Grounding ethernet shield

[testtube44]

I'm trying to run shielded ethernet cable to my modem downstairs, and the project requires low RF noise. I don't know how to ground the shield, or if grounding of the shield is necessary at all. I think it's going to become an antenna, but grounding it would just transfer the electrical noise to the ground wires in my house. What should I do?

[T3sl4co1l]

Shield is returned to the RF ground of the device.

Galvanic ground is only relevant to safety, and irrelevant for noise purposes (indeed, often worse for noise purposes, considering things like ground loop).

If your project requires low noise but a shield will bring noise into it, you're screwed either way, because the Ethernet pair is made of conductive wire, shield or not, and it will act as an antenna regardless.  (Mind that Ethernet itself can be a source of noise, since it is digital data, after all.)  It would then seem, the project needs to be housed inside a full metal shield, with all shielded cables returned to that first, before penetrating the housing.

Tim

[JS]

The shield should be connected only to one end to prevent ground loops, you usually want the receiving device to have it connected to but not essential and ethernet goes either way so connect the shield to the easier end. If you can choose the better grounded one.

The problem with shielded wires is the capacitance from the signal to ground is higher, so longer runs bring new problems. Ethernet is ok for pretty long runs but likely not so ok for such long runs with shielded wires.

I never have an issue with ethernet wiring so far, I found myself in the first one not so long ago, I asked around and it was an old wire that was placed in a different place, I made replace the run and all good. Twisted pairs are pretty good rejecting noise, even better with the fully differential and balanced scheme used in ethernet, so the shielding is not as critical as it is in other signals like analog audio where balanced is not so balanced, differential is not so differential and environmental noise inside the band is much higher, like power lines running inside the band of interest.

JS

[testtube44]

I'm trying to keep the digital noise from the ethernet inside the ethernet cable. Is there no way to do this without extensive modification of the cable? I can't let the RF escape into the living space. And what levels of noise do escape, considering what JS said about the ethernet cables being differential?

[David Hess]

The problem crops up all the time in low frequency instrumentation applications where you want to keep outside noise from getting into the signal.  The usual solution is to only ground the shield on the receiving end where connecting both ends would just create a ground loop which would likely make things worse but like T3sl4co1l says, sometimes there is no good solution and you are screwed either way.

Ethernet has the great advantage of being galvanically isolated at both ends preventing common mode problems when the grounds are at different levels which can be a real problem over 100 meters or more.  While not a complete solution, this suggests tying the shield to chassis ground at both ends through an RF capacitor with a low reactance at the frequencies to be shielded and high reactance at power line frequencies which is easy enough.

If that does not work well enough, then there is a sure fire solution which is actually pretty economical; use Ethernet over optical cable like 1000BASE-SX and there will be no emission (and no susceptibility) at all.  An Ethernet to optical bridge which is essentially a two port switch can be used at either end with the existing wired Ethernet ports if necessary.

[T3sl4co1l]

The shield should be connected only to one end to prevent ground loops

Yes, let's go to the bother of buying shielded cable, and connecting up one end, then injecting exactly 100% of the noise we hoped to block, at the conveniently-opened end!

Yeah no, always ground both ends.  A galvanic ground isn't required, but an RF ground is.  If you have ground loop problems, use capacitors.  Multiple capacitors, stitching around the shield, to maintain RF shield integrity.

The problem with shielded wires is the capacitance from the signal to ground is higher, so longer runs bring new problems. Ethernet is ok for pretty long runs but likely not so ok for such long runs with shielded wires.

Uh?

More capacitance would imply a lower characteristic impedance.  But the manufacturers know this, so they increase the insulation diameter to compensate.

At best, velocity would be lower, due to the higher fraction of dielectric (less air space).  Loss might be higher by the same amount, or it can simply be made with better (bigger?) wire.

When you buy CAT6 or whatever cable, shielded or not, and its datasheet says so-and-so impedance, velocity factor and loss... you're getting what it says.  That's simply it.

I never have an issue with ethernet wiring so far, I found myself in the first one not so long ago, I asked around and it was an old wire that was placed in a different place, I made replace the run and all good. Twisted pairs are pretty good rejecting noise, even better with the fully differential and balanced scheme used in ethernet, so the shielding is not as critical as it is in other signals like analog audio where balanced is not so balanced, differential is not so differential and environmental noise inside the band is much higher, like power lines running inside the band of interest.

Ethernet itself is extremely robust.  OP appears to be more concerned about Ethernet radiating noise in a very sensitive environment.  Note that Ethernet transformers are only rated to 30 or 40dB worth of CMRR, which means, out of a ~1V signal level, a good ~10mV of noise might be expected, at unlucky frequencies (usually switching edges).  In practice, it's better than this -- as it must, because most commercial EMI/RFI regulations are in the single or sub-mV range.

Note that shielded Ethernet isn't particularly useful unless your entire lab space is shielded.  Ambient sources of noise: computers, power supplies, phones, commercial radio, etc., are all in the same ~mV range, and aren't coming in on wires.  There's only one way to block those.

Tim

[testtube44]

Do you mean that other RF sources would cause issues on the shield, as in bouncing? How bad would it really be if the cable is less than an inch thick? That's not really alot of surface area to reflect off of.

[testtube44]

I'm just wanting to keep the ethernet digital noise out of the living space, don't worry about other RF sources. I don't know how to set this up and I'm a noob at electronics in general.

[JS]

The shield should be connected only to one end to prevent ground loops

Yes, let's go to the bother of buying shielded cable, and connecting up one end, then injecting exactly 100% of the noise we hoped to block, at the conveniently-opened end!

Yeah no, always ground both ends.  A galvanic ground isn't required, but an RF ground is.  If you have ground loop problems, use capacitors.  Multiple capacitors, stitching around the shield, to maintain RF shield integrity.

So at half the wire length the noise doesn't come in?

Even at HF, if you have both ends connected any CM between the ends will flow through the shield and can become a problem. In this case, if there's a more sensitive end that's the one which should be shielded, but connecting both ends with a low impedance path at whatever frequency can bring a problem at such frequency. I understand the shield will be less effective at high freq if not connected but also will bring less problems in itself. There's a reason why 99.9% of ethernet is wired without shield, price could be one but I guess there's something else.

Uh?

More capacitance would imply a lower characteristic impedance.  But the manufacturers know this, so they increase the insulation diameter to compensate.

At best, velocity would be lower, due to the higher fraction of dielectric (less air space).  Loss might be higher by the same amount, or it can simply be made with better (bigger?) wire.

When you buy CAT6 or whatever cable, shielded or not, and its datasheet says so-and-so impedance, velocity factor and loss... you're getting what it says.  That's simply it.

Do you still belive in DS? 
I know there are some true behind them but in a direct comparison nature always win. If you use thicker conductors and insulators the twisted pair as less as twisted as the center of the cores are further apart, so what you win from the shielding you might be loosing somewhere else. I'm not saying the rough numbers in datasheets are always crap but there are several things they don't put in datasheets and can be a trap ("for young players")

Ethernet itself is extremely robust.  OP appears to be more concerned about Ethernet radiating noise in a very sensitive environment.  Note that Ethernet transformers are only rated to 30 or 40dB worth of CMRR, which means, out of a ~1V signal level, a good ~10mV of noise might be expected, at unlucky frequencies (usually switching edges).  In practice, it's better than this -- as it must, because most commercial EMI/RFI regulations are in the single or sub-mV range.

Note that shielded Ethernet isn't particularly useful unless your entire lab space is shielded.  Ambient sources of noise: computers, power supplies, phones, commercial radio, etc., are all in the same ~mV range, and aren't coming in on wires.  There's only one way to block those.

Tim

  As suggested. twisted pairs for balanced, differential signals generate much lower radiation than a single wire would. The shielded labs I seen are usually composed for a faraday cage inside a bigger room, in which case the network wires could run outside as soon as they leave the instrument being the useful thing rather than the wire shield. For the short run from the shielded wall to the instrument a flexible tube could be the solution, as I've seen some post around building the lab (for the mains wiring) as detected to pick radiation from the back of the instruments, which is other source of noise in a sensitive environment and likely much higher than very short runs ethernet wiring, from the instrument to the wall.

JS

[Mr.B]

What "ethernet digital noise" are you concerned about and what do you think it will be affecting in your "living space"?
Does your living space contain some particularly sensitive test equipment?

[Brumby]

What "ethernet digital noise" are you concerned about and what do you think it will be affecting in your "living space"?
Does your living space contain some particularly sensitive test equipment?

This was the question buzzing around in my head, too.  As I understand it, twisted pair conductors with balanced currents tend to radiate very little.  I'm just wondering if the OP is trying to "get on top" of a problem that really doesn't exist in any practical sense.

[David Hess]

Just to be clear on the whole ground loop issue with shielded Ethernet, if both endpoints use the same power panel which will almost always be the case in a home or small commercial building, this is not a problem unless something else is already wrong.  Just use a shielded Ethernet cable with the shield connected at both ends and declare victory.

The economical alternative I mentioned is something like this:

https://www.amazon.com/Ethernet-Converter-single-mode-1000Base-LX-1000Base-Tx/dp/B06XZ6CV6W

[Kjelt]

Most if not all pro network equipment need a grounded powersupply and internally the ground is hard wired to the shield of the RJ45 connectors, at least that is what my stuff has. So all you really need to do is make sure the networkswitch and computer and/or other networkdevice are grounded eg connected to a grounded powersupply. Consumer network switches often have these isolated wallwart powersupplies, in that case you have to look what you will do to solve it. Check if the shield is connected through all sockets, if that is the case ground the two devices as stated above and leave the switch as is.

[David Hess]

I have a few small commercial network switches which use isolated 9 or 12 volt power inputs.  But they include a separate ground lug on their metal chassis for making the ground connection.

If shielded Ethernet patch cords are used between the network switch and devices, then this also connects the chassis grounds together but this should not be relied on when grounding is required for safety.

[coppercone2]

Does anyone want to redo their ethernet cables with really durable braid or possibly armor core and then use nice military connectors so it can all be nicely and safely grounded off floating supplies? God damn computer should be grounded through the ethernet. It's clearly all built completely wrong. Our treasure troves of high end test equipment are being compromised by cheap grounding solutions.

BTW won't the longer shield still reduce external interference, even if open at one end? Since there is less exposed conductor? Not that the noise levels in your LAN interface matter.

Could you also raise the impedance of the receiver circuit to reduce the pulse currents if everything is well shielded from external interference (then the standard allows?)

Considering the kind of insane conditions it works under, you might be able to tweak it, possibly using a amplifier at the source with lower current levels in the wire?

[Kjelt]

Does anyone want to redo their ethernet cables with really durable braid or possibly armor core and then use nice military connectors so it can all be nicely and safely grounded off floating supplies?

You put it jokingly as I understand it but I can tell you that for my garage where I have a IP controlled CNC machine I did upgrade to EtherCon plugs from Neutrik.
I don't know who designed the original RJ45 connector but it sure was not made to last in that environment without decent mechanical decoupling of plug and cable. It is actually toy grade, but it works if you don't touch the cables.
Another thing I have done three years ago is in my home lay a backbone from floor to attick of 10Gb optical cable. I had already three RJ45 cables running and needed another one and than thought enough is enough. The 10Gb switches were on sale back then for $200 a piece and in the meantime my PC and NAS have 10Gb NICS.

[coppercone2]

Well their like washing machines the volume of production is just completely INSANE, there is no reason not to put a ethernet jack on something or not to own a washing machine if you can possibly own one (who the FUCK goes to a laundry mat where your clothing can get stolen and you need to sit there for a fucking hour guarding shit in a humid lair filled with weirdos wearing over turned dirty wife beaters??)

No matter what processor, chinese resistors, etc you use, you still need to put the same ethernet jack on everything.. so some guy instantly got freaking high on the notion that saving 0.1 cents translates to what, millions? (even if there are billions to be made)

I think extreme mass production causes different psychological response to traditional accounting.

I mean USB is still pretty damn good compared to LAN and Phone. That little plastic tab is just.... . I wonder why USB is built so nut and bolts in comparison to the fucking LAN connector though. It boggles the mind.

[digsys]

1 more vote for - Ethernet to Fiber converters .. one one job, we use 100s kilometers - 1000+ runs, in very noisy / dangerous environments .. never ever had
an issue in 20+ yrs. (It's at a shipping port). Converters can be quite cheap, and made up fiber (or terminate your own) .. easy peasy.

[madires]

SFP ports prefered Try to avoid old school converters if possible.

[David Hess]

Does anyone want to redo their ethernet cables with really durable braid or possibly armor core and then use nice military connectors so it can all be nicely and safely grounded off floating supplies? God damn computer should be grounded through the ethernet. It's clearly all built completely wrong. Our treasure troves of high end test equipment are being compromised by cheap grounding solutions.

The big advantage of Ethernet is the galvanic isolation it provides so unless there is a specific reason to use them, it is better to avoid shielded cables.

Does anyone want to redo their ethernet cables with really durable braid or possibly armor core and then use nice military connectors so it can all be nicely and safely grounded off floating supplies?

You put it jokingly as I understand it but I can tell you that for my garage where I have a IP controlled CNC machine I did upgrade to EtherCon plugs from Neutrik.
I don't know who designed the original RJ45 connector but it sure was not made to last in that environment without decent mechanical decoupling of plug and cable. It is actually toy grade, but it works if you don't touch the cables.

I have seen industrial stuff where the RJ45 (and Molex) connectors were inside of a large screw-on connector housings.  Something similar can be used for RJ45s connectors in wash-down and marine environments

SFP ports prefered Try to avoid old school converters if possible.

I was surprised how inexpensive new network interface cards with SFP ports are but I suggested external Ethernet to fiber bridges because they will work with anything and the cost is comparable anyway.  The difficult part appears to be finding long pre-made fiber cables.

[testtube44]

So how balanced is it then? What kind of EMF would be measured, say 1 foot away from an active unshielded cable?

[testtube44]

Maybe I should just cave and get fiber and some cheap converters at the device ends. But what if it's too bulky having a piece of equipment like that hanging of the end of my computer and router?

[T3sl4co1l]

So how balanced is it then? What kind of EMF would be measured, say 1 foot away from an active unshielded cable?

I don't remember measuring above ambient, here in my own environment.  Which includes other potential sources like several computer PSUs, surprisingly.

Tim

[testtube44]

That sounds very promising, I'm planning on doing my own measurements anyways, and I will test a very long ethernet cable for any EMF to see for myself. But from what you're saying, it sounds like I won't have to worry about the electrodynamics of this project anymore.