[sdiy] Triangle-to-saw conversion

[David G. Dixon]

> The lack of square wave can be remedied by setting up a differentiating
> circuit (integrator with R and C interchanged) and then put a comparator
> (against 0V) on that output to form a proper square. Then just continue
> as the general plan, i.e. a CMOS-switch controlled by the square to
> provide +1/-2 switched gain. Not too complex.

Doesn't the existence of a triangle wave always imply the existence of a
square wave (or, at least, a square-ish wave) to drive the oscillator?