Technical question for IR testing

[ManchesterKirchoff]

So I’m an apprentice and I’m just trying to get a better understanding of the science behind this because the electricians who have said it have been quite vague and not properly explained.

Recently testing L-E for a circuit which was 0.02 ohms and the electrician said I will have to remove the neutral from the neutral bar of that circuit as it may be “getting a low reading via the neutral”. After doing so, it improved (not by much but beyond 1M ohm). Basically my first question is why is this the case?

The 2nd is it’s standard practice at my company when testing N-E to test between both neutral bar and earth bar, with no conductors removed as this will tell you overall the condition of the circuits and if the reading is bad you can start removing conductors to single out the faulty circuit.

However, some electricians at the company do NOT test between N-E because they say if you are clear between L-E it will be okay N-E. To me this does not seem right, as otherwise that test would not exist and it takes a matter of minutes to test and record readings. So my 2nd question would be is this a cop out to doing an extra test or a legitimate reasoning?

 

[Pretty Mouth]

Recently testing L-E for a circuit which was 0.02 ohms and the electrician said I will have to remove the neutral from the neutral bar of that circuit as it may be “getting a low reading via the neutral”. After doing so, it improved (not by much but beyond 1M ohm). Basically my first question is why is this the case?

It's because of installed equipment:

You are apparently IR testing L-E of one circuit.

However, if there is equipment in this circuit (for example a lightbulb), then the L has some sort of connection to N, which is then connected to the N bar, which is connected to the N's of all other attached circuits, which may well be connected to their respective L's via other equipment. Most, if not all, live conductors and equipment are connected in some way to the L you are testing.

So, as long as the N of the circuit you are testing is connected to the N bar, then you are probably testing between most, if not all of the installation, and earth, not just the circuit that you intended to test. This is likely to give a lower IR result, especially if there are faults on other circuits.

Also, if upstream double pole switcheswill have are closed (eg main switch, RCDs), then the N bar is connected the the incoming N, all the way back to the transformer, and the earth electrode, and the entire mass of planet earth. Now you will get a very low IR result.

 

[pc1966]

Recently testing L-E for a circuit which was 0.02 ohms and the electrician said I will have to remove the neutral from the neutral bar of that circuit as it may be “getting a low reading via the neutral”. After doing so, it improved (not by much but beyond 1M ohm). Basically my first question is why is this the case?

It would seem very unlikely to be 0.02 ohms, but 0.02 Mohm would fit.

If this is a TPN board then normally the neutral is not switched, so it remains linked to E somewhere (the very definition of neutral) so if you have any loads present L-N such as lights or even neon indicators, then if you IR test L-E it finds a path L-N and thus to E.

In a domestic single phase installation the incomer switch has to be DP so if it is off, the neutral will also be isolated from the supply. However, you can get cases of RCBO with a "functional earth" that means there is a path L/N-E through them.

The 2nd is it’s standard practice at my company when testing N-E to test between both neutral bar and earth bar, with no conductors removed as this will tell you overall the condition of the circuits and if the reading is bad you can start removing conductors to single out the faulty circuit.

However, some electricians at the company do NOT test between N-E because they say if you are clear between L-E it will be okay N-E. To me this does not seem right, as otherwise that test would not exist and it takes a matter of minutes to test and record readings. So my 2nd question would be is this a cop out to doing an extra test or a legitimate reasoning?

You have to check N-E somehow, otherwise you might have a fault. That might be unnoticed until something else goes wrong (like open cable) and potentially you then have a danger of L-N current diverting to E which is now open due to a 2nd fault. It will also trip RCD if fitted, but not always in a simple and obvious manner due to the low N-E voltage normally there.

So the usual first sanity check on any installation is to do a "global IR" where you isolate and link L+N together (all three L if needed) and IR test L+N -> E to show if you have any sort of unexpected path. If that is low, below say 0.5M then you have a problem.

It might be functional earths (which ought to be disconnected for testing normally), or it might be a lot of old/dirty/damp cables & fittings all OK in the ~1M region but paralleled they are showing a low value, but something is needing more investigation. If you get a decent IR that way, say over a few M then you know that every circuit has to be at least the global value, and so it is OK.

So good global IR means no more testing?

Maybe, what you don't achieve that way is to know that circuits are isolated from each other. That can be a danger such as the "borrowed neutral" sometimes seen in upstairs/downstairs lights and so on, or it could be cross-linked RFC which is a real danger (as you now have 64A of OCPD on 2.5mm cable), so if doing a proper check then seeing if circuits are isolated is a point that is now mandatory in the ROI for testing. The cross-linking RFC is easy to test, open all MCB/RCBO and check L-L on each RFC, they might show modest resistance if loads still present (and no neutral-switching RCBO used), but if cross-linked you will see only a couple of ohms or less between a pair of MCB that is BAD.

 

[pc1966]

Ah @Pretty Mouth beat me too it!

 

[ManchesterKirchoff]

It would seem very unlikely to be 0.02 ohms, but 0.02 Mohm would fit.

If this is a TPN board then normally the neutral is not switched, so it remains linked to E somewhere (the very definition of neutral) so if you have any loads present L-N such as lights or even neon indicators, then if you IR test L-E it finds a path L-N and thus to E.

In a domestic single phase installation the incomer switch has to be DP so if it is off, the neutral will also be isolated from the supply. However, you can get cases of RCBO with a "functional earth" that means there is a path L/N-E through them.

You have to check N-E somehow, otherwise you might have a fault. That might be unnoticed until something else goes wrong (like open cable) and potentially you then have a danger of L-N current diverting to E which is now open due to a 2nd fault. It will also trip RCD if fitted, but not always in a simple and obvious manner due to the low N-E voltage normally there.

So the usual first sanity check on any installation is to do a "global IR" where you isolate and link L+N together (all three L if needed) and IR test L+N -> E to show if you have any sort of unexpected path. If that is low, below say 0.5M then you have a problem.

It might be functional earths (which ought to be disconnected for testing normally), or it might be a lot of old/dirty/damp cables & fittings all OK in the ~1M region but paralleled they are showing a low value, but something is needing more investigation. If you get a decent IR that way, say over a few M then you know that every circuit has to be at least the global value, and so it is OK.

So good global IR means no more testing?

Maybe, what you don't achieve that way is to know that circuits are isolated from each other. That can be a danger such as the "borrowed neutral" sometimes seen in upstairs/downstairs lights and so on, or it could be cross-linked RFC which is a real danger (as you now have 64A of OCPD on 2.5mm cable), so if doing a proper check then seeing if circuits are isolated is a point that is now mandatory in the ROI for testing. The cross-linking is easy to test, open all MCB/RCBO and check L-L on each RFC, they might show modest resistance if loads still present and no neutral-switching RCBO used, but if cross-lined you will see only a couple of ohms or less between a pair of MCB that is BAD.

Yes I meant 0.02M ohms that’s my mistake.

It’s all domestic work I do so everything’s SP,

It's because of installed equipment:

You are apparently IR testing L-E of one circuit.

However, if there is equipment in this circuit (for example a lightbulb), then the L has some sort of connection to N, which is then connected to the N bar, which is connected to the N's of all other attached circuits, which may well be connected to their respective L's via other equipment. Most, if not all, live conductors and equipment are connected in some way to the L you are testing.

So, as long as the N of the circuit you are testing is connected to the N bar, then you are probably testing between most, if not all of the installation, and earth, not just the circuit that you intended to test. This is likely to give a lower IR result, especially if there are faults on other circuits.

Also, if upstream double pole switcheswill have are closed (eg main switch, RCDs), then the N bar is connected the the incoming N, all the way back to the transformer, and the earth electrode, and the entire mass of planet earth. Now you will get a very low IR result.

I appreciate your reply that makes a lot of sense! Thank you

 

[ManchesterKirchoff]

It would seem very unlikely to be 0.02 ohms, but 0.02 Mohm would fit.

If this is a TPN board then normally the neutral is not switched, so it remains linked to E somewhere (the very definition of neutral) so if you have any loads present L-N such as lights or even neon indicators, then if you IR test L-E it finds a path L-N and thus to E.

In a domestic single phase installation the incomer switch has to be DP so if it is off, the neutral will also be isolated from the supply. However, you can get cases of RCBO with a "functional earth" that means there is a path L/N-E through them.

You have to check N-E somehow, otherwise you might have a fault. That might be unnoticed until something else goes wrong (like open cable) and potentially you then have a danger of L-N current diverting to E which is now open due to a 2nd fault. It will also trip RCD if fitted, but not always in a simple and obvious manner due to the low N-E voltage normally there.

So the usual first sanity check on any installation is to do a "global IR" where you isolate and link L+N together (all three L if needed) and IR test L+N -> E to show if you have any sort of unexpected path. If that is low, below say 0.5M then you have a problem.

It might be functional earths (which ought to be disconnected for testing normally), or it might be a lot of old/dirty/damp cables & fittings all OK in the ~1M region but paralleled they are showing a low value, but something is needing more investigation. If you get a decent IR that way, say over a few M then you know that every circuit has to be at least the global value, and so it is OK.

So good global IR means no more testing?

Maybe, what you don't achieve that way is to know that circuits are isolated from each other. That can be a danger such as the "borrowed neutral" sometimes seen in upstairs/downstairs lights and so on, or it could be cross-linked RFC which is a real danger (as you now have 64A of OCPD on 2.5mm cable), so if doing a proper check then seeing if circuits are isolated is a point that is now mandatory in the ROI for testing. The cross-linking RFC is easy to test, open all MCB/RCBO and check L-L on each RFC, they might show modest resistance if loads still present (and no neutral-switching RCBO used), but if cross-linked you will see only a couple of ohms or less between a pair of MCB that is BAD.

It’s all domestic for me and it’s confirmed that testing N-E should really be done and not excluded from testing. If I was doing tests on a lighting circuit at the board what would be the red flags of a borrowed neutral or a ring inside of a ring scenario if I was testing a RFC?

In regard to a lighting circuit borrowed neutral, it’s been suggested to find the neutrals of both lighting circuits and IR between N-N to see if there’s a connection (if I get a dead short it would prove there is a connection between the two) but honestly I never see anyone do this.

 

[westward10]

Sometimes your work colleagues may understand it in their heads but cannot express this understanding to you so feel free to ask away on here.

 

[pc1966]

It’s all domestic for me and it’s confirmed that testing N-E should really be done and not excluded from testing.

Definitely. The global IR to begin with is a good check for any work, even just adding a circuit or something else, as it helps rule out some unexpected issues.

If I was doing tests on a lighting circuit at the board what would be the red flags of a borrowed neutral or a ring inside of a ring scenario if I was testing a RFC?

The linked RFC is easy to check with by checking L-L on low ohms between all possible pairs of RFC circuits (with MCBs off as all should be sort of isolated, good few ohms or more even if big loads present).

In regard to a lighting circuit borrowed neutral, it’s been suggested to find the neutrals of both lighting circuits and IR between N-N to see if there’s a connection (if I get a dead short it would prove there is a connection between the two) but honestly I never see anyone do this.

The borrowed neutral is harder to check as here you have a load such as a light between L1 and N2 so if someone isolated L2 they might get a shock when disconnecting N2 due to this cross-linking load.

It is trickier to check for at it might depend on the position of two-way switches, etc, as to whether the load is connected or not. I don't really do much domestic work, etc, but on occasions where I have, I make sure the lights are all on first, then check the disconnected circuit pairs for any cross-linking.

If fitting RCBO boards then you find out shortly after as they trip when hall lights are tested!

I know this sort of testing has been mandated in the ROI, but I am not sure if there was some specific reason such as a high profile death or similar that led to it, or just a push for good practice as the normal route. Some of our ROI members like @Risteard or @LastManOnline might know.

 

[Pretty Mouth]

In regard to a lighting circuit borrowed neutral, it’s been suggested to find the neutrals of both lighting circuits and IR between N-N to see if there’s a connection (if I get a dead short it would prove there is a connection between the two) but honestly I never see anyone do this.

Borrowed N for lighting circuits is where a light has the L from one circuit and the N from another. So the two circuits are interconnected via the light. It may not show up under a N-N IR test between two circuits.

Usually found in older properties (esp. pre-mid 70's), it was standard practice at one time to have the landing light taking its switched live from the downstairs lighting circuit via the 2-way switching, and the neutral from the upstairs lighting circuit. You can find it in other places too, and sometimes interconnected with socket circuits etc.

How I test for borrowed N's:

Before killing the power, I go round and switch on every light in the house, and all other equipment. I want everything in circuit. I then isolate at the main switch. I disconnect both L and N of one circuit from the board, and link them together. I then IR test at 250V between that circuit, and all other L's and N's. A very low IR (say <0.05 M ohms) probably indicates a borrowed N. Then I disconnect the L and N of another circuit, and repeat the test, and again until all circuits have been tested. This way you can figure out which circuits are interconnected with which.

Once I have identified a connection between two circuits, to find which light has the borrowed N, I open the light switches one by one, IR testing between the two circuits each time. Opening the switch for the dodgy light will cause the IR to rise much much higher.

 

[littlespark]

Remember you are testing the fixed wiring… not the loads on the end, not even a lamp… and you’re not resting the devices in the board either.

So all it is is a bunch of cables. You’re testing there are no shorts where there shouldn’t be, and continuity of cables where there should be.

Which means testing L to E, N to E and L to N…. Last one is where you have to make sure nothing is connected / plugged in.

 

[pc1966]

Which means testing L to E, N to E and L to N…. Last one is where you have to make sure nothing is connected / plugged in.

Ideally, and tested at 250V just in case!

 

[Pretty Mouth]

Remember you are testing the fixed wiring… not the loads on the end, not even a lamp… and you’re not resting the devices in the board either.

So all it is is a bunch of cables. You’re testing there are no shorts where there shouldn’t be, and continuity of cables where there should be.

Which means testing L to E, N to E and L to N…. Last one is where you have to make sure nothing is connected / plugged in.

I feel this comment perhaps give an incomplete picture.

Installed equipment should be tested to ensure that exposed conductive parts are connected to the system earth (typically an R2 test), and that there are no faults between live parts and earth (IR test L+N to earth). Without confirming these, a fault between L and an an un-earthed exposed conductive part could go undetected, leaving a C1 fault.

 

[littlespark]

I feel this comment perhaps give an incomplete picture.

Installed equipment should be tested to ensure that exposed conductive parts are connected to the system earth (typically an R2 test), and that there are no faults between live parts and earth (IR test L+N to earth). Without confirming these, a fault between L and an an un-earthed exposed conductive part could go undetected, leaving a C1 fault.

Yes, I did simplify things somewhat… and what I said was really aimed at initial testing, before equipment is connected.

I was also trying to convey the fact that testing N-E was as important as L-E, which was part of the OP query.