strange one

[tbs1983]

Hi there, plumber phoned me during a boiler fit in a house saying RCD had tripped (nothing to do with boiler circuit), but I suspect he's disturbed something.
It seems to hold for 10 mins or so then trips.
I have reached a point where ive found the approx mid point of the ring, disconnected both legs and with both legs disconnected at db, I'm getting clear IR.
but here's the strange bit- if I iR across the neutral of one leg and the earth of the other leg at the db, i get 0.6 Mega ohms. same reading if I do it on the N & E of the opposite legs.
Any ideas???? I'm baffled!

 

[Dave 85]

"I also think the 0.6 megohms is a red herring as far as the trip is concerned, although it might lead you to the actual fault by way of being across an unexpected combination of conductors. Something similar cropped up on a recent thread - to trip a 30mA RCD it would have to be a thousand times lower or thereabouts, as it's N-E."

I was waiting for someone to point this out

 

[Dipzy]

Had the snarl-up a couple of months ago, compressed cables in a 25mm back box = RCD tripping. Solution as kitchen was due to be tiled used longer screws on S/O

 

[Rockingit]

"I also think the 0.6 megohms is a red herring as far as the trip is concerned, although it might lead you to the actual fault by way of being across an unexpected combination of conductors. Something similar cropped up on a recent thread - to trip a 30mA RCD it would have to be a thousand times lower or thereabouts, as it's N-E."

I was waiting for someone to point this out

0.6 Meg = 600 Kilo = 600,000
230/600000 = 0.000383A
= 0.38mA

 

[Rockingit]

Had the snarl-up a couple of months ago, compressed cables in a 25mm back box = RCD tripping. Solution as kitchen was due to be tiled used longer screws on S/O

No, by the sounds of it the solution was that you bodged an already bodge job and didn't repair a damaged conductor from a screw

 

[mark ju]

Hi there, plumber phoned me during a boiler fit in a house saying RCD had tripped (nothing to do with boiler circuit), but I suspect he's disturbed something.
It seems to hold for 10 mins or so then trips.
I have reached a point where ive found the approx mid point of the ring, disconnected both legs and with both legs disconnected at db, I'm getting clear IR.
but here's the strange bit- if I iR across the neutral of one leg and the earth of the other leg at the db, i get 0.6 Mega ohms. same reading if I do it on the N & E of the opposite legs.
Any ideas???? I'm baffled!

A mate had something similar a few years back. The ring circuit had been pinned inside a stud wall somehow or another. Point is the offending screw or nail pierced both t&e's effectively shorting cpc on one side to neutral of the other.
apparently proved a bugger to find.
might not be this but you never knows......

 

[tbs1983]

thank you for all the advice, much appreciated, determined to find the fault now! will post my findings later. Thanks again

 

[joelypops]

Last time i got something like this it was because the walls were made with tin gypliner and foil backed plasterboard. Screwed through cable on one side of the room and another one t'other. Because of metal studding i was getting around 0.1M ohms between the two circuits, and also between N and the back boxes where they were touching the foil backed board.. definately took some finding and we are still arguing with the boarding company about costs etc...

 

[Rampantchilli]

I've had a tripping issue on a ring previously, only occasionally, some times after 10 minutes sometimes after 2 days. Can't remember the exact readings, but low IR across conductors, however tracing the fault took a full day, as when all sockets were disconnected all cables individually tested clear on IR, but when reconnected, fault reappeared, even with sockets pulled forward. Turned out to be a nail in the wall to hang a picture on, but I didn't investigate further, just replaced cable via a different route. I'd suggest that the nail was in the live but nothing else, hence the IR being clear, but with all cables connected and flying leads to back boxes connected the circuit was made.
And I've had this similar fault recently again
i would check carefully where the plumber has been working, he may of only disturbed a previously undid entities fault.

 

[Lucien Nunes]

0.6 Meg = 600 Kilo = 600,000
230/600000 = 0.000383A
= 0.38mA

Indeed, and that is from line to earth with 230V across the fault to drive current through it. With the few volts usually found between neutral and earth the current would be more like a few microamps. It really does take a low IR to trip an RCD but it is often another symptom of the real cause.

 

[tbs1983]

Fault Rectified !!! This evening I disconnected each leg of the ring, plus the spurs from jb's. IR'd every individual leg, every single one clear. Reconnected one leg at db and 1st socket, carried out ramp test and got 26mA. connected the next socket, ramp tested and got 26mA. so on and so on until after connecting another leg the ramp tester would not trip, gave >51mA as a reading. left this leg out of the ring, reconnected everything else, ramp tested 26mA and plugged in the loads, happy days. Rewired faulty leg and bobs your uncle problem solved.

What I've learnt from this fault -
1. I need to buy a ramp tester as I would never have found the fault otherwise.
2. I hate faults !
3. Insulation resistance alone doesnt always identify a faulty cable (although this one is still baffling me why the leg IR'd clear yet it was causing the fault)
.

Thank you for all the help and advice on this one.

 

[Dipzy]

No, by the sounds of it the solution was that you bodged an already bodge job and didn't repair a damaged conductor from a screw

Yes it became apparent it was a bodged job and not done by me.

There were 4 x 2.5mm T & E in a 25mm backbox for a double s/o; 2 for the rfc, 1 for the ufh heating and 1 for the boiler spur.
Quite why the ufh and boiler spur cables had been routed through this backbox is a mystery.

The problem was caused by loose conductors which hadn't been terminated properly, no damage to any conductors from screws.
Re-terminated them and used 50mm accessory screws to fasten the faceplate. Once tiled, the tiles would allow a bit of room to ease the compression of the cables.

I was only there to second fix an external light but got asked to look at this problem as you do.....

 

[Lucien Nunes]

I'm glad you found the fault! I was thinking, though, about having found it with the ramp test. I'm not knocking your work, just musing on what the actual fault was and how it might have been found by testing when dead rather than energised.

The ramp test is responding to the symptom, i.e. that the RCD misbehaves, rather than the fault itself, i.e. that the cable is damaged. If you'll allow me an analogy, it's like finding a water leak by changing pipes length by length until your water bill drops. But the leak itself should be visible, and so might the cable fault have been detectable with the insulation and continuity tests.

Did you at any point test from all conductors of the faulty RFC (disconnected at the CU) to the rest of the installation globally, including earth? Or from all cores of the isolated faulted leg to the rest of the RFC? If you did not have a definite clear reading of those, it is possible that there is still a fault between the disconnected faulty leg of the RFC and another cable, although it might no longer show up from the other circuit because the RFC side of the fault is disconnected.

Put another way, a cable that passed (within reason) a IR test between all its conductors, but which causes random RCD tripping and also causes an RCD test to fail, must be significantly faulted to something else. As yet I don't think there is any proof that the fault is only to earth (although that is the most likely) but there is some evidence that it is not to earth (your cross-legged IR results and the fact it causes the ramp test to fail without an N-E short).

Perhaps you had other results that proved it one way or another, would be interested to hear.

 

[chrish_1977]

I would agree, if you had a loose connection then your end to end readings should have shown this up, then just break the circuit down to find the fault, rather than keep energising a faulty circuit to find the fault

 

[Sharpend]

This requires the fine art of true testing, to trace and rectify the fault, an art I'm afraid many sparks struggle with IMO.

 

[telectrix]

This requires the fine art of true testing, to trace and rectify the fault, an art I'm afraid many sparks struggle with IMO.

1. test at 250V. 2. test at 500V. 3. shove 1000V up it. 4. get the hammer out. nothing hard there. :behead: