Re: [sdiy] diode war

[jhaible at debitel.net]

about output impedance: You normally have a large resistor
across the diodes. This sets a (rather high) initial gain, which is
then reduced as th ediodes conduct. No problems with output
resistance here - th eloop is always closed.

About the gain-bandwidth-product: This is a small signal parameter,
and certainly doesn't fit for a distortion circuit once the diodes
conduct. I'd say you have to choose a time domain description for
analyzing this case.
For very low levels, before the diodes conduct, the high closed loop
gain would certainly mean that an opamp with low GBP would
cause some HF rolloff and/or distortion. (Which might be wanted or
unwanted.)

Many people claum there is a difference in sound for certain opamp/diode
based distortion circuits, depending on the opamp that is used.
I'm ready to believe this, but I have not analyzed it. Here are some
thoughts nevertheless:

(1) If the opamp is driven so hard (and if the circuit allows it), that
the opamp gets into saturation when the output approaches the supply
voltage, or if it goes into slew limiting, then a *very* different sound
from different opamp types is obvious.

(2) If the opamp is driven hard, and a diode bounding circuit prevents
the output from case (1) - this should be most diode clippers! - then
I expect no big difference from opamp to opamp.

(3) At low levels (diodes not - or barely - conducting), the GBP of th eopamp
and the maximum closed loop gain (from resistor across the diodes) of the
circuit will have a certain influence.

Does this make sense?

JH.