I have a 125A GE PowerMark Gold Load Center "subpanel" at my detached garage. Subpanel has 2 hots, one neutral with no "connected" ground between panels. The subpanel ground bus bar is tied to a local 8ft subterranean 5/8" grounding rod. 50A branch breaker at main feeding 6/3 wire about 100ft tied directly to the subpanel inlets (H/H/N). Per NEC, the neutral and ground in the subpanel are not bonded (i did not screw in the green bonding screw that came with the panel).

GE Load Center Image

All the outlets including GFCIs all work correctly. When I test panel buses with my meter in the garage I have 120v between neutral and each hot and have 120v between ground and each hot. Sounds good right? Then why do I have continuity between neutral bus bar and ground bar also between neutral bus bar and panel box itself?!? I thought that if I don't insert the screw, then neutral should be "separate" from ground and panel metal box? Or is this normal?

  • 1
    @SeanS -- why on earth is that grounding conductor not used!? And is that UF in the conduit, or NM!? – ThreePhaseEel Sep 22 '20 at 1:28
  • 1
    @EdBeal No, it isn't wired to code. The picture they posted and the description they gave is enough to figure that out. The feeder conduit is plastic and the ground wire is cut at both ends. In addition to not being to code, though, (because of the missing ground) there appears to be a connection between the neutral and the ground rod ground in the sub from all the information OP has given. I'd like them to confirm with some measurements, but OP does not yet have the right answer here. They should absolutely be reading 10s or 100s of ohms between neutral and rod-ground and they're not. – J... Sep 22 '20 at 16:33
  • 3
    OP I am concerned about several things about the installation, and would much rather see photos of the panel, cable entry, and various neutral and ground bars. I am concerned perhaps the N-G bond was not pulled. And I want to see if that ground wire in the feeder is recoverable; deleting that was sheer negligence. – Harper - Reinstate Monica Sep 22 '20 at 19:00
  • 2
    “For detached, it was to be eliminated and the ground was to be tied to a dedicated ground rod at the subpanel” Code never said that... since like 1990, you need both. – Harper - Reinstate Monica Sep 23 '20 at 22:23
  • 2
    @SeanS Hold on... are you sure that's an L6-30R or any of the NEMA 6 family receps? Those don't have neutral. That third pin is safety ground. Not only that, but the ground screw/lug/terminal is internally connected to the receptacle's yoke, as you are measuring. Yes, problem found, I'd say! If you want a receptacle that is H-H-N, you don't. You don't want a receptacle that is H-H-N. They're dangerous, and they've been outlawed since '96. Use NEMA 14, which is H-H-N-G. – Harper - Reinstate Monica Oct 7 '20 at 1:09

Edit: Found it! It's a miswired NEMA 6 recep.

NEMA 6 family receptacles provide Hot, Hot and ground. Someone had miswired a NEMA 6 recep by connecting a neutral wire to the recep's ground lug.

The ground screw and metal chassis of the socket are internally connected, just like any other ground screw on any other receptacle or switch.

One should never wire a neutral wire to a NEMA 6. That violates a lot of stuff, but also 110.3(B) requiring you follow instructions and labeling.

There's no such thing as a receptacle that provides hot-hot-neutral. That is an illegal connection -- anything big enough to need 30A also needs a ground wire. The only options are NEMA 6, or NEMA 14 (hot-hot-neutral-ground - 4 pins in total).

If someone has a legacy appliance that needs hot-hot-neutral, then that needs to be wired up NEMA 14. If it's a range or dryer, it will have a "bootleg jumper" internally typing neutral to ground; that must be removed, as per its instructions.

Current on the ground wire

This is a very serious matter that HAS to be found.

To start with, we need to know that the panel is isolating the neutral and ground bars. Lift all your neutrals including the feeder (remove them from the bar) and then use a plain ohmmeter to check between the neutral bar and the ground bar. That must be infinity ohms. If not that's a problem with the panel config, like a N-G bonding screw that you missed.

Now with all the breakers off, reconnect branch circuits neutrals one at a time, and check again. Should still be infinity ohms. If adding a branch circuit causes it to go to low resistance, that branch has a ground fault (possibly neutral-ground). Chase that down.

I suspect you'll find something here. That is the likeliest explanation for current on ground.

Now I have an alternative explanation. It's possible your wiring is correct. Suppose your neighbor has a lost neutral. Neutral will get back to the transformer via their neutral-ground bond, their ground rods, dirt, your garage's ground rods, your feeder ground wire, your main panel's neutral-ground bond and your neutral back to the transformer you share. This current isn't yours in this case. Your neighbor has the problem.

Neutral bar

OK so that's a GE panel, it looks like it has 14 real spaces, 10 of which allow thin breakers (note croses on buses) so max 24 circuits.

I see a left-side neutral bar with about 24 usable screws. That would seem to suffice, so I'm unclear on the purpose of the right side bar. The documentation describes it as a "grounding bar", which settles the matter. It has its own chassis-grounding screw. So it appears this bar is unfit to be a neutral bar (a high-current path does not exist between it and the neutral lugs) and GE has simply insulated the ground bar for no apparent reason. While also providing an auxiliary location for an additional TGL2 grounding bar kit. Together they would provide 24 grounds, but that's usually not an issue since you're allowed to double-tap or triple-tap grounds.

This panel is definitely a puzzle. I can see why you asked.

Broken off ground

The broken off ground wire is a serious defect. You MUST have both proper grounding rods, AND a wired ground wire back to the main panel. That's mandatory.

This installation appears to be cable in conduit, which has "amateur hour" written all over it. If the conduit is continuous, the cable can be yanked out, and replaced with individual wires with an intact ground wire. Fair chance the installation is improper anyway; that's either NM cable (illegal outside) or UF cable (illegal in 1.5" conduit; requires 2" conduit); so that would need to be done. This would be legal if it was round SE cable.

The quickest way to restore the ground wire is to pull the cable back enough to cut the conduit and fit a steel junction box immediately above the panel with an EMT nipple connecting it to the panel. The steel/EMT will be a continuous ground. It's the quickest way of making the panel "taller" so the ground wire can reach an appropriate land (in the new box). You can't use the 1" stub of ground wire as it is; it needs to come 6" into the box to be Code.

  • 1
    The other possibility is that somewhere in the branch circuitry there is a neutral/ground connection. I think OP should disconnect all wires from the panel and check that the ground and neutral bars are isolated, then progressively add circuits back onto the busses, checking isolation at each point. Something is shorting ground and neutral at or past the sub. Either that or OP's ground rod has a buried line going all the way back to the house or something crazy like that. – J... Sep 22 '20 at 20:41
  • 1
    And true on the "Amateur Hour" observation. I am an IT guy, not an electrician. To be honest, I just watched some you tube videos and installed this entire thing "hot". Crazy of course. I wouldn't do that again. Anyway yes it was difficult getting that 6/3 Romex (NM) into 100+ feet of flex and conduit. – Sean S Sep 23 '20 at 19:16
  • 2
    @SeanS Sorry, it never occurred to me the conduit job was your work. Well, now you know; you can pull THHN. Unfortunately, you'll have to do it again because of the NM-outdoors or UF-too-big problem... or if SE, use one of the other remedies I discussed. – Harper - Reinstate Monica Sep 23 '20 at 19:59
  • 1
    @SeanS How did you measure that 3.2A? Between ground from the panel and the ground rod? With nothing else going on? That does seem like an awful lot, but it would depend on what you've measured. – J... Sep 23 '20 at 21:56
  • 1
    Yikes, ok well it will be easier to pull THHN if I have to redo. I am guessing i can pull the Romex apart and just keep those loose wires in the 100' flex/conduit? From panel it goes NM to under the house, then flex ~80 feet tied to subfloor, then thru garage, still same flex, then NM conduit underground for 15' to sub. For measurement, I used a clampmeter around Hot leg coming into the sub and around the Neutral and the 12ga ground. I had 8.1 on hot leg, 6.2 on Neutral and ~1.7 on ground 12ga the first time i measured. it varied slightly from there. – Sean S Sep 23 '20 at 22:39

The feed for your sub panel comes from your main panel the grounded conductor neutral and grounding conductor equipment ground are connected in the main panel so measuring continuity in a sub is normal even when the grounded conductor is isolated from the box and the grounding conductor. This is normal the resistance value is usually quite low just a couple of ohms or close to zero. The length of wires and any resistance at the connections is all there is.

  • Correct, I measured 0.01 ohm between sub-panel neutral bus and ground bus then the same - 0.01 ohm between neutral bus and panel box. Do you see the dropbox link i posted earlier of the inside of subpanel? Remember no ground tie between panels, only dedicated ground rod tied to subpanel ground bus. Does that match your description of connectivity above? So it's normal? – Sean S Sep 22 '20 at 1:02
  • 1
    Not only normal but required in the main. I have installed hundreds of panels that green bonding screw you did not use in the sub is required in the main or a jumper that ties both buss bars together. I had a customer tear a sub apart when he found the same totally destroyed the panel trying to find the short that’s why I said pull the neutral in the sub and measure it so you did not start taking things apart. It is fine and you have very good connections. – Ed Beal Sep 22 '20 at 1:12
  • Thanks Ed, You explained "measuring continuity in a sub is normal even when the grounded conductor is isolated from the box and the grounding conductor". I trust and will take your advice and leave things alone. – Sean S Sep 22 '20 at 3:55
  • It's actually the resistance of the earth that is being measured. The garage ground goes to a rod in the earth and the house ground comes from a rod or piping in the earth. What you measure between the neutral (house ground) and the garage ground (ground rod) is the resistance of the earth between those two grounding points in the earth. 0.01 ohms is way too low for this to be the case. Something is shorting the garage ground and the house ground or the 0.01 ohm was not measured correctly. – J... Sep 22 '20 at 12:33
  • 2
    @EdBeal I don't think you understand what OP has described in their panel. At the garage there is a rod in the dirt and all grounds at the garage tie to this rod in the dirt. At the house is a neutral, tied to a different ground in the house. There is no ground wire between the house and the garage. Measuring resistance between the neutral in the garage and the ground system in the garage is measuring the resistance between the dirt at the house where the ground is bonded to neutral and the dirt at the garage where the rod is stuck in the ground. This can't be 0.01 ohms, even on a cheap meter. – J... Sep 22 '20 at 15:29

It looks like the original feeder ground IS connected (well touching)[see the pic] the case and providing a ground path for the grounds back to the main panel where ground and neutral are bonded together (Ed's comment). This is certainly not a reliable feeder ground tie in; but maybe you can pull a bit of the wire to get more working wire and nut it to another longer wire and attach it to the ground leg of the box? Im not sure where the conduit with just a single bare copper wire on the right goes, is that out to the 8'copper earth ground? (but its not relevant to this discussion). enter image description here

  • It's not touching - look closer. It's not even close. – J... Sep 22 '20 at 21:14
  • The #8 ground does go to the ground rod outside. It's connected to the ground bar which does have the green bonding screw inserted, so the ground rod is connected to the ground bar and to the chassis (via the green screw and also the AC flex and incidental contact). The ground in the conduit is also clipped at the house (not connected to the inside panel). – J... Sep 22 '20 at 21:17
  • Ok, so to get Ed's "normal" condition where the sub panel's ground and sub panels neutral reference each other at <1ohm, there must be something tied together in one of the circuits attached. like a dryer plug with the neutral and ground tied...? or some other miswired plug. (pulling breakers wont id this, OP would have to pull plugs form outlets) – mark f Sep 22 '20 at 21:26
  • Im continuing this thread even though the OP closed it based on Ed's answer because I think the OP may have an unsafe condition, masked because it gives a "normal" impedance measurement for a properly set up 4 wire feeder. One of his Outlet Neutrals seems to be tied to ground. (other wise in THIS configuration he should get >>10Ohm grd to neutral – mark f Sep 22 '20 at 21:45
  • 1
    My concern (and Ed's if that cut off bare copper really isnt providing a connection) is that SOMETHING is attaching your ground to your neutral, and it COULD be miswiring in one of the plug/outlets. The confusing part is that normally, this would be the desired condition where the bond tie at the house provides this path, but since you only have a 3 wire feeder with NO ground to the house, something else is tying them together. See @Js idea to disconnect each neutral from the bus, and measure the box's neutral to ground impedance/resistance (WITH THE PANEL OFF) – mark f Sep 23 '20 at 22:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.