BMP sustain pot affects collector voltage?

[nocentelli]

I was tinkering with a fuzz circuit on the breadboard, and ended up with the schematic below:



Q2 is a building block that seems to feature in many of Devi Ever's fuzzes, but i decided to put the emitter resistor and cap on Q1's emitter (I think I vaguely recalled Gus railing against grounded emitter "designs" on this forum in the past). I didn't bother for Q2, thinking it would just give a nastier fuzz with more harsh (I was quite comfortable with that possibility).

I was very puzzled when checking the collector voltages, because although both stages gave around 3v when I had tried them separately, with the BMP sustain pot now inserted between them, Q2 was giving 8.5ish volts. I ran through all the usual checks, resistors values, trying different transistors, swapping transistors over etc etc. For a while i thought I was losing my mind, because I tried both circuits out separately and they gave the 3ish volts i was expecting, but when I put it back together as per schematic, Q2 was now giving 3 volts too.... until i turned the sustain pot....

So it seems the sustain pot dials the Q2 collector voltage between three volts at minimum and 8.5v at max, despite the fact it has those DC-blocking caps. I tried different pots, different caps with no change. I then decided to use the same 1k resistor + cap arrangement from Q1 on Q2's emitter, and sure enough, the pot no longer affected the collector voltage.

This is a reliable + repeatable foible - I put a ground link to Q2e on a switch and the sustain pot definitely diddles with the bias if the emitter is shorted to ground, but not if it has the 1k resistor + cap there.

So, what's happening here? I had always thought it was about DC paths to ground or +9v, and that a cap would stop any path to ground affecting the bias. I feel like everything I thought I understood about transistor bias (not that much, tbh) has crumbled to dust.

[swamphorn]

Your circuit looks good to me in theory. How are you measuring the DC bias? If the circuit is oscillating or amplifying noise a voltmeter might only measure the average voltage. With the sustain pot at its maximum value the small-signal AC resistance of Q₂ will load Q₁'s collector resistor which would effect the resulting waveform and voltage swing and, thus, the average voltage.

Is the capacitor going to Q₁ grounded or floating? Does the behavior change when you add additional base-emitter capacitance to either transistor? Does the behavior change when you add additional power supply bypassing?

[nocentelli]

How are you measuring the DC bias?

Is the capacitor going to Q1 grounded or floating? Does the behavior change when you add additional base-emitter capacitance to either transistor? Does the behavior change when you add additional power supply bypassing?

I'm measuring voltage between collector and ground.
Q1 capacitor is grounded.
Haven't tried base-emitter capacitance - will try next.
Breadboard has a 100u electro and 100n ceramic across the rails

[MaxPower]

Lose the switch and use jumper wires to test because the pot shouldn't affect DC bias/voltages as drawn in the schematic.

[MaxPower]

Have you checked to see if you have any DC voltage across the pot and 1k resistor?

[anotherjim]

It could be oscillating or picking up noise. Ground the signal input - what happens then?

[nocentelli]

It could be oscillating or picking up noise. Ground the signal input - what happens then?

We have a winner. I started to suspect it might be related to the amount of gain from the Q2 because (contrary to what I posted last night) adding the 1k to Q2's emitter killed the effect, but adding the electro cap then brought it back.

So I'm guessing when I plug it in, I'll get whistling oscillation at the point on the sustain pot where the Q2 voltage starts to rise.

[anotherjim]

Any non-inverting circuit with enough gain can self oscillate. When you plug in, the source impedance of the guitar might stop the trouble by loading out the stray coupling from output to input. Even an anti-pop resistor for the input cap can be enough. I think it's worth fitting the antipop when breadboarding - even a 10M adds some protection from stray coupling. The circuit you have probably won't mind 1M.

To get a change in a DC reading, it also has to be clipping unequally. If the signal was perfectly symmetrical about the normal no-signal DC level, then a DC reading won't show much change, because the signal swings up and down by the same amount the average voltage is the same as no-signal due to the DC measurement doing the averaging. As you had the collector at 3v, the swing is greater up towards 9v than it is down towards 0v so the average level with signal will be higher. If you had biased it closer to 1/2 supply - you might not measure much change because the signal swings are more equal and the average is closer to the bias level.

When you box it up and plug in, you might not have the problem.

[nocentelli]

....I think it's worth fitting the antipop when breadboarding - even a 10M adds some protection from stray coupling.

It has a 2M2 pull-down that I forgot to include in the schematic.

Sounds OK, still tinkering

[MaxPower]

Interesting. Something to store in the old memory bank.