Stef Mientki wrote:
> Bob Axtell wrote:
>   
>> Alan B. Pearce wrote:
>>   
>>     
>>>> I've found that if I put a nop after writing to TXREG, before going
>>>> back to my TXIF wait loop, I increase the chances of successful xmission
>>>> a whole lot (though not 100%).  Without that nop, I loose a character
>>>> probably 95% of the time I write a string of text out.  With it, I
>>>> have a clean transmission 95% of the time.
>>>>     
>>>>       
>>>>         
>>> I have experience only with the 16F876, but I seem to remember that there is
>>> a finite time between loading the character into TXREG, and the TXIF showing
>>> a correct status, it is probably one clock at the 16x baud rate, or whatever
>>> your multiple is. It didn't affect me as I used Olin's macros to use a FIFO
>>> to stuff characters into, then had the UART send them under interrupt
>>> control. There was no loop inside the interrupt, so there was probably a 30
>>> or so instructions minimum before the next character was loaded.
>>>
>>>   
>>>     
>>>       
>> Gee, I never see any of this on my 877A and 876A designs.
>>   
>>     
> If you use
>   a high level language with a bad optimization
> or
>   a badly written assembler program ;-)
> you will never have such a tight loop,
> and thus will never see the problem.
>
> Stef Mientki
>
>   

Well, I only write in assembler, however I always have 15-20 
instructions minimum before
I test again.

--Bob
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist