Talkbox amp with safe bypass

[mdh]

I just built a little amplifier to drive my talkbox (also a recent project of mine), but I'm having trouble figuring out how to bypass it properly.  The amp is basically a Little Gem MkII from runoffgroove.com.  The basic idea is that the amp should have one input and two outputs:  output 1 sends the amplified signal to the talkbox, and output 2 sends an unamplified signal to another amplifier, possibly by way of more effects.  So it's an amp with its own A/B switch built in, I guess you could say.  However, the way I have the bypass wired right now is leading to squealing oscillations being sent to the talkbox when the amp is bypassed.  Here's how it looks right now:



Obviously this glosses over the internal details of the amp, but should illustrate the bypass problem.  I wired the second pole of the DPDT to ground the "effect" input on bypass, as suggested in the FAQ for high gain effects, but this does not prevent the squealing.   At this point the only other idea I have is that instead of using the second pole to ground the input, maybe I should just use it to cut the power to the amplifier.  Oh, and the amp is battery powered.

I'm also wondering if the common ground is a problem.  Do I really need to go with a complex A/B/Y box kind of circuit (e.g., those on geofex.com), or can I do this right with a single DPDT switch?  Or 3PDT, if absolutely necessary?  I really just need A/B, A/B/Y is not necessary, and I don't really care about an indicator LED at this point.  Any thoughts?

[Peter Snowberg]

Welcome to the forum.

I would change the switching a little so that when one was selected the other would be grounded, like this:



From what you just said it looks like the squeal is from something else though. It shouldn't be a ground problem. I'm guessing you have a problem with the power for the 386. Is the battery fresh? The buffer may be getting some feedback from the 386s through the power supply.

If you don't have one in there, I would first attach a good sized cap (470uF or larger) between +9 and ground to make sure the power supply has a low impedance path to ground for any AC. Next, I would break the wire between the JFET drain and +9V. Put a resistor in there of something like 470 ohms to 1K, and then attach a cap of, let's say, 47uF or larger between the JFET drain and ground. that will provide a filter for the power which should stop any oscillations from internal feedback.

[mdh]

Thanks for the suggestions.  I will play around with some of those when I get the chance.  Looks like you're right about the power, though.  I must admit that I had been running it off of 12V (pack of 8 AA's) without any modification to the circuit.  On 9V, I don't get audible oscillations.  FWIW, there's a 100uF cap between Vcc and ground, but I can easily go bigger.  I also recall reading that a cap from pin 7 to ground can help prevent the LM386 from oscillating; Ruby has a 100nF cap there, LG MkII doesn't.