Tone bender mkII pro, leaky germ trannies are good?

[m_charles]

Hi all.
Was hoping some of you with experience building the two knob MKII tone bender could help.
Here is the dilemma: built up a mkII bender. Q1, 2, 3 hfe's: 60, 70, 110 respectively. All very low leakage. I did replace the Q2 collector (100k) with a 39k to pull the Q3 collector voltage down to around 5v or so. I've seen some recommend that you keep the 100k and expect the voltage to be around 8v at the Q3 collector. I've done this, but the pedal just doesn't sound very good IMO.
Anyways, built up the pedal, and it sounds fantastic. However, doing a rollback on the volume causes a huge hiss to creep up. With the guitar volume set to zero, the volume of the guitar is off, but the hiss is very loud. Not a normal, fuzz pedal loud, but a "something isn't right" loud. So my immediate thought is: impedance issues. So I remove Q1, check it, all good... scratch my head.
Getting frustrated, I say F-it, it's a pretty simple circuit, and just build up another. Fortunately, a while back I bought a big bag of germs on the cheap, which actually turned out to be good and about 60% good transistors, which allows me to be able to experiment.
So I build up another, plug it into a breadboard so I don't have to wire it all up again, and bam, sounds great, volume rollback is what I consider normal as far as hiss, and it's all good.
Now I set about figuring out why one is all good, but the other is hissy only on the volume rollback. 
I can't find a difference. Nothing. This makes me batty!
So... Being the nut case that I am, I build a third. It turns out the same as #1! Sounds great, hissy with roll-back on guitar vol.
Arrrg! So I do some research. I see online that some write the hiss issue off as a "hissy pedal" but my theory is that they built one like I did, found the hiss thing on the rollback, and just moved on. But since #2 is all good, I know it's possible to have one that responds like a good fuzzface.
One thing I have seen hinted at, is that it can be good to have a LEAKY transistor in Q1??? I haven't tried putting in an intentionally leaky trannie in there yet, but wanted to ask if any of you could tell me a reason that this may be true? Or is it just BS? Perhaps my #2 build had a leakier tranny than I realized, and that is why it sounds good? I know I could go in and start cutting, and measuring, but seeing as how I have now built 3, I figured it was time to pop a question on the forum.
Any help would be GREATLY, GREATLY, appreciated. This is making me nuts...
Thx so much in advance.
-chuck

Schem is the first one on top of page:
http://fuzzcentral.ssguitar.com/mkII.php

[yeeshkul]

The Q1 needs some leakage current to bias itself - as you can see, its base is not connected to DC. However saying you need a really leaky tranny is an overstatement. They are all leaky enough, use any one, maybe even a piece that is not-that-leaky.

[LucifersTrip]

I did replace the Q2 collector (100k) with a 39k to pull the Q3 collector voltage down to around 5v or so.

should be 7.5 - 8.5+, then go from there. possibly adjust the B to ground resistor on Q1 since you're using low leakage transistors

edit: are you using a battery? is the hiss at all levels of fuzz?  are the other voltages close to working voltages of known good ones? are the voltages the same in #2 as #1 and #3?

[Arcane Analog]

Some germanium is inherently hissy. This is device specific so auditing them individually is essential. No two devices are the same - even if two devices "measure" identical more often than not there will be differences.

This is a MKII not a Fuzz Face - Q3's collector should be at 8V+ as is found in the original units. If it does not sound "good" you are either doing something wrong on you do not like the MKII.  

Never rely on Fuzz Central for anything. Ever. The site is full of flaws and misinformation. These forums are much better.

[bwanasonic]

It is always stated that the TB likes leaky transistors, but I've never seen the leakage quantified. >100uA?

[midwayfair]

It is always stated that the TB likes leaky transistors, but I've never seen the leakage quantified. >100uA?

I think ElectricWarrior gave numbers for the leakage at one point (possibly even measured from one of his original units), but I'm not sure it was on this forum. I seem to remember something closer to 200uA-300uA being called for in Q1 at least and perhaps Q2. Also, I'm almost positive LucifersTrip has given his leakages in a well-built unit in a post around here.

[Electric Warrior]

I never unsoldered the transistors in my vintage unit. But yeah, leakage is good in my experience. I've only ever built the OC75 version of the circuit, though (10k between Q1's base and ground, 47k on Q2's collector). I don't know how the OC81D version (with the 100ks) behaves.
OC75s leak a lot. Mine are mostly between 200 and 500µA.

With too little leakage (I tried 2SB175s) a MKII is usually hissy as hell. And if you do manage to introduce some gating it will sound horrible. Once you have enough leakage the gating will sound much sweeter and you can dial in just the right amount of it by swapping transistors (Q2 has the most effect on this - you could also use a trim pot on its collector, but there's no need, really). You'd probably want to set it to clean up nicely but still not be hissy at room temperature. As it gets warmer it will become more gated, colder will make it more hissy.

You'd probably want to ignore all advices regarding "correct" hfes. A lot of different setups can work. Just swap transistors and use your ears.

Here's some voltages from my vintage unit:

Battery: -9.67V
Q1 C -9.02 B -0.03 E 0
Q2 C -0.17 B -0.08 E 0
Q3 C -8.44 B -0.17 E -0.11

Not hard to get into that ballpark with the right transistors. In fact I never got any voltages that were far off when I used OC75s.

That said: It can be made to work with really little leakage as well. Some original units came with Impex S3-1Ts, which leak next to nothing. They found a clever way to set up hfes that worked with the transistors they used. I tried to set up my clone work the same way with low leakage OC types, but never managed to make it sound good and behave well at the same time. Going high leakage is the easier route by far.  

[m_charles]

Thx so much guys.
Especially Electric Warrior!
I am going to go and experiment more with it tomorrow and report back.
Learn something new about germaniums everyday.
Not a joke question: were germanium trannies used in the first space missions? If my electronics history is right, I think they must have.
Must have been crazy trying to deal with leakage and such designing a rocket!

[Electric Warrior]

Silicon transistors were available back then. And so were ICs: http://www.ieeeghn.org/wiki/index.php/Integrated_Circuits_and_the_Space_Program_and_Missile_Defense

[PRR]

> were germanium trannies used in the first space missions? If my electronics history is right, I think they must have.
Must have been crazy trying to deal with leakage and such designing a rocket!

Silicon was barely available for the first space-flights but not proven in long-term use. Yes, Apollo et al did jump to Silcon ICs.

Exquisitly detailed reports on Telstar are available.

Being a Bell Labs project, Bell used the WELL-proven point-contact technology.

For those who missed it: http://en.wikipedia.org/wiki/Telstar

For added fun: nobody told Bell that the US had just set-off a nuclear bomb at high altitude, and later the USSR did the same thing. Radiation rates in orbit were hundreds of times higher than Bell had expected. After some months the point-contact transistors tended to latch-up and misunderstand commands. They ticked the logic various ways, and got spurts of operation, until a mis-understood command turned-off the main power relay.

Bell System Technical Journal articles:

http://www3.alcatel-lucent.com/bstj/vol42-1963/articles/bstj42-4-739.pdf Summary of results

http://www3.alcatel-lucent.com/bstj/vol42-1963/bstj-vol42-issue04.html Dedicated issue of BSTJ

http://www3.alcatel-lucent.com/bstj/vol42-1963/articles/bstj42-4-1749.pdf Electronics package

[yeeshkul]

"With too little leakage (I tried 2SB175s) a MKII is usually hissy as hell."

I am not sure i understand this statement. Thought the leakege adds to hiss.

[LucifersTrip]

"With too little leakage (I tried 2SB175s) a MKII is usually hissy as hell."

I am not sure i understand this statement. Thought the leakage adds to hiss.

It depends on the circuit...in this case, if the leakage is higher, Q3 collector will rise closer to the supply (9v). When it gets near supply, it'll get closer to gating and get quieter.  You can test that easy. Hook the DMM to Q3C and hold the transistors with your fingers (simulate leakage) and watch the voltage rise.

You might think that Q3C voltage should drop, but don't forget about Q2 which is hooked straight to Q3B.  Q2C voltage will drop with leakage making Q3B voltage smaller and Q3C larger. That's why choosing Q2 is so important for the gate you want. 

[yeeshkul]

Thank you, that makes sense.

[m_charles]

Just to reiterate what has been said, I popped in a silicone trans in Q1 just for the heck of it, and the circuit would not even pass a signal.
Weird stuff!

[midwayfair]

Just to reiterate what has been said, I popped in a silicone trans in Q1 just for the heck of it, and the circuit would not even pass a signal.
Weird stuff!

Silicon does not have leakage (well, it's a very, very, very, very, very tiny amount of leakage), so there's no necessary positive bias on the base. That's why you need a germanium in that position.