Genome a écrit :
Here's something I'm hacking.

Unlike Jim I am scum.

However, what is it and am I wasting my time?


That's a multiplier, given that Ic(Q21) is your output.

One pb I see is that you want to current drive your Q7-Q8 and the Q9-Q10
logging pairs as well as the Q11-Q12 diff pair tail, which is posing
some pb for the collectors/bases nodes voltage that have to you fix
through R16-R17-R6-R9.

What I'd do is to set the Q7-Q8 and Q9-Q10 collectors voltage, which
will in turn establish the Q11-Q12 and Q13-Q14 diff pairs base voltages.
And also I'd directly bias Q21 with an emitter current source, instead
of using the R12/R13=2 to bias it at 100uA which will run the diff pairs
at different collector currents.



--
Thanks,
Fred.