Control/Dimming - DIY DMX Splitter

[Fatcatpro]

I am simply looking to create a 1 in 2 out DMX Splitter but cannot afford a professional opto splitter. If I were to wire some loose Xlr ends 1-1 2-2 3-3, similar to a mic splitter would this work? Or is there something special about an opto splitter that you need to have?

[mstaylor]

No, DMX doesn't like this. What are you trying to do?

[Fatcatpro]

Get dmx to both SR and SL without daisy chaining

[metti]

As mstaylor said, you can just do a hard split like that. It is possible to DIY an opto splitted but it would probably about as expensive and way more work compared to Enttec's $100 opto.

[MNicolai]

As mstaylor said, you can't just do a hard split like that. It is possible to DIY an opto splitted but it would probably about as expensive and way more work compared to Enttec's $100 opto.

The splitter mentioned by metti is this one. It only costs $100, and has (4) outputs and can be purchased with either (4) 5-pin outputs or (2) 5-pin and (2) 3-pin. It's much more inexpensive than other equivalent options from other manufacturers.

[ccm1495]

DMX is data not an analog signal like audio. So if you where to just split the cable you are going to have reflections in the data and all kinds of other nasty ness. A DMX Opto splitter also electrically isolates the signals from one another, and amplifies the data signal. Metti suggested the ENTTEC $100 OPTO Splitter, its not the best but it works I use one of them and it fits my needs.

[mstaylor]

Get dmx to both SR and SL without daisy chaining

Are you trying to save cable or do you have electrics that have to move indepentantly? If it is the former, it would only cost you a single cable for each jump back. If it the latter then you will need at least the $100.00 splitter shown. A DIY opto splitter, a little tough.

[JD]

By the time you bought the chassis mount connectors and the box, you would already be approaching the cost of the ENTTEC. And that would be if your labor was free!

The splitter serves two other useful functions:
1) A voltage spike on the data line kills the splitter instead of your board.
2) It serves to buffer your signal at the stage end, which helps cut down on data line noise, while increasing the allowable cable run.

[WooferHound]

I don't understand why you can't just make a Y cable or a splitter box. Isn't the signal split that way in every lighting fixture that it is daisy chained through?

[Lotos]

I don't understand why you can't just make a Y cable or a splitter box. Isn't the signal split that way in every lighting fixture that it is daisy chained through?

I like to hope that in most, if not all, gear the full DMX packet is received, buffered, and re-transmitted down the line. At least, at the price point I'm paying for my intelligent gear, I should hope someone had the foresight to throw an extra RS485 chip in there. Or at the bare minimum, optically isolate the output from the input.
(The fact that I'm certain there's cheap gear out there that does, literally, Y the signal... Notwithstanding)



Going back to the OP... It boils down to this:

Yes, simply splitting the DMX signal with a few dollars worth of components will work... Sort of.
However, it will work unpredictably, and do nasty things including (but not limited to) voltage drop, signal reflection and increased 'noise' in the signal.
And it will fail (oh boy, you can count on it) and when it does, you can bet every last dollar that you didn't spend on a proper opto-splitter that it will be at the absolute worst possible moment.

DMX devices are 'dumb' recievers... They hear a start code, and then listen to 512 bytes... Paying attention to the ones they think they need.
Reflections can cause havoc, because they'll transpose data from somewhere else in the packet overtop of the correct data... Or there may simply not be enough signal by the time it reaches the device for it to be read properly... Or it could just be riddled with noise... Or, Or, Or...

To put it bluntly: Your moving heads will wig out unexplainably, your scrollers will get the shakes, and your hazer may just decide to not haze.
(Anyone who's worked with DMX has likely seen any and all of these symptoms occur with proper gear that has had a component within the DMX transmit/recieve module start to fail... And wouldn't, voluntarily, wish this on anyone.)

For $100 you can save yourself (and everyone who inherits the space from you) a world of DMX gremlins, unexplainable random errors, and nastiness...


Sidenote: you probably shouldn't Y audio cable either.
Why Not Wye?

[derekleffew]

I don't understand why you can't just make a Y cable or a splitter box. Isn't the signal split that way in every lighting fixture that it is daisy chained through?

I like to hope that in most, if not all, gear the full DMX packet is received, buffered, and re-transmitted down the line. At least, at the price point I'm paying for my intelligent gear, I should hope someone had the foresight to throw an extra RS485 chip in there. Or at the bare minimum, optically isolate the output from the input.
(The fact that I'm certain there's cheap gear out there that does, literally, Y the signal... Notwithstanding)

Does the fixture pass DMX without being powered? With the fixture disconnected from power and data, is there pin to pin continuity between data ins and outs? If yes, it doesn't do any processing or re-transmitting the signal. The data output is wired in parallel with the input, and there is a stub out to the fixture's receiver circuit (a T-split, not a Y-split, not that there's much difference). But, the thing is, that T-stub is so short (from 1/2" on a PCB to a 3" jumper) so reflection is not an issue, and it's only driving one receiver. What happens inside the fixture, stays inside the fixture. Not so with an external Y-cable.

From Splitting a DMX line :

Provided you use an opto-splitter, and not a Y-cable or two-fer, no problems at all.

From DFD's DMX512 Primer:

Splitting Never split a DMX signal with a "wye" cable. Use a splitter device that buffers each line separately or daisy chain from device to device.
The problem with splitting the signal with a wye cable is that the signal going up one leg of the wye is reflected back down and corrupts the signal going down the other leg. The longer the legs of the wye, the worse the problem. A wye where each leg is 10 feet may work fine but may fail if the legs are increased to 100 feet. A lot has to do with the quality of the cable, the strength of the signal at the wye point, and your relationship with God.
If you must split the signal, and you don't have a buffered splitter, do it as close to the source as possible, i.e. at the console. If you have an isolator available, place the wye at the output of the isolator. If you have two isolators available, place the wye at the input to the two isolators (which is like making your own splitter).

[JD]

The key to the in-fixture split is that the monitoring circuit is high impedance (around 10k ohms), where as the transmission line is low impedance (120 ohms.) As such, it does not load the line, or provide a significant backboard for an echo to bounce off of. Still, there are practical limits to how many times you can do that, which is the reason for the 32 limit. If you split the line you disrupt the "transmission line." The result is that instead of the signal hitting the terminator at the end, it bounces off two ends and may set up interference patters that make the data unreadable.