[sdiy] SAW core VCO flyback time

[Richie Burnett]

> On a ramp wave, the pulses get wider in one direction only, whereas with a 
> triangle, they get wider in both directions symmetrically.
> In the ramp wave case, this means that there's a phase shift going on. And 
> a changing phase shift is a frequency modulation, so it warbles.
> At least, that's my understanding of it. If that's not right, perhaps 
> others can correct me.

That's right Tom.

Also, if you drive a flip-flop divider with the pulse waveform to implement 
a sub-oscillator then it's important that it triggers on whichever edge 
*isn't* the one being modulated.  Otherwise the sub oscillator, will suffer 
vibrato with a perceived wave-shape that is the differential of the PWM 
LFO's waveform.  So, for example, a triangle LFO sweeping the PWM, would 
result in the sub oscillator playing a square-wave trill between two pitches 
if it was generated from the modulated edge of the PWM waveform.

If you're generating your PWM from a perfect triangle waveform, then the 
resulting PWM waveform doesn't have any pitch modulation because it's 
leading and trailing edges are modulated equally and oppositely.  However, 
any sub-oscillator divider that triggers off either edge would suffer some 
pitch modulation.  I guess the easiest way to mitigate this would be to just 
drive the sub-osc divider from whatever logic in the tri-core that 
periodically switches the direction of integration.

-Richie,