1. PIN NAMES ARE FOR THE CMOS PIN CONFIGURATI ON ONLY;
SEE FIGURE 7 FOR LVDS PIN NAMES.
PDWNOEB
DCOB
DCOA
D15A (MSB)
TO
D0A (LSB)
D15B (MSB)
TO
D0B (LSB)
ORA
CLK–
CLK+
ORB
VIN+A
VCM
RBIAS
VIN–B
VIN+B
VIN–A
VREF
SENSE
AD9650
ADC
16
16
08919-001
Data Sheet
FEATURES
1.8 V analog supply operation
1.8 V CMOS or LVDS output supply
SNR
82 dBFS at 30 MHz input and 105 MSPS data rate
83 dBFS at 9.7 MHz input and 25 MSPS data rate
SFDR
90 dBc at 30 MHz input and 105 MSPS data rate
95 dBc at 9.7 MHz input and 25 MSPS data rate
Low power
328 mW per channel at 105 MSPS
119 mW per channel at 25 MSPS
Integer 1-to-8 input clock divider
IF sampling frequencies to 300 MHz
Analog input range of 2.7 V p-p
Optional on-chip dither
Integrated ADC sample-and-hold inputs
Differential analog inputs with 500 MHz bandwidth
ADC clock duty cycle stabilizer
FUNCTIONAL BLOCK DIAGRAM
APPLICATIONS
Industrial instrumentation
X-Ray, MRI, and ultrasound equipment
High speed pulse acquisition
Chemical and spectrum analysis
Direct conversion receivers
Multimode digital receivers
Smart antenna systems
General-purpose software radios
GENERAL DESCRIPTION
The AD9650 is a dual, 16-bit, 25 MSPS/65 MSPS/80 MSPS/
105 MSPS analog-to-digital converter (ADC) designed for
digitizing high frequency, wide dynamic range signals with
input frequencies of up to 300 MHz.
The dual ADC core features a multistage, differential pipelined
architecture with integrated output error correction logic. Each
ADC features wide bandwidth, differential sample-and-hold
analog input amplifiers, and shared integrated voltage reference,
which eases design considerations. A duty cycle stabilizer is
provided to compensate for variations in the ADC clock duty
cycle, allowing the converters to maintain excellent performance.
The ADC output data can be routed directly to the two external
16-bit output ports or multiplexed on a single 16-bit bus. These
outputs can be set to either 1.8 V CMOS or LVDS.
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.
Figure 1.
Flexible power-down options allow significant power savings,
when desired.
Programming for setup and control is accomplished using a 3-wire
SPI-compatible serial interface.
The AD9650 is available in a 64-lead LFCSP and is specified over
the industrial temperature range of −40°C to +85°C.
PRODUCT HIGHLIGHTS
1. On-chip dither option for improved SFDR performance
with low power analog input.
2. Proprietary differential input that maintains excellent SNR
performance for input frequencies up to 300 MHz.
3. Operation from a single 1.8 V supply and a separate digital
output driver supply accommodating 1.8 V CMOS or
LVDS outputs.
4. Standard serial port interface (SPI) that supports various
product features and functions, such as data formatting
(offset binary, twos complement, or gray coding), enabling
the clock DCS, power-down, and test modes.
5. Pin compatible with the AD9268 and other dual families,
AD9269, AD9251, AD9231, and AD9204. This allows a
simple migration across resolutions and bandwidth.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700 www.analog.com
AD9650 Data Sheet
TABLE OF CONTENTS
Features .............................................................................................. 1
AD9650BCPZ-25 AD9650BCPZ-65 AD9650BCPZ-80 AD9650BCPZ-105
Parameter1 Temp Min Typ Max Min Typ Max Min Typ Max Min Typ Max Unit
SIGNAL-TO-NOISE RATIO (SNR)
fIN = 9.7 MHz 25°C 83 83 83 82.5 dBFS
fIN = 30 MHz 25°C 81.5 82 82 82 dBFS
Full 81.8 81.5 81.6 80.5 dBFS
fIN = 70 MHz 25°C 79.5 81 81 80 dBFS
fIN = 141 MHz2 25°C 79.5 80 80 dBFS
SIGNAL-TO-NOISE-AND-DISTORTION
(SINAD)
fIN = 9.7 MHz 25°C 82.2 82 82 82 dBFS
fIN = 30 MHz 25°C 80 81.2 82 80.4 dBFS
Full 81.5 81 80.7 80 dBFS
fIN = 70 MHz 25°C 78 79.2 78.5 78.8 dBFS
fIN = 141 MHz2 25°C 75 75.1 75.5 dBFS
EFFECTIVE NUMBER OF BITS (ENOB)
fIN = 9.7 MHz 25°C 13.5 13.5 13.5 13.3 Bits
fIN = 30 MHz 25°C 13.0 13.2 13.2 13.2 Bits
fIN = 70 MHz 25°C 12.7 13.0 13.0 13.0 Bits
fIN = 141 MHz2 25°C 12.9 13.0 12.3 Bits
WORST SECOND OR THIRD HARMONIC
fIN =9.7 MHz 25°C −95 −94 −95.5 −91 dBc
fIN = 30 MHz 25°C −85 −93 −92 −90 dBc
Full −91.5 −88 −87 −87 dBc
fIN = 70 MHz 25°C −87 −86 −86 −92 dBc
fIN = 141 MHz 25°C −79 −79 −80 dBc
SPURIOUS-FREE DYNAMIC RANGE (SFDR)
fIN = 9.7 MHz 25°C 95 94 95.5 91 dBc
fIN = 30 MHz 25°C 85 93 92 90 dBc
Full 91.5 88 87 87 dBc
fIN = 70 MHz 25°C 87 86 86 92 dBc
fIN = 141 MHz 25°C 79 79 80 dBc
WORST OTHER (HARMONIC OR SPUR)
fIN = 9.7 MHz 25°C −110 −105 −105 −100 dBc
fIN = 30 MHz 25°C −102 −105 −105 −101 dBc
Full −97 −97 −97 −94 dBc
fIN = 70 MHz 25°C −97 −97 −97 −97 dBc
fIN = 141 MHz 25°C −97 −97 −88 dBc
TWO-TONE SFDR
fIN = 7.2 MHz (−7 dBFS ), 8.4 MHz
(−7 dBFS)
fIN = 25 MHz (−7 dBFS ), 30 MHz
(−7 dBFS)
fIN = 125 MHz (−7 dBFS ), 128 MHz
(−7 dBFS)
CROSSTALK3 Full −105 −105 −105 −105 dBFS
ANALOG INPUT BANDWIDTH 25°C 500 500 500 500 MHz
1
See the AN-835 Application Note, Understanding High Speed ADC Testing and Evaluation, for a complete set of definitions.
2
Measurements made with a divide-by-4 clock rate to minimize the effects of clock jitter on the SNR performance.
3
Crosstalk is measured with a 170 MHz tone at −1 dBFS on one channel and no input on the alternate channel.
25°C 87
25°C 84 90 87 87 dBc
25°C 83 83 84 dBc
Rev. A | Page 4 of 44
Data Sheet AD9650
Input Voltage Range
Full
AGND
AVDD
V
Input Capacitance
Full 1 pF
High Level Input Voltage
Full
1.22 2.1
V
Low Level Input Current
Full
38 128
µA
Input Resistance
Full 26 kΩ
DIGITAL SPECIFICATIONS
AVDD = 1.8 V, DRVDD = 1.8 V, maximum sample rate, VIN = −1.0 dBFS differential input, 1.35 V internal reference, and DCS enabled, unless
otherwise noted.
Table 3.
Parameter Temperature Min Typ Max Unit
DIFFERENTIAL CLOCK INPUTS (CLK+, CLK−)
Logic Compliance CMOS/LVDS/LVPECL
Internal Common-Mode Bias Full 0.9 V
Differential Input Voltage Full 0.3 3.6 V p-p
Input Common-Mode Range Full 0.9 1.4 V
High Level Input Current Full −100 +100 µA
Low Level Input Current Full −100 +100 µA
Input Capacitance Full 9 pF
Input Resistance Full 8 10 12 kΩ
SYNC INPUT
Logic Compliance CMOS
Internal Bias Full 0.9 V
Input Voltage Range Full AGND AVDD V
High Level Input Voltage Full 1.2 AVDD V
Low Level Input Voltage Full AGND 0.6 V
High Level Input Current Full −100 +100 µA
Low Level Input Current Full −100 +100 µA
Input Resistance Full 12 16 20 kΩ
LOGIC INPUT (CSB)1
Low Level Input Voltage Full 0 0.6 V
High Level Input Current Full −10 +10 µA
Low Level Input Current Full 40 132 µA
Input Resistance Full 26 kΩ
Input Capacitance Full 2 pF
LOGIC INPUT (SCLK/DFS)2
High Level Input Voltage Full 1.22 2.1 V
Low Level Input Voltage Full 0 0.6 V
High Level Input Current (VIN = 1.8 V) Full −92 −135 µA
Low Level Input Current Full −10 +10 µA
Input Resistance Full 26 kΩ
Input Capacitance Full 2 pF
LOGIC INPUT/OUTPUT (SDIO/DCS)1
High Level Input Voltage Full 1.22 2.1 V
Low Level Input Voltage Full 0 0.6 V
High Level Input Current Full −10 +10 µA
Input Resistance Full 26 kΩ
Input Capacitance Full 5 pF
LOGIC INPUTS (OEB, PDWN)2
High Level Input Voltage Full 1.22 2.1 V
Low Level Input Voltage Full 0 0.6 V
High Level Input Current (VIN = 1.8 V) Full −90 −134 µA
Low Level Input Current Full −10 +10 µA
Input Capacitance Full 5 pF
Rev. A | Page 5 of 44
AD9650 Data Sheet
IOH = 0.5 mA
Full
1.75
V
Output Offset Voltage (VOS), ANSI Mode
Full
1.15
1.25
1.35
V
Parameter Temperature Min Typ Max Unit
DIGITAL OUTPUTS
CMOS Mode—DRVDD = 1.8 V
High Level Output Voltage
IOH = 50 µA Full 1.79 V
Low Level Output Voltage
IOL = 1.6 mA Full 0.2 V
IOL = 50 µA Full 0.05 V
LVDS Mode —DRVDD = 1.8 V
Differential Output Voltage (VOD), ANSI Mode Full 290 345 400 mV
Differential Output Voltage (VOD), Reduced Swing Mode Full 160 200 230 mV
Output Offset Voltage (VOS), Reduced Swing Mode Full 1.15 1.25 1.35 V
1
Pull up.
2
Pull down.
Rev. A | Page 6 of 44
Data Sheet AD9650
SWITCHING SPECIFICATIONS
AVDD = 1.8 V, DRVDD = 1.8 V, maximum sample rate, VIN = −1.0 dBFS differential input, 1.35 V internal reference, and DCS enabled,
tDS Setup time between the data and the rising edge of SCLK 2 ns min
tDH Hold time between the data and the rising edge of SCLK 2 ns min
t
Period of the SCLK 40 ns min
CLK
tH Hold time between CSB and SCLK 2 ns min
t
SCLK pulse width high 10 ns min
HIGH
t
SCLK pulse width low 10 ns min
LOW
t
EN_SDIO
Time required for the SDIO pin to switch from an input to an output relative to the
10 ns min
SCLK falling edge
t
DIS_SDIO
Time required for the SDIO pin to switch from an output to an input relative to the
10 ns min
SCLK rising edge
1
See Figure 93.
Timing Diagrams
Unit
Figure 2. CMOS Default Output Mode Data Output Timing
Figure 3. CMOS Interleaved Output Mode Data Output Timing
Rev. A | Page 8 of 44
Data Sheet AD9650
t
PD
t
SKEW
t
CH
t
DCO
t
CLK
CH A
N – 12
CH B
N – 12
CH A
N – 11
CH B
N – 11
CH A
N – 10
CH B
N – 10
CH A
N – 9
CH B
N – 9
CH A
N – 8
N – 1
N + 1
N + 2
N + 3
N + 5
N + 4
N
V
IN
CLK+
CLK–
CH A/CH B DATA
DCOA/DCOB
t
A
08919-003
SYNC
CLK+
t
HSYNC
t
SSYNC
08919-004
Figure 4. LVDS Mode Data Output Timing
Figure 5. SYNC Input Timing Requirements
Rev. A | Page 9 of 44
AD9650 Data Sheet
CLK+, CLK− to AGND
−0.3 V to AVDD + 0.2 V
CSB to AGND
−0.3 V to DRVDD + 0.2 V
D0A/D0B Through D15A/D15B to
−0.3 V to DRVDD + 0.2 V
150°C
ABSOLUTE MAXIMUM RATINGS
Table 6.
Parameter Rating
Electrical1
AVDD to AGND −0.3 V to +2.0 V
DRVDD to AGND −0.3 V to +2.0 V
VIN+A/VIN+B, VIN−A/VIN−B to AGND −0.3 V to AVDD + 0.2 V
SYNC to AGND −0.3 V to AVDD + 0.2 V
VREF to AGND −0.3 V to AVDD + 0.2 V
SENSE to AGND −0.3 V to AVDD + 0.2 V
VCM to AGND −0.3 V to AVDD + 0.2 V
RBIAS to AGND −0.3 V to AVDD + 0.2 V
SCLK/DFS to AGND −0.3 V to DRVDD + 0.2 V
SDIO/DCS to AGND −0.3 V to DRVDD + 0.2 V
OEB −0.3 V to DRVDD + 0.2 V
PDWN −0.3 V to DRVDD + 0.2 V
AGND
DCOA/DCOB to AGND
−0.3 V to DRVDD + 0.2 V
Environmental
Operating Temperature Range
−40°C to +85°C
(Ambient)
Maximum Junction Temperature
Under Bias
Storage Temperature Range
−65°C to +150°C
(Ambient)
1
The inputs and outputs are rated to the supply voltage (AVDD or DRVDD) +
0.2 V but should not exceed 2.1 V.
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those 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
The exposed paddle must be soldered to the ground plane for
the LFCSP package. Soldering the exposed paddle to the PCB
increases the reliability of the solder joints and maximizes the
thermal capability of the package.
Typical θ
is specified for a 4-layer PCB with a solid ground
JA
plane. As shown in Tab l e 7, airflow improves heat dissipation,
which reduces θ
. In addition, metal in direct contact with the
JA
package leads from metal traces, through holes, ground, and
power planes reduces θ
.
JA
Table 7. Thermal Resistance
Package Type
64-Lead LFCSP
(CP-64-6)
1
Per JEDEC 51-7, plus JEDEC 25-5 2S2P test board.
2
Per JEDEC JESD51-2 (still air) or JEDEC JESD51-6 (moving air).
1. THE EXP OSED THERMAL P AD ON THE BOT TOM OF THE PACKAGE
PROVIDES THE ANALOG GROUND FOR THE PART. THIS EXPOSED
PAD MUST BE CO NNE CTED TO GROUND FOR PRO P E R OPERATION.
08919-005
1
CLK+
Input
ADC Clock Input—True.
32
D6A
Output
Channel A CMOS Output Data.
PIN CONFIGURATIONS AND FUNCTION DESCRIPTIONS
Table 8. Pin Function Descriptions (Parallel CMOS Mode)
Pin No. Mnemonic Type Description
ADC Power Supplies
10, 19, 28, 37 DRVDD Supply Digital Output Driver Supply (1.8 V Nominal).
49, 50, 53, 54, 59,
AVDD Supply Analog Power Supply (1.8 V Nominal).
60, 63, 64
0
AGND,
Exposed Pad
ADC Analog
51 VIN+A Input Differential Analog Input Pin (+) for Channel A.
52 VIN−A Input Differential Analog Input Pin (−) for Channel A.
62 VIN+B Input Differential Analog Input Pin (+) for Channel B.
61 VIN−B Input Differential Analog Input Pin (−) for Channel B.
55 VREF Input/output Voltage Reference Input/Output.
56 SENSE Input Voltage Reference Mode Select. See Table 11 for details.
58 RBIAS Input/output External Reference Bias Resistor.
57 VCM Output Common-Mode Level Bias Output for Analog Inputs.
2 CLK− Input ADC Clock Input—Complement.
Digital Input
3 SYNC Input Digital Synchronization Pin. Slave mode only.
Digital Outputs
25 D0A Output Channel A CMOS Output Data (LSB).
26 D1A Output Channel A CMOS Output Data.
27 D2A Output Channel A CMOS Output Data.
29 D3A Output Channel A CMOS Output Data.
30 D4A Output Channel A CMOS Output Data.
31 D5A Output Channel A CMOS Output Data.
The exposed thermal pad on the bottom of the package provides the analog
ground for the part. This exposed pad must be connected to ground for proper
operation.
Rev. A | Page 11 of 44
AD9650 Data Sheet
41
D14A
Output
Channel A CMOS Output Data.
7
D3B
Output
Channel B CMOS Output Data.
13
D8B
Output
Channel B CMOS Output Data.
20
D14B
Output
Channel B CMOS Output Data.
Pin No. Mnemonic Type Description
33 D7A Output Channel A CMOS Output Data.
34 D8A Output Channel A CMOS Output Data.
35 D9A Output Channel A CMOS Output Data.
36 D10A Output Channel A CMOS Output Data.
38 D11A Output Channel A CMOS Output Data.
39 D12A Output Channel A CMOS Output Data.
40 D13A Output Channel A CMOS Output Data.
42 D15A Output Channel A CMOS Output Data (MSB).
43 ORA Output Channel A Overrange Output.
4 D0B Output Channel B CMOS Output Data (LSB).
5 D1B Output Channel B CMOS Output Data.
6 D2B Output Channel B CMOS Output Data.
8 D4B Output Channel B CMOS Output Data.
9 D5B Output Channel B CMOS Output Data.
11 D6B Output Channel B CMOS Output Data.
12 D7B Output Channel B CMOS Output Data.
14 D9B Output Channel B CMOS Output Data.
15 D10B Output Channel B CMOS Output Data.
16 D11B Output Channel B CMOS Output Data.
17 D12B Output Channel B CMOS Output Data.
18 D13B Output Channel B CMOS Output Data.
21 D15B Output Channel B CMOS Output Data (MSB).
22 ORB Output Channel B Overrange Output
24 DCOA Output Channel A Data Clock Output.
23 DCOB Output Channel B Data Clock Output.
SPI Control
45 SCLK/DFS Input SPI Serial Clock/Data Format Select Pin in External Pin Mode.
44 SDIO/DCS Input/output SPI Serial Data I/O/Duty Cycle Stabilizer Pin in External Pin Mode.
46 CSB Input SPI Chip Select (Active Low).
ADC Configuration
47 OEB Input Output Enable Input (Active Low) in External Pin Mode.
48 PDWN
Input
Power-Down Input in External Pin Mode. In SPI mode, this input can be
configured as power-down or standby.
1. THE EXP OSED THERMAL P AD ON THE BOT TOM OF THE PACKAGE
PROVIDES THE ANALOG GROUND FOR THE PART. THIS EXPOSED
PAD MUST BE CO NNE CTED TO GROUND FOR PRO P E R OPERATION.
Table 9. Pin Function Descriptions (Interleaved Parallel LVDS Mode)
Pin No. Mnemonic Type Description
ADC Power Supplies
10, 19, 28, 37 DRVDD Supply Digital Output Driver Supply (1.8 V Nominal).
49, 50, 53, 54, 59,
AVDD Supply Analog Power Supply (1.8 V Nominal).
60, 63, 64
0
AGND,
Exposed Pad
Ground
The exposed thermal pad on the bottom of the package provides the analog
ground for the part. This exposed pad must be connected to ground for proper
operation.
ADC Analog
51 VIN+A Input Differential Analog Input Pin (+) for Channel A.
52 VIN−A Input Differential Analog Input Pin (−) for Channel A.
61 VIN−B Input Differential Analog Input Pin (−) for Channel B.
55 VREF Input/output Voltage Reference Input/Output.
56 SENSE Input Voltage Reference Mode Select. See Table 11 for details.
58 RBIAS Input/output External Reference Bias Resistor.
57 VCM Output Common-Mode Level Bias Output for Analog Inputs.
2 CLK− Input ADC Clock Input—Complement.
Digital Input
3 SYNC Input Digital Synchronization Pin. Slave mode only.
Digital Outputs
5 D0+ Output Channel A/Channel B LVDS Output Data 0—True (LSB).
4 D0− Output Channel A/Channel B LVDS Output Data 0—Complement (LSB).
7 D1+ Output Channel A/Channel B LVDS Output Data 1—True.
6 D1− Output Channel A/Channel B LVDS Output Data 1—Complement.
9 D2+ Output Channel A/Channel B LVDS Output Data 2—True.
8 D2− Output Channel A/Channel B LVDS Output Data 2—Complement.
Rev. A | Page 13 of 44
AD9650 Data Sheet
21
D7+
Output
Channel A/Channel B LVDS Output Data 7—True.
30
D10+
Output
Channel A/Channel B LVDS Output Data 10—True.
33
D12−
Output
Channel A/Channel B LVDS Output Data 12—Complement.
40
D15−
Output
Channel A/Channel B LVDS Output Data 15—Complement (MSB).
48
PDWN
Input
Power-Down Input in External Pin Mode. In SPI mode, this input can be
Pin No. Mnemonic Type Description
11 D3− Output Channel A/Channel B LVDS Output Data 3—Complement.
14 D4+ Output Channel A/Channel B LVDS Output Data 4—True.
13 D4− Output Channel A/Channel B LVDS Output Data 4—Complement.
16 D5+ Output Channel A/Channel B LVDS Output Data 5—True.
15 D5− Output Channel A/Channel B LVDS Output Data 5—Complement.
18 D6+ Output Channel A/Channel B LVDS Output Data 6—True.
17 D6− Output Channel A/Channel B LVDS Output Data 6—Complement.
20 D7− Output Channel A/Channel B LVDS Output Data 7—Complement.
23 D8+ Output Channel A/Channel B LVDS Output Data 8—True.
22 D8− Output Channel A/Channel B LVDS Output Data 8—Complement.
27 D9+ Output Channel A/Channel B LVDS Output Data 9—True.
26 D9− Output Channel A/Channel B LVDS Output Data 9—Complement.
29 D10− Output Channel A/Channel B LVDS Output Data 10—Complement.
32 D11+ Output Channel A/Channel B LVDS Output Data 11—True.
31 D11− Output Channel A/Channel B LVDS Output Data 11—Complement.
34 D12+ Output Channel A/Channel B LVDS Output Data 12—True.
36 D13+ Output Channel A/Channel B LVDS Output Data 13—True.
35 D13− Output Channel A/Channel B LVDS Output Data 13—Complement.
39 D14+ Output Channel A/Channel B LVDS Output Data 14—True.
38 D14− Output Channel A/Channel B LVDS Output Data 14—Complement.
41 D15+ Output Channel A/Channel B LVDS Output Data 15—True (MSB).
43 OR+ Output Channel A/Channel B LVDS Overrange Output—True.
42 OR− Output Channel A/Channel B LVDS Overrange Output—Complement.
25 DCO+ Output Channel A/Channel B LVDS Data Clock Output—True.
24 DCO− Output Channel A/Channel B LVDS Data Clock Output—Complement.
SPI Control
45 SCLK/DFS Input SPI Serial Clock/Data Format Select Pin in External Pin Mode.
44 SDIO/DCS Input/output SPI Serial Data I/O/Duty Cycle Stabilizer Pin in External Pin Mode.
46 CSB Input SPI Chip Select (Active Low).
ADC Configuration
47 OEB Input Output Enable Input (Active Low) in External Pin Mode.
Figure 49. AD9650-105 Single-Tone FFT with fIN = 9.7 MHz
Figure 50. AD9650-105 Single-Tone FFT with fIN = 30.3 MHz
Figure 52. AD9650-105 Single-Tone FFT with fIN = 141 MHz
Figure 53. AD9650-105 Single-Tone FFT with fIN = 30.3 MHz @ −6 dBFS
with Dither Disabled
Figure 51. AD9650-105 Single-Tone FFT with fIN = 70.1 MHz
Figure 54. AD9650-105 Single-Tone FFT with fIN = 30.3 MHz @ −6 dBFS
with Dither Enabled
Rev. A | Page 24 of 44
Data Sheet AD9650
120
0
20
40
60
80
100
–1000–10–20–30–40–50–60–70–80–90
SNR/SFDR
INPUT AMPLITUDE (dBFS)
SNR (dB)
SNR (dBFS)
SFDR (dBc)
SFDR (dBF S )
08919-158
120
115
110
105
100
95
90
85
80
75
70
–1000–10–20–30–40–50–60–70–80–90
SNR/SFDR (dBFS)
INPUT AMPLITUDE (dBFS)
SNR (dBFS) DITHER OFF
SFDR (dBFS) DITHER OFF
SNR (dBFS) DITHER ON
SFDR (dBF S ) DITHER ON
08919-159
100
65
100
90
80
70
60
50
40
30
20
10
0
70
75
80
85
90
95
050100150200250300
SNR (dBFS)
SFDR (dBc)
INPUT FRE QUENCY (MHz)
SNR (–40°C)
SFDR (–40°C)
SNR (+25°C)
SFDR (+25°C)
SNR (+85°C)
SFDR (+85°C)
08919-160
SNR
SFDR
105
75
80
85
90
95
100
125105859095100110120115
SNR (dBFS) , SFDR (dBc)
SAMPLE RATE (MSPS)
08919-258
SFDR
SNR
0
–140
–120
–100
–80
–60
–40
–20
05040302010
AMPLITUDE (dBFS)
FREQUENCY (MHz)
105MSPS
30.8MHz @ –7dBFS
25.4MHz @ –7dBFS
SFDR = 86.6dBc ( 93.6dBFS)
08919-162
0
–140
–120
–100
–80
–60
–40
–20
–90–10–20–30–40–50–60–70–80
SFDR/IMD3
INPUT AMPLITUDE (dBFS)
IMD3 (dBc)
SFDR (dBc)
IMD3 (dBF S )
SFDR (dBF S )
08919-163
Figure 55. AD9650-105 Single-Tone SNR/SFDR vs. Input Amplitude (AIN)
= 30.3 MHz
with f
IN
Figure 56. AD9650-105 Single Tone SNR/SFDR vs. Input Amplitude (AIN)
= 30.3 MHz with and Without Dither Enabled
with f
IN
Figure 58. AD9650-105 Single-Tone SNR/SFDR vs. Sample Rate (fS)
= 30 MHz
with f
IN
Figure 59. AD9650-105 Two-Tone FFT with f
= 25.4 MHz and f
IN1
IN2
= 30.8 MHz
Figure 57. AD9650-105 Single-Tone SNR/SFDR vs. Input Frequency (fIN)
Rev. A | Page 25 of 44
Figure 60. AD9650-105 Two-Tone SFDR/IMD3 vs. Input Amplitude (AIN)
with f
= 25.4 MHz, f
IN1
= 30.8 MHz, fS = 105 MSPS
IN2
AD9650 Data Sheet
0
–140
–120
–100
–80
–60
–40
–20
05040302010
AMPLITUDE (dBFS)
FREQUENCY (MHz)
105MSPS
124.8MHz @ –7dBFS
128.3MHz @ –7dBFS
SFDR = 83.8dBc
08919-164
0
–140
–120
–100
–80
–60
–40
–20
–90–10–20–30–40–50–60–70–80
SFDR/IMD3
INPUT AMPLITUDE (dBFS)
IMD3 (dBc)
SFDR (dBc)
IMD3 (dBF S )
SFDR (dBF S )
08919-165
900
800
700
600
500
400
300
200
100
0
25145125105856545
TOTAL P OWER (mW)/CURRENT (mA)
SAMPLE RATE (MSPS)
TOTAL POWER LVDS (mW)
LVDS I
DRVDD
(mA)
LVDS AND CMOS I
AVDD
(mA)
TOTAL POWER CMOS (mW)
CMOS I
DRVDD
(mA)
08919-263
1200000
0
200000
400000
600000
800000
1000000
NUMBER OF HI TS
OUTPUT CODE
N + 8
N
N – 1
N + 1
N + 2
N + 3
N + 4
N + 5
N + 6
N + 7
N – 2
N – 3
N – 4
N – 5
N – 6
N – 7
N – 8
08919-168
6
4
2
0
–2
–4
–6
INL ERROR ( LSB)
OUTPUT CODE
0100002000030000400005000060000
DITHER DIS ABLED
DITHER ENABL E D
08919-169
2.0
–2.0
–1.5
–1.0
–0.5
0
0.5
1.0
1.5
DNL ERROR (L S B)
OUTPUT CODE
0100002000030000400005000060000
08919-170
Figure 61. AD9650-105 Two-Tone FFT with f
= 124.8 MHz and f
IN1
= 128.3 MHz
IN2
Figure 64. AD9650-105 Grounded Input Histogram
Figure 62. AD9650-105 Two-Tone SFDR/IMD3 vs. Input Amplitude (A
f
= 128.3 MHz, f
IN1
Figure 63. AD9650-105 Power and Current vs. Sample Rate
= 124.8 MHz, fs = 105 MSPS
IN2
) with
IN
Figure 65. AD9650-105 INL with f
= 9.7 MHz
IN
Figure 66. AD9650-105 DNL with fIN = 9.7 MHz
Rev. A | Page 26 of 44
Data Sheet AD9650
100
40
50
60
70
80
90
0.800.850.900.951.001.051.101.151.20
SNR/SFDR
COMMON-MODE VOLTAGE (V)
SFDR (dBc)
SNR (dBFS)
TYPICAL V
CM
08919-171
Figure 67. SNR/SFDR vs. Input Common Mode (VCM)
= 30.3 MHz
with f
IN
Rev. A | Page 27 of 44
AD9650 Data Sheet
VIN±x
08919-007
AVDD
CLK+
CLK–
0.9V
10kΩ10kΩ
08919-008
DRVDD
PAD
08919-009
26kΩ
350Ω
DRVDD
SDIO/DCS
08919-010
26kΩ
350Ω
DRVDD
SCLK/DFS
OR OEB
08919-011
350Ω
AVDD
SENSE
08919-012
26kΩ
350Ω
DRVDD
CSB
08919-013
6kΩ
AVDD
VREF
08919-014
26kΩ
350Ω
PDWN
08919-015
EQUIVALENT CIRCUITS
Figure 68. Equivalent Analog Input Circuit
Figure 69. Equivalent Clock Input Circuit
Figure 70. Digital Output
Figure 73. Equivalent SENSE Circuit
Figure 74. Equivalent CSB Input Circuit
Figure 75. Equivalent VREF Circuit
Figure 71. Equivalent SDIO/DCS Circuit
Figure 72. Equivalent SCLK/DFS or OEB Input Circuit
Figure 76. Equivalent PDWN Input Circuit
Rev. A | Page 28 of 44
C
PAR1
C
PAR1
C
PAR2
C
PAR2
S
S
S
S
S
S
C
FB
C
FB
C
S
C
S
BIAS
BIAS
VIN+x
H
VIN–x
08919-034
Data Sheet AD9650
THEORY OF OPERATION
The AD9650 dual-core analog-to-digital converter (ADC) is
used for digitizing high frequency, wide dynamic range signals with
input frequencies of up to 300 MHz. The user can sample any f
frequency segment from dc to 300 MHz using appropriate lowpass or band-pass filtering at the ADC inputs with little loss in
ADC performance. The ADCs can also be operated with
independent analog inputs.
In quadrature applications, the AD9650 can be used as a baseband
or direct down-conversion receiver, in which one ADC is used
for I input data, and the other is used for Q input data.
Synchronization capability is provided to allow synchronized
timing between multiple devices.
Programming and control of the AD9650 are accomplished
using a 3-wire, SPI-compatible serial interface.
ADC ARCHITECTURE
The AD9650 architecture consists of a dual front-end sampleand-hold circuit, followed by a pipelined, switched-capacitor
ADC. The quantized outputs from each stage are combined into
a final 16-bit result in the digital correction logic. The pipelined
architecture permits the first stage to operate on a new input
sample and the remaining stages to operate on the preceding
samples. Sampling occurs on the rising edge of the clock.
Each stage of the pipeline, excluding the last, consists of a low
resolution flash ADC connected to a switched-capacitor digitalto-analog converter (DAC) and an interstage residue amplifier
(MDAC). The MDAC magnifies the difference between the
reconstructed DAC output and the flash input for the next stage
in the pipeline. One bit of redundancy is used in each stage to
facilitate digital correction of flash errors. The last stage simply
consists of a flash ADC.
The input stage of each channel contains a differential sampling
circuit that can be ac- or dc-coupled in differential or singleended modes. The output staging block aligns the data, corrects
errors, and passes the data to the output buffers. The output buffers
are powered from a separate supply, allowing digital output noise to
be separated from the analog core. During power-down, the output
buffers go into a high impedance state.
ANALOG INPUT CONSIDERATIONS
The analog input to the AD9650 is a differential switchedcapacitor circuit that has been designed for optimum performance
while processing a differential input signal.
The clock signal alternatively switches the input between sample
mode and hold mode (see Figure 77). When the input is switched
into sample mode, the signal source must be capable of charging
the sample capacitors and settling within ½ of a clock cycle.
/2
S
A small resistor in series with each input can help reduce the
peak transient current required from the output stage of the
driving source. A shunt capacitor can be placed across the inputs
to provide dynamic charging currents. This passive network
creates a low-pass filter at the ADC input; therefore, the precise
values are dependent on the application.
In intermediate frequency (IF) undersampling applications, any
shunt capacitors should be reduced. In combination with the
driving source impedance, the shunt capacitors limit the input
bandwidth. Refer to the AN-742 Application Note, Frequency Domain Response of Switched-Capacitor ADCs; the AN-827
Application Note, A Resonant Approach to Interfacing Amplifiers to Switched-Capacitor ADCs; and the Analog Dialogue article,
“Transformer-Coupled Front-End for Wideband A/D Converters,”
for more information on this subject (visit www.analog.com).
Figure 77. Switched-Capacitor Input
For best dynamic performance, the source impedances driving
VIN+x and VIN−x should be matched, and the inputs should
be differentially balanced.
An internal differential reference buffer creates positive and
negative reference voltages that define the input span of the ADC
core. The span of the ADC core is set by this buffer to 2 × VREF.
Input Common Mode
The analog inputs of the AD9650 are not internally dc biased.
In ac-coupled applications, the user must provide this bias externally. Setting the device so that VCM = 0.5 × AVDD (or 0.9 V) is
recommended for optimum performance, but the device
functions over a wider range with reasonable performance
(see Figure 67). An on-chip, common-mode voltage reference
is included in the design and is available from the VCM pin.
Optimum performance is achieved when the common-mode
voltage of the analog input is set by the VCM pin voltage
(typically 0.5 × AVDD). The VCM pin must be decoupled to
ground by a 0.1 µF capacitor, as described in the Applications
Information section.
Rev. A | Page 29 of 44
AD9650 Data Sheet
ADC CORE
DITHER
DAC
PN GEN
DITHER ENABL E
AD9650
VIN
DOUT
08919-058
VIN
76.8Ω
120Ω
0.1µF
200Ω
200Ω
90Ω
AVDD
33Ω
33Ω
15Ω
15Ω
5pF
15pF
15pF
AD9650
VIN–x
VIN+x
VCM
ADA4938-2
08919-035
2V p-p
49.9Ω
0.1µF
R1
R1
C1
AD9650
VIN+x
VIN–x
VCM
C2
R2
R2
C2
08919-036
Common-Mode Voltage Servo
In applications where there may be a voltage loss between the VCM
output of the AD9650 and the analog inputs, the common-mode
voltage servo can be enabled. When the inputs are ac-coupled and
a resistance of >100 Ω is placed between the VCM output and the
analog inputs, a significant voltage drop can occur and the
common-mode voltage servo should be enabled. Setting Bit 0 in
Register 0x0F to a logic high enables the VCM servo mode. In
this mode, the AD9650 monitors the common-mode input level
at the analog inputs and adjusts the VCM output level to keep the
common-mode input voltage at an optimal level. If both channels
are operational, Channel A is monitored. However, if Channel A
is in power-down or standby mode, the Channel B input is
monitored.
Dither
The AD9650 has an optional dither mode that can be selected
for one or both channels. Dithering is the act of injecting a known
but random amount of white noise, commonly referred to as
dither, into the input of the ADC. Dithering has the effect of
improving the local linearity at various points along the ADC
transfer function. Dithering can significantly improve the SFDR
when quantizing small-signal inputs, typically when the input
level is below −6 dBFS.
As shown in Figure 78, the dither that is added to the input of
the ADC through the dither DAC is precisely subtracted out
digitally to minimize SNR degradation. When dithering is
enabled, the dither DAC is driven by a pseudorandom number
generator (PN gen). In the AD9650, the dither DAC is precisely
calibrated to result in only a very small degradation in SNR and
SINAD.
typically at very low levels and do not limit SFDR when the
ADC is quantizing large-signal inputs, dithering converts these
tones to noise and produces a whiter noise floor.
Small-Signal FFT
For small-signal inputs, the front-end sampling circuit typically
contributes very little distortion, and, therefore, the SFDR is likely
to be limited by tones caused by DNL errors due to random component mismatches. Therefore, for small-signal inputs (typically,
those below −6 dBFS), dithering can significantly improve
SFDR by converting these DNL tones to white noise.
Static Linearity
Dithering also removes sharp local discontinuities in the INL
transfer function of the ADC and reduces the overall peak-topeak INL.
In receiver applications, utilizing dither helps to reduce DNL errors
that cause small-signal gain errors. Often this issue is overcome
by setting the input noise 5 dB to 10 dB above the converter
noise. By using dither within the converter to correct the DNL
errors, the input noise requirement can be reduced.
Differential Input Configurations
Optimum performance is achieved while driving the AD9650
in a differential input configuration. For baseband applications,
the AD8138, ADA4937-2, and ADA4938-2 differential drivers
provide excellent performance and a flexible interface to the ADC.
The output common-mode voltage of the ADA4938-2 is easily
set with the VCM pin of the AD9650 (see Figure 79), and the
driver can be configured in a Sallen-Key filter topology to
provide band limiting of the input signal.
Figure 78. Dither Block Diagram
Large-Signal FFT
In most cases, dithering does not improve SFDR for large-signal
inputs close to full scale, for example, with a −1 dBFS input. For
large-signal inputs, the SFDR is typically limited by front-end
sampling distortion, which dithering cannot improve. However,
even for such large-signal inputs, dithering may be useful for
certain applications because it makes the noise floor whiter.
As is common in pipeline ADCs, the AD9650 contains small
DNL errors caused by random component mismatches that
produce spurs or tones that make the noise floor somewhat
randomly colored part-to-part. Although these tones are
Rev. A | Page 30 of 44
Figure 79. Differential Input Configuration Using the ADA4938-2
For baseband applications in which SNR is a key parameter,
differential transformer coupling is the recommended input
configuration. An example is shown in Figure 80. To bias the
analog input, the VCM voltage can be connected to the center
tap of the secondary winding of the transformer.
The signal characteristics must be considered when selecting
a transformer. Most RF transformers saturate at frequencies
below a few megahertz (MHz). Excessive signal power can also
cause core saturation, which leads to distortion.
At input frequencies in the second Nyquist zone and above, the
noise performance of most amplifiers is not adequate to achieve
the true SNR performance of the AD9650. For applications in
which SNR is a key parameter, differential double balun coupling
is the recommended input configuration (see Figure 81). In this
configuration, the input is ac-coupled, and the CML is provided
to each input through a 33 Ω resistor. These resistors compensate
for losses in the input baluns to provide a 50 Ω impedance to
the dr iver.
In the double balun and transformer configurations, the value of
the input capacitors and resistors is dependent on the input frequency and source impedance and may need to be reduced or
removed. Tabl e 10 displays recommended values to set the RC
network. At higher input frequencies, good performance can be
achieved by using a ferrite bead in series with a resistor and
removing the capacitors. However, these values are dependent
on the input signal and should be used only as a starting guide.
Table 10. Example RC Network
Frequency
Range
(MHz)
R1 Series
(Ω Each)
C1 Differential
(pF)
R2 Series
(Ω Each)
C2 Shunt
(pF Each)
0 to 100 33 5 15 15
100 to 200 10 5 10 10
100 to 300 101 Remove 66 Remove
1
In this configuration, R1 is a ferrite bead with a value of 10 Ω at 100 MHz.
An alternative to using a transformer-coupled input at frequencies in the second Nyquist zone is to use the AD8352
differential driver. An example is shown in Figure 82. See the
AD8352 data sheet for more information.
Figure 82. Differential Input Configuration Using the AD8352
Rev. A | Page 31 of 44
AD9650 Data Sheet
VREF
V
SELECT
AD9650
SELECT
LOGIC
0.1µF1.0µF
VIN–A/VIN–B
VIN+A/VIN+B
ADC
CORE
SENSE
08919-040
0
–3.0
–2.5
–2.0
–1.5
–1.0
–0.5
01.41.21.00.80.60.40.2
REFERENCE VOLTAGE E RROR (%)
LOAD CURRENT ( mA)
08919-188
0
–6
–5
–4
–3
–2
–1
–509070503010–10–30
REFERENCE VOLTAGE E RROR (mV)
TEMPERATURE (°C)
08919-189
Internal Reference
AGND to 0.2
1.35
2.7
VOLTAGE REFERENCE
The AD9650 can be configured for a stable 1.35 V internal
reference or a user-applied external reference. The input range
of the ADC always equals twice the voltage at the reference pin
(VREF) for either an internal or an external reference. Ta ble 11
shows a summary of the internal and external reference
connections.
Internal Reference Connection
A stable and accurate 1.35 V reference is built into the AD9650,
allowing a 2.7 V p-p full-scale input. To configure the AD9650
for an internal reference, the SENSE pin must be tied low. In
addition, to achieve optimal noise performance, it is recommended
that the VREF pin be decoupled by 1.0 µF and 0.1 µF capacitors
close to the pin. Figure 83 shows the configuration for the internal
reference connection.
Figure 84. Reference Voltage Error vs. Load Current
External Reference Operation
The use of an external reference may be necessary to enhance
the gain accuracy of the ADC or improve thermal drift characteristics. Figure 85 shows the typical drift characteristics of the
internal reference in 1.35 V mode.
When the SENSE pin is tied to AVDD, the internal reference is
disabled, allowing the use of an external reference. An internal
reference buffer loads the external reference with an equivalent
6 kΩ load (see Figure 75). The internal buffer generates the
positive and negative full-scale references for the ADC core.
Therefore, the external reference must be limited to a maximum
of 1.35 V.
If the internal reference of the AD9650 is used to drive multiple
converters to improve gain matching, the loading of the reference
by the other converters must be considered. Figure 84 shows
how the internal reference voltage is affected by loading.
Table 11. Reference Configuration Summary
Selected Mode SENSE Voltage (V) Resulting VREF (V) Resulting Differential Span (V p-p)
For optimum performance, the AD9650 sample clock inputs,
CLK+ and CLK−, should be clocked with a differential signal.
The signal is typically ac-coupled into the CLK+ and CLK− pins
via a transformer or capacitors. These pins are biased internally
(see Figure 86) and require no external bias. If the inputs are
floated, the CLK− pin is pulled low to prevent spurious clocking.
Figure 86. Equivalent Clock Input Circuit
Clock Input Options
The AD9650 has a very flexible clock input structure. Clock input
can be a CMOS, LVDS, LVPECL, or sine wave signal. Regardless of
the type of signal being used, clock source jitter is of the most
concern, as described in the Jitter Considerations section.
Figure 87 and Figure 88 show two preferred methods for clocking
the AD9650 (at clock rates up to 625 MHz). A low jitter clock
source is converted from a single-ended signal to a differential
signal using either an RF balun or an RF transformer.
The RF balun configuration is recommended for clock frequencies
between 125 MHz and 625 MHz, and the RF transformer is recommended for clock frequencies from 10 MHz to 200 MHz. The
back-to-back Schottky diodes across the transformer/balun’s
secondary windings limit the clock excursions into the AD9650
to approximately 0.8 V p-p differential.
This limit helps prevent the large voltage swings of the clock
from feeding through to other portions of the AD9650 while
preserving the fast rise and fall times of the signal that are critical
to a low jitter performance.
If a low jitter clock source is not available, another option is to
ac couple a differential PECL signal to the sample clock input
pins, as shown in Figure 89. The AD9510/AD9511/AD9512/
AD9513/AD9514/AD9515/AD9516/AD9517/AD9518 clock
drivers offer excellent jitter performance.
Figure 89. Differential PECL Sample Clock (Up to 625 MHz)
A third option is to ac couple a differential LVDS signal to the
sample clock input pins, as shown in Figure 90. The AD9510/
AD9511/AD9512/AD9513/AD9514/AD9515/AD9516/AD9517/
AD9518 clock drivers offer excellent jitter performance.
Figure 90. Differential LVDS Sample Clock (Up to 625 MHz)
In some applications, it may be acceptable to drive the sample
clock inputs with a single-ended CMOS signal. In such applications, the CLK+ pin should be driven directly from a CMOS gate,
and the CLK− pin should be bypassed to ground with a 0.1 μF
capacitor (see Figure 91).
Figure 87. Transformer-Coupled Differential Clock (Up to 200 MHz)
Figure 88. Balun-Coupled Differential Clock (Up to 625 MHz)
Input Clock Divider
The AD9650 contains an input clock divider with the ability to
divide the input clock by integer values between 1 and 8. For
divide ratios of 1, 2, 4, or 8, the duty cycle stabilizer (DCS) is
optional. For other divide ratios, divide-by-3, -5, -6, and -7, the
duty cycle stabilizer must be enabled for proper part operation.
The AD9650 clock divider can be synchronized using the external
SYNC input. Bit 0 to Bit 2 of Register 0x100 allow the clock
divider to be resynchronized on every SYNC signal or only on
the first SYNC signal after the register is written. A valid SYNC
causes the clock divider to reset to its initial state. This synchronization feature allows multiple parts to have their clock dividers
Rev. A | Page 33 of 44
aligned to guarantee simultaneous input sampling.
Figure 91. Single-Ended 1.8 V CMOS Input Clock (Up to 200 MHz)
AD9650 Data Sheet
)10/(
LF
SNR/
82
80
78
76
74
72
70
1432
SNR (dBFS)
CLK DIVIDE RATIO
08919-293
1V p-p CL K AM P LITUDE
2V p-p CL K AM P LITUDE
Clock Duty Cycle
Typical high speed ADCs use both clock edges to generate a variety
of internal timing signals and, as a result, may be sensitive to clock
duty cycle. The AD9650 requires a tight tolerance on the clock
duty cycle to maintain dynamic performance characteristics.
The AD9650 contains a duty cycle stabilizer (DCS) that retimes
the nonsampling (falling) edge, providing an internal clock signal
with a nominal 50% duty cycle. This allows the user to provide
a wide range of clock input duty cycles without affecting the
performance of the AD9650. Noise and distortion performance
are nearly flat for a wide range of duty cycles with the DCS
enabled.
Jitter in the rising edge of the input is still of paramount concern
and is not easily reduced by the internal stabilization circuit. The
duty cycle control loop does not function for clock rates of less
than 20 MHz, nominally. The loop has a time constant associated
with it that must be considered in applications in which the clock
rate can change dynamically. A wait time of 1.5 µs to 5 µs is
required after a dynamic clock frequency increase or decrease
before the DCS loop is relocked to the input signal. During the
time period that the loop is not locked, the DCS loop is bypassed,
and internal device timing is dependent on the duty cycle of the
input clock signal. In such applications, it may be appropriate to
disable the duty cycle stabilizer. In all other applications, enabling
the DCS circuit is recommended to maximize ac performance.
Jitter Considerations
High speed, high resolution ADCs are sensitive to the quality
of the clock input. For inputs near full scale, the degradation in
SNR from the low frequency SNR (SNR
frequency (f
SNR
) due to jitter (t
INPUT
= −10 log[(2π × f
HF
JRMS
INPUT
) at a given input
LF
) can be calculated by
× t
)2 + 10
JRMS
]
In the equation, the rms aperture jitter represents the clock input
jitter specification.
Improvements in SNR can be achieved for IF undersampling
applications by minimizing the effects of aperture jitter. This
can be accomplished by applying a high frequency clock input
and using the integrated clock divider to achieve the desired
sample rate of the ADC core. Inherently, the jitter performance
of the AD9650 improves as the frequency of the clock increases.
This is a result of the slew rate of the clock affecting the noise
performance of the ADC, where fast transition edges result in
the best performance. Figure 92 shows the improvement in SNR
for the different clock divide ratios for the 1 V p-p and 2 V p-p
sinusoidal clock inputs. Measurements in Figure 92 were taken
for the AD9650BCPZ-105 where the input frequency was
141 MHz. The same analysis can be performed for the various
speed grades of the AD9650 family of parts.
Figure 92. SNR vs. CLK Divide Ratio for f
Sample Rate of 105 MSPS
The clock input should be treated as an analog signal in cases in
which aperture jitter may affect the dynamic range of the AD9650.
Power supplies for clock drivers should be separated from the
ADC output driver supplies to avoid modulating the clock signal
with digital noise. Low jitter, crystal-controlled oscillators make
the best clock sources.
Refer to the AN-501 Application Note and the AN-756 Application
Note (visit www.analog.com) for more information about jitter
performance as it relates to ADCs.
CHANNEL/CHIP SYNCHRONIZATION
The AD9650 has a SYNC input that offers the user flexible
synchronization options for synchronizing the clock divider.
The clock divider sync feature is useful for guaranteeing synchronized sample clocks across multiple ADCs. The input clock
divider can be enabled to synchronize on a single occurrence of
the SYNC signal or on every occurrence.
The SYNC input is internally synchronized to the sample clock;
however, to ensure that there is no timing uncertainty between
multiple parts, the SYNC input signal should be externally synchronized to the input clock signal, meeting the setup and hold
times shown in Tab l e 5. The SYNC input should be driven using
a single-ended CMOS-type signal.
POWER DISSIPATION AND STANDBY MODE
The power dissipated by the AD9650 varies with its sample rate.
In CMOS output mode, the digital power dissipation is determined
primarily by the strength of the digital drivers and the load on
each output bit.
The maximum DRVDD current (IDRVDD) can be calculated as
IDRVDD = VDRVDD × C
where N is the number of output bits (32 plus two DCO outputs
in the case of the AD9650).
This maximum current occurs when every output bit switches on
every clock cycle, that is, a full-scale square wave at the Nyquist
frequency of f
/2. In practice, the DRVDD current is estab-
CLK
LOAD
= 141 MHz and a
IN
× f
× N
CLK
lished by the average number of output bits switching, which is
Rev. A | Page 34 of 44
Data Sheet AD9650
Voltage at Pin
SCLK/DFS
SDIO/DCS
AVDD
Twos complement
DCS enabled
determined by the sample rate and the characteristics of the
analog input signal.
Reducing the capacitive load presented to the output drivers
reduces digital power consumption.
By asserting PDWN (either through the SPI port or by asserting
the PDWN pin high), the AD9650 is placed in power-down
mode. In this state, the ADC typically dissipates 3.3 m W.
During power-down, the output drivers are placed in a high
impedance state. Asserting the PDWN pin low returns the
AD9650 to its normal operating mode.
Low power dissipation in power-down mode is achieved by
shutting down the reference, reference buffer, biasing networks,
and clock. Internal capacitors are discharged when entering powerdown mode and must be recharged when returning to normal
operation.
When using the SPI port interface, the user can place the ADC
in power-down mode or standby mode. Standby mode allows
the user to keep the internal reference circuitry powered when
faster wake-up times are required.
DIGITAL OUTPUTS
The AD9650 output drivers can be configured to interface with
1.8 V CMOS logic families. The AD9650 can also be configured
for LVDS outputs (standard ANSI or reduced output swing mode)
using a DRVDD supply voltage of 1.8 V.
In CMOS output mode, the output drivers are sized to provide
sufficient output current to drive a wide variety of logic families.
However, large drive currents tend to cause current glitches on
the supplies that may affect converter performance.
Applications requiring the ADC to drive large capacitive loads
or large fanouts may require external buffers or latches.
The default output mode is CMOS, with each channel output
on a separate bus, as shown in Figure 2. The output can also be
configured for interleaved CMOS via the SPI port. In interleaved
CMOS mode, the data for both channels is output through the
Channel A output pins, and the Channel B output is placed into
high impedance mode. The timing diagram for interleaved
CMOS output mode is shown in Figure 3.
The output data format can be selected for either offset binary
or twos complement by setting the SCLK/DFS pin when operating
in the external pin mode (see Ta b l e 12).
As detailed in the AN-877 Application Note, Interfacing to High
Speed ADCs via SPI, the data format can be selected for offset
binary, twos complement, or gray code when using the SPI control.
The AD9650 has a flexible three-state ability for the digital output
pins. The three-state mode is enabled using the OEB pin or
through the SPI. If the OEB pin is low, the output data drivers and
DCOs are enabled. If the OEB pin is high, the output data drivers
and DCOs are placed in a high impedance state. This OEB
function is not intended for rapid access to the data bus. Note
that OEB is referenced to the digital output driver supply
(DRVDD) and should not exceed that supply voltage.
When using the SPI, the data outputs and DCO of each channel
can be independently three-stated by using the output enable
bar bit (Bit 4) in Register 0x14.
TIMING
The AD9650 provides latched data with a pipeline delay of
12 clock cycles. Data outputs are available one propagation
delay (t
The length of the output data lines and loads placed on them
should be minimized to reduce transients within the AD9650.
These transients can degrade converter dynamic performance.
The lowest typical conversion rate of the AD9650 is 10 MSPS.
At clock rates below 10 MSPS, dynamic performance can
degrade.
Data Clock Output (DCO)
The AD9650 provides two data clock output (DCO) signals
intended for capturing the data in an external register. In CMOS
output mode, the data outputs are valid on the rising edge of DCO,
unless the DCO clock polarity has been changed via the SPI. In
LVDS output mode, the DCO and data output switching edges
are closely aligned. Additional delay can be added to the DCO
output using SPI Register 0x17 to increase the data setup time.
In this case, the Channel A output data is valid on the rising
edge of DCO, and the Channel B output data is valid on the
falling edge of DCO. See Figure 2, Figure 3, and Figure 4 for a
graphical timing description of the output modes.
The AD9650 includes built-in test features designed to enable
verification of the integrity of each channel as well as facilitate
board level debugging. A BIST (built-in self-test) feature is included
that verifies the integrity of the digital datapath of the AD9650.
Various output test options are also provided to place predictable
values on the outputs of the AD9650.
BUILT-IN SELF-TEST (BIST)
The BIST is a thorough test of the digital portion of the selected
AD9650 signal path. When enabled, the test runs from an internal
pseudorandom noise (PN) source through the digital datapath
starting at the ADC block output. The BIST sequence runs for
512 cycles and stops. The BIST signature value for Channel A or
Channel B is placed in Register 0x24 and Register 0x25. If one
channel is chosen, its BIST signature is written to the two registers.
If both channels are chosen, the results from Channel A are placed
in the BIST signature registers.
The outputs are not disconnected during this test; therefore, the
PN sequence can be observed as it runs. The PN sequence can
be continued from its last value or reset from the beginning,
based on the value programmed in Register 0x0E, Bit 2. The
BIST signature result varies based on the channel configuration.
OUTPUT TEST MODES
The output test modes are shown in Ta ble 17. When an output
test mode is enabled, the analog section of the ADC is disconnected from the digital back end blocks and the test pattern is run
through the output formatting block. Some of the test patterns are
subject to output formatting, and some are not. The seed value for
the PN sequence tests can be forced if the PN reset bits are used
to hold the generator in reset mode by setting Bit 4 or Bit 5 of
Register 0x0D. These tests can be performed with or without
an analog signal (if present, the analog signal is ignored), but
they do require an encode clock. For more information, see the
AN-877 Application Note, Interfacing to High Speed ADCs via SPI.
Rev. A | Page 36 of 44
Data Sheet AD9650
DON’T CARE
DON’T CAREDON’T CARE
DON’T CARE
SDIO
SCLK
CSB
t
S
t
DH
t
CLK
t
DS
t
H
R/WW1W0A12A11A10A9A8A7
D5D4D3D2D1D0
t
LOW
t
HIGH
08919-052
SERIAL PORT INTERFACE (SPI)
The AD9650 serial port interface (SPI) allows the user to configure
the converter for specific functions or operations through a
structured register space provided inside the ADC. The SPI
gives the user added flexibility and customization, depending on
the application. Addresses are accessed via the serial port and
can be written to or read from via the port. Memory is organized
into bytes that can be further divided into fields, which are documented in the Memory Map section. For detailed operational
information, see the AN-877 Application Note, Interfacing to High Speed ADCs via SPI.
CONFIGURATION USING THE SPI
Three pins define the SPI of this ADC: the SCLK/DFS pin, the
SDIO/DCS pin, and the CSB pin (see Ta ble 14). The SCLK/DFS
(a serial clock) is used to synchronize the read and write data
presented from and to the ADC. The SDIO/DCS (serial data
input/output) is a dual-purpose pin that allows data to be sent
to and read from the internal ADC memory map registers. The
CSB (chip select bar) is an active-low control that enables or
disables the read and write cycles.
Table 14. Serial Port Interface Pins
Pin Function
SCLK
SDIO
CSB
Serial clock. The serial shift clock input, which is used to
synchronize serial interface reads and writes.
Serial data input/output. A dual-purpose pin that
typically serves as an input or an output, depending on
the instruction being sent and the relative position in the
timing frame.
Chip select bar. An active-low control that gates the read
and write cycles.
The falling edge of the CSB, in conjunction with the rising edge
of the SCLK, determines the start of the framing. An example of
the serial timing and its definitions can be found in Figure 93
and Tabl e 5.
Other modes involving the CSB are available. When the CSB is
held low indefinitely, which permanently enables the device,
this is called streaming. The CSB can stall high between bytes to
allow for additional external timing. When CSB is tied high, SPI
functions are placed in high impedance mode. This mode turns
on any SPI secondary pin functions.
During an instruction phase, a 16-bit instruction is transmitted.
Data follows the instruction phase, and its length is determined
by the W0 and W1 bits.
In addition to word length, the instruction phase determines
whether the serial frame is a read or write operation, allowing
the serial port to be used both to program the chip and to read
the contents of the on-chip memory. The first bit of the first byte in
a multibyte serial data transfer frame indicates whether a read
command or a write command is issued. If the instruction is a
readback operation, performing a readback causes the serial
data input/output (SDIO) pin to change direction from an input to
an output at the appropriate point in the serial frame.
All data is composed of 8-bit words. Data can be sent in MSB-first
mode or in LSB-first mode. MSB first is the default on power-up
and can be changed via the SPI port configuration register. For
more information about this and other features, see the AN-877
Application Note, Interfacing to High Speed ADCs via SPI.
Figure 93. Serial Port Interface Timing Diagram
Rev. A | Page 37 of 44
AD9650 Data Sheet
AGND (default)
Outputs enabled
HARDWARE INTERFACE
The pins described in Ta ble 14 comprise the physical interface
between the user programming device and the serial port of the
AD9650. The SCLK pin and the CSB pin function as inputs
when using the SPI. The SDIO pin is bidirectional, functioning
as an input during write phases and as an output during
readback.
The SPI is flexible enough to be controlled by either FPGAs or
microcontrollers. One method for SPI configuration is
described in detail in the AN-812 Application Note, Micro-controller-Based Serial Port Interface (SPI) Boot Circuit.
The SPI port should not be active during periods when the full
dynamic performance of the converter is required. Because the
SCLK signal, the CSB signal, and the SDIO signal are typically
asynchronous to the ADC clock, noise from these signals can
degrade converter performance. If the on-board SPI bus is used for
other devices, it may be necessary to provide buffers between
this bus and the AD9650 to prevent these signals from transitioning at the converter inputs during critical sampling periods.
Some pins serve a dual function when the SPI is not being used.
When the pins are strapped to AVDD or ground during device
power-on, they are associated with a specific function. The
Digital Outputs section describes the strappable functions
supported on the AD9650.
CONFIGURATION WITHOUT THE SPI
In applications that do not interface to the SPI control registers,
the SDIO/DCS pin, the SCLK/DFS pin, the OEB pin, and the
PDWN pin serve as standalone CMOS-compatible control pins.
When the device is powered up, it is assumed that the user
intends to use the pins as static control lines for the duty cycle
stabilizer, output data format, output enable, and power-down
feature control. In this mode, the CSB chip select bar should be
connected to AVDD, which disables the serial port interface.
When the device is in SPI mode, the PDWN and OEB pins
remain active. For SPI control of output enable and power-down,
the OEB and PDWN pins should be set to their default states.
Tabl e 16 provides a brief description of the general features that
are accessible via the SPI. These features are described in detail in
the AN-877 Application Note, Interfacing to High Speed ADCs via SPI. The AD9650 part-specific features are described in detail
following Tab le 17, the external memory map register table.
Table 16. Features Accessible Using the SPI
Feature Name Description
Mode
Clock
Offset
Test I/O
Output Mode
Output Phase Allows the user to set the output clock polarity
Output Delay Allows the user to vary the DCO delay
VREF Allows the user to set the reference voltage
Allows the user to set either power-down mode
or standby mode
Allows the user to access the DCS, set the
clock divider, set the clock divider phase, and
enable the sync
Allows the user to digitally adjust the
converter offset
Allows the user to set test modes to have
known data on output bits
Allows the user to set the output mode,
including LVDS
Rev. A | Page 38 of 44
Data Sheet AD9650
MEMORY MAP
READING THE MEMORY MAP REGISTER TABLE
Each row in the memory map register table has eight bit locations.
The memory map is roughly divided into four sections: the chip
configuration registers (Address 0x00 to Address 0x02); the
channel index and transfer registers (Address 0x05 and
Address 0xFF); the ADC functions registers, including setup,
control, and test (Address 0x08 to Address 0x30); and the digital
feature control register (Address 0x100).
The memory map register table (see Tab le 17) documents the
default hexadecimal value for each hexadecimal address shown.
The column with the heading Bit 7 (MSB) is the start of the
default hexadecimal value given. For more information on this
function and others, see the AN-877 Application Note, Interfacing to High Speed ADCs via SPI. This application note details the
functions controlled by Register 0x00 to Register 0xFF. The
remaining register, Register 0x100, is documented in the Memory
Map Register Table section.
Open Locations
All address and bit locations that are not included in Tabl e 17
are not currently supported for this device. Unused bits of a
valid address location should be written with 0s. Writing to these
locations is required only when part of an address location is
open (for example, Address 0x17).
Default Values
After the AD9650 is reset, critical registers are loaded with
default values. The default values for the registers are given in
the memory map register table, Ta bl e 17.
Logic Levels
An explanation of logic level terminology follows:
•“Bit is set” is synonymous with “bit is set to Logic 1” or
“writing Logic 1 for the bit.”
•“Clear a bit” is synonymous with “bit is set to Logic 0” or
“writing Logic 0 for the bit.”
Transfer Register Map
Address 0x08 through Address 0x18 and Address 0x30 are
shadowed. Writes to these addresses do not affect part
operation until a transfer command is issued by writing 0x01 to
Address 0xFF, setting the transfer bit. This allows these registers
to be updated internally and simultaneously when the transfer
bit is set. The internal update takes place when the transfer bit is
set, and the bit autoclears.
Channel-Specific Registers
Some channel setup functions, such as the signal monitor
thresholds, can be programmed differently for each channel. In
these cases, channel address locations are internally duplicated for
each channel. These registers and bits are designated in Tabl e 17
as local. These local registers and bits can be accessed by setting
the appropriate Channel A or Channel B bits in Register 0x05.
If both bits are set, the subsequent write affects the registers of
both channels. In a read cycle, only Channel A or Channel B
should be set to read one of the two registers. If both bits are set
during an SPI read cycle, the part returns the value for Channel A.
Registers and bits designated as global in Tab le 17 affect the entire
part or the channel features for which independent settings are not
allowed between channels. The settings in Register 0x05 do not
affect the global registers and bits.
Rev. A | Page 39 of 44
AD9650 Data Sheet
MEMORY MAP REGISTER TABLE
All address and bit locations that are not included in Tab l e 17 are not currently supported for this device.
Table 17. Memory Map Registers
Default
Address
(Hex)
Chip Configuration Registers
0x00 SPI port
0x01 Chip ID
0x02 Chip grade
Channel Index and Transfer Registers
0x05 Channel
0xFF Transfer Open Open Open Open Open Open Open Transfer 0x00 Synchronously
ADC Functions
0x08 Power
0x09 Global clock
0x0B Clock divide
Register
Name
configuration
(global)
(global)
(global)
index
modes (local)
(global)
(global)
Bit 7
(MSB)
0 LSB first Soft reset 1 1 Soft reset LSB first 0 0x18 The nibbles
register is set,
the test data
is placed on
the output
pins in place of
normal data.
outputs and
the format of
the data.
selection of
clock delays
into the input
clock divider.
Rev. A | Page 41 of 44
AD9650 Data Sheet
MEMORY MAP REGISTER DESCRIPTIONS
For additional information about functions controlled in
Register 0x00 to Register 0xFF, see the AN-877 Application Note,
Interfacing to High Speed ADCs via SPI.
SYNC Control (Register 0x100)
Bits[7:3]—Reserved
Bit 2—Clock Divider Next SYNC Only
If the master SYNC enable bit (Address 0x100, Bit 0) and the clock
divider SYNC enable bit (Address 0x100, Bit 1) are high, Bit 2
allows the clock divider to synchronize to the first SYNC pulse it
receives and to ignore the rest. The clock divider SYNC enable bit
(Address 0x100, Bit 1) resets after it synchronizes.
Bit 1—Clock Divider SYNC Enable
Bit 1 gates the SYNC pulse to the clock divider. The SYNC
signal is enabled when Bit 1 is high and Bit 0 is high. This is
continuous SYNC mode.
Bit 0—Master SYNC Enable
Bit 0 must be high to enable any of the SYNC functions. If the
SYNC capability is not used, this bit should remain low to
conserve power.
Rev. A | Page 42 of 44
Data Sheet AD9650
APPLICATIONS INFORMATION
DESIGN GUIDELINES
Before starting design and layout of the AD9650 as a system,
it is recommended that the designer become familiar with these
guidelines, which discuss the special circuit connections and
layout requirements that are needed for certain pins.
Power and Ground Recommendations
When connecting power to the AD9650, it is recommended that
two separate 1.8 V supplies be used. Use one supply for the analog
outputs (AVDD ); use a separate supply for the digital outputs
(DRVDD). For both AVDD and DRVDD, several different
decoupling capacitors should be used to cover both high and
low frequencies. Place these capacitors close to the point of
entry at the PCB level and close to the pins of the part, with
minimal trace length.
A single PCB ground plane should be sufficient when using the
AD9650. With proper decoupling and smart partitioning of the
PCB analog, digital, and clock sections, optimum performance
is easily achieved.
LVDS Operation
The AD9650 defaults to CMOS output mode on power-up.
If LVDS operation is desired, this mode must be programmed,
using the SPI configuration registers after power-up. When the
AD9650 powers up in CMOS mode with LVDS termination
resistors (100 Ω) on the outputs, the DRVDD current can be
higher than the typical value until the part is placed in LVDS
mode. This additional DRVDD current does not cause damage
to the AD9650, but it should be taken into account when considering the maximum DRVDD current for the part.
To avoid this additional DRVDD current, the AD9650 outputs
can be disabled at power-up by taking the OEB pin high. After
the part is placed in LVDS mode via the SPI port, the OEB pin
can be taken low to enable the outputs.
Exposed Paddle Thermal Heat Slug Recommendations
It is mandatory that the exposed paddle on the underside of the
ADC be connected to analog ground (AGND) to achieve the
best electrical and thermal performance. A continuous, exposed
(no solder mask) copper plane on the PCB should mate to the
AD9650 exposed paddle, Pin 0.
The copper plane should have several vias to achieve the lowest
possible resistive thermal path for heat dissipation to flow through
the bottom of the PCB. These vias should be filled or plugged to
prevent solder wicking through the vias, which can compromise
the connection.
To maximize the coverage and adhesion between the ADC and
the PCB, a silkscreen should be overlaid to partition the continuous
plane on the PCB into several uniform sections. This provides
several tie points between the ADC and the PCB during the reflow
process. Using one continuous plane with no partitions guarantees
only one tie point between the ADC and the PCB. For detailed
information about packaging and PCB layout of chip scale
packages, see the AN-772 Application Note, A Design and
Manufacturing Guide for the Lead Frame Chip Scale Package
(LFCSP), at www.analog.com.
VCM
The VCM pin should be decoupled to ground with a 0.1 μF
capacitor, as shown in Figure 80.
RBIAS
The AD9650 requires that a 10 kΩ resistor be placed between
the RBIAS pin and ground. This resistor sets the master current
reference of the ADC core and should have at least a 1% tolerance.
Reference Decoupling
The VREF pin should be externally decoupled to ground with
a low ESR, 1.0 μF capacitor in parallel with a low ESR, 0.1 μF
ceramic capacitor.
SPI Port
The SPI port should not be active during periods when the full
dynamic performance of the converter is required. Because the
SCLK, CSB, and SDIO signals are typically asynchronous to the
ADC clock, noise from these signals can degrade converter
performance. If the on-board SPI bus is used for other devices,
it may be necessary to provide buffers between this bus and the
AD9650 to keep these signals from transitioning at the converter
inputs during critical sampling periods.
Rev. A | Page 43 of 44
AD9650 Data Sheet
COMPLIANT TO JEDEC STANDARDS MO-220-VMMD-4
0.22 MIN
TOP VIEW
8.75
BSC SQ
9.00
BSC SQ
1
64
16
17
49
48
32
33
0.50
0.40
0.30
0.50
BSC
0.20 REF
12° MAX
0.80 MAX
0.65 TYP
1.00
0.85
0.80
7.50
REF
0.05 MAX
0.02 NOM
0.60 MAX
0.60
MAX
SEATING
PLANE
PIN 1
INDICATOR
7.55
7.50 SQ
7.45
PIN 1
INDICATOR
0.30
0.23
0.18
FOR PROPER CONNECTION OF
THE EXPOSED PAD, REFER TO
THE PIN CONFIGURATION AND
FUNCTION DESCRIPTIONS
SECTION OF THIS DATA SHEET.