Shielded CAT5e

Greetings,

There appears to be a debate about weather or not to terminate BOTH ends of the shield of STP CAT5e cable.

Normally with signal cable, you only terminate one end to avoid ground loops.

I have had two "Industrial Ethernet" product vendors tell me that the shield must be tied to ground at both ends. Is this true?

Greg

Depends on the resistance between the two grounds.

If the resistance is very low then you can terminate on both ends. Your next question might be "what is considered very low? Im not really sure, maybe less than 3 ohms?

If there is any doubt, just ground one end. Make sure it is a good ground connection though, and make sure that your ethernet switch supports STP.

If you dont get a good connection, you may be introducing more noise by using STP.

Hope this helps

Depends on the resistance between the two grounds.

Click to expand...

It als depends on whether there is a high voltage measured between the two points. If there is 100 volts, then connecting a small shield wire at both ends will result in a high current flow, even if the resistance between the two grounds is low.

How many ends to ground depends on the frequency of the noise in the area. The higher the frequency (above 60 Hz), the more reason to connect both ends of shield wires.

In plants where there have been two or three expansions, often the grounding is not equal in each of the sections. The Ufer concrete-encased electrode is by far the best grounding system. If all you have is a grounding rod, then in many locations you don't have an effective low-resistance grounding point.

In any case shielded connector should be on one end only to prevent ground loops, you don't want ANY current through the shield

What end? According to N-Tron's presentation shield should be on switch side because they have designed switch this way.

So I always tell to our customers: shielded cable end at the switch side, unshielded - device.

The general 'rule of thumb' for shielding is "Analog Signals (including encoders) - One End", "Communications cabling - Both Ends".

I generally disagree with the "Communications Cabling - Both Ends", unless I can absolutely control the source and target. Example, Comms cables between drives in a single enclosure, where all are bonded to a proper ground buss, I do ground both ends.

That being said, er, well, we have literally miles of CAT5/CAT6 UNshielded cable in our plants, tieing together a pretty large LAN. I haven't seen a single incident in the two years that I've been monitoring the LAN that a single packet was ever dropped due to noise.

Regular twisted-pair ethernet cabling, of CAT5 and above, is fully capable of rejecting just about all noise, as it appears as common-mode, on both conductors at a time, and is just ignored by the differential line drivers in any device.

We even have quite a number of relatively long runs of cabling, just tossed in the ceiling ("Temporary - meaning we'll fix it sometime in the next decade... Maybe). No noise issues there either.

I suppose, somewhere, STP Cable, STP connectors, STP NIC's, and STP Switches might justify the cost over UTP, but I haven't found that somewhere yet.

I once did the programming for a project where a similar question came up: how/where to ground the shield wire on a VFD (variable frequency drive) cable.

The electrical designer on the project explained it to me like this:

For an analog signal, you want to keep the noise OUT of the cable - so ground one end of the shield to avoid ground loops, etc.

For a VFD cable, you're more interested in keeping the noise IN - so ground both ends. (obviously this doesn't address voltage differences between grounds, but that's really a separate issue anyway)

Seemed to make sense!

I have never tried this, but may someday to see what happens, though I have a pretty good idea.

Put a .01 cap at the "other end", or perhaps at both ends.

Those who have never grounded the shields at both ends should sometime for fun, in a noisey or erratic (not erotic) environment.

Troubleshooting ground loops or why the "cable" is hot to the touch can be amusing and/or educational.

In my opinion, signal shielding should not be grounded on both ends. I have seen as much as 30VAC and almost as much DC between earth grounds only 100 feet away from each other. Signal shields are not meant to conduct current and grounding both ends in such conditions would be troublesome.

On the other hand, as in VFD/motor leads, these are not signal grounds and are often accompanied by ground conductors of larger gauge wire, often the same gauge as the motor leads themselves. Then, a foil wrap or braided shield is added to enclose the cable. The cables are intended for significant current and must force the ground at the motor to the same potential as the ground at the drive. If that is not done, the drive will frequently fault on ground fault, current imbalance, or other nuisance faults. These are due to the motor flopping around at a different ground potential from the drive and is unacceptable for best fault protection. If this grounding can't be done, won't be done, or if the power supply is floating or high-resistance grounded (grimace), then the fault protection in the drive has to be turned off to avoid the nuisance trips. Not a good situation.

Earth ground is not earth ground in many places and that must be carefully reviewed when grounding sensitive equipment spaced a longer distance apart.