thermal noise estimation in switched-capacitor circuits
DESCRIPTION
THERMAL NOISE ESTIMATION IN SWITCHED-CAPACITOR CIRCUITS. Jos é Silva and G ábor C. Temes School of Electrical Engineering and Computer Science Oregon State University. Outline. Thermal noise effects in an SC integrator Switch (kT/C) noise Opamp thermal noise - PowerPoint PPT PresentationTRANSCRIPT
12/21/2004 [email protected] 1/22
THERMAL NOISE ESTIMATION IN THERMAL NOISE ESTIMATION IN
SWITCHED-CAPACITOR SWITCHED-CAPACITOR
CIRCUITSCIRCUITS
José Silva and Gábor C. Temes
School of Electrical Engineering and
Computer Science
Oregon State University
12/21/2004 [email protected] 2/22
Outline
• Thermal noise effects in an SC
integrator
– Switch (kT/C) noise
– Opamp thermal noise
• Noise calculations in ADCs
• Noise calculation Example
12/21/2004 [email protected] 3/22
Noise Effects In SC Integrators• The thermal noise sources are the switches and the opamp.
• Flicker noise is negligible if fcorner << fs. If not, techniques such as correlated double sampling or chopper stabilization can be used.
C1
vout’vin
C2
vout2
21
1
1
S1
S2 S3
S4
mINg
kTPSD
3
16
onkTRPSD 4
12/21/2004 [email protected] 4/22
Switch Noise• Noise charge power in C1 (assuming ideal opamp):
vn1 C1Ron1
Ron3
vn3
vn4C1Ron4
Ron2
vn2
C2
vout
• MS noise charge into C2, in every clock period:
12
2,12
1,121 2kTCqqq
• This can be represented by an equivalent input voltage noise source vn1 with MS value:
1
21
2
C
kTvn
1
21
2
C
kTvn
End of 1 End of 2
12
1,1 kTCq 12
2,1 kTCq
12/21/2004 [email protected] 5/22
Op-amp Noise (1)
• Simplified method, ignoring switch resistance during Φ2=1
• Charge drawn by C1 from C2 in every clock period: C1 v-. This effect can be represented by equivalent input noise source vn2 = v-
C1
C2
vout
2
21
1
vn,eq
v-
- vC2 +
12/21/2004 [email protected] 6/22
Opamp Noise (2)• To find v-, assume a single-pole model for the op-amp with ωu=gm1/CL
2
12 13
4
C
C
C
kTv
Lout LCCC
kTv
21
2
13
4
• MS voltage noise of v-, by voltage division:• Output MS voltage noise:
• C1 will extract a charge C1v- from C2 in every clock period. This effect can be represented by an input source vn2.
• Since vout = v- + vC2, an output equivalent source vn3 is also required.
It represents the unity-gain output noise during Φ1=1
• Using = 1/(1+C1/C2) as the feedback factor:
L
n C
kTv
3
422
Ln C
kTv
3
422
Ln C
kTv
3
423
Ln C
kTv
3
423
12/21/2004 [email protected] 7/22
Integrator Noise Model
• Combining the effects of switch and opamp thermal noise:
vn1 vn2 vn3
vin vout
1
2
C
kT LC
kT
3
4
LC
kT
3
4
• Insert these noise sources into every integrator and SC branch.
12/21/2004 [email protected] 8/22
More Accurate Model
x
kTC
/111
)
1
6/11(2 1
21
x
kTCq
11 2,13
4mongRx
x
kTC
Total power acquired in Φ1 and Φ2:
Assuming an output-compensated opamp,
Noise power acquired by C1 during Φ2=1:
From Ron :
from vno :
During Φ2 :
,
12/21/2004 [email protected] 9/22
Design Considerations
1
3/ 122 x
CkTvn• In the model, now .The other two sources
remain unchanged. The overall input-referred noise is now about 2dB lower than that obtained from the simpler model.
• For several input capacitors, the noise charge powers add. For
two C1 branches, there is a 3 dB noise increase.
• The input – referred noise voltage v2c1 = q2
1 / C21 is minimized
for gm1 >> 1 / Ron. This costs added power. For gm1 << 1 / Ron,
the noise power is only 7/6 ~ 1.17 times larger.
12/21/2004 [email protected] 10/22
An Efficient Design Algorithm
)23/7(121
xv
kTgm
c
11)/12( CgR mon
1])1)(2ln2[( sfN
;2 2,
)1(21 effin
Nc vv
12 1 mongRx
onR
1mg
1C )1/(11 xgC m 1mg
21cv
From previous relations, at the end of Φ2 =1,,where
is the settling time constant of the stage, and is the
input-referred noise power .To minimize ,1. Choose ,the largest value allowed
for settling to N-bit accuracy; 2. Choose 3. Choose 4. Find and from eqs. above;
5. Find from
12/21/2004 [email protected] 11/22
Example (Integrator)
Calculated and simulated integrated noise powers at the output of the integrator. Simulation used [7].
• Let C1 = C2 = 2CL = 1 pF, gm1 = 4 mA / V, RL = 250kΩ, 2Ron = 0.5kΩ, fs = 100MHz. Then 2Ron C1 = C0 / (βgm1)= 0.5ns = 1 / 20fs , allowing for accurate settling.
• Integrating the output noise PSD over 0 to fB = fs / (2 · OSR) gives
f(Hz)
12/21/2004 [email protected] 12/22
Noise Calculations in ADCs
• Step 1: Identify noise sources in the topology.
• Step 3: Calculate transfer function from each noise source to output.
• Step 4: Integrate each noise PSD over desired bandwidth.
• Step 5: Total noise power is sum of all contributions
N
jojn vv
TOTAL1
22
OSRsf
dfNTFvf
v jjs
oj
2
0
222 2
• Step 2: Calculate PSDs of noise sources.
12/21/2004 [email protected] 13/22
Noise Budget
kT/C(50%)
opamp(25%)
other(20%)
Quantization(< 5%)
22222
OTHEROPAMPswqTOTAL nnnnn vvvvv
Quantization noise
Noise from other sources (digital, etc)
Switch (kT/C) noise
Opamp thermal noise
2
2
maxmax
TOTALn
u
v
vSNR
• Maximum SNR of a data converter:Maximum input
signal power
Total noise power
• The total noise power includes contributions from several sources:
• A good balance between these contributions:
12/21/2004 [email protected] 14/22
Example: Low-Distortion Topology
• Integrators do not process input signal, only quantization noise.
No signal No distortion.
• For MASH structures, quantization noise can be tapped directly from yi2.
• In conventional topologies, integrator nonlinearities are attenuated by loop.
For low oversampling ratios, this is not effective.
• Distortion can be avoided by making STF = 1:
121
212
2
HH
HHSTF
H(z)H(z) H(z)H(z)
u
vyi1 yi2
2
DACDAC
q
To next stage
212
121
1
zHH
NTF
Feedforward paths
12/21/2004 [email protected] 15/22
Noise Calculation Example (1)• Step 1: Identify thermal noise sources in the topology:
21
1
1
Ci1
CS1CS2
2 1
2
Ci2
1
2
CF3
2
1
CF2
CF1
2
1
QuantizerQuantizer
2.v 2.v
VREF+ VREF-
u
vq
12/21/2004 [email protected] 16/22
Noise Calculation Example (2)• Step 2: Calculate PSDs of noise sources:
H(z)H(z) H(z)H(z)
u
vyi1
yi2
2
DACDAC
q
1
3/2
1
1
1
21
x
CkT
C
kTv S
Sni
Lno C
kTv
3
421
1
3/2
2
2
2
22
x
CkT
C
kTv S
Sni
LF
FFFno C
kT
C
CCCkTv
3
42
1
32122
1niv 1nov
2nov
2niv
• Noise powers:
• The PSDs are obtained by dividing noise powers by fs/2.
12/21/2004 [email protected] 17/22
Noise Calculation Example (3)• Step 3: Calculate transfer function from each noise source to output:
H(z)H(z) H(z)H(z)
u
vyi1
yi2
2
DACDAC
q
1niv 1nov
2nov
2niv
• Assuming H(z) = z-1/(1 – z-1):
212
2
1 21
2
zzH
HHzNTFi
1122 1
1
zz
H
HzNTFi
2122 1
1
1)(
z
HzNTFo
1122 21
1
2
zzH
HzNTFo
12/21/2004 [email protected] 18/22
Noise Calculation Example (4)• Step 4: Integrate each noise PSD over desired bandwidth:
(Tip: To save time, use a symbolic analysis tool such as Maple™)
0 0.1 0.2 0.3 0.4 0.50
5
10
15
20
25
30
35
40
Mag
nitu
de
Normalized Frequency
OSR
fs2
|NTFi1(ejT)|2
|NTFo2(ejT)|2
|NTFo1(ejT)|2
|NTFi2(ejT)|2
OSR
v
OSROSRvdfNTF
f
vN ni
niis
nii
OSRsf
212
1
0
2
1
212
1 sin452 2
12/21/2004 [email protected] 19/22
Noise Calculation Example (5)
• Step 5: Finally, total noise is sum of all contributions:
22
22
21
21
2oioiTOTAL NNNNN
3
22
22
222 3
sin22
OSR
v
OSROSRvN ninii
5
22
42
222 5
sin8
cossin26
OSR
v
OSROSROSROSRvN nonoo
• Similarly:
3
21
22
121 3
sin18
sincos414
OSR
v
OSROSROSROSRvN nonoo
12/21/2004 [email protected] 20/22
Numerical Example
5.88 v2n = 8.25
8.25 0.75
1.2 pF
• Since Cs1 and CC1 dominate the noise performance, ignore the rest. • Assume maximum input power is 0.25V2 ,and 16-bit noise
performance is desired. Total allowed noise power:
• Assume the loop operates with OSR=256, and choose x=2. Then,
• Allocating 75% of total noise to the thermal noise:
2122 1038 rmstotal VN
12/21/2004 [email protected] 21/22
Additional Considerations• Wideband operation (OSR < 16):
– Input stage does not dominate noise. Optimization algorithms should be used to minimize total capacitance while meeting the noise target.
• Fully-differential circuits:– Use same total capacitance as for single-ended circuit. Noise power
increases by 3 dB for each side, and by 6 dB for differential mode. Signal power also increases by 6 dB, so total SNR is the same.
• MASH topologies:– Transfer functions should be calculated for whole system.
– Quantization noise is cancelled, and so is the noise from some sources.
• Calculations assume brick-wall decimation filter.– For more accuracy, actual transfer function of the decimation block can be
included in calculations.
• |STF(f)| = 1 was assumed. – Calculations are output referred. Signal power is affected by STF.
12/21/2004 [email protected] 22/22
References