[sdiy] Temperature Compensated VCO attempt - help?

[Magnus Danielson]

From: Scott Bernardi <sbernardi at attbi.com>
Subject: [sdiy] Temperature Compensated VCO attempt - help?
Date: Wed, 29 Jan 2003 13:39:54 -0800

Scott,

> I've made an attempt at an electronically temperature compensated
> exponential generator and VCO. I built it into a thru-zero FM VCO based
> on an Electronotes design. The VCO works, but the exponential generator
> suffers from linearity problems. I can tune it to a nice 1v/octave
> response using my VCO tuner
> (http://home.attbi.com/~sbernardi/elec/og2/vco_tuner.html) at 0v CV, but
> as soon as I deviate from that either higher or lower in frequency, the
> scaling is off.
> I've got a pretty detailed theory of operation and schematic at
> http://home.attbi.com/~sbernardi/elec/og2/tempcomp_vco_theory.htm.
> Jim (Patchell), I was especially interested in your temp compensated
> VCO, and notice you changed the circuit because the previous one was
> "bogus".  I believe on your previous circuit you were using a delta Vbe
> bandgap reference like mine - was that giving you problems? I also
> notice you are using a different configuration for your multiplier OTA
> that "linearizes" it.  I suspect it might be my OTA multiplier that's
> causing the problems.
> Anybody care to comment?

Well, yes...

I have also been thinking along the lines of Bandgap references and other
means to do semiconductor-based compensation.

One thing I do notice thought is that you don't mention the rBE, the bulk
resistance, which eat out voltage on VBE. This modifies your equation to
become

            I
      kT     c
V   = -- ln -- - r  I
 BE   q     I     BE c
             s

The main theory these days is to keep rBE low as hell, and money is tossed at
MAT-02s and its like. We've been discussing various means to compensate that
out and the addition of a resistor and a diode in the servo-loop makes minor
wonders, but doesn't really cut it that well (some of my points on that took
some time to land I can remember).

The classic offset trimmer use for LM13600 and it's like isn't a very good one.
It basically sucks, since you changes the input signal gain prior to hitting
the input stage when you are attempting to adjust the offset. Toss a pair of
resistors to ground and put a trimmer through a resistor to the positive input
and you have independent trimming. You will be happy you did it. I have still
to find someone who given VCA trimming the propper respect and handling!!!
If you are to cancel both input and control leakage, you get a little more
complex trimming setups, but once you've done that you can trim it by very
simple means.

Cheers,
Magnus