Neutral Earth fault

[Fuenkli]

I have a Neutral Earth fault in my house I can not find. The fault was only identified when we moved all the lights onto the RCD protected UPS section off the DB (before they were wired without an RCD which I guess is the local standard to safe cost). I have narrowed down the section of the house where the problem is (some lights still on even with all neutrals disconnected). Because the wiring of this section of the house is almost inaccessible (very low roof space) we are about to give up. What would be the risk of leaving it as it is and just separate the UPS section of the DB board into an RCD protected section for the plugs and one without for the lights? 

[___]

12 minutes ago, Fuenkli said:

some lights still on even with all neutrals disconnected

Sounds like you have a leak from Neutral to earth. Those are the most pesky to find. This is a really long shot, but I discovered in my house that cheap surge arrestors can cause such leakage. Usually in the order of mere milliamps, so often not noticed until you have a few of them plugged in. The way I trace them is to unplug everything, that probably means unscrewing light fittings as well, and then check if the fault persists. If the fault goes away, the problem is in an appliance or a light, not in the wiring. That is good: It means you don't have to get into the roof.

[Jaco De Jongh]

19 minutes ago, Fuenkli said:

The fault was only identified when we moved all the lights onto the RCD protected UPS section off the DB (before they were wired without an RCD which I guess is the local standard to safe cost).

Please make sure you have moved the Neutral as well, moving the live without the neutral will cause a trip. 

Db normally has 2 neutral bars, one for equipment that is not connected to the Earth leakage, the other for equipment not on the earth leakage circuit. Its a very common mistake to move the live and forget about the Neutral. 

[Jaco De Jongh]

32 minutes ago, Fuenkli said:

one without for the lights? 

To answer this part of your question. The normal SANS standard allows lights, geysers and stove  not to be part of the EL circuit as long as they (As in all metal parts) are permanently connected to earth.

I don't see a risk if the circuit protection (MCB) is in place and of the correct size. Typically that will be a 10A MCB on a light circuit. 

[pilotfish]

4 minutes ago, plonkster said:

This is a really long shot, but I discovered in my house that cheap surge arrestors can cause such leakage.

He says all plug cct previously on RCD so this scenario, while common, is unlikely in this instance.

3 minutes ago, Jaco de Jongh said:

Please make sure you have moved the Neutral as well, moving the live without the neutral will cause a trip.

Another very common problem when migrating from off/on RCD, but he says ALL neutrals disconnected but lights remain on - if correct then this indicates to me that one section has a major N-E fault allowing the current to pass to earth keeping lights burning.

This is not a diy exercise unless you are very familiar with electrical work. You would need to open all light fittings that are burning while neutrals are disconnected, separate all the neutrals, and then test to see which neutral is down to earth and rectify it.

One common problem is mounting screws through neutral wires.

[Jaco De Jongh]

32 minutes ago, Fuenkli said:

I have narrowed down the section of the house where the problem is (some lights still on even with all neutrals disconnected)

Is this with all the neutrals disconnected? or only the light circuit neutrals disconnected?

In  adding unto a house, i have seen electricians  pull the neutral from lets say another plug circuit, or even in worse case from an Earth. if either one is true, the El will trip constantly. 

[pilotfish]

3 minutes ago, Jaco de Jongh said:

I don't see a risk if the circuit protection (MCB) is in place and of the correct size. Typically that will be a 10A MCB on a light circuit. 

One issue that you will encounter at some stage is that for a valid COC you need min. 1 Mohm insulation resistance which I doubt you currently have, so it will need to be rectified if you ever plan to sell the house.

Sort it out now is the best solution.

[___]

9 minutes ago, pilotfish said:

He says all plug cct previously on RCD so this scenario, while common, is unlikely in this instance. 

I know :-) The suspicion is that the fault is on a lights circuit. Could it be that there is a similar device on the lights circuit? If you have GU10 downlights for example, there might be such a thing in those.

11 minutes ago, pilotfish said:

major N-E fault allowing the current to pass to earth keeping lights burning

That is my diagnosis too.

If I am not mistaken, it is common practice for the neutral to be looped, ie at any light fitting you may find two neutral wires, one coming from the supply side and the other one going to the next fitting. By going from fitting to fitting and disconnecting your neutrals, you may get a better idea where the fault lies.

[Jaco De Jongh]

14 minutes ago, pilotfish said:

major N-E fault allowing the current to pass to earth keeping lights burning.

 

43 minutes ago, Fuenkli said:

The fault was only identified when we moved all the lights onto the RCD protected UPS section off the DB

 

43 minutes ago, Fuenkli said:

(some lights still on even with all neutrals disconnected)

Okay, now I am confused, If the live was moved to the EL circuit, the rest of the theory is impossible.  Especially with a N-E fault. The RCD wont allow the light to burn.  

[___]

3 minutes ago, Jaco de Jongh said:

The RCD wont allow the light to burn

I suspect he temporarily removed the RCD and the neutrals to debug it.

[Jaco De Jongh]

My last thoughts on this. I have to hit the road.... will catch up on this thread later.. 

The section with the lights burning is fed from a plug circuit or something in that lines and isnt part of of the circuits that was moved. Please switch off circuits in the old db to see if that lights are still burning. 

I guess the part with the fault was added. The electrician forgot to bring a Neutral from the DB and connected to another circuits Neutral or even to an Earth, saw it was tripping and took it off the Earth leakage circuit to hide his error. 

I wish I was there. I just love neutral faults. I really do, Most of my electrician friends phone me when they come across a neutral fault and then the bets start running on how long it will take me to solve the issue. 

[Fuenkli]

7 minutes ago, plonkster said:

I suspect he temporarily removed the RCD and the neutrals to debug it.

yes this is what we did. Also measured the resistance between neutral and ground (with the DB board completely isolated from neutral) and we have a total short. I agree we have to sort this out. Told the sparky to keep on looking. Will use some of your suggestions.

[___]

24 minutes ago, Jaco de Jongh said:

I just love neutral faults

You would love the one I'm dealing with. Time will still tell if I have it figured out.

The Victron Multigrid inverter tests its earth/neutral bonding by passing a few DC milliamps through that loop. Some RCDs really don't like this, and if you already have a small standing loss this is often the last straw. What makes this particular case so interesting is that it doesn't trip only the RCD on the input side, it trips the one on the output side too! I thought about this long and hard, it made no sense for the RCD on the output to trip. Then I realised why. It looks as if the test is done with the backfeed relay closed, in other words, the inverter already operates grid-parallel and the output is latched to the input, so the neutral is essentially common. It is now passing a small current down the neutral to see if it is properly bonded, but then the RCD on the input disconnects in protest (taking the bond with it). The current now has nowhere to go but out the output side, upsetting that RCD as well. So by avoiding a trip on the input side, the output side should remain unaffected. That is my theory...

[Jaco De Jongh]

2 hours ago, Fuenkli said:

Also measured the resistance between neutral and ground (with the DB board completely isolated from neutral) and we have a total short.

Sorry for asking all these questions, do you have a short between Earth and Supply Neutral? for that is normal, they should be connected as close as possible to your Db and in reality they should be at the same potential. . 

If you measure a dead short between any of the outgoing neutrals and earth its a sign of a problem. 

[Jaco De Jongh]

1 hour ago, plonkster said:

The Victron Multigrid inverter tests its earth/neutral bonding by passing a few DC milliamps through that loop.

Interesting one, I have to look at this process before i try to comment, 

1 hour ago, plonkster said:

Time will still tell if I have it figured out.

Please do report back, would love to know the reason.  

[Fuenkli]

3 minutes ago, Jaco de Jongh said:

Sorry for asking all these questions, do you have a short between Earth and Supply Neutral?

the short between neutral and earth is inside the house. We disconnected the house from the grid and there is still a dead short. We are presently checking every light connection of the lights which are on even without a neutral connection to the DB. Hopefully we will be able to narrow the fault down further.

[flamegrilled]

13 minutes ago, Jaco de Jongh said:

outgoing neutrals

Hi Jaco,

Could you not spot the issue with one of those Ellies red plugs with leds to indicate a lack of or a state of a circuit?Or even make up an extension with that plug in it to test the light circuits individually?

Edited September 7, 2018 by flamegrilled

[___]

There is something to be said for some regulations, eg I believe in some parts of Australia each circuit (at least plug circuits) gets its own RCBO (ie RCD with overload protection). Costs a bit more, but each circuit is protected on its own and a fault doesn't take down the whole house. No shared neutral bus.

[Jaco De Jongh]

9 minutes ago, flamegrilled said:

Could you not spot the issue with one of those Ellies red plugs with leds to indicate a lack of or a state of a circuit?

Nice idea, I did something similar with an EarthLeakage tester with crocodile clips before to check polarity . 

[Fuenkli]

We found  the short between neutral and ground. What a mission! We followed Plonkster's advice to go from light fitting to light fitting disconnecting the neutrals until we found this (see attached picture). The wire holding clamp put to much pressure on the neutral wire and connected it with the ground wire. I think this is a design problem of the fitting. It should not be possible to put so much pressure on the wires. I learned a lot about neutral earth fault finding. Thanks for all the input.