Diagnosing low insulation resistance (lighting circuits)

[McKenna32]

Hello all,

Before cracking into my lighting circuits (tacking a few things on up & downstairs) I've been doing some checks beforehand like a good boy scout. First up was insulation and I've come a cropper

Earthing is TN-C-S and the meter is a Fluke 1587FC. Measuring per instructions in OSG & GN3 with only the breaker for the circuit in question closed - all other switches open (including all the 2way switches in the rooms)

DC@250V (unless in brackets, in which case it's just on normal ohm range for info)

Downstairs
---L/N = 0 Mohm (680kohm)
---L/cpc = >275 Mohm
---N/cpc = 0.4 Mohm

Upstairs
---L/N = 0 Mohm (1.2Mohm)
---L/cpc = 0.4 Mohm
---N/cpc = 0.4 Mohm

I'm hypothesizing that a cable got crushed upstairs and a borrowed neutral too (stairway up/down switches wiring is a bit of a joke), but I have not much experience diagnosing this sort of thing. Borrowed Neutral plausible for sure but the crushing idea seems pretty infeasible, squished *just enough* to destroy the isolation but not enough to keep blowing the RCD? I don't think I'm right...

Does this smack of anything in particular to you knowledgeable folks out there?

There *are* dimmers upstairs, but none downstairs...(although god knows one might've been plastered into a wall in the past knowing this place...)

Hmm I guess next step is tie the L and N and test to cpc?

I could get hold of a capacitance meter I guess to check for power electronics on the bus but not until Monday and I was really hoping to get my new stuff hooked up...

Any ideas welcome! Thanks in advance

Al

 

[jelliottelectrical]

I would be looking for dimmers, neons. Have you got any fluorescent fittings, transformers

More simply, have you left the lamps in the fittings?

 

[EFLImpudence]

Did you remove all the bulbs and other similar things?
Are there other items on the circuits like fans, boiler etc.

Did you remove the Neutrals from the Neutral bar and the CPCs from the Earth bar?
If not, all the Neutrals are still connected and all the CPCs are still connected.

 

[JohnW2]

First up was insulation .... Measuring per instructions in OSG & GN3 with only the breaker for the circuit in question closed ....

I'm a bit confused here. You presumably did not attempt to measure insulation resistance with the circuit ('L') still connected to a 'closed' (i.e. 'on') breaker. Was the circuit disconnected from the breaker (in which case it wouldn't matter whether it was open or closed!)? ... and, as has been asked, what did you do with the neutral and CPC of that circuit?

Kind Regards, John

 

[flameport]

Measuring per instructions in OSG & GN3 with only the breaker for the circuit in question closed - all other switches open (including all the 2way switches in the rooms)

What exactly have you done when doing the tests - that description is not clear.
There are several methods.
GN3 and OSG contain similar, but not identical information and diagrams.

The results you have most likely indicate an error with testing procedures, rather than faults with the circuit(s).

 

[333rocky333]

I'm a bit confused here. You presumably did not attempt to measure insulation resistance with the circuit ('L') still connected to a 'closed' (i.e. 'on') breaker. Was the circuit disconnected from the breaker (in which case it wouldn't matter whether it was open or closed!)? ... and, as has been asked, what did you do with the neutral and CPC of that circuit?

Kind Regards, John

I noticed that
Wild quess, main switch off and testing from the bus bar side of the main switch

 

[Risteard]

Did you remove all the bulbs and other similar things?
Are there other items on the circuits like fans, boiler etc.

Did you remove the Neutrals from the Neutral bar and the CPCs from the Earth bar?
If not, all the Neutrals are still connected and all the CPCs are still connected.

It is non-compliant with BS7671 to disconnect cpcs from the means of earthing during insulation testing.

 

[EFLImpudence]

That's what it says - but I would do both if the results are at first unsatisfactory.

 

[McKenna32]

Thanks for the quick replies, good questions.

More simply, have you left the lamps in the fittings?

Yes, but all 2W switches open so no volts across...aha...OK so N and cpc still hooked up to the devices yes. Should only affect N/cpc result though? no?

Are there other items on the circuits like fans, boiler etc.

Did you remove the Neutrals from the Neutral bar and the CPCs from the Earth bar?
If not, all the Neutrals are still connected and all the CPCs are still connected.

yes 1 fan on downstairs, on a timer come to think of it...

nooo...I did not remove those things, ok so N/cpc result is garbage (tested the whole install there, still low though, mind, but there's all sorts plugged into the ring ATM) and I need to fish the N conductor out of the (cobwebs first then) bar

I'm a bit confused here. You presumably did not attempt to measure insulation resistance with the circuit ('L') still connected to a 'closed' (i.e. 'on') breaker.

I did indeed - see below

I noticed that
Wild quess, main switch off and testing from the bus bar side of the main switch

Precisely, only one breaker closed, shouldn't matter which side of the breaker I go and I just had the tester hooked up on the bar so I could flip circuits in/out easily

The results you have most likely indicate an error with testing procedures, rather than faults with the circuit(s).

Agreed

It is non-compliant with BS7671 to disconnect cpcs from the means of earthing during insulation testing.

Sounds like that makes sense. I suppose you'd only feel the need to remove cpc from the bar if you had a strong suspicion a live conductor was touching an earth of a device not fed from the circuit in quesion...?

So not removing the N from the bar is by far the most embarrassing of the things I didn't do It's an ugly nest of vipers to root through but will sort that and report back...

Thanks again

 

[EFLImpudence]

You should really have the light switches closed - and for two-way switching, a test in each configuration.

 

[JohnW2]

I did indeed - see below ... Precisely, only one breaker closed, shouldn't matter which side of the breaker I go and I just had the tester hooked up on the bar so I could flip circuits in/out easily

Fair enough (assumimg the main switch was open ), but ...

So not removing the N from the bar is by far the most embarrassing of the things I didn't do

OK, but that would not explain the upstairs L/cpc result, would it?

Kind Regards, John

 

[RF Lighting]

Anything in any circuit with say a L>E suppression capacitor for example, and an effective short between L>N such as a lamp filament or a fan winding on both the upstairs lighting and another circuit with the neutrals commoned together will read as a low IR reading on the upstairs lights as displayed in the OP

 

[McKenna32]

You should really have the light switches closed - and for two-way switching, a test in each configuration.

Yes, I felt like I was sneaking up on any issues by testing first with all light switches open, trying to establish if everything CU side of them was OK first... but do intend to test my new bits with switches closed (will get EICR for whole place when done). Good pointer on the 2w switching, and I was using the term wrong

Fair enough (assumimg the main switch was open ), but ...

It was open yes I did that bit right at least!

OK, but that would not explain the upstairs L/cpc result, would it?
Kind Regards, John

No I see your point. I suppose prime suspect would be something hooked up that's not able to be switched out when I'm testing (fan on timer, etc., as pointed out^)

 

[JohnW2]

No I see your point. I suppose prime suspect would be something hooked up that's not able to be switched out when I'm testing (fan on timer, etc., as pointed out^)

I forgot to ask - do you have any RCDs and, if so, were they open or closed?

On reflection, given that your installation is TN-C-S, the neutral bars surely should be effectively connected to the CPCs (via an RCD, if present and closed), so, if (as you say was the case) you left the neutrals connected, one would have expected near-zero (not 0.4 MΩ) N-cpc readings, wouldn't one?

However, that could explain the low upstairs L-cpc reading, if you had left some L-N load connected in that circuit - since, as above, a low L-N resistance should also be a low L-cpc one (if N is effectively connected to cpcs.

Kind Regards, John