BIASING TB MKIII Q3

[The Surgeon]

Hello!

Can anyone please explain how bias in TB MKIII Q3 works?

I built one and used Smallbear´s transistor set and it worked fine. So fine that my friend insisted to buy it and I thought it was a bit
"heavy" for me so I sold it but now I built another one using same components and PCB but transistors that I allready had.

I used 2N1305112 germanium (Hfe 83 leakage 30 mA) in Q3 position. Here´s what I tried in darlington pair: AC126 (poor sustain, awful sound), AC122 (sustain was there but no bottom end), GT308V (decent sound but poor sustain in some notes). At last
a pair of 2N2904A low gain silicons gave good fuzz sounds and almost the same sustain than Smallbear set. Tone pot also works fine.
All trannies were selected and biased well.

It seems that circuitry is very picky with transistors in darlington stage but I think to get the best out of the sustain now I have to do
something with Q3. I have built well-working Maestro FZ1 where the leakage is big part of system but I can´t figure out what is the deal with that Q3 in MKIII. All parts were checked and I chanced germ. diode too with no effect.

There are lot of good explanations about FF and TS but no about MKIII. I really want to learn my lesson well so anyone, please?

[slideman82]

What a coincidence! I wanted one of those fuzzes, as I have built a MKII, but I couldn't make the MKIII sound right. I tried several Ge trannies, the first option was AC128, but the clipping was really bad... just 5 min of prooves were enough to unsolder all the parts from the little PCB I design... what a pitty!
Try with biasing the first pair too (replace that 10k resistor for a 25k pot). Also, you could reduce the 10K resistor from -9V to the 3rd tranny collector resistor, and then use a 25k pot for 3rd tranny bias. With that 10k resistor, even with 0 resistance on the pot I couldn't reach to 4.5 V... take a look at the FFace schemo, that resistor is only 470ohms, that means low output volume, if you rise it, it starts to climb and it does sound better (try 2k for example...).

[The Surgeon]

Thanks slideman, you got me to think trimmer and yes that solved my problem.

I´ve been busy but now I had some time to finish this project. I chanced 18k resistor on Q3
collector to 20k trimmer. That helped me to adjust Vc3 4,5 V but Vc at darlington pair raised
too high so I put 120k resistor instead of 220k to the base Q1 like Mac told before. I wanted to see how those silicons
work on darlington pair and yes they work fine.

I ended up with Vc1,2 3,50V and Vc3 4,49 V, battery was 9,2 V. Also put 4700 pF to the tone control
instead of 2200pf and input cap 0,22uF.

Now the tone and sustain are what they should be, no insane highs and tonepot works well and it has
a hint of FF in it when tonepot is down and gain almost maxed.

This is very versatile pedal but seems tricky one to get sound right but now I´m happy with it and
if I have to build one again I know how to make it sound right, maybe save those germaniums, too.

[mac]

I have a Ge MKIII in my breadboard right now.

I'm using  2SA52 Toshiba PNP Ge at Q1 & Q2, hfe=50, leakage=20uA=0.02mA. They bias at 4.0v using the original resistors. Fine! My guess is that Q1 & Q2 must be in the range hfe:(40,60), leakage:(20uA,50uA). IMHO the darlington collector voltage is fine between 3.8v to 4.3v. If the transistors are leaky, one can always tweak the collector resistor or the 47k or the 220k to set the bias. And temperature is not a problem.

Q3 IS a problem. Very low leakage and it won't bias, as a silicon, not enough current to the base, collector near the 10k's junction voltage. Too much leakage and the collector will be too close to gnd. And the diode leakage interacts too! And temperature send the voltage up and down. NIce!

I tested with similar hfe transistors but different leakage, that is, hfe near 70 but leakage from 20uA to 150uA, and different Ge diodes, 1n34, 1n60, aa117, oa91 and unknown ones X.
For example, a 2sa102 Matsushita (67, 159uA) bias at 2.5v with diode X, 2.0v with aa117, but 0.1v with 1n60!!!
On the other hand, a 2sa52 Toshiba (74, 30uA) bias only with a resistor from B to C. 1.5v with the 1n60 and 1M resistor. The higher the resistor the higher the voltage. I mean the collector voltage, not the voltage at the 10k's junction.
Since it is summer here, voltages will increase in winter most likely. The diode works to some extent to prevent thermal runaway. Parallel diodes may increase the thermal surface. Have to experiment.

How to bias?
If the transistor is very low leakage, < 20uA - 50uA, then a big resistor from B to C is needed, 220k to 2.2M. Think silicon.
If it has moderate leakage, say 100uA - 200uA, then it is a matter of finding a suitable diode, or parellel/serial diodes, or tweaking the 18k. If it has high leakage maybe playing with the 18k is the only way to make it work.

Which is the best voltage for Q3 collector?
Above 0.5v or so it starts to bias (I mean "sound ok") until it gets too close to the 10k junction voltage, which is always below 4.5v because some current goes to the transistor via the 18k. I read from 3.5v to 4.3v at the 10k's. I had best results at 3.7v - 3.9v.
I suggest to set the the collector voltage at  half the 10k's voltage, that is 1.5v - 2.0v to prevent thermal problems.

About the resistor from Q3 B to C. When used, it decreases the final gain, which helps the gain pot that does not works as it should. BTW I prefer a 250k-B pot.

One interesting mod is to replace the 3.3k resistor at the darlington pair with a 5k pot and a 10uf to the center lug. Then adjusting the collector resistor to set bias. It clears the distortion and can turn the thing into a treble booster at low gain settings. It also can produce many distortion textures if the original gain pot is used.

Well, I'm tired of writting...

mac

[slideman82]

Hi Marcelo! Great work with those Ge trannies! I have many of those Matsu and Toshi (and many others too!), but never measure it's leakage (and I have the Geofex hFe-meter almost built!). I think I will give other shot to the MKIII...

Who does it sounds? Have you ever tried the MKII pro?

[mac]

Hi Marcelo! Great work with those Ge trannies! I have many of those Matsu and Toshi (and many others too!), but never measure it's leakage (and I have the Geofex hFe-meter almost built!). I think I will give other shot to the MKIII...
Who does it sounds? Have you ever tried the MKII pro?

The MK3 sounds different from the MK2. I like both, maybe more used to the MK2.

Compare your results with the table at the bottom of this page:
http://www.diystompboxes.com/biascalc/

mac

[The Surgeon]

Thanks Marcelo!

Hmm...maybe I have to play with the darlington Vc although pedal sounds fine.
Hfe Q1/2 were 75 and 82. I tried Hfe 135 on Q2, way too much gain.
Do you think that hfe`s are too high, those silicons seemed to vary 75 to 110?
It seems that my Q3 leaks quite little, 30 uA. I have a bunch of GT308V´s so I can
try them too but they seemed to be very low leakage..I´m not sure how good they are for
this purpose.  I socketed trannies so it´s easy to do.Maybe I bredboard this to make some experiments.

Next build will be MK1, I´ve built Maestro FZ1 3V model and it was fine but I gave it away
so I want to try 9V.

[mac]

Maybe I bredboard this to make some experiments.

+1

I took the output from the darlington pair and it distorts a little, much of the fuzz is done at Q3. Fuzzcentral suggest the use of silicons at Q1,2 for this reason. But I'd use low gain silicons since chances are that the original ge transistors were low gain, 70-100.

AFAIK, those GTs are low leakage, you may need a 1M or so between B and C.
Did you tried the ACxxx at Q3? Yesterday I put a AC121 (85, 130uA) and with one of the many diodes I have it sounded good, collector at near 2.5v.

mac

[The Surgeon]

Yep, tried first AC126 and then AC188K, which worked well in FZ1 maybe because of
higher than average leak. They didn´t sound good. Now I have more 20+ some 188K´s, didn´t
check them yet.

AC126´s were rather disappoitment anyway, just like "newbie" AC128.

Have to go to work now...by the way, sorry for my english...haven´t written it for a long time.

[mac]

Have to go to work now...by the way, sorry for my english...haven´t written it for a long time.

don't worry, I speak one language, spanish, and very bad!!!

mac

[slideman82]

The FZ1 and MKI sound is quite different to the MKII, FF or other well-known fuzzes. These first era fuzzes sound quite farty! Just listen to some The Who's first albums, sounds horrible! Also, search in the forum for "country fuzz" and you'll get a post that takes you to a blog where are posted a lot of old 60's country songs, with lots of fuzz! Some sound awful! Well, it was the first fuzz, the MKI, in some.

Damn, Marcelo! I'll have to measure my trannies' leakage!

[mac]

Damn, Marcelo! I'll have to measure my trannies' leakage!

Uh! Make lots of coffee
I hate measuring ge.

mac

[The Surgeon]

That´s why I built FZ-1, wanted some primitive power! It was real surprise how good it worked. It had sound of it´s own
but warm, long sustain, worked real well with single coils. I sorted trannies by leakage (yes I read all I could find from the forum).
With 3V, I couldn´t believe my ears. The guy who works as a movie sound producer wanted it so I have to build new one and
I´m going to try 9V model, they are very similar 3V FZ-1/TB MK1 and Zonk Machine? Now we have MK1 cirquit on the forum
which is very nice, I´m going try it.

I think AC188K fits well in that cirquit, never tried anything  else. Sounded better than Satisfaction- riff, which is pretty crappy.
I think sound will vary very much depending on what trannies is used and what their leakage are.

I´m after those early times sounds, I´ve built couple of Fuzz Fazies, Axis Face Silicon, Range Master, Phase 45, some tubescreamers,
Ross/dynacomps and newer pedals but I allways return to those old sounds.

[mac]

I´m after those early times sounds, I´ve built couple of Fuzz Fazies, Axis Face Silicon, Range Master, Phase 45, some tubescreamers,
Ross/dynacomps and newer pedals but I allways return to those old sounds.

I like old tones too. And I guess most users here do.
Fuzz Face, TBMII, Rangemaster, Phase 90, Red Fuzz to push tubes are the ones I always use. Dist+, Faim Fuzz, Bazz Fuss not so often.  I built other pedals that I finally sold.
I'm needing some stuff like ring mods, octave and noise makers, chorus and delays. I know it is easier to go to a store and buy a Boss digital delay or chorus, or a Vox wah, but I prefer DIY projects.

mac

[slideman82]

Yeah! I love fuzztones too! I've just finished a Fuzz Face with a couple of AC128 that are quite perfect, it sounds soooooooooooooooooooo good!
I've made a couple of FZ1-A, it does sound really good, a bit trebly, and the best trannies that worked were AC125/126, but I never biased them.
MKI must sound similar to FZ1...

[mac]

I've made a couple of FZ1-A, it does sound really good, a bit trebly, and the best trannies that worked were AC125/126, but I never biased them.
MKI must sound similar to FZ1...

There is a thread about MKI, FZ1 and other sharing the same topology. It list the differences, minimal, between circuits.

mac