CEM Oscillator woes

[klosmon]

Decided to unearth & recommission my old Curtis CEM 3340 vcos -- have two
on Paia experimenter cards and one on a homemade circuit from POLYPHONY --
the weird thing is that they ALL have the same problem:
altering the pulse width has an effect on the pitch.  It is the same effect
on all three waveforms;  measuring the control voltage going into the chip,
I see a change of about .06 volts moving the pulse width control from one
extreme to another.  this same effect occurs using external pwm voltages.
I've checked all I can think of to check (on all three units);  I've
switched power supplies.

This is all news to me.  I'd conclude that it's a systemic problem with the
CEM chips, but they work just fine in my Pro-One.
Anybody (who's still around, and not carousing for the three-day weekend)
ever encounter this?