Multiple Consoles on a single DMX line
- Thread starterg15
- Start date
A DMX line *is* an RS-485 line. The DMX standard specs:
* a physical layer: XLR5 connectors, particular types of shielded twisted pair cables
* a signalling layer: differential voltages as specified by EIA RS-485 at 250kbaud
* a data layer: min 88μs break, 1 start byte, up to 512 channel data bytes
Of course the data layer can really be carried by any medium you want--ie, packed up and sent through IP over ethernet, or heck even USB or carrier pigeons--but a 'vanilla' traditional "DMX line" is wholly and completely an RS-485 line. Yes, it can be point-to-point if you have only two devices on the line, but the moment you add a third device, the topology chnages from point-to-point to linear multidrop by definition.
EDIT: the idea that DMX is not multidrop may come from a (common, I think) misconception that receiving devices receive and retransmit DMX data. They do not. The male XLR5 on a non-isolating receiving device is wired straight through to the female XLR5.
* a physical layer: XLR5 connectors, particular types of shielded twisted pair cables
* a signalling layer: differential voltages as specified by EIA RS-485 at 250kbaud
* a data layer: min 88μs break, 1 start byte, up to 512 channel data bytes
Of course the data layer can really be carried by any medium you want--ie, packed up and sent through IP over ethernet, or heck even USB or carrier pigeons--but a 'vanilla' traditional "DMX line" is wholly and completely an RS-485 line. Yes, it can be point-to-point if you have only two devices on the line, but the moment you add a third device, the topology chnages from point-to-point to linear multidrop by definition.
EDIT: the idea that DMX is not multidrop may come from a (common, I think) misconception that receiving devices receive and retransmit DMX data. They do not. The male XLR5 on a non-isolating receiving device is wired straight through to the female XLR5.
Last edited:
00AVD
Active Member
Of course DMX is multi-point when viewed at system level, but the individual cable segments are all point to point. You should never see a DMX cable with more than two connectors on it. If it has a mid-point connector, that would have to be only used for a device with zero (or very short) stub length.
Going back to the original point, creating a cable with a mid-point that a lighting console can be plugged into is still a bad idea. You can't terminate the line correctly in that situation as the mid-point would be terminated, leaving one end of the cable "flapping in the breeze".
Going back to the original point, creating a cable with a mid-point that a lighting console can be plugged into is still a bad idea. You can't terminate the line correctly in that situation as the mid-point would be terminated, leaving one end of the cable "flapping in the breeze".
MarshallPope
Well-Known Member
First, let me say that this gave me an extremely...interesting...mental image. Would bats work? we have plenty of them in our theatre.
Back closer to topic - Can someone please explain exactly what terminating an unused branch of the circuit accomplishes? I understand why a regular line would be terminated, but why does this need it? To me, it seems like the signal wouldn't have anywhere to go on that branch in the first place, for lack of a better way to explain what I'm thinking.
00AVD
Active Member
With RS485 systems, each end of the overall circuit needs to be terminated to avoid signal reflections. In a DMX system, the lighting console is at one end and always terminated.
The other end of the first cable goes the first fixture. If that's the only fixture, you terminate there and that's the end of the story. If there are other fixtures, the signal passes in and out of all of them and terminates on the last fixture.
With a cable that has a console mid-point, you can't terminate the line correctly as it should only be applied in two places - one at each end. But the lighting console is always terminated, so what can you do? You can leave the unused male end of the chain unterminated and hope for the best, or you make up a dodgy terminator using a female plug and end up with 3 places of termination. Neither are the right solution.
High speed RS485 does not permit stubs (branches) as they break all the rules about termination. For low speed data, you can get away with it as the reflections are not significantly time delayed compared to the bit time of the slow data.
Just because there's nothing connected to the unused stub (branch) doesn't mean that the signal isn't going there. High speed data can behave very differently to what you might believe.
The other end of the first cable goes the first fixture. If that's the only fixture, you terminate there and that's the end of the story. If there are other fixtures, the signal passes in and out of all of them and terminates on the last fixture.
With a cable that has a console mid-point, you can't terminate the line correctly as it should only be applied in two places - one at each end. But the lighting console is always terminated, so what can you do? You can leave the unused male end of the chain unterminated and hope for the best, or you make up a dodgy terminator using a female plug and end up with 3 places of termination. Neither are the right solution.
High speed RS485 does not permit stubs (branches) as they break all the rules about termination. For low speed data, you can get away with it as the reflections are not significantly time delayed compared to the bit time of the slow data.
Just because there's nothing connected to the unused stub (branch) doesn't mean that the signal isn't going there. High speed data can behave very differently to what you might believe.
Whether or not having the console in the middle of the line violates termination rules depends on whether or not the console terminates the line. If it does, then yes, you would have termination in the middle of the line and at one end, which would not be ideal. However, there is not strictly a need for a console to terminate the line, since it is expected to be the only transmitter, and in the single transmitter case, you don't need to terminate at the transmitter location because the transmitter will always be driving the line there. So I'm not sure which if any consoles terminate, since there's no real need and they can save a fraction of a cent on a resistor there. I sent a note to ETC to ask about their products, but haven't heard back.
So that said, if you move a non-terminating console to the middle of the line, then plugging a terminator into the male at the normal 'head end' of the line *would* be the right solution, even if it means a female plug terminator that defies our normal ideas of gender roles for connectors.
On the other hand, you are right that long stubs should be avoided in general, but it's more of a question of actual driving and receiving characteristics and cable lengths as to whether or not a particular stub size will work. There's no practical way to completely eliminate stubs after all.
So that said, if you move a non-terminating console to the middle of the line, then plugging a terminator into the male at the normal 'head end' of the line *would* be the right solution, even if it means a female plug terminator that defies our normal ideas of gender roles for connectors.
On the other hand, you are right that long stubs should be avoided in general, but it's more of a question of actual driving and receiving characteristics and cable lengths as to whether or not a particular stub size will work. There's no practical way to completely eliminate stubs after all.
00AVD
Active Member
I would assume all consoles are terminated as convention says they are always at one end of the DMX chain. It would be easy enough to measure the resistance between pins 2 & 3 on the console's output when turned off.
A reading of 100 - 120 Ohms should indicate termination, assuming the manufacturer has not used a more exotic termination method.
A reading of 100 - 120 Ohms should indicate termination, assuming the manufacturer has not used a more exotic termination method.
00AVD
Active Member
Not at all. Don't confuse termination with a simplex vs half-duplex protocol. They are two quite separate issues.
Termination is there to keep the data lines from developing "ringing" and data corruption from occuring as a result.
As long as the DMX chain is terminated once at each far end, it makes no difference which node (console or RDM device) is transmitting.
Termination is there to keep the data lines from developing "ringing" and data corruption from occuring as a result.
As long as the DMX chain is terminated once at each far end, it makes no difference which node (console or RDM device) is transmitting.
DMX termination consists of placing a 120Ω resistor across pins 2 and 3 at the end of each DMX branch. Most people buy or make terminators that are just a resistor soldered inside an XLR connector.
It is a simple device that is required by the DMX standard, but many people go without. Most of the time you can get by without termination, but every now and then having unterminated lies can cause big issues. Some devices have a termination switch on them that you can flip to engage or disengage termination depending on where the device is in the chain.
So let's say you have a system with the console at one end of the line, a number of receiving devices, and then proper termination at the end, with proper wiring throughout. If the system does not use RDM, then the console does not need to be terminated because the console will always be the only transmitter on the network, so signals travel only in one direction (away from the console) and will be absorbed by the terminator at the opposite end. No signal (valid or reflected) will travel in the opposite direction.
If that same physical system *is* designed to use RDM, you now can have signals traveling in either direction at any given time on any given segment of the network: the console sends towards the 'foot' end of the line, the last receiving device can send towards the 'head' end of the line, and a device in the middle will send towards the head and foot ends simultaneously. So now you need a terminator at both ends. This is the case where the console *should* in fact terminate the line, so hopefully consoles designed to support RDM do so. Which means that an RDM console should not be placed in the middle of the network, because then you'd have termination in the middle, and that's bad.
00AVD
Active Member
I guess it boils down to this:
If the console (RDM or non-RDM) has no termination resistor internally, you could plug it into the middle of a DMX chain, as long as you also terminate the head end (where the console would normally be plugged into) with a "reverse sex" terminator.
If the console does have a termination resistor internally, you're out of luck and it can only go at the "head" end.
Now the trick is, do console manufacturers clearly specify this?
Bad juju above discussed above - and a better read of it I need so as to better understand current technology.
Still though back in like 92' I was in college and for designing our lights for class projects we had like a 24 or 48 channel Leprecon two scene peset linked up with a ETC Vision light board. You could use the faders on the Leprecon to set your levels in the live and move over to the Vision so as to save them. I was a scenic designer back than and lighting was only my minor thus I don't remember details - but I do remember it worked. Adjust the fader manually on the two scene preset and it would both adjust the lights and adjust the scale on the monitor of the Vision for dimmer levels. The Vision could also take control in a sort of "last being prescidence" type of way. How did this work than given the above? Don't think it was analog or was it?
Still though back in like 92' I was in college and for designing our lights for class projects we had like a 24 or 48 channel Leprecon two scene peset linked up with a ETC Vision light board. You could use the faders on the Leprecon to set your levels in the live and move over to the Vision so as to save them. I was a scenic designer back than and lighting was only my minor thus I don't remember details - but I do remember it worked. Adjust the fader manually on the two scene preset and it would both adjust the lights and adjust the scale on the monitor of the Vision for dimmer levels. The Vision could also take control in a sort of "last being prescidence" type of way. How did this work than given the above? Don't think it was analog or was it?
Didn't the Vision have a DMX in as well as out? The idea being that you could use the Vision to capture DMX values from another source, just like you were doing in cuing, and I guess have the option of failover if one board or the other failed while running for some reason.
Yes. I suppose I wasn't very clear that this is a point I was trying to make earlier, but it seems we are in agreement now.
I don't think so. IIRC, the original Vision predated DMX, and could be spec'd to output CD-80, D192, LMI, or analog. I suspect the term eluding ship is "analog readback." I vaguely recall using a MicroStar set-up this way, that was analog-in/analog-out. I think the Vision was analog-in/whatever-out, but am not positive.
There are also DMX Combine units. Fleenor makes one, ETC makes one, I'm sure others do as well. I use it all the time. It allows multiple consoles to be on a single DMX line. Highest takes priority. Mine, a Fleenor, has 3 ins and 1 out and is set to remember the last DMX command indefinitely in the case of lost input.
