Fault finding!!!!!!

[HappyHippyDad]

Hello all..

I have a problem with a tripping RCBO and it has me stumped!

Yesterday I fitted 10 RCBO's into a CU which currently had an RCD main switch. This was a sub circuit from a house CU which feeds various circuits for a cattery.

4 of the RCBO's are for the cattery pens. Pens 1,2,3 and 4 (each housing 16 seperate units). Pens 1-3 are 1st fix only (awaiting heat lamps). Pen 4 has heat lamps in.

When any heat lamp in pen 4 is energised the RCBO trips, however if the conductors in the CU for pen 4 are added to the RCBO for Pen 1 all is ok. However, when I IR test between lives of Pen1 and pen4 I get >1000MΩ?

IR tests for pen 4 heat lamps test ok, all >1MΩ.

When I disconnect the neutral fly lead in the CU for Pen 1 I get 197V L-N at Pen 4 and 240V L-E. When I reconnect neutral fly lead for Pen 1 the voltage increases to 240V for Pen 4 so Pen 4 must be utilising the neutral path of Pen 1 but it has >1000MΩ of resistance between neutrals!!

Luckily the cats are not cold at the moment!

Any suggestions would be appreciated :frown:

 

[telectrix]

cat curries?

 

[Pete999]

Daft question, but any chance of borrowed neutrals???

 

[telectrix]

Daft question, but any chance of borrowed neutrals???

but are the cats already neutered?

 

[Pete999]

but are the cats already neutered?

Yeah but did another cat borrow any?? went to the wrong Vet didn't I??

 

[HappyHippyDad]

Daft question, but any chance of borrowed neutrals???

Thats what stares at you Pete but I have put all new wiring in (all tested) bar a short length of existing 19 core SWA which I tested every single one (IR tested between all of them and to earth, found a fault to earth on one of them but have removed this from service so it has nothing to do with present situation).

Whats with the decreasing voltage when I remove the neutral fly lead from pen 1? Why am I not getting 240V with the correct neutral from Pen 4? I shall have to check continuity but i'm sure its fine! But it cant be otherwise I'd get 240V!?

I think I am going to have to return and disconnect every single RCBO and start testing from scratch!

 

[davesparks]

Thats what stares at you Pete but I have put all new wiring in (all tested) bar a short length of existing 19 core SWA which I tested every single one (IR tested between all of them and to earth, found a fault to earth on one of them but have removed this from service so it has nothing to do with present situation).

Are you saying you have disconnected a single core of a cable which has a fault, or disconnected a whole cable which has a fault?

 

[HappyHippyDad]

Are you saying you have disconnected a single core of a cable which has a fault, or disconnected a whole cable which has a fault?

There are 19 cores. Core number 13 read a poor resistance to earth so I have not used it in the design, I have connected both ends to earth. I also tested between core number 13 and all other cores and got acceptable IR readings. I dont think this is related to the problem but am certainly open to suggestion.

 

[davesparks]

There are 19 cores. Core number 13 read a poor resistance to earth so I have not used it in the design, I have connected both ends to earth. I also tested between core number 13 and all other cores and got acceptable IR readings. I dont think this is related to the problem but am certainly open to suggestion.

So there is a fault between one of the conductors which is layed up in the outermost layer and the armour? And you don't think maybe that cable is damaged?
Have you established what the fault is?

 

[Deleted member 26818]

Sounds like the heaters are causing some earth leakage.
Could be due to damp.
One solution would be to replace the RCBOs with MCBs temporarily to allow the heaters to heat up enough to drive away any moisture.
Is RCD protection a requirement?

 

[Octopus]

Time to do a few Rn + R2 readings.

what is the rating of the rcd main switch?

 

[PEG]

There are 19 cores. Core number 13 read a poor resistance to earth so I have not used it in the design, I have connected both ends to earth. I also tested between core number 13 and all other cores and got acceptable IR readings. I dont think this is related to the problem but am certainly open to suggestion.

This would be my starting point...regardless of the one single identified fault,it is one cable,ergo;you have a faulty cable...
You would struggle to test ALL the permutations of other possible faults,with that number of cores...

How long is this cable,and could you bridge out to test?

 

[Deleted member 9648]

Thats what stares at you Pete but I have put all new wiring in (all tested) bar a short length of existing 19 core SWA which I tested every single one (IR tested between all of them and to earth, found a fault to earth on one of them but have removed this from service so it has nothing to do with present situation).

Whats with the decreasing voltage when I remove the neutral fly lead from pen 1? Why am I not getting 240V with the correct neutral from Pen 4? I shall have to check continuity but i'm sure its fine! But it cant be otherwise I'd get 240V!?

I think I am going to have to return and disconnect every single RCBO and start testing from scratch!

Think you've answered your own question there HHD!

 

[timo1]

Hi HHD! was the installation on an RCD before? Is there a big ish start up current which may require a C type RCBO? If it was on an RCD before it should hold?

 

[HappyHippyDad]

So there is a fault between one of the conductors which is layed up in the outermost layer and the armour? And you don't think maybe that cable is damaged?
Have you established what the fault is?

Yes, that cable may be damaged and most probably touching the armour so I am not using it and have connected to earth. Can you explain what you are getting at Dave?

Sounds like the heaters are causing some earth leakage.
Could be due to damp.
One solution would be to replace the RCBOs with MCBs temporarily to allow the heaters to heat up enough to drive away any moisture.
Is RCD protection a requirement?

I have IR tested all lamps and all test ok.

Time to do a few Rn + R2 readings.

what is the rating of the rcd main switch?

I have removed the RCD main switch Murdoch as all are now on RCBO's. Previously it was an 80A rated switch, now just a standard double pole 100A main switch. Not sure what you're getting at though?

This would be my starting point...regardless of the one single identified fault,it is one cable,ergo;you have a faulty cable...
You would struggle to test ALL the permutations of other possible faults,with that number of cores...

How long is this cable,and could you bridge out to test?

I have tested all permutations Peg and it took a long time. I did this before changing to RCBO's to avoid this problem! :smiley2:
I have considered bridging and may well do that, thanks.

Think you've answered your own question there HHD!

Yes, I think I have haven't I

Hi HHD! was the installation on an RCD before? Is there a big ish start up current which may require a C type RCBO? If it was on an RCD before it should hold?

Yes Timo, It had an RCD as main switch.

 

[davesparks]

What I am getting at is that the cable is damaged, regardless of how many cores appear to be unaffected you cannot sensibly use that cable for any mains voltage knowing it is damaged.
If you nail through the centre of a 3core and earth cable would you just connect the danger core to earth and carry on using the other two cores?

 

[Deleted member 26818]

HHD, how many circuits utilise the old multi-core cable?
Of these, which circuits are causing problems?

 

[Richard Burns]

I would be tending to the damaged cable scenario.
But have you tried pen 4 in another RCBO, just in case that one is faulty.
Do you get any voltage on Pen 1 conductors if Pen 1 is disconnected but Pen 4 is connected and does the RCBO trip?

 

[HappyHippyDad]

What I am getting at is that the cable is damaged, regardless of how many cores appear to be unaffected you cannot sensibly use that cable for any mains voltage knowing it is damaged.
If you nail through the centre of a 3core and earth cable would you just connect the danger core to earth and carry on using the other two cores?

HHD, how many circuits utilise the old multi-core cable?
Of these, which circuits are causing problems?

I would be tending to the damaged cable scenario.
But have you tried pen 4 in another RCBO, just in case that one is faulty.
Do you get any voltage on Pen 1 conductors if Pen 1 is disconnected but Pen 4 is connected and does the RCBO trip?

With regards the 'damaged' cable, I cant remember the IR results precisely however they were a little below 1MΩ to earth. I am unsure of what I could do regards this cable as the 19 core is laid under a shed and concrete, it is about 15m's in length and all other cores are in use. What would you do Richard? This cable is required and the client is not going to agree to the cattery being dug up, there is also no viable option to lay a new cable. Thinking about it further, it will not be touching the armour as the IR results were only just below 1MΩ, still a fail though!

I have not tried Pen 4 in another RCBO by itself, I will do. I did ramp the RCBO and it was fine. Also I will try your suggestion regards seeing if there is voltage on pen1 with pen 4 connected.

I had been at the cattery for 10 hours and I had got to the point where my brain was not working properly so I left. This is one of the few times when a fault is getting the better of me and I'm finding it quite frustrating. I shall go back with a clear head and slowly and methodically test.

Edit... Sorry, quoted too many people!
Spin.. All circuits utilise existing 19 core.
Dave.. my reply to Richard half answers your question. I realise you are a stickler for doing things correctly, but would you really insist to the owner that he dismantle his shed, rip up the concrete etc for an IR result of a little below 1MΩ?

 

[Richard Burns]

I shall go back with a clear head and slowly and methodically test.

That is probably the answer!

Nothing one could do about the cable now, it should have been replaced at the time of installation if the fault was present then, or replaced once the fault was found but before finishing the cattery, both of which were presumably not viable options.
A value of 1MΩ will not affect a circuit noticeably; even though it is a fail as it is indicative of excessive deterioration either by damage or age.
Probably just damp through the sheath rotting the armour away!

 

[davesparks]

With regards the 'damaged' cable, I cant remember the IR results precisely however they were a little below 1MΩ to earth. I am unsure of what I could do regards this cable as the 19 core is laid under a shed and concrete, it is about 15m's in length and all other cores are in use. What would you do Richard? This cable is required and the client is not going to agree to the cattery being dug up, there is also no viable option to lay a new cable. Thinking about it further, it will not be touching the armour as the IR results were only just below 1MΩ, still a fail though!

I have not tried Pen 4 in another RCBO by itself, I will do. I did ramp the RCBO and it was fine. Also I will try your suggestion regards seeing if there is voltage on pen1 with pen 4 connected.

I had been at the cattery for 10 hours and I had got to the point where my brain was not working properly so I left. This is one of the few times when a fault is getting the better of me and I'm finding it quite frustrating. I shall go back with a clear head and slowly and methodically test.

Edit... Sorry, quoted too many people!
Spin.. All circuits utilise existing 19 core.
Dave.. my reply to Richard half answers your question. I realise you are a stickler for doing things correctly, but would you really insist to the owner that he dismantle his shed, rip up the concrete etc for an IR result of a little below 1MΩ?

15mtrs of PVC or XLPE insulated cable must be pretty badly damaged to get such an IR reading. Don't just blindly follow a number in a book, think about what it actually indicates in real terms.

No I wouldn't insist on dismantling a shed or ripping up concrete. I would find an alternative route for replacement cables

 

[marconi]

With regards the 'damaged' cable, I cant remember the IR results precisely however they were a little below 1MΩ to earth. I am unsure of what I could do regards this cable as the 19 core is laid under a shed and concrete, it is about 15m's in length and all other cores are in use. What would you do Richard? This cable is required and the client is not going to agree to the cattery being dug up, there is also no viable option to lay a new cable. Thinking about it further, it will not be touching the armour as the IR results were only just below 1MΩ, still a fail though!

I wonder if the 19 core cable is being gnawed by rats and their damage to core 13 is the first indication of more to follow. Rats are well known to live under sheds and even under concrete.

 

[Deleted member 26818]

So all the circuits utilise the existing 19 core cable, but only those with heaters connected are having problems.
Heaters are well known for causing RCDs to trip.

 

[HappyHippyDad]

So all the circuits utilise the existing 19 core cable, but only those with heaters connected are having problems.
Heaters are well known for causing RCDs to trip.

Spin, they work fine when in the same RCBO as Pen 1.

 

[ian.settle1]

cat curries?

Don't forget the furry hats as well.

 

[Lazlo]

Quote to run in new supply cables, get customer to arrange quote from groundworker for laying ducts underground to your spec. Nice job really, they may need to build a small meter cupboard/room either end for your switch/dist gear, again just provide them with a spec. Not your problem if they have a damaged cable. If they don't want to do it walk away....

 

[Leesparkykent]

Quote to run in new supply cables, get customer to arrange quote from groundworker for laying ducts underground to your spec. Nice job really, they may need to build a small meter cupboard/room either end for your switch/dist gear, again just provide them with a spec. Not your problem if they have a damaged cable. If they don't want to do it walk away....

I think HHD installed the cable.

 

[Octopus]

Hello all..

I have a problem with a tripping RCBO and it has me stumped!

Yesterday I fitted 10 RCBO's into a CU which currently had an RCD main switch. This was a sub circuit from a house CU which feeds various circuits for a cattery.

4 of the RCBO's are for the cattery pens. Pens 1,2,3 and 4 (each housing 16 seperate units). Pens 1-3 are 1st fix only (awaiting heat lamps). Pen 4 has heat lamps in.

When any heat lamp in pen 4 is energised the RCBO trips, however if the conductors in the CU for pen 4 are added to the RCBO for Pen 1 all is ok. However, when I IR test between lives of Pen1 and pen4 I get >1000MΩ?

IR tests for pen 4 heat lamps test ok, all >1MΩ.

When I disconnect the neutral fly lead in the CU for Pen 1 I get 197V L-N at Pen 4 and 240V L-E. When I reconnect neutral fly lead for Pen 1 the voltage increases to 240V for Pen 4 so Pen 4 must be utilising the neutral path of Pen 1 but it has >1000MΩ of resistance between neutrals!!

Luckily the cats are not cold at the moment!

Any suggestions would be appreciated :frown:

Are you 100% sure you have the correct neutrals in the correct RCBO's? Having read this again thats where I would start.

 

[Hellmooth]

I think HHD installed the cable.

He says he installed all new cable bar the 19core SWA.

 

[UNG]

I don't know how long you have spent messing around with this 19 core but given the IR readings I would suggest you are flogging a dead horse and wasting time and effort thinking it is worth saving. Even if you do get it working how long will it last before you start getting call backs as the cable deteriorates even more and the customer blames you and wants it fixing at your expense. I think the customer has to bite the bullet and install duct(s) for a new cable and any possible future changes and do the job properly from the start even if there is some disruption while the work is carried out

 

[Lucien Nunes]

FWIW I once met a cable that was intended to be re-used, I can't recall exactly what it was, 12 core 2.5 possibly, that tested generally OK but a couple of cores had a few megohms to earth. Turns out someone had put a core drill through about 8 of the cores, which had then been peeled away from the cable and the breaks bridged out with 3" lengths of wire with a choc-block at each end. But the whole lot was in contact with damp concrete and the blocks were corroded and turning green and leaking to earth, and one had melted and the insulation burned away along the cable. The surprising thing is that the continuity was fine at first inspection.

My point here (and Davesparks') is that the fault you detected on the cable may not in itself be a problem in your application but it is an indicator of an unknown degree of damage. It could be a small screw through the cable, it could be seriously mangled, you don't know. But you do know for certain that it is damaged.

The correct procedure for testing insulation is to connect all cores and the armour together to start with. Disconnect one core, test it to the remainder, then reconnect and move to the next core. 20 tests required (don't forget the armour to all the cores) and each result shows as near as possible the quality of insulation of that core alone. Any fault will show up exactly twice.

When I disconnect the neutral fly lead in the CU for Pen 1 I get 197V L-N at Pen 4 and 240V L-E

What testing device are you using? Don't forget that a voltage reading on a supposedly disconnected core is totally meaningless! If your meter has a high input impedance, capacitive coupling can raise the voltage very close to supply voltage, especially as in 19-core armoured it's possible for one core to be completely surrounded by cores at line potential throughout its length. Even with an undamaged cable I would not be surprised to see an unused core float up to 230V on my meter, although of course it won't light a test lamp.

 

[Deleted member 26818]

Getting rather confused here.
Circuits are fine on another RCBO?

 

[marconi]

Dear HHH, I am interested in this from the point of view of 'electrical detective work' which keeps me amused and the grey cells working. But may be I might be helpful too. The big question mark is against the 19 core armoured cable. Like the other commentators I think that unless you are able to confirm beyond any doubt that the cable is sound - mechanically, electrically and materially - then you have no option but to replace it in toto or find where it is faulty and replace the defective section.

What follows is how one might go about investigating the 19core cable (19CC) to arrive at some useful information and an idea on locating the fault.

1. Disconnect the cable completely at both ends, all the conductors and the glands and remove the glands from the cable ends. Take a close look at the state of the cable that was within the glands.

2. At each end of the cable, put all the conductors separately into a plastic connector strip.

3. At each end link all the conductors together but leave the armouring isolated.

4. Divide a piece of A4 paper down the middle, and mark the left division supply end and the right division load end.

5. At each end of the cable measure the IR between all 19 conductors and the armour. Record results and look for any differences.

6. Connect armour and 19 bundle at one and measure continuity of loop armour - bundle at the other. proves electrical continuity of armour (not mechanical though).

7. Disconnect linkage between armour and bundle. Remove conductor 13 at one end. At the other end measure continuity of conductor 13 -bundle loop. Shows whether conductor 13 is continuous or not

8. Disconnect 13 at both ends. Measure IR of conductor 13 to a)armour and b) bundle. Do this at both ends of cable.

9. At supply end remove one conductor at a time and measure loop continuity with bundle. Replace conductor after test. Checks continuity of each conductor.

10. If a conductor at para9 is found not continuous or of higher resistance then for this/these conductors carry out this taking measurements at both ends of the cable. Otherwise, starting with conductor 1, disconnect at both ends and then measure IR to a)Armour and b)bundle. Replace conductor after test.

11. We now have a table of information on continuity, and IR of each conductor to armour and all other conductors in the bundle. Anything odd?

How to home in on cable fault associated with conductor 13? - An idea:

SAFETY NOTICE - LIVE WORKING

My idea is to create a potential drop along the length of conductor 13 and then use the way the potential divides up and down stream of the fault along conductor 13 to create a ratio of conductor lengths. Don't think it will work for a spark gap type of fault but may be for TREE on XLPE - not sure.

Test set up: At supply end connect conductor 13 to neutral and a known satisfactory conductor to live at the load side of an RCBO. Keep armour disconnected both ends. Connect a high current load at far end of 19CC - so if 2.5mm2 conductors say 4-5kW. Connect voltmeter between start of conductor13 and armour. Turn on RCBO - measure potential difference PD1. Now connect voltmeter to conductor 13 and armour at far end. Avoid turning off RCBO between measurements.

Ratio PD1/PD2 approximately ratio distance (start cable to fault)/fault to end of cable).

Some experimentation and scaling of load obviously required to suit rating of cable size and RCBO. But tests short lived and so ohmic heating of 19CC not a concern.

 

[Lucien Nunes]

Don't think it will work for a spark gap type of fault but may be for TREE on XLPE - not sure.

It won't be anything as exotic as a tree. I've seen plenty of screws through armoured cable but never a PE tree at LV.

use the way the potential divides up and down stream of the fault along conductor 13 to create a ratio of conductor lengths

This is an approximation to the Murray loop test, which can locate high-resistance faults in low-resistance circuits as it is a bridge method in which the fault carries no current at null. However, because we're dealing with a buried SWA cable, an insulation fault from any core suggests it has been punctured, in which case the armour will probably be equally or more leaky to earth (I don't think the OP tested this, one normally wouldn't). As your proposed test uses an earth-referenced supply (the mains) any leakage from the armour to earth will spoil the result. Also, deliberately energising a known faulty conductor above ELV, with an armour deliberately not earthed, is poor practice.

A better method would use an isolated LV supply - safer and unaffected by leakage to earth. I have a test set for exactly this purpose, you can pump out 1, 10 or even 100A, whatever you need to get a decent potential between the ends, and so long as a few microamps escape at the fault you can find it, provided the conductor itself is mostly intact.

Although core 13 is not required in the installation, it can still serve a purpose to physically locate the fault, enabling a decision on whether it is possible to salvage the cable. Faults often occur at penetrations etc, so it might be possible to access it for repair without too much work.

 

[marconi]

Lucien Nunes has pointed out a couple of flaws in my test rig - thank you.

Undaunted, I have thought about how you could use items readily available to an electrician to engineer a low voltage test rig. Now, before I go further you could/can hire the low voltage current source and that would be the neatest way to proceed. What follows is an idea on how you can use your van's 12Volt battery and a 50m reel of 2.5mm2 conductor instead.
Some physics:
Resistance of 18m length of 2.5mm2 csa copper conductor = resistivity x length/csa = 1.7 x 10exp-8 x 18/(2.5 x 10exp-6) = 4.896 x 10exp-2 = 0.05 Ohms.
Resistance of 50m of 2.5mm2 is (same equation above but replace 18 by 50) = 0.14 Ohms
Some Ohm's Law.
Connect the 50m of 2.5mm2 conductor to the two cores of 19CC and apply 12Volts (briefly) and a current of approximately 12/(0.14 + [2 x 0.05]) = 50 Amps. This ignores internal resistance of battery so current will be a little less - Rint is of the order 0.003 Ohms.
We need to cause reasonably high PD1 and PD2 to measure without damaging the conductor 13. For 50 Amps that would be a voltage drop along conductor 13 of 50 x 0.05 = 2.5 Volts.
Power dissipated in 2 conductors in 19CC is 50 x 50 x 0.1 = 250 watts or 250/18 = 7 Watts per metre of its length. Not a worry for this short test by which I mean less than 60seconds.
Power dissipated in cable reel is 50 x 50 x 0.14 = 350 watts which is a problem when coiled on a reel. So uncoil it completely.
Revised test rig: [12 volt battery]- [car jump leads]-[100amp wylex isolator switch] - [conductor13 and one other good conductor of 19CC] - [50 m of 2.5mm2 conductor uncoiled]
Voltmeter measurements as before but only one end at a time. Take several pairs(PD1 and PD2) of readings (say 10 off) and take average of each. Allow 60second pause between tests. Calculate ratio and apply to 18m length of 19CC.
Again, one would need to recalculate for different conductors csas - I have assumed 2.5mm2 in 19CC and 50m conductor load.
Best carried out with some help from another to control the switching and just in case to be ready to disconnect jump leads. (I am not sure how an ac wylex isolator will perform switching dc - better to use a dc high amperage PV installation isolator switch).

 

[davesparks]

The cable is damaged and buried under a concrete slab with a building on it.
Is there really any point trying to establish whereabouts the damage is if the ultimate soloution is going to be installing new cables via a seperate route?

 

[Lucien Nunes]

We don't know - maybe there is a screw through the back of the cable where it's above ground. It's less likely to be punctured all the way through a core within the concrete. I'd have given it 20 minutes of my time to try and locate the fault, and in the absence of anything else this is the sort of test I'd rig up.

But I wouldn't use those wonky numbers... Marconi check your multiplication! FWIW the tabulated resistance of 2.5 is 7.4 - 7.5 milliohms per metre.

 

[marconi]

Lucien Nunes has pointed out a couple of flaws in my test rig - thank you.

Undaunted, I have thought about how you could use items readily available to an electrician to engineer a low voltage test rig. Now, before I go further you could/can hire the low voltage current source and that would be the neatest way to proceed. What follows is an idea on how you can use your van's 12Volt battery and a 50m reel of 2.5mm2 conductor instead.
Some physics:
Resistance of 18m length of 2.5mm2 csa copper conductor = resistivity per metre x length = 0.0075 x 18 = 0.135 Ohms.
Resistance of 50m of 2.5mm2 is (same equation above but replace 18 by 50) = 0.375 Ohms
Some Ohm's Law.
Connect the 50m of 2.5mm2 conductor to the two cores of 19CC and apply 12Volts (briefly) and a current of approximately 12/(0.375 + [2 x 0.135]) = 18.6 Amps. This ignores internal resistance of battery so current will be a little less - Rint is of the order 0.003 Ohms.
We need to cause reasonably high PD1 and PD2 to measure without damaging the conductor 13. For 18.6 Amps that would be a voltage drop along conductor 13 of 18.6 x 0.135 = 2.511 Volts.
Power dissipated in 2 conductors in 19CC is 18.6 x 18.6 x 0.27 = 93.4 watts or 93.4/18 =5.18 Watts per metre of its length. Not a worry for this short test by which I mean less than 60seconds.
Power dissipated in cable reel is 18.6 x 18.6 x 0.375 = 129.7 watts which may be a problem when coiled on a reel. So uncoil it completely.
[Check that power in = 12 x 18.6 = 223.2W should equal power out = 93.4 + 129.7 = 223.1W. yes]
Revised test rig: [12 volt battery]- [car jump leads]-[100amp wylex isolator switch] - [conductor13 and one other good conductor of 19CC] - [50 m of 2.5mm2 conductor uncoiled]
Voltmeter measurements as before but only one end at a time. Take several pairs(PD1 and PD2) of readings (say 10 off) and take average of each. Allow 60second pause between tests. Calculate ratio and apply to 18m length of 19CC.
Again, one would need to recalculate for different conductors csas - I have assumed 2.5mm2 in 19CC and 50m conductor load.
Best carried out with some help from another to control the switching and just in case to be ready to disconnect jump leads. (I am not sure how an ac wylex isolator will perform switching dc - better to use a dc high amperage PV installation isolator switch).

Thank you to Lucien Nunes - again. Calculations corrected and the figure of 7.5 mOhm per m used for resistivity of 2.5mm2 conductor.

 

[HappyHippyDad]

FWIW I once met a cable that was intended to be re-used, I can't recall exactly what it was, 12 core 2.5 possibly, that tested generally OK but a couple of cores had a few megohms to earth. Turns out someone had put a core drill through about 8 of the cores, which had then been peeled away from the cable and the breaks bridged out with 3" lengths of wire with a choc-block at each end. But the whole lot was in contact with damp concrete and the blocks were corroded and turning green and leaking to earth, and one had melted and the insulation burned away along the cable. The surprising thing is that the continuity was fine at first inspection.

My point here (and Davesparks') is that the fault you detected on the cable may not in itself be a problem in your application but it is an indicator of an unknown degree of damage. It could be a small screw through the cable, it could be seriously mangled, you don't know. But you do know for certain that it is damaged.

The correct procedure for testing insulation is to connect all cores and the armour together to start with. Disconnect one core, test it to the remainder, then reconnect and move to the next core. 20 tests required (don't forget the armour to all the cores) and each result shows as near as possible the quality of insulation of that core alone. Any fault will show up exactly twice.



What testing device are you using? Don't forget that a voltage reading on a supposedly disconnected core is totally meaningless! If your meter has a high input impedance, capacitive coupling can raise the voltage very close to supply voltage, especially as in 19-core armoured it's possible for one core to be completely surrounded by cores at line potential throughout its length. Even with an undamaged cable I would not be surprised to see an unused core float up to 230V on my meter, although of course it won't light a test lamp.

Thank you so much for reminding me of the ghost voltage Lucien. This is exactly what it was! At one point on the run it was actually reading 209V L-N, I then used the Drummond and no light was present so no voltage. It was then straight forward that I must have connected the neutrals incorrectly at some point. In the CU I had all 4 neutrals for the 4 pens the wrong way around, ie 1 in 4, 2 in 3 etc (thanks Murdoch as well as I think you said this).

The neutral problem was obvious but the ghost voltage was throwing me and I was very tired. 30 mins with a clear head (and a little help) and problem solved.

 

[HappyHippyDad]

Are you 100% sure you have the correct neutrals in the correct RCBO's? Having read this again thats where I would start.

Oh dear