0

I moved into an older house with mostly ungrounded circuits. I do plan to upgrade the electrical system to install a proper ground, but I needed to get my workstation up and running right away for work, so I used a "cheater" plug to hook my power strip to the 2-prong outlet in my new office room.

I noticed that, with my "better" power strip (a large 12-outlet unit with USB charging ports and all that), I get mild shocks off my PC tower's metal casing. It's not static electricity, it's definitely AC voltage (has that tingle to it). But I noticed that things plugged into another cheapo power strip I have on a "cheater" do not cause shocks on the casing.

I could be completely wrong, but I suspect that the issue is that my bigger power strip has an LED that lights to indicate that the circuit ground is good. I wonder if what is happening is that when I touch my PC's case, which would be connected to the ground of the power strip via the power cable, the current (trying to) to light that LED is passing through me to ground. (The strip is under the desk, and I'm not really keen on deliberately causing a shock just to see if the LED flashes, and even then my body is probably adding enough resistance to prevent the LED from lighting anyway - I might just be feeling a very small amount of current, since it's not really painful, just "uncomfortable".)

My questions are: is my theory correct (is this why I get shocks when using a power strip with a ground LED but not when using a cheaper one without such an LED) and is there anything I can do about it other than get some cheaper power strips until I am able to rewire the system with a proper ground? I considered replacing the outlet with a GFCI, but that doesn't really give you a true ground, it just protects you from a ground fault - which might mean I would trip the GFCI every time I touch my PC case, which wouldn't be a good thing...

Improve this question
8
  • The power surge strips(not plain power strips) require a real ground to work. The GFCI protects you from getting cooked. It might be the LED or there might be a real ground fault on the case. Maybe place a UPS between the case and GFCI.
    – crip659
    Commented Jul 11, 2023 at 18:10
  • 3
    Your "better" power strip probably has some EMI filtering components, and these rely on there being a real ground connection. With no ground connection, some of the capacitors in that EMI filtering circuit will effectively put about 60V (or half your mains supply) onto the ground pin - that's where your tingle comes from.
    – brhans
    Commented Jul 11, 2023 at 18:21
  • 2
    Interesting theory. FYI a GFCI trips around 5 mA of current. The LED in the strip is probably configured to draw 5-10 mA normally, but with your body resistance in the circuit, it may not actually rise high enough to trip a GFCI. Perception threshold depends on many factors but is nominally around 1 mA. NEC is quite liberal about retrofitting ground - you can get some bare or green insulated THHN and string it to the nearest good electrical ground (there are some workmanship and wire size requirements, of course).
    – Greg Hill
    Commented Jul 11, 2023 at 18:22
  • @GregHill Yeah, I might do that for a quick solution. I should have access to at least one of the outlets from the basement, the tricky part would be re-running new wire from the first outlet on the circuit to all the rest of them. But even if just the first outlet is grounded that might be enough to get me by for now.
    – mngeek206
    Commented Jul 11, 2023 at 18:53
  • @brhans That makes sense - that's most likely what it is then. It seems kinda weird to me that GFCI is seen as a code-compliant alternative to wiring a ground - sticking 60V on the ground pin wouldn't be stopped by a GFCI, and as another poster said, if the current is low enough it wouldn't even trip the GFCI. Might just try to wire a retrofit ground to the first outlet on the circuit for now, alternative would just have to be using really cheap power strips with no surge protection/filtering.
    – mngeek206
    Commented Jul 11, 2023 at 18:54

2 Answers 2

Reset to default
0

Pretty much anything electronic these days uses switch mode power supplies. To prevent interference, these have suppression capacitors between the hot, neutral and ground lines.

Even in normal operation, these capacitors leak a small current to ground. If your outlet is properly grounded, that's not a problem. But with no ground, the cases of any electrical appliances can float at a voltage up to half the supply voltage. To make it worse, with every appliance you add to the power strip, that leakage current increases.

The only proper solution is to ground the outlet.

Improve this answer
2
  • Yeah, this seems to be the consensus also from more research I've done on this. Also seems a lot of power strips have EMI filtering caps that have the same effect of sticking about half mains voltage on the ground pin, which is confirmed when I tried a contact voltage tester and it read 55-60V. In the basement I can easily replace the circuit with grounded cabling back to the box, but the main level will be trickier - even if I can "steal" a ground with some THHN into the first outlet on a circuit, that won't fix all the outlets, only the first one.
    – mngeek206
    Commented Jul 24, 2023 at 17:10
  • Funnily enough this also explains why "power line networking" adapters insist you never plug them into power strips - those EMI filtering caps filter out the signals they use by shunting them to ground. Learn something new every day...? :D
    – mngeek206
    Commented Jul 24, 2023 at 17:12
1

The shocks would be happening with any power strip.

They are coming from ground faults within your PC or other equipment which is plugged into that circuit, and/or ground faults in other equipment sharing an "island of grounds".

An "island of grounds" is when you have a ground pin which is not actually connected to ground at the panel which is connected to ground rods, and then, you extend that to more than one outlet by extending wires in walls, using a grounded power strip, etc. The result is that you have a group of appliances whose grounds are connected to each other, yet not to the ground rods.

When you have an island of grounds, if any one appliance has a ground fault, it energizes every ground in the island.

So a likely possibility is that you have a faulting appliance of some kind. Test that with measuring instruments, not your body, unless you really need a Darwin award.

I could be completely wrong, but I suspect that the issue is that my bigger power strip has an LED that lights to indicate that the circuit ground is good.

To sell it legally here, it needs to be UL listed. UL listed devices include instructions and labeling. If it says "use only on grounded circuits" you're not following them, and the reason they would demand that is what I said above.

If I had to run that way, I'd use 2-prong extension cords to split out power without islanding the ground.

I considered replacing the outlet with a GFCI, but that doesn't really give you a true ground, it just protects you from a ground fault - which might mean I would trip the GFCI every time I touch my PC case, which wouldn't be a good thing...

I don't think "good thing" means what you think it means :)

Do not fall into some sense of immortality here. Just because you got a tingle instead of a killing shock, doesn't mean it's "magically" current-limited. Something else e.g. your shoes might have been the X-factor to keeping current in safe limits, and that could change at any time, and kill you. Really.

Improve this answer
5
  • Or it burns your house down with you in it while you sleep, given that the shock means at least one appliance is failing. It could easily be multiple appliances.
    – Nelson
    Commented Jul 12, 2023 at 1:01
  • I don't think anything is failing. Based on more research and other responses, I think it's filtering capacitors putting some very minimal current on the ground bus. I got a contact voltage tester with a digital readout, and touching the grounded case shows 55V, which would coincide with everything I've read about this. Since a GFCI won't actually provide a true ground, and GFCIs typically won't even trip below about 5mA and these leakage currents can be as low as 1mA, even a GFCI won't solve the issue. I'm surprised that an ungrounded GFCI is allowed under code given all this.
    – mngeek206
    Commented Jul 24, 2023 at 17:04
  • @fdmillion The science around electrical safety is well-understood, Yes, it's new to you, but somehow you're twisting it as you learn it, almost like "confirmation bias" toward a preferred outcome (which you've plainly stated: will not tolerate GFCI trips and will not fix defective gear.) That's the road to perdition, or to be more precise, death.
    – Harper - Reinstate Monica
    Commented Jul 24, 2023 at 18:36
  • @Harper-ReinstateMonica You're commenting on safety, but the point of this post is to understand why the shocks occur. You state "it would happen with any power strip", but the shocks do not happen with all power strips. A "cheap" power strip does not result in case shocks. I also plugged a few of the "better" power strips into an ungrounded outlet and read 55V on the ground pin with nothing attached. Of course the right solution is to get a proper ground, but I was intrigued that a GFCI with no ground is code compliant when you can get shocks from ungrounded GFCI + power filtering strip.
    – mngeek206
    Commented Aug 14, 2023 at 4:57
  • Also to clarify, there is no ground in the wall. The strip is connected via a cheater plug, so only the grounds of the connected PC, monitor, etc. would be connected. But as I said, connecting the strip to an ungrounded outlet and probing the ground pin gave 55V with nothing attached to the strip, so it's clear the strip is the source of the ground leakage. I can even further postulate this since I have seen IT-grade rack power supplies that explicitly say on them that a ground is required due to filtering caps.
    – mngeek206
    Commented Aug 14, 2023 at 5:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.