[sdiy] filtering out a saw-reset glitch in a triangle shaper

[David G. Dixon]

> Well, what you could do, beyond filtering is something I haven't seen
> anyone attempt... blend in just the right amount of the reset-pulse from
> the oscillator reset-cursuit. The signal is there if you only care to
> think of it as an alternative waveform. Purhaps you better lowpass
> filter it for better shaping.

Sorry, Magnus, but I think that this scheme is unworkable.  The reset pulse
from the comparator is actually a fairly broad trapezoid, and the glitch is
just a little bitty spike.  To make the reset pulse more spikey requires
cutting the positive feedback capacitor back to practically nothing (5pF or
so) which has other effects, and even then it is still very difficult to add
the two signals in a way which they cancel each other out.

I'm actually liking the active full-wave rectifier plus 5.1V zener method
more and more.  Slap a 15pF cap across the feedback loop of the final level
shifting and inverting opamp, and the final -5 to +5V triangle looks
virtually perfect.