DC servo a good or a bad idea

Status
Not open for further replies.
Jump to Latest
D

danny92

Member
Joined 2012
#1
Hi everyone,

I was reading about DC servos using JFET opamps to reduce the output offset voltage, without using electrolytic capacitors. The idea consists in using one opamp to integrate the output voltage of the amplifier with an inverting integrator, and then inverting the output of the integrator with another opamp, this correction signal is injected in the inverting input of the power amplifier thru a high value resistor.
I wanted to know your opinions about the use of this technic and what are the best circuit topologies and opamps to use.

Best regards,
Daniel Almeida
 
#2
You have to form your opinion of the feedback shunt electrolytic and it's possible associated THD.

I like DC servos personally. For op amps I'd go for JFET input low noise type.
I use the two op amp with low pass filters on input and output of the servo.
 
#4
1) I'm not that well versed with op amp models. It's more important to deal with the implementation than the op amp type. Like ensure that the psu decoupling is correct and supply lines are quiet. Remember this servo can send noise straight into the feedback node of the amp.

I'm using a plain old TL072 in my prototype. I don't see any logic in using extreme audiophile op amps for a DC servo.

Have no means of THD measurement at present, so, right now any DC servo op amp experimentation would just be guess work.

2) Input resistor forms a filter with the first op amps feedback capacitor. Going for a high value like 1Meg means that the feedback cap does not need to excessively large.

The output resistor is determined by the amount of offset that needs to corrected and the values of the feedback network. It should as large as it safely can be without the op amp driving to its rails.
 
#9
danny92 said:
Hi everyone,

I was reading about DC servos using JFET opamps to reduce the output offset voltage, without using electrolytic capacitors. The idea consists in using one opamp to integrate the output voltage of the amplifier with an inverting integrator, and then inverting the output of the integrator with another opamp, this correction signal is injected in the inverting input of the power amplifier thru a high value resistor.
I wanted to know your opinions about the use of this technic and what are the best circuit topologies and opamps to use.

Best regards,
Daniel Almeida
Click to expand...
transconductance amps work better than integrators. Integrators dither but transconductance amps settle in with no 'wobble'.

G²
 
#12
Servos want good linearity, so film caps tend to be preferred. In particular, a 1.0uF ECPU(A) and a 100k resistor is a good starting point---hard to beat the cap on price or size for small signal handling, a 1.6Hz cutoff is low enough for most purposes, and the impedance is low enough not to be too sensitive to op amp errors. Resistor selection's insensitive but Susumu's RR series is a good general purpose default.

TL07x or 8x are fine. The TL03x will have somewhat smaller error terms but it's not a big deal---do the maths and decide if you care.
 
#13
Tantalum caps can be good on regulator outputs, because of their high ESR, for damping.

Use a differential integrator, for the servo.

The final output should not just be injected into the amp input. Use a voltage divider on the output, first. You can also add a cap, there, for one more stage of low-pass filtering. Also use voltage dividers on the inputs to the integrator, so that the input is small-enough to not saturate the integrator. Also, that way, you can use a lower rail voltage for the op amps, and servo a higher-voltage power amplifier.

There is an old thread, here, where Klaus and I experimented with quite a few servo schematics...




If you want it to react fast for large DC offsets,
 
Status
Not open for further replies.
Top Bottom