Tri 2 Sin, plus Hal Chamberlin book plug

[Martin Czech]

> A few questions for all you synth-DIY gurus:  Could you create a lowpass
> filter module that does through-zero FM while resonating?  I know that I
> read in some Electronotes issue about modifications to the 4-pole
> lowpass filter (4 OTA+buffer lowpass stages with feedback around whole
> circuit) that allowed the sine wave to be synced; I think it involved
> using a FET in conjunction with the capacitor in the last lowpass stage
> that would drain the capacitor when triggered.  Now, it seems that
> through-zero FM involves a combination of oscillator sync and waveform
> reversal.  Anyone have any ideas?  It would be so cool to have a module
> that could be a filter, and a through-zero linear FM quadrature
> oscillator.

A no-Guru answer:
I know that a filter is an implementation of a differential equation,
this is also true if the filter oscillates (in above case 4th order).
If we assume that the filter still works reasonable linear (how else
could we get a sine from it) the following holds true: The state of an
n-th order linear system (ie. the four capacitors voltage or charge)
determine the future. Ie. if you set all integrators to some level and
then let go, the system behaves in a determined way and I can see no
reason why waveform reversal should not be possible, but it could turn
out that it is very complicated. Smells like measuring the current
state and computing the new initial out of it. 

OTOH it could well be that simple polarity switching at the integrator
inputs (some or all) is sufficient. That's the way how single
integrator osc. (tri osc.) switch direction for through zero fm. But
then stability has to be observed...

m.c.