Title: Optimize Your SAR ADC Design
1Optimize Your SAR ADC Design
- Bonnie Baker
- Senior Applications Engineer
- bonnie_at_ti.com
Special Thanks for Inputs from Tim Green Rick
Downs Miro Oljaca
2SAR ADC System Design
SAR
OpAmp Signal Bandwidth Slew Rate Output
Impedance
Filter Charge Reservoir Capacitor Load
Isolation Noise Filtering
ADC Acquisition Time Data Rate Resolution ADC
Input Topology ADC Ref In
3The Design Tools We Will Use
- Data Sheet Parameters .
- Rules of Thumb .
- Tricks and Tips .
- Testing .
4Design Procedure
V
V
CC
REF
Op Amp
A/D
Filter
-
D
OUT
0 V
C
FLT
R
FLT
V
S
51. Define Input Signal
VCC
VREF
Op Amp
- Highest Frequency
- 1 kHz (single channel)
- Largest Voltage Swing
- 0 to 4.096 V
- High Accuracy
- 62.5 mV LSB size or 16-bit with range of 4.096
A/D
Filter
-
DOUT
0 V
CFLT
RFLT
VS
62. Selecting the ADC
- Things we need to know
- Sampling frequency gt 50 ksps
- Full-scale input range (FSR) 4.096V
- Highest Resolution 16-bit
- SAR Architecture no latency
7ADS8320 Application Specs
- 16-bit, 100 kHz Micropower, Sampling
Analog-To-Digital Converter - Throughput Rate (Sampling Rate) 100 ksps
- tACQ (min) 1.88 µs
- Input VFSR VREF 4.096 V
- CSH (input sample hold capacitance) 45 pF
- Secondary specifications
- SNR 88 dB _at_ 1 kHz
- THD -86 dB _at_ 1 kHz
- SINAD 84 dB _at_1 kHz
- SFDR 86 dB _at_1 kHz
8A/D Converter Terms
- Acquisition Time (tACQ)
- The time the internal A/D sample capacitor is
connected to the A/D input analog signal - Conversion Time (tCONV)
- The time the A/D requires to convert the sampled
analog input to a digital output after the
acquisition time (tACQ) is complete - Throughput Rate Sampling Rate
- Maximum frequency at which A/D conversions can be
repeated - Is equal to the Acquisition time plus the
Conversion time (tACQ tCONV) - i.e. 100 ksps Throughput Rate Sampling Rate
implies that an input analog signal may be
converted every 10µs
9SAR Acquisition and Conversion Time
S1
10Single-pole, Time Constant Multiplier
11ADC System Model
A/D
RFLT
RSW
?
100W
CSH
S1
CFLT
20pF- 50pF
VS
?
S2
12SAR ADC Input Charge Distribution
- Op amp requirements
- Must charge the ADC cap quickly accurately
ADS8361
13ADC Conclusion
- Key ADC specifications per Input Signal
- Sampling Rate 100 ksps
- Full-scale input range 4.096 V
- In following discussion we will use the ADS8320
- tACQ ADC acquisition time
- CSH ADC input capacitance
- k 16-bit time constant multiplier
- VFSR Full-scale input range of ADC
143. Choosing CFLT
VCC
VREF
Op Amp
- Provides charge to ADC sampling capacitor, CSH
- CFLT is the charge reservoir
A/D
Filter
-
DOUT
0 V
CFLT
VS
15ADC Specs for CFLT Determination
tACQ
ADS8320
- Need to Know CSH, tACQ, k, VFSR
- tACQ 1.88 ms
- CSH (sampling ADC input capacitor) 45pF
- k 12
- Worst case ?V across CSH is VFSR
- VFSR VREF 4.096 V
16Capacitor Charge Transfer
- Prior to acquisition VIN ¹ VSH
- During acquisition
- CFLT and CSH exchange charge
- VSH changes to equal VIN
VIN
Charge Reservoir
17Ideal Value for CFLT
- Charge Transfer Equation Q C x V
- QSH CSH x VFSR
- QSH 45pF x 4.096V 184pC
- IDEAL CFLT
- Charge reservoir fills CSH with 1/2LSB from VS
droop on CFLT - 1/2LSB of VFSR VFSR / 2N (worst case)
- 1/2LSB of VFSR 4.096 V / 216 31.25 µV
- QSH QFLT
- QFLT CFLT x (1/2LSB from FSR)
- 184pC CFLT x (31.25µV) ? CFLT 5.9 µF
-
18IDEAL CFLT 5.9 µF Assessment
- The Driving Op Amp probably
- Can not directly drive a 5.9 mF capacitor
- Circuit may be marginally stable
- Could have transient current problems
- With resistor between Op Amp and capacitor
- Resistor and capacitor still need to meet signal
bandwidth - Resistor may not be large enough to help isolate
CFLT
19CFLT Suggested Modification
- Starting Point Partition the charge reservoir
- 95 from CFLT
- 5 from Op Amp
- CFLT value provides QSH with lt5 droop on CFLT
- QFLT QSH
- QFLT CFLT x (0.05 VFSR)
- 184pC CFLT x (0.05 x 4.096V) ? CFLT 898 pF
- Well use 1000pF or 1 nF
- Ensure CFLT ³ 20 x CSH
- During tACQ the Op Amp must be able to replace 5
VFSR on CFLT
20Capacitor Voltage Coefficient
- Voltage coefficient causes distortion
- C C0 x (1bVCAP)
- Also has non-linear error proportional to
frequency - Voltage and frequency coefficients impact ADC
distortion -
-
-
-
-
-
-
-
Source Murata
21Capacitor THDN versus Frequency
-60
-65
-70
-75
-80
-85
THD N (dB)
System Measurement
-90
-95
-100
-105
-110
200
500
1k
20k
50k
10k
10k
2k
5k
200k
100
Silver Mica
22CFLT Type and Value Requirements
- For our example CFLT 1 nF
- High quality capacitor
- Low Voltage Coefficient
- Low Frequency Coefficient
- Capacitor Type C0G
- CFLT gt 20 x CSH
- gt95 of Charge to ADC from CFLT
- lt 5 of Charge to ADC from OPA
- Dominant load of Op amp is CFLT
- Droop on CFLT lt 5
234. Choosing RFLT
VCC
VREF
- Things we need to know
- tACQ ADC acquisition time (1.88 ms)
- k Single-pole time constant multiplier for
16-bit ADC (12) - CFLT External input capacitor to ADC (1 nF)
Op Amp
A/D
Filter
-
DOUT
0 V
CFLT
RFLT
VS
24First Pass Determination of RFLT Value
- First Pass RFLT Calculation
- tFLT RFLT x CFLT Filter time constant
- tACQ ³ k x tFLT
- 1.88 ms 12 x tFLT
- ? tFLT 157 ns
25RFLT Value with 40 Margin
- Given tACQ ³ k x tFLT
- Design in a margin of 40
- 60 x tACQ k x tFLT
- Margin for
- Op Amp Output Load Transient
- Op Amp Output Small Signal Settling Time
- RFLT ³ (0.60 x tACQ ) / ( k x CFLT)
- RFLT ³ (0.60 x 1.88 ns)/(12 x 1 nF)
- ? RFLT ³ 94.2 ?
- Use RFLT 100?
265. Choosing Op Amp
VCC
VREF
- Things we need to know
- RFLT External input resistor to ADC (100 W)
- CFLT External input capacitor to ADC (1 nF)
- tACQ ADC acquisition time (1.88 ms)
Op Amp
A/D
Filter
-
DOUT
0 V
CFLT
RFLT
VS
27Primary Op Amp Buffer Specs
28Secondary Op Amp Buffer Specs
- Total Harmonic Distortion (THD)
- Low Noise for 16-bit performance
- Closed Loop Gain Error
- Peak Output Current
- Input Cross-over Distortion
29OP Amp Buffer Application Specs
- Application
- Single Supply 5V
- Buffer NO CM Input Crossover
- Slew Rate to track 1kHz Input
- Wideband for good gain flatness 1kHz, G1
- Wideband for fast transient response to Noise
Filter Transients - Low Noise for 16 Bit performance
- RRIO for 65mV to 4.935V Input and Output on 5V
Supply - Best Industry Choice
- OPA363 or OPA364 (OPA363 with Shutdown feature)
- 1.8V, 7MHz, 90dB CMRR, Single-Supply,
Rail-To-Rail I/O
30OPA363/OPA364 Application Specs
- SRMIN (V/µs) 2 p f x VOP (1e-6)
- Minimum Slew Rate to track input sine wave (_at_ lt1
Distortion) - SRMIN 2x p x 1kHz x (4.096Vpp/2) x (1e-6)
0.013V/µs - OPA363/OPA364 5V/µs
- Choose Op Amp SROPA gt 2 X SRMIN
- Gain Error
- ACL AOL/(1Aolß)
- AOL _at_1kHz 80dB 10000
- ß 1 for Unity Gain Follower
- ACL 10,000/(1100001) 0.99990001
- 0.009999 Gain Error _at_ 1kHz
- 12 Bit (1/2 LSB Accuracy)
- Calibrate gain error at system level
- Many systems are more concerned about relative
changes than absolute
31Systems Transients
- Input Transient Load Transient
- Input Step Voltage Output Step Voltage
- Output Voltage Slew Rate Output Step Current
Line Transient
Load Transient
IT
OPA364
32Calculating the Op Amp Bandwidth
- Calculate Unity Gain BW
- Select Op Amp BW
- GBWP gt 4 X fFLT f-3db
- fFLT f-3db 1/2?RFLT x CFLT
- fFLT f-3db 1/2? x ?100W x 1 nF 1.6 MHz
- Op Amp GBWP gt4 x 1.6 MHz 6.4 MHz
33OpAmp Filter Small Signal
VFLT
VOA
ADS8320
-
RFLT
107 W
100 W
CFLT
1 nF
OPA364
34OpAmp Filter Frequency Response
35OpAmp Filter Stability
- Buffer Closed Loop Gain Bandwidth as modified by
RFLTCFLT - fCL 3.2 MHz
- Stability Check
- At fCL 3.2 MHz Rate-of-closure is 20dB/decade
? fZX cancels fPX before fCL - fCL gt 2 x fFLT-3dB
- fPX and fZX are lt decade apart
- Phase of pole will be cancelled by phase of
zero - RO 9 x RFLT
- fFLT-3db 1/2?RFLT x CFLT 1.6MHz
36Amplifier Selection Guidelines
- Stability insured and Bandwidth exceeds input
signal - GBWP gt 4 x fS(MAX)
- GBWP ³ 2 / (p CFLT x RFLT)
- RO 9 x RFLT
- fCL gt 2 x fFLT-3dB
- Slews fast enough for the input signal
- SROPA gt 4p x fS(MAX) x VOP x (1e-6)
37Final OPA-SAR Circuit Design
VCC 5V
VREF 4.096V
VCC
RFLT 100W
-
IN
OPA364
Digital Out
ADS8320
- IN
CFLT 1 nF
VS
38ADS8320 On Test System
ADS8320 SNR 88dB ENOB 14.33
Input Frequency 1kHz
Input Amplitude -0.25 dB
39OPA364 Filter ADS8320
Op Amp Filter ADS8320 SNR 87.78dB ENOB
14.19
Input Frequency 1kHz
Input Amplitude -0.25 dB
40Comparison of Tests
ADS8320 SNR 88dB ENOB 14.33
ADS8320 Only
OPA364, Filter, ADS8320
Op Amp Filter ADS8320 SNR 87.78dB ENOB
14.19
41Design Equations SAR-ADC System
- Filter Capacitor and Resistor Values
- CFLT gt 20 x CSH
- RFLT ³ (0.60 x tACQ ) / ( k x CFLT)
- Select Amplifier
- GBWP gt 4 x fS(MAX)
- GBWP ³ 2 / (p CFLT x RFLT)
- RO 9 x RFLT
- fCL gt 2 x fFLT-3dB
- SROPA gt 4p x fS(MAX) x VOP x (1e-6)
42Summary Design Steps
V
V
CC
REF
Op Amp
A/D
Filter
-
D
OUT
0 V
R
FLT
C
FLT
V
S
43REFERENCES
- Reference 1 Baker, B., Oljaca, M., Use External
components to Improve the Accuracy of a SAR ADC,
EDN 2007 - Reference 2 Green, Tim, Operational Amplifier
Stability Part 3 of 15 RO and ROUT,
AnalogZone Acquisition Zone - Reference 3 Oljaca, M. and McEldowney, J.
(2002.) Using a SAR Analog-to-Digital Converter
for Current Measurement in Motor Control
Applications. Burr-Brown Application Report
SBAA081. - Reference 4 Downs, R. and Oljaca, M. (2005)
Designing SAR ADC Drive Circuitry Part II.
AnalogZONE Acquisition Zone. - Reference 5 Downs, R., Oljaca, M., Designing
SAR ADC Circuitry Part 1 A Detailed Look at
SAR ADC Operation, AnalogZone Acquisition Zone - Reference 6 Downs, R., Oljaca, M., Designing
SAR ADC Drive Circuitry Part 3 Designing the
Optimal Input Drive Circuit for SAR ADCs,
AnalogZone Acquisition Zone - Reference 7 Green, T. (2005) Operational
Amplifier Stability - Part 6 of 15
Capacitance-Load Stability RISO, High Gain CF,
Noise Gain AnalogZONE Acquisition Zone. - Reference 8 Baker, Bonnie, A Glossary of
Analog-to-Digital Specifications and Performance
Characteristics, Texas Instruments, SBAA147
44Optimize Your SAR ADC Design