oscillation and negative/positive ground theory

[Foxpaw]

I have discovered, that there are several topics about fixing oscillating pnp/npn fuzzes with negative/positive grounds.

I would like to know why this oscillation happens? Is there a theory or a rule for it? What can "control" this type of oscillation?

Thank you for the answers.

[R.G.]

There are many theories, both proven and disproven about why this should happen in general. IMHO, there is no clear rule about why it happens, and no single prescription, however complex, for keeping it from happening except to use positive ground in certain circuits.

I have a lot of experience at taming oscillations and unwanted funnies like this. I have had instances where I could not cure this oscillation no matter what fixes I applied, except for reconversion back to positive ground.

My current thinking is that there are several things that interact to cause it.

Oscillation always requires feedback which becomes positive at some frequency. Negative feedback with phase shifts does this at frequencies where the total phase shift from input back to input is an integer multiple of 360 degrees, and positive feedback does it at all frequencies. You can get this a lot of ways: stray capacitances, ground impedance, power supply impedance, inductive coupling, leakage currents, lots of them.

If the cure was simple, you wouldn't be reading about it being a problem, would you?

[Foxpaw]

Switching to positive to negative ground means, that you simply change the grounds to +, and + to ground and reverse the battery.
Is this correct?

[R.G.]

You have to reverse connections to signal ground in the circuit as well. For instance pulldown resistors, pots to ground, etc. Otherwise, the impedance of the power supply and any junk on it is carried into these points.