Proper Grounding Techniques of Shielded Ethernet Cable

[cjf]

Hello,

I've recently taken on the daunting task of building out a make shift Data Closet on the second floor of my 120yr old home in an effort to better organize and accommodate my needs. Being an IT guy by trade does help in this effort in terms of how to hook up everything and make it work but there is one aspect of this project that has left me scratching my head no matter how much research I do on the topic due to so much conflicting information on the Internet.

The topic of this thread is meant to focus on how to deal with the shields of an Ethernet cable. Some will say that this kind of cable isnt really needed in a home environment while others disagree. I myself thought using a shielded cable made sense, at the time, but the more I read the more I find myself on the fence.

A couple of points to mention up front. This will be in a home environment that has X 1 primary electrical feed to a 200a breaker box in the basement. There is also X 1 main ground point involved from the breaker box to a ground rod in the earth. All AC receptacles involved are of the 3 Prong type.

So with that information out of the way I will mention what information I have gathered thus far and hopefully the folks here can chime in on what they know of this subject and why one method would be better than another.

So lets start with the hardware involved:
Ethernet switches = Ubiquiti US‑8‑150W POE ( 8 Port, 1GB, 150w POE Power, 2 SFP Slots for Fiber). These switches use a 3 Prong Mains cable. These switches use shielded ETH Ports.
Ethernet Cable = Ubiquiti "Tough Cable" Pro CAT5e Outdoor rated, Foil Shielded
Ethernet Connectors = Ubiquiti Shielded TOUGHCable Connectors
1 X Possible 12 Port Patch Panel (Shielded or Not)

Starting on the second floor and working our way down to the basement lets assume there are X 6 runs of the ETH cable already in place but without any terminations connected to the wire yet. The cable itself runs within 6-8 inches of several lengths of romex that feed various outlets on the first floor. Once the cable length reaches the basement it will then need to traverse other runs of Romex that feed various other items of the house that require electric.

From what I have gathered thus far about possible ways to handle the shields of this Ethernet cable run are as follows:

Switch to Switch or Switch to Device "Home Run" connection
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected. The theory here is that the source switches 3rd Prong will supply the ground for the shield of the ETH cable and by not connecting both ends of the shield potential ground loops should be eliminated.

Switch to Shielded Patch Panel
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected.
2. Ground the Patch Panel itself to earth ground
3. For the ETH cable leaving the Patch Panel, connect the shield at the Patch Panel End and leave the destination end shield unterminated/unconnected.

Switch to UN-Shielded Patch Panel
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected where it is punched down at the Patch Panel
2. Use an unshielded ETH cable for all destinations after the Patch Panel OR use the same cable with no shields connected at both ends.

In addition to the above there is also talk about how a shielded ETH cable could act as an antenna if it is not connected properly. Another thing I am not clear on about this is part is if you do not connect/terminate the shield at either end of a shielded cable (ie..remove it, cut it back, don't use it at all) and pretend it didn't exist could that shield be a potential antenna still or does this theory only apply if there is a physical connection to the shield itself?

Please discuss..Thanks

 

[amirm]

I have only heard of usage of shielded Ethernet cabling in industrial environment with large motors and such creating tons of EMI. I can't imagine any home having problem with standard cat6. I have miles of it in our home and I can run for months and months with zero errors.

Is there some motivation here I don't understand?

 

[RayDunzl]

I didn't work with Ethernet very much.

One problem we had was with a 50ft cable that was 35ft long on one end but only 15ft long on the other end.

 

[Blumlein 88]

I didn't work with Ethernet very much.

One problem we had was with a 50ft cable that was 35ft long on one end but only 15ft long on the other end.

Well that is worse than when your 50 ft cable is 25 feet long on both ends. No decisions need be made with them symmetrical.

 

[cjf]

Hello folks,

The initial reason for picking up shielded cable was because some devices being connected call for that type of cable to be used. In addition, there will be some runs of cable also extending outside the premises which in that case usually almost always calls for shielded cable to be used.

I'm on the fence about the interior cable though. If it turns out to be too much of a PITA to deal with the shields inside I wouldn't be opposed to just not connecting them but that assumes in doing so I wouldn't be introducing other issues like the Antenna effect I read about. I'm still unclear how an antenna will be created by an unconnected shield if indeed that is possible.

 

[Speedskater]

We have to remember that the Safety Ground/Protective Earth circuit is one thing, the ground rod connection to mother earth is another and a shield is yet a third thing. While there is continuity from one circuit to another, each has it's own function.
As to shields, all shields should attach to the chassis at the cable's connector (there are a few hybrid connections). Ethernet circuits operate at very high frequencies and AC power at very low frequencies, so closeness to a correctly wire Romex® cable should not be a concern.

 

[cjf]

Popular topic

..Thanks for the replies though

@Speedskater...It sounds like you are in the camp of attaching the shield at both ends of the cable then, correct? Any thoughts on how, if possible at all, an unattached shield on both ends of an ETH cable could result in an Antenna effect, picking up grunge in mid air from other devices,wires,wireless signals, radio waves...etc in the same vicinity?

In terms of the chassis being used as a grounding point, would that not result in the third prong of the mains cable from the switch coming into play which in a properly wired system find its way to the ground rod outside?

Thanks

 

[March Audio]

What you havent mentioned is the the relevance to your hifi.

Do you have an ethernet connected device as a dac?

Assuming you do, bottom line is that if you have your ethernet shield grounded somewhere, and it is ultimately connected to single ended audio components, you have the potential for ground loops and problems.

My understanding is ethernet data lines are galvanically isolated. In a domestic environment I would have no concern about ethernet data integrity, no need for cat 6. It has zero benefit in terms of audio and is more of a PITA to terminate.

At the end of the day, do what ever you like with the ethernet wiring, it wont affect the audio so long as the last link to your audio device doesnt have the shield connected.

 

[Jinjuku]

Here it is in a nutshell: You don't need shielded cable. Standard CAT 6 UTP is noise immune to 30Mhz. If you had the environment that required shielded CAT6 you wouldn't be able to hear your stereo due to all the industrial machinery.

 

[Speedskater]

Popular topic
..Thanks for the replies though
@Speedskater...It sounds like you are in the camp of attaching the shield at both ends of the cable then, correct? Any thoughts on how, if possible at all, an unattached shield on both ends of an ETH cable could result in an Antenna effect, picking up grunge in mid air from other devices,wires,wireless signals, radio waves...etc in the same vicinity?

I have never worked with shielded Ethernet cables so I can't help much.
But a shield not attached at either end will not act as an antenna.
A shield attached to a chassis at either end will not act as an antenna.
A shield attached to a chassis at both ends will not act as an antenna.
A shield attached to a to an internal point rather than the chassis will act as an antenna.

In terms of the chassis being used as a grounding point, would that not result in the third prong of the mains cable from the switch coming into play which in a properly wired system find its way to the ground rod outside?
Thanks

The ground rod has nothing to do with noise and interference problems. Mother earth does not act as a sink or sump for bad electricity.

 

[Jinjuku]

On shielding:

Floated shield prevents vampire loads (inductance) when you have a large bundle of CAT cables in a run from interfering with each other.

Shielding chassis side helps mitigate external noise from getting into the cable. This means shield is floated at the end point equipment and tied to ground bar at the rackmount.

These are the two most common techniques. Again you are in a home environment. Just use CAT6 UTP and if you do use CAT6 STP float the shield. Please heed the others and my advice you do not need STP cabling.

 

[cjf]

I appreciate the info folks.

One final question just so I'm clear @Jinjuku

When you say "float the shield" does that mean at one end of the cable or does this mean not attaching it at all at either end? I already own the spool of CAT5e which is run in wall from Point A to B in a bundle of 6 cables zip tied together. At this point since the run is already made I intend to just use it and am just looking at what to do about those shields already present in the cable.

As a side note, I have 12 strands of fiber also run in the same wall which I will be using as the Primary backbone connection between floors. The ETH cable was added for items requiring POE and as an additional convenience. The ETH cabling will be used in a much smaller capacity compared to the Fiber which by nature is immune to all this potential nastiness.

Thanks

 

[Jinjuku]

I appreciate the info folks.

One final question just so I'm clear @Jinjuku

When you say "float the shield" does that mean at one end of the cable or does this mean not attaching it at all at either end? I already own the spool of CAT5e which is run in wall from Point A to B in a bundle of 6 cables zip tied together. At this point since the run is already made I intend to just use it and am just looking at what to do about those shields already present in the cable.

As a side note, I have 12 strands of fiber also run in the same wall which I will be using as the Primary backbone connection between floors. The ETH cable was added for items requiring POE and as an additional convenience. The ETH cabling will be used in a much smaller capacity compared to the Fiber which by nature is immune to all this potential nastiness.

Thanks

Floated is no shield connected.

 

[Jinjuku]

I have never worked with shielded Ethernet cables so I can't help much.
But a shield not attached at either end will not act as an antenna.
A shield attached to a chassis at either end will not act as an antenna.
A shield attached to a chassis at both ends will not act as an antenna.
A shield attached to a to an internal point rather than the chassis will act as an antenna.


The ground rod has nothing to do with noise and interference problems. Mother earth does not act as a sink or sump for bad electricity.

Yep. Google "Seimons The Antenna Myth" if anyone is interested in reading more on this.

For perspective I have now tested 315 foot of CAT5e UTP against a $340 AQ Vodka 3 foot RJE, $380 WireWorld Starlight 12 foot CAT8 RJE, and a $699 Nordost Heimdall II 3 foot RJE.

In the case of the Nordost Heimdall 2 Ethernet cable review I put the 315 foot CAT5e directly underneath a running microwave and captured some tracks. No one could reliably tell any difference. We are talking $0.30/ft cabling into a $4000 Cary Audio DMS-500 and captured with a $1600 RME UFX.

 

[Speedskater]

Help me out, need more info to find:
"Seimons The Antenna Myth"

I found several antenna myth pages, but no Seimons.

 

[RayDunzl]

Maybe this one?

http://files.siemon.com/share-white_papers-pdf/13-08-01-screened-and-shielded-cabling.pdf

Links to more pages in the paper at top right...

(link edited, and may not be exactly the same as the original material)

 

[Jinjuku]

Help me out, need more info to find:
"Seimons The Antenna Myth"

I found several antenna myth pages, but no Seimons.

Sorry, I flip-flopped the e and the i. Siemons "The Antenna Myth"

 

[cjf]

Floated is no shield connected.

Thanks for the clarification

 

[cjf]

Maybe this one?

http://www.siemon.com/us/standards/Screened_and_Shielded_Guide_7_Antenna_Myth.asp

Links to more pages in the paper at top right...

Wow..that article explains and covers it all quite perfectly. It seems that my understanding of which cable could act as an "Antenna" was reversed. Interesting!

Thanks

 

[John99]

Hello,

I've recently taken on the daunting task of building out a make shift Data Closet on the second floor of my 120yr old home in an effort to better organize and accommodate my needs. Being an IT guy by trade does help in this effort in terms of how to hook up everything and make it work but there is one aspect of this project that has left me scratching my head no matter how much research I do on the topic due to so much conflicting information on the Internet.

The topic of this thread is meant to focus on how to deal with the shields of an Ethernet cable. Some will say that this kind of cable isnt really needed in a home environment while others disagree. I myself thought using a shielded cable made sense, at the time, but the more I read the more I find myself on the fence.

A couple of points to mention up front. This will be in a home environment that has X 1 primary electrical feed to a 200a breaker box in the basement. There is also X 1 main ground point involved from the breaker box to a ground rod in the earth. All AC receptacles involved are of the 3 Prong type.

So with that information out of the way I will mention what information I have gathered thus far and hopefully the folks here can chime in on what they know of this subject and why one method would be better than another.

So lets start with the hardware involved:
Ethernet switches = Ubiquiti US‑8‑150W POE ( 8 Port, 1GB, 150w POE Power, 2 SFP Slots for Fiber). These switches use a 3 Prong Mains cable. These switches use shielded ETH Ports.
Ethernet Cable = Ubiquiti "Tough Cable" Pro CAT5e Outdoor rated, Foil Shielded
Ethernet Connectors = Ubiquiti Shielded TOUGHCable Connectors
1 X Possible 12 Port Patch Panel (Shielded or Not)

Starting on the second floor and working our way down to the basement lets assume there are X 6 runs of the ETH cable already in place but without any terminations connected to the wire yet. The cable itself runs within 6-8 inches of several lengths of romex that feed various outlets on the first floor. Once the cable length reaches the basement it will then need to traverse other runs of Romex that feed various other items of the house that require electric.

From what I have gathered thus far about possible ways to handle the shields of this Ethernet cable run are as follows:

Switch to Switch or Switch to Device "Home Run" connection
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected. The theory here is that the source switches 3rd Prong will supply the ground for the shield of the ETH cable and by not connecting both ends of the shield potential ground loops should be eliminated.

Switch to Shielded Patch Panel
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected.
2. Ground the Patch Panel itself to earth ground
3. For the ETH cable leaving the Patch Panel, connect the shield at the Patch Panel End and leave the destination end shield unterminated/unconnected.

Switch to UN-Shielded Patch Panel
1. Connect the shield at the source switch end of the cable only. Leave the destination end of the cable shield unterminated/unconnected where it is punched down at the Patch Panel
2. Use an unshielded ETH cable for all destinations after the Patch Panel OR use the same cable with no shields connected at both ends.

In addition to the above there is also talk about how a shielded ETH cable could act as an antenna if it is not connected properly. Another thing I am not clear on about this is part is if you do not connect/terminate the shield at either end of a shielded cable (ie..remove it, cut it back, don't use it at all) and pretend it didn't exist could that shield be a potential antenna still or does this theory only apply if there is a physical connection to the shield itself?

Please discuss..Thanks

Hi
My thoughts on shielded ethernet cable. Ethernet is transformer coupled and so true isolation is achieved.

About the shield possibly acting as an antenna: It can. If you lookup triaxle cable, the inner shield is connected at one end and the outer shield is connected at both ends. The outer shield maintains the Faraday cage if you built a good one, and any current flow in the outer shield that is low enough in frequency to penetrate the shield is blocked by the inner shield that is connected at one end. This is a bulky solution and is not popular. The inner shield if it has a jacket, constructs a capacitor with the outer shield. So the solution provides good low frequency blocking. Most commercial equipment do not have Faraday cage construction. So energy that supposed to stay on the outside goes into the equipment negating the nice cable shield.

If the single ended connected shield lays on grounded cable trays then stray fields are shorted by the grounded tray and the antenna effect is minimized. EMI Labs typically require 2 meters of cables to be maintained 6cm above a ground plane for system testing.

Stopping low frequency current flow: Since each ethernet pair can carry 250Mhz, the end to end shield can be capacitively coupled on one end. If low frequency current flow (60Hz to 400Hz) is not to high, the transformer coupling blocks all low frequency components and capacitive coupling (maybe 200pf) may not be needed.
One author mentioned that unshielded ethernet cables rejects outside signals up to 30Mhz. Seems reasonable.

About myself: I am an Analog electrical engineer who also handles signal integrity issues.