[sdiy] Favourite VCO designs

[cheater cheater]

On Wed, Feb 24, 2010 at 09:42, David G. Dixon <dixon at interchange.ubc.ca> wrote:
>> I believe David Dixon is working on a VCO that incorporates the SSM
>> based exponential converter. Correct me if I'm wrong, David?
>>
>> If it's finished it could turn out to be my favorite VCO.
>> The exponential converter design makes VCOs based on it very very stable.
>
> Thanks for the mention, Damian!

Always happy to refer to good work.

> I just finished emailing Stewart privately
> to extol the virtues of said 2164 Expo VCO.

Forward it to me - I wonder what you talk about there.

> I've built four of them now
> (all sawcore -- two for me and two for a friend), and they are very nice.
> They never seem to drift or go out of tune, and are in tune from the moment
> they are switched on.  I personally will never build any other VCO -- I like
> them that much.  Of course, I really don't have anything else to compare
> them to except my (slightly modified and expanded) Thomas Henry VCO-1s,
> which are also very nice, but require tempcos and are not very stable.
>
> One of these days, I will do a proper layout and have boards made, but for
> now, if you want one you either have to roll yer own PCB or have me send you
> a board.  I can even put together a board kit, if anyone wants one.
>
>


> I'm also working on a tricore version which is functionally identical, but
> which will have a distinctly different, much more mellow sync character.

For a moment there I thought you meant tri-core as in triple core; so
that got me thinking about the trapez core.

Is the sync you'll have there soft-sync?

Have you considered a multi-core version based on my idea that
'hanging up' the oscillator is faster than waiting for the integrator
cap to empty? It could yield improved performance in the highest
ranges. The sync would, I think, sound the same as a normal.

Also: have you tried your oscillators doing FM? How do they work for
that? Can you do FM feedback from an oscillator to itself? Can you do
FM feedback from a slave oscillator to its master? I think a very
stable osc like this could work well for FM.

Finally: I suggest adding a simple mod that could make a lot of
difference, which is a voltage mixer into the comparator reference
input. I seem to remember the design was going to use the REF02 for
the expo, but I'm not sure anymore - is the same reference used for
the integrator? Either way, you could add controlled drift to the VCO,
by mixing in some noise or generally external CV into the integrator
reference, so that the comparator triggers sooner or later. I do think
the vco could sound great this way, and if the signal is symmetric
(for example, noise) then it could drift around and be uneven without
actually going out of tune permanently -- meaning a lot of jitter, but
on average, over an infinite period of time, the error in pitch would
be 0. I strongly believe that 'overperformance' of the comparator
circuit is one thing that makes the modern analogs sound too strict.
One question which is open is, how would a voltage mixer impair the
performance of the voltage reference. But I think there might be a
simple way to have the REF02 just go through two (parallel) resistors
when the CV is not patched in.

Cheers
D.