mode and CLOAD = 5 pF)
RESET
Minimum Low Pulse Width (Relative to fADC)
Temp
Test Level
Min
Full
Full
VI
VI
DRVDD – 0.7
Full
VI
Full
Full
Full
Full
Full
Full
VI
VI
VI
VI
VI
VI
Rev. 0 | Page 6 of 48
DRVDD – 0.7
1.2
1
Typ
Max
Unit
0.4
12
3
V
V
µA
pF
2
1.5/2.3
1.9/2.7
0.7/0.7
1.0/1.0
ns
ns
ns
ns
Clock
cycles
AD9866
SERIAL PORT TIMING SPECIFICATIONS
Table 5. AVDD = 3.3 V ± 5%, DVDD = CLKVDD = DRVDD = 3.3 V ± 10%, unless otherwise noted
Parameter
WRITE OPERATION (See Figure 46)
SCLK Clock Rate (fSCLK)
SCLK Clock High (tHI)
SCLK Clock Low (tLOW)
SDIO to SCLK Setup Time (tDS)
SCLK to SDIO Hold Time (tDH)
SEN to SCLK Setup Time (tS)
SCLK to SEN Hold Time (tH)
READ OPERATION (See Figure 47 and Figure 48)
SCLK Clock Rate (fSCLK)
SCLK Clock High (tHI)
SCLK Clock Low (tLOW)
SDIO to SCLK Setup Time (tDS)
SCLK to SDIO Hold Time (tDH)
SCLK to SDIO (or SDO) Data Valid Time (tDV)
SEN to SDIO Output Valid to Hi-Z (tEZ)
Temp
Test Level
Min
Full
Full
Full
Full
Full
Full
Full
IV
IV
IV
IV
IV
IV
IV
14
14
14
0
14
0
Full
Full
Full
Full
Full
Full
Full
IV
IV
IV
IV
IV
IV
IV
Typ
Max
Unit
32
MHz
ns
ns
ns
ns
ns
ns
32
MHz
ns
ns
ns
ns
ns
ns
14
14
14
0
14
2
HALF-DUPLEX DATA INTERFACE (ADIO PORT) TIMING SPECIFICATIONS
Table 6. AVDD = 3.3 V ±5%, DVDD = CLKVDD = DRVDD = 3.3 V ±10%, unless otherwise noted
Parameter
READ OPERATION (See Figure 50)
Output Data Rate
Three-State Output Enable Time (tPZL)
Three-State Output Disable Time (tPLZ)
Rx Data Valid Time (tDV)
Rx Data Output Delay (tOD)
WRITE OPERATION (See Figure 49)
Input Data Rate (1× Interpolation)
Input Data Rate (2× Interpolation)
Input Data Rate (4× Interpolation)
Tx Data Setup Time (tDS)
Tx Data Hold Time (tDH)
Latch Enable Time (tEN)
Latch Disable Time (tDIS)
Temp
Test Level
Min
Full
Full
Full
Full
Full
II
II
II
II
II
Full
Full
Full
Full
Full
Full
Full
II
II
II
II
II
II
II
Rev. 0 | Page 7 of 48
Typ
Max
Unit
5
80
3
3
4
80
MSPS
ns
20
10
5
12.5
0
3
3
80
80
50
ns
ns
ns
MSPS
MSPS
MSPS
ns
ns
ns
ns
AD9866
FULL-DUPLEX DATA INTERFACE (Tx AND Rx PORT) TIMING SPECIFICATIONS
Table 7. AVDD = 3.3 V ± 5%, DVDD = CLKVDD = DRVDD = 3.3 V ± 10%, unless otherwise noted
Parameter
Tx PATH INTERFACE (See Figure 53)
Input Nibble Rate (2× Interpolation)
Input Nibble Rate (4× Interpolation)
Tx Data Setup Time (tDS)
Tx Data Hold Time (tDH)
Rx PATH INTERFACE (See Figure 54)
Output Nibble Rate
Rx Data Valid Time (tDV)
Rx Data Hold Time (tDH)
Temp
Test Level
Min
Full
Full
Full
Full
II
II
II
II
Full
Full
Full
II
II
II
Typ
Max
Unit
20
10
3
1
160
100
MSPS
MSPS
ns
ns
10
3
0
160
MSPS
ns
ns
Explanation of Test Levels
I: 100% production tested.
II: 100% production tested at 25°C and guaranteed by design and characterization at specified temperatures.
III: Sample tested only.
IV: Parameter is guaranteed by design and characterization testing.
V: Parameter is a typical value only.
VI: 100% production tested at 25°C and guaranteed by design and characterization for industrial temperature range.
Rev. 0 | Page 8 of 48
AD9866
ABSOLUTE MAXIMUM RATINGS
Table 8.
Parameter
ELECTRICAL
AVDD, CLKVDD Voltage
DVDD, DRVDD Voltage
RX+, RX−, REFT, REFB
IOUTP+, IOUTP−
IOUTN+, IOUTN−, IOUTG+,
IOUTG−
OSCIN, XTAL
REFIO, REFADJ
Digital Input and Output Voltage
Digital Output Current
ENVIRONMENTAL
Operating Temperature Range
(Ambient)
Maximum Junction Temperature
Lead Temperature (Soldering, 10 s)
Storage Temperature Range
(Ambient)
Rating
3.9 V max
3.9 V max
−0.3 V to AVDD + 0.3 V
−1.5 V to AVDD + 0.3 V
−0.3 V to 7 V
−0.3 V to CLVDD + 0.3 VS
−0.3 V to AVDD + 0.3 V
−0.3 V to DRVDD + 0.3 V
5 mA max
−40°C to +85°C
Stresses above those listed under the Absolute Maximum
Ratings may cause permanent damage to the device. This is a
stress rating only; functional operation of the device at these or
any other conditions above those indicated in the operational
section of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
THERMAL CHARACTERISTICS
Thermal Resistance: 64-lead LFCSP (4-layer board).
θJA = 24°C/W (paddle soldered to ground plane, 0 LPM air).
θJA = 30.8°C/W (paddle not soldered to ground plane,
0 LPM air).
125°C
150°C
−65°C to +150°C
ESD CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on
the human body and test equipment and can discharge without detection. Although this product features
proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy
electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance
degradation or loss of functionality.
Rev. 0 | Page 9 of 48
AD9866
DRVDD
DRVSS
PWR_DWN
CLKOUT2
DVDD
DVSS
CLKVDD
OSCIN
XTAL
CLKVSS
CONFIG
MODE
IOUT_P+
IOUT_P–
IOUT_N+
IOUT_G+
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
ADIO11/Tx[5]
1
48
AVSS
ADIO10/Tx[4]
2
47
AVSS
46
IOUT_N–
PIN 1
IDENTIFIER
3
ADIO8/Tx[2]
4
45
IOUT_G–
ADIO7/Tx[1]
5
44
AVSS
ADIO6/Tx[0]
6
43
AVDD
ADIO5/Rx[5]
7
42
REFIO
ADIO4/Rx[4]
8
41
REFADJ
ADIO3/Rx[3]
9
40
AVDD
ADIO2/Rx[2] 10
39
AVSS
11
38
RX+
ADIO0/Rx[0] 12
37
RX–
RXEN/RXSYNC 13
36
AVSS
TXEN/TXSYNC 14
35
AVDD
TXCLK/TXQUIET 15
34
AVSS
RXCLK 16
33
REFT
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
DRVSS
CLKOUT1
SDIO
SDO
SCLK
SEN
GAIN/PGA[5]
PGA[4]
PGA[3]
PGA[2]
PGA[1]
PGA[0]
RESET
AVSS
REFB
TOP VIEW
(Not to Scale)
DRVDD
ADIO1/Rx[1]
AD9866
Figure 2. Pin Configuration
Table 9. Pin Function Descriptions
Pin No.
1
2–5
6
7
8, 9
10
11
12
13
14
15
Mnemonic
ADIO11
Tx[5]
ADIO10–7
Tx[4–1]
ADIO6
Tx[0]
ADIO5
Rx[5]
ADIO4–3
Rx[4–3]
ADIO2
Rx[2]
ADIO1
Rx[1]
ADIO0
Rx[0]
RXEN
RXSYNC
TXEN
TXSYNC
TXCLK
TXQUIET
Mode1
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
HD
FD
Pin Function
MSB of ADIO Buffer
MSB of Tx Nibble Input
Bits 10–7 of ADIO Buffer
Bits 4–1 of Tx Nibble Input
Bit 6 of ADIO Buffer
LSB of Tx Nibble Input
Bit 5 of ADIO Buffer
MSB of Rx Nibble Output
Bits 4–3 of ADIO Buffer
Bits 4–3 of Rx Nibble Output
Bit 2 of ADIO Buffer
Bit 2 of Rx Nibble Output
Bit 1 of ADIO Buffer
Bit 1 of Rx Nibble Output
LSB of ADIO Buffer
LSB of Rx Nibble Output
ADIO Buffer Control Input
Rx Data Synchronization Output
Tx Path Enable Input
Tx Data Synchronization Input
ADIO Sample Clock Input
Fast TxDAC/IAMP Power-Down
Rev. 0 | Page 10 of 48
04560-0-002
ADIO9/Tx[3]
AD9866
Pin No.
16
Mnemonic
RXCLK
17, 64
18, 63
19
20
21
22
23
DRVDD
DRVSS
CLKOUT1
SDIO
SDO
SCLK
SEN
24
25–29
30
GAIN
PGA[5]
PGA[4–0]
RESET
31, 34, 36, 39 44, 47, 48
32, 33
35, 40, 43
37, 38
41
42
45
46
49
50
51
52
53
AVSS
REFB, REFT
AVDD
RX−, RX+
REFADJ
REFIO
IOUT_G−
IOUT_N−
IOUT_G+
IOUT_N+
IOUT_P−
IOUT_P+
MODE
54
55
56
57
58
59
60
61
62
CONFIG
CLKVSS
XTAL
OSCIN
CLKVDD
DVSS
DVDD
CLKOUT2
PWR_DWN
Mode1
HD
FD
Pin Function
ADIO Request Clock Input
Rx and Tx Clock Output at 2 × fADC
Digital Output Driver Supply Input
Digital Output Driver Supply Return
fDAC/N Clock Output (L = 1, 2, 4, or 8)
Serial Port Data Input/Output
Serial Port Data Output
Serial Port Clock Input
Serial Port Enable Input
FD
HD or FD
HD or FD
Tx Data Port (Tx[5:0]) Mode Select
MSB of PGA Input Data Port
Bits 4–0 of PGA Input Data Port
Reset Input (Active Low)
Analog Ground
ADC Reference Decoupling Nodes
Analog Power Supply Input
Receive Path − and + Analog Inputs
TxDAC Full-Scale Current Adjust
TxDAC Reference Input/Output
−Tx Amp Current Output_Sink
−Tx Mirror Current Output_Sink
+Tx Amp Current Output_Sink
+Tx Mirror Current Output_Sink
−TxDAC Current Output_Source
+TxDAC Current Output_Source
Digital Interface Mode Select Input
LOW = HD, HIGH = FD
Power-Up SPI Register Default Setting Input
Clock Osc./Synthesizer Supply Return
Crystal Osc. Inverter Output
Crystal Osc. Inverter Input
Clock Osc./Synthesizer Supply
Digital Supply Return
Digital Supply Input
fOSCIN/L Clock Output, (L = 1, 2, or 4)
Power-Down Input
1
HD = half-duplex mode; FD = full-duplex mode.
Rev. 0 | Page 11 of 48
AD9866
TYPICAL PERFORMANCE CHARACTERISTICS
Rx PATH TYPICAL PERFORMANCE CHARACTERISTICS
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = fADC = 50 MSPS, low-pass filter’s f−3 dB = 22 MHz, AIN = −1 dBFS,
RIN = 50 Ω, half- or full-duplex interface, default power bias settings
65
–10
–20
10.5
1MHz
5MHz
10MHz
15MHz
20MHz
62
59
SINAD (dBFS)
–30
–40
–50
–60
–70
9.5
56
9.0
53
8.5
50
8.0
47
7.5
44
7.0
–90
0
6.25
12.50
18.75
41
–6
25.00
0
6
12
FREQUENCY (MHz)
18
24
30
RxPGA GAIN (dB)
36
42
04560-0-006
–80
–100
6.5
48
Figure 6. SINAD/ENOB vs. RxPGA Gain and Frequency
Figure 3. Spectral Plot with 4k FFT of Input Sinusoid
with RxPGA = 0 dB and PIN = 9 dBm
–55
–30
1MHz
5MHz
10MHz
15MHz
20MHz
RBW = 12.2kHz
–40
–60
–50
–60
–65
–70
THD (dBc)
INPUT REFERRED SPECTRUM (dBm)
10.0
ENOB (Bits)
FUND = –1dBFS
SINAD = 61.9dBFS
ENOB = 10BITS
SNR = 64.5dBFS
THD = –65.4dBFS
SFDR = –64.9dBc (THIRD HARMONIC)
RBW = 12.21kHz
0
04560-0-003
REFERRED TO INPUT SPECTRUM (dBm)
10
–80
–90
–70
–75
–100
–110
0
5
10
15
FREQUENCY (MHz)
20
–85
–6
25
0
6
12
18
24
30
RxPGA GAIN (dB)
36
42
48
Figure 7. THD vs. RxPGA Gain and Frequency
Figure 4. Spectral Plot with 4k FFT of 84-Carrier DMT Signal
with PAR = 10.2 dB, PIN = −33.7 dBm, and RxPGA = 36 dB
66
04560-0-007
–130
–80
04493-0-041
–120
–50
65
63
–56
62
SINAD @ 3.14V
SINAD @ 3.3V
SINAD @ 3.46V
51
–80
48
–86
45
–21
–18
–15
–12
–9
–6
INPUT AMPLITUDE (dBFS)
–3
0
–92
56
53
–65
50
–70
47
–75
44
–6
(0dBFS = 2V p-p)
Figure 5. SINAD and THD vs. Input Amplitude and Supply
(fIN = 8 MHz, LPF f−3 dB = 26 MHz; Rx PGA = 0 dB)
–60
0
6
12
18
24
30
RxPGA GAIN (dB)
36
42
Figure 8. SINAD/THD Performance vs. RxPGA Gain
and Temperature ( fIN = 5 MHz)
Rev. 0 | Page 12 of 48
48
–80
THD (dBc)
–74
–55
THD @ +25°C
THD @ +85°C
THD @ –40°C
04560-0-008
THD @ 3.14V
THD @ 3.3V
THD @ 3.46V
SINAD (dBFS)
–68
THD (dBFS)
59
57
54
–50
–62
04560-0-005
60
SINAD (dBFS)
–45
SINAD @ +25°C
SINAD @ +85°C
SINAD @ –40°C
AD9866
Rx Path Typical Performance Characteristics:
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = fADC = 80 MSPS, low-pass filter’s f−3 dB = 30 MHz, AIN = −1 dBFS,
RIN = 50 Ω, half- or full-duplex interface, default power bias settings
–20
5MHz
10MHz
15MHz
20MHz
30MHz
62
59
SINAD (dBFS)
–30
–40
–50
–60
–70
10.0
9.5
56
9.0
53
8.5
50
8.0
47
7.5
44
7.0
ENOB (Bits)
–10
10.5
65
FUND = –1dBFS
SINAD = 62.4dBFS
ENOB = 10.1BITS
SNR = 63.4dBFS
THD = –69.3dBFS
SFDR = –70.5dBc (THIRD HARMONIC)
RBW = 19.53kHz
0
–90
–100
0
10
20
30
41
–6
40
0
6
12
FREQUENCY (MHz)
18
24
30
RxPGA GAIN (dB)
36
42
04560-0-012
–80
04560-0-009
REFERRED TO INPUT SPECTRUM (dBm)
10
6.5
48
Figure 12. SINAD/ENOB vs. RxPGA Gain and Frequency
Figure 9. Spectral Plot with 4k FFT of Input Sinusoid
with RxPGA = 0 dB and PIN = 9 dBm
–30
–55
RBW = 19.53kHz
–60
–50
–60
–65
THD (dBc)
–70
–80
–90
–70
–75
5MHz
10MHz
15MHz
20MHz
30MHz
–110
–120
–130
0
10
20
FREQUENCY (MHz)
30
–80
–85
–6
40
Figure 10. Spectral Plot with 4K FFT of 111-Carrier DMT Signal
with PAR = 11 dB, PIN = −33.7 dBm, LPF's f−3 dB = 32 MHz and RxPGA = 36 dB
6
12
18
24
30
RxPGA GAIN (dB)
42
48
Figure 13. THD vs. RxPGA Gain and Frequency
–50
–56
–68
THD @ 3.14V
THD @ 3.3V
THD @ 3.46V
54
–74
51
–80
48
–86
45
–21
–92
–18
–15
–12
–9
–6
INPUT AMPLITUDE (dBFS)
–3
0
SINAD (dBFS)
57
04560-0-011
–62
THD (dBFS)
61
60
–50
56
–55
53
–60
50
–65
47
–70
THD @ +25°C
THD @ +85°C
THD @ –40°C
44
41
–6
(0dBFS = 2V p-p)
Figure 11. SINAD and THD vs. Input Amplitude and Supply
(fIN = 8 MHz, LPF f−3 dB = 26 MHz; RxPGA = 0 dB)
59
0
6
12
18
24
30
RxPGA GAIN (dB)
–75
36
42
–80
48
THD (dBc)
–40
SINAD @ +25°C
SINAD @ +85°C –45
SINAD @ –40°C
65
SINAD @ 3.14V
SINAD @ 3.3V
SINAD @ 3.46V
63
SINAD (dBFS)
36
04560-0-014
66
0
04560-0-013
–100
04560-0-010
INPUT REFERRED SPECTRUM (dBm)
–40
Figure 14. SINAD/THD Performance vs. RxPGA Gain and Temperature
( fIN = 10 MHz)
Rev. 0 | Page 13 of 48
AD9866
Rx Path Typical Performance Characteristics:
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = fADC = 80 MSPS, low-pass filter’s f−3 dB = 30 MHz, AIN = −1 dBFS,
RIN = 50 Ω, half- or full-duplex interface, default power bias settings
63
–20
–54
62
–25
64.0
–56
61
63.5
–58
60
62.5
62.0
–62
–64
–35
SNR vs. MSPS @ 346VSUP
59
–40
SNR @ 3.13V
58
–45
THD @ 3.13V
THD @ 3.46V
57
–50
THD @ 3.3V
–66
56
–55
61.0
–68
55
–60
60.5
–70
54
–65
60.0
–72
30
0
5
10
15
20
INPUT FREQUENCY (MHz)
35
04560-0-015
61.5
53
20
30
76.6
14
56.6
12
54.7
10
43.8
8
–40°C
6
+85°C
21.9
4
+25°C
2
30
36
RxPGA GAIN (dB)
0
48
42
44
43
42
41
40
39
04560-0-019
16
SNR (dB)
87.5
24
38
0
20
30
40
50
60
70
80
Figure 19. SNR vs. Filter Cutoff Frequency
(50 MSPS; fIN = 5 MHz; AIN = −1 dB; RxPGA = 48 dB)
5
0.5
4
0.4
3
0.3
GAIN STEP ERROR (dB)
2
1
0
–1
–2
–4
0
6
12
18
24
30
36
42
0.2
0.1
0
–0.1
–0.2
–0.3
AD9865: GAIN STEP ERROR @ +25°C
AD9865: GAIN STEP ERROR @ +85°C
AD9865: GAIN STEP ERROR @ –40°C
–0.4
–0.5
–6
48
GAIN (dB)
0
6
12
18
24
30
36
42
RxPGA GAIN (dB)
Figure 17. Rx DC Offset vs. RxPGA Gain
Figure 20. RxPGA Gain Step Error vs. Gain (fIN = 10 MHz)
Rev. 0 | Page 14 of 48
04560-0-020
DEVICE 1
DEVICE 2
DEVICE 3
DEVICE 4
–3
04560-0-017
DC OFFSET (% of full-scale)
10
CUTOFF FREQUENCY (MHz)
Figure 16. Input Referred Integrated Noise and Noise Spectral Density
vs. RxPGA Gain (LPF f−3 dB = 26 MHz)
–5
–6
–70
80
70
45
NOISE SPECTRAL DENSITY (nV/ Hz)
18
04560-0-016
98.5
INTEGRATED NOISE (µV rms)
20
0
18
60
Figure 18. SNR and THD vs. Sample Rate and Supply
(LPF Disabled; RxPGA = 0 dB; fIN = 8 MHz)
109.4
10.9
50
INPUT FREQUENCY (MHz)
Figure 15. SNR and THD vs. Input Frequency and Supply
( LPF f−3 dB = 26 MHz; RxPGA = 0 dB)
32.8
40
THD (dBc)
–60
THD @ 3.14V
THD @ 3.3V
THD @ 3.47V
THD (dBc)
SNR (dBFS)
63.0
–30
SNR vs. MSPS @ 3.0VSUP
04560-0-018
SNR @ 3.14V
SNR @ 3.3V
SNR @ 3.47V
64.5
SNR (dBFS)
–52
65.0
48
AD9866
Rx Path Typical Performance Characteristics:
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = fADC = 50 MSPS, low-pass filter disabled, RxPGA = 0 dB, AIN = −1 dBFS,
RIN = 50 Ω, half- or full-duplex interface, default power bias settings
1408
2048
1280
1792
1152
1536
896
CODE
CODE
1024
1280
1024
768
640
768
04560-0-021
256
0
80
160
240
320
400
480
560
640
04560-0-024
512
512
384
256
720
0
80
160
240
320
TIME (ns)
Figure 21. RxPGA Settling Time −12 dB to +48 dB Transition for DC Input
(fADC = 50 MSPS, LPF Disabled)
640
720
–6dB GAIN
0dB GAIN
+6dB GAIN
–2
–4
FUNDAMENTAL (dB)
–6
–9
–12
–6
–8
–10
+18dB GAIN
+30dB GAIN
+42dB GAIN
–12
–14
04560-0-022
0
5
10
15
20
25
30
35
40
45
04560-0-025
–16
–15
–18
–20
50
0
5
10
15
INPUT FREQUENCY (MHz)
20
25
30
35
40
45
50
INPUT FREQUENCY (MHz)
Figure 25. Rx Low-Pass Filter Amplitude Response vs. RxPGA Gain
(LPF's f−3 dB = 33 MHz)
Figure 22. Rx Low-Pass Filter Amplitude Response vs. Supply (fADC = 50 MSPS,
f−3 dB = 33 MHz, RxPGA = 0 dB)
140
130
TxDAC ISOLATION @ 0dB
420
10
410
9
400
8
RESISTANCE (Ω)
390
110
100
90
7
RIN
380
6
370
5
360
4
CIN
350
3
CAPACITANCE (pF)
120
80
04560-0-023
IAMP ISOLATION @ 0dB
70
60
0
5
10
15
20
25
30
340
2
330
1
320
35
FREQUENCY (MHz)
5
15
25
35
45
55
65
75
85
95
FREQUENCY (MHz)
Figure 26. Rx Input Impedance vs. Frequency
Figure 23. Rx to Tx Full-Duplex Isolation @ 0 RxPGA Setting
(Note: ATTEN @ RxPGA = x dB = ATTEN @ RxPGA = 0 dB − RxPGA Gain)
Rev. 0 | Page 15 of 48
0
105
04493-0-026
AMPLITUDE RESPONSE (dB)
560
0
3.3V
3.0V
3.6V
–3
ATTEN @RxPGA = 0dB (dB)
480
Figure 24. RxPGA Settling Time for 0 dB to +5 dB Transition for DC Input
(fADC = 50 MSPS, LPF Disabled)
0
–18
400
TIME (ns)
AD9866
TxDAC PATH TYPICAL PERFORMANCE CHARACTERISTICS
10
0
0
–10
–10
–20
–20
–30
–30
dBm
10
–40
–50
–50
–60
–60
–70
–80
0
5
10
15
20
04560-0-030
–40
04493-0-072
dBm
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = 50 MSPS and 80 MSPS, RSET = 1.96 kΩ, 2:1 transformer coupled output
(see Figure 63) into 50 Ω load half-or full-duplex interface, default power bias settings
–70
–80
25
0
5
10
FREQUENCY (MHz)
–65
–65
–70
–70
IMD (dBFS)
–75
4dBm
7dBm
–80
10dBm
–85
35
40
–75
4dBm
7dBm
–80
2.5
5.0
7.5
10.0
12.5
15.0
17.5
–90
20.0
04560-0-031
04560-0-028
0
0
5
15
20
25
30
Figure 31. 2-Tone IMD Frequency Sweep vs. Peak Power
with fDATA = 80 MSPS, 2× Interpolation
–70
–70
SFDR (dBFS)
–65
–75
4dBm
7dBm
–80
10dBm
–75
4dBm
–80
7dBm
–85
0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
2-TONE CENTER FREQUENCY (MHz)
–90
04560-0-032
–85
(RELATIVE TO PEAK POWER)
–65
10dBm
10
2-TONE CENTER FREQUENCY (MHz)
04560-0-029
SFDR (dBFS)
30
–85
Figure 28. 2-Tone IMD Frequency Sweep vs. Peak Power
with fDATA = 50 MSPS, 4× Interpolation
(RELATIVE TO PEAK POWER)
25
10dBm
2-TONE CENTER FREQUENCY (MHz)
–90
20
Figure 30. Dual-Tone Spectral Plot of TxDAC's Output
(fDATA = 80 MSPS, 2× Interpolation, 10 dBm Peak Power, F1 = 27.1 MHz,
F2 = 28.7 MHz)
(RELATIVE TO PEAK POWER)
IMD (dBFS)
(RELATIVE TO PEAK POWER)
Figure 27. Dual-Tone Spectral Plot of TxDAC's Output
(fDATA = 50 MSPS, 4× Interpolation, 10 dBm Peak Power, F1 = 17 MHz,
F2 = 18 MHz)
–90
15
FREQUENCY (MHz)
0
5
10
15
20
25
30
2-TONE CENTER FREQUENCY (MHz)
Figure 29. 2-Tone Worst Spur Frequency Sweep vs. Peak Power
with fDATA = 50 MSPS, 4× Interpolation
Figure 32. 2-Tone Worst Spur Frequency Sweep vs. Peak Power
with fDATA = 80 MSPS, 2× Interpolation
Rev. 0 | Page 16 of 48
AD9866
TxDAC Path Typical Performance Characteristics:
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = 50 MSPS and 80 MSPS, RSET = 1.96 kΩ, 2:1 transformer coupled output
(see Figure 63) into 50 Ω load, half- or full-duplex interface, default power bias settings
–30
PAR = 11.4
RMS = –1.4dBm
–30
–40
–50
–50
dBm
–40
–60
–60
–70
–70
–80
–80
04560-0-033
dBm
–20
PAR = 11.4
RMS = –1.4
dBm
–90
–100
0
5
10
15
20
04493-0-081
–20
–90
–100
25
0
5
10
FREQUENCY (MHz)
15
20
25
Figure 33. Spectral Plot of 84-Carrier OFDM Test Vector
fDATA = 50 MSPS, 4× Interpolation)
PAR = 11.4
RMS = –1.4dBm
–40
–40
–50
–50
–60
–60
–70
–70
–80
–80
–90
0
25
50
75
100
125
150
175
04493-0-082
dBm
–30
04493-0-079
dBm
–30
–90
–100
200
0
20
40
FREQUENCY (MHz)
60
80
100
120
140
160
FREQUENCY (MHz)
Figure 34. Wideband Spectral Plot of 88-Subcarrier OFDM Test Vector
(fDATA = 50 MSPS, 4× Interpolation)
Figure 37. Wideband Spectral Plot of 111-Carrier OFDM Test Vector
(fDATA = 80 MSPS, 2× Interpolation)
105
100
100
95
2-TONE IMD
2-TONE IMD
85
80
SNR
75
70
04560-0-035
65
60
55
–24
–21
–18
–15
–12
–9
–6
–3
0
AOUT (dBFS)
85
80
75
SNR
70
65
04560-0-038
90
90
(RELATIVE TO PEAK POWER)
SNR AND 2-TONE IMD (dBFS)
95
(RELATIVE TO PEAK POWER)
40
–20
PAR = 11.4
RMS = –1.4dBm
SNR AND 2-TONE IMD (dBFS)
35
Figure 36. Spectral Plot of 111-Carrier OFDM Test Vector
(fDATA = 80 MSPS, 2× Interpolation)
–20
–100
30
FREQUENCY (MHz)
60
55
–24
–21
–18
–15
–12
–9
–6
–3
AOUT (dBFS)
Figure 35. SNR and SFDR vs. POUT
(fOUT = 12.55 MHz, fDATA = 50 MSPS, 4× Interpolation)
Figure 38. SNR and SFDR vs. POUT
(fOUT = 20 MHz, fDATA = 80 MSPS, 2× Interpolation)
Rev. 0 | Page 17 of 48
0
AD9866
IAMP PATH TYPICAL PERFORMANCE CHARACTERISTICS
AVDD = CLKVDD = DVDD = DRVDD = 3.3 V, fOSCIN = 50 MSPS, RSET = 1.58 kΩ, 1:1 transformer coupled output (see Figure 64
and Figure 65) into 50 Ω load, half- or full-duplex interface, default power bias settings
20
48
RBW = 2.3kHz
15
10
5
2.5MHz
46
5MHz
44
42
–10
–15
–20
OIP3 (dBm)
–25
–30
–35
10MHz
38
20MHz
15MHz
36
–40
–45
34
04493-0-084
–50
–55
–60
40
0
5
10
15
20
04493-0-087
dBm
0
–5
32
30
3.0
25
3.5
4.0
FREQUENCY (MHz)
4.5
5.0
VCM (V)
Figure 39. Dual-Tone Spectral Plot of IAMPN Output (IAMP Settings of
I = 12.5 mA, N = 4, G = 0, 2:1 Transformer into 75 Ω Loader, VCM = 4.8 V)
Figure 42. IOUTN Third-Order Intercept vs. Common-Mode Voltage
(IAMP Settings of I = 12.5 mA, N = 4, G = 0, 2:1 Transformer into 75 Ω Load)
0
42
PAR = 11.4
RMS = 10.3dBm
–10
40
2.5MHz
–20
38
OIP3 (dBm)
–40
–50
5MHz
36
10MHz
34
–60
15MHz
04493-0-085
–70
–80
32
0
5
10
15
20
20MHz
30
3.0
25
3.5
4.0
FREQUENCY (MHz)
4.5
5.0
VCM (V)
Figure 40. Spectral Plot of 84-Carrier OFDM Test Vector Using IAMPN in
Current Mode Configuration
(IAMP Settings of I = 10 mA, N = 4, G = 0; VCM = 4.8 V)
Figure 43. IOUTG Third-Order Intercept vs. Common-Mode Voltage
(IAMP Settings of I = 4.25 mA, N = 0, G = 6, 2:1 Transformer into 75 Ω Load)
0
PAR = 11.4
RMS = 9.8dBm
RBW = 10kHz
0
PAR = 11.4
RMS = 10.4dBm
–10
04493-0-088
dBm
–30
–10
–20
–20
–30
dBm
dBm
–30
–40
–40
–50
–50
–60
04493-0-089
–60
04493-0-086
–70
–70
–80
0
5
10
15
20
–80
25
0
5
10
15
20
25
FREQUENCY (MHz)
FREQUENCY (MHz)
Figure 41. Spectral Plot of 84-Carrier OFDM Test Vector Using IAMP in Voltage
Mode Configuration with AVDD = 5 V
(PBR951 Transistors, IAMP Settings of I = 6 mA, N = 2, G = 6)
Figure 44. Spectral Plot of 84-Carrier OFDM Test Vector Using IAMP in Voltage
Mode Configuration with AVDD = 3.3 V
(PBR951 Transistors, IAMP Settings of I = 6 mA, N = 2, G = 6)
Rev. 0 | Page 18 of 48
AD9866
SERIAL PORT
Table 10. SPI Register Mapping
Address
(Hex) 1
Bit
Breakdown
Power-Up Default Value
MODE = 0 (Half-Duplex)
Description
Width
CONFIG = 0
SPI PORT CONFIGURATION AND SOFTWARE RESET
0x00
(7)
4-Wire SPI
1
0
(6)
LSB First
1
0
(5)
S/W Reset
1
0
POWER CONTROL REGISTERS (via PWR_DWN pin)
0x01
(7)
1
0
Clock Syn.
0x02
(6)
(5)
(4)
(3)
(2)
(1)
(0)
(7)
TxDAC/IAMP
Tx Digital
REF
ADC CML
ADC
PGA Bias
RxPGA
CLK Syn.
(6)
TxDAC/IAMP
(5)
Tx Digital
(4)
REF
(3)
ADC CML
(2)
ADC
(1)
PGA Bias
(0)
RxPGA
HALF-DUPLEX POWER CONTROL
0x03
(7:3)
Tx OFF Delay
(2)
Rx _TXEN
(1)
Tx PWRDN
(0)
Rx PWRDN
0x06
CONFIG = 0
CONFIG = 1
Comments
0
0
0
0
0
0
0
0
0
Default SPI configuration is
3-wire, MSB first.
0
0
0
PWR_DWN = 0
Default setting is for all
blocks powered on.
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1*
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
5
1
1
1
0xFF
PLL CLOCK MULTIPLIER/SYNTHESIZER CONTROL
0x04
(5)
Duty Cycle Enable
1
0
(4)
fADC from PLL
1
0
0x05
MODE = 1 (Full-Duplex)
CONFIG = 1
0xFF
N/A
N/A
0
0
0
0
0
0
(3:2)
(1:0)
PLL Divide-N
PLL Multiplier-M
2
2
00
01
00
10*
00
01
00
01
(2)
(1)
(0)
(7:6)
(5)
(4)
(3:2)
(1)
(0)
OSCIN to RXCLK
Invert RXCLK
Disabled RXCLK
CLKOUT2 Divide
CLKOUT2 Invert
CLKOUT2 Disable
CLKOUT1 Divide
CLKOUT1 Invert
CLKOUT1 Disable
1
1
1
2
1
1
2
1
1
0
0
0
10
0
0
10
0
0
0
0
0
10
0
0
10
0
0
0
0
0
10
0
0
10
0
0
1*
0
0
10
0
1*
10
0
1*
Rev. 0 | Page 19 of 48
PWR_DWN = 1
Default setting* is for all
functional blocks powered
down except PLL.
*MODE = CONFIG = 1
Setting has PLL powered
down with OSCIN input
routed to RXCLK output.
Default setting is for TXEN
input to control power
on/off of Tx/Rx path.
Tx driver delayed by 31
1/fDATA clock cycles.
Default setting is Duty Cycle
Restore disabled, ADC CLK
from OSCIN input, and PLL
multiplier × 2 setting. *PLL
multiplier × 4 setting.
Full-duplex RXCLK normally
at nibble rate.
*Exception on power-up.
Default setting is CLKOUT2
and CLKOUT1 enabled with
divide-by-2.
*CLKOUT1 and CLKOUT2
disabled.
AD9866
Address
(Hex) 1
Bit
Breakdown
Power-Up Default Value
MODE = 0 (Half-Duplex)
MODE = 1 (Full-Duplex)
Description
Width
CONFIG = 0
CONFIG = 1
CONFIG = 0
CONFIG = 1
Comments
Initiate Offset Cal.
Rx Low Power
Rx Filter ON
1
1
1
0
0
1
0
1*
1
0
0
1
0
1*
1
Rx Filter Tuning
Cut-off Frequency
Tx/Rx PATH GAIN CONTROL
0x09
(6)
Use SPI Rx Gain
(5:0)
Rx Gain Code
8
0x80
0x61*
0x80
0x80
Default setting has LPF ON
and Rx path at nominal
power bias setting.
*Rx path to low power.
Refer to Low-Pass Filter
section.
1
6
0x00
0x00
0x00
0x00
0x0A
1
6
0x7F
0x7F
0x7F
0x7F
1
1
1
1
1
0
1
0
0
0
0
1
0
0
1**
0
1
0
1*
0
0
1
0
1*
0
Default setting is RxPGA
control active.
*Tx port with GAIN strobe
(AD9875/AD9876
compatible).
** 3-bit RxPGA gain map
(AD9975 compatible).
2
01
00
01
01
1
0
0
0
0
1
1
1
N/A
0
0
N/A
0
0
0
0
1
0
0
1
Default setting is 2×
interpolation with LPF
response. Data format is
straight binary for halfduplex and twos
complement for full-duplex
interface.
*Full-duplex only.
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0x00
0
0x00
0x44
0x44
Secondary path G1 = 0, 1, 2,
3, 4.
Primary path N = 0, 1, 2, 3, 4.
0x62
0x62
Secondary path stages:
G2 = 0 to 1.50 in 0.25 steps
and G3 = 0 to 6.
Rx PATH CONTROL
0x07
(5)
(4)
(0)
0x08
(7:0)
(6)
Use SPI Tx Gain
(5:0)
Tx Gain Code
Tx AND Rx PGA CONTROL
0x0B
(6)
PGA Code for Tx
(5)
PGA Code for Rx
(3)
Force GAIN strobe
(2)
Rx Gain on Tx Port
(1)
3-Bit RxPGA Port
Tx DIGITAL FILTER AND INTERFACE
0x0C
(7:6)
Interpolation
Factor
(4)
Invert
TXEN/TXSYNC
(2)
LS Nibble First*
(1)
TXCLK neg. edge
(0)
Twos complement
Rx INTERFACE AND ANALOG/DIGITAL LOOPBACK
0x0D
(7)
Analog Loopback
1
0
0
(6)
Digital Loopback*
1
0
0
(5)
Rx Port 3-State
1
N/A
N/A
(4)
1
0
0
Invert
RXEN/RXSYNC
(2)
LS Nibble First*
1
N/A
N/A
(1)
RXCLK neg. edge
1
0
0
(0)
Twos complement 1
0
0
DIGITAL OUTPUT DRIVE STRENGTH, TxDAC OUTPUT, AND REV ID
0x0E
(7)
1
0
0
Low Drive
Strength
(0)
TxDAC Output
1
0
0
0x0F
(3:0)
REV ID Number
4
0x00
0x00
Tx IAMP GAIN AND BIAS CONTROL
0x10
(7)
Select Tx Gain
1
0x44
0x44
(6:4)
G1
3
(2:0)
N
3
0x11
(6:4)
G2
3
0x62
0x62
(2:0)
G3
3
Rev. 0 | Page 20 of 48
Default setting is for
hardware Rx gain code via
PGA or Tx data port.
Default setting is for Tx gain
code via SPI control.
Data format is straight
binary for half-duplex and
twos complement for fullduplex interface.
Analog loopback: ADC Rx
data fed back to TxDAC.
Digital loopback: Tx input
data to Rx output port.
*Full-duplex only.
Default setting is for high
drive strength and IAMP
enabled.
AD9866
Address
(Hex) 1
0x12
0x13
Power-Up Default Value
Bit
Breakdown
Description
Width
(6:4)
(2:0)
(7:5)
(4:3)
(2:0)
Stand_Secondary
Stand_Primary
CPGA Bias Adjust
SPGA Bias Adjust
ADC Bias Adjust
3
3
3
2
4
MODE = 0 (Half-Duplex)
MODE = 1 (Full-Duplex)
CONFIG = 0
CONFIG = 1
CONFIG = 0
CONFIG = 1
Comments
0x01
0x01
0x01
0x01
Standing current of primary
and secondary path.
0x00
Current bias setting for Rx
path’s functional blocks.
Refer to Page 41.
0x00
0x00
0x00
1
Bits that are undefined should always be assigned a 0.
Table 11. SPI Registers Pertaining to SPI Options
REGISTER MAP DESCRIPTION
The AD9866 contains a set of programmable registers described
in Table 10 that can be used to optimize its numerous features,
interface options, and performance parameters from its default
register settings. Registers pertaining to similar functions have
been grouped together and assigned adjacent addresses to
minimize the update time when using the multibyte serial port
interface (SPI) read/write feature. Bits that are undefined within
a register should be assigned a 0 when writing to that register.
The default register settings were intended to allow some
applications to operate without the use of an SPI. The AD9866
can be configured to support a half- or full-duplex digital
interface via the MODE pin with each interface having two
possible default register settings determined by the setting of
the CONFIG pin.
For instance, applications that need to use only the Tx or Rx
path functionality of the AD9866 can configure it for a halfduplex interface (MODE = 0) and use the TXEN pin to select
between the Tx or Rx signal path with the unused path
remaining in a reduced power state. The CONFIG pin can be
used to select the default interpolation ratio of the Tx path and
RxPGA gain mapping.
SERIAL PORT INTERFACE (SPI)
The serial port of the AD9866 has 3- or 4-wire SPI capability
allowing read/write access to all registers that configure the
device’s internal parameters. Registers pertaining to the SPI are
listed in Table 11. The default 3-wire serial communication port
consists of a clock (SCLK), serial port enable (SEN), and a
bidirectional data (SDIO) signal. SEN is an active low control
gating read and write cycles. When SEN is high, SDO and SDIO
are three-stated. The inputs to SCLK, SEN, and SDIO contain a
Schmitt trigger with a nominal hysteresis of 0.4 V centered
about VDDH/2. The SDO pin remains three-stated in a 3-wire
SPI interface.
Address (Hex)
0x00
Bit
(7)
(6)
Description
Enable 4-wire SPI
Enable SPI LSB first
A 4-wire SPI can be enabled by setting the 4-wire SPI bit high,
causing the output data to appear on the SDO pin instead of on
the SDIO pin. The SDIO pin serves as an input only, throughout
the read operation. Note that the SDO pin is active only during
the transmission of data and remains three-stated at any other
time.
An 8-bit instruction header must accompany each read and
write operation. The instruction header is shown in Table 12.
The MSB is a R/W indicator bit with logic high indicating a
read operation. The next two bits, N1 and N0, specify the
number of bytes (one to four bytes) to be transferred during the
data transfer cycle. The remaining five bits specify the address
bits to be accessed during the data transfer portion. The data
bits immediately follow the instruction header for both read
and write operations.
Table 12. Instruction Header Information
MSB
17
R/W
16
N1
15
N0
14
A4
13
A3
12
A2
LSB
11
A1
10
A0
The AD9866 serial port can support both most significant bit
(MSB) first or least significant bit (LSB) first data formats.
Figure 45 illustrates how the serial port words are built for the
MSB first and LSB first modes. The bit order is controlled by the
SPI LSB first bit (Register 0, Bit 6). The default value is 0, MSB
first. Multibyte data transfers in MSB format can be completed
by writing an instruction byte that includes the register address
of the last address to be accessed. The AD9866 automatically
decrements the address for each successive byte required for the
multibyte communication cycle.
Rev. 0 | Page 21 of 48
AD9866
INSTRUCTION CYCLE
SEN
Figure 47 illustrates the timing for a 3-wire read operation to
the SPI port. After SEN goes low, data (SDIO) pertaining to the
instruction header is read on the rising edges of SCLK. A read
operation occurs if the read/not-write indicator is set high.
After the address bits of the instruction header are read, the
eight data bits pertaining to the specified register are shifted out
of the SDIO pin on the falling edges of the next eight clock
cycles. If a multibyte communication cycle is specified in the
instruction header, a similar process as previously described for
a multibyte SPI write operation applies. The SDO pin remains
three-stated in a 3-wire read operation.
DATA TRANSFER CYCLE
SCLK
SDATA
R/W N1
N2
A4
A3
A2
A1
A0
INSTRUCTION CYCLE
SEN
D1N D0N
D71 D61
DATA TRANSFER CYCLE
SCLK
A0
A1
A2
A3
A4
N2
N1
D6N D7N
R/W D01 D11
04560-0-045
SDATA
tS 1/fSCLK
SEN
Figure 45. SPI Timing, MSB First (Upper) and LSB First (Lower)
N1
N0
A0
D7
D6 D1
D0
A1
A0
D7
D6
D1
D0
tS 1/fSCLK
tLOW
tHI
SCLK
SDIO
tDH
R/W
tEZ
N1
A2
A1
A0
tEZ
tDV
SDO
Figure 46. SPI Write Operation Timing
Rev. 0 | Page 22 of 48
D7
D6
D1
D0
Figure 48. SPI 4-Wire Read Operation Timing
04560-0-048
tDS
tH
tDH
R/W
A2
SEN
04560-0-046
tDS
tEZ
N1
Figure 48 illustrates the timing for a 4-wire read operation to
the SPI port. The timing is similar to the 3-wire read operation
with the exception that data appears at the SDO pin, while the
SDIO pin remains high impedance throughout the operation.
The SDO pin is an active output only during the data transfer
phase and remains three-stated at all other times.
SCLK
SDIO
R/W
Figure 47. SPI 3-Wire Read Operation Timing
tLOW
tHI
tDH
SDIO
Figure 46 illustrates the timing requirements for a write
operation to the SPI port. After the serial port enable (SEN)
signal goes low, data (SDIO) pertaining to the instruction
header is read on the rising edges of the clock (SCLK). To
initiate a write operation, the read/not-write bit is set low. After
the instruction header is read, the eight data bits pertaining to
the specified register are shifted into the SDIO pin on the rising
edge of the next eight clock cycles. If a multibyte communication cycle is specified, the destination address is decremented
(MSB first) and another eight bits of data are shifted in. This
process repeats itself until all the bytes specified in the instruction header (N1, N0 bits) are shifted in. SEN must remain low
during the data transfer operation, only going high after the last
bit is shifted in.
tS 1/fSCLK
tDV
tDS
04560-0-047
When the SPI LSB first bit is set high, the serial port interprets
both instruction and data bytes LSB first. Multibyte data
transfers in LSB format can be completed by writing an
instruction byte that includes the register address of the first
address to be accessed. The AD9866 automatically increments
the address for each successive byte required for the multibyte
communication cycle.
SEN
tLOW
tHI
SCLK
AD9866
DIGITAL INTERFACE
The digital interface port is configurable for half-duplex or fullduplex operation by pin-strapping the MODE pin low or high,
respectively. In half-duplex mode, the digital interface port
becomes a 10-bit bidirectional bus called the ADIO port. In fullduplex mode, the digital interface port is divided into two 6-bit
ports called Tx[5:0] and Rx[5:0] for simultaneous Tx and Rx
operations. In this mode, data is transferred between the ASIC
and AD9866 in 6-bit nibbles. The AD9866 also features a
flexible digital interface for updating the RxPGA and TxPGA
gain registers via a 6-bit PGA port or Tx[5:0] port for fast
updates, or via the SPI port for slower updates. See the RXPGA
Control section.
data appears on the bus after a 6-clock-cycle delay due to the
internal FIFO delay. Note that Rx data is not latched back into
the Tx path, if TXEN is high during this interval with TXCLK
present. The ADIO Bus becomes three-stated once the RXEN
pin returns low. Figure 50 illustrates the receive path output
timing.
HALF-DUPLEX MODE
To add flexibility to the digital interface port, several programming options are available in the SPI registers. These options are
listed in Table 13. The default Tx and Rx data input formats are
straight binary, but can be changed to twos complement. The
default TXEN and RXEN settings are active high, but can be set
to opposite polarities, thus allowing them to share the same
control. In this case, the ADIO port can still be placed onto a
shared bus by disabling its input latch via the control signal, and
disabling the output driver via the SPI register. The clock timing
can be independently changed on the transmit and receive
paths by selecting either the rising or falling clock edge as the
validating/sampling edge of the clock. Lastly, the output driver’s
strength can be reduced for lower data rate applications.
The half-duplex mode functions as follows when the MODE
pin is tied low. The bidirectional ADIO port is typically shared
in burst fashion between the transmit path and receive path.
Two control signals, TXEN and RXEN, from a DSP (or digital
ASIC) control the bus direction by enabling the ADIO port’s
input latch and output driver, respectively. Two clock signals,
TXCLK and RXCLK, are used to latch the Tx input data and
clock the Rx output data, respectively. The ADIO port can also
be disabled by setting TXEN and RXEN low (default setting),
thus allowing it to be connected to a shared bus.
Internally, the ADIO port consists of an input latch for the Tx
path in parallel with an output latch with three-state outputs for
the Rx path. TXEN is used to enable the input latch; RXEN is
used to three-state the output latch. A five-sample-deep FIFO is
used on the Tx and Rx paths to absorb any phase difference
between the AD9866’s internal clocks and the externally
supplied clocks (TXCLK, RXCLK). The ADIO bus accepts input
data-words into the transmit path when the TXEN pin is high,
the RXEN pin is low, and a clock is present on the TXCLK pin,
as shown in Figure 49.
tDS
TX0
tDIS
tDH
TX1
TX2
TX3
TX4
04560-0-049
ADIO[9:0]
tEN
RXEN
Figure 49. Transmit Data Input Timing Diagram
The Tx interpolation filter(s) following the ADIO port can be
flushed with zeros, if the clock signal into the TXCLK pin is
present for 33 clock cycles after TXEN goes low. Note that the
data on the ADIO bus is irrelevant over this interval.
The output from the receive path is driven onto the ADIO bus
when the RXEN pin is high, and a clock is present on the
RXCLK pin. While the output latch is enabled by RXEN, valid
tPZL
tOD
tPLZ
tVT
ADIO[9:0]
RX0
RX1
RX2
RX3
04560-0-050
RXEN
Figure 50. Receive Data Output Timing Diagram
Table 13. SPI Registers for Half-Duplex Interface
Address (Hex)
0x0C
0x0D
0x0E
TXCLK
TXEN
RXCLK
Bit
(4)
(1)
(0)
(5)
(4)
(1)
(0)
(7)
Description
Invert TXEN
TXCLK negative edge
Twos complement
Rx port three-state
Invert RXEN
RXCLK negative edge
Twos complement
Low digital drive strength
The half-duplex interface can be configured to act like a slave or
a master to the digital ASIC. An example of a slave configuration is shown in Figure 51. In this example, the AD9866 accepts
all the clock and control signals from the digital ASIC. Because
the sampling clocks for the DAC and ADC are derived internally from the OSCIN signal, it is required that the TXCLK and
RXCLK signals be at exactly the same frequency as the OSCIN
signal. The phase relationships among the TXCLK, RXCLK, and
OSCIN signals can be arbitrary. If the digital ASIC cannot
provide a low jitter clock source to OSCIN, consider using the
AD9866 to generate the clock for its DAC and ADC and pass
the desired clock signal to the digital ASIC via CLKOUT1 or
CLKOUT2.
Rev. 0 | Page 23 of 48
AD9866
AD9866
12
ADIO
[11:0]
Tx/Rx
Data[11:0]
RXEN
RXEN
TXEN
TXEN
DAC_CLK
TXCLK
ADC_CLK
RXCLK
CLKOUT
OSCIN
FROM
Rx ADC
04560-0-051
12
TO
Tx DIGITAL
FILTER
Figure 51. Example of a Half -Duplex Digital Interface
with AD9866 Serving as the Slave
Figure 52 shows a half-duplex interface with the AD9866 acting
as the master, generating all the required clocks. CLKOUT1
provides a clock equal to the bus data rate that is fed to the
ASIC as well as back to the TXCLK and RXCLK inputs. This
interface has the advantage of reducing the digital ASIC’s pin
count by three. The ASIC needs only to generate a bus control
signal that controls the data flow on the bidirectional bus.
DIGITAL ASIC
AD9866
ADIO
[11:0]
Tx/Rx
Data[11:0]
12
12
than the nibble rate. Therefore, the 2× or 4× interpolation filter
must be used with a full-duplex interface.
The AD9866 acts as the master, providing RXCLK as an output
clock that is used for the timing of both the Tx[5:0] and Rx[5:0]
ports. RXCLK always runs at the nibble rate and can be inverted
or disabled via an SPI register. Because RXCLK is derived from
the clock synthesizer, it remains active, provided that this
functional block remains powered on. A buffered version of the
signal appearing at OSCIN can also be directed to RXCLK by
setting Bit 2 of Reg. 0x05. This feature allows the AD9866 to be
completely powered down (including the clock synthesizer)
while serving as the master.
The Tx[5:0] port operates in the following manner with the SPI
register default settings. Two consecutive nibbles of the Tx data
are multiplexed together to form a 10-bit data-word in twos
complement format. The clock appearing on the RXCLK pin is
a buffered version of the internal clock used by the Tx[5:0]
port’s input latch with a frequency that is always twice the ADC
sample rate (2 × fADC). Data from the Tx[5:0] port is read on the
rising edge of this sampling clock, as illustrated in the timing
diagram shown in Figure 53.
tDS
TO
Tx DIGITAL
FILTER
RXCLK
tHD
FROM
Rx ADC
TxSYNC
RXEN
Tx[5:0]
Tx0LSB
Tx1MSB
Tx1LSB
Tx2MSB
Tx3LSB
Tx
2 LSB
Tx3MSB
TXEN
BUS_CTR
TXCLK
04560-0-053
DIGITAL ASIC
Figure 53. Tx[5:0] Port Full-Duplex Timing Diagram
RXCLK
CLKIN
CLKOUT1
FROM
CRYSTAL
OR MASTER CLK
04560-0-052
OSCIN
Figure 52. Example of a Half -Duplex Digital Interface
with AD9866 Serving as the Master
FULL-DUPLEX MODE
The full-duplex mode interface is selected when the MODE pin
is tied high. It can be used for full- or half-duplex applications.
The digital interface port is divided into two 6-bit ports called
Tx[5:0] and Rx[5:0], allowing simultaneous Tx and Rx operations for full-duplex applications. In half-duplex applications,
the Tx[5:0] port can also be used to provide a fast update of the
RxPGA (AD9876 backward compatible) during an Rx
operation. This feature is enabled by default and can be used to
reduce the required pin count of the ASIC (refer to RxPGA
Control section for more detail).
In either application, Tx and Rx data are transferred between
the ASIC and AD9866 in 6-bit (or 5-bit) nibbles at twice the
internal input/output word rates of the Tx interpolation filter
and ADC. Note that the TxDAC update rate must not be less
The TXSYNC signal is used to indicate to which word a nibble
belongs. The first nibble of every word is read while TXSYNC is
low as the most significant nibble. The second nibble of that
same word is read on the following TXSYNC high level as the
least significant nibble. If TXSYNC is low for more than one
clock cycle, the last transmit data is read continuously until
TXSYNC is brought high for the second nibble of a new
transmit word. This feature can be used to flush the interpolator
filters with zeros. Note that the GAIN signal must be kept low
during a Tx operation.
The Rx[5:0] port operates in the following manner with the SPI
register default settings. Two consecutive nibbles of the Rx data
are multiplexed together to form a 10-bit data-word in twos
complement format. The Rx data is valid on the rising edge of
RXCLK, as illustrated in the timing diagram shown in
Figure 54. The RXSYNC signal is used to indicate to which
word a nibble belongs. The first nibble of every word is
transmitted while RXSYNC is low as the most significant
nibble. The second nibble of that same word is transmitted on
the following RXSYNC high level as the least significant nibble.
Rev. 0 | Page 24 of 48
AD9866
Rx1MSB
Rx1LSB
Rx2MSB
Rx3LSB
Rx3MSB
Figure 54. Full-Duplex Rx Port Timing
To add flexibility to the full-duplex digital interface port,
several programming options are available in the SPI registers.
These options are listed in Table 14. The timing for the Tx[5:0]
and/or Rx[5:0] ports can be independently changed by selecting
either the rising or falling clock edge as the sampling/validating
edge of the clock. Inverting RXCLK (via Bit 1 or Reg. 0x0D)
affects both the Rx and Tx interface, because they both use
RXCLK.
OPTIONAL
0x0B
0x0C
0x0D
0x0E
Bit
(2)
Description
OSCIN to RXCLK
(1)
(0)
(2)
(4)
(3)
(2)
(1)
(0)
(5)
(4)
(3)
(2)
(1)
(0)
(7)
Invert RXCLK
Disable RXCLK
Rx gain on Tx port
Invert TXSYNC
Tx 5/5 nibble
LS nibble first
TXCLK negative edge
Twos complement
Rx port three-state
Invert RXSYNC
Rx 5/5 nibble
LS nibble first
RXCLK negative edge
Twos complement
Low drive strength
6
GAIN
Tx[5:0]
Tx Data[5:0]
Rx[5:0]
Table 14. SPI Registers for Full-Duplex Interface
Address (Hex)
0x05
AD9865/AD9866
DIGITAL ASIC
Rx Data[5:0]
RX_SYNC
RXSYNC
TX_SYNC
TXSYNC
CLKIN
10/12
10/12
TO
Rx PGA
TO
Tx DIGITAL
FILTER
FROM
Rx ADC
RXCLK
CLKOUT1
CLKOUT2
OSCIN
FROM
CRYSTAL
OR MASTER CLK
04560-0-055
Rx0LSB
MUX
Rx[5:0]
04560-0-054
tDV
RxSYNC
Figure 55 shows a possible digital interface between an ASIC
and the AD9866. The AD9866 serves as the master generating
the required clocks for the ASIC. This interface requires that the
ASIC reserve 16 pins for the interface, assuming a 6-bit nibble
width and the use of the Tx port for RxPGA gain control. Note
that the ASIC pin allocation can be reduced by 3, if a 5-bit
nibble width is used and the gain (or gain strobe) of the RxPGA
is controlled via the SPI port.
DEMUX
tDH
RXCLK
Figure 55. Example of a Full-Duplex Digital Interface
with Optional RxPGA Gain Control via Tx[5:0]
RxPGA CONTROL
The default Tx and Rx data input formats are twos complement,
but can be changed to straight binary. The default TXSYNC and
RXSYNC settings can be changed such that the first nibble of
the word appears while TXSYNC, RXSYNC, or both are high.
Also, the least significant nibble can be selected as the first
nibble of the word (LS nibble first). The output driver strength
can also be reduced for lower data rate applications.
The AD9866 contains a digital PGA in the Rx path that is used
to extend the dynamic range. The RxPGA can be programmed
over a −12 dB to +48 dB with 1 dB resolution using a 6-bit
word, and with a 0 dB setting corresponding to a 2 V p-p input
signal. The 6-bit word is fed into a LUT that is used to distribute
the desired gain over three amplification stages within the Rx
path. Upon power-up, the RxPGA gain register is set to its
minimum gain of −12 dB. The RxPGA gain mapping is shown
in Figure 56. Table 15 lists the SPI registers pertaining to the
RxPGA.
Rev. 0 | Page 25 of 48
AD9866
48
tSU
42
RXCLK
36
24
Tx [5:0]
GAIN
18
12
04560-0-057
GAIN (dB)
tHD
Tx SYNC
30
GAIN
6
Figure 57. Updating RxPGA via Tx[5:0] in Full-Duplex Mode
0
Updating the RxPGA (or TxPGA) via the PGA[5:0] port is an
option for both the half-duplex3 and full-duplex interfaces. The
PGA port consists of an input buffer that passes the 6-bit data
appearing at its input directly to the RxPGA (or TxPGA) gain
register with no gating signal required. Bit 5 or Bit 6 of Reg.
0x0B is used to select whether the data updates the RxPGA or
TxPGA gain register. In applications that switch between
RxPGA and TxPGA gain control via PGA[5:0], be careful that
the RxPGA (or TxPGA) is not inadvertently loaded with the
wrong data during a transition. In the case of an RxPGA to
TxPGA transition, first deselect the RxPGA gain register, update
the PGA[5:0] port with the desired TxPGA gain setting, and
then select the TxPGA gain register.
04560-0-056
–6
–12
0
6
12
24
30
36
42
48
54
60
18
6-BIT DIGITAL WORD-DECIMAL EQUIVALENT
66
Figure 56. Digital Gain Mapping of RxPGA
Table 15. SPI Registers RxPGA Control
Address (Hex)
0x09
0x0B
Bit
(6)
(5:0)
(6)
(5)
(3)
(2)
(1)
Description
Enable RxPGA update via SPI
RxPGA gain code
Select TxPGA via PGA[5:0]
Select RxPGA via PGA[5:0]
Enable software GAIN strobe –
full-duplex
Enable RxPGA update via Tx[5:0] –
full-duplex
3-bit RxPGA gain mapping –
half-duplex
The RxPGA gain register can be updated via the Tx[5:0] port,
the PGA[5:0] port, or the SPI port. The first two methods allow
fast updates of the RxPGA gain register and should be
considered for digital AGC functions requiring a fast closedloop response. The SPI port allows direct update and readback
of the RxPGA gain register via Reg. 0x09 with an update rate
limited to 1.6 MSPS (with SCLK = 32 MHz). Note that Bit 6 of
Reg. 0x09 must be set for a read or write operation.
Updating the RxPGA via the Tx[5:0] port is an option only in
full-duplex mode.1 In this case, a high level on the GAIN pin2
with Tx SYNC low programs the PGA setting on either the
rising edge or falling edge of RXCLK, as shown in Figure 57.
The GAIN pin must be held high, TxSYNC must be held low,
and GAIN data must be stable for one or more clock cycles to
update the RxPGA gain setting. A low level on the GAIN pin
enables data to be fed to the digital interpolation filter. This
interface should be considered when upgrading existing designs
from the AD9876 MxFE product or half-duplex applications
trying to minimize an ASIC’s pin count.
1
2
Default setting for full-duplex mode (MODE = 1) .
The GAIN strobe can also be set in software via Reg. 0x0B, Bit 3 for
continuous updating. This eliminates the requirement for external GAIN
signal, reducing the ASIC pin count by 1.
The RxPGA also offers an alternative 3-bit word gain mapping
option4 that provides a −12 dB to +36 dB span in 8 dB increments as shown in Table 16. The 3-bit word is directed to
PGA[5:3] with PGA[5] being the MSB. This feature is backward
compatible with the AD9975 MxFE and allows direct interfacing to the CX11647 or INT5130 HomePlug 1.0 PHYs.
Table 16. PGA Timing for AD9975 Backward Compatible
Mode
Digital Gain Setting
PGA[5:3]
000
001
010
011
100
101
110
111
3
4
Decimal
0
1
2
3
4
5
6
7
Default setting for half-duplex mode (MODE = 0).
Default setting for MODE = 0 and CONFIG =1.
Rev. 0 | Page 26 of 48
Gain (dB)
−12
−12
−4
4
12
20
28
36
AD9866
TxPGA CONTROL
0
–1
–2
–3
–4
–5
–6
–7
–8
–9
–10
–11
–12
–13
–14
–15
–16
–17
–18
–19
–20
Table 17. SPI Registers TxPGA Control
Address (Hex)
0x0A
0x0B
TxDACs IOUTP OUTPUT
HAS 7.5dB RANGE
0x0E
IAMPs IOUTN AND IOUTG
OUTPUTS HAS 19.5dB RANGE
0
8
16
24
32
40
04560-0-058
Tx ATTENUATION (dBFS)
The AD9866 also contains a digital PGA in the Tx path distributed between the TxDAC and IAMP. The TxPGA is used to
control the peak current from the TxDAC and IAMP over a
7.5 dB and 19.5 dB span, respectively, with 0.5 dB resolution. A
6-bit word is used to set the TxPGA attenuation according to
the mapping shown in Figure 58. The TxDAC gain mapping is
applicable only when Bit 0 of Reg. 0x0E is set, and only the
4 LSBs of the 6-bit gain word are relevant.
The TxPGA register can be updated via the PGA[5:0] port or
SPI port. The first method should be considered for fast updates
of the TxPGA register. Its operation is similar to the description
in the RxPGA Control section. The SPI port allows direct
update and readback of the TxPGA register via Reg. 0x0A with
an update rate limited to 1.6 MSPS (SCLK = 32 MHz). Bit 6 of
Reg 0x0A must be set for a read or write operation. Table 17
lists the SPI registers pertaining to the TxPGA. The TxPGA
control register default setting is for minimum attenuation
(0 dBFS) with the PGA[5:0] port disabled for Tx gain control.
48
56
64
6-BIT DIGITAL CODE (Decimal Equivalent)
Figure 58. Digital Gain Mapping of TxPGA
Rev. 0 | Page 27 of 48
Bit
(6)
(5:0)
(6)
(5)
(0)
Description
Enable TxPGA update via SPI
TxPGA gain code
Select TxPGA via PGA[5:0]
Select RxPGA via PGA[5:0]
TxDAC output (IAMP disabled)
AD9866
TRANSMIT PATH
0 TO –12dB
AD9865/AD9866
04560-0-059
TXEN/SYNC
TXCLK
2.0
–10
1.5
–20
1.0
0.5
–30
PASS BAND
–40
–50
–1.0
–70
–1.5
–80
–2.0
–90
0
DIGITAL INTERPOLATION FILTERS
Bits [7:6]
00
01
10
11
0.25
0.50
0.75
1.00
1.25
1.50
1.75
NORMALIZED FREQUENCY (Relative to fDATA)
2.5
10
WIDE BAND
Interpolation Factor
4
2
1 (half-duplex only)
Do not use
0
2.0
–10
1.5
–20
1.0
0.5
–30
PASS BAND
0
–40
–1.0dB @ 0.45 fDATA
–50
–0.5
–60
–1.0
–70
–1.5
–80
–2.0
–90
0
The interpolation filter consists of two cascaded half-band filter
stages with each stage providing 2× interpolation. The first stage
filter consists of 43 taps. The second stage filter, operating at the
higher data rate, consists of 11 taps. The normalized wide band
and pass-band filter responses (relative fDATA) for the 2× and
4× low-pass interpolation filters are shown in Figure 60 and
Figure 61, respectively. Note that these responses also include
the inherent sinc(x) from the TxDAC reconstruction process
and can be used to estimate any post analog filtering
requirements.
–2.5
2.00
Figure 60. Frequency Response of 2× Interpolation Filter
(Normalized to fDATA)
WIDE BAND RESPONSE (dB)
Table 18. Interpolation Factor Set via SPI Reg. 0x0C
–0.5
–1.0dB @ 0.441 fDATA
–60
Figure 59. Functional Block Diagram of Tx Path
The input data from the Tx port can be fed into a selectable
2×/4× interpolation filter or directly into the TxDAC (for a halfduplex only). The interpolation factor for the digital filter is set
via SPI Reg. 0x0C with the settings shown in Table 18. The
maximum input word rate, fDATA, into the interpolation filter is
80 MSPS; the maximum DAC update rate is 200 MSPS. Therefore, applications with input word rates at or below 50 MSPS
can benefit from 4× interpolation, while applications with input
word rates between 50 MSPS and 80 MSPS can benefit from
2× interpolation.
0
PASS BAND RESPONSE (dB)
0 TO –7.5dB
IAMP
0
04560-0-060
TxDAC
WIDE BAND
0.5
1.0
1.5
2.0
2.5
3.0
3.5
NORMALIZED FREQUENCY (Relative to fDATA)
–2.5
4.0
PASS BAND RESPONSE (dB)
2-4X
2.5
10
04560-0-061
ADIO[11:6]/
Rx[5:0]
10
IOUT_G+
IOUT_N+
IOUT_N–
IOUT_G–
The pipeline delays of the 2× and 4× filter responses are 21.5
and 24 clock cycles, respectively, relative to fDATA. The filter delay
is also taken into consideration for applications configured for a
half-duplex interface with the half-duplex power-down mode
enabled. This feature allows the user to set a programmable
delay that powers down the TxDAC and IAMP only after the
last Tx input sample has propagated through the digital filter.
See the Power Control section for more details.
WIDE BAND RESPONSE (dB)
ADIO[11:6]/
Tx[5:0]
IOUT_P–
IOUT_P+
The AD9866 (or AD9865) transmit path consists of a selectable
digital 2×/4× interpolation filter, a 12-bit (or 10-bit) TxDAC,
and a current-output amplifier (IAMP), as shown in Figure 59.
Note that the additional two bits of resolution offered by the
AD9866 (vs. the AD9865) result in a 10 dB to 12 dB reduction
in the pass-band noise floor. The digital interpolation filter
relaxes the Tx analog filtering requirements by simultaneously
reducing the images from the DAC reconstruction process
while increasing the analog filter’s transition band. The digital
interpolation filter can also be bypassed, resulting in lower
digital current consumption.
Figure 61. Frequency Response of 4× Interpolation Filter
(Normalized to fDATA)
TxDAC AND IAMP ARCHITECTURE
The Tx path contains a TxDAC along with a current amplifier,
IAMP. The TxDAC reconstructs the output of the interpolation
filter and sources a differential current output that can either be
directed to an external load or fed into the IAMP for further
amplification. The TxDAC’s and IAMPS’s peak current outputs
are digitally programmable over a 0 to −7.5 dB and 0 to −19.5
dB range, respectively, in 0.5 dB increments. Note that this
assumes default register settings for Reg. 0x10 and Reg. 0x11.
Rev. 0 | Page 28 of 48
AD9866
Applications demanding the highest spectral performance
and/or lowest power consumption can use the TxDAC output
directly. The TxDAC is capable of delivering a peak signal
power-up to 10 dBm while maintaining respectable linearity
performance, as shown in Figure 27 through Figure 38. For
power-sensitive applications requiring the highest Tx power
efficiency, the TxDAC’s full-scale current output can be reduced
to as low as 2 mA and its load resistors sized to provide a
suitable voltage swing that can be amplified by a low power opamp-based driver.
Most applications requiring higher peak signal powers (up to
23 dBm) should consider using the IAMP. The IAMP can be
configured as a current source for loads having a well defined
impedance (50 Ω or 75 Ω systems) or a voltage source (with the
addition of a pair of npn transistors) for poorly defined loads
having varying impedance (such as power lines).
G × (I+∆I)
G × (I–∆I)
IOUTG–
I
IOUTG+
±∆IS
IOUTN–
REFADJ
IOUTN+
TxDAC
I
N × (I–∆I)
N × (I+∆I)
Figure 62 shows the equivalent schematic of the TxDAC and
IAMP. The TxDAC provides a differential current output
appearing at IOUTP+ and IOUTP−. It can be modeled as a
differential current source generating a signal-dependent ac
current, when ∆IS has a peak current of I along with two dc
current sources, sourcing a standing current equal to I. The fullscale output current, IOUTFS, is equal to the sum of these
standing current sources (IOUTFS = 2*I).
REFIO
RSET 0.1µF
IOUTP+
I – ∆I
IOUTP–
IOFF1
IOFF1
xN
IOFF2
IOFF2
xN
xG
xG
04560-0-062
I + ∆I
IAMP
Figure 62. Equivalent Schematic of TxDAC and IAMP
The value of I is determined by the RSET value at the REFADJ
pin along with the Tx path’s digital attenuation setting. With
0 dB attenuation, the value of I is
I = 16 ∗ (1.23 / R SET )
Equation 1.
For example, an RSET value of 1.96 kΩ results in I equal to
10.0 mA with IOUTFS equal to 20.0 mA. Note that the REFIO
pin provides a nominal band gap reference voltage of 1.23 V
and should be decoupled to analog ground via a 0.1 µF
capacitor.
The differential current output of the TxDAC is always connected to the IOUTP pins, but can be directed to the IAMP by
setting Bit 0 of Reg 0x0E. As a result, the IOUTP pins must
remain completely open, if the IAMP is to be used. The IAMP
contains two sets of current mirrors that are used to replicate
the TxDAC’s current output with a selectable gain. The first set
of current mirrors is designated as the primary path, providing
a gain factor of N that is programmable from 0 to 4 in steps of 1
via Bits 2:0 of Reg. 0x10 with a default setting of N = 4. Bit 7 of
this register must be set to overwrite the default settings of this
register. This differential path exhibits the best linearity
performance (see Figure 42) and is available at the IOUTN+
and IOUTN− pins. The maximum peak current per output is
100 mA and occurs when the TxDAC’s standing current, I, is set
for 12.5 mA (IOUTFS = 25 mA).
The second set of current mirrors is designated as the secondary path providing a gain factor of G that is programmable
from 0 to 36 via Bits 6:4 of Reg. 0x10 and Bits 6:0 of Reg. 0x11
with a default setting of G = 12. This differential path is
intended to be used in the voltage mode configuration to bias
the external npn transistors, because it exhibits degraded
linearity performance (see Figure 43) relative to the primary
path . It is capable of sinking up to 180 mA of peak current into
either its IOUTG+ or IOUTG− pins. The secondary path
actually consists of 3 gain stages (G1, G2, and G3), which are
individually programmable as shown in Table 19. While many
permutations may exist to provide a fixed gain of G, the
linearity performance of a secondary path remains relatively
independent of the various individual gain settings that are
possible to achieve a particular overall gain factor.
Both sets of mirrors sink current, because they originate from
NMOS devices. Therefore, each output pin requires a dc current
path to a positive supply. Although the voltage output of each
output pin can swing between 0.5 and 7 V, optimum ac
performance is typically achieved by limiting the ac voltage
swing with a dc bias voltage set between 4 to 5 V. Lastly, both the
standing current, I, and the ac current, ∆IS, from the TxDAC are
amplified by the gain factor (N and G) with the total standing
current drawn from the positive supply being equal to
2 ∗ (N + G ) ∗ I
Programmable current sources IOFF1 and IOFF2 via Reg. 0x12 can
be used to improve the primary and secondary path mirrors’
linearity performance under certain conditions by increasing
their signal-to-standing current ratio. This feature provides a
marginal improvement in distortion performance under large
signal conditions when the peak ac current of the reconstructed
waveform frequently approaches the dc standing current within
the TxDAC (0 to −1 dBFS sine wave) causing the internal
mirrors to turn off. However, the improvement in distortion
performance diminishes as the crest factor (peak-to-rms ratio)
of the ac signal increases. Most applications can disable these
Rev. 0 | Page 29 of 48
AD9866
1:1
current sources (set to 0 mA via Reg. 0x12) to reduce the
IAMP’s current consumption.
0x11
(7)
(6:4)
(3)
(2:0)
0x12
(6:4)
(2:0)
Tx PROGRAMMABLE GAIN CONTROL
TxPGA functionality is also available to set the peak output
current from the TxDAC or IAMP. The TxDAC and IAMP are
digitally programmable via the PGA[5:0] port or SPI over a
0 dB to −7.5 dB and 0 dB to −19.5 dB range, respectively, in
0.5 dB increments.
The TxPGA can be considered as two cascaded attenuators with
the TxDAC providing 7.5 dB range in 0.5 dB increments, and
the IAMP providing 12 dB range in 6 dB increments. As a result,
the IAMP’s composite 19.5 dB span is valid only if Reg. 0x10
remains at its default setting of 0x44. Modifying this register
setting corrupts the LUT and results in an invalid gain mapping.
TxDAC OUTPUT OPERATION
The differential current output of the TxDAC can be directed to
the IOUTP+ and IOUTP− pins by setting Bit 0 of Reg. 0x0E.
Any load connected to these pins must be ground referenced to
provide a dc path for the current sources. Figure 63 shows the
outputs of the TxDAC driving a doubly terminated 1:1 transformer with its center-tap tied to ground. The peak-to-peak
voltage, V p-p, across RL (and IOUT+ to IOUT−) is equal to
2*I*(RL//RS). With I = 10 mA and RL = RS = 50 Ω, V p-p is equal
to 0.5 V with 1 dBm of peak power being delivered to RL and
1 dBm being dissipated in RS.
0 TO –7.5dB
IOUT_P–
IOUT_P+
TxDAC
IOUTN+
IOUTG+
IAMP
0 TO –12dB
IOUTN–
IOUTG–
04560-0-063
(3)
(2:0)
Description
TxDAC output
Enable current mirror gain settings
Secondary path first stage gain of 0
to 4 with ∆ = 1
Not used
Primary path NMOS gain of 0 to 4
with ∆ = 1
Don’t care
Secondary path second stage gain of
0 to 1.5 with ∆ = 0.25
Not used
Secondary path third stage gain of 0
to 5 with ∆ = 1
IOFF2, secondary path standing
current
IOFF1, primary path standing current
REFIO
Bit
(0)
(7)
(6:4)
RSET
REFADJ
Address (Hex)
0x0E
0x10
RL
RS
0.1µF
Table 19. SPI Registers for TxDAC and IAMP
Figure 63. TxDAC Output Directly via Center-Tap Transformer
The TxDAC is capable of delivering up to 10 dBm peak power
to a load, RL. To increase the peak power for a fixed standing
current, one must increase V p-p across IOUTP+ and
−IOUTP− by increasing one or more of the following parameters: RS, RL (if possible), and/or the turns ratio, N, of transformer.
For example, the removal of RS and the use of a 2:1 impedance
ratio transformer in the previous example results in 10 dBm of
peak power capabilities to the load. Note that increasing the
power output capabilities of the TxDAC reduces the distortion
performance due to the higher voltage swings seen at IOUTP+
and IOUTP−. See Figure 27 through Figure 38 for performance
plots on the TxDAC’s ac performance. Optimum distortion
performance can typically be achieved by:
•
Limiting the peak positive VIOUTP+ and VIOUTP− to 0.8 V to
avoid onset of TxDAC’s output compression. (TxDAC’s
voltage compliance is around 1.2 V.)
•
Limiting V p-p seen at IOUTP+ and IOUTP− to less than
1.6 V.
Applications demanding higher output voltage swings and
power drive capabilities can benefit from using the IAMP.
IAMP CURRENT MODE OPERATION
The IAMP can be configured for the current mode operation as
shown in Figure 64 for loads remaining relatively constant. In
this mode, the primary path mirrors should be used to deliver
the signal-dependent current to the load via a center-tapped
transformer, because it provides the best linearity performance.
Because the mirrors exhibit a high output impedance, they can
be easily back-terminated (if required).
For peak signal currents (IOUTPK up to 50 mA), only the
primary path mirror gain should be used for optimum
distortion performance and power efficiency. The primary
path’s gain should be set to 4, with the secondary path’s gain
stages set to 0 (Reg. 0x10 = 0x84). The TxDAC’s standing
current, I, can be set between 2.5 mA and 12.5 mA with the
IOUTP outputs left open. The IOUTN outputs should be
connected to the transformer, with the IOUTG (and IOUTP)
Rev. 0 | Page 30 of 48
AD9866
outputs left open for optimum linearity performance. The
transformer1 should be specified to handle the dc standing
current, IBIAS, drawn by the IAMP. Also, because IBIAS remains
signal independent, a series resistor (not shown) can be inserted
between AVDD and the transformer’s center-tap to reduce the
IAMP’s common-mode voltage, VCM, and reduce the power
dissipation on the IC. The VCM, bias should not exceed 5.0 V and
the power dissipated in the IAMP alone is as follows:
PIAMP = 2 ∗ (N + G) ∗ I ∗ VCM
Equation 2.
AVDD
IBIAS = 2 × (N+G) × 1
0.1µF
RSET
AVDD
IOUTG+
0 TO –12dB
IOUTN–
IOUT_P+
RL
TxDAC
IOUTG–
IOUTPK = (N+G) × 1
P_OUTPK = (IOUTPK)2 × T2 × RL
IOUT_P–
IOUTPK
REFIO
IAMP
REFADJ
T:1
IOUTN+
R
DUAL NPN
PHILLIPS PBR951
RS 0.1µF
IOUTG+
IOUTPK
IAMP
0 TO –7.5dB
R
0 TO –12dB
AVDD
TO LOAD
IOUTN–
IOUTG–
RS 0.1µF
Figure 64. Current Mode Operation
A step-down transformer1 with a turn ratio, T, can be used to
increase the output power, P_OUT, delivered to the load. This
causes the output load, RL, to be reflected back to the IAMP’s
differential output by T2 resulting in a larger differential voltage
swing seen at the IAMP’s output. For example, the IAMP can
deliver 24 dBm of peak power to a 50 Ω load, if a 1.41:1 stepdown transformer is used. This results in 5 V p-p voltage swings
appearing at IOUTN+ and IOUTN− pins. Figure 42 shows how
the third order intercept point, OIP3, of the IAMP varies as a
function of common-mode voltage over a 2.5 MHz to 20.0 MHz
span with a 2-tone signal having a peak power of approximately
24 dBm with IOUTPK = 50 mA.
For applications requiring an IOUTPK exceeding 50 mA, set the
secondary’s path to deliver the additional current to the load.
IOUTG+ and IOUTN+ should be shorted as well as IOUTG−
and IOUTN−. If IOUTPK represents the peak current to be
delivered to the load, then the current gain in the secondary
path, G, can be set by the following equation:
Figure 65. Voltage Mode Operation
The peak differential voltage signal developed across the npn’s
bases is as follows:
VOUTPK = R ∗ (N ∗ I )
Equation 4.
where:
N is the gain setting of the primary mirror.
I is the standing current of the TxDAC defined in Equation 1.
The common-mode bias voltage seen at IOUTN+ and IOUTN−
is approximately AVDD − VOUTPK, while the common-mode
voltage seen at IOUTG+ and IOUTG− is approximately the
npn’s VBE drop below this level (AVDD − VOUTPK − 0.65). In
the voltage mode configuration, the total power dissipated
within the IAMP is as follows :
PIAMP = 2 ∗ I ∗ {(AVDD − VOUTPK ) ∗ N
G = IOUTPK / 12.5 − 4
+ ( AVDD − VOUTPK − 0.65) ∗ G}
Equation 3.
Equation 5.
The linearity performance becomes limited by the secondary
mirror path’s distortion.
1
04560-0-065
0 TO –7.5dB
IOUTN+
The voltage mode configuration is shown in Figure 65. This
configuration is suited for applications having a poorly defined
load that can vary over a considerable range. A low impedance
voltage driver can be realized with the addition of two external
RF bipolar npn transistors (Phillips PBR951) and resistors. In
this configuration, the current mirrors in the primary path
(IOUTN outputs) feed into scaling resistors, R, generating a
differential voltage into the bases of the npn transistors. These
transistors are configured as source followers with the secondary path current mirrors appearing at IOUTG+ and IOUTG−
providing a signal-dependent bias current. Note that the IOUTP
outputs must remain open for proper operation.
04560-0-064
TxDAC
0.1µF
IOUT_P–
IOUT_P+
RSET
REFADJ
REFIO
0.1µF
IAMP VOLTAGE MODE OPERATION
The emitters of the npn transistors are ac-coupled to the
transformer1 via a 0.1 µF blocking capacitor and series resistor
of 1Ω to 2 Ω. Note that protection diodes are not shown for
clarity purposes, but should be considered, if interfacing to a
power or phone line.
The B6080 and BX6090 transformers from Pulse Engineering are worthy of
consideration for current and voltage modes.
Rev. 0 | Page 31 of 48
AD9866
The amount of standing and signal-dependent current used to
bias the npn transistors is dependent on the peak current,
IOUTPK, required by the load. If the load is variable, determine
the worst case, IOUTPK, and add 3 mA of margin to ensure that
the npn transistors remain in the active region during peak load
currents. The gain of the secondary path, G, and the TxDAC’s
standing current, I, can be set using the following equation:
100
90
80
ISUPPLY (mA)
70
IOUTPK + 3 mA = G ∗ I
60
IAMPN OUTPUT
50
40
30
The voltage output driver exhibits a high output impedance, if
the bias currents for the npn transistors are removed. This
feature is advantageous in half-duplex applications (for example,
power lines) in which the Tx output driver must go into a high
impedance state while in Rx mode. If the AD9866 is configured
for the half-duplex mode (MODE = 0), the IAMP, TxDAC, and
interpolation filter are automatically powered down after a Tx
burst (via TXEN), thus placing the Tx driver into a high
impedance state while reducing its power consumption.
20
10
04560-0-066
TxDACs AVDD
Equation 6.
1
2
3
4
5
6
7
8
9
10
11
12
13
I (mA)
Figure 66. Current Consumption of TxDAC and IAMP in Current Mode
Operation with IOUTN Only (Default IAMP Settings)
150
140
130
IOUTG OUTPUT
120
110
100
90
80
70
60
50
TxDAC AVDD
40
30
IOUTN OUTPUT
20
10
1.0
04560-0-067
The Tx path’s analog current consumption is an important
consideration when determining its contribution to the overall
on-chip power dissipation. This is especially the case in fullduplex applications, where the power dissipation can exceed the
maximum limit of 1.66 W, if the IAMP’s IOUTPK is set to high.
The analog current consumption includes the TxDAC’s analog
supply (Pin 43) along with the standing current from the
IAMP’s outputs. Equation 2 and Equation 5 can be used to
calculate the power dissipated in the IAMP for the current and
voltage mode configuration. Figure 66 shows the current
consumption for the TxDAC and IAMP as a function of the
TxDAC’s standing current, I, when only the IOUTN outputs are
used. Figure 67 shows the current consumption for the TxDAC
and IAMP as a function of the TxDAC’s standing current, I,
when the IOUTN and IOUTG outputs are used. Both figures
are with the default current mirror gain settings of N = 4 and
G = 12.
ISUPPLY (mA)
IAMP CURRENT CONSUMPTION CONSIDERATIONS
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
I (mA)
Figure 67. Current Consumption of TxDAC and IAMP in Current Mode
Operation with IOUTN Only (Default IAMP Settings)
Rev. 0 | Page 32 of 48
AD9866
RECEIVE PATH
CLKOUT_1
CLKOUT_2
CLK
SYN.
2M CLK
MULTIPLIER
ADIO[11:6]/
Rx[5:0]
OSCIN
XTAL
10/12
RXEN/SYNC
RXCLK
ADC
80MSPS
SPGA
RX+
2-POLE
LPF
1-POLE
LPF
RX–
0 TO 6dB –6 TO 18dB –6 TO 24dB
∆ = 1dB ∆ = 6dB
∆ = 6dB
4
SPORT
GAIN
MAPPING
LUT
6
REGISTER
CONTROL
AD9865/AD9866
04560-0-068
PGA[5:0]
Figure 68. Functional Block Diagram of Rx Path
RX PROGRAMMABLE GAIN AMPLIFIER
The RxPGA has a digitally programmable gain range from
−12 dB to +48 dB with 1 dB resolution via a 6-bit word. Its
purpose is to extend the dynamic range of the Rx path such that
the input of the ADC is presented with a signal that scales
within its fixed 2 V input span. There are multiple ways of
setting the RxPGA’s gain as discussed in the RxPGA Control
section, as well as an alternative 3-bit gain mapping having a
range of −12 dB to +36 dB with 8 dB resolution.
The RxPGA is comprised of two sections: a continuous time
PGA (CPGA) for course gain and a switched capacitor PGA
(SPGA) for fine gain resolution. The CPGA consists of two
cascaded gain stages providing a gain range from −12 dB to
+42 dB with 6 dB resolution. The first stage features a low noise
preamplifier (< 3.0 nV/rtHz), thereby eliminating the need for
an external preamplifier. The SPGA provides a gain range from
0 dB to 6 dB with 1 dB resolution. A look-up table (LUT) is used
to select the appropriate gain setting for each stage.
To limit the RxPGA’s self-induced input offset, an offset
cancellation loop is included. This cancellation loop is
automatically performed upon power-up and can also be
initiated via SPI. During calibration, the RxPGA’s first stage is
internally shorted, and each gain stage set to a high gain setting.
A digital servo loop slaves a calibration DAC, which forces the
Rx input offset to be within ±32 LSB for this particular high
gain setting. Although the offset varies for other gain settings,
the offset is typically limited to ±5% of the ADC’s 2 V input
span. Note that the offset cancellation circuitry is intended to
reduce the voltage offset attributed to only the RxPGA’s input
stage, not any dc offsets attributed to an external source.
The gain of the RxPGA should be set to minimize clipping of
the ADC while utilizing most of its dynamic range. The
maximum peak-to-peak differential voltage that does not result
in clipping of the ADC is shown in Figure 69. While the graph
suggests that maximum input signal for a gain setting of −12 dB
is 8.0 V p-p, the maximum input voltage into the PGA should
be limited to less than 6 V p-p to prevent turning on ESD
protection diodes. For applications having higher maximum
input signals, consider adding an external resistive attenuator
network. While the input sensitivity of the Rx path is degraded
by the amount of attenuation on a dB-to-dB basis, the low noise
characteristics of the RxPGA provide some design margin such
that the external line noise remains the dominant source.
The nominal differential input impedance of the RxPGA input
appearing at the device RX+ and RX− input pins is 400 Ω//4 pF
(±20%) and remains relatively independent of gain setting. The
PGA input is self-biased at a 1.3 V common-mode level allowing maximum input voltage swings of ±1.5 V at RX+ and RX−.
AC coupling the input signal to this stage via coupling capacitors (0.1 µF) is recommended to ensure that any external dc
Rev. 0 | Page 33 of 48
8.0000
4.0000
2.0000
1.0000
0.5000
0.2500
0.1250
0.0625
0.0312
0.0156
0.0100
–12
04560-0-069
ADIO[11:6]/
Tx[5:0]
offset does not get amplified with high RxPGA gain settings,
potentially exceeding the ADC input range.
FULL-SCALE PEAK-TO-PEAK INPUT SPAN (V)
The receive path block diagram for the AD9866 (or AD9865) is
shown in Figure 68. The receive signal path consists of a 3-stage
RxPGA, a 3-pole programmable LPF, and a 12-bit (or 10-bit)
ADC. Note that the additional 2 bits of resolution offered by the
AD9866 (vs. the AD9865) result in a 3 dB to 5 dB lower noise
floor depending on the RxPGA gain setting and LPF cutoff
frequency. Also working in conjunction with the receive path is
an offset correction circuit. These blocks are discussed in detail
in the following sections. Note that the power consumption of
the RxPGA can be modified via Reg. 0x13 as discussed in the
Power Control and Dissipation section.
–6
0
6
12
18
24
30
36
42
48
GAIN (dB)
Figure 69. Maximum Peak-to-Peak Input vs. RxPGA Gain Setting
that Does Not Result in ADC Clipping
AD9866
1.30
0.25
Although the default setting specifies that the LPF be active, it
can also be bypassed providing a nominal f−3 dB of 55 MHz.
Table 20 shows the SPI registers pertaining to the LPF.
Table 20. SPI Registers for Rx Low-Pass Filter
Address (Hex)
0x07
0x08
Bit
(0)
(7:0)
Description
Enable Rx LPF
Target value
GAIN (dB)
The low-pass filter (LPF) provides a third order response with a
cut-off frequency that is typically programmable over a 15 MHz
to 35 MHz span. Figure 68 shows that the first real pole is
implemented within the first CPGA gain stage, and the complex
pole pair is implemented in the second CPGA gain stage.
Capacitor arrays are used to vary the different R-C time
constants within these two stages in a manner that changes the
cut-off frequency while preserving the normalized frequency
response. Because absolute resistor and capacitor values are
process-dependent, a calibration routine lasting less than 100 µs
automatically occurs each time the target cut-off frequency
register (Reg. 0x08) is updated, ensuring a repeatable cut-off
frequency from device to device.
NORMALIZED GAIN RESPONSE
–0.25
1.20
–0.50
1.15
–0.75
1.10
–1.00
1.05
–1.25
1.00
–1.50
0.95
–1.75
0.90
–2.00
0.85
0.80
–2.25
NORMALIZED GROUP DELAY
–2.50
0.75
0.70
–2.75
–3.00
NORMALIZED GROUP DELAY
TIME RESPONSE (GDT)
1.25
0
04560-0-071
LOW-PASS FILTER
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.65
1.0
0.9
NORMALIZED FREQUENCY
Figure 71. LPF’s Normalized Pass-Band Gain and Group Delay Responses
The −3 dB cut-off frequency, f−3 dB, is programmable by writing
an 8-bit word, referred to as the target, to Reg. 0x08. The cut-off
frequency is a function of the ADC sample rate, fADC, and to a
lesser extent RxPGA gain setting (in dB). Figure 72 shows how
the frequency response, f−3 dB, varies as a function of the RxPGA
gain setting.
3
The normalized wideband gain response is shown in Figure 70.
The normalized pass-band gain and group delay responses are
shown in Figure 71. The normalized cut-off frequency, f−3 dB,
results in −3 dB attenuation. Also, the actual group delay time
(GDT) response can be calculated given a programmed cut-off
frequency using the following equation:
–6dB GAIN
0dB GAIN
+6dB GAIN
+18dB GAIN
+30dB GAIN
+42dB GAIN
FUNDAMENTAL (dB)
0
Actual GDT = Normalized GDT /(2.45 ∗ f −3dB )
–3
–6
–9
–12
–15
5
–18
0
0
5
10
15
20
25
30
35
40
45
50
INPUT FREQUENCY (MHz)
–5
Figure 72. Effects of RxPGA Gain on LPF Frequency Response
( f−3 dB = 32 MHz (@ 0 dB and fADC = 80 MSPS)
–10
The following formula1 can be used to estimate f−3 dB for a
RxPGA gain setting of 0 dB:
–15
–20
–25
f −3dB _ 0dB = (128 / target ) ∗ ( f ADC / 80) ∗ ( f ADC / 30 + 23.83)
–30
Equation 8.
04560-0-070
GAIN (dB)
04560-0-072
Equation 7.
–35
0
0.5
1.0
1.5
FREQUENCY
2.0
2.5
3.0
Figure 73 compares the measured and calculated f−3 dB using this
formula.
Figure 70. LPF’s Normalized Wideband Gain Response
1
Empirically derived for a f−3 dB range of 15 MHz to 35 MHz and fADC of 40 MSPS
to 80 MSPS with an RxPGA = 0 dB.
Rev. 0 | Page 34 of 48
AD9866
35
ANALOG TO DIGITAL CONVERTER (ADC)
33
The AD9866 features a 12-bit analog-to-digital converter
(ADC) capable of up to 80 MSPS. Referring to Figure 68, the
ADC is driven by the SPGA stage, which performs both the
sample-and-hold and the fine gain adjust functions. A buffer
amplifier (not shown) isolates the last CPGA gain stage from
the dynamic load presented by the SPGA stage. The full-scale
input span of the ADC is 2 V p-p, with the full-scale input span
into the SPGA adjustable from 1 V to 2 V in 1 dB increments,
depending on the PGA gain setting.
FREQUENCY (MHz)
31
29
27
25
80 MSPS MEASURED
23
80 MSPS CALCULATED
21
17
04560-0-073
19
50 MSPS MEASURED
50 MSPS CALCULATED
15
48
64
80
96 112 128
144
160
176
192
208
A pipelined multistage ADC architecture is used to achieve high
sample rates while consuming low power. The ADC distributes
the conversion over several smaller A/D subblocks, refining the
conversion with progressively higher accuracy as it passes the
results from stage to stage on each clock edge. The ADC
typically performs best when driven internally by a 50% duty
cycle clock. This is especially the case when operating the ADC
at high sample rate (55 MSPS to 80 MSPS) and/or lower
internal bias levels, which adversely affect interstage settling
time requirements.
224
TARGET-DECIMAL EQUIVALENT
Figure 73. Measured and Calculated f−3 dB vs. Target Value
for fADC = 50 MSPS and 80 MSPS
The following scaling factor can be applied to the previous
formula to compensate for the RxPGA gain setting on f−3 dB:
Scale Factor = 1 − (RxPGA in dB ) / 382
Equation 9.
This scaling factor reduces the calculated f−3 dB as the RxPGA is
increased. Applications that need to maintain a minimum cutoff frequency, f−3 dB_MIN, for all RxPGA gain settings should first
determine the scaling factor for the highest RxPGA gain setting
to be used. Next, the f−3 dB_MIN should be divided by this scale
factor to normalize to the 0 dB RxPGA gain setting (f−3 dB_0 dB).
Equation 8 can then be used to calculate the target value.
The LPF frequency response shows a slight sensitivity to
temperature, as shown in Figure 74. Applications sensitive to
temperature drift can recalibrate the LPF by rewriting the target
value to Reg. 0x08.
The ADC’s power consumption can be reduced by 25 mA, with
minimal effect on its performance, by setting Bit 4 of Reg. 0x07.
Alternative power bias settings are also available via Reg. 0x13,
as discussed in the Power Control and Dissipation section.
Lastly, the ADC can be completely powered down for halfduplex operation, further reducing the AD9866’s peak power
consumption.
35
FREQUENCY (MHz)
30
FOUT ACTUAL 80MHz AND –40°C
FOUT ACTUAL 80MHz AND +25°C
25
The ADC sampling clock path also includes a duty cycle
restorer circuit, which ensures that the ADC gets a near 50%
duty cycle clock even when presented with a clock source with
poor symmetry (35/65). This circuit should be enabled, if the
ADC sampling clock is a buffered version of the reference
signal appearing at OSCIN (see the Clock Synthesizer section)
and if this reference signal is derived from an oscillator or
crystal whose specified symmetry cannot be guaranteed to be
within 45/55 (or 55/45). This circuit can remain disabled, if the
ADC sampling clock is derived from a divided down version of
the clock synthesizer’s VCO, because this clock is near 50%.
FOUT ACTUAL 80MHz AND +85°C
15
96
04560-0-074
20
112
128
144
160
176
192
208
224
240
TARGET-DECIMAL EQUIVALENT
Figure 74. Temperature Drift of f−3 dB for fADC = 80 MSPS and RxPGA = 0 dB
Rev. 0 | Page 35 of 48
AD9866
AGC TIMING CONSIDERATIONS
REFT
TO
ADCs
C1
0.1µF
C3
0.1µF
When implementing a digital AGC timing loop, it is important
to consider the Rx path latency and settling time of the Rx path
in response to a change in gain setting. Figure 21 and Figure 24
show the RxPGA’s settling response to a 60 dB and 5 dB change
in gain setting when using the Tx[5:0] or PGA[5:0] port. While
the RxPGA settling time may also show a slight dependency on
the LPF’s cutoff frequency, the ADC’s pipeline delay along with
the ADIO bus interface presents a more significant delay. The
amount of delay or latency depends on whether a half- or fullduplex is selected. An impulse response at the RxPGA’s input
can be observed after 10.0 ADC clock cycles (1/fADC) in the case
of a half-duplex interface and 10.5 ADC clock cycles in the case
of a full-duplex interface. This latency along with the RxPGA
settling time should be considered to ensure stability of the
AGC loop.
C2
10µF
C4
0.1µF
REFB
1.0V
TOP
VIEW
C4
C2
04560-0-075
C3
C1
Figure 75. ADC Reference and Decoupling
The ADC has an internal voltage reference and reference amplifier as shown in Figure 75. The internal band gap reference
generates a stable 1 V reference level that is converted to a
differential 1 V reference centered about mid-supply (AVDD/2).
The outputs of the differential reference amplifier are available
at the REFT and REFB pins and must be properly decoupled for
optimum performance. The REFT and REFB pins are conveniently situated at the corners of the CSP package such that C1
(0603 type) can be placed directly across its pins. C3 and C4 can
be placed underneath C1, and C2 (10 µF tantalum) can be
placed furthest from the package.
Table 21. SPI Registers for Rx ADC
Address (Hex)
0x04
0x07
0x13
Bit
(5)
(4)
(4)
(2:0)
Description
Duty cycle restore circuit
ADC clock from PLL
ADC low power mode
ADC power bias adjust
Rev. 0 | Page 36 of 48
AD9866
CLOCK SYNTHESIZER
The data rate, fDATA, for the Tx and Rx data paths must always be
equal. Therefore, the ADC’s sample rate, fADC, is always equal to
fDATA, while the TxDAC update rate is a factor of 1, 2, or 4 of
fDATA, depending on the interpolation factor selected. The data
rate refers to the word rate and should not be confused with the
nibble rate in full-duplex interface.
XTAL
XTAL
C1
÷2N
2M CLK
MULTIPLIER
OSCIN
TO ADC
TO TxDAC
C2
CLKOUT2
÷2L
04560-0-076
CLKOUT1
÷2R
Figure 76. Clock Oscillator and Synthesizer
The 2M CLK multiplier contains a PLL (with integrated loop
filter) and VCO capable of generating an output frequency that
is a multiple of 1, 2, 4, or 8 of its input reference frequency, fOSCIN,
appearing at OSCIN. The input frequency range of fOSCIN is
between 20 MHz and 80 MHz, while the VCO can operate over
a 40 MHz to 200 MHz span. For the best phase noise/jitter
characteristics, it is advisable to operate the VCO with a frequency between 100 MHz and 200 MHz. The VCO output
drives the TxDAC directly such that its update rate, fDAC, is
related to fOSCIN by the following equation:
f DAC = 2
M∗
(default setting) or a divided version of the VCO output (fDAC).
The first option is the default setting and most desirable, if fOSCIN
is equal to the ADC sample rate, fADC. This option typically
results in the best jitter/phase noise performance for the ADC
sampling clock. The second option is suitable in cases where
fOSCIN is a factor of 2 or 4 less than the fADC. In this case, the
divider ratio, N, is chosen such that the divided down VCO
output, fDAC, is equal to the ADC sample rate, as shown in the
following equation:
f DAC = 2 ( M −N ) ∗ f OSCIN
Equation 12.
where N = 0, 1, or 2.
Figure 77 shows the degradation in phase noise performance
imparted onto the ADC’s sampling clock for different VCO
output frequencies. In this case, a 25 MHz, 1 V p-p sinewave was
used to drive OSCIN and the PLL’s M and N factor were
selected to provide an fADC of 50 MHz for a VCO operating
frequency of 50, 100, and 200 MHz. The RxPGA input was
driven with a near full-scale, 12.5 MHz input signal with a gain
setting of 0 dB. Operating the VCO at the highest possible
frequency results in the best narrow and wideband phase noise
characteristics. For comparison purposes, the clock source for
the ADC was taken directly from OSCIN when driven by a
50 MHz square wave.
0
DIRECT
VCO = 50MHz
VCO = 100MHz
VCO = 200MHz
–10
–20
–30
–40
dBFS
The AD9866 generates all its internal sampling clocks, as well as
two user-programmable clock outputs appearing at CLKOUT1
and CLKOUT2, from a single reference source as shown in
Figure 76. The reference source can be either a fundamental
frequency or an overtone quartz crystal connected between
OSCIN and XTAL with the parallel resonant load components
as specified by the crystal manufacturer. It can also be a TTLlevel clock applied to OSCIN with XTAL left unconnected.
–50
–60
–70
f OSCIN
–80
04560-0-077
–90
Equation 10.
–100
where M = 0, 1, 2, or 3.
–110
2.5
M is the PLL’s multiplication factor set in Reg. 0x04. The value
of M is determined by the Tx path’s word rate, fDATA, and digital
interpolation factor, F, as shown in the following equation:
M = log 2 (F ∗ f DATA / f OSCIN )
Equation 11.
Note that, if the reference frequency appearing at OSCIN is
chosen to be equal to the AD9866’s Tx and Rx path’s word rate,
then M is simply equal to log2(F).
4.5
6.5
8.5
10.5
12.5
14.5
16.5
18.5
20.5
22.5
FREQUENCY (MHz)
Figure 77. Comparison of Phase Noise Performance when ADC Clock Source
Is Derived from Different VCO Output Frequencies
The CLK synthesizer also has two clock outputs appearing at
CLKOUT1 and CLKOUT2. They are programmable via
Reg. 0x06. Both outputs can be inverted or disabled. The voltage
levels appearing at these outputs are relative to DRVDD and
remain active during a hardware or software reset. Table 22
shows the SPI registers pertaining to the clock synthesizer.
The clock source for the ADC can be selected in Reg. 0x04 as a
buffered version of the reference frequency appearing at OSCIN
Rev. 0 | Page 37 of 48
AD9866
CLKOUT1 is a divided version of the VCO output and can be
set to be a submultiple integer of fDAC (fDAC/2R, where R = 0, 1, 2,
or 3). Because this clock is actually derived from the same set of
dividers used within the PLL core, it is phase-locked to them
such that its phase relationship relative to the signal appearing
at OSCIN (or RXCLK)can be determined upon power-up. Also,
this clock has near 50% duty cycle, because it is derived from
the VCO. As a result, CLKOUT1 should be selected before
CLKOUT2 as the primary source for system clock distribution.
Table 22. SPI Registers for CLK Synthesizer
Address (Hex)
0x04
0x06
CLKOUT2 is a divided version of the reference frequency, fOSCIN,
and can be set to be a submultiple integer of fOSCIN (fOSCIN/2L,
where L = 0, 1, or 2). With L set to 0, the output of CLKOUT2 is
a delayed version of the signal appearing at OSCIN, exhibiting
the same duty cycle characteristics. With L set to 1 or 2, the
output of CLKOUT2 is a divided version of the OSCIN signal,
exhibiting a near 50% duty cycle, but without having a
deterministic phase relationship relative to CLKOUT1 (or
RXCLK).
Rev. 0 | Page 38 of 48
Bit
(4)
(3:2)
(1:0)
(7:6)
(5)
(4)
(3:2)
(1)
(0)
Description
ADC CLK from PLL
PLL divide factor ( P)
PLL multiplication factor (M )
CLKOUT2 divide number
CLKOUT2 invert
CLKOUT2 disable
CLKOUT1 divide number
CLKOUT1 invert
CLKOUT1 disable
AD9866
POWER CONTROL AND DISSIPATION
POWER-DOWN
HALF-DUPLEX POWER SAVINGS
The AD9866 provides the ability to control the power-on state
of various functional blocks. The state of the PWRDWN pin
along with the contents of Reg. 0x01 and Reg. 0x02 allow two
user-defined power settings that are pin selectable. The default
settings1 are such that Reg. 0x01 has all blocks powered on (all
bits 0), while Reg. 0x02 has all blocks powered down excluding
the PLL such that the clock signal remains available at
CLKOUT1 and CLKOUT2. When the PWRDWN pin is low,
the functional blocks corresponding to the bits in Reg. 0x01 are
powered down. When the PWRDWN is high, the functional
blocks corresponding to the bits in Reg. 0x02 are powered
down. PWRDWN immediately affects the designated functional
blocks with minimum digital delay.
Significant power savings can be realized in applications having
a half-duplex protocol allowing only the Rx or Tx path to be
operational at any instance. The power-savings method depends
on whether the AD9866 is configured for a full- or half-duplex
interface. Functional blocks having fast power on/off times for
the Tx and Rx path are controlled by the following bits:
TxDAC/IAMP, Tx Digital, ADC, and RxPGA.
Table 23. SPI Registers Associated with Power-Down and
Half-Duplex Power Savings
Address (Hex)
0x01
0x02
0x03
Bit
(7)
(6)
(5)
(4)
(3)
(2)
(1)
(0)
(7)
(6)
(5)
(4)
(3)
(2)
(1)
(0)
(7:3)
(2)
(1)
(0)
1
Description
PLL
TxDAC/IAMP
TX Digital
REF
ADC CML
ADC
PGA BIAS
Rx PGA
PLL
TxDAC/IAMP
TX Digital
REF
ADC CML
ADC
PGA BIAS
Rx PGA
Tx OFF Delay
Rx PWRDWN
via TXEN
Enable Tx
PWRDWN
Enable Rx
PWRDWN
Comments
PWRDWN = 0
Default setting is all
functional blocks
powered on.
PWRDWN = 1
Default setting is all
functional blocks
powered off
excluding PLL.
Half-duplex power
savings.
In the case of a full-duplex digital interface (MODE = 1), one
can set Reg. 0x01 to 0x60 and Reg. 0x02 to 0x05 (or vice versa)
such that the AD9866’s Tx and Rx path are never powered on
simultaneously. The PWRDWN pin can then be used to control
what path is powered on, depending on the burst type. During a
Tx burst, the Rx path’s PGA and ADC blocks can typically be
powered down within 100 ns, while the Tx paths DAC, IAMP,
and digital filter blocks are powered up within 0.5 µs. For an Rx
burst, the Tx path’s can be powered down within 100 ns, while
the Rx circuitry is powered up within 2 µs.
The TXQUIET pin can also be used with the full-duplex
interface to quickly power down the IAMP and disable the
interpolation filter by setting this pin low. This is meant to
maintain backward compatibility with the AD9875/AD9876
MxFEs with the exception that the TxDAC remains powered if
its IOUTP outputs are used. In most applications, the interpolation filter needs to be flushed with 0s before or after being
powered down. This ensures that, upon power-up, the TxDAC
(and IAMP) have a negligible differential dc offset, thus
preventing spectral splatter due to an impulse transient.
Applications using a half-duplex interface (MODE = 0) can
benefit from an additional power savings feature made available
in Reg. 0x03. This register is effective only for a half-duplex
interface. Besides providing power savings for half-duplex
applications, this feature allows the AD9866 to be used in
applications that need only its Rx (or Tx) path functionality
through pin-strapping, making a serial port interface (SPI)
optional. This feature also allows the PWRDWN pin to retain
its default function as a master power control, as defined in
Table 10.
The default settings for Reg. 0x03 provide fast power control of
the functional blocks in the Tx and Rx signal paths (outlined
above) using the TXEN pin. The TxDAC still remains powered
on in this mode, while the IAMP is powered down. Significant
current savings are typically realized when the IAMP is
powered down.
With MODE = 1 and CONFIG = 1, Reg. 0x02 default settings are with all
blocks powered off, with RXCLK providing a buffered version of the signal
appearing at OSCIN. This setting results in the lowest power consumption
upon power-up while still allowing AD9865 to generate the system clock via a
crystal.
For a Tx burst, the falling edge of TXEN is used to generate an
internal delayed signal for powering down the Tx circuitry.
Upon receipt of this signal, power-down of the Tx circuitry
occurs within 100 ns. The user-programmable delay for the Tx
Rev. 0 | Page 39 of 48
AD9866
55
path power-down is meant to match the pipeline delay of the
last Tx burst sample such that power-down of the TxDAC and
IAMP does not impact its transmission. A 5-bit field in Reg.
0x03 sets the delay from 0 to 31 TXCLK clock cycles, with the
default being 31 (0.62 µs with fTxCLK = 50 MSPS). The digital
interpolation filter is automatically flushed with midscale
samples prior to power-down, if the clock signal into the
TXCLK pin is present for 33 additional clock cycles after TXEN
returns low. For an Rx burst, the rising edge of TXEN is used to
generate an internal signal (with no delay) that powers up the
Tx circuitry within 0.5 µs.
50
IAVDDTxDAC (mA)
45
35
30
25
04560-0-078
20
15
The Rx path power-on/power-off can be controlled by either
TXEN or RXEN by setting Bit 2 of Reg. 0x03. In the default
setting, the falling edge of TXEN powers up the Rx circuitry
within 2 µs, while the rising edge of TXEN powers down the Rx
circuitry within 0.5 µs. If RXEN is selected as the control signal,
then its rising edge powers up the Rx circuitry and the falling
edge powers it down. It is possible to disable the fast powerdown of the Tx and/or Rx circuitry by setting Bit 1 and/or Bit 0
to 0.
10
0
1
2
3
4
5
6
7
8
9
10
11
12
13
ISTANDING (mA)
Figure 78. Reduction in TxDAC’s Supply Current vs. Standing Current
65
60
4× INTERPOLATION
55
IDVDD (mA)
50
POWER REDUCTION OPTIONS
45
40
2× INTERPOLATION
35
30
1× (HALF-DUPLEX ONLY)
25
04560-0-079
The power consumption of the AD9866 can be significantly
reduced from its default setting by optimizing the power
consumption versus performance of the various functional
blocks in the Tx and Rx signal path. On the Tx path, minimum
power consumption is realized when the TxDAC output is used
directly and its standing current, I, is reduced to as low as 1 mA.
Although a slight degradation in THD performance results at
reduced standing currents, it often remains adequate for most
applications, because the op amp driver typically limits the
overall linearity performance of the Tx path. The load resistors
used at the TxDAC outputs (IOUTP+ and IOUTP−) can be
increased to generate an adequate differential voltage that can
be further amplified via a power efficient op amp based driver
solution. Figure 78 shows how the supply current for the
TxDAC (Pin 43) is reduced from 55 mA to 14 mA as the
standing current is reduced from 12.5 mA to 1.25 mA. Further
Tx power savings can be achieved by bypassing or reducing the
interpolation factor of the digital filter as shown in Figure 79.
40
20
15
20
30
40
50
60
70
80
INPUT DATA RATE (MSPS)
Figure 79. Digital Supply Current Consumption vs. Input Data Rate
(DVDD = DRVDD =3.3 V and fOUT = fDATA/10)
Power consumption on the Rx path can be achieved by reducing the bias levels of the various amplifiers contained within the
RxPGA and ADC. As previously noted, the RxPGA consists of
two CPGA amplifiers and one SPGA amplifier. The bias levels
of each of these amplifiers along with the ADC can be controlled via Reg. 0x13 as shown in Table 24. The default setting
for 0x13 is 0x00.
Table 24. SPI Register for RxPGA and ADC Biasing
Address (Hex)
0x07
0x13
Rev. 0 | Page 40 of 48
Bit
(4)
(7:5)
(4:3)
(2:0)
Description
ADC low power
CPGA bias adjust
SPGA bias adjust
ADC power bias adjust
AD9866
210
Because the CPGA processes signals in the continuous time
domain, its performance versus bias setting remains mostly
independent of the sample rate. Table 25 shows how the typical
current consumption seen at AVDD (Pins 35 and 40) varies as a
function of Bits (7:5), while the remaining bits are maintained at
their default setting of 0. Only four of the possible settings result
in any reduction in current consumption relative to the default
setting. Reducing the bias level typically results in a degradation
in the THD versus frequency performance as shown in
Figure 80 due to a reduction of the amplifier’s unity gain
bandwidth, while the SNR performance remains relatively
unaffected.
65.0
205
200
IAVDD (mA)
01
195
00
190
185
10
180
04560-0-081
11
175
170
20
30
–20
40
50
60
70
80
ADC SAMPLE RATE (MSPS)
–25
Figure 81. AVDD Current vs. SPGA Bias Setting and Sample Rate
60.0
–30
65
–54
57.5
–35
64
–56
55.0
–40
63
–50
–55
45.0
–60
42.5
–65
THD_RxPGA = 36dB
40.0
000
001
010
011
–70
100
04560-0-080
THD_RxPGA = 0dB
47.5
–66
THD-00
THD-01
THD-10
THD-11
–68
–70
56
Table 25. Analog Supply Current vs. CPGA Bias Settings at
fADC = 65 MSPS
Bit 5
0
1
0
1
0
1
0
1
59
∆ mA
0
−27
−42
−51
−55
27
69
27
The SPGA is implemented as a switched capacitor amplifier.
Therefore, its performance versus bias level is mostly dependent
on the sample rate. Figure 81 shows how the typical current
consumption seen at AVDD (Pins 35 and 40) varies as a
function of bits (4:3) and sample rate, while the remaining bits
are maintained at their default setting of 0. Figure 81 shows how
the SNR and THD performance is affected for a 10 MHz sine
wave input as the ADC sample rate is swept from 20 MHz to 80
MHz.
55
20
–62
–64
57
Figure 80. THD vs. fIN Performance and RxPGA Bias Settings (000,001,010,100
with RxPGA = 0 and +36 dB and AIN = −1 dBFS, LPF set to 26 MHz
and fADC = 50 MSPS)
Bit 6
0
0
1
1
0
0
1
1
60
58
CPGA BIAS SETTING-BITS (7:5)
Bit 7
0
0
0
0
1
1
1
1
61
–60
–72
30
40
50
THD (dBc)
50.0
–58
SNR-00
SNR-01
SNR-10
SNR-11
62
60
70
–74
80
04560-0-082
–45
SNR (dBc)
SNR_RxPGA = 36dB
52.5
THD (dBc)
SNR (dBFS)
SNR_RxPGA = 0dB
62.5
SAMPLE RATE (MSPS)
Figure 82. SNR and THD Performance vs. fADC and SPGA Bias Setting with
RxPGA = 0 dB, fIN = 10 MHz. AIN = −1 dBFS
The ADC is based on a pipeline architecture with each stage
consisting of a switched capacitor amplifier. Therefore, its
performance versus bias level is also mostly dependent on the
sample rate. Figure 83 shows how the typical current consumption seen at AVDD (Pins 35 and 40) varies as a function of bits
(2:0) and sample rate, while the remaining bits are maintained
at their default setting of 0. Setting Bit 4 or Reg. 0x07 corresponds to the 011 setting, and the settings of 101 and 111 result
in higher current consumption. Figure 84 shows how the SNR
and THD performance are affected for a 10 MHz sine wave
input for the lower power settings as the ADC sample rate is
swept from 20 MHz to 80 MHz.
Rev. 0 | Page 41 of 48
AD9866
30.8oC/W, if the heat slug remains unsoldered.) If a particular
application’s maximum ambient temperature, TA, falls below
85oC, the maximum allowable power dissipation can be determined by the following equation:
220
101 OR 111
210
200
000
190
IAVDD (mA)
001
180
PMAX = 1.66 + (85 − TA ) / 24
010
170
Equation 13.
160
100
150
011
Assuming that the IAMP’s common-mode bias voltage is
operating off the same analog supply as the AD9866, the
following equation can be used to calculate the maximum total
current consumption, IMAX, of the IC:
101
04560-0-083
140
130
120
20
30
40
50
60
70
I MAX = (PMAX − PIAMP ) / 3.47
80
SAMPLE RATE (MSPS)
Equation 14.
Figure 83. AVDD Current vs. ADC Bias Setting and Sample Rate
65
–54
64
–56
63
–58
61
60
SNR-000
SNR-001
SNR-010
SNR-011
SNR-100
SNR-101
–60
–62
–64
59
–66
THD (dBc)
SNR (dBc)
62
With an ambient temperature of up to 85°C, IMAX is 478 mA.
If the IAMP is operating off a different supply or in the voltage
mode configuration, first calculate the power dissipated in the
IAMP, PIAMP, using Equation 2 or Equation 5, and then recalculate IMAX, using the following equation:
I MAX = (PMAX − PIAMP ) / 3.47
Equation 15.
57
56
55
20
–68
THD-000
THD-001
THD-010
THD-011
THD-100
THD-101
30
40
–70
–72
50
60
70
–74
80
04560-0-084
58
Figure 78, Figure 79, Figure 81, and Figure 83 can be used to
calculate the current consumption of the Rx and Tx paths for a
given setting.
MODE SELECT UPON POWER-UP AND RESET
SAMPLE RATE (MSPS)
Figure 84. SNR and THD Performance vs. fADC and SPGA Bias Setting with
RxPGA = 0 dB, fIN = 10 MHz, AIN = −1 dBFS
A sine wave input is a standard and convenient method of
analyzing the performance of a system. However, the amount of
power reduction that is possible is application dependent, based
on the nature of the input waveform (such as frequency content,
peak-to-rms ratio), the minimum ADC sample, and the minimum acceptable level of performance. As a result, it is advisable
that power-sensitive applications optimize the power bias
setting of the Rx path using an input waveform that is representative of the application.
POWER DISSIPATION
The power dissipation of the AD9866 can become quite high in
full-duplex applications in which the Tx and Rx paths are
simultaneously operating with nominal power bias settings. In
fact, some applications desiring to use the IAMP may need to
either reduce its peak power capabilities or reduce the power
consumption of the Rx path, so that the device’s maximum
allowable power consumption, PMAX, is not exceeded.
PMAX is specified at 1.66 W to ensure that the die temperature
does not exceed 125oC at an ambient temperature of 85oC. This
specification is based on the 64-pin LFSCP having a thermal
resistance, θJA, of 24oC/W with its heat slug soldered. (The θJA is
The AD9866 power-up state is determined by the logic levels
appearing at the MODE and CONFIG pins. The MODE pin is
used to select a half- or full-duplex interface by pin strapping it
low or high, respectively. The CONFIG pin is used in conjunction with the MODE pin to determine the default settings for
the SPI registers as outlined in Table 10.
The intent of these particular default settings is to allow some
applications to avoid using the SPI (disabled by pin-strapping
SEN high), thereby reducing the implementation cost. For
example, setting MODE low and CONFIG high configures the
AD9866 to be backward compatible with the AD9975, while
setting MODE high and CONFIG low makes it backward
compatible with the AD9875. Other applications must use the
SPI to configure the device.
A hardware (RESET pin) or software (Bit 5 of Reg. 0x00) reset
can be used to place the AD9866 into a known state of operation as determined by the state of the MODE and CONFIG
pins. A dc offset calibration and filter tuning routine is also
initiated upon a hardware reset, but not with a software reset.
Neither reset method flushes the digital interpolation filters in
the Tx path. Refer to the Half-Duplex Mode and Full-Duplex
Mode sections for information on flushing the digital filters.
Rev. 0 | Page 42 of 48
AD9866
A hardware reset can be triggered by pulsing the RESET pin low
for a minimum of 50 ns. The SPI registers are instantly reset to
their default settings upon RESET going low, while the dc offset
calibration and filter tuning routine is initiated upon RESET
returning high. To ensure sufficient power-on time of the
various functional blocks, RESET returning high should occur
no less than 10 ms upon power-up. If a digital reset signal from
a microprocessor reset circuit (such as ADM1818) is not
available, a simple R-C network referenced to DVDD can be
used to hold RESET low for approximately 10 ms upon
power-up.
ANALOG AND DIGITAL LOOP-BACK TEST MODES
The AD9866 features analog and digital loop-back capabilities
that can assist in system debug and final test. Analog loop-back
routes the digital output of the ADC back into the Tx data path
prior to the interpolation filters such that the Rx input signal
can be monitored at the output of the TxDAC or IAMP. As a
result, the analog loop-back feature can be used for a half- or
full-duplex interface, to allow testing of the functionality of the
entire IC (excluding the digital data interface).
For example, the user can configure the AD9866 with similar
settings as the target system, inject an input signal (sinusoidal
waveform) into the Rx input, and monitor the quality of the
reconstructed output from the TxDAC or IAMP to ensure a
minimum level of performance. In this test, the user can also
exercise the RxPGA as well as validate the attenuation characteristics of the RxLPF. Note that the RxPGA gain setting should
be selected such that the input does not result in clipping of the
ADC.
Digital loop-back can be used to test the full-duplex digital
interface of the AD9866. In this test, data appearing on the
Tx[5:0] port is routed back to the Rx[5:0] port, thereby
confirming proper bus operation. The Rx port can also be
three-stated for half- and full-duplex interfaces.
Table 26. SPI Registers for Test Modes
Address (Hex)
0x0D
Rev. 0 | Page 43 of 48
Bit
(7)
(6)
(5)
Description
Analog loop-back
Digital loop-back
Rx port three-state
AD9866
PCB DESIGN CONSIDERATIONS
Although the AD9866 is a mixed-signal device, the part should
be treated as an analog component. The on-chip digital circuitry has been specially designed to minimize the impact of its
digital switching noise on the MxFE’s analog performance.
To achieve the best performance, the power, grounding, and
layout recommendations in this section should be followed.
Assembly instructions for the micro-lead frame package can be
found in an application note from Amkor at:
http://www.amkor.com/products/notes_papers/MLF_AppNote
_0902.pdf.
COMPONENT PLACEMENT
If the three following guidelines of component placement are
followed, chances for getting the best performance from the
MxFE are greatly increased. First, manage the path of return
currents flowing in the ground plane so that high frequency
switching currents from the digital circuits do not flow on the
ground plane under the MxFE or analog circuits. Second, keep
noisy digital signal paths and sensitive receive signal paths as
short as possible. Third, keep digital (noise generating) and
analog (noise susceptible) circuits as far away from each other
as possible.
To best manage the return currents, pure digital circuits that
generate high switching currents should be closest to the power
supply entry. This keeps the highest frequency return current
paths short and prevents them from traveling over the sensitive
MxFE and analog portions of the ground plane. Also, these
circuits should be generously bypassed at each device, which
further reduces the high frequency ground currents. The MxFE
should be placed adjacent to the digital circuits, such that the
ground return currents from the digital sections do not flow in
the ground plane under the MxFE.
The AD9866 has several pins that are used to decouple sensitive
internal nodes. These pins are REFIO, REFB, and REFT. The
decoupling capacitors connected to these points should have
low ESR and ESL. These capacitors should be placed as close to
the MxFE as possible (see Figure 75) and be connected directly
to the analog ground plane. The resistor connected to the
REFADJ pin should also be placed close to the device and
connected directly to the analog ground plane.
POWER PLANES AND DECOUPLING
While the AD9866 evaluation board demonstrates a very good
power supply distribution and decoupling strategy, it can be
further simplified for many applications. The board has four
layers: two signal layers, one ground plane, and one power
plane. While the power plane on the evaluation board is split
into multiple analog and digital subsections, a permissible
alternative would be to have AVDD and CLKVDD share the
same analog 3.3 V power plane. A separate analog plane/supply
may be allocated to the IAMP, if its supply voltage differs from
the 3.3 V required by AVDD and CLKVDD. On the digital
side, DVDD and DRVDD can share the same 3.3 V digital
power plane. This digital power plane brings the current used
to power the digital portion of the MxFE and its output drivers.
This digital plane should be kept from going underneath the
analog components.
The analog and digital power planes allocated to the MxFE may
be fed from the same low noise voltage source; however, they
should be decoupled from each other to prevent the noise
generated in the digital portion of the MxFE from corrupting
the AVDD supply. This can be done by using ferrite beads
between the voltage source and the respective analog and
digital power planes with a low ESR, bulk decoupling capacitor
on the MxFE side of the ferrite. Each of the MxFE’s supply pins
(AVDD, CLKVDD, DVDD, and DRVDD) should also have a
dedicated low ESR, ESL decoupling capacitors. The decoupling
capacitors should be placed as close to the MxFE supply pins as
possible.
GROUND PLANES
The AD9866 evaluation board uses a single serrated ground
plane to help prevent any high frequency digital ground
currents from coupling over to the analog portion of the
ground plane. The digital currents affiliated with the high speed
data bus interface (Pins 1–16) have the highest potential of
generating problematic high frequency noise. A ground
serration that contains these currents should reduce the effects
of this potential noise source.
The ground plane directly underneath the MxFE should be
continuous and uniform. The 64-lead LFCSP package is
designed to provide excellent thermal conductivity. This is
partly achieved by incorporating an exposed die paddle on the
bottom surface of the package. However, to take full advantage
of this feature, the PCB must have features to effectively
conduct heat away from the package. This can be achieved by
incorporating thermal pad and thermal vias on the PCB. While
a thermal pad provides a solderable surface on the top surface
of the PCB (to solder the package die paddle on the board),
thermal vias are needed to provide a thermal path to inner
and/or bottom layers of the PCB to remove the heat.
Lastly, all ground connections should be made as short as
possible. This results in the lowest impedance return paths and
the quietest ground connections.
SIGNAL ROUTING
The digital Rx and Tx signal paths should be kept as short as
possible. Also, the impedance of these traces should have a
controlled characteristic impedance of about 50 Ω. This
prevents poor signal integrity and the high currents that can
Rev. 0 | Page 44 of 48
AD9866
occur during undershoot or overshoot caused by ringing. If the
signal traces cannot be kept shorter than about 1.5 inches,
series termination resistors (33 Ω to 47 Ω) should be placed
close to all digital signal sources. It is a good idea to seriesterminate all clock signals at their source, regardless of trace
length.
The receive RX+ and RX− signals are the most sensitive signals
on the entire board. Careful routing of these signals is essential
for good receive path performance. The RX+ and RX− signals
form a differential pair and should be routed together as a pair.
By keeping the traces adjacent to each other, noise coupled
onto the signals appears as common mode and is largely
rejected by the MxFE receive input. Keeping the driving point
impedance of the receive signal low and placing any low-pass
filtering of the signals close to the MxFE further reduces the
possibility of noise corrupting these signals.
Rev. 0 | Page 45 of 48
AD9866
EVALUATION BOARD
An evaluation board is available for the AD9865 and AD9866.
The digital interface to the evaluation board can be configured
for a half- or full-duplex interface. Two 40-pin and one 26-pin
male right angle headers (0.100 inches) provide easy interfacing
to test equipment such as digital data capture boards, pattern
generators, or custom digital evaluation boards (FPGA, DSP, or
ASIC). The reference clock source can originate from an
external generator, crystal oscillator, or crystal. Software and an
interface cable are included to allow for programming of the
SPI registers via a PC.
The analog interface on the evaluation board provides a full
analog front-end reference design for power line applications. It
includes a power line socket, line transformer, protection
diodes, and passive filtering components. An auxiliary path
allows independent monitoring of the ac power line. The
evaluation board allows complete optimization of power line
reference designs based around the AD9865 or AD9866.
Alternatively, the evaluation board allows independent
evaluation of the TxDAC, IAMP, and Rx paths via SMA
connectors. The IAMP can be easily configured for a voltage or
current mode interface via jumper settings. The TxDAC’s
performance can be evaluated directly or via an optional dual
op amp driver stage. The Rx path includes a transformer and
termination resistor, allowing for a calibrated differential input
signal to be injected into its front end.
More information on the AD9866 evaluation board can be
found at:
http://www.analog.com/Analog_Root/productPage/
productHome/0%2C2121%2CAD9866%2C00.html.
Rev. 0 | Page 46 of 48
AD9866
OUTLINE DIMENSIONS
9.00
BSC SQ
0.60 MAX
0.60 MAX
49
48
PIN 1
INDICATOR
TOP
VIEW
8.75
BSC SQ
0.30
0.25
0.18
PIN 1
INDICATOR
64
1
7.25
7.10 SQ*
6.95
BOTTOM
VIEW
0.45
0.40
0.35
33
32
17
16
0.25 MIN
7.50 REF
1.00
0.85
0.80
12° MAX
0.80 MAX
0.65 TYP
0.05 MAX
0.02 NOM
0.50 BSC
SEATING
PLANE
0.20 REF
* COMPLIANT TO JEDEC STANDARDS MO-220-VMMD
EXCEPT FOR EXPOSED PAD DIMENSION
Figure 85. 64-Lead Lead Frame Chip Scale Package (LFCSP)
[CP-64-3]
Dimensions shown in millimeters
ORDERING GUIDE
Model
AD9866BCP
AD9866BCPRL
AD9866CHIPS
AD9866-EB
Temperature Range
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
25°C
Package Description
64-Lead LFCSP
64-Lead LFCSP
Chip
Evaluation Board
Rev. 0 | Page 47 of 48
Package Option
CP-64-3
CP-64-3
NOTES
© 2003 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
C04560–0–11/03(0)
Rev. 0 | Page 48 of 48