[sdiy] VCO - sine output - why bother?

[Tom Wiltshire]

I'd guess because it's hard to make the straight line waveforms from a sine.

With the standard cores, going from Ramp->Tri or Tri->Ramp isn't difficult. Getting Squares is dead simple as you've noted. The sine is the hardest, and that's not that hard, at least if you accept "reasonable audio purity" rather than wanting "test equipment quality".

But from a sine core, you can easily get a square wave only. Getting a ramp from a square can be done with an integrator, but throws up the same problem that the classic DCOs had - the amplitude depends on the frequency. Twice as high means half as loud. Then you have to compensate. That's not *too* difficult in a DCO where everything is controlled by a uP anyway, but it's harder in a analog VCO. Still, it could be done by feeding the Freq CV to a exponential VCA that boosted the gain by 6dB for each octave. It's at this point that I start to see any potential benefits disappearing into the far distance….;)

Perhaps there's a better way to do Sine->Ramp or Sine->Triangle I haven't spotted that would make this make more sense. Could you use a Tri->Sine OTA or Diff-pair shaper in reverse to convert the Sine directly to triangle? (or otherwise bodge the transfer function you'd need).

Tom
 

On 26 Aug 2016, at 12:36, Quincas <quincas at gmail.com> wrote:

> I have another question about the sinewave issue (btw I love my sines, use them all the time, but only the filters in oscillation give me perfect sines)
> Is there not a way to make a sine core oscillator and derive the other waves from it? Like a filter with fixed high resonance for the sine, a comparator for the square etc...
> Why doesn't anybody do that?
> 
> Quincas Moreira
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20160826/c0982030/attachment.htm>