Circuit end mismatches panel connection

Question

Some years ago, before finishing my basement, I ran a Romex 12/3 cable from my main unit's panel to the kitchen to supply the dishwasher and garbage disposal, each on its own 20A circuit. I ran the cable through the ceiling of the basement, which is the floor of the main unit. I connected the red wire to one breaker and the black to another.

I am now working on this kitchen and I have discovered, by testing, that the black wire from the panel side comes out as white (neutral) on the kitchen end, which means the white is hot and black is neutral. The red wire matches fine. I arrived to that conclusion by using a plug tester with three lights as well as a multi meter.

Following logic and parsimonious reasoning, the most realistic explanation for this oddity is that somewhere along the route of the cable, there is a connection, as in a junction box or, god forbid, nut splice behind the wall, that connects the panel black to the kitchen white and vice versa (scenario 1). To think that the black inside the cable magically becomes white and vice versa (scenario 2) would be highly unparsimonious and akin to superstitious thinking, right?

Scenario 1 would be easy to accept if this cable was legacy that came with the house, done by some drunk dude who knows when. The problem is that it was done by me and that there is no way in hell I would splice an illegal connection behind a wall AND, worse yet, I cannot locate a junction box where this nefarious connection takes place.

Is there a way for me to test that the same continuous cable that starts at the panel is or is not the one that comes out on the kitchen end? Like can I look for a serial number of the cable on each side? If not, how would I go about finding where this transition takes place? It is not excluded the possibility that I connected it in another receptacle used for something else, like a light or smoke detector. Essentially, I'm looking for a way to audit the route of my cable behind the wall as its end doesn't correspond to its beginning.

If I can't find it, which is a scenario I don't want to accept, which would be like a crashed airplane without a black box, should I just treat the white as black and vice versa in the kitchen because it just works that way? I hate black box scenarios but I have limited time to dedicate to resolving this issue.

I know I should have kept documentation where each cable goes and all the connections it makes but stuff got chaotic over the years and there is a lesson to be learned.

UPDATE: As per the suggestion of @Harper, here are some "raw data" voltage testing results, conducted on loose wires on the kitchen end with a multimeter.

KITCHEN SIDE VOLTAGES BETWEEN WIRES (rows) UNDER 
DIFFERENT BREAKER STATES ON PANEL SIDE (columns)
-----------------------------------------------------
     | Both On | R-On/B-Off | B-On/R-Off | Both Off |
-----------------------------------------------------
W-G  | 120     | Low        | 120        | 0        |
-----------------------------------------------------
B-W  | 120     | Low        | 120        | 0        |
-----------------------------------------------------
B-G  | 0       | 0          | 0          | 0        |
-----------------------------------------------------
R-G  | 240     | Low        | Low        | 0        |
-----------------------------------------------------
R-W  | 120     | 120        | Low        | 0        |
-----------------------------------------------------
R-B  | 240     | Low        | Low        | 0        |
-----------------------------------------------------

Here are some "raw data" voltage testing results, conducted on loose wires on the panel end with a multimeter. Data points differing from the kitchen end are marked with (*)

PANEL SIDE VOLTAGES BETWEEN WIRES (rows) UNDER 
DIFFERENT BREAKER STATES (columns)
-----------------------------------------------------
     | Both On | R-On/B-Off | B-On/R-Off | Both Off |
-----------------------------------------------------
W-G  | 0 *     | 0 *        | 0 *        | 0        |
-----------------------------------------------------
B-W  | 120     | Low        | 120        | 0        |
-----------------------------------------------------
B-G  | 120 *   | Low *      | 120 *      | 0        |
-----------------------------------------------------
R-G  | 120 *   | 120 *      | Low        | 0        |
-----------------------------------------------------
R-W  | 120     | 120        | Low        | 0        |
-----------------------------------------------------
R-B  | 240     | Low        | Low        | 0        |
-----------------------------------------------------

Answer 1

You should not leave the wires reversed this would be a code violation and may get someone hurt in the future. As all junctions are required to be in boxes and the boxes being accessible , you may want to invest in a toner and trace the circuit with the breaker turned off. Simple toners are not horribly expensive and put a signal on the wire that the hand held amplifier picks up so you know the wire is the right one.

Answer 2

I ran a Romex 12/3 cable ... to supply the dishwasher and garbage disposal, each on its own 20A circuit.

Those aren't 2 circuits. That is a multi-wire branch circuit, and it has a bunch of special rules, and we'll just pretend your work predates those rules.

However, the takeaway here is that MWBCs are fairly complex beasts with interesting side-effects that may challenge the novice.

I connected the red wire to one breaker and the black to another.

And there are rules about where you must place those breakers. Generally you must use either a 2-pole breaker, or 2 handle-tied breakers in the exact same positions as a 2-pole breaker will occupy. Note the handle-ties are required, and forbid use of a double-stuff breaker. All manner of woe will follow if you put them on a tandem.

the black wire from the panel side comes out as white (neutral) on the kitchen end, which means the white is hot and black is neutral. The red wire matches fine. I arrived to that conclusion by using a plug tester with three lights as well as a multi meter.

Not frickin likely.

This is a false reading due to the oddities of MWBCs, intersecting with notoriously bad testers (magic 8-ball testers are particularly notorious in giving you laughably wrong indications), intersecting with a lack of the depth of experience you need to troubleshoot MWBCs.

The wires are not transposed at all. That did not happen.

You did not do an illegal connection.

You did not leave yourself 18" of slack somewhere so that you were able to insert a legal connection that you plumb forgot about.

Treat the 3 above points as absolute fact (because they are) and cross off any conclusion that does not agree with those facts. Peel the label off the magic 8-ball tester and throw it in the trash, and cross off any theories that flowed from anything that label said.

Things actually worth looking for are bad connections at terminations, breakers which indicate on but are off, breakers mis-phased (there MUST be 240V between red and black; this is mandatory!), faulty bus stabs, shorted (nail through them) wires, and broken wires.

Voltage must be:

• Black-N: 120V
• Red-N: 120V
• Black-Red: 240V (or 208V if NYC)
• N-G: less than 3 volts

This is a more accurate label.

Answer 3

You've installed a multi wired branch circuit. The black wire on one breaker and the red wire on another breaker. The two share a neutral and the breakers should be next to each other with handle ties or a double pole breaker. Voltage between the black and red wires should be 240V and 120V from black or red to white or ground. Start out by checking this whole circuit again. You can get weird readings if you're not totally sure of how this circuit works. You can check for continuity in the wires by disconnecting them from the breaker/breakers and grounding one end and checking the other end with an ohm meter or continuity light.

Answer 4

The only option other than Ed's that I can offer is (assuming you left all junction boxes accessible as required) go back through every JB you can find in the house, particularly in the ceiling of the finished basement, that has even a remote chance of being on the circuit.
Also, could there be other outlets that you put on the circuit and forgot about? Shut down the breaker and see if any other outlets lose power. That might give you a clue of where the problem might be and what else might be on the circuit.

Answer 5

Did you test to an actual ground source or to the ground wire? I think it may be possible you have a hammer/staple damaged cable that the ground wire is broken and possibly penetrating the black wire, or the insulation on the black has a slight tear and some moisture has cause corrosion that is putting low current on the ground wire.

First thing I would do is remove all wires off the breakers and buses, and do a insulation resistance "megger" test on the wires.

If you don't have easy access to a megger I would be interested to see testing results with a low impedance solenoid "Wiggy" tester, they need more current than a high impedance electronic tester and a corroded connection may not pass enough current to pull the solenoid.

No wiggy either? I would also be interested in seeing a voltage test on the ground wire loosened from the panel measured to ground in the panel and would like to see what voltage readings are at the end of the cable with an open ground.

Answer 6

Probably irrelevant, but I'm recalling that about 30 years ago I was replacing a thermostat in our electric water heater and I decided that the wire between thermostat and heating element needed replacing.

So I took a 3-foot section of romex and stripped off the outer sheath. Inside I discovered that the black wire had been spliced in the factory, apparently to join two rolls of the wire. The wires were joined with a crimped gizmo, bare copper all around. Had this segment of wire been installed in a wall it would have shorted out to the ground wire, and the reason for the short would have been very much unclear. I'm guessing that factory protocol was that measurements would be made and the splice cut out after manufacture, but there was a screw-up.

It's highly unlikely, but there's a vague possibility that some mega version of this screw-up could have occurred somewhere in your cabling (maybe something intentional by a disgruntled employee), causing black and white to get swapped.