Fuzz Face negative ground problems explained?

[DuncanM]

OK, I've been thinking about this for a while and I think I understand why some people suffer from oscillations in theit neg. ground Fuzz Faces.

Here's the usual Positive Ground Circuit:



Notice that the guitar pickup connects between the base and emitter of Q1...

Now the Negative Ground circuit - I've redrawn this for clarity (I hope)...



Now the pickup is connected across the base and collector of Q1 - more about this in a minute.
SO, point C on the diagram is our output - current flowing through R4 and R5 produces our output voltage across R4.

Now, as the signal varies, so does the current through R4 & R5 which comes from the battery - and as we know, the battery has an internal resistance. So, by Ohm's law, the more current being drawn from the battery, the lower the voltage the battery produces.

Accordingly, the voltage at point B will vary with current flow.

Point A will be fixed at 0.1V lower than Point B due to the forward voltage of the b/e junction.

OK, the difference between the circuits is that, in the pos. ground version, the pickup is directly connected between the emitter and the base of Q1 and the emitter is grounded.
However, in the Neg. Ground version, variations in supply voltage will appear at the base of Q1 (with reference to ground through the pickup) and be amplified, which will cause varying current draw from the battery, which will cause supply votage to change, which appears at the base of Q1 due to the constant BE voltage and is amplified, which..............

Instant positive feedback loop, depending on how much gain is in the circuit.
No wonder it oscillates...
And it explains why sticking a big capacitor across the battery to nail down variations in the supply voltage cures it...

Now, I may have oversimplified this and others may wish to chip in, but I believe this is the core of the problem.

So, what do you all think.?

[RedHouse]

Seem to me folks (not me) mostly have trouble with converting to Negative Ground from a PNP topology.

[Pedal love]

I think you have the battery turned around, still the concept sounds interesting. pl

[R.G.]

If that was all that was going on, putting an appropriately sized cap or set of caps across the battery would kill the oscillation by making the power supply be a low impedance at the frequency of oscillation.

There are cases where this will not work. Sometimes it's fine, other cases where it's not enough.

[brett]

Hi
I'm no expert, but I agree that you've indicated one of the main problems:

Accordingly, the voltage at point B will vary with current flow.

Consequently, to solve the problem, the voltage at B needs to be as constant as possible and the impedance to the flow of any signal from B to ground needs to be quite low.

As long as the battery is good, B is very DC stiff, and AC impedance to ground is low (tens of ohms ?).  If you use a large cap (100uF or more), B is also DC stiff and AC impedance to ground is only a few ohms except at very low frequencies.  If you use an external supply (as environmetally freiendly guitar players do), then B is extremely DC stiff and AC impedance to ground is only an ohm or two because they usually have very large caps in them (usually 1000uF or more). 

For the mathematically minded: the AC voltage at B is the input signal times the Z of supply divided by the Z of the base-emitter junction of Q1* (a simple voltage divider).  Typically, the signal at B would have an amplitude of the input signal x a few ohms/20k.  So there's not much happeneing except with a flat or cheap battery and a small or no bypass cap.

PS the big input cap is a shocker for encouraging oscillations because the cutoff frequency for inputs to Q1 is super low.  At these frequencies (a fe Hz), bypass caps need to be hundreds rather than tens of uF to be effective. 

Thanks for your cool explanation and good artwork.
cheers

*Voltages at the emitter of Q2 can be ignored as far as oscillations are concerned, because they are out of phase with the input