Electronics > Beginners
Op amp stability problem
(1/1)
Greg Robinson:
Hi everyone.
I've been working on a project, but having some oscillation issues with this part of the circuit. The pot allows adjustment of the amount of crossover distortion in the (audio) signal. This forms a rudimentary noise gate, but more importantly a desired effect for signal processing. This is not intended for accurate reproduction, only the effect, for guitar. Debate among yourselves whether you feel this is a desirable effect, but lets just accept that it's what I want.

My problem is that while the circuit is stable when the pot is adjusted from low to about 50% of its resistance, it breaks out into oscillation at ~135kHz as the wiper approaches the output of the first op amp. This then lets the signal through relatively cleanly (as though the diodes were in the feedback loop, which they no longer completely are), modulated by the oscillation frequency of course.

I have tried introducing some series resistance from the pot wiper to the non-inverting input of the second op amp, removing the zener diodes (which are intended to clamp the output swing of the first op amp to provide more graceful clipping in the case of overload and to avoid signal inversion that may happen when hitting the rails), and altering the circuit layout, without any success.

The op amp has 100n MLCC local bypassing, and no other part of the circuit suffers from oscillation.
The loop area of the negative feedback loop is larger than ideal due to user interface/mechanical constraints, and as yet I've only prototyped it on the breadboard and matrix/vero, I have not tried a proper PCB with groundplane. Changes to lead-dress or layout (I've built it twice on matrix board) have not altered the oscillation frequency though, so I've largely discounted this as the cause.

I have also not yet tried making the input op amp inverting, as I require a high input impedance and do not have a spare op amp and would like to avoid an extra buffer.

So, does anyone have any suggestions? Any help would be appreciated.
Kleinstein:
The 2nd OP as a buffer adds some phase shift to the feedback loop, when at the GBW limit. The TL072 is known to not have very much phase margin and can thus start oscillating. However just 135 kHz is a relatively low frequency - this gives the question if the OP is a real TL072 or a Chinese fake (e.g. LM358 like).

One could try add in a resistor (some 220 Ohms) at the output of the 2 nd OP and a small capacitor (some 500pF - 1 nF)  as direct feedback to OP1. So the high frequency FB would be directly and the special function would be mainly for the lower frequencies.

Usually there is little need for OP2, so one could also just skip the buffer and directly use the wiper of the pot - maybe with some 1 K in series for protection.

Too much capacitive load (this may be just 1-2 m of coax cable) at the output could also be a problem. So maybe add another 100-220 Ohms to the output for isolation.
Greg Robinson:
Thanks Kleinstein,

The op amps are presumed genuine, I only source silicon from authorised resellers - element14, RS Components, Digikey, Mouser, though they have all been known to get the occasional grey-market parts. I've switched out the op amps for a different batch though with no change.

I'll try your suggestion of a build-out resistor on the second op amp, though I did forget to mention I have already tried high-frequency bypassing around the first op amp. I'll try a combination though, thanks.

I'm sure there was a reason I had included the buffer in the loop, but now come to think of it, I can't recall why. I'll try bypassing that too.

There's no significant capacitive loading on the output, it's DC coupled to another op amp in the same package.

Thanks again for your suggestions, you've given me some things to think about. I think I've just been scratching my head about this one for so long I can't see the forest for the trees.
Greg Robinson:
Removing the buffer from the loop seems to have solved the issue.

I now recall why I had added the buffer to the loop, when I had previously used the input op amp in inverting configuration, to maintain the same gain level throughout the range of the pot. I needed non-inverting this time and left it in without giving it enough thought.

Thanks once more.
Navigation
Message Index
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod