John Walshe wrote:
>
> Roman,
>     Let me see if I have this right. I adjust the period until I get bit
> change in the counter. However the pulse width will be constant (actually a
> very very slowly changing - to the extent of being constant) so effectively
> changing the period will only change duty cycle. Maybe I'm being obtuse but
> I can't see how this will improve the resolution if I'm running with a
> system clock of 1uS. Initially I thought you were suggesting a the use of
> dither to get the resolution, but as far as I remember dither will help
> resolve and additional 1/2LSB which will gain only 500uS. For the
> application I would really need 250nS resolution on the width of the pulse.
>
> I'll give what I was proposing a shot and I'll let you know how it gets on.


Hi John, this was your original idea am I
correct?:

>   I'm considering a situation where I would clock timer1 in external mode,
>driven by a copy of the Xtal output pin, in order to gain a bit of
>resolution when capturing the width of a pulse using the capture register.
>The Xtal frequency is 4Mhz, giving a pic clock of 1us which just a little
>too slow. I am hoping to have a resolution of 250nS by using the above
>method.

I'm not sure it can work. If you are clocking the
timer from the crystal itself (obviously with a
buffer) you will get 4MHz ok and the timer will
handle that speed. But at some point the PIC has
to read the timer, and I believe PICs do that on
a specific step every cycle. My datasheet says
external clock inputs are read during Q3.
That would always read the timer at count 3?

Measuring the period of a very slowly changing
pulse shouldn't be that hard to do at your
required resolution.

How long is the pulse? Could you use a faster
PIC? Maybe if you generate a random element that
adds a small random delay to the pulse start,
you could read the dither and get very fine
resolutions that way too.
-Roman

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