Fast VCOs for wavetable

[Martin Czech]

> 
> Then, you can rescale the VCO to make it go faster, chaning the
> capacitor C2 and the resistor R17 is just one way. The main reason the
> VCO will get stuck at high frequencies is the saturation of the
> exponential transistor Q2 which can not deliver much current above 10 mA.
> 
> One way to double the frequency of a sawtooth is to waveshape it. You
> can do this by consider this form of waveshaping.
> 
> VCO shape:
> 
>  /|        /|
> / |       / |
>   |      /  |
>   |     /   |
>   |    /    |
>   |   /     |
>   |  /      |
>   | /       | /
>   |/        |/
> 
> Comparator output for detection of half level, detector output reduced
> to half level amplitude:
> 
> ---    |-----
>   |    |    |
>   |    |    |
>   |    |    |
>   |-----    |--
> 
> VCO shape subracted with the Comparator output signal:
> 
>  /|    /    |
> / |   /|   /|
>   |  / |  / |
>   | /  | /  | /
>   |/   |/   |/
> 
> Amplify up with 2 and you got a full-swing double frequency ramp.
> 
> This method allows for recursive approach, thus you can cascade it
> into become 4, 8, 16, 32 and 64 time the frequency. There is an issue
> of speed of op-amps if you want to get higher speeds.
> 
> Note that when you casecade these, you can use the output of the
> comparators as the output of a counter ;)
> 
> I migth drop a schematic of such a waveshaper if people requests it.
> 
> Cheers,
> Magnus
> 
> 
Also, you could try to full-wave rectify a tri wave, this has no sharp
edges , so may be it is better at high frequencys. Normal op-amp designs
suffer from slew-tae bandwidth problems however, you won't get much about
some kHz. But a single transistor, or discrete differential amp design
might be able to deliver a good 2x frequency tri wave without too much
waveform distortion.

If one would be able to make a sine osc (state var with limited feedback)
one could try an analog multiplier (MC1496) approach. I have never tryed
this, but it seems to be reasonable.

m.c.