Tonebender Mk 3- nonsense voltages?

[Bricksnbeatles]

I've got a Tonebender MK3 that I built a few years ago on a Fuzzdpg board when I was first starting to, but never boxed up. I modded it to have an external bias on q3 for a greater range of tones by replacing R9 with a 10K pot in series with a 1k resistor.
I revisited it and I'm not making sense of the voltages I'm getting:

Q1
C -3.5V
B -1.59V
E -1.5V

Q2
C -3.5V
B -1.49V
E -1.4V


Q3 [Bias cw/100% (11K)]
C -101mV
B -60mV
E 0v

Q3 [Bias ccw/0% (1K)]

C -2.8V
B -68.8mV
E 0v

SO... Q1 and Q2 are within the typical expected bias range, but what has me baffled is that when Q3 has the bias fully CCW, so the r9 value is only 1K, I'm getting fairly typical mk3 Q3 voltages, other that the Base voltage which is around 1/10 of what it should be, but whatever. What has me confused is that with the pot fully clockwise so the R9 value is ~11k (just a hair over the typical 10k value), my voltages are absurdly tiny compared to what they should typically be, but the sound I'm getting is more along the lines of what is to be expected of a Mk3, while the tone with the bias pot fully CCW so the R9 value is at 1K is more what you might expect of a TB with a starved Q3. Pretty much, it functions mostly as you would expect (the range of the bias pot isn't what I was hoping for, so I might sub some stuff so I can get a wider range of voltages on Q3-- at the time I was trying to get more of a gated sound like when the balance is down low on a buzzaround), but the voltages make no sense. thoughts?
Am I correct in being confused by this, or does everything make sense and I'm just confusing myself by once again working on pedal stuff late at night when I should be catching up on sleep?

Schematic and sound clip (with timestamps for different fuzz and bias levels) below

https://soundcloud.com/user-598888574%2Fmk3-test

0:01- Bypass


0:15- Fuzz 100%; Tone 50%; Bias 100%

0:35- Fuzz 100% Tone 50%; Bias 0%

1:07- Fuzz 0%; Tone 50%; Bias 100%

1:24- Fuzz 0% Tone 50% Bias 0%

1:44 Fuzz 0%; Tone 50%; Bias sweep from 100 to 0

1:52 Fuzz 100% Tone 50%; Bias sweep from 100 to 0

2:16 Fuzz 0%; Tone 100%; Bias sweep from 100 to 0

2:27 Fuzz 100%; Tone 100% Bias sweep from 100 to 0

2:43 Bypass again

 

[fig]

What transistors are you using, and were their gains and leakages measured? I think Q1 & Q2 should be 60-70 hfe with lower leakage. Q3 should double hfe @120 and a higher leakage ~200-400uA

 

[Coda]

Whatever is going on, that is an excellent fuzz tone…though not quite your average MK3…

 

[Bricksnbeatles]

What transistors are you using, and were their gains and leakages measured? I think Q1 & Q2 should be 60-70 hfe with lower leakage. Q3 should double hfe @120 and a higher leakage ~200-400uA

A trio of ac128s (it was a matched Tonebender mk 3 set from small bear, circa 2015 I think)
Q1 and 2 are both around 60hfe with low leakage (don’t remember exactly what— I forgot to write it down and I don’t know where the test box I put together for the RG Keen method is).
Q3 is around 105hfe (a bit low, but still fairly acceptable) and ~250uA iirc. I’ll throw a keen tester together on the breadboard later to double check. Even still, they’re still all close enough to expected spec for the numbers I’m seeing to be a bit baffling. Not sure why the Q3 collector voltage would be just about right when the R9 value is way down to just 1k, but be just 5% of the expected voltage when r9 has a stock value of 10k yet sound closer to expected than when the Uc is at a normal reading.

I mean, hey, it works and sounds cool— no issue in that regard, but the numbers I’m seeing don’t at all match what I’m hearing. Not necessarily trying to “fix” anything, but rather understand why, what’s on paper, is so different from what’s sitting on my bench.

 

[fig]

Yeah those are spot on. TBH, I’ve never heard two identical fuzz circuits..but I haven’t auditioned them in that respect either. I’ve compared FF builds and found differences in bias resistance but figured that due to the performance of the Ge transistor when heated.

 

[Bricksnbeatles]

Yeah those are spot on. TBH, I’ve never heard two identical fuzz circuits..but I haven’t auditioned them in that respect either. I’ve compared FF builds and found differences in bias resistance but figured that due to the performance of the Ge transistor when heated.

Yeah, it's odd. for one thing, the diode in a mk3 is supposed to have to do with temperature stability or something like that from what I've heard, and besides, it currently like 70ºF in the room I've checked the values in, so the temperature shouldn't be messing with the bias anyway– not like I was directly handling the transistors-- they've been in the sockets on the board and I was just using problems leads to check, so my body temp wouldn't have warmed the transistors up.

 

[Chuck D. Bones]

Q3's collector voltage will change when you apply signal because C4 will charge up and alter Q3's bias. If you want a gated sound, then Q3 has to be either cutoff (Q3-C = -4.5V) or saturated (Q3-C < -50mV) with no signal.

Are you sure D2 is facing the right way?

To get to saturation with no signal, you need Q3 to have higher leakage or try a 1M pot from B to C on Q3.

I think that's a better way to fiddle the bias.