Can I confirm how to test for and how to find a borrowed neutral?

[HappyHippyDad]

During some testing I found a N-E IR fault on the kitchen lights of 0.1Mohms. I also found a N-E IR fault on the ring sockets, again 0.1Mohms.

This could have been a coincidence, but I disconnected both neutrals from the neutral bar (i.e the kitchen lights and sockets) and tested between them. They showed continuity (0.97 ohms). Just FYI, there was no continuity with these neutrals to the neutrals of the other circuits.

This is a dual RCD board. K lights and sockets on the same side.

Does the fact the neutrals showed continuity mean this is a borrowed neutral? I have not come across continuity between neutrals before. Is this the correct test for them? Does it actually mean they are connected together somewhere, or am I missing other possibilities?

Also, as the neutrals have been connected together somewhere in the house between these 2 circuits, how do I find this join? The house is full of shoddy DIY alterations. I would like to offer at least a guess as to how you find this join, but for the life of me I cannot think of a way.

 

[timhoward]

What possible reason could there be for an actual neutral connection between lights and sockets!

I think this one is highly likely to be two sources of supply to something.

I like the idea mentioned of using a wander lead and checking sockets neutral to lighting points, and finding lowest reading, then checking lighting neutral to sockets neutrals and finding lowest reading. In theory you are then have the two joined points.
Otherwise, breaking down the sockets circuit won't take long.
Good luck!

 

[Dave OCD]

With the two circuit CPCs separated [removed from the earth bar] do you still have the very low reading between the neutrals ?

 

[HappyHippyDad]

With the two circuit CPCs separated [removed from the earth bar] do you still have the very low reading between the neutrals ?

That's an interesting point Dave. I'm afraid the cpcs were still in the MET. I should have tried with them out. I got a minor ticking off from @westward10 some years ago for IR testing without the cpcs in the earth bar and now its second nature to leave them in. Of course, during fault finding it would be fine, even necessary to remove, so that you get more information.

That was a pretty long winded way of saying...I didn't do it.

 

[Lucien Nunes]

The test with the CPCs disconnected would be mainly to find out whether the N-N fault is actually a manifestation of two N-E shorts, one on each circuit, linking the neutrals together via the CPCs. However, we already know that there are no solid N-E shorts, only that 0.1MΩ that drew your attention in the first place. So while a test with disconnected CPCs might help trace the low IR, it won't shed any light on the N-N interconnection.

Returning to the original title and the following comment:

In your example. If the line had been taken from a switch on the lighting circuit and the neutral from a nearby socket, how would that show continuity between neutrals? I realise there will be a resistance figure showing up between L and N as not all loads will be disconnected, but are there any appliances that would be as low as 0.97ohms?

This observation is bang on the money. An N-N resistance of an ohm or two involving a lighting circuit is not going to be a borrowed neutral because no load will have such a low resistance, and a simple borrowing puts two loads in series between N-N as per @Pretty Mouth. The lowest load resistance one is going to encounter in a domestic environment is likely to be a large motor or transformer of a few ohms - saw bench, welder etc. - or a shower at half a dozen, but these are the exceptions.

An accidental interconnection of two lighting neutrals is quite easy to visualise in a system with assorted junction boxes but it's harder to see why a socket-outlet circuit would be involved if only because the meeting of 2.5mm² and 1.0 or 1.5 is likely to draw attention. Lights fed from a fused spur tangled up with the normal lighting circuit? Central heating controls? Extractor fan? Automation? Floorboard nail through two cables?

 

[timhoward]

Mad idea - as there are two N paths back, two resistances in parallel, would putting mA clamp meter on the socket neutral and turning each light on in turn result in the highest reading when the lowest resistance path is increasingly in use?
The low power consumption of modern LED bulbs might make this idea a non starter….
There also may not be a material difference between the ‘wrong route’ and the higher resistance path of the smaller cable that is the correct route.

 

[LukeD]

Mad idea - as there are two N paths back, two resistances in parallel, would putting mA clamp meter on the socket neutral and turning each light on in turn result in the highest reading when the lowest resistance path is increasingly in use?
The low power consumption of modern LED bulbs might make this idea a non starter….
There also may not be a material difference between the ‘wrong route’ and the higher resistance path of the smaller cable that is the correct route.

They are good points. I hate "Borrowed neutrals " so much !

 

[Lucien Nunes]

In theory any test that can be done using load can also be done dead using resistance, and as a dead test that ought to be preferred, however the load test might involve less disturbance of the fittings.

As the lighting circuit is a radial, as the load is moved away from the CU towards the spurious interconnection, the lighting circuit neutral at the CU will receive a lower fraction of the load and the socket circuit neutral an increasing fraction. Once the interconnection point has been passed, the ratio will remain constant. It would be easier to apply a heavy load at a socket, but the ring configuration will complicate interpretation of the readings obtained.

As a reminder for anyone who hasn't read the whole story so far, the fault in the thread is seemingly not a borrowed neutral but a direct interconnection or fault between the neutrals of two circuits. A borrowed neutral is specifically the situation where a load is connected between the line of one circuit and the neutral of another.

 

[HappyHippyDad]

I've had a think about the fault finding aspect. I just think I'm going to get in to a pickle trying to use a wander lead every where and relying on very small decreases/increases in resistance through a long (very unreliable) wander lead.

I like @davesparks method. I'e what i always believe to be the standard way of finding an IR fault.

I remove all the suspect (x3) neutrals from the bar in the CU. I split the ring somewhere. Test for continuity between socket neutral and lighting neutral in the CU. Only one of the socket neutrals should now show continuity. Liven up the side that doesn't show continuity and every thing that works is removed from the suspect list. However, if the fault/join is in a JB then that means you can't rule out one of the cables in the actual socket you have split the ring in, so I always keep a record of which sockets I have made the split at, until of course they have been ruled out by making a split elsewhere.

I think this is either going to be an easy find on the kitchen lights as they all look dodgy, or.... it's going to be horrible!

The 80 year old, former senior civil engineer client continuously likes to get involved and believes he has a grasp of electrics. He doesn't. Whenever I try and explain something to him (because he has asked), he says 'yes, yes I know that, I was a civil engineer you know'. He then explains it to his wife (wrongly). I absolutely cringed when she asked him a question and he replied 'you don't need to worry your pretty little head about it'. I think he's still living in the 50's. He didn't say it in a mean way, it appeared to be just standard speech between him and his wife. Scary.

 

[timhoward]

I remove all the suspect (x3) neutrals from the bar in the CU. I split the ring somewhere. Test for continuity between socket neutral and lighting neutral in the CU. Only one of the socket neutrals should now show continuity. Liven up the side that doesn't show continuity and every thing that works is removed from the suspect list.

having split the ring I'd do a basic wander lead continuity test between the lighting N and N at every socket. (plug in lead)
No continuity = ruled out.
Continuity = take socket apart and try both N's in that socket. Rule more out. Rinse and repeat.

I think that's the fastest way.
Walk around with an electric screwdriver, you'll be home for afternoon tea! Best of luck....

 

[Lucien Nunes]

I like @davesparks method. I'e what i always believe to be the standard way of finding an IR fault.

Nothing wrong with that, but I'd take advantage of the known continuity of the ring to save opening and disconnecting stuff. I would inject a current around it from a small ELV benchtop power supply that can operate in constant-current mode into a short-circuit.

First I would go to the CU and hook up the PSU to circulate say 5A around the ring neutral from leg A to leg B. I would tie the ring line to neutral at leg A so that the voltage along the neutral can be read at every socket without a wander lead, just by plugging in a voltmeter with a 13A plug adaptor. Each metre of run between CU leg A and the socket would add 38-40mV to the reading.

First I would measure the voltage between leg A neutral and the lighting neutral at the CU, test (A) in the sketch, to find out the voltage present on the ring neutral at the point of the interconnection. Then I would walk around the ring doing test (B) with the plug adaptor, looking for a socket showing the nearest voltage to test (A), which would be at the point of interconnection. Or, if it's midway between two sockets, those with the nearest voltages above and below test (A).

At this point you will know the location of the problem on the ring without having to access any lighting points. If it's not then obvious which lighting point is involved, it would also be possible to do the same test using VD along the lighting circuit, but it would need shorting L-N at the far end or a wander lead running out to it, to serve in the same current injection role as leg B of the ring.

This probably sounds a bit long-winded but I would say with a dozen points on each circuit, so long as I could get to the sockets and lampholders / roses, I would guarantee to have found the nearest point to the interconnection within 20 minutes, without undoing a single screw other than to open the CU. Although the PSU would be 'energising' the circuit, it would only be with a volt or two, so none of the tests would be hazardous. It would not matter if a few small hidden loads remained connected to the circuits.

In another thread I will tell how I used this method to trace a fault in a completely inaccessible circuit with 20 points, on location on the beach in Amalfi, using a brake light bulb and an alarm battery as my constant-current source.

 

[ipf]

An accidental interconnection of two lighting neutrals is quite easy to visualise in a system with assorted junction boxes but it's harder to see why a socket-outlet circuit would be involved if only because the meeting of 2.5mm² and 1.0 or 1.5 is likely to draw attention. Lights fed from a fused spur tangled up with the normal lighting circuit? Central heating controls? Extractor fan? Automation? Floorboard nail through two cables?

Anything is possible with some 'kitchen engineers'.

 

[westward10]

Maybe they both have neutral/earth IR faults hence the apparent connection between them. Have you done a continuity test between neutral/earth.

 

[mainline]

Maybe they both have neutral/earth IR faults hence the apparent connection between them. Have you done a continuity test between neutral/earth.

The test with the CPCs disconnected would be mainly to find out whether the N-N fault is actually a manifestation of two N-E shorts, one on each circuit, linking the neutrals together via the CPCs. However, we already know that there are no solid N-E shorts, only that 0.1MΩ that drew your attention in the first place. So while a test with disconnected CPCs might help trace the low IR, it won't shed any light on the N-N interconnection.

Can I confirm how to test for and how to find a borrowed neutral? | Page 2 | on ElectriciansForums

Find Can I confirm how to test for and how to find a borrowed neutral? Advice and Help. How-to Can I confirm how to test for and how to find a borrowed neutral? in the UK Electrical Forum advice boards on ElectriciansForums.net Est.2006 | Free Electrical Advice Forum. Can I confirm how to...

www.electriciansforums.net

 

[Lucien Nunes]

Oops I think that's a quote from my post #19!

 

[HappyHippyDad]

Maybe they both have neutral/earth IR faults hence the apparent connection between them. Have you done a continuity test between neutral/earth.

Yes, 0.1Mohms (with CPC's still in the bar)