Locating discontiniuty in CPC - how? Help?!

[Lobster]

So, today the plan was to swap in the sparkly new 17th edn consumer unit, for the old circuit-breaker job (see my previous thread about the selection of the aformentioned CU!)

Unfortunately, having removed the old CU, when I ran some tests on the existing ring main I found that although there was continuity between L and N on the two legs of the ring main, the two CPCs showed open circuit.

Shit.

I've been trying to track down the discontinuity for most of the day without success. I've opened up every socket and junction box that I'm aware of, and all connections are fine, so the only possibilities I can come up with are (a) a damaged cable, or (b) a loose connection in a junction box etc that I don't know about.

Unfortunately the ring main serves the entire (two storey) house, so there's a lot of area to check. (Not ideal, but not planning on upgrading the electrics at the moment, except I've just wired a socket to a blank way in the new CU, so that's currently the only one in the house! which is unoccupied BTW).

I'd really welcome advice on how best to track down the fault. What's the best way forward - I was thinking of removing all the plug sockets, trying to work out the route the cable takes between them, and then by temporarily shorting the conductors and CPC at one socket position, then check continuity at the next socket position, and try to work out which 'leg' has the problem. Does that sound about right, or is there a better way?

Thanks David

[A.Lee]

You need an accurate meter. In the OSG, there is a table of expected resistance values per metre for a variety of cables . It can be worked out, just about, how many metres from the CU that a fault is, if you do a little bit of testing. This will not trace the fault, but if it is at, say, 5 metres, it would be far easier to trace if the ring was 50 metres long.

You can take readings from each socket with a long roving lead. The one that shows a far different ohm reading from the next nearest one has probably got a fault between it and the next one in the ring. Or, join up line and earth at the cu, then measure at each socket, and again, any fault will show up with a far higher reading at one of the sockets - or it will go gradually higher, then drop down suddenly.

Alan.

[John MacLeod]

Well, one answer if the circuits are all dead at the moment is to the traditional "ring out" test with a bellset. However, if you have power to the circuits, then there's no need to open up boxes to determine what's connected to which leg. Disconnect one leg and use a simple neon socket tester and the neon connected between live and earth will show whether there's a connection from that particular socket to the leg that's actually connected.

[cynic]

Isolate and open the ring by removing both cable ends at the consumer unit. Short the L & CPC of one of the ring cable ends (not both) and progressively work around each socket checking for L-CPC continuity until you find which ones are not showing continuity and note them. This will get you into the right area. The open circuit may not be at the socket but at the next connection in line or damage to the cable between. Note also there may be more than one fault so having identified the good sockets open the connection L-CPC you made at the consumer unit and connect the other cable end similarly. then go round the rest of the sockets and test. Hopefully you will find all the others bar one are "good" You may have a hidden joint box and electricians are nothing if not inventive in their routing.

[Andrew Mawson]

between L

socket

What's

sockets,

Long length of wire connected to one earth tail, and do a continuity test to the earth of each socket working away from the consumer unit. When continuity ceases you are in the right area. You can then do the same using the other earth tail which hopefully will lead you to the other side of the fault area. Somewhere between is your fault!

AWEM

[Tim Watts]

Way to do this quickly is:

a) Disconnect all wires for that circuit at the CU. b) Connect L+E of one end only. c) Continuity test at every socket until you find the first one continuity ceases at (hint binary chop/search pattern makes this quick if you have any idea how the ring is run physically. d) Your fault is now pinned down to "near this socket" e) For a bit more finess, remove the shorting link and short L-E of the other end of the ring. f) You should now have the fault pinned down to being between 2 particular sockets for sure.

I know this works - I used it to locate a broken neutral.

[Dave Plowman (News)]

It would be quicker to get a long length of wire - anything would do - and use that to extend one of the leads to your meter. Would save a lot of disconnecting.

I have a cable tracer device - it injects a signal at one end and a receiver follows it along the cable. That will show where the break is pretty accurately - assuming you can keep the receiver reasonably close to the cable.

[Tim Watts]

There's no disconnecting in my (and the other chap's) method, except for the CU whch you'd need to do anyway.

agreed - if you had that it would certainly help with tracing the physical run.

[Ronald Raygun]

You don't need to open the sockets. As seen from the CU, you obviously have two different ring earths. Connect one of them to ring live and the other to ring neutral.

Then visit *every* socket in the house and test for continuity between earth and, in turn, live and neutral. Thus you can classify each socket as to which of the two ring earths (if any!) it is connected to.

Probably the easiest way to do this test is by using a short length of 3 core flex wired to a plug, and wrap the earth wire round one of the buzzer's probes and hold it in place with tape, then touch the other probe to the two remaining wires in turn.

Don't test the one socket which is still connected to the supply!

Ideally all socket earths will be connected to one or other of the earths, although if you're unlucky you may find the odd few which are connected to neither. By observing where the sockets change from "live earth" to "neutral earth", you'll get a better idea where the break is.

[dennis]

Don't people make things hard..

If the power is still on..

switch it off and unplug everything. disconnect the CPC off one leg at the CU. power up and run around the sockets with a simple mains socket tester. This will tell you the faulty and the working sockets. It should be easy from there to turn the power off and buzz the cores between the last working and the first non working socket.. Don't forget to reconnect the CPC before you put any appliances back.

If you don't have power a battery across the CPC and L/N and a 13A plug with a bulb wired across E and L/N will do the job. Don't leave the plug and bulb around for a kid to plug in a live socket.

[george [dicegeorge]]

This method is the best so far: as you're using a mains tester plug there's no danger of putting mains into a low voltage tester...

Then disconnect the other earth, reconnect the first, do the test again, mark the sockets with different coloured sticky tape depending on results (4 colours?) and make notes on a cipboard paper too, noting which colour tape means what!

[g]

[Lobster]

Thanks very much for all the replies. I used the above method which was very straightforward - wish I'd worked that out before all the head-scratching yesterday and not getting very far!

I was able to localise the aocket where the earth 'flipped' from ring live to ring neutral; however it's proved a bugger to take things any further. The two cables from the socket in question emerge under the bathroom and various stud partitions and as far as I can tell, the damaged leg is probably the most inaccessible in the house, buried in plaster; snaking under the bathroom floor and finally buried under tiles in the kitchen. I think.

Sadly I reckon I'll have to get a pro in to look at it - can't afford any more time on this. Maybe the damaged leg will need replacing with cable run in trunking (ugh)? But it will be a bit awkward, given the d-i-y half-installed CU though... lead balloon time methinks?

David

[<me9]

I did a similar set of tests at a friend's house where all 3 conductors showed OC. Eventually it was pinned down to between two bedroom sockets. Floorboards up revealed the wires disappearing up the wall into the plaster in the upstairs hall. investigating further found a back box with unstripped cable inside, covered over by plaster, obviously overlooked at second fix. A lack of socket on the upstairs landing has now been rectified by a new socket outlet.

It took many hours to find the fault as several hidden junction boxes were hidden under floors. I wonder what testing (if any) took place on the original installation.

I have yet to convince the owner of the need for further work to remove other departures from IEE regs as were several decades ago. I think the installation was 13th or 14th edition, now well past its sell by date!

[dennis]

Make into two 20A radials, it will just take a couple of breakers and a few snips at the cable to remove the faulty bit. Is there a lot of load to prevent this?

[Lobster]

Well it's a thought... however the entire house is wired with just one ring main - this one - and converting it to two radials would (IIRC, given the location of the fault) leave one radial with a single socket and the rest of the installation on the other radial. So I think not...

David

[ARWadsworth]

Voltage drop due to the cable length would also have to be taken into consideration.

[Mike Barnes]

Can you easily put in a new cable from the far end of the 'rest' to the CU?

[Ronald Raygun]

Ah, but you're not going to get much voltage drop on the CPC, are you? I know it's not in the spirit of the regs, but is there anything actually in the regs to say you can't mix ring topology for live and neutral with twin radials for the earth? So long as the earth actually gets to all the sockets, does it really matter where it comes from and how it gets there?

:-)

I've just had a thought. Is any of the bathroom pipework bonded to earth? If so, is it possible that it has been spliced in to the ring, and that one of the two connections has come adrift?

[ARWadsworth]

You will if someone puts a nail into the cable.

Ther are some rules, however I do believe the OP is not wanting a bodge job.

Possible, but most unlikely.

[Ronald Raygun]

So what do you think has happened here? A nail, which has managed to chop the earth conductor and neither of the other two, and which has subsequently been removed?

Presumably all three conductors will be taking the same physical route by virtue of being part of the same T&E cable. Would it be worth re-connecting the ring to the supply in order to use one of those cable detector devices which pick up the mains "hum" even from cables buried deep in plaster?

He could then (carefully!) connect the supply live to the two ring earths in turn, and thus (agin with said device) pinpoint the exact location where the break is.