DC Servo question...

[Dxvideo]

Hi all,

I am still working on a op-amp preamplifier + headphone amplifier project for my LM4702 power amp.
I have some LM4562, OPA134 and NE5534 to use that project. I dont use any capacitors on signal path in LM4702 amplifier. And also dont want to use any capacitor in preamplifier.
So I decided to use DC servo in preamp. I saw Damian's LM4562 preamplifier for gainclones on somewhere. Its very simple and looks goodworking one. However as you can see it uses capacitor on signal path.
I have two alternatives;
- I can use that circuit without capacitor,
- I can add a DC servo to it.
But the op-amps used for DC servo are generally expensive and low availabililty ones. I want to have my preamp very quickly, and I can find some LF411 and LF412 s in Turkey. May I use them for that project as DC servo? Because I saw that op-amps in original BPA200 project used for DC servo for LM3886s..
What do you say? May I use LF411 as DC servo for LM4562 or OPA134 (also for LT1210 in headphone amp) ? And works fine?
Thx in advance..

[gootee]

I haven't tried using an LF411, for that. But it seems like the LF411 might be quite good, for a DC servo's opamp, since it is listed as having low offset (0.5mV max), low offset drift (10uV/degC max), and a fairly-high slew rate (10 uV/sec min). [Looking at the datasheet from http://www.national.com , it looks like the figures I quoted above are actually for the LF411A, and are slightly better than those for a plain LF411.] The LF412 looks similar, although I haven't studied either datasheet in detail.

I have a schematic and downloadable LTspice simulation of a pretty-good DC Servo circuit that I designed, at:

http://www.fullnet.com/~tomg/gooteesp.htm

That circuit should still work very well, with an LF411.

If in doubt, you could use an LF411 spice model and simulate it. And you could easily add ideal sources at the opamp inputs, to simulate the various offset voltages and currents, even including temperature effects.

- Tom Gootee

http://www.fullnet.com/~tomg/index.html

[AndrewT]

Hi,
of all the old types of opamps the LF411 is nearly ideal for DC servo duty. I think the two parameters, low offset and low offset drift, are the important ones.

There are new types that can do the job better but they are generally more expensive and often difficult to source.

Remember that the servo only NEEDS to handle frequencies below audibility. Make the servo's job easier by filtering the input signal to attenuate the audio frequencies.
Remember to also filter the output of the servo to attenuate opamp/psu noise and to attenuate the remains of the audio that passed the input filter. The servo output is usually connected directly to the inverting input and any extra signal added here will be multiplied and appear at the amplifier output.

Go and read Gootee's website for more information.

The same rules apply to a pre-amp servo as to a power amp servo. Just take care of noise injection appropriately.

[Dxvideo]

Hi,

I examined the Gootee's web site.. However I am confused.
I will use 3 x gain for preamp. So for a shortcut solution;
As Damian's application for LT1201..
I want to use Damian's circuit with OPA134 instead of LT1201 and LF411 (or 412) instead of OPA604..
Do you think this is suitable for that opamps?
I am not sure, because in Spice analyse I could run this circuit with ideal op-amp sub circuits, however with OPA134 and LF411 it doesnt look like working. I dont know the reason but thats the situation.

Another way is to use MKP type caps on signal path. Do you recommend that?

Thx

[SY]

That circuit will work fine with LF411, though you'll need to capacitively couple the input. There are one or two things you should do, though. First, eliminate R7/C5 and take the noninverting input of the 411 directly to ground. Second, split R6 into two 22-27k resistors in series, then take a 100nF cap to ground from their junction.

The latter mod adds extra passive filtering of the output as Andrew suggested.

The other suggestion is to return the servo to the bottom of R2.

[Onvinyl]

Hi Ozgur,
I think there are a few issues with your last circuit.

1) I wouldn't inject the servo output into the same node where your signal goes. Look at gootee schematic how he did it.

2) If you for some reason want to keep that scheme, you can omit C5/R7 and simply ground the non-inverting input.

3) Keep an eye that you use an FET-input opamp. The LF411 is FET-input, so will work good. A more modern choice would be AD820.

4) Output filtering. You need to know what your worst case offset is. The lower (without servo) the better. If your worst possible offset is say 50mV+/- you can figure what the maximum output voltage swing of the LF411 is. Lets assume it's 13 V+/-, you have an ratio of 260:1. So you can set the voltage divider R6 : (R2||RinLT1210). Allow a safety margin and you might get by with R6 as big as 200k, which enhances filtering of the noise that is coming from your servo.

@gootee: what advantages do you see in having an active low pass filter after the integrator?

Rüdiger

[darkfenriz]

No one seems to care about inputs phase reversal in DC-servo op-amps?

[Onvinyl]

Originally posted by darkfenriz
No one seems to care about inputs phase reversal in DC-servo op-amps?

care to explain?
Rüdiger

[Dxvideo]

Hi,

As mentioned,
I've examined the Gootee's site. However its a bit theorical... He starts from the modelling and goes with calculations.... I'm sure thats the way it should be.
However, until last month I even didnot know anything about the DC servos. Now I am still very newbie on this matter.
I dont want you to design a DC servo for me! But I want to try it, I am not sure that DC blocking cap causes any kind of bad affect on sound. But theyre very big, expensive and hard to find. So if there is a better way to solve the DC offset problem then I would make it.
Now,
I am talking what I understand:
- To connect DC servos output to the NI input directly is not a good idea.
- To use NI input of DC servo op-amp is also not a good idea.

But in Gootee's example, he uses two stages DC servo, while I coulnt understand one stage DC servo, how could I adapt it to my s h i t t y preamp project?????

More help for that poor diyer pls...

[darkfenriz]

Some opamps, including lf41x reverse the plarity of inputs when overdriven. I don't know, but I suppose it can bring catastrophic results if DC servo changes polarity... So I ask if anyone should care?