DMX on shielded CAT5; a shield query?

[RonHebbard]

If you're running DMX on shielded category cable, what are you doing with the shield?
Grounding it only at the source end?
Grounding it only at the receiving end?
Grounding it at both ends?
Ignoring it, insulating / isolating it at both ends?
I've done a little Googling finding various, inconsistent, answers?
The application is entirely within EMT and four x 24, immediately adjacent, Johnson Systems ICON dimmer racks.
The plan is to reuse the existing Belden 9841 from the booth to the racks as it's in perfect condition and has given zero problems for more than a decade. Connecting Belden's shielded pair is no problem. My only query is how to handle the shield on the very short lengths of category cable within the four immediately adjacent racks.
Why are we using shielded CAT cable in the first place? (You may be wondering.) Because it's what's on hand.
I'm leaning towards using Fleenor / Pathway's pairing / color code and grounding the shields at the source ends ONLY to avoid circulating currents / ground loops on the shields.
This run is dedicated to dimmers only and terminates at the last, fourth, rack. A second run of 9841 routes the board's second output to a 1x4 opto splitter in FOH then, via the 'loop out' on to a second 1x4 opto serving on stage electrics where this second 'loop out' is terminated.
Your thoughts please.
Toodleoo!
Ron Hebbard

 

[porkchop]

My understanding is that optimal shielding for data transmission is grounded at the source only, but I have done very little testing or looking into proving that idea.

 

[VRommel]

My understanding is that optimal shielding for data transmission is grounded at the source only, but I have done very little testing or looking into proving that idea.

Correct.

 

[Wheezy]

Agreed. Cable shield grounded at both ends is one source of ground loop. Terminate the source end only.

 

[RonHebbard]

When using using 8 conductor UTP for DMX, an 8 contact RJ45 is an OK connector.
When using 8 conductor STP, an Ethercon is excellent as it provides a 9th path permitting the shield to be bonded to metallic panel ground rather than having to connect the shield to DMX common which, although "common" is not "Ground" (Nor should it be. Ask the reputable manufacturers)
There are other RJ45 connectors intended specifically for use with shielded, 8 conductor STP, but, if you'll pardon the expression, they're pretty cheesy. (Sp?)
If I were to connect 8 conductor STP's shield to DMX common at the source end, (either end for that matter) I believe I'd be "contaminating" DMX common. (Equipment manufacturers are telling me this.) The further I've dug into my query, with more and more equipment manufacturers, the more I'm being told the grounding of 8 conductor / 4 pair STP's shield on any of a standard 8 contact connector's contacts is a bigger problem in the industry as a whole, not only when using theses cables and connectors for DMX. I'm not trying to change any quality manufacturers' products. I'd just like to see a pin-out detailing specifically where / how to handle the termination of 8 conductor / 4 pair STP's shield and I'd like to see this on a reputable site; USITT, Pathways, Doug Fleenor, Johnson Systems, Control Booth, etcetera. Maybe what I'm really trying to do here is change the world by embarrassing manufacturers of "value engineered" products into stepping up their game.
Can I get a discussion going here PLEASE?
Toodleoo!
Ron Hebbard.

 

[RickR]

That's what I've specified for STP. Ground loop/noise is irrelevant if it's not part of the data path, which the shield usually is, but is not if the shield and common are not connected.

Think of it as double shielding.

 

[RonHebbard]

That's what I've specified for STP. Ground loop/noise is irrelevant if it's not part of the data path, which the shield usually is, but is not if the shield and common are not connected.

Think of it as double shielding.

Thanks Rick. Exactly. Understood and agreed. How are you connecting STP's shield to ground when all a given manufacturer provides is a normal 8 contact connector?
Toodleoo!
Ron Hebbard

 

[Wheezy]

Ron, it sounds as if you are talking a home made solution. If that is the case, I would simply solder a solid conductor wire to the shield, then terminate that to any case screw of the rack or device, assuming the case is grounded.

 

[RonHebbard]

Ron, it sounds as if you are talking a home made solution. If that is the case, I would simply solder a solid conductor wire to the shield, then terminate that to any case screw of the rack or device, assuming the case is grounded.

Understood and thanks Wheez'. I'm not asking for a home-brew solution, as we've all got our approaches and, with the addition of a little Teflon tubing and heatshrink, yours would be one of mine. I'm just hoping, with sufficient support, it may be possible to push the "valued engineered" (cheap) manufacturers towards a 9 contact solution, such as Neutrik's ethercon and, if not the ethercon, at least they could add specific instructions for the termination of shields to their sites to reduce the number of individuals who equate 'DMX Common' with ground. (Which it isn't and shouldn't be.) I guess I'm just 'tilting at windmills' (Again) but that's me. Thanks again.
Toodleoo!
Ron Hebbard.

 

[Wheezy]

I see what you're getting at now. Would be nice.

 

[FMEng]

The way to avoid problems it to adhere to the carefully designed DMX standard. The more manufacturers stray from it, the more unreliable things will get. When followed strictly, the standard prevents ground loops. Using an Ethercon creates the chance for ground loops because the shell is used to make the shield connection. A group loop will result in contamination of the data, sometimes fatal and sometimes not. If you want to use shielded Cat 5, no problem but don't mess with oddball connectors, and don't split the pairs. Tie the shield to pin 1 at both ends of cable connectors, per the standard.

 

[RonHebbard]

The way to avoid problems it to adhere to the carefully designed DMX standard. The more manufacturers stray from it, the more unreliable things will get. When followed strictly, the standard prevents ground loops. Using an Ethercon creates the chance for ground loops because the shell is used to make the shield connection. A group loop will result in contamination of the data, sometimes fatal and sometimes not. If you want to use shielded Cat 5, no problem but don't mess with oddball connectors, and don't split the pairs. Tie the shield to pin 1 at both ends of cable connectors, per the standard.

Perhaps, with sufficient support, we can force all manufacturers to adhere to the standard. Even manufacturers violating the standard by using XLR contacts 4 & 5 for 'non-standard' functions along with standard violators using XLR-3's. This would be an UP HILL battle but I can dream can't I?
Toodleoo!
Ron Hebbard.

 

[JD]

One thing about XLR connectors- Pin #1 is designed to make contact first. This is true on both 3 and 5 pin connectors. In audio, is is useful to reduce the "pop" that may occur if there is a stray charge on the line (static.) For the same reason, it is better to use the 5 pin XLR connectors on DMX runs. No reason to let a little stray static pop a transceiver and wreck your night!

 

[RonHebbard]

One thing about XLR connectors- Pin #1 is designed to make contact first. This is true on both 3 and 5 pin connectors. In audio, is is useful to reduce the "pop" that may occur if there is a stray charge on the line (static.) For the same reason, it is better to use the 5 pin XLR connectors on DMX runs. No reason to let a little stray static pop a transceiver and wreck your night!

Understood and agreed John; You wouldn't believe how many people say 'B.S. All the pins are the same length' without realizing the magic lies in the females where contact 1 is set back a little while all other contacts from 2 on up to 7 are set back just that little bit further such that contact 1 always makes first and breaks last. You're 'preaching to the choir and I THANK YOU for doing so.
Toodleoo!
Ron Hebbard.

 

[RickR]

How are you connecting STP's shield to ground when all a given manufacturer provides is a normal 8 contact connector?

There are RJ45 style connectors with a case ground. You can tell them by the metal surrounding the entire plug. Requires a patch bay with grounded jacks.

 

[RonHebbard]

There are RJ45 style connectors with a case ground. You can tell them by the metal surrounding the entire plug. Requires a patch bay with grounded jacks.

Understood Rick and thank you. This harkens back to FM's and John's posts regarding manufacturers deviating from the carefully engineered and clearly specified "standard" in terms of the specified connectors, pin-outs, cable types, et al.
The word "standard" implies an expectation of standardization that's not always adhered to. I guess I'm just being a naive idealist, tilting at windmills in hope of dragging the 'value engineered' back to the higher ground. (Please pardon the obvious pun.)
Toodleoo!
Ron Hebbard.

 

[MikeJ]

There are RJ45 style connectors with a case ground. You can tell them by the metal surrounding the entire plug. Requires a patch bay with grounded jacks.

Understood Rick and thank you. This harkens back to FM's and John's posts regarding manufacturers deviating from the carefully engineered and clearly specified "standard" in terms of the specified connectors, pin-outs, cable types, et al.
The word "standard" implies an expectation of standardization that's not always adhered to. I guess I'm just being a naive idealist, tilting at windmills in hope of dragging the 'value engineered' back to the higher ground. (Please pardon the obvious pun.)
Toodleoo!
Ron Hebbard.

Umm. I really don't understand what is going on in this thread. RJ connectors, as Rick has pointed out, are available specifically for shielded cat 5/6 cable. It IS an industry STANDARD, and has nothing to do with the connector being inside of an ether-con shell.

Take a look at the 3 versions of a "sneak snake" on TMB's website, if you want to understand more about DMX over cat5/6.

 

[RonHebbard]

Umm. I really don't understand what is going on in this thread. RJ connectors, as Rick has pointed out, are available specifically for shielded cat 5/6 cable. It IS an industry STANDARD, and has nothing to do with the connector being inside of an ether-con shell.

Take a look at the 3 versions of a "sneak snake" on TMB's website, if you want to understand more about DMX over cat5/6.

I believe what some of us geezers are recalling / lamenting is the original standard being 5 contact 'XLR style' connectors. I should 'shut up' now. (As if I shouldn't have shut up long ago. I guess candles and gas lights are gone forever.)
Toodleoo!
Ron Hebbard.

 

[FMEng]

Umm. I really don't understand what is going on in this thread. RJ connectors, as Rick has pointed out, are available specifically for shielded cat 5/6 cable. It IS an industry STANDARD, and has nothing to do with the connector being inside of an ether-con shell.

Take a look at the 3 versions of a "sneak snake" on TMB's website, if you want to understand more about DMX over cat5/6.

It boils down to this: The DMX specification is for receivers (inputs) to have a high impedance between signal common and ground. That is mechanically impossible to do with an Ethercon mounted on a metal chassis, because the Ethrercon shell makes an electrical connection from the cable shield to the metal jack body, and the metal jack body is connect to the chassis. A plastic jack body would solve the problem until the cable connector shell touches some other piece of metal, such as another light or piece of truss, and causes odd DMX issues. So yes, the type of connector used for DMX matters considerably.

 

[JD]

RJ45 is an Ethernet connector.
5 pin XLR is a DMX connector.

If you are running an Ethernet node system, great, use RJ45. Although both cables are about 110 ohms impedance, that is where the similarities end. In Ethernet adapters, you have a pair for "send" and a pair for "receive." Both signals are isolated electrically from the adapter circuit by a small hi-pot tested transformer. Ground does not matter short of a lightning strike.

If you are running DMX, the you have active (and fragile) electronic components tied right to pins 2 and 3. (Usually a transceiver chip) Although it is a differential one-way signal, the "common" conductor does play a role in preventing an excessive voltage spike from hitting that chip. Indeed, if there is enough noise between the electronic "common" on two units, (more that about 1.5 volts of noise) the transceiver will fail to be able to read the DMX signal. For this reason, the electrical "commons" of all units are tied together via pin 1. Do not confuse this with the frame ground which should not be connected! The 5 pin XLR connector achieves both needs by 1) making the common connection FIRST on insertion, therefore discharging stray voltage. and 2) by keeping the case ground isolated from the pin 1 common.

The use of shielded cable, with the shield tied to pin 1 at one end of the cable is not part of the protocol. Is it a good idea? Not part of the question. (My own opinion would be yes.)
So, can you use Cat5 or Cat6 cable on DMX? Of course.
Can you use RJ45 connectors and call it DMX? No.
Will it work? Yes.
Would I do it? No.
Will manufacturers at some point do it anyway? Yes.
The last "yes" is based on one simple fact- RJ45 is a lower cost alternative to 5 pin XLR. But, in my opinion any fixture that is running DMX protocol but is equipped with RJ45 should be called "DMX compatible" as it is not true DMX protocol.