I, for one, can’t stand the thought of having an electrolytic, unipolar or bipolar, in the feedback path of an amplifier. Here is a simple servo which allows you to get rid of the DC blocking capacitor in the feedback path. Is this cheaper than an offset pot? No. But it is a bit more elegant.
Some of the features are very small size and simple power requirements.
The opamp is an Analog Devices CMOS AD8603. Very low current requirement, good precision and reasonably quiet. Note that it is not chopper stabilized. It needs a power supply of +/- 2.5V (max) and is in a 5 pin SOT-23 sized package only. LEDs D1 and D2 (0603 size) set the power supply to around +/- 2.2 volts, the forward drop of a yellow LED. C2 and C3 are supply bypass caps, to be placed close to the opamp. Integrator C1 is a 1uf/6.3 Volt ceramic cap, X7R, 0603 size. A film cap would be fine but larger sized. R2 and R1 are 0603 thin film resistors. R1 limits the current/voltage applied to the feedback point. R3 and R4 (1/2 watt to be safe) are sized for a +/- 36 volt supply. LED current is around 7 milliamps, enough to be fairly bright. Opamp supply current is less than 100 ua. You may need to change R1, R3 and R4 to suit your circuit.
Most amplifiers, discrete or chip, need very little current or voltage applied to hold the output voltage at precisely zero volts. I typically see at most .3 volts at the output of the opamp in simulation. This is why, I think, you can get away using an opamp with this very low voltage supply. Why people run their servo opamp at +/- 12 V, I’ll never understand. Typical no signal output offset is less than 1 millivolt with this servo. This servo should also work well in preamp designs too.
Here is the AD Spice model that works with LTspice. Use the opamp2 symbol on your schematic and rename it AD8603.
* AD8603 SPICE Macro-model
* Typical Values
* Soufiane Bendaoud, ADI Silicon Valley
* Rev. 0
* Copyright 2003 by Analog Devices
*
* Refer to "README.DOC" file for License Statement. Use of this
* model indicates your acceptance of the terms and provisions in
* the License Statement.
*
* Node Assignments
* noninverting input
* | inverting input
* | | positive supply
* | | | negative supply
* | | | | output
* | | | | |
* | | | | |
.SUBCKT AD8603 1 2 99 50 45
*
* INPUT STAGE
*
M1 14 7 8 8 PIX L=1E-6 W=42E-6
M2 16 2 8 8 PIX L=1E-6 W=42E-6
M3 17 7 10 10 NIX L=1E-6 W=42E-6
M4 18 2 10 10 NIX L=1E-6 W=42E-6
RC5 14 50 1E+5
RC6 16 50 1E+5
RC7 99 17 1E+5
RC8 99 18 1E+5
C1 14 16 0.8E-12
C2 17 18 0.8E-12
I1 99 8 4E-6
I2 10 50 4E-6
V1 99 9 0.3
V2 13 50 0.3
D1 8 9 DX
D2 13 10 DX
EOS 7 1 POLY(3) (22,98) (73,98) (81,98) 40E-6 1 1 1
IOS 1 2 0.05E-12
*
* CMRR 90dB, ZERO AT 15kHz, POLE AT 2MHz
*
ECM1 21 98 POLY(2) (1,98) (2,98) 0 0.5 0.5
CCM1 21 22 3.54E-10
RCM1 21 22 30000
RCM2 22 98 1
*
* PSRR=100dB, ZERO AT 300Hz
*
EPSY 98 72 POLY(1) (99,50) 0 1
CPS3 72 73 5.30E-9
RPS3 72 73 100E+3
RPS4 73 98 1
*
*
* VOLTAGE NOISE REFERENCE OF 20nV/rt(Hz)
*
VN1 80 98 0
RN1 80 98 16.45E-3
HN 81 98 VN1 20
RN2 81 98 1
*
* INTERNAL VOLTAGE REFERENCE
*
EREF 98 0 POLY(2) (99,0) (50,0) 0 .5 .5
GSY 99 50 (99,50) 1.9E-6
EVP 97 98 POLY(1) (99,50) -0.6 0.5
EVN 51 98 POLY(1) (50,99) 0.6 0.5
*
* GAIN STAGE
*
G1 98 30 POLY(2) (14,16) (17,18) 0 1.25E-5 1.25E-5
R1 30 98 4.69e7
CF 45 30 50E-12
D3 30 97 DX
D4 51 30 DX
*
* OUTPUT STAGE
*
M5 45 46 99 99 POX L=1E-6 W=1.61E-3
M6 45 47 50 50 NOX L=1E-6 W=2.15E-3
EG1 99 46 POLY(1) (98,30) 0.3778 1
EG2 47 50 POLY(1) (30,98) 0.3771 1
*
* MODELS
*
.MODEL POX PMOS (LEVEL=2,KP=10E-6,VTO=-0.328,LAMBDA=0.01,RD=0)
.MODEL NOX NMOS (LEVEL=2,KP=10E-6,VTO=+0.328,LAMBDA=0.01,RD=0)
.MODEL PIX PMOS (LEVEL=2,KP=10E-6,VTO=-0.328,LAMBDA=0.01,Kf=10E-35,AF=1,TOX=100E-3)
.MODEL NIX NMOS (LEVEL=2,KP=10E-6,VTO=+0.328,LAMBDA=0.01,KF=10E-35,AF=1,TOX=100E-3)
.MODEL DX D(IS=1E-14,RS=5)
.ENDS
Some of the features are very small size and simple power requirements.
The opamp is an Analog Devices CMOS AD8603. Very low current requirement, good precision and reasonably quiet. Note that it is not chopper stabilized. It needs a power supply of +/- 2.5V (max) and is in a 5 pin SOT-23 sized package only. LEDs D1 and D2 (0603 size) set the power supply to around +/- 2.2 volts, the forward drop of a yellow LED. C2 and C3 are supply bypass caps, to be placed close to the opamp. Integrator C1 is a 1uf/6.3 Volt ceramic cap, X7R, 0603 size. A film cap would be fine but larger sized. R2 and R1 are 0603 thin film resistors. R1 limits the current/voltage applied to the feedback point. R3 and R4 (1/2 watt to be safe) are sized for a +/- 36 volt supply. LED current is around 7 milliamps, enough to be fairly bright. Opamp supply current is less than 100 ua. You may need to change R1, R3 and R4 to suit your circuit.
Most amplifiers, discrete or chip, need very little current or voltage applied to hold the output voltage at precisely zero volts. I typically see at most .3 volts at the output of the opamp in simulation. This is why, I think, you can get away using an opamp with this very low voltage supply. Why people run their servo opamp at +/- 12 V, I’ll never understand. Typical no signal output offset is less than 1 millivolt with this servo. This servo should also work well in preamp designs too.
Here is the AD Spice model that works with LTspice. Use the opamp2 symbol on your schematic and rename it AD8603.
* AD8603 SPICE Macro-model
* Typical Values
* Soufiane Bendaoud, ADI Silicon Valley
* Rev. 0
* Copyright 2003 by Analog Devices
*
* Refer to "README.DOC" file for License Statement. Use of this
* model indicates your acceptance of the terms and provisions in
* the License Statement.
*
* Node Assignments
* noninverting input
* | inverting input
* | | positive supply
* | | | negative supply
* | | | | output
* | | | | |
* | | | | |
.SUBCKT AD8603 1 2 99 50 45
*
* INPUT STAGE
*
M1 14 7 8 8 PIX L=1E-6 W=42E-6
M2 16 2 8 8 PIX L=1E-6 W=42E-6
M3 17 7 10 10 NIX L=1E-6 W=42E-6
M4 18 2 10 10 NIX L=1E-6 W=42E-6
RC5 14 50 1E+5
RC6 16 50 1E+5
RC7 99 17 1E+5
RC8 99 18 1E+5
C1 14 16 0.8E-12
C2 17 18 0.8E-12
I1 99 8 4E-6
I2 10 50 4E-6
V1 99 9 0.3
V2 13 50 0.3
D1 8 9 DX
D2 13 10 DX
EOS 7 1 POLY(3) (22,98) (73,98) (81,98) 40E-6 1 1 1
IOS 1 2 0.05E-12
*
* CMRR 90dB, ZERO AT 15kHz, POLE AT 2MHz
*
ECM1 21 98 POLY(2) (1,98) (2,98) 0 0.5 0.5
CCM1 21 22 3.54E-10
RCM1 21 22 30000
RCM2 22 98 1
*
* PSRR=100dB, ZERO AT 300Hz
*
EPSY 98 72 POLY(1) (99,50) 0 1
CPS3 72 73 5.30E-9
RPS3 72 73 100E+3
RPS4 73 98 1
*
*
* VOLTAGE NOISE REFERENCE OF 20nV/rt(Hz)
*
VN1 80 98 0
RN1 80 98 16.45E-3
HN 81 98 VN1 20
RN2 81 98 1
*
* INTERNAL VOLTAGE REFERENCE
*
EREF 98 0 POLY(2) (99,0) (50,0) 0 .5 .5
GSY 99 50 (99,50) 1.9E-6
EVP 97 98 POLY(1) (99,50) -0.6 0.5
EVN 51 98 POLY(1) (50,99) 0.6 0.5
*
* GAIN STAGE
*
G1 98 30 POLY(2) (14,16) (17,18) 0 1.25E-5 1.25E-5
R1 30 98 4.69e7
CF 45 30 50E-12
D3 30 97 DX
D4 51 30 DX
*
* OUTPUT STAGE
*
M5 45 46 99 99 POX L=1E-6 W=1.61E-3
M6 45 47 50 50 NOX L=1E-6 W=2.15E-3
EG1 99 46 POLY(1) (98,30) 0.3778 1
EG2 47 50 POLY(1) (30,98) 0.3771 1
*
* MODELS
*
.MODEL POX PMOS (LEVEL=2,KP=10E-6,VTO=-0.328,LAMBDA=0.01,RD=0)
.MODEL NOX NMOS (LEVEL=2,KP=10E-6,VTO=+0.328,LAMBDA=0.01,RD=0)
.MODEL PIX PMOS (LEVEL=2,KP=10E-6,VTO=-0.328,LAMBDA=0.01,Kf=10E-35,AF=1,TOX=100E-3)
.MODEL NIX NMOS (LEVEL=2,KP=10E-6,VTO=+0.328,LAMBDA=0.01,KF=10E-35,AF=1,TOX=100E-3)
.MODEL DX D(IS=1E-14,RS=5)
.ENDS
the +-2.5V supplies severely limits the DC correction (error) that can be applied.
Better to use +-18Vdc, or even +-22Vdc (411a).
There is no filter on the output of the servo. This is typical of the way a -ve servo is implemented.
I think you really need to filter the servo output. Otherwise all the crud that comes out of the servo is injected straight into +IN and cannot be removed once added.
I much prefer a +ve servo with filtering of the output from the amp and then post filtering to remove opamp artifacts.
Have you seen Gootee's servo postings? Well worth reading !!!!!!!!
Better to use +-18Vdc, or even +-22Vdc (411a).
There is no filter on the output of the servo. This is typical of the way a -ve servo is implemented.
I think you really need to filter the servo output. Otherwise all the crud that comes out of the servo is injected straight into +IN and cannot be removed once added.
I much prefer a +ve servo with filtering of the output from the amp and then post filtering to remove opamp artifacts.
Have you seen Gootee's servo postings? Well worth reading !!!!!!!!
The one transistor servo from the ASKA100 amplifier was momentarily discussed here: http://www.diyaudio.com/forums/solid-state/169590-mongrel-supersym-ii-21.html#post2406997
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.