[sdiy] How is the feedback loop of the exponential converter analysed?

[René Schmitz]

On 28.01.2020 18:53, David Manley wrote:
> Hi Spiro,
> 
> I haven't seen a reply to your question, so I'll just point you in the 
> general direction and hope that helps.  I'm sure one of the analog gurus 
> will step up and correct me.  :-)
> 
> An opamp can affect not only the gain of a signal it is processing, but 
> also its phase.  If the output phase is shifted over 180 degrees, that 
> signal is feed back to the input, and the opamp has gain at that 
> frequency then it will oscillate.  Look up 'gain and phase margin'.
What happens here is that the transistor provides gain. While we are 
mainly interested in the current typically (and the current gain is 
close to 1), if we break up the loop, at the - input of the opamp, and 
look at the collector as the output, we find a bona fide born-and-bread 
common base circuit, with a voltage amplification of roughly 
R(collector)/R(emitter).

That gain in the loop eats up the phase margin because the gain is in 
the order of 1000 here. An you need to overcompensate the opamp to keep 
the loop stable.

In practice there are also some effects due to the various capacitances, 
of the transistor but they will likely get reduced if the collector is 
kept at a constant potential.

Best,
  René

--
synth at schmitzbits.de
http://schmitzbits.de