[sdiy] Additive VCO

[Grant Richter]

> And if you can achieve pure sinewave of fundamental frequency, you can use
> Tschebyshev polynominals to make hundreds (yeah, tell me I'm crazy) of
> harmonics. That'll require one multiplier like AD633 and other minor stuff
> like opamp for each harmonic.
> 

Mike Furman wrote a software program for the Mini-Wave to do this. You give
it a table of harmonic coefficients and it generates a non-linear transfer
function the will have those harmonic coefficients when fed a sine wave of
unit amplitude.

If you vary the sine amplitude from zero to one, you do get spectral
animation, but nothing to write home about.

Bank 14 of the Mini-Wave has all the harmonic partials (2-15) upper partials
are more noisy due to zippering. In THEORY, simply reading multiple pages of
the PROM and routing to seperate analog outputs, would give an additive
series of 15 partials. Hardware complexity increases rapidly.

The Buchla 148 used a series of waveshapers to bend a triangle into 10
harmonic partials. Full wave rectification (FWR) with amplitude correction
gives a frequency multiplication of 2. Non-linear transforms (NLT) were used
for the primes 3, 5 and 7. The partials were built by this method (IIRC):

2   FWR (x2)
3   3NLT (x3)
4   FWR->FWR (2x2)
5   5NLT (x5)
6   3NLT->FWR (3x2)
7   7NLT (x7)
8   FWR->FWR->FWR (2x2x2)
9   3NLT->3NLT (3x3)
10  5NLT->FWR (5x2)