> Does this mean that if I program them (OTP's) with a Picstart Plus some may
> not work across the entire 2.5-5.5V range?? I have a very small run that I'd
> rather not buy the PROMATE II parts. I am still not clear on the subject. Is
> it the erasing, the programming or both that is important to ensure
> operation over the full range??

On the current parts that Microchip is shipping, a memory cell which is
fully programmed will read "0" at any valid supply voltage.  A memory
cell which is fully erased will read "1" at any valid supply voltage.  A
memory sell which is "in between" may read zero at lower supply voltages
and "1" at higher voltages.

The most common place that this can cause a problem with with window parts
in low voltage devices.  If you try to program a part that hasn't been
in the eraser long enough, some of the bits may be read as "1" by a +5v-only
programmer, but may be read as "0" by the chip's internal log when VDD is
less than 5 volts.

In practice, none of this is going to matter when programming the OTP dev-
ices that Microchip is shipping today.  All such devices are thoroughly
blank when they leave the factory (if they weren't they'd be useless since
you couldn't remedy the situation!) so low-voltage operation won't pose a
problem; because the parts are hit with additional programming pulses even
after they readback as "1", high-voltage operation (5.1-6.0v) shouldn't
pose a problem either.

The one major caveat here is that Microchip could start using a new memory
design which did not behave this way and where multi-voltage verification
was necessary to ensure correct operation even on OTP's.  In practice,
though, I doubt that this will ever happen.