OT - Strange GFCI Behavior - Deadly!

[DerbyDad03]

Just to save everyone some typing, I know a GFCI works and I've installed dozens of them over time, both as standalone devices and as protection for downstream receptacles. Line vs. load, Pre-2006 devices, Post-2006 devices, etc.

Anybody ever heard of a case where the GFCI acts like this:

Press the test button and the GFCI "trips" - the button pops out and the downstream receptacles go dead. However, the GFCI itself is still live.

A tester shows that the GFCI is wired correctly, at least as far as hot, neutral and ground are concerned. I can only assume that something is wrong internal to the GFCI but if that's the case, that's a really dangerous failure mode.

The homeowner doesn't know when the device was installed and isn't a family member or close enough friend for me to start messing with wiring in a 1940's era house. I shut off the breaker and told her to call an electrician.

I've seen GFCI's fail, but not like that.

[Clare Snyder]

Line and load switched???? Power to the downstreem and downstreem cvircuit to the "line" side??? That's what it sounds like to me.

[Puckdropper]

DerbyDad03 [e-mail] wrote in news: [e-mail]:

I saw someone mention that on a home improvement TV show. They said when the line and load were swapped, the GFCI would still shut off the downstream outlets but would leave the GFCI outlet hot.

Hopefully they've redesigned them now.

Puckdropper

[Leon]

I suspect an internal short.

[DerbyDad03]

I bench tested that set-up last night. With power coming into the load side, my GFCI won't even supply power to it's own receptacles. Won't set, etc. No power downstream either.

It's possible that the GFCI in question is pre-2006 and may work differently. The one I tested is definitely post-2006.

I removed the switch plate at the neighbor's house and the box is jammed packed with the GFCI, a dimmer switch and a bunch of 12g wires. I couldn't get the GFCI out more than 1/4" and I wasn't about to start removing the dimmer, yanking on wires, etc.

[DerbyDad03]

Internal Short = Dangerous Failure Mode :-O

[Clare Snyder]

Can't redesign them to do what they are not supposed to do. They MUST be properly installed to work properly

[DerbyDad03]

They were redesigned in mid-2006 to be made safer by failing in the "off" condition.

Yes, they must be wired correctly, but I've proven via benching testing that swapping line and load will not cause a post-2006 GFCI to act like the one I'm talking about. If the incoming power is placed on the load side, the device won't ever go live, therefore it can't "trip" when tested. You can't trip what can't be set.

I don't have a pre-2006 GFCI to test, so I don't know if mis-wiring an older GFCI would cause this specific symptom.

[Leon]

While this sounds unlikely, and yes I know now that you don't want to pull that mess out of the box, but is there a chance that the receptacle has two screws to hold two hot wires from different circuits? And yes it would not make sense for that to be relevant if the GFCI is working properly, I would think.

Where am I going with this? Some receptacles have a break off jumper that connects both bottom and top portions in the receptacle. So the top portion and bottom portion of the unit can be fed by separate circuits. And this would possibly only meet any code if the GFCI itself was out of range of a water hazard, basically it would be protecting other outlets but one of its positions would still be live from another circuit.

I'm probably speculating a lot here and if that is not the case the GFCI is likely toast for safety reasons.

[Joe Gwinn]

Just being packed that tight likely violates the electrical code as well.

Joe Gwinn

[DerbyDad03]

GFCI's do not have the bridge. There's no way to split a GFCI. Well, other than with a hacksaw.

BTW another common reason to remove the bridge is when a switch is involved. You can switch the lower receptacle and leave the top one live, even on the same circuit.

Pop Quiz: Why do I say "switch the lower"?

[Clare Snyder]

Because that is the standard - lower "lamp" side switched - upper "live" unswitched.

[DerbyDad03]

Assuming there is an actual "standard" involved here, there is usually a reason that a standard is written/accepted.

What is that reason?

[Joe Gwinn]

The National Electrical Code is the place to start.

Joe Gwinn

[DerbyDad03]

Help me out. What NEC article describes which half (top or bottom) of a split receptacle should be switched?

[Joe Gwinn]

I don't offhand know, so some research is in order. Older versions of NEC are available gratis online. As are training courses. Electricians are supposed to know these things, so one can ask them as well.

The other possibility is NEMA's standards on plugs and outlets, although I don't recall seeing that.

Joe Gwinn

[Clare Snyder]

if your going to plug in a table lamp, put the switched at the bottom. That way the constant hot is easier to access to plug in other devices. Many electricians also install half-switched (and even full switched) outlets upside-down so they are easily identified as switched outlets.

[DerbyDad03]

Ding Ding Ding - Give the man a cigar!

[DerbyDad03]

I've done the research and that's why I asked for the help.

I don't recall seeing the "bottom switched, top always hot" mentioned any where in the wiring methods section of the NEC. Nor do I recall seeing "left or right" when it comes to horizontal switched receptacles.

I'd be very surprised to learn that the NEC cares which receptacle of a duplex is switched. Why would they?

[Joe Gwinn]

Ahh. I was not saying that I knew it to be in the NEC, only that that would be the place for it to be were it mandated.

Joe Gwinn