Superfuzz troubleshooting - dies after warmup

[Martin Beer]

I have a Univox Superfuzz clone I've just finished putting together, and I have an unusual problem with it. When it is first switched on, it works exactly as expected and sounds great - all controls working correctly. After a short time (about 1-5 minutes) the output gets quieter before disappearing entirely. If I disconnect the power supply for a few minutes, it repeats this behaviour. It does the same regardless of whether a battery or power supply is used.
I'll check the voltages at my transistors when first switched on and after 10 minutes when I get time, and I'm not really asking anyone to walk me through troubleshooting this. I'm more interested in hearing ideas on what, generically, might cause this type of failure where it works correctly at first, then dies after it's warmed up for a while.

[Steve Conner]

Germanium transistors?

[Mark Hammer]

Typically, anything that takes a predictable period of time to manifest some difficulty is either dealing with heat or with bad caps (or cap-related charging up, or drifting/loss of bias point). Heat is much less likely in this instance, so check all your caps.

[52 Bill]

I have seen a diode react thermally and work like a slow pulsing on & off switch, but that rarely happens, especially in a low voltage circuit like a superfuzz.

Before and after voltage readings will probably show the source of the problem.

[Steve Conner]

NOS germanium transistors can be incredibly temperature sensitive. The reason is that the germanium transistors available nowadays are mostly the left-over rejects that were too leaky to use back in their heyday. The leakage current varies greatly with temperature, and therefore so does the circuit's DC bias point.

[Martin Beer]

The superfuzz uses Silicon transistors, (new 2N3904 in my clone) which should be less problematic. There are a pair of Ge clipping diodes in the circuit but the behaviour is the same with those removed, other than sounding less clipped. The schematic I used is here; http://pedalparts.co.uk/wp-content/u.../SuperFuzz.pdf
I have measured the voltages at the transistor legs and found some measurements from an original Superfuzz to compare them to. All voltages on mine are in the right ballpark except for Q1 collector and Q2 base. These are at around 6v in an original unit, but drop down to below 2v in mine. The behaviour is the same when the Q1 and Q2 are replaced, so those appear to be OK. I think I'll explore Mark's suggestion that a predictable time-based behaviour is likely capacitor related, and replace the caps associated with Q1 and Q2, and see what happens.

[Jazz P Bass]

I would suggest 'monitoring' the base voltage on transistor Q1.
If it is creeping high, then of coarse Q1 collector will drop.

[Martin Beer]

I've replaced the four capacitors which look like they could affect Q1/Q2 biasing (C2, C3, C8, C15). The circuit is now not passing signal at all, the time related-behaviour has now disappeared and the voltages at Q1/Q2 are:
Q1; C 0.5v B 1.0v E 0.37V
Q2; C 9.65v B 0.5v E 0.0v
In an original, working Superfuzz, voltages have been reported as
Q1; C 6.0 B 0.7 E 0.1
Q2 C 8.9 B 6.0 E 5.4

Q3 to Q6 in mine are within 10% of the values from an original, so I'm presuming the rest of the circuit is not the problem.
So something is obviously not right here. I've checked all the resistor values and can't see any unintentional shorts. I'm a bit stumped by this. I'm going to put it away for today, do some reading on how transistors are supposed to behave (which I'm a little shaky on being from a non EE background) and have another look at it tomorrow.

[Steve Conner]

At the risk of "teaching my granny to suck eggs": Are you sure the capacitors are good and you put them in with the correct polarity?

[Martin Beer]

Capacitors are untested but new and their polarity matches the schematic, but yes, it's possible there is some small detail in my assembly that I'm missing. I'm going to put it down for a while and come back to it fresh. Thanks for your help, everyone.

[nickb]

If Q2E is 0v and Q1B is 1V then you have a short. Checking the voltage at the top of C8 might give you a clue, look for solder splashes and clean any flux off (flux cleaner or isopropyl alcohol).

[g1]

At the risk of "teaching my granny to suck eggs":

Had to look that one up Steve. It said that saying is just a modern version of this: "A swyne to teach Minerua, was a prouerbe, for which we sai: Englyshe to teach our dame to spyne."

[Jazz P Bass]

Q1; C 0.5v B 1.0v E 0.37V
Q2; C 9.65v B 0.5v E 0.0v
.

Explain why Q1's Collector no longer has voltage.
It did before.

[Martin Beer]

Explain why Q1's Collector no longer has voltage.
It did before.

I've got it working reliably and measuring as expected now - I think that when I swapped the caps out, I had a short from one of the PCB pads to the ground plane, which was messing up the biasing, and was a bit of a red herring in identifiying the problem in the OP (which may or may not have been cap related - I don't know).
Thanks for you advice, everyone. It helps me to think a bit more clearly when working this stuff out!