Help with Signal To Noise in this op amp pan control/amplifer schematic

[Se7en_Costanza]

Hey guys,
I'm working on a 2 input pan control/crossfade and sum to 1 output, but my main problem is that im getting a huge amount of white noise on the output, was wondering how and why im getting that.

The schematic ive provided is the first portion of a larger schematic im working (so thats why it ends on an inverting op amp, there is meant to be more haha) but this is the portion of the circuit producing the noise problem.
Is it due to the 100k blend control? ive gotten the potentiometer idea from RG's pan control pdf, but changed it intentionally to a larger pot with smaller resistor values so that the first 50% of the pot wont reduce the volume but the other half will. Also, the wiper on his pan control is set to 4.5v, but mine is to ground (with caps wither side, so there isn't crackle when you turn the pot). Does this make a difference in noise also?


Schematic:


RG's Pan control (for reference)
http://www.geofex.com/Article_Folders/panner.pdf

[Se7en_Costanza]

Edit: 56k in the last op amp stage is meant to be 33k, (it just makes the output sort of unity with the input). Still the noise problem is there though.

[ElectricDruid]

The only things I can spot are:

1) No bias resistor on the lower input op-amp

2) Unnecessary decoupling caps around the panner.

For point (2), you've got decoupling caps both *before* the panner network and *after* too. Assuming you need them, either one or the other would be enough. But actually, in this case, the signals from your input op-amps are already correctly biased around 4.5V, so you don't need them at all. I'd get rid of all three 1uF caps.

If this doesn't fix it, I think you've just got an error somewhere. The schematic is basically fine.

[PRR]

> one or the other would be enough.

If you float the pan-net at 4.5V, then the pot wiper *must* go to ground through a cap.

That can work but leaks bass unless the cap is very large (or bass-leakage is accepted). IMHO several modest signal caps and re-biasing is preferable to a monster cap. But it could go either/any way.

Agree I see no reason to hiss.

[ElectricDruid]

> one or the other would be enough.

If you float the pan-net at 4.5V, then the pot wiper *must* go to ground through a cap.

Ah yes, PRR's quite right. I'd missed that.