I live in Maryland in a house built in 2007. The house has 400A service so I have two main panels of 200A each. There are two electrical service lines, one for each panel and two solid copper ground lines in addition to a gang of ground wires that are part of the service lines. I also have a 20KW generator with an Automatic Transfer Switch connected to a large circuit breaker. Everything has worked fine until now.

I'll call the two main panels #1 and #2. Panel one has lots of 120VAC breakers with two or three 240VAC breakers. Panel #2 has more 240 connections and the ATS breakers in on #2 as well.

The problem started when I went to do some welding in my garage/workshop - despite trying several methods, my TIG welder wouldn't arc. Then my 240V built-in microwave failed and blew a fuse. Next, we had a hurricane blow through; we lost power, the generator cranked up, but the ATS didn't transfer to generator power. All three devices are on panel #2

With three 240v devices having trouble I got a little suspicious that the panel might have issues. Suspecting a grounding issue, I took out my clamp meter and started looking for current flow on the neutrals and grounds. I removed all panel covers and put a good amount of load on both panels before I tested. Panel #1 showed zero current flow on the neutral and the ground, but panel #2 showed near zero on the neutral but close to 1A on the ground. I unplugged the welder and microwave to isolate them, but did not throw the breaker to the generator - got the same readings. Devices like welders (I have a high power transistor (IGBT) welder) and microwaves are jammed full of sensitive electronics - they need very clean grounds; ditto for the control board in the generator.

The connections in panel #2 all look tight (not planning to play around in the panel with a screwdriver) and bonding looks correct. I'm not sure whether the problem is inside the panel (my problem) or the meter box (power company's problem). I don't know if the grounds go through the meter box or are fed directly from a connection to the buried cable. The panels are inside my garage/workshop which is insulated and drywalled so I can't see the routing.

Is there any way to isolate the source of the problem so I know who to call?

  • 1
    Voltage readings at different outlets. All should be ~120v, test with and without 240 breakers being on/off. Hot to neutral and hot to ground. Does your welder use neutral or just two hots and ground?
    – crip659
    Commented Oct 9, 2023 at 17:15
  • The power company swung by - the code doesn't allow the underground bar to pass through meter box and the box checked out perfectly - so the problem is on my side. I'll check the outlets per your suggestion. The welder is two hots and a ground - 3 prong NEMA 6-50. Commented Oct 9, 2023 at 19:46
  • 1
    Take readings while you turn off one switch after another. Does it change?
    – Martin
    Commented Oct 9, 2023 at 21:41

2 Answers 2


Is there any way to isolate the source of the problem so I know who to call?

If you're not going to try to DIY this, call an electrician. No further information is required. If you were having the same problem on both panels, you might have a lost neutral from your utility which would mean you should call the utility and report an outage, but that's not the case.

Devices like welders (I have a high power transistor (IGBT) welder) and microwaves and jammed full of sensitive electronics - they need very clean grounds; ditto for the control board in the generator.

This is not true. Ground is #1 for human safety (exposed metal is grounded so that if a hot wire touches it, the resulting short circuit will trip a circuit breaker or blow a fuse instead of electrifying the metal and shocking someone, possibly fatally) and #2 enclosures are grounded to keep electrically-noisy devices from emitting RF noise that may interfere with other devices / to keep ambient RF noise from interfering with sensitive devices (like radio receivers and computers). The devices you list would function just fine without ground (although using them would not be safe due to #1).

panel #2 showed near zero on the neutral but close to 1A on the ground.

Since you only turned on 240V loads, this doesn't tell us a whole lot. The expected current on neutral with 240V loads would be 0A anyway. The 1A you're seeing on the ground wire could just be due to a poorly-calibrated meter, or it could be due to small 120V loads in the "240V" appliances and a broken neutral feeder wire to panel #2.

A better test would be to turn on a large 120V load or two (two hair dryers or space heaters would be good, if you have them). Make sure the loads are both on the same leg of service if you activate loads on different circuits (i.e. on breakers that are not in adjacent rows). You should see the same current on one of the feeder hot wires and the feeder neutral, and nothing on the other feeder hot and nothing on ground. If you see no current on the feeder neutral and say 12A or 24A on ground, you definitely have a bad neutral connection between your meter and panel #2 - the current is making its way over the ground wire, back to panel #1, then via panel #1's neutral wire back to the meter (and power company). As ground bonding wires aren't required to be anywhere near as big as your 200A feeder neutral wire, this could - in the pathological case of 160A of 120V load all on one leg - overload the ground wire, possibly causing a fire.

  • 1
    The power company had a truck just down the street so their tech popped over to ensure all was good on their side. Commented Oct 9, 2023 at 21:23
  • Thanks for clarifying the ground issue regarding sensitive electronics. When I first moved in I had a lot of induction problems that were traced to a line running to a heat pump draped over a bunch of smaller lines - the problem disappeared when the line was rerouted. Had ground loop issues as well. I did have a few 120 loads on within Panel #2 when I tested, but as you point out not enough to prove anything. I'll add bigger loads and repeat per your directions. Commented Oct 9, 2023 at 22:07
  • To help me get clear, and explain what I saw inside the Meter box when the tech pulled the cover - the 200A feeders are two thick hots and a load of twisted copper wire - I assume the twisted wires are the neutral and are indeed routed to the neutral bus. The ground is a 1/4" solid copper rod attached to the ground bus. Neutral white wires and ground wires are mixed on both busses, not the neatest work I've ever seen but it passed inspection. Thanks also for you comment about my clamp meter - it's ancient. I just ordered a Fluke to hopefully eliminate erroneous readings. Commented Oct 9, 2023 at 22:07
  • I just received my new clamp meter and went to test panel #2. I don't have many 120 breakers in this panel but I have two on one leg - a string of flood lights on one, and I plugged in an area heater on the other. Together they pull about 15A. I turned off all breakers on the other leg. I see the 15A on the feeder but I get about 12A on the neutral and around 3A on the ground... almost as if the current was splitting the two paths. But there should be zip on the ground. When I turn off the two 120 breakers most of the current on the ground drops off except for a .25A. Time for repairs. Commented Oct 12, 2023 at 18:39

Electrician has a lot of work to do as the problem turned out to be multiple issues. Almost every connection in the main panels was loose, with a few burned wires as a result. The ATS suffered from a very poor bus ground. The incoming earth ground is one wire running to panel #1 and looped to panel #2 (would have preferred an independent earth ground to each main panel). The service panel on the dock was badly corroded compromising its isolation from ground.

Proper grounding of the ATS instantly restored reliable generator swithch-over. The multiple fixes to the various panels did assure separation of neutral and ground - every circuit in the house was checked and the system is free of any signification voltage on neutral or ground.

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.