About stability and feedback

In another discussion with Iain Churches, the subject of stability with
feedback came into the conversation.

The basic rules for stability with feedback applied are quite simple.
Namely that the feedback should not become 'in phase' with the input
(due to phase lag issues typically in the circuit in question) AND have
a value of greater than one.

Now, how does the LOAD affect this phase lag (or lead) ? In a typical
semiconductor based amplifier which I am of course rather more used to
working with, using an emitter follower output, quite simply it dosen't
very much since the ouput Z is inherently low compared to the load. So
this topology suffers relatively little from load related instability
problems.

However some circuit topologies place the load in a different place such
as in the collector circuit and this DOES require careful attention.

It occurs to me that this is analogous to the typical tube P-P output
stage. In short, this topology is especially 'vulnerable' to load
related instability since the reactance of the load itself can cause
huge changes in phase margin.

Now, an interesting way of dealing with the problem of load related
instability would be to do something analogous to the semiconductor
solution and use a cathode follower output stage.

It seems that this was done many years ago by EV and they called it the
'Circlotron'.
http://circlotron.tripod.com/

And now I see what PPP is !

Curiously, I don't see any mention of stability mentioned in the online
sources about this arrangement but maybe this wasn't the intended
benefit at the time ? Certainly it would seem that larger amounts of NFB
than is typical with ordinary P-P could be safely applied to such a
stage.


Graham