RDM Controller timing clarification

[Dan Scheurell]

I’d like to verify my understanding of the RDM Controller Packet Spacing Times. These are given in Table 3-2 of the RDM spec.

For discovery, a controller sends the command and then waits up to 2.8 ms. If no response, then there are no responders and we move on to the next branch. If there is a response within 2.8 ms, then receive responses until a total of 5.8 ms has passed. So, from the controller's point-of-view, a discovery response always takes either 2.8 ms (no responses) or 5.8 ms (responses).

For normal operation, the controller waits up to 2.8 ms for the response. If no response is received, then wait another 200 us before sending any other packet.

Is this a correct reading of the spec?

Thank you,
Dan

[sblair]

Dan,

Your understanding is pretty close I'd say. My reading of the document would say that as a Controller you must always wait 5.8mS after sending a Discovery packet before transmitting any other packet, including another Discovery Packet.

I'm not sure why we stated in that manner though other than to make sure there isn't any missed traffic as a result of collisions happening. Worst case, if the responders waited until right up until the end of their window to start transmitting and then you have collisions, it might be possible to miss the response and then incorrectly infer that there was nothing in that branch.

Scott

[gthaliath]

Hi Scott,
We have been working on responders and while doing a discovery with different RDM controllers they all come up with diffrerent numbers.

We had a test setup of 32 nodes on a network(1km long) and the controller used where

1.Controll Freak(ADDICT)
2.ENTTEC(RDM USB pro)
3.DMXter 4 RDM

out of which the second two worked fine.
But the ADDICT was showing numbers all over the place.(26 to 34)(34 when it was showing address not on the network, no way or cross talk from any other devices just a random address)

how could we check if the responders on the network are working fine in respect to timing.

Regards
George

[ericthegeek]

Originally Posted by gthaliath

But the ADDICT was showing numbers all over the place.(26 to 34)(34 when it was showing address not on the network, no way or cross talk from any other devices just a random address)

This is a common problem with RDM controllers that don't implement discovery properly.

During discovery, you have lots of responders trying to respond at once. Sometimes these collisions will line up in such a way that it looks like a valid discovery response for a non-existent responder UID.

When it sees a valid discovery response, the controller is supposed to mute the UID from that response, and then only add it to the list of discovered devices if the responder replies to the mute request. But some poorly implemented controllers add the non-existent UID even when there's no response.

Make sure you're running the newest firmware on the ADDICT. If you still have the problem get in touch with the manufacturer and work with them to get the problem solved. It might help to point them to these forums.

Originally Posted by gthaliath

how could we check if the responders on the network are working fine in respect to timing.

I'd start with your DMXter. Run a full discovery. In the "RDM Controller" menu, there is a "View Responder Timing" option, this will show you details about discovery timing.

[gthaliath]

Thanks Erick