Strange fault

[RF Lighting]

I'm thinking I might transfer one circuit at a time onto the other RCD until that RCD starts to trip and then at least I know for sure which circuit I'm chasing.

It's gone 2 days without tripping and then off it goes again. I hate faults like this.

 

[Peet]

, and if it isn't that what could cause an RCD trip on an unearthed circuit?

It's the difference in the L & N currents that causes it to trip. That difference may or may not have gone to earth.

You say 'but got progressively worse.' in your OP. That makes me think of something marginal or breaking down or rubbing/moving etc etc. Might sound silly, but have you thought about jumping on floors? Causing a bit of vibration? Or tapping things? Sometimes tapping kit shows up stuff that's otherwise hard to find.

Have you got a cable tracer that you can trace out the cable route of the one you suspect? May help you find the hidden joint too.

 

[bernardgreen]

Is there space to fit temporarily an RCD and suitable MCB adjacent to the CU ( a garage CU springs to mind ) and use that to supply the suspect circuit with a temporary feed from the CU but before the CU's RCD.

That should give a clear indication if the fault is on that circuit.

 

[RF Lighting]

I went back today. I transferred two circuits from the RCD which is tripping. One to the RCD which has never tripped, and one to a stand alone RCD I have installed temporarily.

Hopefully this should start to narrow things down quite quickly.

 

[EFLImpudence]

I'm not sure why you have not concentrated on the two way light as you read 0.012MΩ when it was switched on - suggesting it is the switched line (or a strapper if only when on in one way) or the borrowed neutral - or the lamp, of course.

0.012MΩ is 20mA @ 240 and the RCD ramped at 22.5mA

As the RCD ramp reading was with and without the wiring it would suggest no other faults although the IRs aren't brilliant.

 

[RF Lighting]

The two way was a red herring. The wiring has been completely disconnected and still it trips. The poor IR came from reading though a CFL to the borrowed neutral which is in turn tied to earth on the DNOs network.

I had the line and neutral for this circuit disconnected at the board so when I got a poor reading line > earth I didn't consider a borrowed neutral, and assumed a genuine earth fault.

 

[RF Lighting]

Just a bit of an update.

I went back last Wednesday morning to do some more investigation and found the IR for the suspect lighting circuit had risen from 9MΩ on my first visit to about 300MΩ on Wednesday which suggests to me something had dried out.

Anyway, seems I didn't quite test the outside light properly first time around. The light has a SL&N from an unknown junction box, but has had a photocell added to it afterwards.

I stripped the photocell and IR'ed the light from there, but who ever installed this has cut the CPC for the photocell at the light, hence no bad reading.

Also as this is supplied from a class 2 circuit the original installers have borrowed an earth from somewhere to supply the light.

Here it is:

Anyway I bunged it all up with mastic compound as a temporary measure and so far, touch wood it hasn't tripped since, despite 2 days of heavy rain on Fri / Sat.

This is going to be rewired properly as part of some other scheduled work in a few weeks. Might even get the pyro out

 

[JohnW2]

Just noticed this - sorry to jump in a few days late ...

It's the difference in the L & N currents that causes it to trip. That difference may or may not have gone to earth.

If 'the difference' had not gone to earth, where else do you feel it might have gone?

That difference current would have to have found its way back to either L or N on the supply side of the RCD and, apart from going via 'earth' to the N, one would surely have to postulate some extraordinarily unlikely scenario for it to have found any other path?

Kind Regards, John.

 

[securespark]

'Kin hate cr@ppy outside lights like that!!!

 

[ban-all-sheds]

And it doesn't help when such a disgusting job is done of installing them.

 

[bernardgreen]

one would surely have to postulate some extraordinarily unlikely scenario for it to have found any other path?

A borrowed neutral, borrowed from the flat upstairs or the house on the other side of the party wall ?.

 

[RF Lighting]

Just noticed this - sorry to jump in a few days late ...

It's the difference in the L & N currents that causes it to trip. That difference may or may not have gone to earth.

If 'the difference' had not gone to earth, where else do you feel it might have gone?

That difference current would have to have found its way back to either L or N on the supply side of the RCD and, apart from going via 'earth' to the N, one would surely have to postulate some extraordinarily unlikely scenario for it to have found any other path?

Kind Regards, John.

I think he was referring to current flowing to earth via the installations CPC. It is possible for the current to flow to earth via damp stonework, metallic pipework etc, or it could flow to neutral via a borrowed neutral on a non-RCD circuit, all of which would cause an RCD to operate without putting a current into the installations earthing system.

 

[RF Lighting]

Turns out the outside light was also a red herring as the RCD has again tripped a few times since I gummed it up.

I think I've cracked it this time.

Found this in the loft connected to the circuit which showed low IR

Mmmm yummy ELV transformer

 

[JohnW2]

one would surely have to postulate some extraordinarily unlikely scenario for it to have found any other path?

A borrowed neutral, borrowed from the flat upstairs or the house on the other side of the party wall ?.

Sure, but this RCD had been in place for about a year, and was just tripping occasionally. That is not the really the picture of a wiring fault such as a borrowed neutral.

Kind Regards, John

 

[JohnW2]

If 'the difference' had not gone to earth, where else do you feel it might have gone?

I think he was referring to current flowing to earth via the installations CPC. It is possible for the current to flow to earth via damp stonework, metallic pipework etc, ...

Fair enough - if that's what he meant, then I agree.

... or it could flow to neutral via a borrowed neutral on a non-RCD circuit,

True - but, as I just wrote to Bernard, one wouldn't normally expect a wiring fault such as a borrowed neutral to result in just occasional trips of an RCD that had been in place for about a year.

Kind Regards, John.