Design Questions about DMX Cable Shields, Earth Ground, and Signal Ground

DIYLED

Active Member
Hi, thank you for all the good info on this forum. In the past, I've occasionally stumbled across this place from Internet searches and found useful advice.

Now I'm doing electrical work for an upcoming art installation for an event. After doing research, I'm still wondering about a few details that may or may not affect the reliability of the lighting system. It would be nice to be confident ahead of time that everything will work properly without problems from RF noise or ground loops, instead of having to fix issues during the installation. Maybe someone here can offer some advice. Here's the general setup:

DMX will be sent through shielded cat5e cable. It's a foil shield with drain wire, not a braided shield. Cable will be running in the same ceiling and wall spaces as Romex power wire. There is no metal conduit.

The generator will be at least 50 feet from the structure.

Opto-splitter: it's unclear whether/how signal ground is connected to earth. When it arrives, I can measure the direct current resistance.

There will be 4 DMX lines of 175 to 275 feet each, with 4 to 8 receivers on each line and terminators at the ends.

On some receivers, the signal ground is directly connected to the shield. It would be a pain to change this, involving a dremel tool and epoxy, but not impossible. I would likely break a few cutting wheels because it's a tight space between the jacks and the circuit board.

On other receivers, I can choose how or whether to connect signal ground to earth ground. For example, a 10nF cap, nothing, or direct connection at 1 receiver if nothing else on the same run makes that connection.

DC switching power supplies: signal ground is connected to earth ground via a capacitor of unspecified value in the schematic.

From what I've been reading, it's ideal to connect signal ground to earth ground with low impedance (less than 20 ohms) at the DMX transmitter only and high impedance (greater than 22 Mega ohms) connections at the receivers, although direct connections at receivers are sometimes okay but discouraged because of ground loop issues. It's unclear to me whether the shield should be connected to earth ground at one location only or if it can or should connect at every fixture.

If all these details are too time-consuming to analyze for some random stranger on a forum (me), I would be curious to just hear personal experience on when RF noise and ground loops actually cause problems and when it's generally okay to not bother following the strict recommendations. Of course, technical advice would be greatly appreciated as well.

Thanks in advance.
 
The best source of information on all of this is at http://tsp.esta.org/tsp/documents/published_docs.php (free) "ANSI E1.27" parts 1 & 2 cover cabling for portable and permanent installations. "Recommended Practice for DMX512: A guide for users and installers" (at the bottom) is a bit easier to read.

I got tired of explaining the details to installing electricians, so I wrote a summary for the guys on the job. "DMX for Electricians"

What you've posted sounds very typical unless you are building the DMX transmitters and receivers yourself. I would hesitate to depend on DIY DMX devices. There are great products readily available to solve virtually any requirements. Both www.dfd.com and www.pathwayconnectivity.com are top class and participate here frequently.

The short story is that DMX512 is so robust that grounding/RF issues rarely cause problems. DMX cables go to DMX receivers and don't get grounded otherwise. DMX lights get power and ground just like any non-DMX fixtures.
 

Attachments

  • DMX for electricians.pdf
    170.3 KB · Views: 1,010
The short story is that DMX512 is so robust.....

I like to think of it as very tolerant, rather than robust. Google definitions-

Robust: strong and healthy; vigorous.

Tolerant: showing willingness to allow the existence of opinions or behavior that one does not necessarily agree with.

:)
 
Shield should only be wired to pin 1, not the case. in other words, signal grounds and frame grounds should be kept separate. On equipment and cables that have pin 1 and the shell connected together, you may want to insert a ground lift cable. Basically, a short section of cable that does not have pin 1 and the shell connected. Some equipment may appear to have pin 1 and the case connected but it may not actually be the case. Often, there may be an inductor or low ohmage resistor actually separating them. In most cases, such as opto-splitters, it does not actually matter as the cables plugged into it don't have the shell connected, and the power supply is a wall-wart with no ground pin. Just be careful if the splitter has a metal case that it is not in contact with anything. The system will probably work fine even if that connection is there, but if you are having problems, it would be high on my list of things to eliminate.
 
Thanks, Rick. I was wondering how to access some of the official documents. I'll take a look at those and the other guide you recommended. Your guide implies that signal common and shield are always connected together. Now things are making sense.

Sounds like ground loops and RF noise are usually a bigger issue with audio, possibly really long DMX runs. Sounds like "tolerant" means typical ground loops of a few millivolts are insignificant in a 2.5V swing or the common mode rejection eliminates it. Whereas with audio, the ripples would get amplified in the op-amps. I read somewhere (I think on this forum) someone recommended shielded cable if it runs near power cable and in parallel, especially with DC switching power supplies.

And thanks JD. That clears things up a bit. In my case, signal common would be hard connected to the shield at many of the receivers but not hard connected to earth ground because each DC power supply uses a capacitor between earth and signal common. I believe the cap is a very low value because when I measure resistance it doesn't show even a little anything on the meter as the cap fills up. In this scenario, it looks like I might want to make 1 connection on each line between shield and earth ground if the opto-splitter doesn't make that connection. And on the receivers that don't connect signal common to shield, it looks like I should make that connection.
 
If the DMX devices adhere strictly to the electrical specifications, they have already isolated pin 1 on inputs. Only outputs would have pin 1 connected to ground. So the standard prevents ground loops. However, we all know that a few manufacturers violate standards.
 
As per my knowledge the shield should only be wired to pin 1, not the case because signal grounds and frame grounds should be kept separate.
A short section of cable that does not have pin 1 and the shell connected.
Some equipment may appear to have pin 1 and the case connected but it may not actually be the case.
Often, there may be an inductor or low ohm resistor actually separating them.
 
Thanks for all the help! It worked perfectly. I was able to ground the shield at the controller and opto-splitter output by soldering ground wires to shielded cat5e couplers. There was even 150ft of coiled cat5e next to the generator and no problems.
 

Users who are viewing this thread

Back