Germanium vs. silicon vs. Schottky diodes for a fuzz

[VSmolsky]

Hi all, it's my first message here.
I hope I will not violate any forum rule 
Diodes are definitely a hotly debated topic, I think it's been talked about hundreds of times, but my request is a little more specific.
I have an EHX Satisfaction fuzz which as you may know can easily be modded as the PCB has holes for new components. The first mod is the "Bass Mod" by adding a cap in C7 (the most used is the 22nF one, I'll do this). the second one concerns the diodes. The PCB has 4 holes to add two diodes in D2 and D3 (see pic attached)

I've read that many use silicon diodes to bring it closer to the Jordan Boss Tone. I, on the other hand, want to bring it closer to the Selmer Buzz Tone, a 60's fuzz that uses germanium transistors related to the Maestro FZ-1 Fuzz Tone (I read they use the same transistors, 2n2614 or RCA 2N2613).
I know they are different pedals with different layout and different transistors, I just want to take the Satisfaction into that direction without any presumption of making it sound identical. Just making it reminiscent of that gritty and mid-focused tone.
For this reason, I think silicon diodes are not a good choice.
Suggestions for a couple of good germanium diodes? I had thought about the 1N34A, but I have no idea of which could be the second one.
Secondly, how about if I use a Schottky diode? What sound would I get? Can I put it in this circuit?

[Mark Hammer]

Welcome on board. 

The role of diodes is often misunderstood.  In the position they occupy in this circuit, they not only clip, but limit the maximum output.  If one uses diodes with a lower forward voltage, such as germanium or Schottky, you will get serious clipping more easily, but will sacrifice a lot of output level, in comparison to using silicon diodes. 

Now, the Bosstone circuit has plenty of output, so maybe that's not an issue, right?  But maybe you might want to insert some sort of filtering or tonestack between the diodes and output level control.  Such a passive circuit would eat up signal, and perhaps require the input sensitivity (remember that gain is fixed in this circuit) to be dimed all the time, in order to have any appreciable volume level at the output.  So my own vote would be to leave the silicon diodes in there, and add some filtering.

Alternatively, a pair of Ge diodes, with a 1k pot between the diodes and ground would offer variable limiting/clipping, that you may find useful.  More sizzle and bite can be achieved by reducing C3 (1n5 to 1n) and dropping C2 (22n to maybe 18 or 15n). 

Finally, if you've never toyed with it, consider adding some optional crossover distortion to the clipping the circuit provides.  Clipping sets the ceiling on the positive and negative excursions.  Crossover distortion creates a "wall" that the signal has to cross over before it passes.  I'll note that crossover distortion is crucial to the sound of the Boss HM-2 and all its imitators.  To implement it, you would stick a back-to-back pair of schottky or Ge diodes in series with the signal, just before the volume pot.  Note that they WILL deduct some voltage from the output, which is yet another reason to stick with the silicon clipping diodes.  My experience is that crossover distortion LOVES bridge humbuckers/P90s.

[VSmolsky]

Hi Mark, thanks for the welcome.
I didn't know that the diodes in that position also limited the output volume. It depends on how much the pedal lowers the volume, whether by a little or by a lot. I can turn up the volume on the amp.
I'm thinking of using sockets like these and experimenting with different diodes (Si, Ge, Schottky). I don't want to make too many changes to this pedal (crossover, additional caps, etc...) once i find a good combination of diodes i just keep that ones.
https://www.banzaimusic.com/SIL-20-pins.html

The Si + Ge combination seems interesting (1N4148+1N34A)
https://youtu.be/bifKfe04TBc?t=60

[VSmolsky]

Excuse me Mark, what are the parameters from which I can get the output of a diode?
Is the forward voltage enough, or are there others?

[idy]

Really FV is it. Normal two diodes ("anti-parallel") and the peak to peak is twice the FV. About .6 for 2 Ge, about 1.2 for 2 Si, etc.

[VSmolsky]

Really FV is it. Normal two diodes ("anti-parallel") and the peak to peak is twice the FV. About .6 for 2 Ge, about 1.2 for 2 Si, etc.

Thank you. Is there any contraindication to using two different types, for example a Si and a Ge or a Si and a Schottky?
I would like to do some experiments.

In case of 2 different diodes, each with its FV, is the total FV given by the sum of the two (e.g. 0.3+0.6=0.9)?

[Steben]

The exact output is always "about" the Vref...
The impedance/resistance of the signals around influence this.

Ge and Schottky diodes are usually labeled as the same region of forward voltage but this can change a lot depending on the circumstance.

Just as silicon power transistors usually have less Vbe interval at the same current as low signal transistors.

[MikeA]

Really FV is it. Normal two diodes ("anti-parallel") and the peak to peak is twice the FV. About .6 for 2 Ge, about 1.2 for 2 Si, etc.

Thank you. Is there any contraindication to using two different types, for example a Si and a Ge or a Si and a Schottky?
I would like to do some experiments.

Two diodes anti-parallel with different forward voltages will give you an asymmetrical positive/negative waveform, with one direction clipped more than the other and/or with a differently distorted shape. 

[Phend]

The only important clipping is the ear test.
(Funny how we like Fuzz in our ears, but anyway)
Looking at waveforms is like looking at the menu at the ice cream shop.
Lots of different flavors shown in pictures,  make mine vinillia.
Maybe some like sprinkles.
Thus experimenting with diodes is like tasting ice cream without the weight gain.
Find your best combo and you will be happy.
PS if you plan on switching between two sets, use one set high and one set low.
Else you may not hear much difference hence rendering the switch more or less useless.

[Mark Hammer]

I didn't know that the diodes in that position also limited the output volume. It depends on how much the pedal lowers the volume, whether by a little or by a lot. I can turn up the volume on the amp.

The level-affecting role of diodes depends on whether there is any gain recovery, or severe attenuation, associated with them.

The old MXR Distortion+ used a pair of germanium diodes to ground. Between that and the too-low value of the original volume pot, you had to dime the thing to get anything much louder than bypass.  That's fine if it's the sound you want.  But if one wants a mildly distorted tone AND a bit of a volume boost for solos, it wasn't going to happen.

Germanium diodes to ground are going to limit the maximum output to about roughly +/-300mv, depending on the specific diodes.  A pair of silicons (as was used in the DOD250 and similar) will get you +/- roughly 550-650mv, which will let you push the amp, or any other stages between the pedal and amp, harder and also allow for achieving milder clipping with a decent output level.  Alternatively, what some circuits do is include a "gain recovery" stage after any clipping that severely limits the output.  It can be an op-amp, or a single transistor stage, like one sees in the Big Muff.

[VSmolsky]

I didn't know that the diodes in that position also limited the output volume. It depends on how much the pedal lowers the volume, whether by a little or by a lot. I can turn up the volume on the amp.

The level-affecting role of diodes depends on whether there is any gain recovery, or severe attenuation, associated with them.

The old MXR Distortion+ used a pair of germanium diodes to ground. Between that and the too-low value of the original volume pot, you had to dime the thing to get anything much louder than bypass.  That's fine if it's the sound you want.  But if one wants a mildly distorted tone AND a bit of a volume boost for solos, it wasn't going to happen.

Germanium diodes to ground are going to limit the maximum output to about roughly +/-300mv, depending on the specific diodes.  A pair of silicons (as was used in the DOD250 and similar) will get you +/- roughly 550-650mv, which will let you push the amp, or any other stages between the pedal and amp, harder and also allow for achieving milder clipping with a decent output level.  Alternatively, what some circuits do is include a "gain recovery" stage after any clipping that severely limits the output.  It can be an op-amp, or a single transistor stage, like one sees in the Big Muff.

I know, unfortunately. I have both the Distortion+ and the DOD250, and the only possible setting for the Distortion+ is full volume and gain.
So if I use two germanium diodes, will I probably have a pedal that works like the Distortion+?
That's why I also put silicon diodes in my cart...

[VSmolsky]

Thank you. Is there any contraindication to using two different types, for example a Si and a Ge or a Si and a Schottky?
I would like to do some experiments.

Two diodes anti-parallel with different forward voltages will give you an asymmetrical positive/negative waveform, with one direction clipped more than the other and/or with a differently distorted shape.

This one doesn't bore me much, it's full of asymmetrical clipping pedals that sound good. The Boss SD-1 is probably the most famous, but also the Fulltone Fulldrive and some variants of the ProCo RAT. My goal is not to have asymmetrical clipping, but to have a bit more output volume than 2 germanium diodes.

[Mark Hammer]

I know, unfortunately. I have both the Distortion+ and the DOD250, and the only possible setting for the Distortion+ is full volume and gain.
So if I use two germanium diodes, will I probably have a pedal that works like the Distortion+?
That's why I also put silicon diodes in my cart...

If you mean germanium diodes in the design risk the same disadvantages as a Distortion+, then my answer is "Yes".  But again, a gain recovery stage can sidestep those disadvantages.

[idy]

Asymmetric is one cheap trick that often does wonders. Simplest is just one diode. Other direction gets opamp clipping only. I have an old D+, messed with it and added silicon diodes... tried to clone it, never sounded right, finally noticed one of the diodes the glass was cracked.
Asymmetric can be like lightening in a bottle. The unclipped (or later-clipped) side gets big enough to make other pedals or your amp clip. You get distortion but still hear your guitar. Then at higher gain something else starts to clip (like your amp) then at higher still the opamp starts to clip... Or just playing harder those various things happen.

[Mark Hammer]

Asymmetric is one cheap trick that often does wonders. Simplest is just one diode. Other direction gets opamp clipping only. I have an old D+, messed with it and added silicon diodes... tried to clone it, never sounded right, finally noticed one of the diodes the glass was cracked.
Asymmetric can be like lightening in a bottle. The unclipped (or later-clipped) side gets big enough to make other pedals or your amp clip. You get distortion but still hear your guitar. Then at higher gain something else starts to clip (like your amp) then at higher still the opamp starts to clip... Or just playing harder those various things happen.

Somewhat true, but depends on the gain staging.

[matopotato]

Welcome to the forum!
I agree that experimenting is what gets you closer to something you will like rather than what the consensus might be (ideally they coincide). If possible some clips or wires and try out different combinations. or perhaps single socket legs. Then you can keep them in and swap around later if you should change your mind.

[VSmolsky]

If possible some clips or wires and try out different combinations. or perhaps single socket legs. Then you can keep them in and swap around later if you should change your mind.

I'll take this route for now. I chose about ten transistors (at 0.15-0.20€ each...) and some sockets to test several of them, because I can't predict in advance how they will sound.
A curiosity: the PCB is designed for 2 transistors, but can I also put just one?

[bluebunny]

A curiosity: the PCB is designed for 2 transistors, but can I also put just one?

I don't see how that would work.  Here's the alleged schematic.

[antonis]

It could work, by turning Q1/Q2 Sziklai pair into a single BJT CE amp, but it should need re-bias 'cause Q1 Collector current isn't determined by Q2 V_BE anymore..

[matopotato]

If possible some clips or wires and try out different combinations. or perhaps single socket legs. Then you can keep them in and swap around later if you should change your mind.

I'll take this route for now. I chose about ten transistors (at 0.15-0.20€ each...) and some sockets to test several of them, because I can't predict in advance how they will sound.
A curiosity: the PCB is designed for 2 transistors, but can I also put just one?

Oh, I thought you were trying different diodes?
If that is what you meant, I suppose you can use one diode only, but it might not be very interesting.
You could try (if diodes) with 3 of a kind. One single as intended and the other two in series. A 1+2 clipping set. Others can explain better but it should affect even/odd harmonics. Somehow. But if you really mean the transistors, then I have no idea.
Keep us posted.