build your own tempco resistors ??

[Joachim Verghese]

On Thu, 12 Dec 1996, Haible_Juergen#Tel2743 wrote:

> (1) A different expo converter topology! Instead of the usual
> differential pair with a current regulation loop around the
> transistor that serves as a reference, they keep the
> current constant by *separating* the pair with a voltage
> buffer! The idea looks nice; are there any drawbacks?

I guess both topologies are equally sensitive to op-amp input offset
voltage drift...I don't know if it's a drawback really, but in the
Transcendent design, Vcb of the exponentiating transistor is not
constant...I'm thinking of the Early effect.  And, of course, 1M into
-12V still isn't a perfect current source...but these result in very
small inaccuracies, so I wouldn't be surprised if both circuits 
work equally well in practise.

BTW, wasn't this what Pearlman presented in his patent form 1965?

-joachim