What could cause a power drop on one leg in 500' of buried cable (and a possible solution)?

Question

I'm having a voltage drop that appears and disappears in a 500' underground cable.

The setup: We have a 400 amp service coming into the house from the meter. It goes through a transfer switch for the generator, then two two panels. In one panel is a 100 amp breaker and coming out of that breaker is heavy duty wire (forgot which gauge) that runs about 60' to a j-box. At the j-box that wire is joined to 4/0 wire. (This is all 3 cables plus the ground.) While I don't remember the gauge of the cable from the panel to the j-box (and most of that run is underground), I do remember working out the math with more than one electrician and making sure it could handle the load.

I think it's standard labeling. The common line is marked with white tape, one leg with red tape, and the other is black (no tape).

The 4/0 cables go from the j-box, then underground for about 500', to our barn. There they come up and into the wall and, on the other side of the wall, into the breaker box in the barn. This has worked fine for several years, since we renovated the barn. (It's also passed inspection.) This Saturday we had a late season temperature drop and snow. It was cold, but not as cold as it's been and it didn't stay cold too long. Down below freezing for a few hours Saturday and Sunday night, but in the 40s during the day.

Also, of course, the house is properly grounded. Ground wires run with the 4/0 cable to the barn and connect to the panel ground. Also, the barn itself is properly grounded according to code and that ground connects to the panel and the ground line from the house.

The problem: I went down to the barn on Saturday (the snowy day) and some lights were dim and HVAC was not working. The outlet checker showed some outlets with 97VAC and some with 84VAC, others with 127VAC.

For troubleshooting we (including a licensed electrician friend) removed the cover panels on the electrical panel in the barn and on the panel in the house that feeds the barn. We also turned off the master switch on the barn panel.

At the house, and at the j-box near it, we get 247VAC on the cables from leg to leg. Down in the barn, it reads 237VAC or 214VAC at times. It goes back and forth from 237 to 214 and back. We can't tell yet if it's at regular intervals. There is no temperature correlation. (In other words, it doesn't go down when it gets cold or up when it warms up.) The voltage at the house and nearby j-box remains constant. In the barn, measuring from the black leg to common, we get 127VAC, and that stays constant. From red to common, we get either 112VAC or 90VAC. The total voltage, leg to leg, is always within a volt or two and is the same as the total of measuring both legs separately.

It's possible that one leg has always been a little lower. I remember at the time going over numbers with the electrician when we wired the barn and that the numbers were at least nominal, but one may have been lower than the other. However, the extra 20VAC we lose at times (and get back) is my concern now.

This cable is buried. What could be causing this voltage drop? I've considered moles or groundhogs, but, again, the voltage goes up and down, so something changes at times, at least several times a day.

Possible solution: My electrician and I were talking this over and are worried that something seems to be happening and changing somewhere along 500' of underground cable. That's not easy to find!

We have talked about turning off the power to the underground cables at the house. Then, since it's the red cable that's bad, switching the red and the white cables, both at the house and in the barn. That way the problem cable will become the neutral. Since the neutral is connected to the ground, and the panel is grounded at the barn and connected to the ground from the house, if the problem is a hole in the insulation of the red cable (and, as of now, there's no GFCI on either end - and inspectors were okay with that), then we figure that cable may have a very tenuous connection to the earth ground along the way. Since the neutral connects to the earth ground anyway, it seems like that would be okay.

I'm not an electrician and I have to admit, that's more my idea, from lack of experience, than the electrician's idea. He wants to think it over.

Does that possible solution sound risky or does it make sense?

Are there other thoughts, besides the obvious (digging up cable and testing it!) for a way to deal with this fluctuating voltage drop?

Note: I've already touched on some of this in another question.

Addendum: I forgot to mention that if my solution were to work, I would feel it would still be necessary, as early as I can, to dig another long trench and run replacement cables to the barn. Whatever is happening to the current red cable could get worse or be part of a bigger problem that could hurt the other cables, too. But that's a LOT of work and not something I can do soon or quickly. (We need the barn, both of us work out of it.)

Addendum 2: This is ongoing. I'm getting a LOT of help in chat from one person. I haven't picked an answer yet because I'm learning a lot and, at this point, there are points I've learned that aren't in any of the answers.

Also, a note on conduit, since one answer and some comments have brought that up: This line goes through the woods. Conduit is seriously problematical, since there is a limit of 360° to the number of turns one can have in conduit and 180° of those is used up by the two 90° elbows that are needed to go from the horizontal run to vertical to bring the conduit to the surface. I've talked with building inspectors. (I have a good relationship with the Inspections Department in our county.) Buried line is considered safe and rarely breaks. We've found one break and I'm pretty sure I know the cause of it. It's not animals or rocks. (I was careful to keep rocks and sharp objects out of the way.) It's pretty much impossible to run a straight line through the woods or limit even part of the line to 180° of turns. (And pulling 4/0 through even 180° of turns takes a LOT of strength and means the insulation is going to be rubbing against ALL of corners in the conduit.)

Answer 1

TLDR: Disassemble, clean and reassemble every wire connection on the 4/0 subpanel feeder. Torque the lugs with a torque wrench. Don't take any shortcuts here. Don't "visually inspect and declare to be fine" since it isn't fine. Make sure the splice terminals are aluminum-rated.

Assume that it wasn't properly torqued when done, so re-do everything.

You can do a bunch more measurements and paper troubleshooting if you really want to, but I wouldn't bother.

In the barn, measuring from the black leg to common, we get 127VAC, and that stays constant. From red to common, we get either 112VAC or 90VAC. The total voltage, leg to leg, is always within a volt or two and is the same as the total of measuring both legs separately.

You lost a hot

One of your hot wires is having a bad day. Neutral and the other hot are fine.

It might be an intermittent or partial connection. We know it's not a total loss of connection because that would cause an inversion - the 240V loads would leak phase L2 onto phase L1, causing L1 to be at maybe 90V but in L2's phase. (or 30V between them).

So I think you have a loose connection but not a total break.

Is it somewhere along the cable?

People think that, because the largest area of exposure risk is there. But no - that is vanishingly unlikely, especially for individual wires which per Code must be inside conduit.

Almost always, wire problems are at or near terminations. Since you have been inside the intermediate splice box looking, I would check there first since it it outdoors. Also at the panel and any other splice.

When I say "check a splice" there are methods you can use thermally with a sustained load, but really, the best/safest way is to physically unscrew the splice, remove the wire, examine the condition of wire and splice device, clean up anything that is nasty, add no-Al-ox paste (optional but helpful), reassemble, and torque to the specification. The last bit is often overlooked by DIYers, and science has proven it causes many connection failures on copper wires and aluminum is not immune to the same problems.

Aluminum wire is perfectly fine, and is really the only wise choice for a long run like this. However you must use terminals properly rated for aluminum wire (although almost any terminal designed for 4/0 ought to be, since the vast majority of 4/0 installations are aluminum).

So make sure all that stuff is correct including the torques.

This cable is buried. What could be causing this voltage drop? I've considered moles or groundhogs...

Not likely. Direct burial is under at least 24" of cover. If in plastic conduit, it's at least 18" of cover. That's one enterprising rodent. Thick-wall steel conduit can be 6" cover but no rodent is getting through that.

Neutral and ground are supposed to be isolated at the subpanel

That way the problem cable will become the neutral. Since the neutral is connected to the ground, and the panel is grounded at the barn and connected to the ground from the house, if the problem is a hole in the insulation of the red cable (and, as of now, there's no GFCI on either end - and inspectors were okay with that), then we figure that cable may have a very tenuous connection to the earth ground along the way. Since the neutral connects to the earth ground anyway, it seems like that would be okay.

No. Neutral and ground are totally unrelated systems. They are isolated from each other in all subpanels (if the subpanel was installed correctly) and are only bonded at the main panel. The way they are often bonded is quite confusing and makes people think they're the same thing, but they're not.

Neutral is designed to carry normal return current. Ground is not expected to have any current at all except during a fault condition; the rest of the time its job is to be safe. It can't be safe and carry normal return current; that's why 3-wire feeds to dryers and ranges were banned.

You need the long ground wire to return fault current, to assure you get a breaker trip and don't just float the grounding system at a high voltage. You need the ground rod to arrest lightning and ESD and protect you from natural voltage gradients across the land (e.g. from nearby lightning strikes).

Note that when the hot wire is enlarged for voltage drop reasons, the ground must be too. You enlarged the hot wires from the mandatory #1 for 100A, to #0000, so -4 numerical sizes. The normal ground for 100A is #8 copper so #6 aluminum, so a proportional ground increase is #4Cu or #2Al.

Switching the "dead" wire to neutral would make things much worse. A "lost neutral" is more damaging than a lost hot, because without anything to keep neutral in the middle, both legs will add up to 240V but will vary wildly - one leg being well over 120V.

So if I plug a computer into an outlet, it might get 145 instead of 120? And something else could get 95 instead because it's on the other leg?

Much worse than that. I'd expect numbers like that for a house near other houses, because at the main panel, there's the ground rod/dirt as an alternate path. On a subpanel there is no alternate path, so swings will be extreme.

Transformers

Is there something I can use that can take the two legs as input and output two legs and a neutral?

For now, 100 amps in the barn. At some point, we could change that to 200 amps.

Well, that 4/0 wire for 500' + ??? for 60' is not going to work for 200 amps. Your best bet is a pair of "step-up" transformers. (or to be more precise, your least bad bet). They would need to be 50 KVA, but they can be commodity 240/480V-120/240V transformers which are readily available and occasionally even seen on Craigslist.

At the house you would step up from 240V to 480V. At the barn you would step down from 480V to 120/240V.

This connection only needs 2 conductors, so your lost conductor won't be a problem.

Doing it again? Conduit.

You mentioned individual wires and then have not said a word about conduit. One of the few ways damage can occur "in-line" on a wire is when the ground shifts and settles, and sharp rocks dig into the wires. Of course you do your level best to minimize this, by laying the cable in a bed of several inches of sifted/fine dirt or sand, and covering it with several inches more before backfilling. But it can still happen, e.g. from freeze/thaw or with vehicle traffic.

So if you find yourself digging it up, consider conduit the next time around.

Answer 2

I don't see anything in your narrative about simultaneous measurements at the panel in the barn and at the junction box where the 4/0 cable starts near the house. Two people on their phones should measure voltage at the same time at both locations using two voltmeters.

In addition, when the measurement at the panel in the barn indicates a large voltage drop, measure current (amps) on each conductor using a clamp-on ammeter at the junction box near the house. This will help distinguish between a short to ground in the buried cable (high amps) versus a bad connection or a near-break in the cable (no amps).