Unwanted signal in an op-amp circuit (inductive pickup?)

[Omegatron]

I've got a simple audio circuit that I wouldn't expect to have any signal coming out of it (the resistor at the input is a pot and is being grounded out).

But the previous stage's signal is still coming through:

http://img399.imageshack.us/img399/5804/opampinterferenceaz8.png

* It's not the contact resistance of the pot; I've soldered the terminals directly together and it's still there.
* 1/2 Vcc pin is perfectly flat relative to ground.
* Increasing the electrolytic cap didn't seem to have any effect
* There are two of these, and the crosstalk on one is higher amplitude than the other, though all the values are the same
* The crosstalk increases as frequency increases

So I'm suspecting some kind of inductive pickup in the pot (20k resistance)? Does that sound likely?

 

[NoTime]

Do you have long wires going to the pot?
Could be inductive or capacitive coupling.
Use sheilded cable.

Or feed back through to power supply.
Did you check the power pins relative to ground?

 

[Omegatron]

Do you have long wires going to the pot?

Just traces going an inch or so.

Could be inductive or capacitive coupling.
Use sheilded cable.

That wouldn't work here; they're on the same board pretty close to each other.

Or feed back through to power supply.
Did you check the power pins relative to ground?

I didn't actually check it, but I was assuming it was fine, especially because there are two channels on the same dual op-amp and they have different output levels.

 

[NoTime]

If the 1/2vcc is flat and you shorted the c1 - to ground then power supply seems probable with 1" traces.
I'd check it.

Is it an audio amp?

 

[berkeman]

Smells like a miswire. Double-check your pinouts and wiring. Also replace the IC, in case it's a bad chip (preferably from a different source -- I once wasted a day with multiple bad opamp chips from the same vendor/lot). Definitely check the supply as NoTime says, but the PSRR of most (non-bad) opamps would generally preclude that path.

 

[Darryl]

dollars to donuts, you've built yourself a very nice high frequency oscillator.

op amps have HUGE gain, they will oscillate at the drop of a hat.. you probably need to put a low pass filter in the feedback loop, and possibly reduce the gain of the circuit.

it could well be oscillating at quite a high frequency as well, we'll 1 or 2Mhz, we'll above audio range.
(thats my guess).

 

[Omegatron]

If the 1/2vcc is flat and you shorted the c1 - to ground then power supply seems probable with 1" traces.
I'd check it.

I will.

Is it an audio amp?

Yep.

Smells like a miswire. Double-check your pinouts and wiring.

Definitely a possibility.

Also replace the IC, in case it's a bad chip (preferably from a different source -- I once wasted a day with multiple bad opamp chips from the same vendor/lot). Definitely check the supply as NoTime says, but the PSRR of most (non-bad) opamps would generally preclude that path.

That's what I was thinking.

dollars to donuts, you've built yourself a very nice high frequency oscillator.

op amps have HUGE gain, they will oscillate at the drop of a hat.. you probably need to put a low pass filter in the feedback loop, and possibly reduce the gain of the circuit.

Well, there's the standard frequency compensation cap in the feedback loop, limiting it to a 3 dB point around 30 kHz probably.

it could well be oscillating at quite a high frequency as well, we'll 1 or 2Mhz, we'll above audio range.
(thats my guess).

The signal I'm seeing is in the audio range, though, and the same as the previous stage, so I think it's just crosstalk of some type. That's why I suspect it's coupling into the pot. There are a lot of things in the vicinity; I just lack knowledge of how to model interference sources like that.

If I were getting capacitive coupling into my pot, for instance, it would appear like a capacitor connected from some component in the previous stage to somewhere on the pot?

 

[Darryl]

if its the same signal as the previous stage, and you have truly shorted out the input singnal, at the pot, ie you've tied it to zeor volts.

but the signal is still at the output, and the same as the previous stage, what happens when you remove the input signal to the previous stage.

but that original signal is not coming from the input pot from the previous stage, and if its not self oscillation. its has to be getting into your circuit somehow. then i would guess, its because you run a common power supply, and you don't have enough supply decoupling in your circuit.

what happens if you
1, remove the input signal to the previous stage. ?
2. vary the amplitude or frequency of the signal into the previous stage.
3. does your signal follow your changes, ie is it really the input sigal or its own oscillations.

how are you deriving your 1/2 VCC supply, ?

good luck, I am sure you'll get it worked out, divide and conqure.