pic16f876a pin c3 coding bug when with real time interrupt

codeboy2016 · Joined: 15 Dec 2016 Posts: 2

Hi all, i wrote a program and use pcwh to compile and download program into pic16f876a. I found strange output on pin_C3. Maybe it is compiler bug.

Here is what i wrote:

Ttelmah · Joined: 11 Mar 2010 Posts: 19328

1) When asking questions like this always tell us your compiler version number. There are now something over 300 CCS versions. We have no hope of knowing if you may have a faulty version if you don't tell us what one you are using....
However this is not actually likely to be the problem unless you have one of the 'beta' versions that some peopel seem to have.

2) ADC_CLOCK_INTERNAL is _not_ recommended above 1MHz CPU clock rate. If you intend to use the ADC in the future, worth reading the data sheet and correcting.

3) If using the hardware UART, you _must_ always have 'ERRORS' in your RS232 setup, unless you are adding the code to handle errors yourself.

4) When developing code, don't use 'PROTECT'. Only enable this when your code is finished. Having it enabled, increases the wear and tear on the chip's flash memory, meaning that the programmer has to erase every cell in the chip, each time even one byte is changed.

5) Bit problem is what happens at the end. Your code is dropping off the end, which means the chip will go to sleep. Now the exact timing in the code sequence of this I'd have to sit down and calculate, but my guess is that the extra time involved in the interrupt code, is letting it just drop off the end at the wrong time. CCS code is 'standalone'. There is no OS for the code to go 'back' to. Your code should always ensure that it doesn't do this.

Mike Walne · Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK

Learn to use the code button, preserves your formatting, makes code easier to read.

What happens if you skip (comment out) these lines?

codeboy2016 · Joined: 15 Dec 2016 Posts: 2

I am using version 4.023

Sorry for confuse you.

Basically the question of the project has nothing deal with uart or analog input. But problem is on the real time clock interrupt.

Please read on the line 14 and line 18.
You will see 'output_high(pin_C5);' and also'output_low(pin_c5);'.
when these 2 lines are DISABLED, the led at pin_c3 and pin_b4 works perfectly.

But when either of the lines enabled, then only b4 works perfectly (follow 3 sec on 3 sec off). But pin c3 is only turned on less than 0.5sec then auto turn off.

I only connect to a led and through resistor 1kr. So nothing with reset or unstable. Because b4 works perfectly, and also c3 works perfectly when line 14 and 18 disabled.

Here is the new coding, please read again.

Ttelmah · Joined: 11 Mar 2010 Posts: 19328

Ouch....
Do you really mean 4.023.

That was at best an 'early beta' version.
I don't actually have a V4 version this low. I only keep versions once they are proven to be nearly working. At this time I was still using 3.249, since V4 wasn't yet running. The earliest version I kept was 4.027, which still had major problems, then versions starting a little later....

Big problem is you have LVP enabled. This will cause problems unless the LVP pins are biased off.

A test with 4.027 (with LVP off), shows it running fine, except for the LED being about 20% slow (since it is spending so much time in the interrupt - remember you are calling this every 256 instructions and it uses about 70 instruction)....