Help with an envelope follower used as a gate to mute a channel

[BluffChill]

Hi everyone,

I'm working on this unusually specific project which involves plugging two guitars into the circuit on two separate channels, and when one is playing, the other is muted.
I came up with the idea of using a CMOS envelope follower that I have used before in other things, to be able to output a logic state whenever it detects signal on one channel. I use this to supply current to the base of a transistor, the collector of which is connected to the other channel. When current is supplied, the signal goes straight to ground through the transistor instead of through the op amp to the output.

Surprinsingly, it works - I fine tuned the threshold to be very sensitive so even playing guitar 1 very gently is enough for the envelope generator to spit out a 1, activate the transistor on the other guitar, and mute the note. The problem is, it is slightly noisy, and as the note fades and the other channel opens again, the note distorts and sounds weird. It's not a clean transition.

Is there any way I can clean this up a little and stop this from happening? Would this have anything to do with the fact I'm doing this with a battery rather than a regulated supply?

Here's the schematic, slightly simplified, but it uses a pretty standard 9v single supply with a 2x100k/10uF voltage divider for VR.

[iainpunk]

if i assume correctly that the triangle GND is actual ground and the stripe GNDREF is 4,5v, you aren't ''dumping signal to ground'' you are ''fuccing with the bias''.
you are essentially mixing two signals with the opamp, one is guitar with 100k and the other is -4.5v with less than 100 Ohm,
which generates a gain of 1000, giving the opamp incentive to put out 4,5kV, but it can't so it just clips to the positive rail...

try this:


i really like that logo by the way

cheers, Iain

[Fancy Lime]

Hey, interesting project. Sounds like you get a bit of ripple on the decaying envelope. The buffering inverters U1A and B then amplify that to rail to rail ripple. Try turning U1A + U1B into a Schmitt Trigger by running a resistor from pin 2 to pin 5 of U1. Use a 1 M pot and see what resistance works ( if it does). If my guess-analysis of your problem happens to be right, this should fix it, I think.

GNDREF and GND are both half supply, right? If GND is negative supply, I wonder how this works at all.

Cheers,
Andy

[BluffChill]

Thanks both for your replies - yes, striped ground (on op amp pins 3 and 5) is half supply, triangle ground is circuit ground. Do I need to take transistor emitter to half supply as well?

I'll try both of your suggestions - Iain, what exactly is different about yours? Just the placement of the transistor?

[Fancy Lime]

I think all points that are now GND should actually go to half supply.

[BluffChill]

Yeah, I think you're right. I think at some point (there is more to the circuit than just this component) I got my grounds mixed up.

Thanks again!