Fast analog sw. as VC resistor

Mark Smart smart at nn.com
Tue Jan 7 17:21:01 CET 1997


> Mark - 
> 
> Thanks for doing the bench work for your experiment.  I'm closely following
> this thread and I wonder if maybe in your first trial you are debugging too
> much circuit.  By this I mean don't use any comparator in the test circuit,
> but rather use your function generator to produce a calibrated and
> predictable duty cycle. 

I first tried doing this, but my function generator won't do this properly.
There is a knob to adjust pulse width in microseconds, rather than % of 
cycle. If you increase the pulse width beyond the period of the oscillator,
it halves the frequency of the oscillator instantly! What possible use this
could be, I don't know, but it's very annoying. Sometimes even at one
frequency it's impossible to get it to sweep over the whole range from 5%
to 95%. Unfortunately, this is the only function generator available to
me. This is the reason why I built the comparator right off the bat.

> I wonder if maybe you are using your comparator
> control voltage as a true indication of duty cycle instead of actually
> measuring the duty cycle directly.  What I'm suggesting is that maybe the
> CV to PW convert is the culprit and not the switched resistor.  I would
> think it would be pretty hard to accurately measure duty cycle with
> anything but a really good measurement o-scope (but maybe you are?)
>

I was looking at the duty cycle of the comparator output on a scope, but not a
real fancy one (Tektronix 475). To get some idea of the output curve I put
the scope in x-y mode, and it drew the curve I drew in my last note as I
swept the pot up and down. The nonlinearity is severe enough that you can
see it obviously with this setup.

 
> Anyway I think the theory is right on, you just gotta tame the beast.
> 
> John Speth (johns at oei.com)
> Object Engeering, Inc.
> Vancouver, WA



************************************************
*     Mark Smart                               *
*     Network Technician                       *
*     University Communications Inc. (UCI)     *
*     smart at medusa.nn.com                      *
************************************************



More information about the Synth-diy mailing list