110 V between L and CPC

[KeenPensioner]

I am in training and decided to do some dead tests using my second hand Fluke MFT 1653B on an upstairs ring. I did the tests at a socket rather than at the CU. Results end to end L - L 0.74 ohms, end to end N-N 0.76 ohms, end to end cpc >2000 ohms. I checked all the sockets on the ring and couldn't find a break in the cpc. With the circuit energised I'm getting 110 V between L and cpc at the test socket as well. No idea what's going on. The cu is fully RCBO. Any help would be appreciated.

 

[Julie.]

Sounds like the cpc is broken in two places, and you are picking up capacitive/leakage voltage on the floating portion.

Carrying out dead tests, link one end of the cpc with the corresponding L then do a continuity test at each socket outlet by just plugging in the mft checking L-cpc. As you work around you will find the last one with cpc, and the next without (so a break between them), then repeat for the other leg of the ring. You should be able to find the cpc breaks quite easily.

 

[Avo Mk8]

Re your 110V reading on the energised circuit, if there is anything plugged in (or even something forming part of the fixed installation) that has an interference filter in it (which typically includes capacitors from L and N to E), then that can give half line voltage on a disconnected earth when measured with a hight impedance tester. It implies that part of the CPC is open circuit.

 

[Lucien Nunes]

As above, a floating section of CPC will tend to take up a voltage somewhere between line and neutral/earth potentials. The more symmetrical the leakage pathways, the nearer it approximates to a 2:1 voltage divider and if cable inter-core capacitance is the main factor, the longer the cable run the lower the source impedace of the 'ghost' voltage. With T+E, where the capacitance from to each live core is very similar and the stray capacitance to the otside world is minimised by the layout of the conudctors, a long run can produce quite a convincing 50% of line volts, usually pulled down a bit by the input resistance of the meter and the stray to real earth.

 

[KeenPensioner]

Thank you all for replying, I appreciate that. I followed Julie's advice and linked the cpc and L from each leg and did get continuity from different sockets depending on legs. There are 8 sockets on the ring (and 1 spur). From one leg I got 6 giving continuity and 2 with >2000 ohms. Swapping the legs I got exactly the opposite (2 working and 6 not). Unfortunately I have no idea how the cable runs between each socket. Is all I'm left with is lifting boards and going in the loft? Also, I got perfectly acceptable Zs figures for each socket - don't understand that. Anyway, thanks again - every day is a school day.

 

[timhoward]

You can do a similar test to find the wiring route or at least get some idea, take both ends out, Link L and CPC again on one leg and measure the resistances, the resistance will increase the longer the wire is and therefore this gives a good idea as to the order the sockets are wired.
If there are no spurs, the break should be between the sockets with the highest resistance reading when you link each end in turn. The other leg needs to not be joined for this to work.
Hope that makes sense, it's easier to do than describe!

 

[Julie.]

Thank you all for replying, I appreciate that. I followed Julie's advice and linked the cpc and L from each leg and did get continuity from different sockets depending on legs. There are 8 sockets on the ring (and 1 spur). From one leg I got 6 giving continuity and 2 with >2000 ohms. Swapping the legs I got exactly the opposite (2 working and 6 not). Unfortunately I have no idea how the cable runs between each socket. Is all I'm left with is lifting boards and going in the loft? Also, I got perfectly acceptable Zs figures for each socket - don't understand that. Anyway, thanks again - every day is a school day.

Well it sounds like you know where the break is, i.e. between which two socket outlets, now is the difficult part - finding the cable route, if there is a spur, then it may be there is a jb in this portion of the run, finding it is a matter of working out the path - how the cable leaves the socket, - which is the logical path to the next socket etc.

This needs to be addressed, you can't leave the ring as is, either sort it, or change it to two radials (max 20A RCBO not 32A), and remove any socket with no proper cpc connection.

 

[KeenPensioner]

Thanks again. I have a spare way in the cu so I think I'll just split it into 2 (20A) radials

 

[brianmoooore]

You can't just leave the faulty cable connected. Both ends of it must be completely disconnected from their respective sockets.
If the bank of MCBs is protected by a RCD, and the max amps rating of that RCD is lest than the upstream fuse protecting it, then the sum of your individual MCB ratings, added together, including the extra 12A you are introducing, must be less or equal to the RCD's rating.

 

[KeenPensioner]

Can't thank you enough for your help. I'm not qualified but have loads of theory. I managed to trace the fault in the cpc - it was at the cu with an over tightened terminal that had broken one of the conductors. I'm chuffed to bits.

 

[DPG]

Can't thank you enough for your help. I'm not qualified but have loads of theory. I managed to trace the fault in the cpc - it was at the cu with an over tightened terminal that had broken one of the conductors. I'm chuffed to bits.

Great to hear.