VCDO

[Buck Buchanan]

>That remainds me about my idea for vco bank to be implemented in my future polysynth.
>Each voice employs few PLLs driven by a single DCO (a programmable frequency divider).
>Each PLL contains its own programmable divider and frequency shifter (e.g. the one
>designed by Scott Gravenhorst) within the loop. The loop filter is voltage controlled 
>
>This structure is aimed at perfect longterm stability of the VCO frequency ratios, yet 
>it allows the short-term microdrifts to achieve that organic feel of analogue oscilators. 
>I have already experimented with 4046. Yes, that's true the useful range is 
>too limited to provide full tracking over 10 octaves, but my idea is 
>to insert a programable prescaler:

>I think the best results might be achievable using a typical widerange synth VCO,
>but I also consider 8038. Another point is if exp VCO is suitable here (or even 
>better than linear). Did anybody tried anything similar ?

Hi Mark and all,

First off, I'd like to say this sounds like a really neat idea.  I expecially like the voltage controlled loop filter.  But I do have a couple of questions here:

1) Given the low frequencies feeding the phase detector, is it even possible to get BOTH accurate loop tracking AND fast tracking?  I would think you might have to sacrafice one or the other.

2) Initially, I would vote for the linear VCO so that loop behavior would be fixed over frequency.  However, is it possible that an exp VCO could allow a wider loop bandwidth?.... This might cause phase noise/jitter to increase with frequency - maintaining the same percentage of frequency error per octave across the band (instead of improving as with a linear VCO).  Could this buy a wider loop bandwidth for better tracking in the bottom end?  Combined with some kind of dynamic loop control (with the VC loop filter!) this could rip!

Ok, now I'm guilty of both getting a little over my head and mental masturbation.  ;-)  

I'd love to know how this turns out!  Good luck!

Buck

-