GFCI with shocking result....

[meternerd]

I installed a couple of LED grow lights in my garage (built in 2013 to Code). They came with a switched connection cord, two prong polarized. They came with a jumper for connecting two lights to one cord. The jumper and the associated sockets were three prong. I hung the lights on a PVC rack, and plugged 'em in to the garage GFCI receptacle. Used to light a hydroponic herb garden...NO, NOT Marijuana..... Yesterday, I was leaning over the trays, standing on a damp concrete floor in my stocking feet and accidentally touched the metal housing of one of the lights with my forehead. I got the SNOT knocked out of me! The GFCI never tripped, polarity of the receptacle is correct and I haven't dissected the cord yet to see what went wrong. But how do you wire a three prong jumper safely to a two prong power cord. And, why did the GFCI not trip when I sent what felt like jiggawatts thru my skull? No UL label on the lights, but sold thru Amazon. Shouldn't the GFCI have tripped? I don't want to go into an in-depth investigation, but take this more as a warning. Don't be stupid and ground yourself on a bare concrete floor when working around electricity! I'm not a newbie....I should know better. I did run a jumper from the light housings to the receptacle ground and the problem disappeared.

 

[mopowr steve]

You may not have yet hit the threshold of 5ma that it should trip at, you may have gotten a 4ma shock. Ouch

 

[JFletcher]

I am having a hard time envisioning the extension cord and jumper set up... Could you post a picture?

If you ran a ground wire to an effective ground, and it did not trip the GFCI, then the leakage current is or should be less than 6 milliamps with a properly functioning GFCI receptacle. 5 milliamps can still be a painful shock if it hits you the wrong way.

GFCI devices do not need or care about a ground wire or how or if it is installed. Nevertheless, I would check the receptacle to make sure its polarity and line in/ line out wiring is correct, and obviously test and reset it.

 

[gar]

181205-2250 EST

meternerd:

GFCIs are not shock preventers. There just supposed to prevent a lethal shock. But if you are on a ladder, get shocked, and fall off, bust your head, then you may die.

.

 

[LarryFine]

GFCIs are not shock preventers.

In fact, you must receive a shock for them to trip.

 

[Adamjamma]

And, it could be a case of the shock you got was more from the same type of differences as sometimes happens with barns or pools... shock from the difference in ground potentials... the equipment is operating at a different ground potential than your concrete floor is. Perhaps the rebar 8n your garage floor was not bonded to the rebar in the foundations so has its own floating ground potential... which means you may want to put some rubber floor mats around the grow station.

 

[meternerd]

I am having a hard time envisioning the extension cord and jumper set up... Could you post a picture?

If you ran a ground wire to an effective ground, and it did not trip the GFCI, then the leakage current is or should be less than 6 milliamps with a properly functioning GFCI receptacle. 5 milliamps can still be a painful shock if it hits you the wrong way.

GFCI devices do not need or care about a ground wire or how or if it is installed. Nevertheless, I would check the receptacle to make sure its polarity and line in/ line out wiring is correct, and obviously test and reset it.

I tried a few pics but they really don't show much...too spread out. Just a two prong cord with an inline switch to the first light, then a three wire jumper to daisy chain more lights. I used a receptacle tester and the GFCI showed proper wiring and the test button tripped the GFCI. My surprise is that when I grounded the metal housings of the LED fixtures, there was no spark, no GFCI trip, but also no shock when I touched it. It gave a pretty healthy tingle when I touched it before. I haven't bothered to measure the voltage on the lamp housing to ground. I have a third set of regular CFL bulbs in a metal bathroom light fixture I bought at the thrift store for a couple of bucks. It's wired with a three prong cord with the metal housing bonded to the EGC. Supposedly, some of the herbs do better with white light instead of just the UV from the LED's. Growing like crazy, so I guess that's true. I just clipped a jumper from that frame to the LED frames. I suppose I could disassemble the LED fixtures to see how they're wired, but really not worth a lot of effort, just surprised. Thanks for the comments. Stay safe out there.

 

[ptonsparky]

Send that stuff over. I have Monday, Wednesday and Fridays off, unless I happen to go into work. I won’t be able to tell you why, but I’ll have fun and keep busy.

 

[junkhound]

Guess is that if you take one of the lights apart there is NO (or broken) internal connection of the metal housing to ground, as problem went away when you grounded the housing externally.

If that is the case, what zapped your forehead was likely static discharge vs. 60 Hz leakage. Static produced no 60 Hz differential current so no GFCI trip.

UV will charge an isolated conductor. IF a cold dry day where you are and the PVC is not wet, then the housing could charge to a few 10's of kilovolts.


Just yesterday was working on a 222nm UV light used to test UV degradation of plastics, cold and dry here yesterday in a non-esd controlled area. Me in rubber soled shoes. 10-15 seconds of UV light exposure charged me enough to cause static discharges when I touched a ground.

Thus a conjecture is that when you got zapped the lights had been on for a few minutes and the housing was well charged. When you hooked up the housing ground the housing had not been UV static charged.

 

[JPinVA]

If that is the case, what zapped your forehead was likely static discharge vs. 60 Hz leakage. Static produced no 60 Hz differential current so no GFCI trip..

Excellent observation!!

When you have a two prong cord, it's usually because the housing is double insulated. That eliminates the need for an EGC low impedance path to clear a fault between hot and (inner) housing because you have touch protection with the double insulation. But when that path is eliminated (by going with a two prong cord) the other benefit of the housing ground (bleeding off static build up) is no longer achieved.

I learned about this in my studies regarding machinery (usually belted machinery)...but never came across one like this...assuming that's what it is. But even if it's not the reason...it could be the reason...and that alone is a good teaching moment.

Now my mind is swirling....how much of this is traceable to the capacitors (or whatever) used in LED lighting? Could there be an engineering moment here. Any case studies out there regarding LED, double insulated shells, and static build up? Maybe a three prong should be part of the design of LED glow lights?

 

[meternerd]

Is U/L required....

Is U/L required....

I was under the impression that any electrical devices sold in the US required U/L testing. Is that really true? When I was in the power industry, U/L always showed up at the EUSERC meetings, so you could ask any questions you had. They had demos of fake U/L stickers, changes to U/L tests, etc. Even tours of U/L labs. Kinda miss that part of working for a living....but no, I won't give up retirement! I would have thought U/L would have rejected anything that can zap you, but maybe that's just the nature of LED fixtures. If I get really bored and decide to investigate further, I'll let y'all know what I find. I just don't want to deprive my new garden of light because it's really cool to watch things grow from nothing but water and a little plant food. Here's a pic if I did it right.....