measure the width of a pulse

[lgeorge123]

I use 16f877 with 32768Hz crystal to measure a pulse width.

the pulse is as follows :

high to low transition at A, after 9 ms there is a low to high transition B.,
so there is a length of time 9ms btn AB. the code is as follows but no success !!!

#include <16F877.h>
#fuses lp,NOWDT,NOPROTECT,NOLVP
#use delay( clock=32768)
#include "c:\lcd.c"
int count;
long time,time1,time2;
#int_ccp1
void time_isr()
{
time1=get_timer1();
count++;
setup_ccp1(ccp_capture_re);
clear_interrupt(int_ccp1);
}
void main()
{ count=0;
time=time1;
lcd_init();
setup_ccp1(CCP_CAPTURE_FE); // Configure CCP2 to capture fall
setup_timer_1(T1_INTERNAL); // Start timer 1
enable_interrupts(INT_CCP1);
enable_interrupts(GLOBAL);
while(1)
{
if (count==2)
{
time2=time1-time;
printf(lcd_putc,"%LX",time2); // time2/32768/4 is the time
}
}
}

external pulse is connect to ccp1. Can anyone give me a hand ????
thanks very much !!!!!

[PCM programmer]

CCS has an example file, though it uses both CCP1 and CCP2.
Here's the file location:

[Bill Boucher]

I can see a few problems...

Aside from the obvious that this program will only capture 1 pulse properly because it never sets itself back to look for falling edges... but that it will respond to each subsequent rising edge and print a value for every time count hits 2 again... assuming that pulses were to continue arriving on infinitum...

1/ var time1 is not initialized with a value on startup but var time is assigned the value of time1 and so var time remains unknown.

2/ var time never gets updated with timer1 value so there is no reference for the final calculation. The result will be random because timer1 is free running and var time contains a fixed value and so provides no reference to calc the value stored in var time2.

3/ the var time1 gets the value of timer1 twice, once at the leading edge and again at the trailing edge so time1's first value gets overwritten by the second.

In your main(), add this: