Voltage up to 100v on CPC even when disconnected

[chrisses]

Hi
I've never posted in one of these forums before, but have read lots of threads and found them extremely helpful.
So it makes sense seeing as I know the knowledge on this sight is very impressive to ask you guys for help.
I have a problem with a job and can't work out what the problem is I've been back there two days in a team of two and haven't fixed the problem and I'm starting to feel a bit stupid

It all started when I got a call from a customer I completed a full rewire and extension for last year, She claimed to be getting small shocks from the metal light switches in one room of the house. I turned up expecting it to be something simple, Maybe even static.
Turns out at that switch there was 44v on the earth wire.
Ok so at this point I'm thinking that some where on the circuit there is a damaged cable, not enough to trip but slight damage
I strongly suspect that the most likely cause is a plate screw slightly intruding on the insulation of a live so I unscrew all the plates.
I should add at this point that I went to the fuse board in the garage and discovered that they had a new automatic garage door fitted and the rails ran right in front of the fuse board, Giving me just enough space to get the fuse board cover off but not to safely work in it or test properly.
However back to the point I discovered that every earth on a feed in or out cable has voltage on it some as high as 100v!. When I turn the breaker off to this circuit this voltage disappears completely.
I disconnect the earth of this circuit out of the board and this earth has voltage, So its not coming from the mains, This is where it gets confusing to me, I disconnect all the earths from each other and from the board expecting to find which cable has the fault. All the earths are still live!
This doesn't seem possible to me and I'm hoping someone will make me feel stupid and point out something I've missed
I'm assuming the neutrals must be involved because otherwise it wouldn't be possible for all earths to be live even when disconnected from each other.
As a temporary solution I have removed all the metal plates and replaced them with plastic. Obviously the customer is not happy with this and I need to go back and remove the garage door bracket to get to the board and test properly.
Any replies of help would be most welcome (I'm a desperate man) please feel free to make me feel stupid I have no doubt in my ability as an electrician but I am also under no illusions that there are electricians which have forgotten more than I know! Seriously though I really appreciate you taking the time to read this and would appreciate ALL suggestions, I really want to sort this because how else do we learn

 

[Sintra]

I would start by confirming voltages between L&N L&E and N&E both at the incomer and then on the final circuits.

 

[spark 68]

Where/what are you measuring your earth to ?

what is your reference ?

 

[chrisses]

Ok last time I was there

L-E 230v
L-N 230v
N-E 63v (at board) upto 100v in other parts of the circuit

 

[spark 68]

What type of earthing system ? is the next question.

 

[chrisses]

I measured earth to neutral first then to an earth of another circuit

- - - Updated - - -

Its a TN system

 

[chrisses]

Its a TN system

 

[spark 68]

TN what though ? TN-S or TNC-S, have you done any IR testing ?

 

[Marvo]

There's a few things I don't fully understand from your post, especially the part about separating the earths and the neutrals being involved.

A few questions first;
Did you test the incoming supply?
What is the earth impedance of the supply?


I can give you some general advice, if you get ashock from something then there at least two faults present. Firstly there's an insulation problem causing current to leak down the CPC to earth. Secondly there's an earthing problem which allows a high touch voltage to develop because the earth impedance is too high.

Tracing this problem by measuring the touch voltage isn't the best way, rather use your mega tester to find the low insulation fault and then use your earth impedance tester to find the earthing issue.

 

[jon4381]

Borrowed neutral or maybe an earth leakage from a boiler

 

[spark 68]

I was thinking on the same lines as Marvo, ie. two faults, maybe a supply Earthing problem, and a fault on the house wiring.

 

[NMc]

When you completed the rewire was everything tested out ok, if so, see what has been added since as a starting point.

 

[ifindoubtdont]

You have mentioned an automatic garage door has been fitted since you rewired, I would start looking in that area first. The people who fitted it may have only been garage door fitters "cus it's only free wires innit, wot cun goo rong."

 

[telectrix]

if you are getting a voltage N-E with your CU main switch off, i'd strongly suspect a fault on the supply. need Ze readings for a start, then IR tests on the installation. could there be a problem with a neighbour's installation back feeding your earth?

 

[chrisses]

Lots of questions to answer so please let me know if I miss any
first of all let me clarify that if I remove the cpc for this circuit from the earthing terminal in the consumer unit I still have voltage on all the earths including the disconnected one in the consumer unit but not from the main earth bar.
secondly I then disconnect all the cpc's from each other at each switch (fed at switches not 3 plate) I still have voltage on the earth on the incoming and outgoing feed even though they are disconnected from each other
The initial verification test was mainly performed by myself however dead tests were done by someone else and I did the final test in a two man team
i have spoken to the second man and he has admitted to me that he tried to take an earth loop but the tester wouldn't let him so he worked it out from the R1 + R2
the the earth impedance of the supply is recorded as 0.05
the garage door is just on a plug top plug into the garage ring
i think I've covered it all?

 

[chrisses]

Oh and to clarify as well when I switch off the breaker to this circuit only the earth voltage disappears

 

[Tiny Spark]

Have you dont IR tests on all circuits?

 

[spark 68]

i have spoken to the second man and he has admitted to me that he tried to take an earth loop but the tester wouldn't let him

Oh dear!, so instead of investigating this issue (above 50V touch ?), the installation was put into service with a possible inherent defect.

 

[telectrix]

Oh and to clarify as well when I switch off the breaker to this circuit only the earth voltage disappears

so, the stray voltage must be coming from that MCB. dis all that circuit's cables from CU. do continuity and then IR tests on that circuit with all loads disconnected. bet you the IR tests will show a L-E fault on that circuit.

 

[spark 68]

It does look as though there are two faults, the initial one (earthing) not picked up, and whatever is leaking back to the CPC.

 

[chrisses]

Oh dear!, so instead of investigating this issue (above 50V touch ?), the installation was put into service with a possible inherent defect.

Yes oh dear I have dealt with the issue accordingly

 

[chrisses]

so, the stray voltage must be coming from that MCB. dis all that circuit's cables from CU. do continuity and then IR tests on that circuit with all loads disconnected. bet you the IR tests will show a L-E fault on that circuit.

Yeah I think your right but am I not right in thinking if that was the case then I would be able to disconnect the earths and only one would still be live?

 

[telectrix]

not if the fault is transmitting through a load to N and then onto the earth via the TNC-S.

 

[Octopus]

Oh and to clarify as well when I switch off the breaker to this circuit only the earth voltage disappears

In which case I'd be back with a copy of the EIC you did for the rewire and comparing the results with what you find now.

 

[Octopus]

so, the stray voltage must be coming from that MCB. dis all that circuit's cables from CU. do continuity and then IR tests on that circuit with all loads disconnected. bet you the IR tests will show a L-E fault on that circuit.

But wouldn't the rcd trip?

 

[spark 68]

I would be more concerned about how the Earth is rising above 50V, and why the ELI test could not be performed first and foremost, then try to find the other fault.

Something doesn't add up here, I would be inclined to start re-testing from the beginning.

 

[chrisses]

not if the fault is transmitting through a load to N and then onto the earth via the TNC-S.

I've arranged to go there tomorrow I'll let you know how I get on.

- - - Updated - - -

Thanks to everyone who has taken the time to give advice on this I will let you know how I get on

 

[Des 56]

You need to clear your head with the confused state you are at


Approach the job with a fresh mind as if it was the initial verification,do the appropriate tests and in the correct sequence
The fault or faults can and should become apparent

Personally I would find faffing around taking voltage readings a waste of time and effort and it will likely confuse ( as it has done with your investigating)

Ignore these voltage readings that you seem fixated on,clear the system using insulation.continuity and loop impedance tests
Get those done correctly with acceptable results and you will sort the problem yourself

 

[Marvo]

I would be more concerned about how the Earth is rising above 50V, and why the ELI test could not be performed first and foremost, then try to find the other fault.

Something doesn't add up here, I would be inclined to start re-testing from the beginning.

I agree, start from scratch and be methodical. Test the incoming supply to prove all the faults are on the premises, find out why the earthing is high impedance hance allowing a rise in touch voltage and find the IR fault(s). Finally test the RCD to make sure it's still functional.

 

[HandySparks]

TN what though ? TN-S or TNC-S...

I may have missed it, but have we got an answer to this?

 

[HandySparks]

I like a bit of fault-finding.

I plan to be in Guildford tomorrow. PM me if you like and I may be able to put a second pair of eyes on the job.

 

[Engineer54]

I then disconnect all the cpc's from each other at each switch (fed at switches not 3 plate) I still have voltage on the earth on the incoming and outgoing feed even though they are disconnected from each other

I wouldn't mind betting that at least one fault, is in one of the switch back boxes that's jammed packed full of cables!!

 

[Deleted member 72666]

Re: Voltage up to 100v on CPC even when disconnected

Originally Posted by spark 68
Oh dear!, so instead of investigating this issue (above 50V touch ?), the installation was put into service with a possible inherent defect.



Yes oh dear I have dealt with the issue accordingly




How, may I ask?​

 

[chrisses]

I may have missed it, but have we got an answer to this?

TNC-S sorry forgot to answer that

 

[HandySparks]

TNC-S sorry forgot to answer that

OK. I think it's worth checking the supply first. If that's not right, nothing else will make sense if you're live testing.

What's the resistance between incoming neutral tail to the CU main switch and the earth bar in the CU?

And worth checking polarity on the supply. Does a volt stick light up on the line, but not on the CU neutral or earth bars?

 

[chrisses]

I cracked it
the problem was actually quite simple in the end, the solution was a bit tougher though
i approached it as a new installation and tested the whole lot, turns out the guy who said he'd done r1 + r2 obviously hadn't
there was a break in the earth somewhere in the circuit leaving most of the lighting circuit unearthed!
with no earth the cpc was getting endured current from the live
couldn't find the break in the earth but managed to refeed the circuit from the board!
thanks to everyone for the helpful advice I'll try and return the favour in the future

 

[Octopus]

Trying to do a Zs on the circuit may have helped highlight the fault. Worth remembering for future.

 

[Des 56]

I cracked it

guy who said he'd done r1 + r2 obviously hadn't

These apparent complicated faults,they generally have a very basic cause
Well done,

 

[Engineer54]

I cracked it
the problem was actually quite simple in the end, the solution was a bit tougher though
i approached it as a new installation and tested the whole lot, turns out the guy who said he'd done r1 + r2 obviously hadn't
there was a break in the earth somewhere in the circuit leaving most of the lighting circuit unearthed!
with no earth the cpc was getting endured current from the live
couldn't find the break in the earth but managed to refeed the circuit from the board!
thanks to everyone for the helpful advice I'll try and return the favour in the future

Another valuable lesson learned, ... ''Never'' take as gospel what others tell you, Always check for yourself!! That is especially the case when in fault finding mode!! lol!!

 

[GT1]

Had this sort of induced current before with similar voltages.
What you have to factor in is not just the voltage but the current behind it too.
It may well be 100 odd volts or so but may only be milliamps 'worth..
Satisfy yourself that it's all compliant, that OCPDs and RCDs are working correctly then, let the earthing system do it's job and dispose of it.