Thanks for your reply.
I am using a 10.000 MHz ceramic resonator with an accuracy of +/- .5%
more or less I am trying to figure out how to do timing as accurately as
possible.  I know there is error because using my delay function I calculate
the delay I would need to make an LED blink at 1 Hz...(.5 seconds) using the
method I mentioned earlier:

10000000 cycles/sec * 1 inst./4 cycles = 2500000 inst./sec = 2500000 TMR0
Incs/sec

2500000  TMR0 incs/sec * 1 loop/X TMR0 incs = 2 loops/sec

X = 1250000 TMR0 incs = 5000 * 250 TMR0 incs =  .5 secs/loop

here is the code:

#include <16f84.h>
//================================
#pragma config |= 0x3FFF,WDTE=off,FOSC=HS
#pragma bit servo  @ PORTA.1
//================================

//do times*cycles TMR0 Increments.
void delay(long int times, int cycles)
{
        long int i;
        for(i = 0;i < times;i++)
        {
                TMR0 = 0;
                while(TMR0 < cycles);
        }
}
//=============================================================
void main(void)
{
        INTCON = 0x00;
        OPTION = 0x00;
        TRISA  = 0x00;
        while(1)                                //Create a 1 Hz pulse
        {
                servo = 0;                      //Turns LED on (PORTA.1 used to sink current)
                delay(5000,250);                //delays .5 sec
                sevro = 1;                      //Turns LED off)
                delay(5000,250);                //delays .5 sec
        }

}
//=============================================================


I could tell just by looking at the LED that it was not blinking at 1Hz.  It
was going much more slowly than that, closer to .5 Hz.  Tomorrow I will be
able to go down to the Electronics lab and use the HP O-Scopes to see
exactly what is going on.  But somewhere there is considerable error in my
ways.  The only thing that I can think of is.. that I should not set TMR0 =
0, because by the time I get there a certain about of cycles have already
occurred.  I would just like to know the best way to solve this problem.

Thanks again,
Samuel






> -----Original Message-----
>
>        First question is what is the accuracy of your resonator? You have
>imposed a requirement of 400ppm.
>All that aside at the frequency your are running at  the time that it takes
>to call the function is a very small percentage of the error.
>Can you give us more of an explanation, what is the error  that you are
>observing etc
> Sam