Electric shock from my network cable

[Bobber]

Okay, this one's got me stumped.
Google returns no results for my search of "electric shock from a network cable" (nor if I exclude the word "a") so I really don't know what to do.

I've had my computer running on a wired network connection before, but have been running wirelessly for a couple of weeks since I've been missing a cable. I bought ten meters of "FTP 26AWGx4P CAT 5E patch cord cable" (pretty and yellow), but when I plug it into my computer and then the switch (from which two other computers are getting a working connection), the status lights don't even come on on either device.

Furthermore, when I plug the cable into my computer and touch the opposite end, I get a quite powerful electric shock - it's right on the border between a very powerful tingle and being somewhat painful. This happens both when the other end is unplugged and if I touch it while it's plugged into the router.

What can I do? Just buy a new cable? Did I in some strange way actually buy the wrong network cable?

Edit: damn, this is hardware, not the help desk. Can I please ask that it be moved?

[2.71828183]

Power over Ethernet is 48VDC nominal, which would deliver a shock as you describe. That being said, there's something wrong with the cable insulation if it's shocking you just by touching the ends. You should have to actually bridge two contacts to do that.

Ordinary ethernet produces a maximum of +/-2V, which is not enough to shock you.

It seems unlikely to me that a home device would support PoE, but I don't know anything about it other than what's in that Wikipedia article.

The other possibilities are that the network switch ports are floating above ground for some reason, which would shock you if you were grounded and made contact with the cable terminals, and that there's a serious problem with the network switch and it's supplying line voltage to the cable. Either way, though, what metal parts are you touching to get shocked? The contacts on the cable (at best difficult to touch accidentally) should be the only thing that carries any sort of power. Indeed, most of the cables I've seen are plastic on the outside.

[Bobber]

The ends look like this, and it's the metal parts of these that shock me.

Another computer is now available to me, so I tried plugging it in there. It still doesn't work, but at least it doesn't electrocute me.

Sounds like I got the wrong wire, or at least a seriously faulty one.

[Sero]

Woops. I was about to ask if it was possible it was static discharge, but this sounds like a persistent thing, I misread your original post. A rather amusing incident the other day wherein a classmate spent a few minutes investigating whether his earbuds had a short, before I suggested he might have just grounded himself on the metal mesh covering the speaker when he put them in his ears.

That is a decidedly odd looking ethernet cable, I do admit. I've never seen one with that sort of metal casing, though I suppose it might be for interference shielding, and otherwise be standard cable?

[Axman]

It does sound like static discharge because it takes something like 1500 volts to jump a couple of millimeters, and the jackets around the jacks are most definitely ground.

It's quite possible that you're building up static from the plenum insulator if it's a different material than your previous cable. Also, new carpet, new shoes, new chair, new air filters, all these cause static. Touch other things that you know are ground, like faucets. Do they spark?

Why do you keep touching it if there's a spark? Sparks are bad.

In any case, sounds like your computer's Ethernet adapter got toasted, unless it works with another port or cable.

Make sure the adapter's enabled in the software, though.

[2.71828183]

A static shock is a very different feeling from being shocked by medium-voltage DC or line current. For one thing, static shocks discharge once and don't keep happening, unless you're hooked up to a Van de Graaf generator or something.

If what you're getting was actually a continuous current, then the metal shield on the cable, which is supposed to be grounded, is at a different potential than earth ground, to the tune of several tens of volts. Is your house miswired (hot/neutral or hot/ground swapped)? If not, then there's something very seriously wrong with one of the devices the cable hooks up to.

[phillipsjk]

There is no way the cable (glorified phone wire), by itself, can give you an electric shock. Device hooked up to ethernet are supposed to be transformer isolated to the tune of 1000V AC or DC: that is why you can use the data wires for Power over Ethernet.

In theory, "Ethernet killers" should not work

Edit: After mentioning Power Over Ethernet, I am wondering if the same isolation ratings hold: probably not....

[section9_bateau]

the cable you bought is listed as FTP, or foil twisted pair. It is indeed a variant of shielded cable, and not generally recommended for home machines. It is very likely your router is NOT grounded if it is a SOHO one, rather it just has a +/- DC supply, and the voltage difference between that and a true ground, such as on your machine/machine's nic is causing enough current to flow that the nic is assuming the cable is damaged. (as many cheap devices will negativ eor positive ground their case)

What router are you using with? (I bet something with a transformer block). Swap the cable for one that is not between ungrounded devices on your lan, ground your router if possible (some dlink devices I know have a tab to connect a grounding wire to if desired, despite using transformers, etc.)

[zap360]

Hi Bobber,

possibly you've got a kind of grounding problem with your PC or your electrical installation.
You've got an FTP-cable, which is usually a good thing. That means your cable has a conductive shielding which electrically connects both of the metal cases of the LAN plugs. When you plug your LAN cable into your PC, the metal parts of your cable should have earth potential by being connected to the PE (protective earth [http://en.wikipedia.org/wiki/Earthing_system]) of your PC's mains socket. That means, that there should be no voltage between the bare metal parts and any other surface nearby.
Unfortunately, in some (faulty) electrical installations, separate rooms or sockets sometimes have different earth potentials, which causes a voltage (and a current flow) between them. In your case, between the bare metal parts of your PC (also the LAN cable), the LAN socket of your router and your body, when you're physically standing in another room. This causes strange sensations and possibly device malfunctions.

As an experiment, if you try to connect the router and the PC while they are powered via the very same mains socket, you should not get shocked.

Maybe you should also consider summoning an electrician to get your electrical installation checked. Faulty protective-earth lines can be dangerous.

[Bobber]

Thank you all for the answers, you were eight times more helpful than the bastards at the Icrontic forums.

My computer is indeed not grounded, and neither is the router. I'll just ground it and check again if it works. One moment.

EDIT: FUCK the electrician who decided that only two sockets in my house accept should three-pronged plugs (and thus be grounded), and that one should be in the bathroom and the other in the kitchen. That's Denmark for you - we don't care if our stuff is grounded. I had to run a wire from my computer to the heater instead. It produces very very tiny sparks when I touch the computer with that wire, so I'm positive that it's now grounded. I tried another cable between the same port on the router and the same port on my computer (I have two) and that works.

And the FTP cable STILL doesn't work, though it no longer shocks me, so the grounding definitely helped in that regard.

I give up. I'll buy another cable. Unless you guys have any further ideas?

[hintss]

is there a chance you accidentally got a crossover cable. forgot your first post, not paying attention to problems other than zapping

[phillipsjk]

In theory, anything using a "wall wart" (transformer block) does not need grounding. Such devices "float" such that if you come into contact with 220V, that becomes ground (the other wire goes to -220V, relative to you).

If you are using shielded cable, the shield is only supposed to be connected to ground at one end to prevent "ground loops." Even good wiring can have a significant potential difference (less than 12V?) over the length of the cable run. I would suggest unshielded cable to take advantage of the transformer isolation. Edit: If the cable is designed as I suggest, the shield connecting to the router should be the "floating" one because the router is also "floating". You can check if the shields are isolated from each other using a multimeter.

Your wiring sounds scary. I don't think the ground wire is supposed to carry enough current to "cause sparks." It is a safety feature. During normal operation, your device is supposed to use the neutral wire as ground. Is your heater properly grounded? It is not clear if you are talking about a steam or electric radiator.

[2.71828183]

The problem is almost definitely your house wiring. If your computer case produces sparks when it's connected to ground (the case is supposed to be grounded to earth via the power supply, not left floating), then the outlet it's plugged into has the hot and ground wires switched. This is a serious safety hazard, and I'm surprised you haven't noticed it before now (because grabbing the case in this state should shock you with line current if you're grounded at the same time).

Alternatively, if the outlet you're plugging into is unpolarized in addition to ungrounded (equipment with a grounded metal case shouldn't even have a plug compatible with unpolarized sockets, because that's a big safety hazard in its own right), flip the plug over and plug it in upside-down.

See, the deal with floating power supplies is, their local ground can float wherever it wants to in relation to earth ground (until you start getting into electrostatic effects, but that's beside the point), but there's no current flow there. Once you ground it to earth, ground is ground is ground. Current doesn't flow from one ground into the other, because there's no way for it to. There can be small differences between connected grounds enough to interfere with data signals, but again, current (and voltage too in this case) is very low. Sparks indicate to me that there's something that you think is ground that's actually a supply line, or powered from a supply line (maybe a high resistance connection or something).

Can you measure voltage or amperage between the two ground connections? Something still doesn't add up, because if you shorted a line straight to ground through a normal-sized wire, you'd be flowing lots of current and you'd definitely know it. But you shouldn't be getting sparks like that. I strongly suggest you get an electrician to look at your house wiring. I know there's something wrong with it, just from the fact that you apparently have a computer plugged into an outlet that's not grounded.

[hintss]

try painting the metal part of the connector on one end if you're not worried about house wiring

[phillipsjk]

See, the deal with floating power supplies is, their local ground can float wherever it wants to in relation to earth ground (until you start getting into electrostatic effects, but that's beside the point), but there's no current flow there. Once you ground it to earth, ground is ground is ground.

I suppose that implies my comment about switching the shielded cable around was incorrect. Shielding floating at ground minus (say) 12V (DC or even AC) will still work: Ethernet uses a high-pass filter. My concern was not with plugging the shielded cable into a properly-grounded case, but plugging it into another cheap device powered by a "wall wart" with a floating ground, possibly tied to one side of the transformer. If virtual ground A is at 12V and virtual ground B is at -5V, you can get conflict if they are tied together, then another port is grounded to actual ground. Without having a break in the shielding, your "virtual" ground may not be able to float anymore. Edit: Nevermind; with transformer isolation for the Ethernet signal, ground is ground is ground still applies (to the shielding).

Can you measure voltage or amperage between the two ground connections? Something still doesn't add up, because if you shorted a line straight to ground through a normal-sized wire, you'd be flowing lots of current and you'd definitely know it. But you shouldn't be getting sparks like that. I strongly suggest you get an electrician to look at your house wiring. I know there's something wrong with it, just from the fact that you apparently have a computer plugged into an outlet that's not grounded.

It is possible that OP did not scrape away paint before connecting the ground wire. Enamel paint acts as an insulator.

[Bic Finepoint]

All I can say about this one is if it shocks me and I don't expect it I investigate until I know why. If it shocks me and I expected it than shame on me. If I were you I'd spend some quality time with an electrician friend.

[2.71828183]

I suppose that implies my comment about switching the shielded cable around was incorrect. Shielding floating at ground minus (say) 12V (DC or even AC) will still work: Ethernet uses a high-pass filter. My concern was not with plugging the shielded cable into a properly-grounded case, but plugging it into another cheap device powered by a "wall wart" with a floating ground, possibly tied to one side of the transformer. If virtual ground A is at 12V and virtual ground B is at -5V, you can get conflict if they are tied together, then another port is grounded to actual ground. Without having a break in the shielding, your "virtual" ground may not be able to float anymore. Edit: Nevermind; with transformer isolation for the Ethernet signal, ground is ground is ground still applies (to the shielding).

The deal with ground loops (and things that aren't grounded but should be) is that they can pick up RFI and cause trouble. For example, just connecting a loose alligator clip to the probe of my oscilloscope and dangling it in the air gives me a rough 60Hz sine wave about half a volt P-P. That could be from wiring in the walls, from fluorescent light ballast, etc. Ground the wire and it goes away. The current passing through the grounded wire is utterly negligible (to give an indication of how negligible it is, just touching the end with a bare finger is enough to ground it, or at least to let it float in sync with what your body is doing at the time, which you can also see on the display), but the voltage when it's allowed to float is small but not negligible. And that's just a foot of wire. With the right frequency and sensitivity settings (assuming an oscilloscope sensitive enough), I'm sure I could pick up stuff like cell phone signals, radio broadcasts, and so forth. So, if you have a long floating wire that's intended to be earthed providing the reference ground for your Ethernet, it could very well be interfering with the signal. That's a separate issue from the sparking, though.

It is possible that OP did not scrape away paint before connecting the ground wire. Enamel paint acts as an insulator.

Insulator yes, but not a very good resistor. A visible spark indicates a circuit was made; since the plasma of the spark is much more conductive than air and paint, I'd expect it to pass a lot of amperage if it was a dead short from line to ground.

[phillipsjk]

I probably wasn't grounded but I have seen the opposite: oscilloscope shows dirty 60Hz when I touch the probe. To pick up cell phone signals you would need an oscilloscope designed for radio frequency work (converts the signal to a lower frequency): no oscilloscope (that I know of) has a bandwidth of 900Mhz.

If the shielding is floating at one end, the cable is exposed to that induced voltage only at the gap in the shielding. Therefore, interference from strong electric fields along the cable run is still minimized. Although, I heard a gap in shielding is designed to force any induced voltage to use equipment ground and wiring to get rid of the potential difference. All bets are off if everything is floating.

[zap360]

My computer is indeed not grounded, and neither is the router.

Well, if your Computer has got a metal case, it needs to be grounded. Inside your PC's power supply, there are capacitors between the case insulation and the power line. This is part of the line filter. The capacitor causes a quite high voltage on your PC case via Electrostatic induction. This is the cause of the sparks you noticed and also your shock. But, because of the high impedance, the current flow is low and usually not dangerous.
I don't know if this is an allowed installation in denmark, but you could install a 3 pronged outlet, where the PE (protective earth) conductor is connected to the N (neutral) conductor inside the outlet. This is, according to Wikipedia, called a TN-C earthing system. This should fix at least this problem.

And the FTP cable STILL doesn't work, though it no longer shocks me, so the grounding definitely helped in that regard.

Hmm, that's strange. Maybe you really bought a crossover cable. You can check the colours inside the plug inside the clear plastic. When aligning the plugs side-by-side (not face-to-face), the wires' colour sequence needs to be identical. If it differs, you've got a cross cable. You can also see that on the photo you previously posted.

[Bobber]

Thank you so much for all the answers and the incredibly thorough explanations!
I have to admit that this is becoming too advanced for me. I am genuinely trying to understand this problem but you are using a lot of terms that (you couldn't possibly know that) I don't know.

By floating, I take it that this means that the voltage difference between, say 5V and ground in the device is actually not 5V if measured between 5V in the device and the Earth's actual ground. Is this correct?

My computer does have a metal case. I connected it to a steam heater, and I did connect it to an unpainted brass part to make sure I wasn't getting problems from the paint insulating.

The thing about my computer really needing to be grounded is a big surprise to me. In Denmark, few outlets support the third prong. If you get a three-pronged plug and you only have a two-holed outlet, often here you'll just say "bah" and saw off the third prong. I don't think any of my computers the past fifteen years have been grounded, ever. To elaborate: the power supply in my computer has three prongs, and the power supply cable has three holes in one end which go into the computer, but only two prongs in the other end to go into the socket! And this is the cable that came with the power supply!

Our sockets aren't phased (that's what you called it right) so it doesn't matter how you turn the plug, i.e. turning it upside down would make no difference. About the crossover thing, I'll give that a look when I get home.

It seems like the best suggestion so far would be to paint the metal casing connected to the wire shielding on one end to see if that would help. But, I would again like to point out that grounding my case DID stop me from being shocked when I connected the cable to my computer and touched it.

I hope I've understood all this right, but again, I admit that I have a hard time following the explanation about floating ground and floating transformers and impedance and whatnot. All I know of impedance is that it's used for sound devices but I don't know what a higher or lower impedance would do to such a device.

I have one final question. I have a USB headset with a microphone. When the microphone is turned on, both myself and people on the other end of a Skype call hear a strange "knocking" static, it sounds pretty low frequency and could very possibly be 60 hz (which is what we use for wiring here in Denmark). Question: could these two problems (60 hz static, improper wiring) be related?

Edit: Uh, a third thing: when I play 3D games or watch a full-screen video in a Flash player, my entire computer (cursor, sound, everything) lags in perfect sync with what I once believed to me received packets on the network, but I now realize that it might be in sync with the same static knocking heard in my headset. This might be far-fetched, but could it be that I actually have a plethora of problems with my computer all caused by improper wiring?
Since it only happens when my graphics card is running in 3D mode, an explanation could be that it draws power in another way when it does this, which enables the problem. Again, I admit that this is a very far fetched theory that I am mostly pulling out of my ass, and it's probably just wishful thinking that I could solve all my computer problems of the past year or so in one swift move.

[Axman]

Well, if the increased power load of the GPU exceeds the power supply's power factor correction it could introduce an equally disruptive harmonic current (both of which will be fucked with by improper grounding) which would could introduce all kinds of problems. Getting a power supply with better PFC might help a great deal... but.

But I can't figure out how you would notice a 60Hz "lag"--it'd be practically constant. Here are test tones for audio equipment. 60Hz is a single, low-tone hum. A tick doesn't seem consistent with that frequency.

[Bobber]

I realize my explanation was ambiguous. It's not a continuous 60 hz lag, but rather intermittent periods of freezing. I didn't mean that the lag happened at this frequency, but just at the same time as when I heard the (maybe) 60 hz static, which comes in small "splots" of maybe 70 ms every 400 ms or so.

Edit: I've read this through a couple of times and my words seem inadequate. If only I could wave my hands around and make silly noises with my mouth to better explain what I mean!

[Bobber]

I took my computer apart and reassembled it. The only thing which seemed like it might cause bridging problems was a PCI card which I removed, but it didn't help. Here are some pictures.
As you can see, I've grounded the case pretty well, the power cable for my computer doesn't support grounding (two pronged, and we don't have a prong on any of our sockets here like in.. uh, France I think?), and the cable is not a crossover cable.

[phillipsjk]

"Floating" does not refer to the actual voltage: You can have "ground" (meaning power return) on a "floating" device equal to the voltage of actual ground. "Floating" simply means that voltage is free to move around. "Grounding" means you force the device's ground-plain to match Earth ground. In case it is not clear, I don't completely understand grounding issues, and was not really expecting you to either. You can get "weird" behavior if your "floating ground" drifts to unexpected voltages.

• Yes that is a straight-through cable. Modern equipment can auto-switch anyway.
• You can see that the plug you have is polarized. Notice one of the prongs is labeled 'N' for "neutral." I think the ground is supposed to contact a plate or prong protruding from your socket. Yes, that would be a french socket: http://en.wikipedia.org/wiki/AC_power_plugs_and_sockets You seem to be describing a type C socket.

Even though I said Ethernet is "supposed to" be transformer isolated, doesn't mean it is. Maybe you damaged the port. You can try a different Ethernet Card.

[hintss]

it dosen't seem like the uninsulated wire is touching the pipe...

[Bobber]

Okay, I admit the picture isn't as good as I'd thought it would be, but I promise you that it is indeed touching the pipe. At least I get loud static in my speakers when I touch my case with the wire from the heater, and I don't think I would if it were just a wire dangling into the air.

All sockets in my apartment look like this:

Thanks for all the help, but I'm just going to buy another cable.This is way way way too complicated.

[phillipsjk]

Short version:

Alternatively, if the outlet you're plugging into is unpolarized in addition to ungrounded (equipment with a grounded metal case shouldn't even have a plug compatible with unpolarized sockets, because that's a big safety hazard in its own right), flip the plug over and plug it in upside-down.

... or your neutral connection isn't as good as it should be. Hence the "consult an electrician suggestion": do different (non-polarized) sockets exhibit the same behaviour (with the same equipment)?

If you want a more secure ground connection to the case you can try loosening one of the screws holding the power supply in place and wrapping the wire around that (then re-tightening the screw).

[hintss]

how long is the exposed wire on the heater pipe end

[Bobber]

About a third of an inch. I laid it around the pipe in a J shape, then wound the wire around both the pipe and the wire end to hold it in place. I just removed some more of the insulation and fastened it even better (not in the crease as on the picture, but along the actual pipe) and I also fastened it differently to the computer case to allow for more surface area. But, lo and behold, it didn't help.

[Axman]

Wrap the exposed grounding wire around a case screw and screw it in tight.

[Bobber]

Just tried it, didn't work.
I'm still worried about the sparks I get when I touch my PC case with the grounding wire.
I must not have been thorough enough when I dissembled and reassembled it.

[m05am]

Hi guys..
i am facing the same problem however i am running a cable from the server to a beam saw machine and while the machine is on, without the cable being connected to the machine or server, recieve a quite shocking shock. i used CAT5 cable, a heavy duty cable as well as CAT6 cable. is there anyway of overcoming this problem? im not sure but perhaps its the static field around the machine which is causing the cable to shock, since it only happens while the machine is on..
any ideas on how to overcome this?
thanx..

[mosc]

People get obsessed with grounding. It's generally not necessary. I would go so far as to say for many sensitive analog electronics (any kind of amplifier which means a lot of signal receivers even when the signal itself is digital because sometimes it gets amplified before decoding) it can be harmful. Grounding is needed when you have devices that generate static. Computers should not be generating static. Three pronged connectors becoming more ubiquitous in homes has more to do with the wall an the outlet than what you plug into it. They are worried about houses burning down from substandard wiring or a lightning strike.

Can you describe the physical layout of the devices in question? Relative position? Power circuit between the two? It's hard to give you much of an answer. The only thing that comes to mind is that the beam saw machine may have sufficient reactance to ruin your AC power wave to devices on the same circuit. Electric isolation is the only thing that will help with that. Larger machines will have their own transformers to help with this. Massive ones can even have their own three phase power tap with accompanying adjustable capacitors to ensure they don't ruin the power waveform for those around them.