At 09:37 AM 9/10/97 +0100, you wrote:
>I'd like to use one of the CCP modules on the 16C74A to decode a frame
>of pwm data (not unlike R/C equipment, but with 16 pwm pulses in the
>frame)
>
>To capture the mark/space ratio, I need to capture on the leading and
>the trailing edge of each pulse.
>
>I thought of changing the CCPCON register as soon as the first positive
>edge had caused an interrupt, so that it'd capture again on the trailing
>edge.
>
>The data book says that doing this can cause an unwanted interrupt - is
>it enough to clear out the CCP1IF bit again before returning from the
>first interrupt?

Andy Warren's "yes" is correct.

I do it this way myself in my robotic controller (which, btw, does 16
channels in/out too, at 1uS resolution on 4 MHz).

Basic flow:

in the ISR:

        if CCP1IF
                record captured width
                set CCP1 to other edge
                CCP1IF = 0


To save time later, I test to see if the measured edge was the rising or
falling edge.  If rising edge, I clear TMR1 before continuing so that I end
up with a 0~2500uS value in my array of widths, and on falling edges just
make sure the high period was valid.  I process (validate, modify, etc) the
array of widths in the foreground.

Andy

==================================================================
Andy Kunz - Montana Design - 409 S 6th St - Phillipsburg, NJ 08865
          Hardware & Software for Industry & R/C Hobbies
        "Go fast, turn right, and keep the wet side down!"
==================================================================