Scratchy sound when turning potentiometer

[tk@halmi]

This is really wierd. I have used capacitor coupling (470u/35V) on my preamp and the feedback terminal is capacitor coupled (4700u/6V) to ground as well, a very standard circuit IMHO. As I understand this should eliminate problems with DC shift. However, when using discrete opamps there is a scratchy noise when changing gain in a certain region of the pot. This does not happen when using chip opamps in the same circuit. I am wondering what is going on. What measurements would help me troubleshoot this?

Thanks,
Tamas

 

[gyraf]

Capacitor leakage currents are never zero. Your discrete opamp probably has much higher offset voltage than the integrated ones, thus pushing more DC current into your pot...

Jakob E.

 

[tk@halmi]

Thanks Jakob!

Is there a remedy for this? Three out of the five DOAs I have make this noisy affair.
Perhaps a little adjustment circuit as shown on this app note?
http://www.jensen-transformers.com/as/as018.pdf

Tamas

 

[tk@halmi]

Let me continue my monolog. The John Hardy 990 document has a very nifty little circuit for adjusting DC offset also.

 

[gyraf]

A DC Servo perhaps?

 

[Samuel Groner]

I would start with the "Hardy" compensation...

Samuel

 

[tk@halmi]

Jakob,

The DC servo will take me back where I started. Oddly, the problem was existent with the servo in place at the beginning. It did help a little, but it may not have been fast enough to compensate for quick gain changes.

Perhaps I should try it again and reduce the RC network on the servo a little. I was using 1MOhm/470nF on the first stage and 2MOhm/470nF on the second stage.
Through experiments it has become appearant that the DC offset in the first stage always dominates.

Thanks,
Tamas

 

[Winston OBoogie]

...

 

[Samuel Groner]

ddly, the problem was existent with the servo in place at the beginning.

A DC servo, as you have tried it, will not help you. It will force the output of your amp to zero (OK, close to zero) DC, but it will not cancel any offset present (due to input-currents) on the pot/minus-input. A servo applied as in the Jensen note or the Hardy circuit will reduce the offset on the pot.

I don't have your circuit at hand, but if it's somewhat based on a discrete opamp with variable feedback, this should apply.

Samuel

 

[tk@halmi]

[quote author="Winston O'Boogie"]
6/ Rip the bugger out and smash it with the heel of your shoe.
[/quote]
LOL. I have come close that. Good to see you around again!

[quote author="Samuel"]
It will force the output of your amp to zero (OK, close to zero) DC, but it will not cancel any offset present (due to input-currents) on the pot/minus-input.
[/quote]

I will try the Hardy circuit on the next project designed onto the PCB.
For now, I am going with chip opamps (OPA627) because they sound darn good and have no offset problems. The second stage opamp is buffered by an LT1010 for current drive into the output transformer. (Tamas' infamous composite opamp)

Thanks Everyone, this has been a very educational project for me.

Cheers,
Tamas

 

[Bauman]

Hey Winston, nice to see you man !!