Interpreting IR test

[timhoward]

I'm doing testing ahead of a board change. It's a cramped Wylex rewirable installed by Spiderman himself. (I still haven't worked out why there seem to be 4 lives and 5 neutrals for lighting circuits, but that mystery will have to wait until the point of no return)
I eventually got the RFC lives and neutrals untangled and removed for an IR test. The cpc's are still connected unfortunately. L+N together to E at 250v is giving 8Mohms and slowly rising (reached 10 and still rising when I let go), but at 500v it's giving 0.5 Mohms.
Other than looking harder for connected equipment that I've missed is there anything this instantly suggests please?
I was suspecting outside lights on a FCU or something but can't see any.
(Zs is ok for all circuits. RFC continuity end to end is good too. Current main concern is this IR test result)
TIA

 

[Strima]

IR should be done with the CPCs connected to earth.

I'd double check for connected equipment, things like switched mode power supplies can really mess your readings up, so can extension leads with surge protectors.

 

[loz2754]

Sounds like a connected USB socket, or other load that contains a switched mode power supply.

 

[pc1966]

That sort of thing is common for SPD in something like an extension block, etc, but usually more extreme than that.

It would bother me as well if I could not find a reason, as the point of 500V testing is to uncover weaknesses in cables!

 

[pc1966]

As above the CPC should be connected for this sort of test.

But if you see something dodgy, which you do, then try doing the test again but L+N to isolated CPCs. If much better then it really is a fault from cable to something else somewhere (i.e. leak to building via damaged sheath, not via cable's own CPC to connected equipment).

 

[loz2754]

Also a low testing ohm meter across line to neutral would indicate if there are connected loads.

 

[timhoward]

As above the CPC should be connected for this sort of test.

But if you see something dodgy, which you do, then try doing the test again but L+N to isolated CPCs. If much better then it really is a fault from cable to something else somewhere (i.e. leak to building via damaged sheath, not via cable's own CPC to connected equipment).

Cheers, I did a mind jump - I meant unfortunately I don't think I can get CPC's out as that was what my instinct said to do as well.

 

[timhoward]

Firstly what a great bunch you are - didn't expect replies within 10 mins! Thanks all.
I've just found a USB charging socket that I missed and a bathroom mirror that looked innocent but turns out to have lights and a bluetooth speaker in it with a FCU behind it.
I've had to pack up for the night now but will be back tomorrow.
Thanks again.

 

[timhoward]

Think I've found it. Single socket behind sofa with a surge protected extension lead in. Really should have noticed that but customer had assured me there were only two sockets in the lounge. She was quite excited to find this one!
Thanks again.