[sdiy] How DCOs work

[Tom Wiltshire]

On 13 Oct 2016, at 00:28, David G Dixon <dixon at mail.ubc.ca> wrote:

>> There was(is) a mono synth that actually did this - Fredrik 
>> Carlqvist's Minod Vorga:
>> 
>> http://www.minod.com/products/minodvorga/vorga.html
>> 
>> There may be others, I don't know. If I was building a DCO 
>> synth, I'd certainly include it.
>> 
>> It's very simple to do, and I once experimented with the idea 
>> on the breadboard. You can set the ramp core up with a 
>> comparator that resets the integrator like normal, and also 
>> feed in a reset pulse form the counter. Setting the control 
>> current correctly gives you the normal ramp wave, reset by 
>> the counter reset pulse. Pushing the control current higher 
>> forces additional resets from the comparator and gives you 
>> sync sounds.
> 
> Right, so if I understand you correctly: You basically build a bog-standard
> sawcore VCO, but you have a digitally controlled reset pulse as well which
> ensures "perfect" tracking.  Presumably, if the tracking is poor, the
> digital reset pulse will occur slightly before the comparator pulse (which
> is late because the tracking is off).  However, if you increase the
> integrator current, the integrator will still reset "normally" perhaps
> several times before the next digital pulse comes through.

Yes, exactly.

> What if the comparator and digital reset pulses come at exactly the same
> time?  I guess it wouldn't really matter, except that you may be a slightly
> longer-than-normal interval at 0V.

Yes, that's what happened on my breadboard. I had the two reset pulses connected together with a diode OR. As you say, at the point that the sync effect goes just pst 100% (or 200%, 300% etc) the comparator resets and then the counter resets, so it's possible for a "double length" reset to occur.
In practice, it didn't sound bad, and in the middle of a sync bend, you don't notice.

You don't necessarily have to do it like this. Fredrik used a comparator on the same PIC18 he used for the counter, so the OR of the two types of reset pulse was done in software. That gives you some more flexibility about how you handle the situation.

> By this philosophy, any VCO could be made into a DCO without too much fuss,
> no?

I suppose so!

Tom