Debugging Zonk Machine build

[Reg18]

I got fed up with my tchula build yesterday and decided to build a Zonk Machine mk2 fuzz.
Pic added, I am getting a tiny amount of signal through when engaged, but nothing that sounds like fuzz.
8.7v across battery, with a 10k bias resistor I was getting 1v at collector of transistor so I added a 20k trim pot instead, bias it to 5v ish however I get the exact same reading on the other transistor collector. If I adjust the trim pot it adjusts both collector voltages weirdly.
Still no working pedal either way however.
Any suggestions on next debugging idea?
I haven't done many pnp circuits so I was going to bypass the footswitch just to make sure my wiring wasn't at fault.

[antonis]

Can't see your wiring clearly enough but I suppose you implement positive ground power supply..

[Reg18]

Can see your wiring clearly enough but I suppose you implement positive ground power supply..

Yes black wire from battery snap is feeding power to the circuit.
I feel like it may be something simple like that though as I'm
Confident the parts are connected as shown on schematic but I'm
Not confident in my off board wiring for pnp

[antonis]

Not confident in my off board wiring for pnp

You have to refer to particular transistors pin-out..

[Rob Strand]

Check for shorts on the transistors.  Especially between b and c of Q2.

[Reg18]

Check for shorts on the transistors.  Especially between b and c of Q2.

I did find a fault, one of my grounds wasn't connected properly on the volume knob.
Now I am getting 8.4v at battery
Q1, C=0.3v,B=0v, E=0v
Q2, C=5.4v, B =0.3v, E=0.3v

Something is not right, not sure what

[antonis]

Q1 Base should stand about 600mV lower than Emitter (for pnp Si BJT)

Maybe 680k resistor isn't properly connected either on Q1 Collector or Q1 Base (or both)

[Reg18]

Q1 Base should stand about 600mV lower than Emitter (for pnp Si BJT)

Maybe 680k resistor isn't properly connected either on Q1 Collector or Q1 Base (or both)

Good suggestion, I took the 680k out and checked to make sure I had the right value, yep it was right.
Tried a few different values going right down to 50k and it raised the base a few points to 0.7 ish but the emitter still was on 0
I've swapped out both transistors incase one was blown, still no luck so far.
No sound at all now, back to the drawing board, should have bread boarded this one before putting it together

[Rob Strand]

I just bashed out some hand calculations assuming a transistor gain of hFE=200 and *silicon* transistors,

With the values show on the schematic, you should get something like,

VC1 = 1.12V
VB1 = 0.55V
VE1 = 0V

VC2  = 6.58V  (with emitter resistor 3k9 and 5k), 
             7.94V (with emitter resistor only 5k)
VB2 = 1.12V  ; same as VC1.
VE2 = 0.53V

A trimpot in place of the 10k resistor will change VC2 but shouldn't affect the other voltages.
For different gains you will see different VC1, VB2, VE2, VC2 voltages.

[Rob Strand]

Measure the resistor values with a multimeter.   I'm not sure they are correct.

[Reg18]

Measure the resistor values with a multimeter.   I'm not sure they are correct.

I've double checked the 680k, 47k and it's a 20k trim
In place of the 10k bias resistor. They all are within tolerance, think the 47k is actually a 54k and the 680k was within 10%. Are there others I should check?
I've just breadboarded the same circuit but with modern components and some other germainium transistors and it works ok. Followed the same schematic.
Oh and I've used germainium transistors for this build not silicon.

[Rob Strand]

Oh and I've used germainium transistors for this build not silicon.

Q1, C=0.3v,B=0v, E=0v
Q2, C=5.4v, B =0.3v, E=0.3v

OK that explains the 0.3V's.  The pedal is designed for silicons.

Changing to Germaniums could be the reason Q1, C=0.3V (VC1).  You need to make the 680k *larger* to increase VC1.
In fact you might be able to de-solder the 680k altogether.  Then VC1 & VB2 should rise-up.  If not the leakage of Q1 might be too high for the 47k collector resistor.

EDIT:
With the 3k9, VC1 might not rise much above 1.6V and without the 3k9 perhaps 2.6V; that's with the trimpot at 10k.    If you get that far it means Q1 is can turn off and maybe Q2 is OK.   Then find a value for the "680k" position that gives you about VC1= 0.9 to 1.1V (roughly).

[Reg18]

Oh and I've used germainium transistors for this build not silicon.

Q1, C=0.3v,B=0v, E=0v
Q2, C=5.4v, B =0.3v, E=0.3v

OK that explains the 0.3V's.  The pedal is designed for silicons.

Changing to Germaniums could be the reason Q1, C=0.3V (VC1).  You need to make the 680k *larger* to increase VC1.
In fact you might be able to de-solder the 680k altogether.  Then VC1 & VB2 should rise-up.  If not the leakage of Q1 might be too high for the 47k collector resistor.

EDIT:
With the 3k9, VC1 might not rise much above 1.6V and without the 3k9 perhaps 2.6V; that's with the trimpot at 10k.    If you get that far it means Q1 is can turn off and maybe Q2 is OK.   Then find a value for the "680k" position that gives you about VC1= 0.9 to 1.1V (roughly).

If I bypass the fuzz control by going straight to ground with the emitter will that still work?
Might eliminate that part of the circuit as the problem perhaps

[Rob Strand]

If I bypass the fuzz control by going straight to ground with the emitter will that still work?
Might eliminate that part of the circuit as the problem perhaps

Unlikely, you might get some fizzly/farty sounds to come though if you are lucky.
If the Q1 isn't blocking the signal then Q2 will most likely end-up blocking the signal.

A better plan might be to lift the base of Q2 from the collector of Q1 then try to get Q1 to bias with
a collector voltage around 1V.  It you can't get that far having Q2 in the picture is just going to make
life difficult to debug.

If you can do that then measuring the voltages on Q1 might be helpful.

[Reg18]

So I converted it to an npn silicon transistor Fuzz with exactly the same components and it works fine! Maybe it is a dodgy germainium transistor, might save the germainiums for a 3 transistor Zonk Machine fuzz which was designed for germanium and leave this one as a silicon for now. 

[Rob Strand]

So I converted it to an npn silicon transistor Fuzz with exactly the same components and it works fine! Maybe it is a dodgy germainium transistor, might save the germainiums for a 3 transistor Zonk Machine fuzz which was designed for germanium and leave this one as a silicon for now.

Excellent.   

Actually a three transistor Zonk Machine I or Tone bender MKI is really worth building.  You could even put a switch at the input to switch between the two cap values.

[Reg18]

Sounds alright too!

[Rob Strand]

Sounds alright too!

Sounds good.   Very useable tone. 

[anotherjim]

If your DMM has a diode test range, you can check any BJT transistor since they look like 2 diodes with the ends joined at the base. For PNP, the black meter probe goes on the base (cathodes) of the junction diodes then try touching the red probe to the collector and emitter (anodes of the diodes). The test reading is usually an "indication" of the forward voltage drop so 300-ish for germanium and 700-ish for silicon. The Emitter usually reads slightly higher than the collector. Reversing the probes should have no reading, if it does, it indicates a leaky junction. Similarly, test both ways with probes on the emitter and collector. There should be no reading unless there is a fault.
Note that faulty Germanium transistors that have at least one good junction diode make great clipping diodes - so don't throw them out!