This discussion is locked.
You cannot post a reply to this discussion. If you have a question start a new discussion

Tricky fault finding and rectification job

Former Community Member
Former Community Member
Hello,

I’d be grateful for some advice on a tricky job.

Got an evening call out to a house where the power had gone off, the RCD kept tripping.

Arrived, checked, it was an old-fashioned fuse board with six circuits, two 30 amp rings, upstairs and downstairs, two 5 amp lighting circuits, up and down, a 16-amp combi boiler and a 40-amp shower fuse. The RCD had been installed at some time on the tails between the meter and board to cover the fuses.

As I investigated, I found some problems. I opened up all the ground floor sockets, most of them were OK, wired correctly, LL, NN, EE, but a switch for an outside light had been wired directly from a socket without a fused spur, I disconnected it. I also found a double socket spur that had been wired directly from the cooker control switch; the neutral terminal inside the double socket had melted, so I disconnected that and made it redundant. Also found a double socket that had partly melted at the front, replaced that with a new one.

When I switched the board back on (after safe isolation for the work) the RCD tripped again. It was an old one, so I replaced it, but tripping continued. I eventually pulled out the fuse for the downstairs ring, and everything else worked fine. Customers don’t have a downstairs ring circuit for the moment, they are plugging into a 10-meter long four-socket extension running from a socket on the upstairs ring.

That’s the scenario, now I have four questions:

One, I’m going back on Tuesday (earliest I can get back) to do a split-ring IR test, problem is, I’ve been out of work for months due to the lock down, and the calibration on my Fluke MF tester is a few months out of date, that’s why I didn’t take it first time, will the IR tests I do be fairly reliable?

Two, I’ve noticed that the cooker control switch is working, as there is power on the cooker socket, but there is no fuse in the board for the cooker, what might be going on there?

Three, I did a belling out test with a long extension link to establish how the downstairs ring is interconnected, and I noticed that a leg of old red and black cable in a front room socket is linked to a leg of new brown and blue cable in one of the kitchen sockets, i.e. when I put the croc clip of my test cable onto the copper end of the red live in the front room, and the other croc clip of the test cable onto one probe of my Kewtech hand tester, the tester beeped as soon as I touched the other probe onto the copper end of a brown live in the kitchen. so I’m assuming that the old and new cables are linked by a hidden junction box; that might be causing the problem. How can I locate the box?

Four, if I do narrow the fault down to a leg of the circuit, then how can I repair that leg? I ask because it’s an old house, built in the 1940’s, and if, say, a leg of the ring has damaged cable, how do I replace that without tearing out part of the wall and ceiling, causing inconvenience to the customers?

Thank you if you can help.

  • The route to this kind of fault is simple divide and conquer. Let's assume you find an insulation fault on the red ring wire, BTW your meter calibration doesn't matter for this as long as it works. Disconnect both reds at the fuse and keep separate. Go to the opposite side of the house and find a socket. Disconnect the reds and see which one has poor insulation resistance. Leave it apart. Find the next nearest socket and repeat. You will find two adjacent sockets with the fault between. Remember this could be a spur, so run a test lead between the two good red ends and check the ring continuity at the fuse. Replace the faulty cable in whatever way you can, surface, mini trunking, whatever the client wants. Fully test the ring (r1+r2 etc.) and check the Zs at each socket, there could be another fault! Job done
  • I'll have a go at a couple of the questions.


    Calibration: unlikely to be miles out, especially if it has just been loafing around doing nothing. If your MFT gives results that you don't expect, that's the time to be questioning the calibration. (Naturally you would have been using your checkbox regularly despite lockdown. ? ) Don't worry about it.


    If there is a damaged cable it has to be replaced or bypassed. There isn't a simple solution, but if a previous sparks has put it under capping, you will bless him. ?


    ETA: cooker socket - there was no mention of a cooker circuit!


  • " I’ve noticed that the cooker control switch is working, as there is power on the cooker socket, but there is no fuse in the board for the cooker, what might be going on there?" Perhaps the electric shower got removed when the combi boiler went in and a shower mixer was installed, maybe that is now the cooker circuit? Or they used the cooker circuit cables to feed the combi, and added a double socket apparently.


  • Just a thought, it would be worth taking the batteries out of MFTs etc. if not using them for a period of time, with another lockdown in mind. They might leak and damage tester.


    Edit; in answer to part 4 of OP, disconnect the damaged part of the ring at the socket outlets, derate the fuse to 20A and create a centre-fed radial circuit, assuming there are not too many loads.
  • " One, I’m going back on Tuesday (earliest I can get back) to do a split-ring IR test, problem is, I’ve been out of work for months due to the lock down, and the calibration on my Fluke MF tester is a few months out of date, that’s why I didn’t take it first time, will the IR tests I do be fairly reliable? "


    If you use your test instrument regularly you begin to know what is correct and what is false. Do you trust your instruments readings?


    Standard test for rfcs is to join outgoing LN then LCPC with their incoming counterparts etc.


    Another rather import check would be a basic continuity test between N and CPC. 

    If there is a low resistance then there has been a badger in during the night doing nocturnal activies on the electrics.

    You see there is no way that you can install an RCD that will not trip on a circuit that has an interconnected N and an earthe cpc or a cross connected neutral usually found on lighting circuits.

    Legh


  • You are over thinking this and getting over involved.


    The RCD has been tripping, so turn off the supply at the main switch in the consumer unit and at the upfront RCD.


    Prove the whole installation dead.


    With all the fuses in place and everything connected, link the live and neutral in the fuse board and test between them and earth at 250 Vdc with the insulation tester, hopefully it will be nearly zero showing a definite fault so you have something to look for.


    Assuming there's a low insulation test result  and you have the insulation tester connected to the live between the main switch and fuses as well as the neutral bar, take the fuses out one at a time then start disconnecting the neutrals from the bar, hopefully then as you disconnect a faulty circuit the insulation test result will improve significantly.


    When you have found the circuit with the low insulation test result and it is fully disconnected start faulting by halves. Guess where the middle of the circuit is and disconnect a fitting dividing the circuit into two, then test each half. Then when you have identified the faulty half, repeat the process guessing where the middle is and finding the faulty half of the half a circuit; and so on and so fourth until you reach the length of cable that is faulty, the faulty fitting or appliance.


    Once you are sure there’s not any appliances to blow up you can increase the test voltage to 500 Vdc.


    Do the repair or remove the faulty appliance, reassemble and test that circuit, reassemble the fuse board, test the RCD, write out a minor works certificate and the invoice, get your card reader out and take payment by card or cash or cheque.


    Comment on the rest of the installation, but if it isn’t lethal leave it alone, you were called out to sort out the tripping RCD and you have done that.


    There is no reason why your tester should suddenly become inaccurate, however you should have some way of checking basic functions yourself with a calibration check boxer the like, which you can make yourself. I bought a Seaward Checkbox 16 many years ago plus a couple of resistor boxes to to interim checks, but to be honest if you use your testers frequently you soon realise if something is amiss. 


    If you don't feel confident  to test your own testers and assure yourself that they are reasonably accurate you should not be using them to test electrical installations.


    Don't get over involved.


    Andy
  • There’s a selection of calibration check cards and boxes on EBay and other websites suitable for checking continuity and insulation test functions on your tester.

    https://www.ebay.co.uk/itm/NEW-LOOK-NICEIC-CalCheck-Calibration-Checkbox-Insulation-Resistance/123632330550?epid=28028987925&hash=item1cc90fa736:g:VnsAAOSwa1ZcWs7C
  • Use the calibration check immediately after the meter has been calibrated and record the results, which may be a bit off from what it says on the card or box, then check the meter doesn’t drift.
  • "Use the calibration check immediately after the meter has been calibrated and record the results "


    Ditto that
  • For fault finding like this, do not get hung up on the meter calibration unless you are in the habit of really throwing it around or have good reason to suspect it is broken. - almost any meter will do, we are not worrying about a decision based on an insulation resistance of 1.01 megohms pass versus 0.99 megohms fail here.

    To trip a 30mA RCD the fault impedance will be very low, less than 30k (0.03 Megohms). If the meter can tell the difference between the test probes shorted and not shorted (always a good test to do that at the start of session, as it is the test leads & their connections that suffer most damage) it is good enough to go for this one. In terms of the precision needed to find a fault like this, almost any old meter and at a push even a battery and doorbell can  get you quite a long way.


    Sometimes the users can unwittingly provide the clues to speed diagnosis- was there any building work or  DIY in the house recently that might have nailed a hidden cable, or a new squeak from an floorboard that is now working loose, any leaks from plumbing (or animals for that matter) that could have got into a joint under a floor ?

    IF you can lift part of the floor, sometimes a mirror and a torch can be used, rather like a periscope,  to see further along into places you cannot get your head.

    I second the 'divide and conquer' approach, but sometimes to jumper around and isolate the defective section  is worth considering, without actually finding the fault, if it is plastered in or under a tiled floor for example.

    It is possible to remove a section of ring, isolating the ends to create a centre-fed radial,  and reduce the fuse to 15-20A, which is another quick fix. No good if the ring is heavily loaded of course.