Strange Super Fuzz issue

[Heemis]

Hey, I know I'm not likely to get too many answers on a Friday evening, but I'm going to throw this out there to see if I can get this fixed tonight.

I built up a superfuzz on the breadboard and it sounds fantastic.  I love everything about it, except this one strange issue.

For some reason, after I play a note/chord/anything and I stop the note quickly, I hear a strange swell in background noise, followed by a dropout, and finally the bg noise levels off.

I know it's hard to imagine that as I explained it so I've recorded a quick clip of the issue in action here:  http://www.box.net/shared/mjjqt92dk5

If anyone has any idea why this is happening I'd be glad to hear it!

In case it matters, this clip was recorded with 2n2222s and 2n4401s for the transistors, and germanium diodes.

[amptramp]

Is this the circuit you are using?

http://www.geofex.com/FX_images/univox1.gif

If so, carrying the drive too high may cause the frequency doubling transistors (the ones with the collectors tied together) to pull excess current and bias the stage off until the emitter voltage drops back to normal.  This sounds similar to "gulp distortion" where an amplifier is driven into cutoff by excess signal and the signal drops out until the normal operating point is restored.

Just a guess, of course.

[Heemis]

For reference, this is the schematic I was working off of:  http://www.univox.org/pics/schematics/superfuzz.gif  They are very similar.

If it's helpful, I can record more clips.  Has anyone else had an issue like this with their superfuzz build?

[Heemis]

I did some experimenting with turning down the gain (expander) pot, and it did seem to make the effect less apparent, but I really like the sound of it full tilt.  Anyone else have any idea what I can try?

[Mark Hammer]

That schematic is actually incorrect, believe it or not.  The 10k and 22k resistors in the filter should be switched around, such that the 10k is first and 22k is second.

[Solidhex]

This is pretty common behavior for the Superfuzz. Don't really think there's much you can do to get rid of it. Usually when the octave is perfectly balanced the general hiss level will lessen but the swell of noise will still be there.

--Brad

[Heemis]

Thanks for that info on the schem Mark, I swapped it up and the notch setting definitely sounds better.

As far as the "gulping" issue, I guess that is just part of having such an insane sound.  I find when I dial the expander pot down to about 8-9 I still get a huge sound and not nearly as much weirdness.

A side note is that I breadboarded this circuit and tried 3 separate sets of transistors.  For each set I used two transistors with identical gain for Qs 4 and 5.  I actually get a perfect octave sound just grounding the 22k resistors without the 10k pot on each set of transistors. 

First up were 2N3904s which I measured all to be in the vicinity of 160-180 Hfe.

Next were 2N2222s and a pair of 2N4401s which all measured 240-260 Hfe.

Last were 2N5088s, which all measured right around 350 Hfe.

To the credit of everyone who said the transistors were not critical in this circuit, you seem to be right on.  I barely noticed a difference between the 2N3904s and the 2N5088s.  The only true difference is the fact that the 2N3904s were very noisy, but the sound of the circuit itself was almost identical between different transistor gain ranges.

If anyone's interested, I could post soundclips of all 3 ranges in action before I actually build it up.  Thanks again to everyone for the help.