Virtual ground mixer

[funk1980]

I'm working on a parallel effects loop for a bass pre-amp. I know about tubes, but opamps are a bit new territory . The schematic below works perfect for my intends, except that the mixing of the two signals is not linear. I tested this by providing the same signal (1khz sine wave) to both inputs of the mixer. The two pots are actually a single stereo pot and is used to blend the output signal from wet to dry.

The differende in output signal when the pot is fully turned one way or at 50% is 2.15dB (3.6V vs 2.8V).I'd expect the output signal to be constant, which is desired because the difference is audible.

Can anyone provide me with some tips or insight why this effect occurs and perhaps how to resolve it?
Thanks

[David_AVD]

Is the "stereo pot" (dual gang) linear or log?

[funk1980]

Is the "stereo pot" (dual gang) linear or log?

Good question, I totally forgot to mention that. It's linear. The lowest output is achieved at a 50% turn.

I did some simulations. As it turns out, replacing the 10k pot with a 1k fixes the problem for the most part. My guess is when the pot is at lets say 50%, the parts of the pots is series with the 10k input resistor become part of the feedback equation, thus lowering the amplification of the whole.

Is my assumption correct?? If so, that means the pot needs to be a fair amount smaller than the input resistors. In that case, would it be better to decrease the pots value, or increase the feedback/input resistors.

[David_AVD]

Sounds like the source impedance of the return input is higher than the buffer.  Ideally you'd buffer the return so it's a known impedance (same as other buffer) for the mixer section.