delay functions inaccurate

[bignick270]

Is it just me or is the delay functions inaccurate. My 200ms delay using the built in ccs function can be more or less than that value, it is different for each restart.

There are no interrupts running at the time these delays are going.

Is there a more accurate way to create delay functions?

[PCM programmer]

For a question like this, you need to post your PIC, compiler version and
test program that shows the problem. Explain how you measured the
pulse length. Did you use a logic analyzer ? Which one ? What settings
did you use on it. This is so we can duplicate your test environment.


We don't remember your PIC and compiler version from one thread to
the next. It might change. You need to tell us again in each new thread.

[bignick270]

my bad...

PIC: 16F886
8 Mhz clock

Compiler is 4.0.74

I have observed the code on a Digiview logic analyzer.

The function uses delays in between changing channel led's. The function is somewhat long so I will not post it on here but basically it goes like this

[asmallri]

[SherpaDoug]

I find the CCS delay functions to be pretty accurate and VERY repeatable, as long as there are no interrupts. Could your crystal be unstable?
_________________
The search for better is endless. Instead simply find very good and get the job done.

[bignick270]

it could be the problem, this chip is using the internal oscillator.

[PCM programmer]

This is why we want a test program. We would notice the INTRC_IO
fuse being used.

[ckielstra]

The internal oscillator for these chips is quiet accurate, 1% at 25C and 2% for 0C to 85C. A one second start up difference will not be caused by this.

Normally the CCS delay functions are quiet accurate and certainly will not cause a variable delay.
For such long delay variations at start up I'm more thinking of power supply problems resetting the CPU.

Like PCM Programmer said, post a test program demonstrating your problem. Now we have to do wild guessing.

[bignick270]

I have found a problem as I have also posted in my other topic dealing with the same chip.

I had an array length that according to the datasheet was still in good memory addresses but come to find out I was over running into reserve locations which was causing the PIC to restart. Once I changed the array spacing the delays seemed to be more constant.