Bizarre variable IR problem at 250V

[pc1966]

Recently changed a CU for a friend and today when visiting one of the RFC's RCBOs tripped. It was due attention anyway as open ring and no time to solve it last week, but due to the trip I checked the IR again and found it was different. Last time around a stable 15M on 250V with stuff attached, this time mostly 200M but flashing down to 15M or so.

Clearly not good, so made the effort to get everything unplugged to properly test at 500V and hopefully decide of wiring fault or appliance. Odd thing was it showed much the same at 250V, but on L-E and N-E but not L-N, though when tested at 500V was sowing 200M (max my MFT shows). Of course I'm am cursing myself for not bringing my Megger that goes to 1G now...

Found the open N on one socket and it seems to be the end point of one leg, but after trying to divine the cable layout and splitting it to see which half is faulty, it is proving elusive

There are times where it seems to need two segments connected to show the effect, as either on its own is not showing it. I wondered if it is due to cable capacitance but trying last then with a new 100m reel of 2.5 T&E in parallel seems not to be there. But now the damn thing seems to have fixed itself

Ideally it would be a faulty socket (BG Nexus metal things recently fitted by my friend) but that seems unlikely. Sadly attempts to provoke it by testing at 1kV also show no problem, 200M again.

TL;DR anyone seen something like this before?

 

[westward10]

You over thinking things again pc.

 

[jay sean]

sometimes you need to disconnect all USB sockets, and you can IR L and N to earth on 500Meg

 

[plugsandsparks]

Might be stating the obvious but i have found IR testing a challenge when there is a fault caused by either damp / corrosion . I have had 0.0 readings and it still energised and have had 10M plus readings and it trips before i can put the tester away. You could try a clamp milliammeter across Ls and Ns at the board.
My worst case you may never find it

 

[SparkySy]

i had a very similar fault about 10 years ago, this was on a radial with two sockets on it only about 10m of cable in an old stone Cottage, did a board change and tested all ok, all RCBO's stayed in until later that evening when the circuit in question tripped.
Went back next day and tested the circuit again at 500v and 1k perfect 999meg reading. Put back into service and was just packing up when it tripped again. So kit back out swapped both sockets tested again and again 999meg. Reset RCBO and it tripped straight away!
Multiple tests on RCBO all ok, swapped to another one, still tripped!
Couldn't figure it out and couldn't access the circuit to find any damage so ended up putting a new cable in. Never did figure out what had happened to it!
Sy

 

[pc1966]

sometimes you need to disconnect all USB sockets, and you can IR L and N to earth on 500Meg

No USB sockets here (fo far...) and this time everything unplugged so I could test at higher voltages.

 

[pc1966]

Might be stating the obvious but i have found IR testing a challenge when there is a fault caused by either damp / corrosion . I have had 0.0 readings and it still energised and have had 10M plus readings and it trips before i can put the tester away. You could try a clamp milliammeter across Ls and Ns at the board.

I have almost put it back as the original RFC now, though it is still showing the odd IR problem when tested as a whole.

Now doing what you suggest with some short 2.5mm and wagos to temporarily "lollipop" the circuit so I can easily get my clamp ammeter on it, So far shows 0.5 to 1.1mA depending on how I wiggle the clamp meter, etc. Now on max-hold to see what I get (before it auto powers down )

My worst case you may never find it

Indeed!

 

[pc1966]

i had a very similar fault about 10 years ago, this was on a radial with two sockets on it only about 10m of cable in an old stone Cottage, did a board change and tested all ok, all RCBO's stayed in until later that evening when the circuit in question tripped.
Went back next day and tested the circuit again at 500v and 1k perfect 999meg reading. Put back into service and was just packing up when it tripped again. So kit back out swapped both sockets tested again and again 999meg. Reset RCBO and it tripped straight away!
Multiple tests on RCBO all ok, swapped to another one, still tripped!
Couldn't figure it out and couldn't access the circuit to find any damage so ended up putting a new cable in. Never did figure out what had happened to it!
Sy

This is my worry, as its an old house (~150 years) and most cable is routed in walls and under the 1st floor floorboards, so replacing most segments won't be easy or practical.

 

[pc1966]

So back today, same IR issue, but now r1 & rN good.

Testing to see if it can be provoked. Yesterday when it tripped they had a jointer in doing some work (always a suspicious coincidence) but he has not fixed new flooring down yet as plumber still to do first fix, so lifted it to look, nothing odd. Wiggled cables now energised, nothing tripped.

However, that circuit had a freezer on it, now on an extension lead to kitchen (not RCD'd as another CU change really needed...) and it occurred to me maybe it has a fault on some deicer that only run one a week or so? Again, should have brought PAT tester but might be worth doing a L+N to E IR test on it.

 

[pc1966]

Just to add, it is showing the flaky behaviour at 500V and 1kV, but not noticeably worse than at 250V.

I split the RFC in a few positions and mapped out the route using resistance of L, still to convert to distance via mOhm/m figure though, but they all added to around the overall r1 reading so good.

While broken up I IR tested each of the segments at 250V to 1kV and none of them showed anything odd, all reading MFT (DiLOG 9118) max of 200M. But together again and odd

 

[SparkyChick]

So, what I've mostly found in situations like this is a pinhole somewhere in the insulation where a cable's been pushed back into a box and scraped on a sharp edge or clipped by a screw etc.

You run the test for along period (on the Megger MFT1731 you can just hold the test button to keep the juice flowing) and what is usually observable in such situations is the reading climbs to a nice high acceptable reading but then out of the blue it drops down to a much lower value before climbing back up again.

I've observed the same behaviour as you've described on many occassions and as I say, it's usually been a pinhole in the insulation of a conductor. It's not manifesting during your investigations because you've probably pulled all the conductors forward out of the box and thus the cable can't generate enough charge to breakdown the insulation/air gap that now exists, but when you put it all back together it's being pushed back in and it can develop a sufficient charge to flash over (the momentary drop of IR reading).

Unfortunately, if you've not made any changes to the circuit or accessories and it's only been found due to the RCBO tripping, it could be a real pain to find it.

 

[plugsandsparks]

There is always the mice issue, watching you and putting his/her paw across the wire just when you are about to leave. They are not as daft as they look. When i get desperate, if i think i have narrowed it down, you can always run a replacement cable across the floor to see if it holds. Only had to do this once or twice a decade , sometimes you cannot let it go and it becomes personal........

 

[pc1966]

So, what I've mostly found in situations like this is a pinhole somewhere in the insulation where a cable's been pushed back into a box and scraped on a sharp edge or clipped by a screw etc.

Some form of damage is my thought, but what is odd here is both L and N seem equally impacted, for usual scrapes, etc, I would be surprised if they got equal "treatment".

You run the test for along period (on the Megger MFT1731 you can just hold the test button to keep the juice flowing) and what is usually observable in such situations is the reading climbs to a nice high acceptable reading but then out of the blue it drops down to a much lower value before climbing back up again.

I've observed the same behaviour as you've described on many occasions and as I say, it's usually been a pinhole in the insulation of a conductor. It's not manifesting during your investigations because you've probably pulled all the conductors forward out of the box and thus the cable can't generate enough charge to breakdown the insulation/air gap that now exists, but when you put it all back together it's being pushed back in and it can develop a sufficient charge to flash over (the momentary drop of IR reading).

I could see it happen with sockets off and forward. Also without any noticeable movement, I would see it if linking two L, but not one L at a time.

My suspicion is the fault would flash over and with enough cable it would establish a big enough arc to allow enough fallow-through to make the MFT notice.

But what is odd is I could not persuade any of the partitioned for test segments to show anything odd even at 1kV. Maybe had I brought my Megger I would have seen some odd changes in the 200M-1000M region, but if it comes back I will be prepared.

Unfortunately, if you've not made any changes to the circuit or accessories and it's only been found due to the RCBO tripping, it could be a real pain to find it.

Alas, I suspect you are right

 

[pc1966]

There is always the mice issue, watching you and putting his/her paw across the wire just when you are about to leave. They are not as daft as they look. When i get desperate, if i think i have narrowed it down, you can always run a

Mice are one option, as in they might have nibbles both L and N over a segment, but most circuits seemed OK so far and no other evidence of an infestation.

replacement cable across the floor to see if it holds.

If it keeps happening (odd trip) then I might try that with one segment at a time. Hoping it is not several faults...

Will get an extra RCBO so if that happens I can partition in to two radials as well allowing finer detection of dodgy segments.

Only had to do this once or twice a decade , sometimes you cannot let it go and it becomes personal........

Hopefully it will be OK and not haunt me to become my MobyDick.

I guess Captain Ahab could get cream for that these days...