Why does my GFCI outlet trip when I touch it to the wall box?

Question

UPDATE: The wall box IS grounded via a wire that is screwed to the back of the wall box, and clamped to the Original 1953 galvanized plumbing in the wall behind the kitchen sink. I think it is the only grounded outlet in the house.

I plan to test the ground, replace the rusty wall box, reconnect the ground wire, and wire up the GFCI properly.

As a test, I pulled the outlet out of the wall box, plugged in the dishwasher, and measured 72V from the door to ground and to the neutral (not ground) wire. That seems odd. There is no measurable voltage on the door when the ground pin is grounded.

More Questions of course…

Is there a practical way to test & prove out the ground connection?

Is it sane to count on 70 year old galvanized kitchen sink plumbing to provide a robust path to ground?

Should i install a new ground rod and wire it to the box/outlet?

I found a copy of the 2020 NFPA chapter 3 electrical code, so i now have 1000+ pages of bedtime reading to add to the pile of books on the nightstand.

Thanks, Paul

————— Original post below —————

I installed a dishwasher in our house for my wife’s birthday. Along the way I encountered two electrical oddities that I can’t figure out. I aim to have a safe and reliable system, and i am concerned that is not the case right now.

When the dishwasher was plugged into the outlet under the sink, running my finger along the door skin made a humming noise. I assume this means some voltage present on the door. In debugging mode, I took apart the 3-prong outlet and found there was no ground wire. Shunting white “neutral” to the ground pin “fixed” the humming, and the door felt like an inanimate object… a good thing.

But the humming bothered me and I decided to install a GFCI outlet under the sink. After wiring it up (black, white, and shunt from ground pin to white), the GFCI behaves as expected as long as it is outside the wall. That is to say, with nothing plugged in, the GFCI can be tested and reset successfully. But, if i go to install the outlet into the wall box, the moment the mount screws of the outlet contact the metal wall box, the GFCI trips.

Can anyone help me understand or diagnose what is causing the GFCI to trip? Is it reasonable to susoect any issue with the door of the dishwasher is either because of a problem within the (new, bosch) dishwasher, or because it is designed to work only with a ground wire in the outlet?

After reading about home wiring and gfci function, I have a theory that the neutral wire is not tied prioerly to griund at the panel. This might be allowing its voltage to float a lottle above ground. When the gfci is grounded to the wall box, the potential difference between neutral wire and ground causes current to flow through the neutral wire. Since nothing is plugged into the outlet, no current flows through the black (hot) wite. The gfci detects the difference in current flow between the two wires, and trips.

Bottom line: Do i need to do any diagnostics on the dishwasher itself? Or is it reasonable and safe to assume if it works properly and the door isn’t powered when plugged into a grounded outlet, that all is well with the dishwasher?

How can i figure out what is causing the GFCI to trip, and how can i fix it?

What other questions or approaches would you suggest i consider?

Thank you, Paul

Answer 1

Shunting white “neutral” to the ground pin “fixed” the humming, and the door felt like an inanimate object… a good thing.

NO.

Just no.

That's called "bootlegging ground". It fixes nothing and breaks the GFCI.

If that "causes it to start working" that is interesting diagnostic but continuing it in service that way is out of the question.

After wiring it up (black, white, and shunt from ground pin to white)

Again NO.

Cauterize that "very bad nogood idea" right out of your brain, before it cauterizes you. Seriously. Neutral Is Not Ground.

All grounds go together and only to grounds. Note that ground is also metal boxes, metal conduit, and a non-flex metal conduit can actually be the grounding path - no ground wire needed.

All neutral wires run with their partner hot wires and never tie to ground anywhere.

Well. Once you have that down, we'll discuss the Neutral-Ground Equipotential Bond which happens at the first disconnect past the meter (e.g. your main service panel), which due to hokey implementation often causes people confusion on the "neutral is not ground" thing.

That is to say, with nothing plugged in, the GFCI can be tested and reset successfully. But, if i go to install the outlet into the wall box, the moment the mount screws of the outlet contact the metal wall box, the GFCI trips.

You did that to yourself with the bootleg ground lol.

Anyway, GFCI protective devices do not take ground. GFCI breakers don't even have access to ground. Of course you are familiar with a GFCI plus two sockets; that takes ground only for the sockets. The internal GFCI module doesn't use it at all.

If you read the instructions, it tells how to put 2 wires under a screw. A common error is to peel (and ignore) the warning tape covering 2 other screws, and use them instead. This tends to make complications. TLDR: Never do that until you have skilled up on using the downline protection, and only use them for that purpose. Heeding that advice will steer you clear of many frustrating problems.

If you reinstall the GFCI with the Load terminals cleared of wires, that should fix up your problem. It sounds like your metal box is grounded, so the GFCI will automagically pick up ground via the mounting screws.

This can confuse your testing, because it "seems to work" without the ground then "fails" with the ground, making the ground seem like the problem. It's not the problem. The problem is elsewhere and breaking the grounding is just preventing it from being detected.

Now if it holds without the dishwasher plugged in, but trips with the dishwasher plugged in, try other appliances. If it only trips with the dishwasher, the dishwasher has a problem.

Note a GFCI receptacle is not allowed to be inaccessible behind the dishwasher. Ask how to solve that problem.

Answer 2

Thanks for the clarification.

Don't connect ground to neutral on the GFCI. I suspect that there's a bit of voltage on the ground of the box you're installing the outlet in. The old outlet had ground connected to the ears, and was running that voltage to the dishwasher. That is probably what is tripping your GFCI as well.

Put a voltmeter from the ground of the box to neutral. I bet you see a voltage reading more than zero. At that point you need to find out where that voltage is coming from in order to be truly safe. Any situation that ties ground to neutral will negate the reason to put the GFCI in, or just make it pop.

Answer 3

You need to check the overall grounding of the house. The starting point for that is your main panel. While retrofitting ground for individual circuits is a legitimate thing to do, it is dependent on the main ground for the house being done properly. Generally speaking this is some combination of:

• Ground wire from main panel to a ground rod
• Ground wire from main panel to metal plumbing
• Ground wire from main panel to an Ufer ground

The general rule, but it varies by jurisdiction, is Ufer ground or a tested ground rod or two ground rods. But the ground to plumbing was far more common in the past, and is still used in many places. Here are some things to consider (based on many Q&A posts here plus my personal experience and stories from my electrician):

• Ground wire from main panel to a ground rod

This depends on being "really good". But checking that requires expensive equipment. So places that allow relying on "just a ground rod" normally require two ground rods, since that is pretty much guaranteed to provide a good enough ground. My electrician told me this is what they do in Howard County, MD because a large percentage of homes there don't have metal water pipes.

• Ground wire from main panel to metal plumbing

This used to be the standard. The catch is that if the connection is not at the plumbing service entrance then if copper or galvanized pipes are replaced with plastic, or even just a single fitting is plastic, then the ground is useless. In my area (Montgomery County, MD) the usual setup now is one ground wire to the plumbing service entrance and one ground rod.

• Ground wire from main panel to an Ufer ground

This is a great solution, but only practical for new construction or major renovation or addition that includes concrete work. Many places require this for new homes.

So the starting point is to figure out what, if any, ground wires are going to the main panel. If there are none (or if they are non-functional), then depending on your local rules, either two ground rods (that can be one wire to two rods, several feet apart) or one ground rod + plumbing will likely be the answer.

Once you have the ground for the house, the next step is the ground wires for individual circuits. The ground wires for individual circuits should be going back to the main panel, not going to the plumbing. The good news is that you can retrofit grounds in a number of different ways. For example, you could run a ground wire daisy-chain from one receptacle or junction box to the next until you get all the way back to the main panel, without following the same route as the hot/neutral cables.

Getting back to the original dishwasher problem (which may have been a blessing in disguise if it leads to making your entire electrical system safer), it sounds like you have a defective dishwasher. Generally speaking, the metal parts of an appliance should be at ground, and ground should be effectively the same as neutral. But it is not 100% clear to me, as ground and neutral really shouldn't be connected at all inside the appliance. In other words, if you don't have ground connected, there shouldn't be a high-resistance connection at 72V, there should be no connection at all between the metal frame and neutral. A little more research on that may be needed, and a call to the manufacturer may be a good idea.