Strange fault

[RF Lighting]

Got called to a house yesterday. Their RCD has been tripping a lot. Customer said it tripped 5 or 6 times on Sunday, with no aparrent correlation to time or activity.

The setup is a new dual split CU installed about a year ago. Aparrently it has been tripping ever since it was fitted, but got progressively worse.

The CU has been moved from one side of the kitchen to the other, and I can't yet find the joint.

I got my megger on the job and got various readings ranging from >500MΩ down to 9MΩ on the individual circuits, and about 3MΩ for all the circuits together (L&N-E) If I disconnect the 9MΩ circuit, the overall reading jumps to about 12MΩ

I have left that circuit disconnected, and the customer rang me today to say the CU hasn't tripped at all for 24hrs now, so I reckon that's the right circuit.

Here's the strange thing. The circuit just feeds 7 lights wired on central joint box with no earth! Did some more testing and got a reading of 0.012MΩ L-E when the top of the stairs light is on, but it turns out this has it's neutral borrowed from another circuit on the same RCD.

I've left the rest of the circuit live with the top of stairs 2 way disconnected for now to see how it goes, but I don't see how that would cause an RCD trip, and if it isn't that what could cause an RCD trip on an unearthed circuit? There is one outside light, but I've stripped it all down and confirmed >1000MΩ on IR.

I still need to find the hidden joint, but it could be anywhere.

Just to add, I ramp tested the RCD and it tripped at 22.5mA both with and without the wiring connected.

Sorry for the long post, any thoughts much appreciated.

 

[Adam_151]

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

Fault between the live and other earthed metalwork such as copper pipes, metallic parts on other circuits

I assume when you were testing IR of the other circuits , you left all the cpcs in the earth bar and just took neutral and phase out of the board?

 

[dingodeano]

The 9M circuit has no earth anywhere at all along it's entirety? As in wired in singles?
If there is some t&e I am wondering if someone borrowed a neutral they may have piggybacked a cpc somewhere also.

 

[RF Lighting]

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

Fault between the live and other earthed metalwork such as copper pipes, metallic parts on other circuits

I assume when you were testing IR of the other circuits , you left all the cpcs in the earth bar and just took neutral and phase out of the board?

Yeah left all the CPCs in the earth bar and tested each L&N to earth

I did wonder about a wire maybe melted to a heating pipe etc. It's going to be one of those jobs I can just tell. Of course it's never tripped while I've been there.

The 9M circuit has no earth anywhere at all along it's entirety? As in wired in singles?
If there is some t&e I am wondering if someone borrowed a neutral they may have piggybacked a cpc somewhere also.

The feed from the CU is T&E and I guess this is jointed out to the old wiring at the old CU position which I haven't yet found. The old wiring is twin with no CPC and DI singles.

 

[dingodeano]

I always feel a bit grubby having to reconnect on to crappy workmanship. Unless you can find those join(s) you are going to be guessing.

 

[securespark]

That cpc MUST short to a line conductor somewhere otherwise you'd get a decent IR reading, surely?

Or have I missed a trick?

I've had faults on two way strappers causing my Alphatek to interpret faults as reverse polarity before now....

Have you located that central JB?

All I can say is try & work your way from A - B belling each leg out...

 

[Adam_151]

I've had faults on two way strappers causing my Alphatek to interpret faults as reverse polarity before now....

That was one of the strappers shorted to earth by a picture hook, TT system and knackered VO ELCB wasn't it?

 

[bernardgreen]

Have you considered that the leakage that is causing the trip is not from Live to the system "earth" but is leakage from a damaged Live (or switched Live) direct to ground via a damp wall or similar path not involving the CPC at all.

 

[securespark]

I've had faults on two way strappers causing my Alphatek to interpret faults as reverse polarity before now....

That was one of the strappers shorted to earth by a picture hook, TT system and knackered VO ELCB wasn't it?

We've had a few now....

But that rings bells.

 

[ban-all-sheds]

Was the bell on 2-way pushbuttons then?

 

[RF Lighting]

It's just tripped again

Back with the megger tomorrow then

I've been thinking about buying a milli-amp meter to help investigate faults like this. Does any one own one, and if so are they any good?

 

[ricicle]

Rob - remember that just clamping the meter around the cpc or earthing conductor is not going to be accurate due to network currents. You must clamp around both L + N conductors of the install or circuit and the meter will read the difference. I'm not sure you will get a milli-amp clamp around 2 x 25mm² tails.
Or am I telling you how to suck eggs ?

 

[RF Lighting]

No not at all. It's something I've heard about but I've never actually used one. I did wonder if you could clamp L and then N and calculate the difference, as was mentioned earlier, the fault might not necessarily be returning via the installation CPCs / earth.

 

[bernardgreen]

The majority of clamp meters are not that accurate / sensitive when looking for a few milli-amps difference between two conductors in the clamp. One reason being the position of the conductor in the clamp ( touching it or right in the middle ) affects the measurement.

What is needed is small core current transformer and a multi-meter to measure the current from its secondary....

The ideal current transformer is the one in many RCDs as it is designed to accurately compare the currents and provide an output proportional to the difference. ( however not all are suitable ).

The professional way would be to use a ring core current tranformer such as

http://www.dk-moriarty.ltd.uk/principle_list/metering.html

with Live and Neutral passing through it opposite directions. Then measure the current in the secondary which will be directly proportional to the difference between Live and Neutral currents. To calibrate the volts per milli-amp could be calculated from transformer data or by passing a series of known currents and measuring the voltages.

 

[echoes]

Have you considered that the leakage that is causing the trip is not from Live to the system "earth" but is leakage from a damaged Live (or switched Live) direct to ground via a damp wall or similar path not involving the CPC at all.

...Or from a similar fault on either the neutral of the circuit without the CPC, or the borrowed neutral?