RCD tripping. TT. where could be the fault? (FIRE ALARM SENSOR)

[mrmisior]

EDIT:
faulty fire alarm sensor spurred within ceiling was causing the RCD to trip.
not mentioned in paperwork.
all other sensors are on a separate ring, so this one is a mystery.
electronic device which caused L-E fault only when powered up so this would have probably came up on a megger (if i was to call an electrician and i was close to doing so), but did not show the fault with a 9V multimeter.
essentially, i dismantled the ring and tested all tails and all sockets, to find a perfectly healthy ring with no faults and still tripping one RCD...
i hope that my experience will help someone.
:EDIT

TT system with earthing electrode.
Split load board with two 30mA RCDs.
One RCD trips.
Identified the tripping circuit to 'downstairs sockets'.
All other circuits protected by this RCD function properly (cooker, lights, fire alarm, central heating, kitchen sockets).
All other circuits protected by another RCD also function properly and the other RCD does not trip.
MCB 'downstairs sockets' does not trip. Only the RCD.
When this MCB 'downstairs sockets' if OFF, RCD stays ON.
When this MCB 'downstairs sockets' is ON, RCD trips instantly (<0.5s).

The 'downstairs sockets' circuit contains 7 double sockets.
All devices disconnected - all plugs removed from sockets. All sockets in OFF position.
Still trips as above.
Opened one socket and measured L-N open, L-E open, N-E ~40 Ohm.
There was no work done around the house, no leaks.

What can I test? What would it be?

 

[jj4091]

Mice under the floor chewing the cables

 

[flameport]

What can I test?

Disconnect the wires for that circuit at the consumer unit, which will be 3 of them for a radial, 6 for a ring.
At a socket about halfway along the circuit, disconnect the wires there and measure N-E. One set of wires will have the 40 ohm fault, the others will not.
Find another socket about halfway between that point and the consumer unit in the direction of the cable that has the fault, split there and see which of those halves has the fault.
Continue to divide until you either find the fault at a socket, or the single piece of cable between 2 sockets which has the fault.
Repair or replace as necessary.

 

[EFLImpudence]

Surely an N to E fault would still cause the RCD to trip when the MCB is off.

 

[mrmisior]

Surely an N to E fault would still cause the RCD to trip when the MCB is off.

on some other forums found that it may not always trip on NE fault.
although, to my experience, working on circuits had to always be careful no to touch the neutral cos i would knock the RCD along with lights.
thinking (hard).
perhaps there is something spurred of the ring?
will find out soon.

 

[flameport]

Depends on the exact routes the current flows, with a TT system that may not be what's expected.

 

[JohnW2]

Surely an N to E fault would still cause the RCD to trip when the MCB is off.

That presumably depends on what, if any, loads are connected to other circuits (assuming, as stated, no loads on the faulty circuit)?

Kind Regards, John

 

[EFLImpudence]

Well, yes, the RCD won't trip with no current on all circuits flowing through it , but that isn't usually the case.

 

[plugwash]

If there are no devices connected and the live is non-faulty then turning on the MCB shouldn't have an impact.

I see two main possibilities here.

1. There is a fault from live to earth in addition to the fault from neutral to earth, but whatever measuring device the OP was using was unable to detect it. Faults can be quite non-linear, a fault that measures off-scale on a multimeter can easilly pass significant current at 240V. That is *why* electricans use 500V insulation testers.
2. There is an undiscovered device on the ring, which has a fault to earth that is behind a single pole switch. When un-powered this shows up as a high-resistance neutral to earth fault, but when the device is powered on it starts passing considerablly more current from live to earth.

In any case I think the "divide and conquer" approach flameport suggested is the first port of call. Even if the neutral fault is not the cause of the tripping it still needs to be found at recitified.

 

[JohnW2]

Well, yes, the RCD won't trip with no current on all circuits flowing through it , but that isn't usually the case.

Probably not, but we do hear of cases in which a N-E fault results in 'intermittent' RCD tripping, presumably implying that loads on other circuits are not always great enough to result in the trip.

Also, don't forget that we are talking about a fault which, if it really is ~40Ω, is far from being 'of negligible impedance', so that only a relatively small proportion of the neutral current from other circuits would probably flow through it.

Finally, it's not necessarily only current in circuits protected by the same RCD that matters, since some of the neutral current from circuits protected by a different RCD may go 'backwards' through the RCD of the faulty circuit, and thence through the fault to earth (creating a current imbalance' in the RCD).

Kind Regards, John

 

[EFLImpudence]

Yes, John (and Flameport) - but the fact remains that if someone just turns off the MCB and then (as the OP has himself sort of said) the N and E touch then nearly always the RCD will trip.

So, as I said:

Surely an N to E fault would still cause the RCD to trip when the MCB is off.

I should have said could instead of would.

 

[JohnW2]

Yes, John (and Flameport) - but the fact remains that if someone just turns off the MCB and then (as the OP has himself sort of said) the N and E touch then nearly always the RCD will trip.

Yes, probably 'nearly always' but, as I said, we hear of (permanent) N-E faults which result in intermittent (rather than 'permanent') RCD tripping (when affected circuit is 'SP isolated' by an MCB) ... but I'm not really disagreeing with you.

Kind Regards, John

 

[bernardgreen]

A Neutral to Earth fault will trip the RCD if there is a load taking current on any the circuits protected by the RCD

The CPC ( green / yellow ) can be earthed

(1) as shown where Neutral and Earth are combined on a single conductor in the district supply network, The Neutral is Grounded at the substation

or

(2) by any direct connection to the ground such as a TT Ground rod or a Earth bonded metallic service pipe ( gas water )

 

[JohnW2]

A Neutral to Earth fault will trip the RCD if there is a load taking current on any the circuits protected by the RCD

Indeed but, as I wrote, I think also even if there is load taking current on a a circuit protected by a different RCD (since some of the neutral current will then flow 'backwards' through the affected RCD and the N-E fault on one of its circuits.

Kind Regards, John

 

[ericmark]

Opened one socket and measured L-N open, L-E open, N-E ~40 Ohm.

With what? I have been in a hurry and measured with a multi-meter and got open, but measured with an insulation tester showed nearly short circuit.

I would first select a socket to test, and clip on meter to show the 40Ω, then remove the 3.5 mm screws on other sockets checking meter as screws removed, then pull each socket away from the wall a little, likely a spider or the like, and once socket where it is hiding is disturbed the fault will go.

It could be rodents and you may need to test neutrals between each socket, but try removing first.