4

I uncovered a metal junction box that was behind a wall (I'll fix this, I know it should be accessible). I looked inside and found the following: a 20-amp circuit (12 ga wire) and a separate 15-amp circuit (14 ga wire). It all looked good at first; however, then I noticed that the ground wire from the 15 amp was connected to the ground wire for the 20 amp. The circuits are not otherwise shared (hot and neutral of each circuit are separated). I think the person did this because the 15 amp circuit coming from the breaker uses a flexible metal tube with 2 wires with fabric coating inside (the house was built in 1940 or so), and the circuit was "extended" using 14-2 w/ ground to include some other receptacles and lights. Technically the metal tube would be clamped to the box and the ground wire for the 15-amp circuit attached to the box with a green screw. Should I fix this or is it ok to have the two circuits share a ground? I did a lot of searching and it looks like they can share as long as the larger gauge can handle 70% of the smaller, which it can (70% of 15 is 10.5). I asked a friend who is an electrician and he said it was ok; however I want a 2nd opinion. Thanks.

Edit: this is different from another, yet similar, question because this involves two different amp circuits, and this one involves BX cable.

Edit again: Thanks for the answers. Running a new ground from the box to the bus is an unrealistic option. Regarding another comment: The 12-ga ground wire isn't connected to the BX except for incidental contact with the box.

Since there's room, should I separate the two in separate boxes? I would ground the 14-2 to the metal box and be sure the BX is clamped good and tight.

enter image description here Picture description: lower right 12-2 connects with yellow 12-2. Armored comes from bottom and connects with the two 14-2 coming out the top. (I traced the armored to the box, so it is grounded the best that stuff can be; built in 1940.) Ground wires (grounding conductor) for both are twisted together. There was a green screw in there, showing evidence that at some point the possibility that the grounds for the 14-2s were connected to the box.

  • @mjohns Standards don't go down. It would be very strange for an inspector to be concerned that you followed code later than thay adopted by the locale. – bib Jul 21 '15 at 22:29
  • 1
    Could you add a photo or two of the wiring? – Tester101 Jul 23 '15 at 10:59
  • 1
    To be clear, the "15-amp circuit" enters the box via two wires inside a "flexible metal tube", and leaves along 14/2 w/ ground nonmetallic sheathed cable. The "20-amp circuit" enters, and leaves using 12/2 w/ ground nonmetallic sheathed cable. Is that correct? – Tester101 Jul 23 '15 at 10:59
  • 1
    Side note, you need some form of cable clamp for your 12/2 that enters the box via the bare knock out. You don't want that cable rubbing against the bare sides of the knock out. – mjohns Jul 24 '15 at 14:22
  • 1
    @SkiDude, your yellow-sheathed 12/2 that enters the box on the right side is not clamped. – mjohns Jul 24 '15 at 16:42
3

It doesn't sound like the #14 and #12 grounding wires being connected is the issue, to me.

EDIT:

NEC 250.148 (C) Metal Boxes. A connection shall be made between the one or more equipment grounding conductors and a metal box by means of a grounding screw that shall be used for no other purpose, equipment listed for grounding, or a listed grounding device.

The OP uploaded a picture showing that the #12 is an independent circuit, in new 12/2 w/ground NM cable. So what you actually have is grounding via the BX armor, presuming that the BX tightly clamped in both this box and the service panel, and grounding through the #12 wire in the 12/2 circuit, and the two should be bonded together to the metal box.

The ground path through the BX is likely to have higher resistance than the ground path through that nice clean #12 copper. The metal box has to be grounded no matter what. I say leave all the grounds bonded together and to the box, and make sure the BX clamp is clean and tight at the junction box and back at the panel. If you're really concerned about the grounding through the BX sheath, try to run another lone ground wire back to the panel, or pull new NM cable to replace the BX.

Also, you would get a much tighter connection between the armored cable and the box using a better clamp than the one built-in to that old box, something like one of the following (that red plastic bushing protects the wires from sharp edges, so you never energize your grounding system accidentally):

enter image description here enter image description here enter image description here enter image description here

END EDIT

All of the grounds in your house are ultimately bonded together, anyway, to zero the voltage differential all over your house (Neutrals are different! The neutrals are only bonded to the grounding bus in the main panel, nowhere else, and neutrals can only be shared by two circuits under specific conditions).

You wouldn't want a large circuit using a too-skinny conductor for grounding, though.

It sounds to me as if the original circuit is possibly BX armored cable, and the flexible conduit itself is the only grounding conductor going back from this junction box to the panel.

This not up to current code if the flexible metal conduit is the only ground path--there can be continuity problems through the flexible armor. BX was actually demarketed for a while, then came back as armored cable with a metal bonding strip that makes contact with the flexible conduit all along it's length to ensure a good ground path.

However; if you can somehow run a #12 bare or green-insulated wire back from that junction box to the panel (route it up through the attic?) and bond it to the grounding bus inside the panel, as well as to both the #12 and #14 wires, and the BX armor in the junction box by using the right metal clamps and bonding the wire to the metal box, you'd be perfectly fine.

Having said all that, I guess somebody could argue that some grounding through the BX armor is better than no grounding. But if you're in there messing around, you want to do it right and bring it up to code.

  • To the point of running a new ground, I would consider running new 12/2 w/ ground to replace the old cable assembly. To me, the NEC exception for running separate grounds like this is questionable, especially when you're talking about taking a completely separate path for the ground. It might not be inherently unsafe, but it's certain to be unexpected by the next guy working on the house. – mjohns Jul 24 '15 at 12:06
  • @mjohns If it's remotely reasonable (e.g. no walls have to be torn out) to run new 12/2, I 100% agree with you. The only redeeming grace here is that the grounding wire isn't actually part of the circuit (totally unlike the neutral, which is a live circuit conductor). – Craig Jul 24 '15 at 15:46
  • I'm a DIY-er, so gotta keep this simple. So, is it ok if they share or should I separate them in different boxes and ground the 14-2 to the metal box? – Ski Dude Jul 24 '15 at 16:08
  • @SkiDude connecting all the grounding wires in that box is okay, even preferable. The grounding all over your house is supposed to be tied together into one equipotential unit. Electricity doesn't flow unless there's a different in voltage potential between two objects (they could both be at 100,000 volts potential, and no current will flow). So you want everything in your house at the same potential. Thus grounding. I think I understand better now that you have a good ground in the 12/2 circuit. You have grounding (though potentially spotty) through that armored cable sheathing... – Craig Jul 24 '15 at 16:59
  • The grounding wires from all the 12/2 and 14/2 cables absolutely have to be connected to the metal box and to the old armored cable (and to the grounding bus in your panel, and from there to your water pipes and a grounding rod or two outside). In other words, you have to connect them all together. The question is whether you need to replace the armored cable with new 14/2 or 12/2, or maybe just run a single ground wire back to the panel to augment the grounding you're getting through the 12/2 and the potentially higher resistance ground you're getting through the armored cable sheath. – Craig Jul 24 '15 at 17:03
-1

I think this is partially common sense. If the total combined load exceeds 2x the branch circuit grounds capacity then you should not piggyback more grounds into the branch circuit.

For example a # 10 copper ground will hold up to 60 amps. This is why #6-3 NM ground is sized as #10, not as a # 6.

That is 2x the capacity of the non-grounded conductors.

Anything over 60 amps is specifically specified in the NEC.

  • A #10 copper wire will actually conduct a good deal more than 60 Amps for quite a while before getting so hot that the wire melts, but somewhere in the interim you have the potential for fire. The wire ratings ensure that a wire never conducts enough current that it gets so hot that it poses a fire risk. The ground should never be carrying the full current of the circuit for more than a moment before a breaker trips. But stuff happens... – Craig Jul 22 '15 at 22:51
  • I'm sure it will. I was just stating what the NEC permits for a circuit <= 60 amps. – Kris Jul 23 '15 at 2:07

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.