New cu first or plaster bare wall?

[JohnW2]

There's a rough as F and potentially dangerous method to check for borrowed neutrals, more than borrowed lines. Shan't be sharing it publicly. Not a guaranteed method but can give an indication.

Yes, I can think of some

Probably the quickest way to 'check', empirically, is to disconnect the neutral of one lighting circuit at the CU and to make sure that all the lights on the other lighting circuit(s) still work. If they don't, then one can investigate/test more formally.

Kind Regards, John

 

[OwainDIYer]

When you say that you test N-N and L-L between circuits, do you only do this for lighting circuits, or for all circuits? If the latter, it could be pretty tedious and time-consuming. For example, with an 8-way board, there would be 28 N-N pairs and 28 L-L pairs to test - a total of 56 tests!

But you don't have to test every pair. You only have to test each wire against all the others.

So assuming you have intact fuses or circuit breakers closed, with the main switch open, connect one meter probe to the L busbar. Open MCB/remove fuse (or remove circuit 1 L from its fuse) and connect probe to load side of fuseholder (or circuit wire). Test then restore connection. Repeat with circuits 2-8.

Then repeat for the Ns removing one wire at a time from the N busbar and test.

That's eight tests for the L and eight tests for the N.

If you get a 'share' between one circuit and busbar you should also get a 'share' between the corresponding circuit and busbar, so those sixteen tests will identify one L-L or a N-N bridge. Only if you get more than two 'shares' do you have to probe the circuit pairs individually to find which circuit shares with which (and you only have to test the affected circuits, not all of them).

To identify an L on one circuit returning as a N on another circuit (typical stair lights borrowed neutral) remove the neutrals from the N busbar and connect into the corresponding circuit fuses. Then repeat the tests, testing between the busbar and the L+N pair. (Ensure all lights are switched on, and that tester will show continuity through the lamps.)

If you don't have an intact consumer unit or if the Ns aren't long enough to stretch into the fuses, then a test lead of a chain of crocodile clips makes quick work of commoning the circuits and disconnecting the one under test. (This would not be GS38 compliant and should not be used for live testing or insulation testing).

 

[JohnW2]

But you don't have to test every pair. You only have to test each wire against all the others.

Yes, you're right, at least in the first instance, one only has to test each wire against all of the others (joined together) - and, as you say, that 'joining together' can be via the L-busbar or the N terminal block. As you say, that would require just 8 tests for L and 8 for N.

In the case of a single cross-connection, two of those tests would presumably show a connection, so one could reasonably conclude that those were the two connected circuits. If more than two of the tests showed cross-connections, then more, pair-wise, tests would have to be done.

Kind Regards, John

 

[PrenticeBoyofDerry]

Thanks. Yes, I think that the only way to do it is to literally test for connections/continuity between the neutrals (and lines) of each circuit, but I was wondering whether that is usually done routinely in the context of an EICR.

I would do this if the customer was asking for a board change, so I can then inform them, if remedial work would is required on the circuits concerned.

You've said that you would do it, but do you think that electricians in general would do it routinely as part of an EICR?
When you say that you test N-N and L-L between circuits, do you only do this for lighting circuits, or for all circuits? If the latter, it could be pretty tedious and time-consuming. For example, with an 8-way board, there would be 28 N-N pairs and 28 L-L pairs to test - a total of 56 tests!

No not as routine part, but I would generally test the lighting conductors and remark on my findings.

 

[JohnW2]

I would do this if the customer was asking for a board change, so I can then inform them, if remedial work would is required on the circuits concerned.

That sounds fair enough. From what you go on to say, do I take it that you would not do this 'routinely' if you were just asked to 'do an EICR' (without any stated plans for a board change)?

Kind Regards, John

 

[PrenticeBoyofDerry]

That sounds fair enough. From what you go on to say, do I take it that you would not do this 'routinely' if you were just asked to 'do an EICR' (without any stated plans for a board change)?

No, I would normally just do the standard test!
If I found something odd, and potentially there was follow up work. I may investigate a little further, after discussion and if requested to do so by the client.