L-N short circuit MCB not tripping

brodders97 · Nov 18, 2022

Hi all,

I’m hoping you can shed some light on a problem I found. Please be gentle, I’m an industrial spark, not a domestic spark.

I’ve been asked to look at a family members electrics as they are renovating a room. A cowboy spark has been in and messed the job up. I was asked to come in and sort it out, my first plan of attack was an EICR on the property. Looking at his work (pictured as he left it) probably a waste of time.

Anyway, I was performing some tests on a non rcd protected board with everything unplugged etc. The circuits tested ok except the ring final, the circuit he has been working on. I found 11 Ohms between L-N from the board. Some sockets were reverse polarity.

When I tested between L-N in the room he had been on, the reading was much lower at 1 Ohm, so he’s done something here.

What I don’t understand is, if there was a short circuit between L-N, why has the MCB not tripped?

They are wanting a board protected by an RCD but obviously that can’t happen until this is sorted. What would your plan of attack be to sort it?

Thanks all

littlespark · Nov 18, 2022

For a reading that low, there must be something still plugged in.... But without first correcting the reversed N-E... you cant trust the readings.

Take off every socket, point, whatever you can find and test the cables so you know which cable runs between which points.
Likely youll find more joints underfloor, and possibly figure of eight rings....

Its going to take time.... if theyve had such a cowboy in beforehand

timhoward · Nov 18, 2022

A couple of principles first
An RCD won't care about a L-N fault, it will only care if any of the L-N current goes missing in action.

I'm already suspicious that a 10 ohm reading at the board turned into a 1 ohm reading in a room as there's no way there is enough cable in a house to explain that.

Go back to first principles.
i.e. Have you got a ring at all, and is there any leakage to earth?

End-to-end tests at board for L, N and CPC . I'm half expecting these to be open circuit and there to be at least one more polarity swap that's also a lousy / loose connection.

IR tests at board, L=N together to CPC at 250v for each end of the ring in turn.
Let's see if that is sane before doing anything else.

From what you said, I'd be surprised if those tests come back with normal numbers.

Final question, are you SURE there is nothing plugged in anywhere, it's really hard to be sure of this and normally any measurement L-N would have me thinking there's a connected load I missed somewhere. But in this case my blind guess is another polarity swap and poor connections.

brodders97 · Nov 18, 2022

timhoward said:
A couple of principles first
An RCD won't care about a L-N fault, it will only care if any of the L-N current goes missing in action.

I'm already suspicious that a 10 ohm reading at the board turned into a 1 ohm reading in a room as there's no way there is enough cable in a house to explain that.

Go back to first principles.
i.e. Have you got a ring at all, and is there any leakage to earth?

End-to-end tests at board for L, N and CPC . I'm half expecting these to be open circuit and there to be at least one more polarity swap that's also a lousy / loose connection.

IR tests at board, L=N together to CPC at 250v for each end of the ring in turn.
Let's see if that is sane before doing anything else.

From what you said, I'd be surprised if those tests come back with normal numbers.

Final question, are you SURE there is nothing plugged in anywhere, it's really hard to be sure of this and normally any measurement L-N would have me thinking there's a connected load I missed somewhere. But in this case my blind guess is another polarity swap and poor connections.

Thanks for your comments guys. It’s a bit of unexplored territory for me, but if I can help them out, I will. Once I come to my limit I’ll call in the proper misters.

On the RFC, end to end of both L and N were 0.5 Ohms. E-E was 1 Ohm, slightly high.

L + N to E tested ok on 250v (18M ohms).

I was confident nothing was plugged in but I could be wrong. It’s an old very big house with lots of additions.

The previous ‘sparky’ had tried to put an RCD board on, reverted back to MCB only when the RCD tripped instantly, im assuming this is due to the reversed polarity in the sockets.

Thanks again

timhoward · Nov 18, 2022

brodders97 said:
The previous ‘sparky’ had tried to put an RCD board on, reverted back to MCB only when the RCD tripped instantly, im assuming this is due to the reversed polarity in the sockets.

From the RCDs point of view, it's sending out live and getting back neutral. It would only be reversed from the point of view of the appliance plugged in. So I don't think that is the cause of the trip.
Getting N and E the wrong way around at a socket would on the other hand result in a working socket that would trip if an RCD was introduced.

So far the clues are the 11 ohms measurement and the fact an RCD tripped.

If the RCD was tripping with a clear L-N to E reading on that circuit, it's either a faulty appliance or interconnected circuits.
- Do continuity tests between the ring neutrals (disconnected), and the other neutrals, and the ring lives (disconnected) and each other live, just to ensure there are no interconnected circuits.
-Consider PAT testing the appliances, or plugging them in one at a time until it trips

To track down this 11 ohms (20 amps which I'm struggling to see how can persist unnoticed)
1 - still look for things plugged in. Kitchen appliances, maybe a Fused spur somewhere, towel rail, fan, boiler etc. TV amplifier in loft (usually on lighting circuit but who knows)
2 - Temporarily connect this theoretically unloaded circuit to a 6A breaker and see if it trips, as it's theoretically pulling 20 amps all the time, so it should. Or use a clamp meter at the board.

If totally stumped, deliberately break the ring in that room, and go back to the board and see if one leg is open circuit and the other is 11 ohms. Then test at a few sockets to get a sense of which half the 11 ohms is on. Then break again in the middle of that lot and repeat.
Absolute worst case is you end up with 3/4 of a ring proving fine, and either disconnecting the remaining section and derating the breakers (last choice) or rewiring that section.

Best of luck, this doesn't sound like a fun one.

brodders97 · Nov 18, 2022

timhoward said:
From the RCDs point of view, it's sending out live and getting back neutral. It would only be reversed from the point of view of the appliance plugged in. So I don't think that is the cause of the trip.
Getting N and E the wrong way around at a socket would on the other hand result in a working socket that would trip if an RCD was introduced.

So far the clues are the 11 ohms measurement and the fact an RCD tripped.

If the RCD was tripping with a clear L-N to E reading on that circuit, it's either a faulty appliance or interconnected circuits.
- Do continuity tests between the ring neutrals (disconnected), and the other neutrals, and the ring lives (disconnected) and each other live, just to ensure there are no interconnected circuits.
-Consider PAT testing the appliances, or plugging them in one at a time until it trips

To track down this 11 ohms (20 amps which I'm struggling to see how can persist unnoticed)
1 - still look for things plugged in. Kitchen appliances, maybe a Fused spur somewhere, towel rail, fan, boiler etc. TV amplifier in loft (usually on lighting circuit but who knows)
2 - Temporarily connect this theoretically unloaded circuit to a 6A breaker and see if it trips, as it's theoretically pulling 20 amps all the time, so it should. Or use a clamp meter at the board.

If totally stumped, deliberately break the ring in that room, and go back to the board and see if one leg is open circuit and the other is 11 ohms. Then test at a few sockets to get a sense of which half the 11 ohms is on. Then break again in the middle of that lot and repeat.
Absolute worst case is you end up with 3/4 of a ring proving fine, and either disconnecting the remaining section and derating the breakers (last choice) or rewiring that section.

Best of luck, this doesn't sound like a fun one.

Thanks mate much appreciated. One of them where you regret saying you’ll get involved. Although I’m doing this favour the right side of Christmas!

default · Nov 18, 2022

brodders97 said:
Thanks for your comments guys. It’s a bit of unexplored territory for me, but if I can help them out, I will. Once I come to my limit I’ll call in the proper misters.

On the RFC, end to end of both L and N were 0.5 Ohms. E-E was 1 Ohm, slightly high.

L + N to E tested ok on 250v (18M ohms).

I was confident nothing was plugged in but I could be wrong. It’s an old very big house with lots of additions.

The previous ‘sparky’ had tried to put an RCD board on, reverted back to MCB only when the RCD tripped instantly, im assuming this is due to the reversed polarity in the sockets.

Thanks again

The ring continuity readings are OK. Cpcs should be around 0.8 but loose connections can make it higher or if the CPC is only 1mm. Maybe getting those readings on the insulation test because he's cut into another ring or spured off to pick up heating controls. As someone has already said, disconnet the sockets. Find your first legs from the DB and painstakingly go from there. You maybe lucky and its at the first point but sods law says otherwise.

default · Nov 18, 2022

timhoward said:
From the RCDs point of view, it's sending out live and getting back neutral. It would only be reversed from the point of view of the appliance plugged in. So I don't think that is the cause of the trip.
Getting N and E the wrong way around at a socket would on the other hand result in a working socket that would trip if an RCD was introduced.

So far the clues are the 11 ohms measurement and the fact an RCD tripped.

If the RCD was tripping with a clear L-N to E reading on that circuit, it's either a faulty appliance or interconnected circuits.
- Do continuity tests between the ring neutrals (disconnected), and the other neutrals, and the ring lives (disconnected) and each other live, just to ensure there are no interconnected circuits.
-Consider PAT testing the appliances, or plugging them in one at a time until it trips

To track down this 11 ohms (20 amps which I'm struggling to see how can persist unnoticed)
1 - still look for things plugged in. Kitchen appliances, maybe a Fused spur somewhere, towel rail, fan, boiler etc. TV amplifier in loft (usually on lighting circuit but who knows)
2 - Temporarily connect this theoretically unloaded circuit to a 6A breaker and see if it trips, as it's theoretically pulling 20 amps all the time, so it should. Or use a clamp meter at the board.

If totally stumped, deliberately break the ring in that room, and go back to the board and see if one leg is open circuit and the other is 11 ohms. Then test at a few sockets to get a sense of which half the 11 ohms is on. Then break again in the middle of that lot and repeat.
Absolute worst case is you end up with 3/4 of a ring proving fine, and either disconnecting the remaining section and derating the breakers (last choice) or rewiring that section.

Best of luck, this doesn't sound like a fun one.

I've been to a fault before with intermittent tripping of the RCD. Everything was on and working fine and then for no reason it would trip. The house was being decorated so while i was thinking about the issue I watched a painter unplug his radio, which was working, to move to another room and the RCD tripped. Checked the socket and N&E were in the wrong terminals. So even though that shouldn't be able to happen it can. Leave no stone unturned.