[sdiy] Buchla 258/Plan B M15 expo converter

[René Schmitz]

Tim Stinchcombe schrieb:

> If it were an op amp instead, I (more or less) don't have difficulty
> believing that the op amp can 'mop up' the 'excess current'. For this
> transistor case, it is quantifying all the gains etc. around the 'servo
> loop' that I want to see, in order to convince myself that it *is* also
> capable of doing the same thing!

Well the loop gain is very low compared to an opamp, and also highly 
nonlinear, I'd suppose. There is no emitter degeneration i.e. current 
feedback (neglecting D1's internal resistance) on Q1, so Q1s gain is 
Rc/re. re being the intrinisic emitter resistance re=Vt/Ic = 25(ohms per 
mA). So the finite loop gain has to be taken into account, together with 
the nonlinearity.

> However, the fog may be lifting with respect to these standard feedback
> configurations - I'm going to work through the simple emitter-follower case
> to see what I can learn, and so see if it is extendable in some way to
> putting such a follower in the feedback loop, which looks to be near enough
> what these setups are doing.

Hard to say without simulations, but you might get some nonlinear 
behaviour, if the transistor is steered into saturation. But I doubt it 
is the case.

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159