> There are many potential gotchas running the PIC at low voltages,
> these are a few that I have found:

Today I got given a couple of motor controller units to evaluate from
a large (in the 1000s) UK production run using the 12C671. I'm told
that a similar circuit has been made for many years (not with a 671
presumably) "without complaints". Apparently 10% of these units,
which are used for golf carts, historically "go funny". The biggest
fault is that despite having 12VDC available, the micro is allowed
a paltry 2.2V via a 2.7V zener and there are no filter caps. No
wonder they "go funny"

Many other things are out-of-spec or very close to it. For example
a 25k pot on the ADC, and half the pot's travel is wasted putting
out too-low duty cycle PWM (at an annoying 150Hz), the pot has
to be hard CCW to completely stop pulses going to the motor (no
off switch - no sleep), no reverse polarity protection, one of the 3
planned-for FETs is out to make room for a thick motor wire. etc
etc etc. And for some reason it shuts the motor down after 10
minutes. It's probably the crappiest piece of junk I've seen for a
long time and it's surprising any of them work at all. Looks like I
might have my next job ;-)))))

(just a wee post prompted by incredulity at both the shoddy work-
personship and price)

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.