SV: [sdiy] QVCO schemo up

[harrybissell]

OK...ok

let me say that I have never seen an LM13600/13700 that did not
have huge offset problems when I activate the diode bias

Don't mean it can't happen... ;^)

I'd add that I usually use equal value bias resistors... then inject a small
offset voltage through a high value (megohms) resistor.

I don't usually use the trimpot as Ian showed... (had bad luck with that
once).
Is that method any better at trimming the offset ???

In my Spice sim, I trimmed only one OTA... and only a couple percent...

H^) harry

Ian Fritz wrote:

> At 03:53 PM 11/11/05, Harry Bissell Jr wrote:
> >Using  the linearizing diodes ~guarantees~
> >that you will have big offset troubles. The simulation
> >can be trimmed easily to remove them.
>
> I wouldn't say "guarantees".  It can happen if the input transistors are
> well matched but the diodes are not.  Makes selecting really good devices
> pretty difficult, though.   When I was working on discrete OTAs I found
> that each pair has to be matched to better than 50 uV.  Then the offsets
> are small and can be trimmed out and the the bias dependence can be quite
> small.  A good idea to use selected 3280's.
>
>    Ian