6

When I plug a high powered (1200W) device into my new GFCI outlet the GF circuit pops after 3 or 4 minutes. It does not pop if I plug the same device into a downstream outlet fed by the load terminals. It does not pop with smaller devices. This happens with a hair dryer and with a space heater.

Any ideas on how to diagnose this?

Some additional detail:

At first I blamed the hair dryer. I thought it might be faulty, or wet or something related to the additional GF breaker in its plug. But the problem also occurs with a space heater that has a plastic chassis, a two-pin plug without its own GFI, sitting on a ceramic tile floor with nobody touching it. It's hard to imagine how a current imbalance would occur with that one. The very strange thing is that it only occurs when the device is plugged directly into the GFCI outlet, not the downstream load ones, and it occurs reliably after a few minutes. Almost as if it's "heating up" although it's not hot to the touch. It resets and pops again quickly unless I let it rest a while.

It's brand new, an Eaton SGF20 on a 20A circuit. The hair dryer is 9A, the heater is 12. (I'm not using the at the same time obviously.) With smaller devices plugged in (maybe 1A) it runs forever.

Adding some test results further to comments in Harper's answer

Temperature scan of outlet before use is 78.8F :

enter image description here

After 6 minutes of use with a hair dryer and about 3 seconds after the GFCI trips, socket temp is 96.3. I attempted to measure the pin temperature of the plug, it was 88-ish but harder to capture with this device and only two hands. The body of the GFCI plug and the power cord were also in the mid to high eighties.

enter image description here

Note Within margin of error of my measurements we could have a 20 degree rise, which could be a trigger for the device's "self tests". Waiting for callback from Eaton.

jay613
  • 37,422
  • 2
  • 51
  • 148
  • 5
    That is very strange, indeed - bad contacts on the receptacle triggering some thermal overload we don't normally see? Is it the same for both sockets of the GFCI itself? I'd say give Eaton a call, it sounds defective. – Ecnerwal Dec 04 '22 at 16:46
  • 2
    Not to blame you, but did you do proper install using correct torque on the screws – Traveler Dec 04 '22 at 18:19
  • @Ruskes go ahead and blame me! I hope it's my fault, that will make it easier to fix. I did torque them, though now you call it out the instructions say 16 lb-in and I have a feeling I used 14. It's not obvious how that would cause this problem but I'll open it up, check the connections and correct them anyway. – jay613 Dec 04 '22 at 18:52
  • 1
    @Ruskes I tried blaming the user of the hair dryer. Can you imagine how THAT went? :) – jay613 Dec 04 '22 at 18:56
  • LOL...LOL... did you have to hide :) – Traveler Dec 04 '22 at 19:01
  • Not that it has to do with anything, but both cycle power on/off – Traveler Dec 04 '22 at 19:03
  • 2
    14 inch-pound is probably in-spec. – Harper - Reinstate Monica Dec 04 '22 at 22:29
  • @Ecnerwal looks like you may have made the right call with your initial comment, "defective". Identical replacement not tripping. I'll give it a month or so before calling that conclusively. – jay613 Dec 19 '22 at 15:13
  • 1
    @Ecnerwal your initial comment (10 minutes after I asked the question) is proven correct: It was defective. An identical unit from the same box is fine after a month of regular use. – jay613 Jan 25 '23 at 17:34

3 Answers3

6

I suspect the problem is thermal, at the plug.

Eaton is notorious for making products that do a bit more than their specifications say they must. For instance, Eaton's newest AFCI and GFCI breakers have the computer also monitor voltage (trip if over 140-ish volts as you get with a Lost Neutral) and over-current (allowing motors to start up, but acting sooner than the thermal trip would in the event of a stalled/stuck motor).

So I suspect your Eaton GFCI has thermal detection at the sockets. These are all the rage of late, and are in fact mandatory for plug-in EV chargers (e.g. the travel units supplied with EVs). Don't look at the temperature of the unit face, look at the temperature of the plug blades.

Or, shrug who knows? It could just be a mystery of science LOL.

Harper - Reinstate Monica
  • 300,628
  • 26
  • 286
  • 734
  • Rules-based thermal trip is plausible. Eaton's doc says it performs periodic self-tests, but not what they are. I'll call them Monday. If I don't get into a soul-sucking marathon of unknowledgeable hand-waving I'll post the results here. Otherwise I may have to go with @kreemoweet's advice and try a different brand or maybe just tell my wife to plug heating gizmos into a downstream outlet. – jay613 Dec 04 '22 at 23:26
  • 1
    I had a hunch it was a thermal trip at the GFCI - makes logical sense. I just didn't know that it is not a bug, it is a feature. You can even another plain receptacle right next to this one if the downstream (could be 4 feet away) receptacle is not convenient. Just pigtail with the other Load wires. – manassehkatz-Moving 2 Codidact Dec 05 '22 at 00:23
  • @manassehkatz-Moving2Codidact yup, two of the existing downstream ones are within reach already, one is inches away. If it turns out to be a "feature" it's a pretty dumb one considering a hair dryer in the bathroom is pretty much the poster child for ground fault protection. Except maybe pool lights. – jay613 Dec 05 '22 at 01:17
  • 1
    It may a little too sensitive. But to be honest, I have seen a lot of space heaters & hair dryers (worst by far is space heaters because they run for hours at a time) that seem to have cords no thicker than a lamp cord. Those cords and plugs often do heat up quite a bit - which could be avoided simply by upsizing the conductors. But except at the high end, these are commodity items and manufacturers are literally counting pennies. Put "We use 14 AWG wire in the power cord!" on the space heater box and it won't sell any more than without it. So the minimum UL/ETL requires is all you get. – manassehkatz-Moving 2 Codidact Dec 05 '22 at 01:24
  • 2
    I don't think it's intentionally thermal -- it could be that heat is shifting some threshold in the electronics and causing it to be excessively sensitive to a small leak as a result? – ThreePhaseEel Dec 05 '22 at 03:15
  • 1
    The space heater cord definitely heats up a little, exactly as Manasseh describes, but very little. Maybe 90F, guessing by touch. If the GFCI is supposed to work in a hot garage in the summer, say 120 or 130F ... that alone can't be what triggers it. Do all Eaton breakers fail in hot weather? I doubt it. There is something in a tech spec on their site about "thermal rise". I wonder if the self-check includes something like X degrees increase in Y minutes. I called Eaton, a helpful person whose job was NOT to get rid of me said he'd have a specialist call me back. – jay613 Dec 05 '22 at 20:09
  • IDK if the hair dryer cord heats up, I'll test that, but I'd be surprised if the hair dryer could transmit much heat into the plug pins THROUGH its own GFCI plug circuitry. It would have to be REALLY hot to do that, I think I would notice. – jay613 Dec 05 '22 at 20:10
  • 2
    @jay613 Hot sockets are very much a thing on plug-in travel EVSEs, and people find using the cheapie $12 14-50 sockets often results in their EVSE reducing power due to hot plug. The $80 Hubbell does not do that. But I would look closest at plug pin temperature. Because if this is getting too hot to touch, then take a close look at plug condition. – Harper - Reinstate Monica Dec 05 '22 at 22:02
  • 1
    @Harper-ReinstateMonica definitely not too hot to touch, I would have noticed that and made it my primary enquiry here rather than the GFCI trip. However see info I added to the question. Surface temp of the outlet rises 20 degrees in six minutes then the GFCI pops. This has to be my main line of inquiry now. Waiting for Eaton to call me back. – jay613 Dec 05 '22 at 23:52
  • If the GFCI heats up that bad but doesn't when using appliance on the LOAD connection, the only place where the issue can be is within the GFCI, between the connectors and the socket part. Definitely a "return/exchange" scenario – Jeffrey Dec 06 '22 at 00:01
  • @jay613 yeah I saw that about surface temp of the side of the outlet casing facing the wall. I'm not overly concerned about that, what matters is the temperature at the prongs. – Harper - Reinstate Monica Dec 06 '22 at 00:18
  • A couple of updates: 1) Eaton would not acknowledge that by design a rapid and sustained thermal rise would trip the GF circuit. They were helpful, spoke to their engineers, and came up blank. 2) Replacing the outlet with an identical one from the same box seems to have solved the problem. I'm going to wait a few more weeks before calling it. IF so the correct answer is somewhere between kreemoweets "they're quirky" and Ecnerwal's "it's defective". – jay613 Dec 19 '22 at 15:10
2

The actual problem in this case was a faulty GFCI device. Replacing it solved the problem, and the replacement was an identical one, in fact from the same box.

jay613
  • 37,422
  • 2
  • 51
  • 148
1

There's nothing to "diagnose". GFCI's have always been quirky, unreliable technology. They trip for endless, mysterious reasons with no connection to their ostensible function, including using them with higher loads, and just turning stuff on and off. Try a different brand.

kreemoweet
  • 3,529
  • 13
  • 13
  • Although frustrating, I am increasingly feeling this is the best answer. – jay613 Dec 05 '22 at 23:54
  • 2
    to be fair, between a regular socket that heats up and a GFCI socket that heats up and trip, I'll take the GFCI any time. Yeah, a regular socket that doesn't heat up is better, but still.... the diagnostic part did its job – Jeffrey Dec 06 '22 at 00:04
  • 1
    @Jeffrey I would like to think that a 20-degree rise in 20 minutes with ongoing increase would trigger some "let's stop this before it gets bad" reflex in the GFCI. I'm a fan of the attitude, "No, it REALLY IS a feature, you idiot, stop regarding it as a nuisance". The thing is, I phoned Eaton and asked them this very question and they did some research and couldn't come up with a "Yes, that's it". Not done yet, I'm going to do some tests they requested. – jay613 Dec 08 '22 at 17:15