[sdiy] function generation

[Grant Richter]

> The Terry 
> Michael's (ASM-1) vco core is much more suitable for pitch
> generations.

Not to quibble with Jim, who is giving good advice about linear VCOs being
quite difficult to control exponentially.

As a note of historical interest. The EML 101 used uA726 exponential voltage
to voltage converters driving linear VCOs. While not preferred now, it can
be made to work, as some thousand EML 101s were produced and sold.

The problem is the tiny voltage differences at the low end of the scale
accentuate any offset or gain problems, along with temperature stability
issues and noise issues which are worse for the linear VCO design.

The approach might work OK for a microprocessor controlled system with a
high resolution DAC. The DAC could use a lookup table to correct for tiny
linearity flaws in the VCO. It would be a whole lot of extra work for only a
minor savings in cost.

John Simonton did pull it off OK in the PAIA Fatman using a low resolution
DAC, and some of his own brand of magic.