Stupid DMX PHY question

Jay Ashworth

Well-Known Member
In the article at

https://erg.abdn.ac.uk/users/gorry/eg3576/DMX-phy.html

it says:

In summary:
  • Receivers must not connect the signal shield/earth to their local ground. This could otherwise result in a ground-loop, adding noise to the received signal.
  • Transmitters should ground the shield/earth.
  • The ground/shield must be connected at both ends of a length of cable so that electrically this is connected along the entire cable segment, otherwise extension cables joined end-to-end would not be shielded.
But that seems circular. If a device has DMX IN and THRU, then by rule one, we don't connect the shield to anything on the IN. But we *must*, by rule 2, ground the THRU. And by rule 3, we must connect the shields together from the IN and the THRU, which forces us to violate rule 1.

Clearly, I am missing something. Could someone tell me what? :)
 
I'd argue the thru is not a transmitter.
So the in and thru get daisy chained and pin 1 carried through.
Then as far as the fixture is concerned, it's an input and left floating...
 
I was wondering if that was the part I was missing. I tend to think of fixtures more than consoles since there are so many more of them, and most fixtures don't have an out.
 
Out does not equal thru.
If you look at the physical traces for 'in' and 'thru' on most devices, you'll see they are directly connected. Thru is not a transmitter, thru is just the continued daisy chain.

Out is a new, or re-buffered signal ready for another 32 units.

To quote an old post on thru:

Most simply connect 1 to 1, 2 to 2, and 3 to 3, and branch off to the board in a simple Y connection. This is an imperfect method and it is best to keep the wiring as short as possible. The board itself should not have a terminator on it. (4 to 4 and 5 to 5 if you are using standard 5 pin. No pcb connection on those.)

Each time this is done, some degradation occurs and it is one of the reasons no more than 32 units should be on any given run. Use opto-isolators with multiple outputs to gain numbers higher than 32. (Star topography.)

Some units buffer the signal and actually have separate in/out connections on the PCB. Sounds good, but actually causes another problem: If one unit goes down, everything downstream is lost. Can't think of any I've opened up where it was that way.

As for the connection between pin 1 and the case, Some connect it directly, but it is better to put a 10 ohm resistor between the two. This helps prevents the data cable from becoming part of the ground circuit between fixtures. If your project case is plastic, no need. If you do use a resistor, purchase a fuse-able resistor and cover it in a small section of fiberglass sheath. That way if there is a catastrophic ground fault, you will not set your DMX cable on fire. (Most important on fixtures with an electrically grounded case.)
 
So the upshot of this is that the DMX shield should never be earthed except at the board, or a repeater/splitter?

And, by extension: if you unplug a DMX universe from its controller, and test for continuity from pin 1 to the shield, it should be open, and if not, you should know why?

Or is it still not that simple?
 
Woops, didn't see that second question.

If everything is following the rules, unplugging the master should result in pin 1 / shield reading open.
 
Woops, didn't see that second question.

If everything is following the rules, unplugging the master should result in pin 1 / shield reading open.
Reading open between pin 1 and the shield or between pin 1, the shield and ground?
One more touch of ambiguity in this thread: In some instances, a poster refers to "board" and is inferring a local PC "board" while at other times the reference is to the output of the DMX source, the lighting "board". Possibly, for posterity, we should be clearer when referring to a board or boards.
Toodleoo!
Ron Hebbard.
 
Assuming i'm remembering XLR correctly, pin 1 is always connected to the shield and the connector. Pin one is an equipment ground in balanced applications, not a signal ground. The end result is to use the cable shield as an extension of the interference protection offered by the host chassis. So if anything, it should read as open circuit in reference to mains ground when unplugged from the host / master / DMX OUT.

If not, then one of the devices down the line is connecting ground, which can result in ground loops as has already been said.

So recap:

Pin 1 is Chassis Ground from HOST.
pin 1 is the same as connector shield.
 
Last edited:
Assuming i'm remembering XLR correctly, pin 1 is always connected to the shield and the connector. Pin one is an equipment ground in balanced applications, not a signal ground. The end result is to use the cable shield as an extension of the interference protection offered by the host chassis. So if anything, it should read as open circuit in reference to mains ground when unplugged from the host / master / DMX OUT.

If not, then one of the devices down the line is connecting ground, which can result in ground loops as has already been said.

So recap:

Pin 1 is Equipment Ground from HOST.
pin 1 is generally the same as connector shield.

Pin 1 should never connected to the connector body of an XLR, either in DMX or audio applications. It is a signal ground, not a chassis ground.
 
Pin 1 should never connected to the connector body of an XLR, either in DMX or audio applications. It is a signal ground, not a chassis ground.

I agree. Bonding the connector body is asking for trouble.

The simple way to think about this is to not think about it at all. Cables should all be wired straight though. If the equipment follows the standard, then nothing special has to be done by the user in terms of preventing ground loops. If a manufacturer does not follow the standard, may all their products fall off the ship from China.
 
I stand corrected.

I should probably check some of my XLR cables, could have sworn I saw the body soldered to pin 1...
 
Last edited:
I stand corrected.

I should probably check some of my XLR cables, could have sworn I saw the body soldered to pin 1...
You probably did. Many imported XLR cables seem to have pin 1 tied to the connector shell; they always think something has to be connected to that lug. Always inspect or test new ones!
 
You probably did. Many imported XLR cables seem to have pin 1 tied to the connector shell; they always think something has to be connected to that lug. Always inspect or test new ones!
If I'm recalling correctly, cables routinely included with Shure SM58's (and likely with many of their other mics as well) used to arrive with an XLR cable female pre-installed on one end with the shield soldered to pin 1 AND the shell. The other end was pre-stripped for the purchaser to install their own connector of choice. In my recollection, I believe this goes back to the days when the east and west coasts couldn't agree on the polarity of contacts 2 & 3 in so far as which of the two contacts was deemed "hot".
Toodleoo!
Ron Hebbard.
 
No, I'm pretty sure I always meant lighting board rather than circuit board. But another new question is raised here: are fixture manufacturers usually pretty good about correctly labeling an out versus a through? I have in mind particularly consumer-grade fixtures or DJ grade, if you will, where they can do DMX mastering from one to another
 
Out does not equal thru.
If you look at the physical traces for 'in' and 'thru' on most devices, you'll see they are directly connected. Thru is not a transmitter, thru is just the continued daisy chain.

Out is a new, or re-buffered signal ready for another 32 units.

To quote an old post on thru:
Within the quote within this post is the following "Most simply connect 1 to 1, 2 to 2, and 3 to 3, and branch off to the board in a simple Y connection." where I thought the reference to "board" was ambiguous and thought possibly the reference was to a PC board within a given fixture.
Toodleoo!
Ron Hebbard.
 
Ah that. On alot of the fixtures I've seen, the DMX portion is on it's own PCB, that is what was being referred to in his quote.
 
Ah that. On a lot of the fixtures I've seen, the DMX portion is on it's own PCB, that is what was being referred to in his quote.
Understood @EdSavoie and that's exactly how I interpreted it. I just felt possibly the thread could/should have been edited for both posterity, and the benefit of any neophytes reading, as the bulk of the time the term "board" was referring to the DMX source, the output of the LX board. Most times things were clearly termed as printed circuit boards, or PC boards, or PCB's except in the instance of this one example where simply typing board read as somewhat ambiguous.
Toodleoo!
Ron Hebbard.
 

Users who are viewing this thread

Back