Multi face with mosfet in q1

[fluffhead]

Hi all
Throwing together my first pedal. After 15 years I'm finally building a germanium fuzz face. To start out, I breadboarded a silicon npn face following small bear. It sounded great. Then I swapped out the silicon for the two best npn germaniums I was able to find at my local shop, it still sounded great! Then I got sucked into the rabbit hole after reading "the many faces of fuzz" and seeing the multi face plan, and spent a night swapping transistors and mosfets (2n2222a, 2n3904, 2n4105, 2n7000, bs170, etc... my local store has a ton of transistors). When I hit upon THE tone: fat, rich, but so fuzzy, and it was... beautiful. Cleaned up perfectly with my volume knob and responded beautifully to my picking. I swear it was a 2n7000 in q1 and my best Ge 2n4105 in q2. Then I got too confident and tried to turn it into a fuzzcentral axis face and...

it.

stopped.

working.

So I stripped it all down and went back to the small bear breadboard, with q1 2n7000 and q2 2n4105. And nothing. It will work with the mosfet in q2, and any of my npn trans in q1, but no sound with it set up like it was. I will swear all day that the combination worked, in fact it was so nice that I bought a few tested npn germaniums online, tHinking it would be even better with a better transistor than the crapshoot ones from the store. But now I don't even know. I know it's possible, right? There's a sound clip of a similar set up in the multi face. Oh yeah, and I still haven't even built the germanium fuzz I set out to build...

TLDR: why isn't 2n7000 working in q1 of a fuzz face? I've searched the forum for an answer but can't really find anything that can help me, besides one post with a vague reference to biasing q1 that goes over my head. Is the solution the 10k trim pot on the emitter of q2 in the multi face? What does that trimmer do? Why is it on the emitter of q2 as opposed to the collector, like I did to bias it as suggested by small bear?
referenced:
http://diy.smallbearelec.com/HowTos/BreadboardSiFF/BreadboardSiFF.htm
http://web.archive.org/web/20050513200126/www.geocities.com/tpe123/folkurban/fuzz/fuzzyfaces.html
http://www.home-wrecker.com/multiface.html

[Kevin Mitchell]

I read the long part and thought "Oh, they're not biasing the thing" and then seen the TLDR part 

This whole mosfet or jfet in a fuzz face is new to me. But from the quick research I've done, you'll have to make some real adjustments to get the device to play along. You need to get the mosfet stable and to it's correct bias point. So start by learning how to bias transistors!

And welcome to the party/forum.

[antonis]

Hi & Welcome..

Do you really want a practically zero leakage current device to act as a Ge one..??

[fluffhead]

Every time I try to look for info on biasing transistors, either it is written for an electrical engineering student  or specifically about fuzz pedals, but only about q2 collector biasing. That's why I have the questions about the trim pots. If someone could point me to something about biasing transistors that's a little more complete without requiring a ton of EE knowledge, that'd be great.

Regarding the leakage of q1, I didn't think that was an issue. The all silicon and Si/Ge fuzzes worked fine and those are also zero leakage going into q2.

Has anyone put a mosfet in q1 and had luck? The only thing I can think of that was different was I had a trim pot on q2 collector, not a fixed resistor, but I didn't think that was affecting the bias of q1.

[Rob Strand]

Are you sure you had Q1 as the MOSFET and Q2 and the germanium. The other way around
definitely works without too much thinking,  I did something similar back around 2000,
I think Jack did it first,

http://www.muzique.com/schem/fuzzface4.gif

(It is possible to get Q1 MOSFET working.)

[fluffhead]

I'm like 99.23% sure it was q1: 2n7000 and q2 2n4105. I immediately wrote a note and circled and underlined it on my idea pad. And the 2n4105s I have are pretty leaky and sound hissy and gated in q1. The other way sounds pretty good, but it's not as strikingly obvious to my ears. Seriously it was like a bolt of lightning.

Idk something's different tho. After redoing the breadboard, my q2 collector is biasing at 3.5 rather than 4.5v like it was when it had THE tone. So maybe it isn't exactly the same.

I found a R.G. Keen paper on biasing mosfets, maybe it'll help. God it gets technical tho. Any tips on how to get it to work, Mr. Strand? Is the multiface layout I linked to the way to go? Will it at least point me towards a solution?

I know I'm not crazy or high, it was working!!

[Rob Strand]

Idk something's different tho. After redoing the breadboard, my q2 collector is biasing at 3.5 rather than 4.5v like it was when it had THE tone. So maybe it isn't exactly the same.

I found a R.G. Keen paper on biasing mosfets, maybe it'll help. God it gets technical tho. Any tips on how to get it to work, Mr. Strand? Is the multiface layout I linked to the way to go? Will it at least point me towards a solution?

I know I'm not crazy or high, it was working!!

The key point about MOSFETs is gate needs to sit at about 2V.  That's much higher than the 0.10V (ge) or  0.6V (si) you see on the base of a transistor.    So that's where I'm not 100% sure what you did when you got it working.   The normal part values in the FF circuit won't bias a Q1 MOSFET correctly.

In order for the gate to sit at 2V on in the FF circuit the voltage across the 1k drive pot needs to be about 2V.  That's somewhat higher than the normal voltage there.

I'm assuming you more or less followed the basic FF circuit, and did not make radical changes.    In this case the only way I can see the circuit working is if you use a very low collector resistor on Q2.   If you had a trimpot on the collector of Q2 you might have set it to a low resistance and that would work.   It might need to be 1k or 2k, something low like that.   Just play with it and you will get it.

You can see some circuits here,
http://www.oocities.org/tpe123/folkurban/fuzz/mosfetface.gif

Notice that in circuit B the Q2 emitter resistor sort of the same size as the Q2 collector resistor.   Here the Q2 collector resistor is 10k and the Q2 emitter resistor has been increase from the normal 1k value.   It is equally valid to leave the Q2 emitter resistor at the common 1k value and adjust  the Q2 collector value to be 1k or 2k.

Oh, with a low value resistor on Q2's collector you might need to lower the resistor on Q1's drain.
I just did some calculations and a 100k drain resistor on Q1 should still work but lower will also work - maybe play with it to change the sound a bit.

[anotherjim]

It's always possible to zap a MOSFET from a static charge to its gate. I think you'd need to be very unlucky for that to happen.
If you want Q2 emitter resistance at it's usual lower value, you might add a bias resistor between Q1 drain-gate. Try 1M to start? If Q1 sits in the off state, it's drain voltage should pull higher and a bias resistor might get the gate more positive to bias Q1 on again.

[fluffhead]

Ugh I hope I didn't zap it... and now I'm really unsure if it was in q1 or not. I definitely didn't tweak it that much, I changed the 330R resistor to 1.2k, and the q2 bias resistor to a trimpot (which was set pretty low). Maybe I got lucky and something clicked. I do know the voltages were a little higher when it worked, now q2 emitter is at 3.5 (and won't go higher, the trimpot is at its lowest) when it was at 4.5 before I redid it.

I did find this post in reference to someone trying mosfets in q1 and q2:

I suspect it's a bias issue. Those two do differ in one place - that being the arrangement of resistors on the emitter/source of Q2. The second has a trimmer for setting the DC voltage on Q2 source, and has the gain control AC coupled through a cap. In the first one, the DC source resistor is fixed.

That means that you can trim in the DC voltage at the source of Q2 to properly bias Q1 in the second version, but have to take pot luck on the modified MOS Face. MOSFETs have minimum gate-source on voltages of 1.6 - 3V, so I suspect that when you put in a MOSFET for Q1, it's not getting enough Vgs to turn on.

Converting the Q2 source resistor to a trimmer and gain pot as in the second version should clean this up.

He's talking about stock MOSface vs. the multiface and "many faces of fuzz" schematics. I've also been comparing fuzz face and rangemaster schematics and an AMZ mosfet boost article to better try to understand the biasing issue. Thanks guys, hopefully something will work. And it better sound as good as it did....

[fluffhead]

I got it working! After trying to convert it to a multi face (and failing...) I just said screw it, I'm just ready to box this up. After all, a working fuzz face is better than none. And I have a three knob tone bender I'm dying to build! So the plan was to breadboard a stock npn face for the final time, then get it on a permanent board with sockets for the transistors so I can always switch around. I decided to try a mosfet in q1 one more time, just to confirm it wouldn't work, aaaaaaand... IT ACTUALLY WORKED! I have a BS170 in there right now and it sounds great. And the cool thing is the bias trimmer actually gives me a range of usable sounds, from that beautiful rich sustaining fuzz that hit my ears like a thunderbolt, to a slight oscillation, to gated glitchy robot guitar with no decay. I'm thinking about making it a full size pot.

This shouldn't work, should it? But it does. And it's awesome.

[Rob Strand]

Good you got it working.

This shouldn't work, should it? But it does. And it's awesome.

If you measure the DC voltages on the transistors it will it makes sense.
Suppose that Q2's collector resistor isn't *that* low and Q2 is stuck on,
then there still can be enough gain from Q1 to drive Q2 on/off perhaps giving
a gated effect.

[fluffhead]

Yeah, I get why it has such a pronounced gating effect at higher resistance. Measuring the DC voltages was the first thing I did after hitting an E chord. Thanks for your help, even tho I ended up ignoring all the advice I was given!
I actually understand the fuzz face circuit now, and biasing, and common emitter amplifier circuits. Plus I learned a lot about mosfets, and a simple booster is now on my project list. Cool little beans, those semiconductors are.

I'll post some pictures/clips after I get it together and in a box