Control/Dimming DMX cable shielding?

klaasvt

Member
Hi DMX sages!
I'm working on a new DMX installation (moving over from Analogue). I have an 80m cable run to make from my desk location to the dimmer packs, and I've heard that it might be best to use unsheilded cable for longer dmx runs... Is that the case? It seems that most cables, even when it is specifically marked as DMX cable (like large rolls of un-terminated cable even), is still sheilded... Can you explain more?

Thanks!
Klaas
 
Mic cables have a higher capacitance that take the digital waveform of 0's and 1's and over a long enough distance blurs the difference between the 0's and the 1's. Once the data corruption reaches a certain point, the devices may not be able to decode some or all of the waveform and cause your fixtures behave erratic or not at all. If you mix and match mic cables and DMX cables, sometimes the different impedance between types of cables will cause a signal reflection that'll come back down the cable run and cause data corruption as well.

You can use mic cables in a pinch but I would never depend on it and would never use it for installed cables. DMX is very forgiving generally so it works right up until suddenly it doesn't.

There can be special ways to selectively ground or lift the shield on installed audio cables to minimize ground hums in sound systems, but that's not pertinent in any way to using mic cables for DMX. Portable mic cables should always have the shield landed on Pin 1 at both ends.
 
I'm wondering if there's a confusion between ground and shield. The DMX cable needs three connections, hot, cold and ground. The shield is separate from the ground. The cable should have 1-1, 2-2, 3-3 at each end and shields connected to the shell at one end only, but my understanding is that the shield shouldn't be connected to the signal ground, as this can cause hum loops and interfere with the signal, just as you get hum loops in audio cables where multiple grounds and shields are interconnected. And yes, mic cable, over long runs, will work right up until it doesn't. UTP works great for installed DMX runs.
 
Pin 1 should always be where the drain wire for the shield or common conductor lands.

There are very rare circumstances where the chassis ground tab of an XLR shell may be bonded to Pin 1, such as in manufactured devices with chassis-mount connectors. Field cabling and portable cables should never have the tab connected. Generally because of the risk of the XLR shell coming in contact with surfaces that have different ground potentials.
 
Pin 1 should always be where the drain wire for the shield or common conductor lands.

There are very rare circumstances where the chassis ground tab of an XLR shell may be bonded to Pin 1, such as in manufactured devices with chassis-mount connectors. Field cabling and portable cables should never have the tab connected. Generally because of the risk of the XLR shell coming in contact with surfaces that have different ground potentials.

Something to consider when using Blizzard's products- Their RokBoxes have 5V power supplied to the connector shell for their wireless dongles. It will short out if you use DMX cables that have the connector chassis grounded. Don't know if this extends to other Blizzard products or if any other manufacturers are doing this, but it will kill the fixture if you use cables with the connector grounded.
 
Pin 1 should always be where the drain wire for the shield or common conductor lands.

There are very rare circumstances where the chassis ground tab of an XLR shell may be bonded to Pin 1, such as in manufactured devices with chassis-mount connectors. Field cabling and portable cables should never have the tab connected. Generally because of the risk of the XLR shell coming in contact with surfaces that have different ground potentials.
I didn't realize this. When a 5pin XLR is disassembled, is the spring that makes the tension within the connector not connected to any pin? I guess that would make sense because neitrik doesn't specify what pins do what and that would be a scary assumption.
 
Correct. You have to wire that tab separately if you want the shell of the connector to be part of the shield or ground path.

In reality, DMX, being very forgiving, doesn't care if 1-2" a few times in a signal chain doesn't have a complete shield around the mated connectors. Same for audio. The potential consequences of wiring those tabs in often far outweigh the marginal signal purity benefits you might be lucky to achieve under ideal circumstances.
 
Confusion comes from lack of understanding. Good on you for investigating! As usual, these things are more complex than first appears.

DMX doesn't need a ground! Sending or receiving devices usually do, but mixing that with the signal reference is not required, and as noted troublesome. A member of the DMX standard committee has said that +90% of the noise rejection is due to the twists of the data pair. That would make shielding almost unnecessary. Even Mr Fleenor has described situations where the signal reference isn't needed or used.

That said, I will follow the published standard and not put my professional reputation at risk.
 
Confusion comes from lack of understanding. Good on you for investigating! As usual, these things are more complex than first appears.

DMX doesn't need a ground! Sending or receiving devices usually do, but mixing that with the signal reference is not required, and as noted troublesome. A member of the DMX standard committee has said that +90% of the noise rejection is due to the twists of the data pair. That would make shielding almost unnecessary. Even Mr Fleenor has described situations where the signal reference isn't needed or used.

That said, I will follow the published standard and not put my professional reputation at risk.

There are 2 reasons for the shield and corresponding drain wire - to prevent the radiation of data noise (RF) from the twisted pair, and to minimize induction of interfering currents from the outside world. If neither of these is perceived as a problem by the end user it is possible to use a twisted pair of the correct capacitance and nominal impedance (like CAT5e or CAT6). If we can shunt that noise/induced current to ground, so much the better.

Following established, published standards is a really good idea. I think Doug Fleenor (correct me if I'm wrong, Janell) has said, to the effect of, "people spend a lot of money to fly me around the world to tell them to use real DMX-spec cables and not put more that 32 devices on DMX string..." and implied that the majority of DMX problems can be traced to the use of either or both in the same system.

Of the DMX-related problems I've tracked down, the cable and fixture count are easy. It's the flaky fixtures - bad solder on the DMX connectors or PC board or physical damage - and bad cable termination that make for more involved troubleshooting. Cold solder joints in audio are mostly annoying but in data they're either corrupting or fatal errors.
 
Following established, published standards is a really good idea. I think Doug Fleenor (correct me if I'm wrong, Janell) has said, to the effect of, "people spend a lot of money to fly me around the world to tell them to use real DMX-spec cables and not put more that 32 devices on DMX string..." and implied that the majority of DMX problems can be traced to the use of either or both in the same system.

Yup. Basically... follow all the big rules, and you can probably get by with breaking small ones. It's when you ignore too many that things start going wonky.

This also applies to life in general. :p
 
Confusion comes from lack of understanding. Good on you for investigating! As usual, these things are more complex than first appears.

DMX doesn't need a ground! Sending or receiving devices usually do, but mixing that with the signal reference is not required, and as noted troublesome. A member of the DMX standard committee has said that +90% of the noise rejection is due to the twists of the data pair. That would make shielding almost unnecessary. Even Mr Fleenor has described situations where the signal reference isn't needed or used.

That said, I will follow the published standard and not put my professional reputation at risk.
As said above, there is confusion between ground, shielding and signal reference. DMX512 is a 3 conductor system.
The third conductor may or may not be a shield, and it can be shown as stated above that the shield provides minimal protection from noise. The signal reference is the third conductor and should be connected thru on Pin 1. The other major problem is ground loops. If the signal reference is connected to ground at more than 1 location, ground loops are possible. This is a negligible problem over 10 feet, but when you start to talk about 80m it is problematic. By connecting the shield at multiple locations, you have now strung a piece of 22awg wire in parallel with the earth ground conductor and general current flow within the building's steel. The only sure method is to use a single point ground, ideally at the transmitter, and the receivers are isolated from ground such that the resistance between Pin 1 and safety earth is on the order of meg-ohms. Generically the standard calls out for 22M which is probably unreasonably high. To do this costs money. It means the DMX transceiver is likely to need its own isolating power supply internally. It's one of the differentiating factors between cheap LED fixtures and good ones.
It's also one of the justifications for using opto-splitters. Install the opto-splitter at the dimmer end of the 80m run and split off to your fixtures from there.
Bob Goddard/Goddard Design Co.
 

Users who are viewing this thread

Back