Electronics > Projects, Designs, and Technical Stuff

Outer Foil Shield/Unshielded Twisted Pair (F/UTP Cable) GROUNDING

(1/2) > >>

mcinque:
Hello all,

at work we are having a discussion on the grounding of F/UTP (Outer Foil Shield/Unshielded Twisted Pair) cable, basically 4 twisted pairs wrapped in thin aluminium tape.

Some say that this type of cable absolutely must be grounded at both ends, with shielded plugs and sockets.

Others say that grounding only one end is ideal, otherwise ground loops can be inserted, which would make things worse.

Others say that shielding with aluminium is not "true" shielding but something in between: it reduces interference and emissions but cannot be considered a "real" shielding, so it should not be treated as such (what's more, it is also fragile and difficult to crimp into connectors).

I always referred to "shielded" when talking of cables with proper copper/alluminium braid AND eventually alluminium foil (not only foil).

All this doubts remains to some, even though a Fluke CableAnalyzer has certified that the cable's performance is perfect, even though it has only been grounded at one end.

Where is the truth? Some manufacturers speak of grounding at both ends, others say "depending on the case". Do you have a paper that can clarify this doubt? is there a specific standard?

Attached pictures of the cable inside the patch panel and the alluminium foil grounded (one side, at the rack).

Many thanks in advance

T3sl4co1l:
What is being shielded?

The analyzer isn't going to tell you anything, it's just going to see UTP like anything else.  The question is, what was the purpose of purchasing shielded cable in the first place if you're not going to use it?

Tim

TERRA Operative:
For those clamps in the image, I would be folding the shield back over the outer insulation and clamping over the whole lot.
How it's shown in the image will give a tenuous ground connection at best.

Also, shielded Ethernet cable is mostly unnessecary in all but high noise environments, industrial etc. But if you have it already, may as well use it.

I used it at home anyway just for fun, and have the shield connected from the network switch, through the grounded jacks in the rack all the way to the metallic Panduit keystone jacks on the wall.
For those connections where the shield connection continued to the grounded end device, I haven't noticed anything but perfect connection.

I'm sure there's a standard out there somewhere for grounding Ethernet cables, but my system works well enough that I haven't looked too hard to find it.

mcinque:

--- Quote from: T3sl4co1l on March 21, 2023, 11:31:08 am ---What is being shielded?

--- End quote ---
network cables for intranet


--- Quote from: T3sl4co1l on March 21, 2023, 11:31:08 am ---The analyzer isn't going to tell you anything, it's just going to see UTP like anything else.

--- End quote ---
The analyzer certifies network speed, and the result is signed by the company that installed the cables.

Question are:

1. both sides should be grounded?
2. just one side should be grounde to avoid ground loops?
3. aluminium foil IS a proper shielding in a cable?

T3sl4co1l:
The other side, what's being shielded out?

Both sides must be grounded to get full value of the shield, but ground loops may occur.  This is the tradeoff.  An AC ground (lots of bypass capacitors) is all that's necessary, which may alleviate some ground loop issues, but I don't know if this is a common item; more of an electronic design option.

If one or neither side is grounded, it's about equivalent to UTP.  You get no value from the shield, but it's no worse than UTP, which does pretty well in most cases.

Even wholly unconnected foil still serves as a shield to very local / intense / pointlike fields, which could otherwise couple into individual twists of the pair(s), corrupting the differential signal directly.  This is pretty minor, but it's still a little advantage.

The other common case, I think, is when cable is simply sourced with foil, whether that foil is desired/needed or not.  In that case, it's no worse to leave it floating.

Foil is inferior versus braid, or multiple layers of either kind, but it's all a matter of degree.  Foil might afford 20-40dB of shielding (under certain specified conditions), while braid does 60-80; with proper dress, bulkhead connectors and multiple layers, over 100dB of shielding is achievable.

20dB would be enough to extend, for example, a marginal link passing through ~30V of noise (typically enough to corrupt 100BASE-T Ethernet, I think), but not 100 or 300V.  The former might arise in an industrial setting, say in the general environment around unfiltered VFDs and switchgear, but not in direct contact with their wiring (especially the output (motor) cables); the latter might arise in contact with such motor cables.

If very low bit error rate is required, quite a bit of additional shielding may be necessary to afford transient immunity.  For example, relays/contactors switching inductive loads typically generate fast transients (10s of ns pulse width, lots of ringing) with amplitudes of several kV.  That environment might be reason to use foil + braid cable and improved connector grounding practice (like coaxial bulkhead connectors).

Tim

Navigation

[0] Message Index

[#] Next page