# Broken underground ground wire

I have underground three-strand #4 aluminum insulated from my service meter to my shop approx 500 ft. At the shop panel the volt meter with no load on either line across the lines 240V plus and from each live to the ground line 120V plus. All looks good, turn lights on one of the line circuit voltage will drop 10 to 20V, the other line with no load will rise by about the same. If I load the other line, the voltage will reverse the readings. Have told this is because ground wire is bad.
So now I have questions:

1. How do I locate the break?

2. Why does the ground to soil rod not taking up the task?

3. My house is 325 ft away from the barn, my redneck engineering says to go from my shop earth ground rod to my house earth ground with bare copper. and save ripping up a paved driveway.

• Run a separate ground wire, next to the driveway, in the grass if possible. Sounds more like you need a "neutral" wire and less a "ground" wire. Cannot use ground steaks - think of the earthworms! Mar 12, 2020 at 14:32
• It sounds more like a broken neutral than a broken ground. In which case you will have to rip the wire out. The ground rod doesn't help because its resistance is too high to do anything useful here. Mar 12, 2020 at 14:48
• The ground wire is a safety wire, and is a smaller gauge than a proper neutral wire. So if you are using the ground wire as the return, then it'll have higher losses and cause an imbalance. Mar 12, 2020 at 16:45
• Lindy, I have a few questions but I also have a possible answer, too. (I probably measured the voltages you've measured... many times before and it's always been the same problem.) (1) How long ago was the 325' run of wire put in the ground? (This usually is NOT the problem. But I have to ask.) (2) Do I understand correctly that the underground line carries L1, L2, and N? All three? (N is usually starred to Earth near the entrance panel at the home.) Again, a question I have to make sure about. (NOTE: 325' of Earth will have quite a resistance to it, so don't depend on that.)
– jonk
Mar 12, 2020 at 17:52
• @rdtsc Technically, once you grind a steak it becomes ground beef.
– DKNguyen
Mar 12, 2020 at 19:30

And this is an emergency. You need to fix this before using the power there.

The safety ground isn't making a difference because that is neutral's job. Your neutral is completely gone.

How do I locate the break?

Now, unless you've been trenching, 99% of the time the problem is at a termination, which is probably just 2 or 4 places at most. I'm guessing you did not see it because you've been checking ground, not neutral. And depending on how much "redneck engineering" you've been doing, it might have been installed incorrectly in the first place.

If you have any ability to measure AC amps, another thing you can test for is current flow on the safety ground, e.g. out to the grounding rods. There should never be any. If you see current flow on the ground, that means neutral is not doing its job.

You never, ever, ever use earth as a current return.

Why does the ground to soil rod not taking up the task?

Same reason they go to all the trouble to mine copper, instead of just using dirt. Dirt is a very poor conductor. The only time dirt ever works as a conductor is with extremely low current (thus extremely high voltages), such as Australian rural distribution or a hack done in the third world with microwave oven transformers.

This work is incorrect by modern standards. A ground wire should have been trenched along with the conductors if installed in the past 20+ years. A ground wire can be retrofitted and does not need to follow the same route.

My house is 325 ft away from the barn, my redneck engineering says to go from my shop earth ground rod to my house earth ground with bare copper. and save ripping up a paved driveway.

I'd run it by the inspector, but in principle, that is compatible with the gist of the retrofit ground rules.

## Neutral is not safety ground

And by that I mean ground in the mains electrical sense, not GND as "common" in the electronics sense. If you're an EE and trained to sink all current to GND/Vss, that thing is what we call neutral. In mains power, you sink current to neutral (or the other hot). You never sink current to ground. Ever. It is never a current return; current never flows on ground except during fault conditions.

Electricity doesn't want to go back to ground; it wants to go back to source. The difference is neutral goes back to the transformer (source for human-made power), and ground goes to the earth (return for naturally made power: lightning and ESD).

A lot of people conflate the two, on the logic that they're tied together at the service panel. That should only be one place in your whole service. The job of the neutral-ground equipotential bond is to a) prevent hots and neutral from floating at thousands of volts (Say due to capacitive coupling inside a supply transformer)... and b) allow fault current moving on ground an easy path back to neutral, so the fault current can flow enough current to overcurrent-trip the breaker. (otherwise fault current flows indefinitely, and simply energizes your grounds).

## One option: rewire it as a main service panel

Now, in the old days, they allowed groundless 3-wire feeders to subpanels in outbuildings. I gather your installation is that old. That's dangerous generally; however with such very long distances there is some sense to treating the line to this outbuilding like it's a Service. Service lines (e.g. your pole drop from the Power Co.) do not carry grounds. Talk to your local inspector about giving you waiver for that... if your very old installation isn't grandfathered. (if it was built to code at the time, you don't need to modernize it).

If you have your grandfathering or waiver, you treat it just like a service and the panel like a main panel: 2 local ground rods at the building (or an Ufer ground or 1 rod that passes the expensive test), which you need anyway, because it's an outbuilding. Then install the one neutral-ground equipotential bond there in the "main" panel.

## Another option: Go 240V with a transformer

If you are not grandfathered and your inspector won't give a waiver, you will need to wire a ground wire from the main panel. Given the cost of trenching, I would say you are better off going with a transformer, which has 2 effects:

• It only needs 2 wires, freeing up your third wire to be used as safety ground.
• It turns this barn panel into a separately derived service which means you do indeed bond neutral and ground.

I mention this because most people spec their wiring runs for 3% drop, and #4 aluminum at 500' only gives you 19A @ 3% drop. So I'm guessing you only intend lights and odd plugs/sockets (dual 20A). Well, that's only 4800 VA, and 5k VA transformers are \$100 on Craigslist.

In a transformer setup, the 3 wires are tasked to 2 hots and 1 safety ground. No neutral. The pole service feeds the 2 hots at 240V via a 20A breaker. At the barn, the 2 hots go into the primary of the 5 KVA transformer. The service panel is then served by the secondary's L1, L2 and neutral. Now you have 20A @ 240V, or dual 20A @ 120V.

If you want more, you could go with a >=7.2 KVA transformer and run 30A that same way... at 4.8% voltage drop, which is acceptable. But the transformer will be more money.

You could do the same with a 10 KVA transformer, running 40A, but voltage drop would be 6.5% at full current, which is really, really pushing it.

If you want much more, you can get 25 KVA transformers at both ends back to back (primary to primary both jumpered for 480V). Then backfeed the transformer at the pole, so it's a 240V-240V pair of transformers with 480V for transmission. The wire would run 480V@50A for only 4.1% voltage drop, perfectly reasonable! That will give you 100A@240V at the garage. So if you want a Tesla 80A charger, you're all set.

If you want even yet more, you can do ditto-ditto- dual 240V/600V transformers at hard wire limit of 65A@600V (39 KVA) at 4.2% drop. That will deliver 163A @ 240V to the shop, though realistically, 175A because you'd need to round up to the next available breaker.

All of the above is a no-trenching option.

As mentioned in several comments, it does sound like the problem is with the neutral line rather than the ground line. The neutral line returns current to the transformer from either of the two 120V phases.

While it is possible to use the earth as a current return path, this is typically done with specially designed high voltage power distribution systems. The resistance of soil is too high to use as current return path for normal loads at 120V.

So you will need repair or replace the neutral return line.

I agree with @rdtsc that running a brand new line is likely easier than locating and repairing the break in the old one.

But if you do want to attempt to repair the line, you will need to locate the break, then dig (which usually requires clearance from local utilities), and then do a splice that can survive being buried again. I've done this myself, and it was a big hassle.

To locate the fault, the best way is to use a Time Domain Reflectometer. These are almost magic devices that shoot a pulse of electricity down the wire and count how long it takes to get a reflection from the break. Using the speed of electricity in the conductor, they can then compute the distance to the break.

Note that I happen to own one of these TDR's, and I am happy to lend it to you if you are near NYC - but, again, I think replace is a better option here than repair.

• I Thank all for the insight, I under stand the "netural" much better. So what "if" Run a netural withe from my barn panel to the netural on the house panel?
– Lindy Sasser
Mar 14, 2020 at 14:03
• @LindySasser As long as the new neutral is big enough and you disconnect the old neutral on both ends and connect the new one instead, and there are no faults with the 2 hot wires, then it should work, but if you are going to run a new neutral it is probably best to run 3 new connectors since whatever killed the old netural might also kill the other conductors soon! Mar 15, 2020 at 16:00
• correction: 3 new conductors (wires) not connectors Mar 15, 2020 at 16:06