[sdiy] Henry Walmsley's page

[David G. Dixon]

You see, this is why I love tuning in to [synth-diy].  Put any circuit up on
the web, and within half a day the guru's have redesigned it to be 200%
better than the original!


> >Walmsley's VCO current source is at the following link:
> >
> >http://homepage.ntlworld.com/henry01/waveform_synth/waveform02-02.pdf
> >
> >It is a standard matched-pair-and-tempco arrangement based on LM394.
> From
> >this one current source, he is tapping three separate currents for
> various
> >uses:
> 
> Four, actually -- AP1 also feeds the HFT circuit.  Does anyone understand
> the purpose of doing the HFT the way he does?  It seems to me to do the
> same thing as the standard Franco compensation, which could easily be done
> with a resistor in series with the timing cap.
> 
> 
> >"INT-CC" drives the interpolator, "HF-CC" drives the high-frequency
> >clock oscillator (this is basically the "core" of this oscillator, as it
> is
> >what defines the frequency of the VCO output), and "TRACK-CC" drives the
> >cutoff frequency of the tracking filter.
> >
> >My inclination would be to replace this current source with the 2164
> Expo,
> >but probably leave the rest of the circuit more or less as is.  Of
> course, I
> >can never leave well enough alone, so I'd probably simulate the whole
> mess
> >and go looking for tweaks, but based on Mr Walmsley's write-up, he seems
> to
> >have thought the circuit through exceedingly well, and the proof is in
> the
> >scope pictures.
> 
> Since the VCO core is just used as a clock, it seems to me that it would
> make *much* more sense to use a F-V chip (charge balance oscillator)
> rather
> than a switched integrator.  I used the LM331 for my DoubleDeka VCO, and
> it
> has phenomenal tracking up to well over 100kHz, and rock steady
> temperature
> compensation using the Pease scheme.
> 
> For the OTA-based VCO it will be difficult to get the core frequency over
> about 40kHz with decent tracking -- and even that will take quite a bit of
> effort. That will leave you with a max frequency of ~2.5kHz.
> 
> As far as needing an interpolator, well I guess that's a matter of
> taste.  I actually like the steps -- it gives the sound its own special
> character, and it's easy enough to filter down to a smoother waveform.  (A
> single pole filter changes a step into a ramp, no?)
> 
>    Ian