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

[David G. Dixon]

I'm designing a triangle shaper for the 2164 expo VCO.  This is a new thing
for me, because all of my previous VCOs are triangle-core.  The basis is a
simple yet elegant precision full-wave rectifier circuit I found on the web
somewhere.  To mostly eliminate the saw-reset glitch at the apex of the
triangle, I'm thinking of filtering the rectifier output with a passive LPF
with a cutoff frequency of 50kHz (3.3k + 1nF).  That, plus a couple of
strategically placed 22p caps in the rectifier circuit itself, seems to all
but eliminate the glitch in my simulations.

So, here's my question: Is there some reason not to do this that I'm not
thinking of, or is there some other way to reduce or eliminate this glitch
which is blindingly obvious to everyone but me?