[sdiy] Imperfect VCO

[Theo]

Don't know the ION but a real 0:100 digital saw wave has a lot of aliasing
going on.
With a 0:100 relation there is no room for bandlimiting, with 2:98 there is.
Put simply, a steep (reset) slope needs higher frequency partials.
The steeper the slope the more partials are needed.
So if you reset time gets smaller more frequencies above 1/2 sample rate
will be contained in your digital saw.
A analogue oscillator would not suffer.

Theo


From: Oakley Sound <oakley at techrepairs.freeserve.co.uk>


> > To get something interesting you really have to draw large distortions
> of the original curve...
>
> But try this one. Alter the rise time vs fall time of a triangular
> waveform. The difference between a 'sawtooth' with a rise time to fall
> time ratio of 1:99 and 2:98 is very noticeable. I tried this with the
> Ion, and was suprised how 'digital' the 0:100 setting was and how much
> warmer the 2:98 setting was. I was hearing [I think] the fundamental
> becoming stronger. Perhaps a significant part of the 'tone' of the
> sawtooth output on VCOs could be controlled by the reset time. But
> thinking about it, is it that we describe a powerful VCO by the strength
> of the fundamental? A perfect sawtooth being that little bit too bright?
>
> I should do some basic fourier maths on sawtooth waveforms with
> differing flyback times.
>
> Also there is quite a difference in tone when there is a discontinuity
> in the falling ramp of the sawtooth. When setting up Polyfusion VCOs
> [triangle based core], you need to adjust the symmetry of the waveshaper
> to give you zero glitching on falling ramp. A small glitch as seen on
> the scope does give a different tone.
>
> Regards,
>
> Tony Allgood                       Penrith, Cumbria, England
>
> Oakley Sound Systems                   www.oakleysound.co.uk
> Modular Synthesisers      www.oakleysound.co.uk/projects.htm
>
>