How to use non-leaky Ge Transistors in a Melos Fuzzder or FZ-1A type circuit?

[Osprey66]

Hi Folks -

I'm new to all of this and hoping someone who's been down this road before can help me out with my first build. I'm building a Melos Fuzzder FS-1 which from what I gather is pretty close to the Maestro FZ-1A. I have an original unit and trying to 'back engineer' it to get a similar sound, however I now know this circuit was designed around leaky 'rejects' and standard low-leak transistors don't work. All you get is a very gated fuzz.

My question is this: Can you make non-leaky Ge transistors behave the same as leaky ones in this circuit? If I put 330K to 1M resistors between Base and Collector on Q1, Q2 and Q3 I can get similar readings as my original unit, but is this really the best approach to work around with non-leaky Ge?

I'm using Matsushita 2SB175 same as original and I can actually get a nice sound this way, however it isn't nearly as nasty or over the top as the original.

Many thanks for any help or advice.

[PRR]

Welcome.

A very high value C-B is a good approximation to leakage.

"Better" might be low-quality diodes "backward" C-B; but these will vary 1000:1 in leakage and you may never find a "right" one.

Also try 10r to 200r in series with E and C pins. "Reject" transistors don't just leak, they may not conduct so well. In small-signal work no difference; but if you are slamming the little things the stray resistance may take an edge off the sound.

I believe Small Bear has huge quantities of Ge parts and may know which are "bad" enough for your use.

[Electric Warrior]

Interesting. I wasn't aware that they used rejects, but buying some germanium types sure can be tricky because they are available with vastly different specs.

How far off are your voltages? Rebiasing may be worth a try. With three transistors to bias it can be messy and confusing, but since you've got a vintage unit to compare your clone with it should be worth a shot.

[PRR]

> wasn't aware that they used rejects

"Reject" may be too strong.

I can buy near perfect pine boards; or I can buy less-perfect rough knotty stuff. In my rustic cabin, the "reject" may suit the mood better, especially considering the price.

Making Ge transistors was messy baking. They came out everything from very-good to hardly-working. The best became Computer Logic, then Computer drivers, then RF radio amplifiers, then they sorted-out IF amps, then Audio amps, then cheap Audio amps, etc. Even non-working transistors found use as diodes. Sorting-out was a big cottage business in the old days. Small Bear is still sorting through old barrels of Ge parts.

[Phoenix]

If you want to be really authentic (to the point of lunacy), use an NTC thermistor from C to E instead of just a plain resistor, to get that temperature dependent leakage

[Osprey66]

Thanks for your responses and suggestions.

To answer your question Electric Warrior, I have about a dozen 2SB175s to choose from hFE 50-100 with none having measured leakage over 200uA and could not get the circuit to work properly until using the C- B resistors.

The measurements on my original Fuzzder FS-1 are:

Q1 hFE 85 c=1.6v , b=.70 , e= .64

Q2 hFE 33 c=.53v , b=.08 , e=0

Q3 hFE 51 c=.88v , b=.08 , e=0

Without the C-B resistors, my 'clone' measurements would be something like:

Q2 c=1.5v b= .04 , Q3 c=1.6v , b=.008 , Q1 was pretty close to original with a transistor around 200ua leak.

I'm guessing these numbers are what you see when the transistor isn't turning on?  And that's why it sounds gated. Still learning all of this.

I plan on trying some trimmer pots to experiment a little and see what will get the best fuzz. Also, I'll try to get measurements from the other components on the original board to see how they look in comparison to spec.

[Electric Warrior]

Battery: 1.59

Q1 C -1.59 B -0.57 E -0.8
Q2 C -0.71 B -0.09 E 0 (C -1.44 at min fuzz)
Q3 C -1.17 B -0.07 E 0

Transistors are all 2sb175 B. It's very similar to this one: http://stompboxes.co.uk/forum/viewtopic.php?f=10&t=1707&hilit=monacor

As the voltage on Q1C is the battery voltage, things tends to look rather similar at first glance. The actual differences need to be observed on B and E.
The fuzz control sets the bias on Q2, so the voltages there may vary a lot. It's probably best to measure at maximum and minimum fuzz

The voltages on Q2 and Q3 being rather high does indicate that they may have too little leakage. How much do the transistors in your fuzzder leak?

[brett]

Thanks for this post and thread.
I note that the 2015-origin thread concerning an FZ-1a not working failed to address a lack of leakage as the reason the base of Q1 and Q3 were barely above/below the emitter voltahes and the transistors were Off.
In particular, nominating a 330k resistor for between Collector and Base is consistent with my results.  220k is a load of leakage, 470k is little.  (At 1V Collector-Base they are about 4 and 1 uA, respectively.  Note, however that Vc-b depends on Rcb, so it's not really a linear relationship).
For better or worse, I'm running a 2V (3x1N4148 diode) power supply, with 1uF coupling caps and GT313 transistors (hFE between 40 and 50, leakage 1-20uA).
Cheers and thanks again.