Ge clipping diodes after boost removes white noise?

[darron]

i have pnp, neg ground, germanium boost on my breadboard now and i am trying to get the noise levels down without much luck. i remember a friend told me that the circuit can be a little bit noisy but running into a fuzz it it's not really noisy. for fun i put a couple of clipping ge diodes at the end and the noise level cut in half!


so the first thing that you may think is that of course the diodes are clipping the level down, but i mean the noise level had dropped while an instrument is plugged in and not even making any noise, therefor not anywhere near the 0.2v to clip. with a loud strum at full volume you can hear a bit of clipping of course with maybe 20% db drop.

what is the effect that is happening here? is there any way that i can emulate it so that the pedal can stay clean? i've tried resistors to ground at the end to load it down a little, but no help. i don't want to bleed off too much high to quieten it down.


i know than ge's are particularly good with finding AM frequencies...? maybe i need some sort of miller effect cap?

[Earthscum]

Diode capacitance? I have a pair of black Ge's from NTE that have no fuzz to them in either a Bazz Fuss or BMP... they just compress the sound. It is kinda neat and just sucks at the same time, lol. I put them in the 2nd clipping stage of a BMP and just got solid compression with a felt lining, to put it visually. That was after Si's in the first clip stage.

[darron]

I played with some capacitors where the diodes are. They don't remove the noise until they dull out the tone, whereas the diodes don't effect the tone until they clip.

Ge's can be pretty discrete in a fuzz. For example if I leave the diodes in and push a driven amp you can't really hear them, but it would be nice to be able to keep it clean too.

I wonder if some larger zeners would still do it... Some sort of ground reference/feedback thing going on???

[Earthscum]

Were they after a cap (decoupled)?

Idea... try using Si and Ge series to increase the headroom and see if it still gets rid of the noise? 2 Ge's if not? If it does get rid of the noise, maybe try stacking another one, and then try a polarity test, maybe one is doing the work and the other is just clipping?

[darron]

Earthscum,

They were after the final output cap, so decoupled. they were even decoupled to ground themselves incase the next effect had DC on its input which would burn out the diode, but it worked without its own cap too.

the diodes in series idea is probably good, which is why i was also thinking about a zener with a higher 'turn on' voltage. that's probably what i will do if nothing else works, but i'm curious to know how this magically removes so much noise?

i also tried switching just one diode on and off for both polarities. in both cases, switching one removes half of the noise fix and it is equal for both sides of the wave.

[Earthscum]

I just tried making a high gain, extra hissy single tranny (2n2222) circuit and dropping in the Ge's... it didn't do anything. This is strange... but, then again, who knows. I tried decoupled and both ways coupled (forward biased and reverse biased). I may pop in a Ge tranny in the morning and see if I can replicate it with that... might be the combination Ge Ge? Or, maybe it's based on PNP... I did use a NPN after all.

This has me interested because I have a couple Si tranny circuits I'd like to board and box, but I want them to STFU without losing my tone, lol.

[darron]

something to do with pnp with negative ground? i can send you a pcb if you want? (:


i already started a thread quite a while ago open for suggestions to get the noise out, nothing made any major difference, but i'm very interested to know what's going on with  diodes and fuzzes afterwards making it quiet even when not clipping....

[Johan]

i already started a thread quite a while ago open for suggestions to get the noise out, nothing made any major difference, but i'm very interested to know what's going on with  diodes and fuzzes afterwards making it quiet even when not clipping....

could be a impedance issiue...so you put the diodes in parrallell with the 1M at the output?..
J

[Earthscum]

Ack! Ok, well... was gonna try it out, but upon searching for my Germs, I realized I'd already used them in a FF for my buddy and keep forgetting to order more.

[darron]

i already started a thread quite a while ago open for suggestions to get the noise out, nothing made any major difference, but i'm very interested to know what's going on with  diodes and fuzzes afterwards making it quiet even when not clipping....

could be a impedance issiue...so you put the diodes in parrallell with the 1M at the output?..
J

Yeah. Somehow a sensitive output? That's right, the diodes are in parallel with the last 1M.

[WGTP]

If I understand what your wanting to do, the Ge (or other) diodes need to be placed "in-line" to reduce white noise.  This is used in several of the Heavy Metal Petals.  It also creates "crossover" distortion that is similar to that of typical 50-100w tube amps.  Using Ge diodes for clipping is the noisiest choice.  Since the diodes limit the output, the noise stays the same.  LED's will have the highest output vs. Noise, but also the least distortion.  The Ge's will have the most.

Also, the 2N5089 and MSA018 are high gain/low noise transistor that may work well for you.

[darron]

hey WGTP. where abouts do you mean for 'in-line'? it makes me think in series with the signal but i would imagine that to gate like crazy. could you please give me a suggestion where from the schematic?


5089 would make sence, but i'd like to keep it suitable for pnp GE so that i can chuck in any old OCxx / ACxxx / sanyo / rayon etc. etc. it's more about mojo and tonal difference, otherwise i'd probably just use a fet boost or something.


it's still on my breadboard. just found a REALLLLY nice sounding oc83 for it, but i'll keep experimenting with my new stock ac128s

[WGTP]

Whoops missed the Ge transistor part thinking about the other.  A 2N5087 would be a similar PNP SI.

The diodes are in the signal path in series with the output of the Transistor where the 1uf cap is.  You may need to put a cap on each side of the diodes, I'm not sure.  I'm don't know how subtle the sonic impact will be, but I swear I smelled an old 50w Fender Bassman the first time I tried this with a fuzz.  In-line diodes clip the smallest signals so they block the noise IIRC.   

[merlinb]

Darron, do you get the same noise reduction if you use silicon diodes??

[WGTP]

You get more noise reduction, but it starts sounding funny. 

[merlinb]

You get more noise reduction, but it starts sounding funny. 

I'm talking about Darron's peculiar situation with shunt diodes, not series diodes.