My barn had a small (6-space) main lug sub panel with four 20A 120V breakers and one 20A 240V breaker. The panel itself was protected by a 50A 240V breaker in the main service panel. It was wired up presumably before 2001 since all of the NM wire going into the panel have white sheathing -- even the 12 gauge. It worked fine until just a few days ago.

On this day, one of the 120V circuits abruptly stopped working without tripping the breaker. I noticed that the panel felt very hot so I cut power to the entire panel until the next day, when I had time to investigate.

It was then that I discovered the carnage inside of the panel. All of the neutral wires had around an inch of melted plastic that was the wire insulation and several inches of blacked/burnt plastic going up the wire. One of the neutrals (the one for the circuit that stopped working) was loose, since it had melted right through -- the hole in the neutral bar is filled with previously molten copper and/or aluminum! The neutral bar itself is very discolored and the plastic support for the bar is thoroughly deformed due to its own melted state.

EDIT: Added photo showing the remaining very-burnt neutrals. The neutral bar lug that is largely black is the one with the fully melted neutral (not pictured). Burnt and melted neutral wires

Notably, none of the circuit breakers ever tripped -- not even the one with the completely melted through neutral. They are all 20+ year old breakers so none are arc-detecting.

I have since ripped the entire panel off the wall and replaced it and all the breakers with a new panel and breakers, just in case. I also cut back the neutrals to leave only non-discolored copper and wire insulation. But what I want to know is why the previous panel failed the way it did, to ensure that something like that doesn't happen again.

One theory is that the melted neutral actually worked itself loose over the decades and at some recent time, was at the right distance from the neutral bar to start arcing to the bar. This would create tremendous heat -- presumably enough to heat up the neutral bar enough to transfer the heat to the other neutrals, causing their insulation to melt and burn... and hot enough to melt the copper conductor in the problem neutral wire itself.

But can a loose neutral wire arc quite that much, especially without ever tripping the breaker?

Is there something else obvious that might be the culprit?

  • 4
    Is this a 3-wire feed or a 4-wire feed? Any chance you took some pictures you can edit in?
    – Ecnerwal
    Sep 9, 2023 at 1:57
  • 2
    Your theory is viable. It could arc "that much". The faulty neutral would tend to burn itself away from the terminal it is arcing, but if the wire is installed to move by gravity or its own springiness back towards that terminal it would keep arcing and heating things up as you describe. It would not trip the breaker because the current is actually reduced, not increased, this way. It could trip a breaker if stray molten copper created a short circuit between two phases.
    – jay613
    Sep 9, 2023 at 2:18
  • @Ecnerwal - I added a photo of the neutral bar showing the remaining neutrals, melted plastic, and everything other than the neutral that melted all the way through since that's swinging free. Sep 9, 2023 at 20:23

3 Answers 3


Because NEC 110.14(D) wasn't followed. A torque wrench was not used to set the screw torques. This is why the rule was added to NEC 2014.

Now if the work predates 2014, it was legal and should not have failed, but unfortunately arc faults can't read :)

The circuit you found loose may have been loose before, and that may have caused all of this. Arcing is serious business.

Series arcing is dealing with potentially as much power as the load is designed to draw. And a resistive load will decrease current as arcing consumes some of the current, but a motor load may draw more, and a constant-power switching power supply will definitely draw more current as the arcing gets worse, creating a race condition. EV charging, a common source of such trouble, will draw constant current despite the arcing.

But can a loose neutral wire arc quite that much, especially without ever tripping the breaker?

Yes. And breakers are not magic everything detectors. They only detect overcurrent.

  • 5
    Okay. It might be notable, then, that the circuit in question did have its usage pattern drastically change recently. For most of the past decade, it maybe had 100W on it at a time for a few minutes a year. In comparison, the past few months have had 1000W to 1300W loads for hours at a time and daily. Maybe it was always loose and it simply didn't matter until now? Sep 9, 2023 at 20:25
  • 1
    @Kurt quite likely, and like I say it depends on the load how it will react to the voltage drop caused by the arcing. Sep 9, 2023 at 20:51
  • 2
    @KurtGranroth -I'm guessing that an electric vehicle was recently purchased? Besides the higher amount of current, EV charging creates a continuous load (longer than 3 hours according to NEC code) which can heat up wiring more than intermittent use does. Not that this caused the arcing, but it's worth making sure that you don't use more than 80% of the rated capacity of the circuit while you are charging your EV, or whatever it is that is drawing the power for hours at a time. For a 15 amp circuit the limit is 12 amp continuous load, for a 20 amp circuit the limit is 16 amp continuous, etc. Sep 10, 2023 at 21:50
  • 1
    Do you possibly mean positive feedback loop rather than race condition? The latter typically refers to situations where the results depend on the precise timing of events, which doesn't seem to be what's happening here...
    – psmears
    Sep 11, 2023 at 10:49
  • Some people are lucky enough to have AFCI breakers, which detect arcing, on circuits that arc.
    – user253751
    Sep 11, 2023 at 13:59

A possible way a loose wire can cause this after appearing fine for a while is a glowing connection. The poor contact creates a hot spot which causes thermal cycling of the wire and busbar. This will loosen the connection further over time.

Glowing connections don't tend to cause a large voltage drop, so go unnoticed. Eventually they fail, or melt enough insulation to cause a short circuit first. Increasing the load will make failure more likely due to increased heating.

Once the connection fails, it can start arcing, and the series arc will be like that of a welder, ballasted by the load on the circuit.

  • 2
    This is exactly why you should use a torque wrench!
    – Ian Kemp
    Sep 10, 2023 at 8:10

I lack reputation to comment on this post, but I must make the following remark;

1 element in the photo that immediately drew my attention, but is not menioned in any comments or answers, is the burned/melted stranded-wire sticking-out of the neutral bar.

I am in the EU, and I am not an electrician or up-to-date with current regulations here in the EU, but as far as I know, stranded-wires are an absolute no-go in wiring panels and sockets. (I am not at all familair with the relevant code/regulations in the US).

Even without the relevant training, I can think of 3 or 4 different ways stranded wires might pose a safety hazard in this context and I'm sure there are more I didn't think of.

So, although I don't have an answer, maybe it's worth investigating the role of these stranded wires in your situation?

  • 2
    Welcome! That's an interesting insight about wiring practice in the EU. Indeed standards are different in the US: we use both solid and stranded wire. Our 8 awg/8.36 mm2 and heavier are almost always stranded whether they're single conductors or part of a cable, and single conductors for use in conduit also are available stranded. In all cases the stranding is more coarse than what's used in portable cords though. Our 14 awg through 8 awg THHN type building wire all use 19-strand construction wheras an SOOW type portable cord in 12 awg might be comprised of 65 strands.
    – Greg Hill
    Sep 11, 2023 at 17:04
  • 2
    Electrical wiring regulations differ between EU member states, so I'm afraid you can't just generalize like this. I know for a fact that at least in the Netherlands, using stranded wire for residential wiring is perfectly normal and entirely legal, as long as you crimp a ferrule onto the end before sticking it into a screw terminal (unless that terminal is rated to accept stranded wire directly).
    – TooTea
    Sep 11, 2023 at 20:12
  • 1
    North American practice uses coarse (7 or 19, sometimes 37 on large wires) strandings routinely on building wire (THHN/TF(F)N/XHHW and inner conductors on cables save for NM/UF 10AWG and smaller which are always solid). It's fine stranded wires that cause the major issues in lugs (and are thus prohibited on most normal screw lug terminals) Sep 12, 2023 at 2:38
  • @TooTea: Interesting to know it's legal in NL, because that's where I'm from originally, though I've never seen it used mself. Anyway, I readily admit having limited knowledge of regulations, anywhere. My main concern, and reason for responding, was the stranded wire in the photo, which looks very shady to me and would be one of the first things I'd look at. For a start, the strands don't look to e bound together, they fan-out towards the bar. This could suggest only a part of the strands were actually tied down, which would be a perfect reason for overheating.
    – arritjenof
    Sep 12, 2023 at 14:52

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.