1

Four years ago we installed a 50A 240V circuit to power a hot tub. There is a standard 50A breaker in the main service panel (a 1990's vintage 200A Murray panel), with 4 wires (L1, L2, neutral and ground) to an Eaton subpanel (located about 50 ft from the main panel) that has 2 circuits - a 240V 30A circuit that powers the hot tub's heater that runs intermittently, and a 120V 20A circuit that powers the rest of the hot tub, including a 40W circulation pump that runs constantly. Both of these subpanel circuits are protected by GFCI breakers. This installation has worked perfectly with no issues, up until now.

Recently we installed a new 60A 240V circuit from the main panel to power a Tesla wall charger (or EVSE),located about 25 ft from the main panel. As required in the installation instructions, this circuit has only 3 wires (L1, L2 and ground) that are directly hardwired from the main panel to the charger. The charger is set to operate at its highest rating, which is 48A constant load.

Now, whenever we are charging the car, both GFCI breakers in the hot tub subpanel will trip. None of the breakers in the main panel trip, nor does the car charger diagnostics report any problems - the car continues to charge at full power. It is worth mentioning that the GFCI breakers do not trip right away, but several minutes to more than a half hour after the car starts charging. If I attempt to reset the breakers while the car is still charging, however, the breakers trip immediately. But once we stop charging the car, the GFCI breakers can be reset and those circuits function normally until we attempt to use the Tesla charger again.

Any clue as to what's triggering the GFCIs to trip after a time during the charging session and how to address it? I don't know of any possible connection between the two circuits but I have yet to open the main panel - I did try derating the Tesla charger to operate at 40A as a test to see if a lower power draw would solve the problem, but the hot tub GFCI breakers still will trip after a time.

Update 7/1: After further examination of the wiring did not reveal any obvious issues, as a test I turned off all breakers in the main panel except for the charger, spa, and a plug circuit that powers my wifi so I could control the car charging remotely. I also turned down the hot tub temp to prevent the spa heater from turning on during a charging session:

  1. Charging the car at 40A trips the breakers within an hour.

  2. When charging the car at 40A, and while the spa breakers were on, I tripped the breakers manually with the test button; the breakers then tripped immediately upon attempting to reset.

  3. I then ran a series of tests similar to 2) but with the car charger drawing different amounts of current. I discovered that the 30A breaker cannot be reset if the car is drawing more than 18A, and the 20A breaker fails to reset if the car is drawing more than 22A. Below 19A I can manually trip and reset the breakers without issue while the car is charging, so I suspect that charging the car above 18A will eventually cause the tripping problem I'm observing at higher amperage.

dponti
  • 11
  • 4
  • Is the subpanel properly installed? Ground and neutral should not be bonded in the panel. The Tesla install only uses the two hot wires and ground wire, no neutral, so whatever is going on is most likely to be related to the neutral and ground in the subpanel. – KMJ Jun 15 '23 at 20:05
  • I will check, thank you. This should be the case for both the 50A circuit connection to the main panel as well as in the subpanel (Eg. ground and neutral must be separated in both panels)? If the subpanel was improperly installed (4 years ago) shouldn't this issue have shown up prior to installation of the Tesla circuit and if not why now? Are there conditions where GFCIs might not trip even where ground and neutral are bonded? Just trying to understand what is going on. – dponti Jun 15 '23 at 20:47
  • 2
    Main panel gets a bonded neutral and ground, every other device and panel hanging off the main panel does not get them bonded. If they're bonded somewhere else that can cause you trouble. If there's a loose connection on one of the subpanel feeders that can cause you trouble as well. I would validate those fundamentals before looking for anything weirder. – KMJ Jun 15 '23 at 20:51
  • 1
    Are the GFCI/breakers tripping due to GFCI fault or overcurrent? There should be an indicator of some sort to tell you the type of trip, which is very important here. – manassehkatz-Moving 2 Codidact Jun 15 '23 at 21:03
  • 1
    When you commission the Tesla Wall Connector, you tell it the breaker size. Either a) re-do the commissioning procedure and say 20A breaker, or b) if you have a soft/console/app control to adjust current, set it to 16A. Run it that way for awhile and let us know if that stops the trips. And if it does, please provide house square footage and the nameplates of your large built-in appliances (don't bother with fridge, dryer, tanked water heater or range, unless unusual). That's so we can look at your Load Calculation. Don't worry, that line of inquiry does not* lead to a service upgrade.* – Harper - Reinstate Monica Jun 15 '23 at 21:23
  • Thank you for the suggestions. I will check the subpanel connections, try to determine if the trips are ground fault or current related, and test with the charger commissioned for a 20A breaker and report back. I am not at the location and won't be for several days but if there are other things I should check, please comment in the mean time. Thank you! – dponti Jun 15 '23 at 21:45
  • Yeah, I could see only three broad reasons for tripping. Legit GFCI imbalance (But it is not immediately clear how charging the car would do that to a hot tub circuit, especially in a way that takes meaningful time to trip initially). Second is overload of the hut tub circuits. I suppose the car charging could cause a voltage dip which could cause those the tub circuits to draw more current, and thus trip. This becomes even more likely if your load is nearing or even exceeding your service limit, which it sounds like Harper is questioning. – Kevin Cathcart Jun 15 '23 at 22:09
  • Lastly is false GFCI tripping. For example some GFCIs can be sensitive to radiated or conducted RF interference (AFCIs are also known to being sensitive to such, but they are another story). It is not impossible for the car's charger to be introducing some such interference. I'm slightly skeptical of this, especially given the potentially long timing after initial charging, but it is not completely impossible. – Kevin Cathcart Jun 15 '23 at 22:12
  • Yeah, also is this place intended as an AirBnB style travel rental? I.E. is the typical guest expected to drive over 300 miles and arrive totally empty? That is the one edge condition where 60A makes sense, but that is achievable if need be. – Harper - Reinstate Monica Jun 16 '23 at 02:31
  • Finally able to address the above questions/comments. Thank you for the comments - I have more details but still a mystery: @KMJ: Subpanel is wired correctly - ground and neutral are not bonded. Connections to breakers are all tight. One hot lead to the spa breaker in the main panel was a bit loose but tightening did not resolve the problem. – dponti Jul 01 '23 at 22:12
  • @manassehkatz: Not sure if breakers are tripping due to overcurrent or fault, but tripping breakers with the test button (to simulate a fault) sends the breaker to the off position, consistent with what I observe when they trip during car charging. I don't know if these breakers trip from overcurrent at a detent stop like standard breakers do. These are Eaton double pole GFCI breakers in the spa subpanel. So all I can say now is that the breakers trip consistent with a fault. I also supect overcurrent is not a problem (see below comments). – dponti Jul 01 '23 at 22:17
  • @Harper: Charging the car for 4 hours at 16A did not trip the GFCI breakers, so power draw from the charger certainly a factor, but see comments below as to why I doubt there's overcurrent, but in case capacity is a problem here are details on my service: 1) 200A Murray panel 24/40 with 30 pole circuits installed so within panel rating. 2) all single pole circuits are 20A and run lights/plugs and no large devices except for standard frig, washer, dishwasher, etc.(all on separate circuits as per code). – dponti Jul 01 '23 at 22:36
  • @Harper: The Panel has 5 double pole breakers (240V) - a) Tesla charger (60A), b) hot tub (50A to a subpanel that has the 30A and 20A double pole GFCI breakers), c) 30A for electric JennAir range (cooktop is gas so oven is the only draw), d) 30A breaker to a subpanel in our pumphouse for a well. That subpanel holds a 20A 240V breaker for a 1 HP Gould pump (max. wattage unknown but presumably 16A or lower), and a 120V 20A circuit to power a filter and lights in the pumphouse, and e) a 50A breaker that powers a Carrier AC unit of mid 2000's vintage. – dponti Jul 01 '23 at 22:41
  • @Harper: To be clear, the issue with the GFCI's have occurred without the AC in operation (we rarely use it), with the oven off, and without the tub in use. The hot tub has a continuously running circulation pump that runs off the 240V 20A GFCI circult and is supposed to draw only 40 vatts. The heater can be a major draw (runs off the 30A 240V GFCI circuit) but only operates intermittently. The jet pump runs off of the 20A circuit but was never running during car charging. – dponti Jul 01 '23 at 22:47
  • I ran this experiment to test for overcurrent possibility: I turned off all breakers in the main panel except for the charger, spa, and a plug circuit that powers my wifi so I could control the car charging remotely. I also turned down the hot tub temp to prevent the spa heater from turning on during a charging session. 1) Charging the car at 40A trips the breakers within an hour. 2) When charging the car at 40A, and while the spa breakers were on, I tripped the breakers manually with the test button; the breakers then tripped immediately upon attempting to reset. – dponti Jul 01 '23 at 22:56
  • I then ran a series of tests similar to 2) but with the car charger drawing different amounts of current. I discovered that the 30A breaker cannot be reset if the car is drawing more than 18A, and the 20A breaker fails to reset if the car is drawing more than 22A. Below 19A I can trip and reset the breakers without issue while the car is charging, so I suspect that charging the car above 18A will eventually cause the tripping problem I'm observing at higher amperage. – dponti Jul 01 '23 at 23:05
  • I wanted to test this with the AC running but not the charger to see if this issue was charger related or simply due to a system power draw of more than 18A. However, we had a grid outage (not caused by me :-) ) and we were out of power for 17 hours. So I'm thinking of 2 possibilities: 1) faulty GFCI breakers, or 2) the breakers are sensitive to RF interference from the car charge as Kevin Cathcart suggested, with this interference increasing with increased power draw. Any additional thoughts or suggestions for a fix? – dponti Jul 01 '23 at 23:13
  • @Harper: Finally, this is our second home and not used as a rental. It is located in a rural area in the foothills of the Sierra Nevada Mtns. currently about 50 miles from the nearest public DC fast charger. We have frequent power outages from snow (downed trees), high winds (fire and preemptive power shutoffs to prevent them), and hot weather that overloads the local grid when tourists are blasting the AC in their AirBNB's :-). So being able to charge up quickly while the iron is hot, so to speak, is desirable. While 48A isn't necessary, being limited to 18A seems rather limiting.., – dponti Jul 01 '23 at 23:23
  • Feel free to [edit] that into your question. Is your testing during peak A/C demand hours? Are these the newer Eaton breakers with small round test button? Is neutral wired to the hot tub subpanel, and if so, how? – Harper - Reinstate Monica Jul 02 '23 at 00:22
  • I was testing about 3 pm on a Saturday but also noticed tripping at other times of the day. Temps were in the low 70's so unlikely there was huge demand. However, there was a grid failure at around 4 pm - i suspect some equipment failure as weather wasn't an issue. Yes, the Eaton breakers are the ones with the small test button. There is a neutral to the subpanel from the main - all the neutrals in the subpanel are bonded to a bar that is disconnected from the panel housing. The ground wires are bonded to a bar that is attached to the housing. – dponti Jul 02 '23 at 01:05
  • A lot of good questions, and great answers. Do you have a volt meter available that you could connect up to your service to confirm that you're not browning out your incoming voltage during this failure scenario, which might cause issues? This is unlikely, but would be an additional datapoint to confirm that you're not over drawing what's coming into your main panel. Also, when the breakers trip, do they feel warm to the touch at all (or do any other breakers feel warm, other than possibly the ones feeding your vehicle charging circuit)? – Milwrdfan Jul 02 '23 at 01:28
  • 1
    Since you have the more modern Eaton type breaker, you should really google up the instructions for your model breaker, and do the test procedure to determine the cause for the previous trip. It stores that. It will tell you if it's a GF, overvoltage, overload or mechanical trip. Yes, overvoltage trips Eaton GFCIs, in which case we might have a Lost Neutral from the utility. – Harper - Reinstate Monica Jul 02 '23 at 18:22
  • All good suggestions, thank you. My breakers do not have an LED, which is what I found when googling Eaton codes, so perhaps I don't have the new ones, despite the small round test button. I'll have to get the specific model number and recheck. Double-triple checking the connections is first on the agenda. – dponti Jul 02 '23 at 19:15
  • 1
    @Milwrdfan - I have an app that monitors voltage/current during car charging and voltages to the charger stays between 237-248V, and current draw is constant, Haven't tested voltages seen with individual circuits but I do have a voltage meter and can check. The breakers that trip are not warm, I haven't checked breakers that are not tripping. The charger monitors wire temperature entering the charger and in the cable to the car and will ramp down current if things get too hot. I've seen temps around 100-110F but the charger continues to function normally. – dponti Jul 02 '23 at 19:51
  • @dponti do you have an update on this? Another user has posted a question (today) implying a significant current flowing through ground from their charger. The potential (pun intended) is there for your two questions being related. https://diy.stackexchange.com/questions/279479/ – pbristow Aug 18 '23 at 03:26

1 Answers1

-1

Nothing suggests this is an overload problem, at least based on the loads you have described.

What GFCIs measure is basically 'is the same amount of return current coming back as I'm sending out?' For 120V that is comparing the load to the neutral. For 240V GFCI breakers, the comparison has to be made between both the hot lines and the neutral as a pair of balanced measurements. In both cases the current should always balance, i.e. the currents one direction and the other direction across the two hots and neutral have to match to within a fairly small amount of current. That's really a situation where nothing else should be able to do much to what the GFCI sees, unless there's something funky going on with the neutral. If one of the hot connections is loose somewhere and low voltage, the current coming out of it is still going to be balanced by the other leg and neutral. In the case of my hot tub, the neutral isn't even connected, which means the only remaining players are the ground (somehow), electrical noise, a defective GFCI, or an actual ground fault.

Go around and triple-check the torque of every connection including panel feed and ground rod connections, make sure every breaker is in place fully, and test the resistance of your ground system. It's not supposed to matter, except it helps to stabilize the neutral to some degree so it has a little bit of influence there. I suspect there's something slightly funky about a neutral somewhere feeding this panel or upstream.

If checking every connection doesn't do it, there's a few more things you can check: does it work with a different car? You might have to rent one to confirm that, unless you have a friend with one. You could exchange the Wall Connector with Tesla in case it's causing problems, they are pretty easy to get warranty service through. You could try the shotgun approach: replace the GFCI breakers. I don't think this will work but it's one of the cheaper items you could swap. And the last option would be to get a power quality monitor put on your system to see exactly what is happening on the line - this will probably be expensive as it's a very specialized tool.

That's the ideas I've got. I'm not highly confident in this answer. This is a super weird one. The fix is ultimately likely to be something simple that we're not even thinking of, or something totally off the wall.

KMJ
  • 11,320
  • 13
  • 43
  • 1
    I do not agree with replacing GFCI breakers just yet. Much more must be examined re. neutrals, grounding and bonding. Also the newer Eaton breakers store the last fault code, and detect many conditions. – Harper - Reinstate Monica Jul 02 '23 at 00:25
  • 1
    Thanks - I will check/retorque all connections as suggested - perhaps it's time to call out an electrician. The breakers are not cheap so I would like to be confident that they are the problem. I have an app that reports charger diagnostics and there are differences in L1 vs. L2 voltages coming to the charger of as much as 0.8 volts so presumably those differences are seen at the spa subpanel as well. Could that degree of variation in voltage contribute to a fault?. Could a noisy source (eg. spikey voltage vatiations) be the problem as well? – dponti Jul 02 '23 at 00:46
  • @Harper-ReinstateMonica that's why I called it the shotgun approach. I don't think it's the breakers either, but there's not many items left. My strongest confidence here is still that it's something mechanical somewhere, i.e. a bad connection. – KMJ Jul 02 '23 at 04:01
  • @dponti what do you mean by L1 vs L2? The two lines coming in to the EVSE have different voltages? I guess that means it's measuring against ground. If that's what is going on, and you don't find a bad connection somewhere, it's going to have to be in the electric provider's equipment. Once everything is connected well and assuming the wires in each spot are the same size for each side of the split phase, there's just not anywhere for the voltage to go. – KMJ Jul 02 '23 at 04:04
  • To put things in priority order: check literally every connection twice first, and the quality of your grounding system. This is assuming you can't get an informative code from the GFCIs. – KMJ Jul 02 '23 at 04:16
  • This answer makes no sense at all. You seem to think there is some sort of fundamental difference between 120V and 240V GFCIs. There is not. They all work the exact same way: all the current-carrying wires are looped through a transformer core, so that any imbalance (i.e. current supplied through the breaker but not coming back through it, such as through a person to ground) is detected. It's exactly the same whether it's 120V L/N, 240V L1/L2/N, 240V L1/L2, or three-phase. – nobody Jul 02 '23 at 14:23
  • Yes, the two hot leads going to the EVSE have slightly different voltages, and yes, measured against ground as there's no neutral to the EVSE. Also, I noticed when looking at the charging history that the line voltage varies by as much as 10V (~238-248V) ) although I don't know if that's unusual or not for service voltages to vary that much, or whether that would cause a problem. – dponti Jul 02 '23 at 19:29
  • 10V of line voltage variance is No Big Deal. My line voltage varies by 15V from highest to lowest during the day. https://www.pge.com/includes/docs/pdfs/mybusiness/customerservice/energystatus/powerquality/voltage_tolerance.pdf gives a bunch of background. – KMJ Jul 02 '23 at 22:32
  • @nobody that's why I called it a shotgun approach - I don't think there's something wrong with the GFCIs and don't see any reason they should be tripping while feeding a hot tub. Maybe they have an electronics problem though. It's unlikely since two of them failed at once, but it's not impossible. The hardest troubleshooting I've had in my life has been when I go 'there's no possible way these could have both failed' and then it turns out they both failed. If you have a better answer, please write it up. I'd love to know what is happening here because it's not adding up and I'm not sure why. – KMJ Jul 02 '23 at 22:33
  • The fundamental difference I see is that a 120V GFCI touches the neutral, and a 240V GFCI only touches both phase wires. It's all about 'what is in common', not about how it does the measurement. A 240V GFCI and a 120V GFCI both measure the same thing, but only one of them touches the neutral wire - and it's after the current measurement, so that drags my thinking here back to 'bad connection somewhere'. – KMJ Jul 02 '23 at 22:36
  • @KMJ 240V GFCIs also use neutral (otherwise they wouldn't be able to power a 120/240V load such as a dryer) – ThreePhaseEel Jul 03 '23 at 01:38
  • @KMJ My point is, your understanding of GFCIs is fundamentally wrong. 240V GFCIs absolutely do connect to neutral (a pure two-hot-only GFCI could exist but would undoubtedly be a waste of manufacturing and distribution capacity because it would rarely be usable). – nobody Jul 03 '23 at 02:00
  • I figured out my confusion: I was thinking of the GFCI built in to the EVSE, not the one in breakers. You two are totally right - a GFCI 240V breaker uses the neutral. The GFCI in the EVSE does not use the neutral, because the EVSE doesn't have a connection to neutral. So it can't use it. I'll revise my answer when I have a minute, the main revision being that it's even more likely IMO to be something weird going on with the neutral which is causing this problem, even though the Tesla EVSE doesn't connect to it. Thanks for pushing, it got me to see another way this could be happening. – KMJ Jul 03 '23 at 03:28
  • So yeah, I was mixing myself up because I was thinking too much about the EVSE and not enough about the hot tub breakers. I was also getting mixed up because my hot tub doesn't use the neutral, just the two hots and ground. Thanks. – KMJ Jul 03 '23 at 03:29
  • I think I got it, have to run off for a bit now or I'd do more. If you see more and think the answer is salvageable, feel free to edit. Or if you've got a totally different idea here, go ahead and write an answer. I may come back later and see if I can make the thru-line of this one more sensible, but not right now. Thanks again. – KMJ Jul 03 '23 at 03:37