Making Shielded Cables

[SgtBuddy]

Search Results matching the keyword shielded
No search results were found back.


I am a network engineer and I am looking for guidance on making shielded CAT5 cables. I build the cables just dandy, but would like the "official" directions on how to handle the shield. Some say only one end has to be grounded, some say both ends. Google has failed me (it does happen) and whatis.com helped very little.

Any others sites? Looking for actual documentation/reference so I can build a local training plan. I am going to find the RFC now and maybe head on over to the EIA/TIA site.

 

[ScottMac]

Shielded networking cables are generally a bad thing. Improper grounding or lack of grounding causes much more noise and interference than the shielding will ever keep out and performance is usually lower.

The panels for shielded cabling will have grounding lugs to clamp the shield/screen. The panels MUST be fully grounded, along with the rack/cabinet and other components. The EIA/TIA spec for grounding should be followed: in a nutshell, no less than 6Ga copper to the service facilities. A single point ground is recommended: since all panels (and rack & cabinets) tie back to the same ground: copper plates cascaded to the facilities ground, each aggregation point connects upstream with larger diameter cable. Both ends of the span are are grounded to the same place (no chance of ground/current loops).

So, both ends are grounded at the panels (panels designed for shielded cable), all grounds are tied through a "structured grounding system" terminating at the facilities ground.

You should also be using shielded jumpers, identified by the metal sleeve around the connector.

Shielded cabling is a PITA, I hope you had a good reason for choosing it. As mentioned above, failure to properly ground the system will reduce system performance. It must be grounded, and grounded properly - you shouldn't just "float" the shielding.

Good Luck

Scott

 

[Oaf357]

What my network guys go is cut the ground off from both ends. Seems to work just fine and I work for a huge worldwide organization. Now, if you had a device and connectors that allowed for a ground then I would definitely ground both ends.

 

[JackMDS]

Originally posted by: Oaf357
What my network guys go is cut the ground off from both ends. .

Is it still considered a shielded wire? Or may be there is another name for it.

 

[ScottMac]

No, it's not considered shielded cable at that point: it's considered a noise antenna with reduced performance for high-speed data.

If you're not going to use the shield, then you should have used UTP to begin with (aside from the fact utp usually less expensive, so you're wasting money).

BAD Things would happen to someone I caught clipping ground leads on any shielded cable in my neighborhood (not likely to happen, using shielded cabling for Ethernet is generally a stupid idea to begin with - it's a "special occasion cable").

IMHO, it's especially ignorant to remove the grounding (assuming it was any good to begin with) after you paid EXTRA for the cable, EXTRA for the panels, and EXTRA for the jumpers. Then you further reduce the cable's capabilitiesand neuter it down to less than Cat 4, maybe Cat3 specs. If the grounding is bad, the solution is to fix the grounding/shielding: leaving "floating" metal that close to your conductors just add lots of noise to the signal.

Happy Thanksgiving Jack! Your shirt will be on the way soon (sorry for the delay).

FWIW

Scott

 

[Oaf357]

Originally posted by: JackMDS

Originally posted by: Oaf357
What my network guys go is cut the ground off from both ends. .

Is it still considered a shielded wire? Or may be there is another name for it.

Yes, the shielding is the aluminum foil like paper inside. A ground cable exsisting in the insulation is simply that. An unused ground cable.

 

[Oaf357]

Originally posted by: ScottMac
No, it's not considered shielded cable at that point: it's considered a noise antenna with reduced performance for high-speed data.

If you're not going to use the shield, then you should have used UTP to begin with (aside from the fact utp usually less expensive, so you're wasting money).

BAD Things would happen to someone I caught clipping ground leads on any shielded cable in my neighborhood (not likely to happen, using shielded cabling for Ethernet is generally a stupid idea to begin with - it's a "special occasion cable").

IMHO, it's especially ignorant to remove the grounding (assuming it was any good to begin with) after you paid EXTRA for the cable, EXTRA for the panels, and EXTRA for the jumpers. Then you further reduce the cable's capabilitiesand neuter it down to less than Cat 4, maybe Cat3 specs. If the grounding is bad, the solution is to fix the grounding/shielding: leaving "floating" metal that close to your conductors just add lots of noise to the signal.

Happy Thanksgiving Jack! Your shirt will be on the way soon (sorry for the delay).

FWIW

Scott

Hmm... yet again I think we're talking about two entirely different things. I know what I'm talking about actually improved my connectivity at home and is being done all over the world at this very moment.

If your jacks, panels, and equipment allow for grounding then I highly recommend using it. However, clipping the ground wire (which actually sits outside the shield) at both ends will not increase interference in any way, shape, or form.

 

[ScottMac]

Yes, of course it does (affect the signals in the cabling) ...

That's exactly why TV antennas have a bunch of elements that aren't "driven:" some are reflectors, some are directors - they affect they way electromagnetic signals are collected by the elements that are connected. They are not grounded, they "float:" the electromagnetic field (the TV signals) are affected because of the way the field propagates in metal versus the way they propagate in free space.

Waves is waves, antennas are antennas. In the CABLING LAB I used to work in, this was tested: ungrounded shield or improperly grounded shield has a detrimental effect on the propagation of high-speed signals. That is also why UTP should be strung through the largest possible metal conduit: the conduit acts a loose shield and has an effect (minimal, but measurable) on the signal carrying characteristics of the UTP (a negative effect).

The fact(?) that you get data through does not mean that the cable is providing an optimum path, as witnessed by the multitueds of newbie home-cablers that don't use the correct pair order: it "works" (they get data through), but it's not working as designed.

Certainly, if you want to trash your cabling system, that's your right or privilege, but keep it to yourself, and stop encouraging people, by providing bad information, to waste their money on proven-bad "solutions."


Have a Safe & Happy Holiday

Scott

 

[Oaf357]

Originally posted by: ScottMac
Yes, of course it does (affect the signals in the cabling) ...

That's exactly why TV antennas have a bunch of elements that aren't "driven:" some are reflectors, some are directors - they affect they way electromagnetic signals are collected by the elements that are connected. They are not grounded, they "float:" the electromagnetic field (the TV signals) are affected because of the way the field propagates in metal versus the way they propagate in free space.

Waves is waves, antennas are antennas. In the CABLING LAB I used to work in, this was tested: ungrounded shield or improperly grounded shield has a detrimental effect on the propagation of high-speed signals. That is also why UTP should be strung through the largest possible metal conduit: the conduit acts a loose shield and has an effect (minimal, but measurable) on the signal carrying characteristics of the UTP (a negative effect).

The fact(?) that you get data through does not mean that the cable is providing an optimum path, as witnessed by the multitueds of newbie home-cablers that don't use the correct pair order: it "works" (they get data through), but it's not working as designed.

Certainly, if you want to trash your cabling system, that's your right or privilege, but keep it to yourself, and stop encouraging people, by providing bad information, to waste their money on proven-bad "solutions."


Have a Safe & Happy Holiday

Scott

Okay. But, I'll leave it up to you to tell the Department of Defense they're wrong.

Me, I'll sit here and enjoy my perfect ungrounded cable. When I get home I surf faster than typical UTP. Mean while, I think you're the one giving out bad advice. But I still don't think we're talking about the same thing.

 

[ScottMac]

They're wrong for using shielded? Or you're wrong for making unauthorized modifications to an installed government-owned infrastructure (if it was my network, I'd call it vandalism and can you in a heartbeat).

I talk to DOD organizations all the time, about infrastructure both active and passive (talked to an Air Force Base today).

The gub'ment don't just drop in cable or components, everything is engineered and planned - overly so if anything. Most folks I talk to (senior folks anyway) are pretty methodical, good networking people- especially the military guys. Oh yeah, there's some AD HOC stuff here & there, but the standing infrastructure is (with rare exception) pretty solid.

If you get caught butchering, you may end up in a lot of trouble. You might wanna quit before you're found out.

FWIW

Scott

 

[rectifire]

Shielded CAT5 is used in areas where EMI could be a problem. However the best thing to use in this kind of situation is fiber optic cable. The shield is also commonly called the "drain", since when it is properly hooked up, it effectively "drains" EMI to ground.

Improperly grounded or shielded CAT5 can cause earthing problems such as ground loops etc. Ground loops are currents induced in ground wires (or shields) by differences in ground potential between two different locations. This is why it is recommended to ground the shield at only one end of the cable, and have all grounds in the building eventually tied to one point.

Any ungrounded metal (think the antenna on your car, or the foil shielding of a Cat5 cable) is an antenna for picking up electromagnetic radiation. In cases where the shield is cut off and/or not properly grounded at either end, you have a situation worse than what you would have with just ordinary Cat5 cabling. The shield would now be a sizeable antenna with the sole purpose of inducing interference into your Cat5 cable.

Shielded cable works wonderfully, and drains most EMI to ground if installed correctly. If installed incorrectly, you end up with a product worse than what you started with. Think of it as a sink with a drain that is plugged up. If the water has nowhere to go (your shield is not properly hooked up), the water (or in our case, EMI) backs up and collects in the sink. This, of course, is a bad situation.

With the right test equipment (such as those used by Cat5 installers), one is able to test Cat5 cable for maximum attainable bandwidth and frequency. If you were to put a really sharp bend in a Cat5 cable, this device can actually let you see how the sharp bend detrimentally affects data carrying capacity. I guarantee you that a shielded Cat5 cable with the shields cut off at each end, would perform worse with this testing tool than an ordinary unshielded Cat5 cable.


Sorry for ranting. I just had to say something. I am a professional electrician, who is certified by Leviton in installing their Cat5 systems.

 

[Oaf357]

Originally posted by: ScottMac
They're wrong for using shielded? Or you're wrong for making unauthorized modifications to an installed government-owned infrastructure (if it was my network, I'd call it vandalism and can you in a heartbeat).

I talk to DOD organizations all the time, about infrastructure both active and passive (talked to an Air Force Base today).

The gub'ment don't just drop in cable or components, everything is engineered and planned - overly so if anything. Most folks I talk to (senior folks anyway) are pretty methodical, good networking people- especially the military guys. Oh yeah, there's some AD HOC stuff here & there, but the standing infrastructure is (with rare exception) pretty solid.

If you get caught butchering, you may end up in a lot of trouble. You might wanna quit before you're found out.

FWIW

Scott

Well considering you missed the entire point of my last post I'll reiterate.

The organization I'm in uses shielded cable. They simply clip the grounds off on both ends, crimp and patch. I don't understand how you think I go along and start clipping grounds, of course I probably shouldn't expect you to interpret things correctly considering it appears you have a kind of obstruction in your mind (anger maybe?). As for me vandalizing things, like I said you haven't interpreted my previous thread. Try re-reading it, it might help.

PS - They can't can me.

 

[NetWareHead]

Originally posted by: rectifire

This is why it is recommended to ground the shield at only one end of the cable, and have all grounds in the building eventually tied to one point.

Originally posted by: ScottMac

So, both ends are grounded at the panels

So which one is it? Ground at both ends...or just one end?

 

[Haden]

From my own experience it's better to ground only one end (if we are talking about long cable), reason was said by rectifire, It's usually
speed drop from 100Mb/s to ~5Mb/s because of double grounding.

Shielded cable is very good if you use it properly (it must be grounded), now bonus
is not only less noise, but also shield gathers and grounds voltage which is faradaiced from lightning strikes (far from cable of course).

 

[tgillitzr]

Originally posted by: Oaf357

Originally posted by: ScottMac
They're wrong for using shielded? Or you're wrong for making unauthorized modifications to an installed government-owned infrastructure (if it was my network, I'd call it vandalism and can you in a heartbeat).

I talk to DOD organizations all the time, about infrastructure both active and passive (talked to an Air Force Base today).

The gub'ment don't just drop in cable or components, everything is engineered and planned - overly so if anything. Most folks I talk to (senior folks anyway) are pretty methodical, good networking people- especially the military guys. Oh yeah, there's some AD HOC stuff here & there, but the standing infrastructure is (with rare exception) pretty solid.

If you get caught butchering, you may end up in a lot of trouble. You might wanna quit before you're found out.

FWIW

Scott

Well considering you missed the entire point of my last post I'll reiterate.

The organization I'm in uses shielded cable. They simply clip the grounds off on both ends, crimp and patch. I don't understand how you think I go along and start clipping grounds, of course I probably shouldn't expect you to interpret things correctly considering it appears you have a kind of obstruction in your mind (anger maybe?). As for me vandalizing things, like I said you haven't interpreted my previous thread. Try re-reading it, it might help.

PS - They can't can me.

So you are cutting of the RJ45 connector that has the shielded jumper (metal sleeve around the connector) as ScottMac called it, and replacing it with a standard RJ45 connector, leaving the sheilding unconnected to anything?

 

[Oaf357]

Originally posted by: tgillitzr

Originally posted by: Oaf357

Originally posted by: ScottMac
They're wrong for using shielded? Or you're wrong for making unauthorized modifications to an installed government-owned infrastructure (if it was my network, I'd call it vandalism and can you in a heartbeat).

I talk to DOD organizations all the time, about infrastructure both active and passive (talked to an Air Force Base today).

The gub'ment don't just drop in cable or components, everything is engineered and planned - overly so if anything. Most folks I talk to (senior folks anyway) are pretty methodical, good networking people- especially the military guys. Oh yeah, there's some AD HOC stuff here & there, but the standing infrastructure is (with rare exception) pretty solid.

If you get caught butchering, you may end up in a lot of trouble. You might wanna quit before you're found out.

FWIW

Scott

Well considering you missed the entire point of my last post I'll reiterate.

The organization I'm in uses shielded cable. They simply clip the grounds off on both ends, crimp and patch. I don't understand how you think I go along and start clipping grounds, of course I probably shouldn't expect you to interpret things correctly considering it appears you have a kind of obstruction in your mind (anger maybe?). As for me vandalizing things, like I said you haven't interpreted my previous thread. Try re-reading it, it might help.

PS - They can't can me.

So you are cutting of the RJ45 connector that has the shielded jumper (metal sleeve around the connector) as ScottMac called it, and replacing it with a standard RJ45 connector, leaving the sheilding unconnected to anything?

In my home I have standard plastic RJ45s crimped onto the outer insulation (you know what I'm talking about). The ground wire is clipped off on both ends. The shield runs from connector to connector. Hopefully that answers your question. How this creates an antenna like effect I don't know. All I'm trying to accomplish is limiting EMI and interference from power sources.

I've seen this done exactly like this in hundreds (if not thousands) of places.

 

[L3Guy]

In my home I have standard plastic RJ45s crimped onto the outer insulation (you know what I'm talking about). The ground wire is clipped off on both ends. The shield runs from connector to connector. Hopefully that answers your question. How this creates an antenna like effect I don't know. All I'm trying to accomplish is limiting EMI and interference from power sources.

Any wire that is not grounded has the potential to become an antenna. Shielded cable works because the outer, grounded shield forms a Faraday Cage. Basically, if the shield is grounded, the electric field is the same inside the shield, blocking external electro-magnetic interference. If the shield is not grounded, EMI will change the voltage of the shield relative to ground and data, potentially interfering with the data.

In general, I have rarely seen UTP have major interference issues. The only time I have seen fields large enough to interfere was at a Railroad Locomotive factory. You are trying it fix something that isn't broken. The problem is it probably works well enough so that you can't tell the difference. However, there are no guarantees that others will share your experience. I recommend that UTP is used in all but the most extraordinary situations, and that those few applications that require shielded twisted pair, that it is installed per specifications, using the recommended grounding.

Doug

 

[igiveup]

Basically what Scottmac and others are saying (if I may paraphrase), is that if you need to use shielded cable to begin with then that foil around the wires needs a path to discharge the electricity it collects (exactly like an antenna). If you don't have a path to ground then that collected signal will not discharge and will actually start building and interfering with your data signals, which is the exact opposite effect that should have happened if you used shielded UTP correctly. It may not be obvious outside of a lab, but there is a point where it becomes glaringly obvious. We all want to run at faster than 10Mbit don't we?

If you don't need shielding and can run the cable around any EMI producing areas then by all means go cheaper and use regular UTP. If not and its not grounded properly then its only a false sense of security, AT BEST. Most companies that know what they are doing are going to list specifications in their contracts they set forth. If they find you incorrectly terminated it later on then you can be held responsible for whatever ill effects they want to attribute to the poor signal.

DON'T do it. Not worth it in the least.

 

[Oaf357]

Okay then. Needless to day I'm not going to start tearing up the 3 to 20 foot runs of the stuff I have but let me ask this.

Where are you guys grounding to?

 

[ScottMac]

The best grounding point is the facilities ground (the same place your power is grounded to). By providing ground from a single source, you greatly reduce the possibility of a ground loop.

Shielded cable should be grounded at both ends (ASSUMING that ALL panels are gounded to the same location, which is the EIA/TIA specification). Otherwise, devices plugged into the ungrounded end do not continue the shielding effect and frequently present an "impedence lump." If you use shielded cabling (usually a silly thing), you must shielded from one endpoint to the other.

A "single point ground" doesn't mean ground only one end of the cable, it means that all ends are grounded to the same place (facilities ground, per spec and code). RF transmission/broadband systems are frequently grounded only at one end because they are using COAX, not twisted pair. In an RF system, the single-ended ground reduces high-freq EMI/RFI. Coax distribution systems are also frequently used between locations (like, separate buildings, campuses, communities) where a grounding differential is likely. In these situations, a single-ended ground is more likely to be effective and reduces the chances of dangerous differences of ground potential and ground loops - but again - these are COAX-based systems, not UTP.

FWIW

Scott

 

[Oaf357]

So you take the ground wire and ground it to the patch panel then the patch panel is grounded to a rack, etc.?

Explain.

 

[ScottMac]

I'll try to find my copy of the spec to be sure (or talk to the Lab guys), but my recollection is that the racks (cabinets) are independently grounded (code, common sense, and convention): the panels have their own ground run. They have to be panels designed for shielded cable, so the contact for the grounding sleeve of the jumper is active (to continue the ground to the ultimate endpoint).

Cascaded grounds may be permissable (that's the part I have to look up): Most common is a series of grounding plates (3/8" copper of whatever dimension you need for all the connections) that are for the local connections. All the area grounding plates tie back to a central grounding plate, which is bonded to the system facilities ground.

The construction of the racks (cabinets) do not guarantee sufficient grounding contact through the mounting screws (the panel could be anodized or painted, the rack/cabinet may be a little oxidized, etc). So there's usually a grounding lug on the panel that gets a connection back to the central grounding system (the copper plates tied to fac. ground).

There is an EIA/TIA spec for a "structured grounding system," I believe it's part of 568A (and probably B, as amended).

If you want to see a picture of the grounding plate, there's one in the Switch picture on my site:

rack.scottmac.net

I had the grounding installed when they built the place. I have 4ga copper to facilities ground, with either 12 ga or 6ga copper grounding the networking equipment and transmission/ satellite/AV system. The Cisco rack and the Rittal Rack with the ATM gear are both also grounded through this system. Single point grounding is a good thing. If you use a water pipe, make sure you put a bypass strap across the water meter.

FWIW

Scott

 

[Oaf357]

So in a house where the ground is subpar at best if even there where would you ground?

Electrical outlets? That wouldn't be wise at all.

 

[tgillitzr]

Originally posted by: Oaf357
So in a house where the ground is subpar at best if even there where would you ground?

Electrical outlets? That wouldn't be wise at all.

He has stated numerous times, only in extreme cases would he use shielded cable. If you're not going to rerun the cabling with unsheilded cabling, I wouldn't worry about it.

 

[ScottMac]

That pretty much covers it. Thanks Tgillitzr.

Scott