Banjo Spam schrieb:
> A couple of notes
>
> 1) I can't tell if the "; do someting" is actual code,
> or filler to make the example smaller.


In my test circuit there are indeed nop's and nothing else. I stay in
Bank 0.

> 2) Your interrupt code will modify counter_hi and
> counter_lo possibly at the same time you are moveing
> to W and send to TXREG.


Yes, I'm aware of this. But for testing I'm not interested in a correct
result but in an working transmission.

> 3) Disable any flow control on the computer
>

??? I'm working with a terminal in HEX mode. If there are no interrupts
I can see two zeros and a <cr> as expected. So when no interrupt is
trigeered, everything works fine.

> 5) You don't mention if 'freezes'



I've added some additional LEDs:

- After each second, an LED (LED1) is toggled by the main loop
- One LED (LED2) is switched on while waiting for TRMT being cleared
after a transmission
- One LED (LED3) is switched on when the interrupt routine is entered
and switched off before the routine is left

Now I can desribe the 'freeze-state' a little more precise: 1 second
after power-on and the transmission of the first byte, the PIC waits for
TRMT to be cleared (LED2 remains on) and LED1 is not toggled any more -
so it waits for ever for TRMT to be cleared. The interrupt is triggered
as before (LED3 flickers).

When I trigger the comparator interrupt slowly by hand with a
potentiometer, the amount of int-calls is transmitted correctly, so the
RS232 part works. But when I turn the poti very fast by hand it can
happen that the PIC freezes as described above. Maybe I generate some
ripples around the hystresis point. In my application I expect about
300-500 impulses per second from an operation amp which last about 125s
and the DC-range is from 1,5 to 3,0V. So I use a comperator voltage of
about 2V. In this case The circuit freezes (see above) after having sent
the first byte.

Does the rate of interrupts affect the clearing/setting of TRMT? The
UART interrupts are disabled.

Can anyone understand this strange behaviour.

Regards
Markus

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu