Mike Harrison wrote:
> On Thu, 09 Nov 2006 16:23:51 -0700, you wrote:
>
>   
>> gus gusset wrote:
>>     
>>> I'm trying to write a receiver for the theatrical
>>> lighting control protocol DMX (8-bit async serial, 1
>>> start, 1 stop, 250kbaud).
>>>       
>
> That should be 2 stop, but this shouldn;t make a difference to a receiver
>
>   
>>> In order to detect break (>> 2x a single word), I am
>>> polling RCSTA.FERR
>>>
>>> For debug, I turn PORTC,0 on when I enter this part of
>>> the code, and clear it shortly afterwards. If were
>>> ever to see an FERR, PORTC,3 would be set. Using a
>>> scope, I can see that I do enter (and subsequently
>>> leave) this part of the code, however, I never see
>>> FERR. Note, I've seen other people's DMX-receiving
>>> code; they use FERR to detect break. Unfortunately, I
>>> can't find a substantive difference between their code
>>> and mine (they are using different PICs).
>>>
>>>       
>
> Are you polling for FERR BEFORE reading the byte from the UART? Reading the byte will clear FERR.
> Are you remembering to take the byte from the uart after checking FERR? 
>    
>   
>> I wrote a DMX handler too. I was not aware that FERR was meant to
>> detect a break.
>>     
>
> Not 'meant to' but a very convenient way to do so. 
> FERR happens if the line is asserted during the stop bit, so this will happen on a break.
> Works fone on AVRs and unless there is something peculiar about the PIC's UART I would expect it to
> work just the same. 
>
>   
>> What I did to detect the break is to simply allow a short timer to timeout
>> whenever serial characters failed to be received; when a character WAS
>> received, the timer was preset. It then counted down using the TIMR1
>> interrupt. It was quite precise, and would detect the break before
>> being too far into it. I set the timer to timeout when two character bit 
>> times
>> failed to arrive. At 250KB, that was 20/250kb.
>>     
>
> I don't think DMX defines maximum inter-character gaps within a frame, so wouldn't this scheme
> detect false breaks if there were gaps between bytes within a frame?
>
>   
It was not defined, but in fact I never saw the DMX sender ever take a 
break between characters. I made my detect circuit fast
because I needed to use the break to catch up on routine data updates.

--Bob
>
>   

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