Industrial ethernet shielding.

Status
Not open for further replies.
L

lquadros

Member
We have lots of equipment, integrated with robots, transfer, touchscreen, PLCs. These are all being run with Ethernet communication. We have been lately instructed, to isolate the shields from device ends of the link and keep the switch end shield connected. This instruction is based on the manual written by ODVA (open devicenet vendor association). The TIA ( telecommunication industries association) does not permit this method. The hardware and cable manufacturers follow TIA standards. So we have a mismatch.
I am looking forward to hear from anyone, if you have any input on this situation and what would be the best way to address this issue. There is a lot of talk about open shield avoids ground loops etc etc. But how to study a ground loop, to know if it actually effects the system before it is actually operational? How to choose a network to be shielded or unshielded?

Any feedback would be greatly appreciated.
 
The ODVA rules are based on being an industrial environment with high power equipment and wiring all around. The TIA rules are based on telephone standards and office environments.
 
Indusrial ethernet

Indusrial ethernet

Jraef said:
The ODVA rules are based on being an industrial environment with high power equipment and wiring all around. The TIA rules are based on telephone standards and office environments.
Click to expand...

TIA-1005 covers telecommunication standards for industrial premises which is also equivalent to ISO/IEC 24702. This covers industrial Ethernet.
 
OK TIA is Telecom standard which uses a specific Isolated Single Point Ground of Equipment topology. Fancy way of saying all equipment is Electrically Isolated from any building structural elements and electrical systems, and all equipment is bonded to a Single Point where no outside or inside currents are allowed to flow in the Ground Plane. For current to flow there must be a entry point, and exit point with a difference in potential between them. With only one connection point, no current can flow. In a SPG application you can bond each end of a shielded Enet cabling at both ends if you insist, but has no purpose and opens you up to the possibility of interference. In a true Isolated Ground Plane no Current will flow in the shields because there is no potential difference between each end of the cable. So in theory you can bond each end, but is not practiced because most Isolated Ground planes are corrupted. Some sparky at some point unfamiliar with Isolated Ground planes does not use isolation bushings on floor anchors, or allows conduit to touch concrete or equipment outside the protected Ground Plane, or installs a SG receptacle inside the equipment room. All that can be eliminated by leaving one end open.

Moro of this story is simple. Bond only one end of the shield to the Host side, not the peripheral side in any application, TIA or Industrial. It works in both Isolated and non-isolated environments. No equalizing currents can flow in an open circuit, thus no common mode noise can develop along the length of the shield and injected or radiated into signal conductors. The whole point of Ethernet and other balanced signal technologies is to eliminate ground referenced signals, and common mode noise. . Secondly if you must have absolutely no chance of interference, use an optical media for transmission and you could care less what if any signal ground topology is used. Optical is immune to electrical interference. There is a very good reason military and critical communication systems have all gone optical. It cannot be detected by the enemy (via radiated signals) and immune to interference. ;)

Hope that helps.
 
Last edited:
ethernet shielding

ethernet shielding

dereckbc said:
OK TIA is Telecom standard which uses a specific Isolated Single Point Ground of Equipment topology. Fancy way of saying all equipment is Electrically Isolated from any building structural elements and electrical systems, and all equipment is bonded to a Single Point where no outside or inside currents are allowed to flow in the Ground Plane. For current to flow there must be a entry point, and exit point with a difference in potential between them. With only one connection point, no current can flow. In a SPG application you can bond each end of a shielded Enet cabling at both ends if you insist, but has no purpose and opens you up to the possibility of interference. In a true Isolated Ground Plane no Current will flow in the shields because there is no potential difference between each end of the cable. So in theory you can bond each end, but is not practiced because most Isolated Ground planes are corrupted. Some sparky at some point unfamiliar with Isolated Ground planes does not use isolation bushings on floor anchors, or allows conduit to touch concrete or equipment outside the protected Ground Plane, or installs a SG receptacle inside the equipment room. All that can be eliminated by leaving one end open.

Moro of this story is simple. Bond only one end of the shield to the Host side, not the peripheral side in any application, TIA or Industrial. It works in both Isolated and non-isolated environments. No equalizing currents can flow in an open circuit, thus no common mode noise can develop along the length of the shield and injected or radiated into signal conductors. The whole point of Ethernet and other balanced signal technologies is to eliminate ground referenced signals, and common mode noise. . Secondly if you must have absolutely no chance of interference, use an optical media for transmission and you could care less what if any signal ground topology is used. Optical is immune to electrical interference. There is a very good reason military and critical communication systems have all gone optical. It cannot be detected by the enemy (via radiated signals) and immune to interference. ;)

Hope that helps.
Click to expand...

Thanks derekbc. But the STP cable manufacturers will not make cables with unshielded end on one side, nor they identify the cables as such. I think we may need to look into making field attachable connectors and identify them this will again add wiring errors etc.. By the way, any idea, on how to decide if the shielding is even needed? I have installed projects with UTP cables with no issues at all. Also we have used dnet, cclink, profibus, profinet, canbus, etc. so far with continuous shield.
 
In general, a twisted pair wiring method is fairly immune to noise, as the noise is introduced equally into each conductor. Shielding also can help but it is only bonded on one end.
So the question is, do your devices have a bonded or grounded RJ45 jack? If not, then you won't need a sheilded connector
 
tom baker said:
In general, a twisted pair wiring method is fairly immune to noise, as the noise is introduced equally into each conductor. Shielding also can help but it is only bonded on one end.
So the question is, do your devices have a bonded or grounded RJ45 jack? If not, then you won't need a sheilded connector
Click to expand...
The host device and field device connectors are bonded RJ45 jacks. But they are eventually grounded to system ground.
 
lquadros said:
The host device and field device connectors are bonded RJ45 jacks. But they are eventually grounded to system ground.
Click to expand...
You did not hear it from me, but the low down of using shielded RJ-45 Connectors is not for electrical purposes, but mechanical and upsell. The Jacks Metal Framing combined with the rigidity and tight Form Factor size of the Plug when mated makes for a very sturdy, secure, and tight fit which can withstand the rigors of a active patch panel from constant adding and removing jumpers. It helps prevent unnecessary outages from daily provisioning operations. It is also a nice pricey upsell.

It is a lot like the 5-Ohm Earth Ground Requirement myth. At least that had a purpose 40 years ago.
 
dereckbc said:
You did not hear it from me, but the low down of using shielded RJ-45 Connectors is not for electrical purposes, but mechanical and upsell. The Jacks Metal Framing combined with the rigidity and tight Form Factor size of the Plug when mated makes for a very sturdy, secure, and tight fit which can withstand the rigors of a active patch panel from constant adding and removing jumpers. It helps prevent unnecessary outages from daily provisioning operations. It is also a nice pricey upsell.

It is a lot like the 5-Ohm Earth Ground Requirement myth. At least that had a purpose 40 years ago.
Click to expand...

derekbc, I was answering tom baker above -"So the question is, do your devices have a bonded or grounded RJ45 jack?"
What i meant to say was, the jack structure is eventually connected to ground. This is what I think.
But after all this, at the end, I am not sure, how can we install a large project, with manufactured ethernet cables, if we are unable to find cables with one end shield and another end open? No one seems to make them and I am totally lost. I cannot find an answer. The only way is to make our own?
 
lquadros said:
derekbc, I was answering tom baker above -"So the question is, do your devices have a bonded or grounded RJ45 jack?"
What i meant to say was, the jack structure is eventually connected to ground. This is what I think.
But after all this, at the end, I am not sure, how can we install a large project, with manufactured ethernet cables, if we are unable to find cables with one end shield and another end open? No one seems to make them and I am totally lost. I cannot find an answer. The only way is to make our own?
Click to expand...

Perhaps I missed something, who is making you use Shielded cables? Who says thou shall use shielded cables? Get my hint?

Sounds like a factory Sales rep buddering his biscuit with up sell commissions. Other than mechanical, I cannot see where a shielded cables do anything with the exception of high energy RF nearby. Years ago in data and analog signalling protocols were Ground Referenced topology like RS-232 and countless other ground referenced signal architectures. It wreak havoc on signal transmission. Lead to using Isolated Single Point Ground designs which are expensive and extremely difficult to keep from being compromised. I made a career out of it.

Today all that ground referenced signal protocols is gone along with all the noise and interference problems it causes. All of those problems were engineered extinct. So what has changed that? Nothing. When shielded Ethernet came around was about mechanical strength in Cross-Connect Patch Panels in data rooms where Cross Connects are daily operations. We used the shielded cables as Patch Cords for Cross-Connects. The wear and tear of daily Cross Connect operations caused unnessecary outages becaus a cable be hung up and getting pulled out of the socket or breaking the plastic plugs. Electrical had nothing to do with it.

As a final comment any piece of equipment with Ethernet Ports have a signal isloation transformer to isolate any galvanic or metallic connection. Guess what the characteristic of any transformer has, even a dry 480 to 208/120 power transformer has? Even the most inexpensive power transformers give you a minimum of 60 dB of common mode noise reduction. That is a ratio of 1,000,000 to 1. A transformer made as a Isolation Transformer like those used in both power and signal have CMNR of 160 db or 10,000,000,000,000,000 to 1. Heck I do not even know what number that is. I cannot comprehend anything that large.

So what does some salesman say a Shielded cable can even come close to that. Maybe 10 Db or 10 to 1 reduction in a really long run in a noisy environment. I do no tknow what they are smoking. All this crazy stuff came up after they legalized Ganja in the USA
 
That's why we don't pull shielded cable unless you specifically ask for it.
To many "that's not how you ground it"
 
ethernet - open shield at load end

ethernet - open shield at load end

qcroanoke said:
That's why we don't pull shielded cable unless you specifically ask for it.
To many "that's not how you ground it"
Click to expand...

dereckbc, I agree with your comment. I also think that we here, need to first understand if shield is required at all. But a document from ODVA has been considered for our specification. This document advices to keep source end of the shield open. We are using TN-S grounding system with single point grounding. I have already isolated the ground on load end of the shield. But I have heard that a tingling is sensed during Ethernet cable disconnection. I have not investigated it yet. Keeping one end open may cause this problem?
 
Status
Not open for further replies.
Top