Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
02/05/08 03:56
Read: times


 
#150341 - Multi master my way
Responding to: ???'s previous message
I did an implementation of this technique a while ago. So, my thoughts may be a bit rusty on this. However, this is how I did it then

Each node has two RS485 transceivers; one is a switched transmitter(turn DIR to transmit) and the other is a permanent receiver. Now, whenever a node wishes to transmit, it just listens for activity on the line(any received character). If there is no activity, shoot out the packet byte by byte and listen in at the same time. If the data you get back is not what you have sent, BACK OFF for a random delay time and retry.
Similarly, all other transmitting nodes will BACK OFF and free the bus for a random time delay when a collision happens. The receivers though will not know about a collision. They will know about a defective packet by computing the packet checksum/CRC that you implement.

This is my way of doing a collision detection. I'm sure there may be other ways to it, but this does not need any additional lines except the RS485. This scheme will not have fixed latencies, but it worked for me.

Jerson

List of 18 messages in thread
TopicAuthorDate
Multi Master communication RS-485 ?            01/01/70 00:00      
   how fast do you need to go?            01/01/70 00:00      
      Use another cable pair - for bus busy            01/01/70 00:00      
         that will not work            01/01/70 00:00      
   How's about using with MAX1490            01/01/70 00:00      
      what good will that do ?            01/01/70 00:00      
         Common solutions            01/01/70 00:00      
            Multi master my way            01/01/70 00:00      
               Still on progress            01/01/70 00:00      
                  wrong chip?            01/01/70 00:00      
   Maybe Token Ring            01/01/70 00:00      
      ARCNET            01/01/70 00:00      
   re:multimaster            01/01/70 00:00      
      what if two masters send 'busy' at the same time?            01/01/70 00:00      
         re:collision            01/01/70 00:00      
            all good and well; however ...            01/01/70 00:00      
   separate the masters for network            01/01/70 00:00      
      sure, but the problems are the same            01/01/70 00:00      

Back to Subject List