Thomas,
I wish I could have a one on one conversation with you. This is not, and has not ever been, personal. I've tried to make it
clear that I am representing information and tried to also
address your questions and points. Clearly I have failed. Do not listen to me because of where I work - I've never asked for that or expect it.
For those that are interested, I will try to
address some of these topics.
In fixtures, and other devices that have a
DMX connection, there is indeed a stub. This is defined as the amount of distance between the transceiver chip and the
DMX transmission
line. In some cases this distance is zero because the lines connect at the input of the chip which means the chip is sitting directly on the data bus. In other products, this distance might be 6" or so as the
DMX In and
Thru connectors will be soldered together at the front panel and then a single twisted pair then goes to the board and chip. Best practice design [elicited by the standard and several use docs] say that any stubs should be kept as short as possible. So yes, at each device that has a non-zero stub, that is a "Y" or "T".
It is true that a
DMX console plugged into the middle of a
line is sending data in two directions, ie a "Y" or "T". Technically, an RS-485
line can be terminated at each end and this will eliminate reflections on the
line. This is not what was done many years ago, nor to this day when such systems are installed. I can tell you, however, that I have installed terminators on both ends in cases where data reflections caused issues. Reflections are based on several parameters, the most notable being distance of the
DMX cable. Time is distance down a
wire and with
wire at the perfect length, data waveforms can be altered or canceled out due to reflections off the unterminated end.
RDM does require specific, non-DMX termination and it must be held correctly at each location due to bi-directional data flow and transmit status. I have found, on a number of installations, that the termination is even more critical.
RDM places transmitters all over the data bus, not just at the ends, which is why data integrity is even more important than
DMX.
Isolation chips are not designed to isolate chatter as they do not know about proper timing. If a reflected signal is the dominate signal, the receiver chip and isolator accept the changed timing and
send it on the CPU for processing. Isolators are designed to provide galvanic or optical isolation to reduce the possibility that an out of tolerance signal will get coupled into the main electronics. It's just like the galvanic isolation that is provided on each port of an ENET
switch where it just passes on anything it gets but doesn't know if it's a real signal or not. Yes, more expensive products do have them and they are wonderful when someone sends the wrong signal down a data
line or there's a nearby lightning
strike.
Somehow I have missed where the OP stated his
DMX In port is at the end of a
DMX chain. Unless we are talking about where the
console is plugged in.
I have scope shots of corrupted
DMX, copies of
DMX specs and best practices, schematics of products and I can post that if that will help people reading this. I want a healthy discourse and to work using standards and best practices. Thomas, you can do what you are saying - I never said you couldn't. I did say I wouldn't and that other methods are more sure.
Anyone that has now spent the time reading through all my dribble and has an interest in knowing more about
DMX practices, I would highly recommend Recommended Practice for
DMX512, by Adam Bennette. It takes the
DMX standard it turns it into a doc that can more easily be interpreted to use cases in the field and with product design.
Thomas, if I have failed to explain things clearly, I would appreciate a phone
call when you have a free moment. It's possible that things are just getting lost in translation.
David
eight eight eight - 908-2142 reaches my
desk toll free