0

I bought a brand new tiny house last year (RVIA certified) and I've been slowly getting a solar inverter and battery system installed. In the process my drill caught some household electrical wiring which I repaired (the wiring being three wire romex cable). However, thereafter, my (F150) truck's generator would trip a ground fault and not power my house (I've isolated the solar inverter from the house's electrical system for the moment... i.e. it is completely disconnected). I used off-the-shelf circuit testers and could find nothing wrong. Eventually, I pulled out a multi meter and found neutral-to-ground faults on three of the four wires my drill got a hold of. The resistance is about 400 kOhms on each.

Running new wiring seems impractical.

I am not an electrician/electrical engineer, however, I thought it might be a good idea to simply put a high impedance resistor at the breaks of the ground wires to trick the truck's inverter into thinking there is nothing wrong. (Basic google search reveals this is done with transformers but I am not sure how equivalent that application is to this)

How stupid is this idea? Is there something better?


Troubleshooting steps followed: My initial troubleshooting steps consisted of removing one romex cable after another from my tiny house's circuit breaker panel. I isolated the problem to three romex cables. Having confirmed that the problematic cables were the ones my drill had caught, I cut the cables, removed the frayed insulation, and butt spliced them together with heat-shrinked butt splices (the splicing is not in the walls; it is under my bed).

I was able to get the truck's GFCI to accept two of three cables I had spliced initially. However, with a bit of fiddling I had done (adding insulation), those circuits then once again started tripping the GFCI. I could never resolve the third one through these "repairs."

At that point I broke out my Hantek 2000 multimeter and tested the three cables closest to the wall where I inflicted the wounds. I confirmed there was infinite resistance between the hot wire and ground as well as the hot wire and neutral. I confirmed that there was a ~400MOhm resistance between the neutral and ground wire. My conclusion was whatever remaining injury was in the wall.

9
  • 3
    If your ground wires are broken, you should replace them. If you can't or don't want to do it, call an electrician. Anything else is on the stupidity scale (which goes from 1 to 10) about a 7, I would say maybe 8.
    – mkeith
    Commented Jan 28, 2022 at 5:57
  • 3
    One your repair was not done right, or two, you have other problems that need to be found and fixed. Damage at one spot can cause damage somewhere else.
    – crip659
    Commented Jan 28, 2022 at 13:22
  • 1
    I'm curious. You say you've got an inverter, so I presume your house is running on AC from the DC produced by the PV panels. How did you go about repairing the wiring? For AC wiring (presuming you're in the US, here) all wire splices must be made in approved junction boxes. Did you cut both ends of the damaged wire back, put them into boxes in the wall, then run new cable between them? If you just wire nutted in the wall, soldered, wrapped in electrical tape, etc., then you've got a fire-starting arc waiting to happen.
    – FreeMan
    Commented Jan 28, 2022 at 13:22
  • 1
    "drill caught some household electrical wiring which I repaired" - if your repair did not consist of pulling a new cable for each one you damaged, but instead joining wires and taping the joints in the walls, then no, you did not "repair" the damage. You bodged it enough to make it sorta-kinda work, and if your generator still sees a ground fault then you didn't even do a good bodge. Running new wires is the only way to fix this safely. Not running new wires almost certainly will void your insurance when something goes wrong.
    – brhans
    Commented Jan 28, 2022 at 15:16
  • 2
    Fair enough, @mkeith. Right up there with pulling the meter back off the base after the PoCo has turned off your power and jumpering the connections with a couple of 16d nails. (I've seen it done.)
    – FreeMan
    Commented Jan 28, 2022 at 18:06

4 Answers 4

5

In addition to "I know there is a potentially dangerous problem that I want to ignore" you also would be changing things in a way that might prevent the truck's GFCI from protecting against a serious unknown problem in the future.

Everyone worries about "nuisance trips" from AFCI and GFCI. There are some real world situations of False Positives. But they are actually relatively limited and most of the time the trips are real. GFCIs really do save lives. I can't say if any ever saved my life. But I can say that when they've tripped in my house it has always been for a really good reason.

3

Any tinkering/messing around/botching by adding mystical resistors to circumvent safety breakers/GFCIs/AFCIs etc. would be a serious and dangerous matter - in the worst case living in a tiny house would be replaced by living in a tiny cell in a shared room arrangement with full board and a very simple, but professionally maintained electric circuit system.

The resistance of only 400k Ohm measured with a standard multimeter may be very misleading.

Reasons:

  1. An increased capacitor (damaged cable) could mean a decreased impedance relevant for 60Hz-Alternative Current, which can't be measured with a multimeter, that only measures Direct-Current-resistances.

  2. Impedances in the real world, and especially the resistors of a damaged/drilled or burnt cable are not necessarily linear. It means, it could have a resistance of 400kOhm if measured with the 9V DC battery of a multimeter. 400kOhm would result in less then 0.5mA current @ 120V, so a 5mA GFCI should not trip.

But with 120V~, the resistance could be much less, f.e. only some 10kOhm, which will result in more than 5mA, thus tripping the GFCI.


Conclusions:

  1. The 400kOhm@9V= is most likely an indicator for a serious problem, which is difficult to identify with a standard multimeter.

  2. People with more knowledge/experience concerning electric engineering should be assigned to this case.

1
  • And the ground and neutral cables are actually connected together in the grid's junction box (done by a professional electrician). Commented Jan 28, 2022 at 18:45
3

Ground faults need to be treated seriously

In particular, don't make it your goal to defeat GFCI protection. The whole point of GFCI protection is to save your life.

Honestly the impulse of most Americans is to blame the GFCI for tripping (shoot the messenger), and they will replace the GFCI 2-3 times before coming here at wit's end, and we explain they have a ground fault. So you are ahead of the game.

Repair wires properly.

In the process my drill caught some household electrical wiring which I repaired.

Noting that you come from the electronics stack, I must caution you that electronics style repairs do not work and are not safe or legal in household wiring.

There are only a limited number of ways to safely and legally splice what I presume is NM/Romex cable. The only inline splice that is approved is the the Tyco in-wall splice kit. Any other methods involve two junction boxes on either side of the damage, and splicing a section in the middle with splices made inside the junction boxes.

Nicked wire is not OK because it reduces the cross-section and thus the practical ampacity before wire overheat.

Eventually, I pulled out a multi meter and found neutral-to-ground faults on three of the four wires my drill got a hold off. The resistance is about 400 kOhms on each.

That's disturbing if you can see it with a normal DVM. Ground faults are typically not linear (resistor-like) and behave more like a VBO (Voltage Break Over) device such as MOV. So they have almost infinite resistance at the very low test voltages that DVMs use, and much lower resistance at line voltage. There's a special type of ohm-meter called a "Megger" specifically for insulation testing. On the other hand, the upside is you have an easy way to measure when you've fixed it.

The neutral-ground bond will instantly trip any GFCI.

What drives me nuts is the grid's electrical panel has ground and neutral connected together (done by a professional electrician).

Well, that's correctly wired for a house's main panel: I.E. if the house will be planted somewhere and fed from an electric meter.

A lot of tiny houses get built with wheels under them to side-step the Building Codes since they're technically an RV... however, that creates an ambiguity of whether a Tiny House is a house or an RV. An RV needs separated neutral and ground. A house needs it bonded right past the meter.

Point 1: NEC 2020 now calls for main breakers at the meter (or to be more precise, disconnects outdoors, but given the cost of the various options, "main breaker at the meter" works the best on cost). The neutral-ground bond happens at that main breaker, 4-wire is brought onward, and neutral and ground are separated past that point.

Point 2: The power company will not attach a meter and service drop to a house with wheels. (look at any "mobile home park" where the utility slaps the meter on a nearby telephone pole, and there's a main breaker or ranch panel right there. In fact there's a special 4-wire cable called "MH Feeder" made just for the pole-MH connection).

These two things combine to mean that most likely any practical installation will require you to pull that bond.

You need exactly one neutral-ground bond in the system and it needs to be on the supply side of any GFCI. If you remove your bond, watch out using it with any generator which also had its bond removed.

1

What's a ground fault, anyway??

A ground fault is either an improper connection between neutral and ground, or else current from a hot wire returning through ground instead of through neutral. The question includes a remark about a multimeter measuring neutral-to-ground faults on 3 of 4 wires. This doesn't entirely make sense (does "wires" mean "separate conductors" or "cables, each having several separate conductors"?). I'll write a few words about how you could improve your test methods to isolate the real problem better.

Eliminate possible hot-to-ground leaks

First, eliminate all possibility of leakage from hot to ground. That's easy: Unplug anything that has a plug, turn off anything that has a switch. Maybe even turn off all the circuit breakers.

Test whether there is a neutral-to-ground leak

Now, let's test whether there's a neutral-to-ground leak inside the house. One way to do it is with the multimeter. At the house power inlet cord, measure resistance between the neutral blade and the ground blade. If it's anything less than infinite, and assuming test procedure is correct, then somewhere there's a leak between a neutral wire and a ground wire.

There's another simple way to test for stray neutral-to-ground leaks.

  1. Start up the truck's inverter and plug in an outlet strip. With all the house loads still powered off and/or disconnected, plug in the house power inlet cord. The truck's inverter should not trip.
  2. Plug in some load to the outlet strip and turn on the load. Current will flow down the hot wire from the inverter, through the outlet strip, and up the appliance cord. The return path should be the reverse: appliance cord neutral, outlet strip neutral, back to the inverter. GFCI should not trip, but..
  3. If the truck's GFCI does trip it means there's a neutral-ground leak in the house. Here's how we reach that conclusion. Current flows down the hot wire from inverter to appliance as described above -- but the return path is wrong. Current returns through the appliance cord's neutral, but then it forks at the outlet strip. Some of the current takes the direct path on neutral back to the inverter as it should. But some of the current -- the fault current -- goes the long way around. It flows along the neutral toward the house, across the neutral-to-ground leak hidden somewhere inside the house, and returns to the inverter on the ground wire. That'll make the truck's GFCI trip. It would be a serious problem that needs to be found and fixed, not hidden by impairing the ground with resistors etc.

If the tests above do not reveal a neutral-to-ground leak (fault) then celebrate! Your repairs, although they might need to "level up" to a higher standard, at least didn't cause a neutral-ground fault.

If the tests do reveal a neutral-to-ground fault then we can investigate that further. Undo the recent repairs, spread out the conductors so you can see space between them, test again. If there's still a fault then the hunt continues.. maybe it's in the recent solar inverter addition -- that might not be so isolated as you believe it to be, for example.

No neutral-to-ground leak? Look for hot-ground leak/fault

If you turned off circuit breakers, turn those on one at a time while monitoring whether the truck inverter's GFCI trips. Then start switching on loads one at a time. Finally start plugging in things that had been unplugged. If the GFCI ever trips, turn off or unplug that most recent thing. Reset the GFCI. Proceed to turn on the other loads/circuits. Ask a new question for help troubleshooting any load(s) that appear to be causing a hot-to-ground fault.

1
  • Very detailed thank you. I am certain at this point that there is a neutral-to-ground fault in the wall. My hair brained thinking probably is routed in my limited understanding of PCB circuit design. I thought that keeping the ground cable connected with a big resistor would ensure that any static electricity on the ground prong would have a route to ground and that it would also present a higher impedance path for current returning on the neutral wire along the ground wire than would be the case along the neutral wire. My thought also was that it would present to the GFCI as big impedan Commented Jan 31, 2022 at 3:58

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.