I’m working this issue now for the XTB-IIR, and will finish that collision tutorial when I get the time. However, I would like some input from those of you who may be really familiar with the X10 protocol.
As most of you know, X10 powerline communication is through a series of 120KHz bursts on the powerline. A burst after a zero crossing is a “1”, and the absence of a burst is a “0”. When two commands overlap, a collision occurs when a burst is sent in a “0” slot. That is easy to identify when repeating a X10 command because the XTB-IIR has a complete copy of the received command, and knows exactly what is in each bit window.
It isn’t quite as easy when the source is local, since the XTB-IIR transmitter doesn’t know what bits are coming. The X10 protocol never has more than two “0” bits in a row until the gap at the end of a command. I’m planning to use a 3 “0” gap to recognize that a local message has completed. I know the protocol specifies a gap of 6 half cycles, but if the XTB-II has terminated its transmission, it can’t do anything about a tailgater stepping into the gap at the end of its transmission.
When collision avoidance is enabled in the XTB-IIR, it would terminate a transmission immediately whenever a collision is detected. It would normally re-transmit when the line has cleared. I’m not sure this is necessary with the digital input because the acknowledgement would not be sent if the complete message is not transmitted. That would be similar to receiving a corrupted acknowledgement from the TW523. If the XTB-IIR automatically re-transmits when the line has cleared, it might confuse the controller. Comments?
Anyone have any insights that I may be missing?
Jeff
Author
Topic: X10 Powerline Collisions (Read 12880 times)