DMX distribution idea

[Aquarius]

So, some background about our theatre:

We have 2 DMX ports in the whole thing. 1 in the booth, which the light board (ETC Express 24/48) is always plugged in to. The other one is on the stage right house light control panel.

The problem:

We have a dance concert coming up, and the choreographer requested fog machines. We recently converted our fog machines to DMX, as we got tired of running them manually and wanted to run them straight from the board. We normally position our fog machines just upstage of the proscenium arch, and have the dressers run them from there. However, now that we converted them to DMX, we would need to either run an XLR line over the arch (way too much time and cable) or over the dance floor (ugly, and a tripping hazard even if it was taped down).

The solution:

The plus side to our theatre is that we have mic ports coming out the wazoo. Each of the foggers is conveniently located right next to a pair of them, and we don't use any mics for dance shows. We came up with this solution: run DMX line from our 1 DMX port SR into the SR fogger. The cable from the XLR out port on the SR fogger runs into one of the mic ports. The input line on the SL fogger gets plugged into another mic port. Up in the booth, we have a patch bay. Using stereo patch cable, we patch together the mic ports that both of the foggers are plugged into.

Voila! We now have a way of extending DMX to anywhere in the theatre that we have a pair of mic ports (which is basically everywhere). And it even works well, despite the length of (hidden) cable involved.

Anyone ever tried this before?

 

[Thefoxygranpa]

Has anyone done it before? Yep, many times. When searching older topics this question comes up a lot and in short, it will work, sort of. I'm running older trackspots down an audio snake line from the booth in back of house to a jack on stage, and it has worked just fine for the past couple of months.

The install wire may not be DMX spec and therefore can create some issues, as well as your patch cords and the wire used on the patchbay themselves may not be on spec with DMX. And as always, you should terminate your signal if you are not already

Hope its working well for you!

 

[Tex]

Voila! We now have a way of extending DMX to anywhere in the theatre that we have a pair of mic ports (which is basically everywhere). And it even works well, despite the length of (hidden) cable involved.

Anyone ever tried this before?

I'm assuming that your patch panel is comprised of XLR connectors? Mine is 1/4" stereo phone plugs that do not allow DMX to pass through.

 

[mstaylor]

It will probbly work but don't be surprised if weird things happen. Standard mic cable isn't made to handle DMX. If you are connecting the two sends in the booth and are sending the signal back to the stage then to the DMX imput on stage then back to the booth, that's a lot of cable length and can create it's own problem.
A better solution may be to run to the pit floor, across the arch and back to the next fogger. The other option is to place your foggers upstage and cross against the upstage wall. Personally I would go over the arch but that's me.

 

[derekleffew]

...Anyone ever tried this before?

Bad idea. Thefoxygranpa has noted the issue of using audio cable to carry a DMX signal, so I'll address another problem.

The system is designed to allow for a console to be placed either in the booth or backstage right. The wiring probably goes from the booth to the SR input, where it is "Tee'd," then on to the dimmers. Using the stage input as an output creates a "Y," which is not permitted per the DMX512/EIA-485 standards.

Yes, I've done it, but only for non-critical applications (smoke machines probably qualify) and always/only with an opto-splitter immediately after the onstage input. Using an input as an output also requires a FF turnaround, which is usually an indication that you're doing something at least unintended, if not outright wrong.

 

[epimetheus]

I'm assuming that your patch panel is comprised of XLR connectors? Mine is 1/4" stereo phone plugs that do not allow DMX to pass through.

What kind of 1/4" do you have that doesn't allow DMX to pass through? I agree that the audio cable isn't DMX spec and that there is probably a 50/50 chance of this working, but a 1/4" TRS has all the same connections as an XLR, unless your patchbay has common shields or something like that.

 

[fredthe]

The wiring probably goes from the booth to the SR input, where it is "Tee'd," then on to the dimmers. Using the stage input as an output creates a "Y," which is not permitted per the DMX512/EIA-485 standards.

This actually brings up another issue... termination. IF there are two points feeding the dimmers (booth and house), then presumably the dimmers are terminating the connection. Which means that you can't terminate the run to the foggers, else you'll be double terminating.

Yes, it might work. No, it's not the right way. (See the thread on "Jerry Rigged")

-Fred

 

[JD]

Be aware that most "pulled" audio cable (what's in the conduit) is 64 ohm. (Often Belden 8761) DMX will work as mic cable, but mic cable should not be used for DMX. Pulling DMX would be a good idea as you can splice new to old and then use the old cable to pull the new one through. (be careful!) Whenever splitting the run, use an active opto-splitter. Each output line should then be terminated. As for using stereo phone jacks for DMX, sounds like a good way to static spike the line and damage something. With XLR cables, pin #1 is designed to make contact first, which drains off any static charge. Phone Jacks hit tip first, so any stray voltage will spike the actual data line.

 

[epimetheus]

Be aware that most "pulled" audio cable (what's in the conduit) is 64 ohm. (Often Belden 8761) DMX will work as mic cable, but mic cable should not be used for DMX. Pulling DMX would be a good idea as you can splice new to old and then use the old cable to pull the new one through. (be careful!) Whenever splitting the run, use an active opto-splitter. Each output line should then be terminated. As for using stereo phone jacks for DMX, sounds like a good way to static spike the line and damage something. With XLR cables, pin #1 is designed to make contact first, which drains off any static charge. Phone Jacks hit tip first, so any stray voltage will spike the actual data line.

Good point about the tip making contact first. I had thought of that. If you were to try to run DMX through a 1/4" patchbay, I certainly wouldn't try to make a connection with the system powered up.

 

[BillESC]

Just a quick note about running digital signals through installed house audio cables.

We were contracted to provide lighting and effects for a dance recital in the brand new 4 plus million dollar auditorium at West Point High School, West Point, VA.

To make things easy I brought a half dozen NSI 4 channel dimmer packs to plug the effects into and plugged the microplex into an audio channel behind the procenium wall. As soon as I powered up the packs, all kinds of things started to blink, wink, strobe, etc.

After trouble shooting the situation, I discovered all the the house audio were run with just two conductors. Nice job audio contractor.

 

[Chris15]

I'm assuming that your patch panel is comprised of XLR connectors? Mine is 1/4" stereo phone plugs that do not allow DMX to pass through.

but a 1/4" TRS has all the same connections as an XLR, unless your patchbay has common shields or something like that.

Be aware that most "pulled" audio cable (what's in the conduit) is 64 ohm. (Often Belden 8761) DMX will work as mic cable, but mic cable should not be used for DMX. Pulling DMX would be a good idea as you can splice new to old and then use the old cable to pull the new one through. (be careful!) Whenever splitting the run, use an active opto-splitter. Each output line should then be terminated. As for using stereo phone jacks for DMX, sounds like a good way to static spike the line and damage something. With XLR cables, pin #1 is designed to make contact first, which drains off any static charge. Phone Jacks hit tip first, so any stray voltage will spike the actual data line.

Good point about the tip making contact first. I had thought of that. If you were to try to run DMX through a 1/4" patchbay, I certainly wouldn't try to make a connection with the system powered up.

Ok, considering a number of things;
To say mic line is 64 ohm is likely wishful thinking. Until digital came along, most manufacturers might list an impedance on the data sheet, but the actual impedance is probably best described as whatever came off the machine...
AES grade audio cable is in general suitable for DMX usage (110R nominal Z).
Patch bays - I don't recall seeing a 6.5mm patch bay in some time that used standard plugs. They normally use longframe plugs that have tips that are smaller and so don't suffer the shorting on insertion syndrome as much. Agreed that in general, patching whilst non live is good. The reality is that you only need 2 conductors for DMX, it's a differential balanced line and industrial control systems run RS485 (which is the underlying transport of DMX-512A) for hundreds of metres without shielding...

Oh and common grounds on anything digital are bad news. AES audio suffers badly when attempted to run on common grounded multicores...

 

[epimetheus]

Ok, considering a number of things;
To say mic line is 64 ohm is likely wishful thinking. Until digital came along, most manufacturers might list an impedance on the data sheet, but the actual impedance is probably best described as whatever came off the machine...
AES grade audio cable is in general suitable for DMX usage (110R nominal Z).
Patch bays - I don't recall seeing a 6.5mm patch bay in some time that used standard plugs. They normally use longframe plugs that have tips that are smaller and so don't suffer the shorting on insertion syndrome as much. Agreed that in general, patching whilst non live is good. The reality is that you only need 2 conductors for DMX, it's a differential balanced line and industrial control systems run RS485 (which is the underlying transport of DMX-512A) for hundreds of metres without shielding...

Oh and common grounds on anything digital are bad news. AES audio suffers badly when attempted to run on common grounded multicores...

The 3rd (5th) wire (the bare wire) is needed for DMX or any other RS-485 signal. This wire provides the ground reference for the balanced signal, without it the signal plus and minus can pick up a voltage offset that will fry an input that isn't properly isolated. This is more of a concern for longer runs than it is for short runs.

BTW, industrial applications do run RS-485/422 for hundreds of meters, but not with shielding/ground wire. Industrial RS-485/422 installations consist of either +, -, and G or RX+, RX-, TX+, TX-, and G. The former, a 2-wire 485 system, is only half-duplex, i.e. one way communication at a time. It is bidirectional, but not at the same time. The latter, a 4-wire 422 system, is full duplex, i.e. bidirectional at the same time, since transmit and receive do not share the differential pair as in 485.

 

[JD]

The 3rd (5th) wire (the bare wire) is needed for DMX or any other RS-485 signal. This wire provides the ground reference for the balanced signal, without it the signal plus and minus can pick up a voltage offset that will fry an input that isn't properly isolated. This is more of a concern for longer runs than it is for short runs.

Correct!
In practice, the RS485 transceiver chip needs ground. It can only compare the + and - inputs if those signals are between its own supply rails (0 and 5 volts) Outside of that range, the output will be random regardless of what ratio is between the data lines. The "ground" line, or actually common, connects the electrical commons (0 volt) on the circuit boards in all of the units. This is also why opto-couplers are so good. The "ground" or pin 1 on the input does not have to be the same as the ground on the output (thus the "opto" part), but the proper reference is maintained on both input and output legs.

Proper operation is best achieved when the + and - data lines are between 0.75 volts and 4.25 volts with reference to ground. This is due to the fact that the transistor junctions on the chip require a junction drop voltage. If the + has a higher voltage then the -, the chip will output a "1". If the - has a higher voltage then the +, the chip will output a "0".
Some examples:
"+" = 3.5 volts "-" = 3.3 volts, output = 1
"+" = 2.2 volts "-" = 3.3 volts, output = 0
"+" = 0.1 volts "-" = 0.3 volts, output = arbitrary
"+" = 4.8 volts "-" = 4.9 volts, output = arbitrary
"+" = 4.5 volts "-" = 0.5 volts, output = arbitrary
"+" = 4 volts "-" = 1 volts, output = 1
In the third, fourth, and fifth cases, the voltages are too close to the chip supply rails for the chip to operate properly.

 

[Knave]

Just a quick note about running digital signals through installed house audio cables.

We were contracted to provide lighting and effects for a dance recital in the brand new 4 plus million dollar auditorium at West Point High School, West Point, VA.

To make things easy I brought a half dozen NSI 4 channel dimmer packs to plug the effects into and plugged the microplex into an audio channel behind the procenium wall. As soon as I powered up the packs, all kinds of things started to blink, wink, strobe, etc.

After trouble shooting the situation, I discovered all the the house audio were run with just two conductors. Nice job audio contractor.

I've had similar experiences. It was the shared ground on the patch panel in our case, but it was disappointing at any rate.

If it works and is reliable enough to run your show, awesome!
Don't count on it when you're planning a show, though.

A spool of Cat5 and some time soldering is far cheaper than a show that's missing critical effects/fixtures, or worse a fog machine that goes off at the wrong time, a mover that can't be controlled...

(If I had a money tree I'd just keep a wireless DMX system in my toolkit for scenarios like this. I bet they aren't too expensive to rent, though.)

 

[Les]

A spool of Cat5 and some time soldering is far cheaper than a show that's missing critical effects/fixtures, or worse a fog machine that goes off at the wrong time, a mover that can't be controlled...

Very well said.

 

[FMEng]

If one of those mic jacks doesn't get properly patched, and the audio console gets turned on, 48V phantom power could be applied to a DMX connection on the equipment. If that happens, there will be smoke, but it won't be coming from the fog machine's normal place.

This is especially dangerous if the patchbay is normalled to the console, because connections are made when no patch cords are inserted.

 

[Chris15]

I rescind my earlier comments about the optional nature of ground for DMX. A truly balanced system ought to be able to be run at any common mode voltage, without a common reference, with the signal being conveyed purely by the difference of the signals relative to one another. It seems RS 485 is not truly balanced...

 

[JD]

It seems RS 485 is not truly balanced...

You are correct! Example, a balanced microphone line. Doesn't matter if there is an offset voltage. In fact, that is exactly what the phantom power is. DMX is not a true balanced system as DC can throw it off. In fact, one of the more common chips (SN75176) ( http://focus.ti.com/lit/ds/symlink/sn75176b.pdf ) input toggles at 2.3 volts with reference to ground. in other words if "-" is 2.4v and "+" is 3.5 volts, the output will still be arbitrary! To build a true balanced system you would encounter problems when the stream hits a long run of "0"'s or "1"'s in that DC decoupling would cause drift. It could be done, but the electronics involved would kill the cost.

 

[Chris15]

To build a true balanced system you would encounter problems when the stream hits a long run of "0"'s or "1"'s in that DC decoupling would cause drift. It could be done, but the electronics involved would kill the cost.

I tend to disagree... You just need to change the encoding method. Look at the way AES audio uses Biphase Mark Coding - to produce a] clock recovery via the data stream, b] zero DC component - ground IS optional on AES.and c] polarity neutrality, you can transpose hot and cold without issues...

 

[JD]

I tend to disagree... You just need to change the encoding method.

Then it would no longer be the DMX standard. We are stuck with what we have.