| ??? 02/09/04 15:27 Read: times |
#64301 - RE: Parallel Async Communication Protoco Responding to: ???'s previous message |
Ebrahim Vakilpour wrote:
------------------------------- It mans that two 8051s should stop their jobs and start talking to each other which I said before that's not what I’m looking for. I want an Async (or whatever you call it) communication so that 8051s do not interrupt each other for communication. I think you are mixing up queueing and sync/async. If you do not have a supervisor chip, your 8051 will have to stop whatever he is doing if a byte is sent for transmission in order to read (and accept) the byte. You will have to queue (read: buffer) your data if you do not want to treat your data immediately, but that is a plain (not always simple) buffer mechanism. AN example: If you are sending data to your printer, does he stop printing to receive the data? I don't think so.. It is receiving data into it's buffer, and yet it is printing.(and very likely using a parallel centronics or IEEE1284 I/F...) So why should both your devices be tied up communicating? It is not so much a matter of sync/async functioning but more a question of throughput and the available cpu-time. If your transmitter transmits a byte every second, then there is no reason for your processor to simply wait for a second, doing nothing, before receiving a second byte. Certainly so if it takes only a few machine-cycles to read a byte and save it in memory. If however, you expect to sent a few Mbit/s (sustained), then you would "tie" up your processor, but then a fifo would be of no help... So, going back to your protocol: If you do not want your 8051's being interrupted for communication, I think you'll need a supervisor chip with a buffer that will store the data until the 8051 is ready to read it. What was the maximum throughput you were thinking of? regards Patrick |



