On my kitchen ceiling, there are two light boxes controlled (I hope?) by one dimmer switch. The boxes are connected to the far wall (opposite the switch) by a conduit that runs along the ceiling; however, some wires stick out directly from the ceiling, not from the conduit.

               add'l wires 
  |]========[boxA]====[boxB]                      |
  |      ^                                        |
  w     conduit                                   w
  a                                               a
  l                                               l    
  l                                               l
  |                                               |
 [|]                                              |
 [|] <- breaker                       dimmer ->  {|
 [|]    panel                                     |
  |                                               |

Box A only had two loose (capped) wires: a black and a white. I tied them to the respective black & white wires of a pendant fixture, and all was well. That light is now controlled by the dimmer switch exactly as I'd expect. For the sake of completeness, I'll note:

  • There was another pair of black Romex running thru box A, but it didn't have any ends exposed, just ran straight thru the conduit toward box B.
  • Box A doesn't appear to be grounded? I ran a green grounding wire (which came with the fixture) from the pendant's canopy to the junction box, but there's no connection to the condo's ground wiring that I can see.

Meanwhile, over in Box B, I found a total of 6 wires, which were joined together into 3 pigtails.

closeup of box B wires

  1. Black Dark Grey + White
  2. Black + Black
  3. Blue + White

There is continuity between the black grey wire in pigtail #1 and the blue wire in pigtail #3. (Note that these are the two wires coming from the conduit; the other four come from the ceiling). No continuity between any other pair of wires.

So, questions:

  • How the heck do I wire my second pendant fixture into box B?
  • Does the wiring in box A need to be modified for proper grounding?

Edit: here's a zoomed out photo

photo of the whole ceiling/wall

^^ From left to right, you can see the wires coming out of box B, the pendant successfully hanging from box A, and the breaker panel underneath the shelves on the far wall. I separated the box B wires for testing (i.e. the pigtails seen in the closeup photo above are no longer connected), so the pendant obviously does not light up anymore.

Followup Edit: by popular request, I've disassembled box A for photographing, and taken a bunch more measurements.

closeup of Box A

^^ Here you can see the single black and white wires coming out of the ceiling, which I'd previously wired into a pendant, and also the conjoined pair of wires (black+grey) running thru the conduit. As for continuity:

  • Box B pigtail #2 black [the one paired with black] == Box A's loose black wire
  • Box B pigtail #3 white [the one paired with blue] == Box A's loose white wire
  • Box B blue == Box B dark grey == Box B ceiling plate == Box A ceiling plate

The main discovery is that with better lighting, I can see that the dark wire from Box B's first pigtail is ever so slightly lighter than the other two black wires. I've relabeled it "grey" above, since it appears to be grounded.

  • Can you provide more photos of the situation? Feb 3, 2018 at 14:35
  • How is the connection from the panel to the first box made? Feb 3, 2018 at 14:37
  • @ThreePhaseEel I added a photo of the overall situation, showing the panel + conduit + boxes. No idea what kind of connections are happening behind the walls, but I do have a multimeter if there's more tests you'd like me to run. Feb 3, 2018 at 17:12
  • Turn the breaker for the circuit off, then see if there's continuity between the neutral wire to the fixture and the conduit, for starters Feb 3, 2018 at 17:56
  • You're going to need to test this extensively, especially if you didn't mark the whites and blacks. Obviously there's a switch involved, a supply involved, and a run to the other lamp (blue for neutral???) By the way those singular wires are actually called THHN, I doubt two THHN and a balky stiff /3 Romex cable would fit in Legrand Wiremold... Feb 3, 2018 at 18:49

1 Answer 1


Well, here's what I think you've got. You know that the lamp in Box A works and once you traced the wires it was easy to chase it back from that.

The wiremold surface conduit brings power (always-hot and neutral) into the system.

Nut 3 is neutral, using (very bizarrely) blue for neutral. Which is the standard in Europe (though they prefer a robin's egg blue) - not legal in North America (white or gray please).

Oddball colors like that are possible in conduit (such as this surface mount) since they are THHN individual wires. Possibly the other wires are multiconductor cable (NM Romex or metallic AC) always black-white.

Nut 2 is switched-hot for your lamp. Black is a legal color for switched-hot (or any hot) but red is ideal.

Nut 1 is always-hot. Note that it's using a white wire in the switch loop, which is actually a Code requirement when using multiconductor cable like NM or AC however the wire must be taped to identify it as a hot.

In this diagram, I go well past the mandatory markings to mark every wire with "tape" whose native color doesn't match its function. All these markings are legal except the blue wire - neutrals must be natively white or gray (even in cable), and blue here is a code violation that isn't fixed by marking it.

enter image description here

My guess is before the condoization, it was powered at the light switch off a circuit that served other areas of the office which are now part of other dwelling units. That breaker would've been in someone else's condo, which violates Code. This work was done competently by an old-head who knew what he was doing... except the blue wire! Perhaps a European immigrant...

As for grounding, I would pop the cover off the "surface conduit starter box" over on the right wall, and see what's behind it. If you see a steel junction box with tubular conduit feeding down to the panel, that EMT or IMC or rigid conduit is a valid grounding path. The back of the box will have a hole tapped #10-32, I'd get a ground screw (they sell them at electrical supply by the one, or big-box by the ten, they're even green so you don't mix them up with general hardware) and some #12 bare ground wire, and bring it to each of the junction boxes via the Wiremold surface conduit.

  • It worked, thanks. I would've never guessed the dark wire from the Wiremold was always-hot, since it had continuity with the ceiling plates. I guess that's due to an incandescent light somewhere on the same circuit... Also, I did not find continuity between the always-hot and switched-hot wires, even when the switch was fully "on". In ohmmeter mode, my probe could tell something was different from truly disconnected pairs of wires, but it got really confused and jumped around in the mega-ohm range before giving up and calling it "infinite". Guess it doesn't like dimmers! Feb 5, 2018 at 1:53
  • Yeah that confusion is typical of when it sees something with capacitance, the ohmmeter test current slowly charges the capacitor to equilibrium, slower as it fills, and the apparent resistance goes upward to infinity. Feb 5, 2018 at 4:04

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.