Excessive noise in big muff build silenced by touching diodes?Did I miss a grnd?

[Bandwagonesque]

Hiya folks. apologies if this is elementary but a bit lost. i have a big muff build on my  breadboard that I tweaked to taste and while it sounds amazing, I can't for the life of me dial out excessive noise that comes right off the bat once i bring the volume pot to just 9 o'clock (but I am diming the sustain, and I did goose values here and there to get things right at the threshold of oscillation, but its still a bit much noise I suppose).

While bashing my head about and starting to schem out my breadboard to see if i missed anything, I noticed that the noise is stripped back a great deal into tolerance without compromising anything else  (though i couldn't tell if this affected anything further down the circuit in regards to controls, wish I had a third hand right now) if I put my fingers across the first pair of diodes while having my other hand across the strings of the guitar.

Now i've checked my grounds but does this mean i missed a ground somewhere? can anyone tell me what this must mean and if I just need to pay closer attention til I find my solution or is this some quirk that has nothing to do with my build? One other thing I noticed is that when I have it in tone bypass mode, the noise still persists and from my recollection, bypassing the tone stack while raising the volume quite a bit, would also lower the noise floor a great deal so i'm just a bit confused as of now as to what is the matter. everything else seems right and controls are good and it passes signal but the noise is just bad til I do that little hand trick.  Any help would be greatly appreciated as i'm so close to finally finishing this up once the noise is sussed out.

[anotherjim]

Sometimes having it on a breadboard leaves it more open to picking up noise.
But, some time ago, member PinkJimPhoton posted a tip. Try adding a germanium diode across one of the clipping diodes at Q2. If you like the effect, one of the silicon diodes can be replaced by the germanium one. The direction of the germanium diode doesn't matter so long as it's opposite to the plain silicon one.

I'll add that tweaking it to the edge of feedback may mean it's too much once in a box when things are closer together and you get squealing - but that can be dealt with.

[Vivek]

I think you have put your finger on the problem !

[Vivek]

Why does that germanium trick work ?

[GibsonGM]

Why does that germanium trick work ?

Would it be because the Ge diode has a lower Vf, so it will turn on first, and the Si one it is in parallel with will not?

[R.G.]

This is a classical description of self-oscillation at radio frequencies. You can't hear the RF, but the circuit picks it up and de-modulates the jitter in the RF as noise. A finger's capacitance lowers the amount of feedback at radio frequencies and it stops. Might be something else, OK, but this is a candidate.
A 50pf cap in various places on the circuit would stop it
[got interrupted while typing...]
if this is the problem.  Try putting one across the diodes, or across one or more collector resistors.

[Vivek]

Why does that germanium trick work ?

Would it be because the Ge diode has a lower Vf, so it will turn on first, and the Si one it is in parallel with will not?

It's understood that Ge will conduct before Si

But is that the reason why noise is curtailed ?

I am wont to believe RGs explaination re parasitic RF oscillation, and the 2 diodes acting as caps for small signals, somehow stopping the oscillations

Proof would be if a small cap in parallel to the Si diodes acted in same way as a Ge in parallel to the Si diodes

Or even 3 Si diodes, 2 pointing one way and one the other.

[Vivek]

I did goose values here and there to get things right at the threshold of oscillation.

If there is no real feedback path between stages in the circuit

and it still oscillates

That could mean that there is unwanted parasitic feedback

which would be different in the breadboard than in the PCB

[GibsonGM]

Why does that germanium trick work ?

Would it be because the Ge diode has a lower Vf, so it will turn on first, and the Si one it is in parallel with will not?

It's understood that Ge will conduct before Si

But is that the reason why noise is curtailed ?

I am wont to believe RGs explaination re parasitic RF oscillation, and the 2 diodes acting as caps for small signals, somehow stopping the oscillations

Proof would be if a small cap in parallel to the Si diodes acted in same way as a Ge in parallel to the Si diodes

Or even 3 Si diodes, 2 pointing one way and one the other.

Yes, what RG said absolutely has much merit....I was merely stating why the mod that Jim discussed works, in response to your question, which you clearly understand, due to Vf.  I didn't take it to be a remedy for the noise, I believe he was just relating a mod to the circuit to make it sound better (for some).   Maybe that was taken the wrong way...   

[antonis]

+1 to what R.G. said..

A small value cap across diode pair should act a a LPF inside NFB loop..
(a calculation of its value in relation with diode(s) forward resistance(*) for any given RF is possible but trial-and-error is more convenient..)

(*) Here comes Vf divided by current..

[anotherjim]

I'm not so sure. In the old threads, I think caps were tried and a few other things - and does a Ge diode have that much capacitance?
Oh, and it worked for my own build BMP and it didn't even have the proper low-noise transistors.

[PRR]

> if I put my fingers across the first pair of diodes while having my other hand across the strings of the guitar.

Then how do you play?

i.e. how do you know you aren't just killing all the gain? Finger on diode connects to finger on strings which connects to ground?

I'm not arguing with the other wise suggestions posted here. In fact oscillations stop if gain is reduced enough.

[Rob Strand]

The best plan of action is to follow the RF oscillations theory.

With so many stages you will need to experiment.
- The most obvious is from the output of stage 4 (the actual output) back the input of stage 1 (the actual output).
A sledge-hammer approach is to put say 470pF/500pF across the circuit's input terminal.
If that works try backing off the cap.
- The next on the list could be the output of stage 3 back to the input of stage 2.
   There's wiring and pots at these point and that promotes feedback/oscillations.
   In this case try the cap at the input of stage 2 (where the pot wiper enters the pcb).

[Bandwagonesque]

thanks a bunchhh for the feedback folks. definitely food for thought as I tackle this issue in the morning.

Soooo its interesting as this morning I decided to swap out an old PCB based big muff builds values with the same hot values I have on my noisey breadboard build and the hiss and noise remains now through both builds. was hoping it was an issue with the breadboard being the source of the noise and a bed for parasitic noise but alas its consistent with both. both builds even have the same weird quirk when the tone section is bypassed in that it doesn't quite up. hate to ask again but my experience with muffs is mostly with the IC circuit which to my recollection realllllly quieted up once the tone stack was bypassed. Am I wrong in my memory or is there something wrong with the noise still carrying through with the tone stack bypassed? thanks again everyone for the help and will definitely report back in a day or two if the diode and capacitance fixes prove to be fruitful.

[R.G.]

Just to be clear, there's no guarantee it's RF oscillation. The description just made my "RF oscillation!!" detector go off. 
As Paul notes, if you reduce gain enough, oscillation quits. Put another way, if you increase gain enough, every circuit oscillates. At high enough gains, the air path around the thing is enough capacitance and resistance and phase shift at some frequency to make it sing; so an explicit feedback path isn't needed. Also, just a common ground path and common power supply make for feedback at some high-enough frequency and gain.
Which reminds me - does your circuit have good power supply bypassing?