Or one could suggest that he could get fairly close to exactly 5ms by
investing in an atomic clock ;-)

----- From: "Andrew Warren" <aiw@
> On 31 Aug 2004 at 8:20, Olin Lathrop wrote:
>
> > Specifying an analog quantity as "exactly" is useless, and immediately
> > tells us that the person doing the speccing doesn't know what he's
> > doing.
>
>     Or maybe it just tells us that we're holding that person to an
>     unreasonably (or, at least, unusually) high standard of precision in
>     his usage of the language.
>
> > Since 5mS can't be achieved "exactly", we can only ignore it and take
> > our own guess about how close is close enough.
>
>     Right.  When someone asks for "exactly five milliseconds", but
>     doesn't mention anything about the tolerance of his oscillator or
>     about theoretical physics, MY guess -- and, I expect, yours -- is
>     that he means "exactly five milliseconds, assuming that my
>     oscillator is running exactly at its nominal frequency".
>
>     It's like someone asking, "How can I get my PIC to just sit and wait
>     forever for a pin to go high?"  I don't think he wants me to answer
>     that by explaining how the Second Law of Thermodynamics makes
>     "forever" impossible; he'd probably prefer that I just draw up a
>     BTFSS/GOTO loop and remind him to disable the watchdog timer.
>
> > Given that, what's the best way to teach someone to time 5mS on a PIC?
> > This is 1000s of instruction cycles even at a modest clock rate, so I
> > think a busy loop shouldn't be the first recommendation.
>
>     How likely is it that he needs to do something else while waiting
>     those five milliseconds?  He didn't SAY that he did...
>
>     If he's just waiting, a busy loop is by far the easiest and most
>     appropriate solution.
>
> > >         DELAY5K:
> > >             CALL    DELAY1K
> > >             CALL    DELAY1K
> > >             CALL    DELAY1K
> > >             CALL    DELAY1K
> > >         DELAY1K:
> > >             MOVLW   249
> > >         LOOP:
> > >             ADDLW   -1
> > >             SKPZ
> > >             BRA     LOOP
> > >             RETURN
> >
> > If you've decided you want a busy loop for whatever reason, this sort
> > of structure is reasonable.
>
>     Thanks.
>
> > However, getting the details just right can be very tricky.  Counting
> > cycles isn't as easy as it may seem, especially for a beginner.
>
>     True.  Good thing the PICLIST is here to help those beginners.
>
>     -Andy
>
> _______________________________________________
> http://www.piclist.com
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

_______________________________________________
http://www.piclist.com
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist