My house has a 200A main panel and receives power from a center tap transformer (240V split into 120V phases). The house consumes more than 10KW of energy. I have noticed in the past that outlets on one 120V phase will sit around 105V-110V and outlets on the other phase will be 120V-125V. Recently the main water line was replaced in the house and the voltage difference on outlets between the two phases spread further while the work was being performed (90V on one phase and 140V on the other). I found that the ground for the main panel runs about 75 feet from the main panel on the back of the house to where the main water line comes into the front of the house, and they had dug up and cut the pipe which was my ground, leaving only about 4 feet of grounded pipe buried. Once the work on the main water line was done voltage differences returned to their original levels.

From what I have read, it seems like the voltage difference between the two phases would be caused by a poor return path back to neutral. However, when the main water line work was being performed my ground was severely diminished which caused an even greater voltage imbalance which makes me wonder if I need a better ground.

I have inspected the connections I can see in the main panel by removing a cover below the meter cover (exposes the 3 main bar connections) and breaker cover (exposes breaker and ground-neutral bonded connections), and I do not see any corrosion or anything that looks like it isn't to code.


1. I thought if you remove the ground from the system completely, the voltage should still be balanced (assuming normal conditions with nothing shorting to ground), so why did losing a good ground cause a larger voltage swing between the two phases?

2. It seems like my problem is in the main panel since loads on one phase of the house cause a voltage swing on the other phase of the house (across all circuits in the house). How do I measure (like with a multimeter) to determine whether I have a grounding problem or a neutral problem?


My uneducated guesses may have led me astray. The important question ...

What is a good procedure to identify the source of the issue?

Answer: use a clamp-on amp meter to measure current through main neutral (I had a reading of 0A but an imablance of more than 20A between the two phases) and measure the current through the grounding electrode (I had 20A going through it). This confirmed that I had a floating neutral which the power company had to come out and fix.

  • Was the previous water line all metal ( ie; copper or galvanized ) and what was it replaced with?
    – BillWeckel
    Commented Oct 10, 2018 at 20:43
  • Original water main is copper, and they cut the original where it comes inside the house leaving the ground clamped to it. They then ran a new copper line a foot higher and use that to re-attach the water to the house. So now I guess I have a grounding electrode which is a copper pipe not attached to the water system which is probably buried for about 10 ft.
    – gregjhogan
    Commented Oct 11, 2018 at 3:45
  • @gregjhogan -- good on them -- that's actually a pretty nice grounding electrode they left you :) Commented Oct 13, 2018 at 0:39

3 Answers 3


Contact your power company and report an outage.

Specifically, a lost neutral. This is a serious problem that needs to be fixed lickety-split.

Your neutral is either gone or fading fast, and differential current is flowing through the neutral-ground bond, grounding electrode system, earth, to the transformer's ground and back to the transformer. Dirt doesn't conduct electricity well enough to do that job, which is why the voltages are going out of balance.

I bet the workers servicing the grounding electrode system were getting "zappity zap zap" while they were working on it. There should never be current flowing on ground wires.

  • Will do, thanks for the advice! I discovered that they actually ran a new copper pipe for the water main about a foot higher through the ground, but left the main panel ground clamped to the old copper pipe (no longer connected to water system). That probably needs to be fixed, and I measured the voltage between the two copper pipes (which are not connected in any way physically) and it is 0.6 VAC. Is that a good way to determine if excess current is flowing through my grounding electrode?
    – gregjhogan
    Commented Oct 11, 2018 at 3:55
  • No, I don't think that will be conclusive. The better way is put a clamp ammeter around the grounding electrode wire. Commented Oct 11, 2018 at 4:01
  • 1
    Power company came out and had to make a new neutral connection up on a pole. Old neutral had a aluminium to copper connection which had so much corrosion that we had no neutral connection. The main water line had 20-30A going through it to the ground, now it has 2A.
    – gregjhogan
    Commented Oct 12, 2018 at 23:22

You have a dangerous situation on your hands

Your service neutral is the wire running from the center tap lug on the utility transformer to the neutral lug on your main distribution panel. The normal function of your 120V circuits depends on this wire being in place solidly connected providing a low impedance path back to the transformer tap.

If the service neutral is not in place, and no low impedance path exists, the 120V loads on the one leg of your service are in series with the 120V loads on the other leg. This can lead to large variations in voltage on the 120V circuits. (Your hypothesis that the voltage swings you're seeing imply that the problem is in your main panel is incorrect.)

However the grounding system can provide a parallel path from your MDP neutral the utility neutral. For example, if your grounding system is bonded to a metal water supply pipe, and the metal water supply is continuous to your neighbor's water supply, and their water supply pipe is bonded to their house's grounding system, and that in turn is bonded to their service neutral, which goes back to the same lug ... there's a parallel path. If it's sufficiently low impedance, you may not see much voltage variation even though your service neutral is not in order.

The dangerous part

There are a number of potential hazards here. In comments it was noted that the ground system connection was not moved to the new metallic water supply. This explains why the voltage imbalance worsened after the work done on the water supply; the ground is no longer connected to a pipe that was providing an alternate low impedance path back to the neutral lug. Who ever fixes this must use extreme caution - the ground wires are generally thought to be harmless and under normal conditions safe to work on, but if someone grabs that new water pipe in one hand and the old ground wire in the other, they'll take a potentially fatal hand to hand shock.

In addition, with your service neutral out of order, the usual path through which ground faults are cleared is missing. Under normal circumstances, if for example your washer's internal wiring is broken and a hot wire comes in contact with the metallic frame of the machine, normally the fault would immediately cause high current to flow through the equipment ground and the breaker would open. In this situation, the ground path may not allow enough current to flow to trip the breaker. This will leave the frame energized, and you would get shocked when you touch the washer.

There are other hazards, ground shield on coax wires carrying current, and the problems / hazards may extend to your neighbors.

Use extreme caution, and get the utility company and / or an electrician to on this as soon as possible

  • Interesting re: water pipes being connected between houses. Also interesting, I found my ground is no longer attached to the water system (see my other comments, I now have a defunct copper water pipe connected as a grounding electrode).
    – gregjhogan
    Commented Oct 11, 2018 at 4:14
  • The metallic water pipe needs to be grounded where it enters the building - that should have been moved. WHO EVER REPAIRS THIS MUST USE EXTREME CAUTION. The ground is normally thought to be harmless but many have been shocked and killed in this situation. Commented Oct 11, 2018 at 9:22
  • Yeah, I mentioned in other comments I thought the ground needs to be moved. However, there is still a ground and the voltage difference between the phases returned to the original levels after the work was complete because the defunct copper water pipe is underground, but buried only about 10 feet now that it is disconnected from the water system.
    – gregjhogan
    Commented Oct 11, 2018 at 20:14

I have seen several homes that the water service lines have been replaced with plastic. This was the only ground to the system so it would not meet code requiring a minimal of 10' of metal in contact in contact with earth. This not a utility problem and can be easily fixed if not in a desert zone by driving 2 new ground rods at least 8' deep 6' apart and running a 6 awg wire to the ground or case of the main panel. Some loads on an unevenly balanced split phase panel can make the imbalance have a larger swing from l1 to ground compared to l2 to ground. Your statement that the metallic plumbing was changed is a BIG RED FLAG! , next to check after driving 2 ground rods and connecting them to the panel. The loads needs to be ballanced l1 compared to l2. The amps on rack leg should be close to the same value. Code only states well balanced but no specific values. I have found a imbalance over 20% not only creates a voltage imbalance but even creates additional ground to neutral voltage that over ~ 2.5 volts ( if I remember correctly) will cause problems with switching power supplies (the most common type of supplies powering computers, cell phone chargers, Flouresents lights, CD/ DVD players) . So first get your ground rods installed a couple of rods and some #6 or larger wire to your main panel is all that is required by code and should do the trick if your panel l1 & l2 have similar amperage. Last if you are in a desert it may take a much deeper ground rod to achieve a low resistance value, but as long as 2 rods are driven code ignores the single rod 25 ohm requirement that is required for a single rod.

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.