Need help with weird-behaving self-built Fuzz Face

[Speeddemon]

I decided to build a Fuzz Face clone, to see what all the fuss (no pun!) was about, as I'm not a fuzz user myself.
I used the directions from Phillip's Fuzzcentral stuff, but I wanted it to be NPN, so I switched the polarity of the electrolytics and the -9V did become +9V.

I used a 1k lineair pot for Fuzz, 500k log for Volume,
2x AC127 NPN Ge transistors and kept the rest of the stuff original (2.2uF and 22uF electrolytics, resistors 33k, 100k, 470 Ohm and 8k2).
I did bias one of the AC127's, by raising the 8k2 to 12k. This gave 4.69V across the collector/+9V (whereas it was 3.97V with the 8k2)

So far so good right?
Not really. While the fuzz works as it should (turn fuzz clockwise=more fuzz, turn volume clockwise=more volume), there are 2 very strange things happening:
1) As the Fuzz control is being increased, the volume drops. And once it's past '2 o' clock', the volume DIVES way below unity gain. Yes, there IS more fuzz happening, but you can hardly hear it, even with the volume cranked.

2) When I roll back the tone knob of my guitar, I expected a more round singing fuzz sound. But it become brighter/clearer and more fuzzy. 

Does anyone have any clue what might be wrong here?
I double checked all the wires/connections already (although I did build it on perfboard) and I even swapped the '1' and '3' wires of the Fuzz-pot, which just reversed the problem.

[Pedal love]

Do a resistance check on the 1k pot. Does it transistion normally? Are there any changes drastically when you move any of the lugs under test? Let me know the results.pl

[Stevo]

Sounds like the pot is not working right!!!  You can check this by simulating the pot with a 1 K resistor.....hooked up as if where maxed to full fuzz and see what kind of volume you get than...that is what I would do!!!
  The pot may have some solder in it that is common to happen......

[Speeddemon]

Are there any changes drastically when you move any of the lugs under test? Let me know the results.pl

What do you mean with 'move any of the lugs'? I mean, the pot is screwed tightly in its hole and none of the lugs are close to the enclosure or anything.
It's an Alpha pot btw, but since it's almost 2am here, I'm going to have to wait 'till tomorrow to change it.

By the way, is it normal for a Ge Fuzz Face to change 'face' drastically when you use your 5-way switch in a HSH equipped guitar?
The fuzz becomes more prominent with singlecoils it seems.

[Pedal love]

I would substitute the 1k pot for another 1k pot for it is obvious to me that this pot or the bypass 22uf capacitors, malfunctioning or not connected somewhere properly.pl

[Pedal love]

Check the 1k pot, check the 22uf capacitor, make sure everything is connected together correctly and try it again.pl

[Speeddemon]

Check the 1k pot, check the 22uf capacitor, make sure everything is connected together correctly and try it again.pl

I checked the 1k pot, and it behaves totally normal (down from 0.3 Ohm up to 1.028k Ohm)
How to check the big cap then? My DMM only has 2 modes for capacitors, one is up to 4nF and the other up to 400nF. 

Does the ECB order change when converting a PNP Fuzzface to a NPN perhaps?

[Joecool85]

When changing from NPN to PNP you have to go from negative ground to positive, and turn all caps around...thats really it though.  (AFAIK)

[R.G.]

Pedal Love's got it. Likeliest cause for what you describe is first a shorted or reversed polarity 22uF cap, then a bum or wrong-value 1K pot. Or both. There are others, but check those first.

[petemoore]

http://www.generalguitargadgets.com/diagrams/fuzzface_sc_nn.gif
  Here's a great place to start, or choose another correct one for your transistors.

[Speeddemon]

I swapped the 22uF electrolytic for a 15uF one, but the problem remained.
I got so fed up with this (soldering on perfboard is a bitch), I socketed the electrolytic.
Went back and forth between the 2, but the weird problem remains.

I'm really out of ideas.  I checked all the connections. They all seem good (because the fuzz also works normally, apart from the drop in volume, when the fuzz control is increased.
I checked it with 9V PSU and battery; problem remains.

These are the moments I really hate electronics/DIY. If only it WAS a bad pot or a bad cap, then I could blame the parts.
Oh, I previously checked with different transistors as well (BC109C, N3904 and now AC127 and AC141). The problem remains. 

[R.G.]

OK, not the obvious problem then.

What do the collector voltages do as you turn the pot?

[Speeddemon]

OK, not the obvious problem then.

What do the collector voltages do as you turn the pot?

From which transistor? (I have to measure between Collector and +9V right?)

[R.G.]

Both collectors, measure between collectors and ground.

[cd]

Aren't you Mr. Hot Wire Audio man?  Why are you asking for help in a DIY forum?

Even if you've double checked that pot, replace it if you haven't already.  I had a similar problem with a 2k pot (in a Fuzz Face, no less), checked out fine with a meter but something was funny with it.  One of the lug traces was bad/intermittent.

[Speeddemon]

Aren't you Mr. Hot Wire Audio man?  Why are you asking for help in a DIY forum?

Yes I am. And perhaps because there are tons of more knowledgeable people on this board.
Like I said, this is my first endavour with a DIY fuzz (and transistors for that matter, since the pedals I built before all work with op-amps).

But if you happen to have a problem with me posting here, I'll be happy to hear it. 

[Speeddemon]

Both collectors, measure between collectors and ground.

O.k., just measured them. I dunno if the behaviour is strange or not, but these are the results:
Fuzz knob @ 0%. Q1 C = 0,438V; Q2 C= 4,51V

Fuzz knob @ 50%. Q1 C = 0,438V; Q2 C= 4,51V

Fuzz knob @ 100%. Q1 C = 0,438V; Q2 C= 4,51V


BUT... While I was measuring, steadily the Q2 C voltage was rising. At the start of the test it was around 4,47V, and at the end it was around 4,88V (!!!). And the Q1 C voltage was slowly dropping (from 0,438 down to 0,423V).

So, while the Fuzz knob didn't change the readings, the readings themselves were slowly changing. (like 0,01V rise each 3 or 4 seconds).

[Speeddemon]

Just checked again. Now the Q2 voltage was all the way up to 6,7V and DROPPING, and the Q1 C voltage was down to 0.267V and RISING! 
A bad cap somewhere or something?! 

[R.G.]

A bad cap somewhere or something?!

Affirmative. Check all the electros. A leaky or reversed input cap might do something like that.

[Speeddemon]

R.G., I just replaced BOTH the electrolytics AND the transistors (N3904 Si instead of the NPN Ge AC141's).
The problem remains. The voltage ramps up and down slowly.