Hi Olin,

> That doesn't make sense.  A 4MHz clock results in a 1MHz
> instruction rate,
> which must be divided by 64 to get 8 bit PWM resolution,
> which results in
> 15.6KHz PWM rate.  8 bit resolution is certainly enough for
> closed loop
> control of a Peltier cooler.  Actually, 6 bits is probably more than
> adequate.
>

Oops, silly me.  How did I ever drop that leading '1' in my calculations,
I'll never know ;)


> > Unfortunately, increasing crystal speed will require a
> > considerable amount of code revision to adjust timing,
> examples being
> > PWM registers, serial registers, bit-bashing clocking loops, and PID
> > timing loops.
>
> This should not be the case for well written code.  You
> should always have
> an assembly constant the specifies the clock speed, then
> everything else
> computed from that.  Take a look at my UART_BAUD and
> TIMER2_USEC macros in
> STD.INS.ASPIC at http://www.embedinc.com/pic for examples of
> how this can be
> done.  Hard coding divider values and other timing parameters
> is just plain
> bad practise.
>

Agreed, but this is a project "in the works", and did not start out as my
own code.   It is also large, over 6k of code space.  It does a lot more
than just PID control of Peltier's.  I will be going through the code and
pulling all of this stuff together once the basic implementation is
complete.  I am still using absolute mode, and not the linker, which does
not help ... but again, this is not originally my code so changing over from
absolute mode *will* require considerable revision.

> > However, I have almost finished a dual PIC PID controller design,
> > runnning at 20MHz, and using interrupt driven bit-bashing to
> > communicate between PICs, using one PIC as the PID controller, and
> > the other as the master. This setup works really well.

This one is completely my own design, done the way *I* like it.  It
implements all the points you have made out, and then some.

> >
> > PS: Olin - 1mF???  Isn't that just 1000uF?
>
> The "m" stands for "milli" meaning 1/1000 and the "u" for
> "micro" meaning
> 1/1000000.  In engineering notation, one tries to keep the
> mantissa part so
> that it has 1 to 3 digits left of the point.  While 1000uF
> does express the
> value, 1mF is the preferred way to write it.
>
> For some reason the standard engineering notation convention was not
> followed for capacitance way back when.  The milli and nano
> ranges were
> skipped so that everything was either micro or pico.  This
> led to cumbersome
> values like .01uF or 10000pF.  Lately this sillyness is being
> abandoned.  I
> make it a point to use proper engineering notation for
> values, including
> capacitance.

Try telling this to parts suppliers, electronics ans schematical texts, and,
well, everywhere really.  I know my engineering notation very well thank you
Olin, but I am definately not used to seeing 1000uF written as 1mF, so I
thought it might have been a typo on your behalf ;)  I can understand the
"non-standardness" with capacitance however, because the standard measure
for 1 Farad is very big.  This problem is not restricted to capacitance
either.  Inductance and the Henry share the same problem.

Thanks and regards
Ian

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body