Short in socket - why/how?

[TheJack]

Hi. I'm in the process of getting my qualifications and am taking on some small jobs for family to try and get some experience in.

Today I was looking at a socket on a final ring circuit. The house was built in the 80s and the socket is almost certainly original to then.

There were 3 cables running through the wall connected to the socket - with all 3 conductors twisted together (ie all 3 lines twisted together and connected to the socket as one).

I presumed 2 would be part of the ring, and wanting to know where the other was spurring to I untwisted everything and with a bit of tugging worked out that one was connected to a socket in an adjacent room.

I then attempted to rewire the socket, but rather than retwisting everything I put the conductors back in individually (it was a bit of a faff as it was tight, and the old cables quite inflexible) but I got there in the end.

However when I re-energised the circuit and plugged the appliances back in there was a crack, a small bang, and a weird (perhaps acidy) smell, which I understand to be a short circuit.

Confusingly nothing tripped.

I'm baffled though as my understanding is that a short would be due to the live conductors touching, however I definitely did not incorrectly rewire it - ie lines and neutrals separate.

I'd be so grateful if someone could shed some light please?

 

[westward10]

Sounds like you have damaged the insulation of a live conductor to the back box. Plugging something into it disturbed it enough to create a live to earth fault.

 

[TheJack]

Sounds like you have damaged the insulation of a live conductor to the back box. Plugging something into it disturbed it enough to create a live to earth fault.

Thanks for replying.

As in the insulation has split (for example) and so part of the live conductor is making contact with the metal back box?

If that's the case is it going to need the cable replaced? Please say no as getting cables in and out of that wall is going to be an absolute nightmare!!!

 

[westward10]

If this is the case and the insulation is damages and most likely the copper conductor it will need replacing. Assuming there is enough length you just need to replace it within the back box using a Wago for example. You need to establish the cause first.

 

[DPG]

Do you have an MFT?

 

[TheJack]

Do you have an MFT?

Hi mate. Yes I do.

 

[TheJack]

So I took it apart again, had a good inspection and couldn't see any damage to the conductors.

I've now put it carefully back together and a plug socket tester is telling me it's all good so hurrah.

I didn't love doing that though as obviously it was a live test and I nearly pooed myself after the firework display caused by yesterday's short.

So what I'm trying to work out is whether there's a dead test that I could have done to check for a fault before re-energising?

The ring final continuity tests I've been taught (End to End "Little R values" / Figure 8 line and neutral / Figure 8 line and CPC) are all done with one of the faces removed so clips can be placed on the conductors. However in this old house I am reluctant to remove another socket face as that's how I got in this pickle to begin.

 

[snowhead]

It may only be that socket that's got short conductors, possibly damaged before 2nd fix during plastering.

If you had fireworks, there was and probably still is a fault.
Maybe live conductor insulation cut when it was knifed to strip the outer and it touched the earth conductor after you'd reconnected.

 

[Julie.]

Since you have a MFT, you should always do an IR test at the very least, although measuring the r1+rn etc should be done as well.

You should have a plug with the MFT, to go directly into a socket with leads to plug into the MFT.

If you don't - just wire a plug accordingly, testing with 240v then 500v should identify ir faults that don't show themselves under normal circumstances.

You could have a small leakage current - just being energised won't show this - your MFT will.

 

[davesparks]

The ring final continuity tests I've been taught (End to End "Little R values" / Figure 8 line and neutral / Figure 8 line and CPC) are all done with one of the faces removed so clips can be placed on the conductors.

Normally this is done at the distribution board, not at a socket.


But whilst that test might show up the fault it will not necessarily be obvious in the results. An IR test is the way to identify such faults.

 

[davesparks]

However when I re-energised the circuit and plugged the appliances back in there was a crack, a small bang, and a weird (perhaps acidy) smell, which I understand to be a short circuit.

No, that is not a short circuit, that is two noises and a smell, these could have been caused by a short circuit or an earth fault.

A short circuit is a connection of negligible resistance between live conductors, an earth fault is a connection of negligible resistance between one or more love conductors and earth.

 

[TheJack]

Morning. So I've now carried out some IR testing at the board:

Neutral to earth = 0.43
Line to earth = 0.43
Line to neutral = 0.00

 

[westward10]

0.43 could be caused by an RCD socket or spur. It could also be a filtered extension lead (have you tested at 250v?).
Testing an existing circuit you should have your neutral conductors disconnected and linked to the line conductors, keep the cpcs connected. Do not test line to neutral unless everything is disconnected.

 

[TheJack]

0.43 could be caused by an RCD socket or spur. It could also be a filtered extension lead (have you tested at 250v?).
Testing an existing circuit you should have your neutral conductors disconnected and linked to the line conductors, keep the cpcs connected. Do not test line to neutral unless everything is disconnected.

Thanks Westward.

So how I've been taught to perform an IR test:

• CB safely isolated
• MFT in Isolation mode at 500v
• Zero by touching probes together
• Test N/E, L/E, L/N
• Probes at at top of breaker (for line) or terminals at top of board (for N & E)

 

[westward10]

That is fine for a new initial verification but this is an existing in use circuit. Your zero across L and N was probably something still connected across the circuit. Test as suggested in #13.

 

[TheJack]

0.43 could be caused by an RCD socket or spur. It could also be a filtered extension lead (have you tested at 250v?).
Testing an existing circuit you should have your neutral conductors disconnected and linked to the line conductors, keep the cpcs connected. Do not test line to neutral unless everything is disconnected.

Hi Westard. I've been back to this installation again.

Re "Testing an existing circuit you should have your neutral conductors disconnected and linked to the line conductors, keep the cpcs connected" - do you mean that at the board I leave the CPC connected and remove the line and neutral conductors for the relevant circuit and connect them with a jump lead? So as per the attached illustration?

Once I'm clear on that and I'm back at the installation I'll perform the test, but in the meantime I corrected a couple of mistakes I realised I had made with the wiring and retested as per post 14 and have different (better?) results:

Neutral to earth: >999
Line to earth: 0.31
Line to neutral: >999

As I understand it my Line to Earth is my issue as less than 2 mega ohms is a fail.

I've been around the ring circuit unplugging everything and identified a few accessories that I hadn't previously been aware of: a spurred electric radiator, its switch (which has a couple of neons), a socket with a USB port. I disconnected all these and - to maintain the ring - joined the relevant conductors (earth to earth, line to line, neutral to neutral) together with jump leads / wagos (which I presume was the correct thing to do???).

This made no difference however, the reading remained at 0.31.

One theory: 4 of the double sockets have tiny neon on/off lights on the switches - could that be responsible for the reading? Or am I simply overlooking something that's still in the circuit?

 

[brianmoooore]

I disconnected all these and - to maintain the ring - joined the relevant conductors (earth to earth, line to line, neutral to neutral) together with jump leads / wagos (which I presume was the correct thing to do???).

If you've broken the ring up, leave it broken and test the individual sections.

 

[TheJack]

Hi Westard. I've been back to this installation again.

Re "Testing an existing circuit you should have your neutral conductors disconnected and linked to the line conductors, keep the cpcs connected" - do you mean that at the board I leave the CPC connected and remove the line and neutral conductors for the relevant circuit and connect them with a jump lead? So as per the attached illustration?

Once I'm clear on that and I'm back at the installation I'll perform the test, but in the meantime I corrected a couple of mistakes I realised I had made with the wiring and retested as per post 14 and have different (better?) results:

Neutral to earth: >999
Line to earth: 0.31
Line to neutral: >999

As I understand it my Line to Earth is my issue as less than 2 mega ohms is a fail.

I've been around the ring circuit unplugging everything and identified a few accessories that I hadn't previously been aware of: a spurred electric radiator, its switch (which has a couple of neons), a socket with a USB port. I disconnected all these and - to maintain the ring - joined the relevant conductors (earth to earth, line to line, neutral to neutral) together with jump leads / wagos (which I presume was the correct thing to do???).

This made no difference however, the reading remained at 0.31.

One theory: 4 of the double sockets have tiny neon on/off lights on the switches - could that be responsible for the reading? Or am I simply overlooking something that's still in the circuit?

Bumping this up