Cheap DMX signal split

[TehBIGrat]

Hi,

I'm looking at purchasing some 5 volt 3 Channel DMX decoders for some RGB LED's,

Basically I'm going to house each one in a translucent white box, ice-cream container, so i have several different 'pixels'.

the units i am looking at purchasing do not have a DMX through/out, Is there a cheap way of splitting the signal between them,

I would rather not buy a $100 octo splitter for every 8 of my 'pixels'.

Thanks

 

[Tex]

Aren't most DMX through/out connections usually just soldered directly to the input?

 

[rm2011]

I believe you are referring to an opto-isolator opposed to an octo-isolator, which as far as I know, doesn't exist. Anyways, I'm pretty sure you will have to buy ONE opto-isolator (that is, one with 8 outputs). I would highly recommend not trying any diy twofer techniques.

Sent from my Galaxy Nexus using Tapatalk 2

 

[len]

DMX splitters are all over (assuming that's really what you want). I have a couple Elations, and a couple Martin products. Fleenor are available as well, but you said "cheap." I hear Chauvet is fine as well. Contact any of the fine dealers who post here.

 

[DaveySimps]

I have used the Elation splitters on short runs, and have been pleased with them. Especially if you are not touring and banging them around in a truck / road case.

And, although I am sure you meant opto splitter, an OCTO does indeed exist.
Pathway Connectivity Inc. - Pathport Octo

~Dave

 

[temper]

Aren't most DMX through/out connections usually just soldered directly to the input?

Yep, every dmx device that I've taken apart (dimmer packs, led fixtures, etc) have the in and thru jacks just soldered or otherwise physically connected across in parallel. And, that's pretty much implied in the dmx spec anyway (when you have a spec that states a total run length for the wire, instead of run length between nodes which require amplification on the thru, it usually means that the devices are hooked in parallel across the control or signal wires).

Here's a little rant, but as I've gotten into dmx technology over the last couple years, it's amazing to see an industry that has completely preyed on people's lack of understanding. That $10 terminator - a 120 ohm resistor across the signal pair. Take a look for some opto-isolator schematics and you should be able to build one for $30 in parts (including jacks and the box), but they regularly go for $2-500. I'm sure a modern, fast switching op-amp would do just as well as an opto-isolator (network switches don't opto-isolate anything yet work at much faster speeds, and 10baseT hubs used to be completely passive), and although I haven't tried it, electrically a simple passive splitter is probably going to be fine for just about anything that isn't running more than a couple hundred feet. So, if it were me, I would just connect them all in parallel and put a terminator on the end of the entire dmx run.

 

[Chris15]

Can you provide some more details on these units that you are looking at that don't have the DMX through?

In most cases, the in and the through will be paralleled inside the unit, and frankly a hundred mm or two of Y split tail is going to make no difference to the integrity of the DMX run...

 

[Chris15]

Here's a little rant, but as I've gotten into dmx technology over the last couple years, it's amazing to see an industry that has completely preyed on people's lack of understanding. That $10 terminator - a 120 ohm resistor across the signal pair. Take a look for some opto-isolator schematics and you should be able to build one for $30 in parts (including jacks and the box), but they regularly go for $2-500. I'm sure a modern, fast switching op-amp would do just as well as an opto-isolator (network switches don't opto-isolate anything yet work at much faster speeds, and 10baseT hubs used to be completely passive), and although I haven't tried it, electrically a simple passive splitter is probably going to be fine for just about anything that isn't running more than a couple hundred feet. So, if it were me, I would just connect them all in parallel and put a terminator on the end of the entire dmx run.

NO NO NO NO NO.
Ethernet is isolated by virtue of the transformers used on every port - they are normally built inside the shell of the connector itself. There is 1.5KV of galvanic isolation required for any device that actually meets the ethernet specs.

A 120R resistor might only be a few cents, but add the cost of a connector and someone's time to assemble it and you've easily got $10.

If you can get a box and connectors for 30 bucks, then they are rubbish. They won't last in light usage and they certainly won't survive the harsh conditions experienced on tour.
Isolated means isolated, so you need N power supplies for the device, where N is the number of outputs plus 2 (1 per output, 1 for the input and one for the stuff in the middle).

If you think long Y legs are OK on a DMX run, you don't know all that much about reflections and standing waves on a signal bus do you?

No op amp known to man will survive a mains short across it, yet an opto will stop the fault propagating.

 

[dramatech]

To emphasize what Len said, there are plenty of Opto splitters and non-opto splitters available for distribution of DMX. While we would all like to have Fleenor splitters, many projects just don't warrant such expense and quality.
I have 2 Elations, 1 Neo-neon, 1 Northlight and some Chinese product that I purchased just to see how good it could be for $32. They all work and even the cheapest has Texas Instrument ICs.
One of my Elations has had a 24 volt regulated supply attached to the top and all of the output connectors changed to 4 pin XLRs and serves as a scroller power supply for 20 scrollers. The Neo-neon has had the 3 pin connectors changed to 5 pin. The cheap Chinese unit has no isolation between outputs, other than the RS485 driver ICs, in this case they are surface mount, have 3 pin connectors, and an external wall wart supply. It sure comes in handy for driving Chauvet color splash Jrs. It also has 8 outputs which is great when driving lots of LED effects.
Are they all perfect, are they top of the line? No!! They work and are installed according to the priority of how important a failure would be. So far after 7 years on all but the cheap Chinese unit, there have been no failures and they are on most all of the time. The cheapo unit has only been around for a little over a year, with no failures. Would I use it on any really important stuff? Absolutely NOT!!
The point is that DMX ISO splitters are several power supplies, Isolation ICs and RS 485 tranceivers, plus a housing and connectors. Most al of the ICs are the same manufacture. The big differences are the circuit board and transformer quality. We sometimes tend to become equipment "snobbs". On the other hand, when the DMX product has some sort of software such as a merger, then the best you can afford becomes the logical rule. Splitters don't have any software, they just receive the signal pass it through an isolation and send it back out with more outputs.
The soldering and circuitboard quality of the units I have vary a bit, but all will work. Of the ones that I have the Elations have the best product quality "under the hood".
From what I have observed, if I am asked which to buy, I generally suggest the Enttec four output for $100, but there are many more choices that will serve the purpose as well.

 

[temper]

NO NO NO NO NO.
Ethernet is isolated by virtue of the transformers used on every port - they are normally built inside the shell of the connector itself. There is 1.5KV of galvanic isolation required for any device that actually meets the ethernet specs.

A 120R resistor might only be a few cents, but add the cost of a connector and someone's time to assemble it and you've easily got $10.

If you can get a box and connectors for 30 bucks, then they are rubbish. They won't last in light usage and they certainly won't survive the harsh conditions experienced on tour.
Isolated means isolated, so you need N power supplies for the device, where N is the number of outputs plus 2 (1 per output, 1 for the input and one for the stuff in the middle).

If you think long Y legs are OK on a DMX run, you don't know all that much about reflections and standing waves on a signal bus do you?

No op amp known to man will survive a mains short across it, yet an opto will stop the fault propagating.

Chris15, not to pick a fight, but yeah, I guess I'm a little on the grumpy side this morning and probably took it a little too far and of course, you are right on all accounts. My overall point wasn't that opto-isolation is bad or that y legs are good practice, etc. It was more along the lines of reacting to someone who is considering buying an opto-isolated splitter to drive each dmx device when in all honesty he probably is fine with putting them all on the same bus and terminating it properly.

Oh, and I guess I just don't love spending $100-200 on a simple 4-way dmx splitter.

Just out of curiosity though, how long would a y leg on a dmx run that is properly terminated on all ends of the bus have to be before it would start to cause problems? It's a question I've wondered about for a while, just never really asked or looked it up.

 

[Chris15]

Likewise, no fights intended.

The OP hasn't given us enough information, what these devices s/he wants to get the DMX to is one thing, how many of these pixels is another...

My gut reaction is that the best solution will be to parallel them in groups of say 20 - 25 and run each of those from an opto split output. If the devices are not the greatest of designs, then loading too close to the theoretical limit of 32 devices on the bus will probably be unwise.

To opto split outputs for things that in all likelihood are going to be fed from the same power supply just seems wasteful...

My gut is to say you want your Y split tails under say half a metre. Reflections are always going to matter more as you go up the frequency spectrum. DMX doesn't run THAT fast, but it's still 250k baud, it's not like the 1200 baud systems used in architectural lighting control that can tolerate almost any wiring configuration...
If it weren't time to go to bed, I'd consider thinking about doing the maths about wavelengths and all that to try and narrow down the length at which it would start to become a problem...

 

[gafftapegreenia]

The ENTTEC D-Split is around $125, whereas a Doug Fleenor 5 way is around $800. A Doug Fleenor splitter survived being run over by a fork lift, whereas I wouldn't want to drop the D-Split off my desk. For me, a splitter adds security and peace of mind. While DMX wye cables may work, I find that anything with a chance of failure will tend to fail 15 minutes before the house opens, even after behaving all through tech.

 

[derekleffew]

... Just out of curiosity though, how long would a y leg on a dmx run that is properly terminated on all ends of the bus have to be before it would start to cause problems? It's a question I've wondered about for a while, just never really asked or looked it up.

From Doug Fleenor Design - DMX Primer (written 20 years ago!):

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.

TehBIGrat, just wiring, parallel your outputs and inputs at each "fixture". Plan on no more than 31 units per stream, and use terminator, DMX s. (What Chris15 said in post#11.) Also as suggested, one can buy an opto-splitter without a case, connectors, or power supply. One (expensive) source: Pathway Connectivity Inc. - #1002 Opto-Splitter . I don't see any on Northlight System's site, but his would probably be less spendy.
.

 

[JD]

"Pixel" - Hummm.... That implies you are about to make hundreds of these things. If that is the case, regular DMX distribution may be prohibitive. Are these going to be assembled in to one big "thing" ? If that is the case, you may want one DMX in and out loop, and treat the whole thing as one fixture.

Now, lets look at what actually happens to the DMX when it hits the jack on your "thing." Since it is one unit, there is no need for opto-splitters. As the name implies, the "opto" part electrically isolates each output. You do not need this type of isolation. What you will need is someone to design a circuit that buffers the heck out of the signal so that each of the modules can read the DMX. Although this can be done quite cheaply, it must be designed correctly. Your buffer is going to be spitting something out that electrically resembles DMX but can no longer be considered DMX at it will be at a much lower impedance. As long as the waveform is not distorted and is of the correct amplitude, you can then load hundreds of pixel units, with their receivers, on it. Although the parts may be quite cheap, hiring someone to do the design work, and fix the bugs will be your greatest expense. Here's the limiter- If this "array" is in one unit, then busing the "super-DMX" signal will not be a problem. If, however, there are many individual units, the signal busing will be your downfall. Each sub-feed would have to be the same length and arranged in a "star" topology, or the signal reflections in the wiring will mutilate the waveform.