[sdiy] Re: VCO reset time for BBD-based PM

René Schmitz uzs159 at uni-bonn.de
Thu Jun 3 01:05:01 CEST 2004


Hi Jürgen, Ian, Harry and List,

> But once you want to clock a BBD for making a Physical Modelling oscillator,
> it looks pretty bad. With a BBD length of 512 stages, you're already in the
> 1ns
> range of reset time for the accuracy quoted above. Or use some compensation
> method and longer reset times.
> 
> Re-introducing Harry's dreaded BBDs into the thread (;->)
> But I'm serious about this application. So who makes a HF VCO that goes
> up into the MHz range and keeps the tracking error in the ns range ??
> HF compensation allowed, PPLs not allowed.

The fastest expo VCO I ever did is the one that is the clock VCO used in 
the three phase generator. It goes up to 500kHz. I guess with a faster 
gate you could push this a little further. Its no stablity or precision 
wonder, but I guess a better expo convertor as the two transistor 
circuit could help here. And I think that one can even make the cap 
somewhat smaller, without getting too close to the stray capacitances.

However, for precision work, I would use a schmitt trigger gate and a 
OTA, and make a _triangle_ oscillator. It doesn't have the reset time 
problem per se.

HF compensation can for example be introduced by Rossum 
(over-)compensation.
(or the variation that I posted some months ago. Grant Richter
posted a PNP variation shortly after that.)

However, I wouldn't rule out PLLs. You could use a "tracking" loop 
filter. (so that settling time does depend on pitch and not on the 
lowest frequency) and use a guiding VCO that runs some octaves above the 
normal range (so you get quicker locking). Then you can use a not so 
precise exponential VCO as the VCO inside the loop (you can then track 
its control voltage to the guiding VCO.) The loop is then not doing the 
entire work of steering the high frequency VCO, but only generating an 
error term which is dependant on the amount of mistracking. This should 
result in smaller settling transients.

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159




More information about the Synth-diy mailing list