Weird oscillation in tube screamer build

[eurekaiv]

I'm working on ST-9 (a "Super Tube Screamer) I was "commissioned" to build and I'm having this really strange issue where after playing a bit then stopping I'm getting a low level oscillation in the background noise.  It's sort of a faint "whu whu whu" sound.  This only happens when I flip a switch to disengage the cap that controls the gain frequency range.  I like this mod as it allows it to be more of a "distortion" device.  Anyway, it will also stop oscillating if I adjust any control on the pedal only to return after playing through it a bit.  It exhibits none of these symptoms with the cap "in the circuit."  See my schematic drawing below for what's going on. Could this be a layout issue?  I never noticed it on the breadboard (wish I hadn't pulled it off already) but then it took a solid week of fiddling with the actual build before I heard it.  It's really really faint but it's also noticeable enough that I can't in good conscience sell it as is.

[bean]

Rather than switching the cap out completely, use the additive property of caps in parallel to get what you want. The problem will probably go away. So, add a cap in your normal "bypass" mode, and use the switch to drop the other one in parallel, thus increasing the overall value.

[John Lyons]

That cap is blocking DC voltage so with no cap there you are getting what people call "motorboating".
I would think that putting a large cap there would do the same thing as having no cap there (without the oscillation of course)

I may be wrong but I'm pretty sure this is the case.

EDIT Bean beat me to it but yeah. Try a .22 and another .22 in parallel and see how that works.
These caps affect the bass roll off of the clipping stage... Bigger = more bass.
But with two .22 in parallel that's about as full range as you'll get.

john

[eurekaiv]

That cap is blocking DC voltage so with no cap there you are getting what people call "motorboating".
I would think that putting a large cap there would do the same thing as having no cap there (without the oscillation of course)

I may be wrong but I'm pretty sure this is the case.

I actually just tried the larger cap value a bit ago and holy schnikees this thing went nuts.  It was "motorboating" so much it sounded like a phaser.  I'll try the parallel thing right now though and see how that does.

Thanks!

[John Lyons]

Hmmm
With a larger cap it should not be oscillating.
The resistor isn't that small so there isn't a ton or gain.
Paralleling the two caps to add up to .44uf may be just the thing.
The cap sets the frequency of that stage so I think that you are just
oscillating at a low frequency which would make sense given the motorboating.

With the .22 cap you are already going very low and passing most everything that needs to be
amplified as far as low end. 

[eurekaiv]

I played around with putting different values in parallel and I get motorboating at any value past .22 and it gets increasingly worse as I go up.  Like I said, two 1uf's go absolutely nuts.  There's defenitely something else going on with this thing and I think I figured it out. It appears the schem I was given to use is wrong.  In the ST-9 this portion of the circuit should be grounded and not given bias voltage as in a standard TS.   

[bean]

I'd say try a 47n in line and switch in a 100 or 180n. So, that gives you midboost vs. low boost.

Or, use a 10kB pot in series with the resistor (try dropping from 4k7 to 2k2). On the middle lug attach a 1uF to Vb. On lug 1 (right side) use a 33 or 39nf to Vb. That gives you a bass roll off, left to right.

You can connect this to ground if you are still motorboating. I doubt you will hear a difference, tone-wise.