Light circuit random tripping mystery

[brickers911]

So, I had this particularly tricky fault to find recently at a friends 4 story townhouse which was last refurbished in the mid 90's. To make matters more complex the 2 lighting MCB's were wired up with 1 supplying the ground floor only and the other supplying the basement, 1st and 2nd floors (3 floors). A quick test on the MFT gave me a low resistance on one of the legs (0.7 Meg) and no continuity on the ring? After lots of interrogation of the owner looking for clues as to a possible location for the fault I was no nearer a solution. There had been loads of metallic light fittings fitted and very difficult to remove panels to access some, most of which had been fitted by DIYer's (eg not earthed) so I started looking at a few of these, no luck. All this time no tripping despite liberal activation of all the switches on affected floors, then a bit later a trip occured whilst I was doing nothing and over the course of more searching (I was there about 6 hrs) more random trips, every time the MCB reset with no issues, I was starting to get a bit desparate when as I switched on the 1st floor bathroom light the MCB actually tripped, checked the light fitting over the shower which was ok however the supply to the extract fan was the culprit with poorly terminated connections in a high moisture environment. See photo.
My question is to all you bright sparks out there is how would you have tested the circuitry to track the fault down faster?

 

[Fleeting]

That sort of insulation resistance is not likely to trip an mcb it won't develop enough fault current.

 

[ProDave]

That looks like a pinched cable so the damaged conductors flash over. Last one I investigated only got found when we worked out it tripped each time you stood on a particular floorboard.

 

[Andy™]

sometimes you can hear the bang

 

[binky]

there is no quick way of finding these sorts of faults. The only thing you can do is start in the middle of a circuit and split the feeds, then test insulation. Spli affected side again, and continue. Trouble is with so many drivers in light fittings it does fuddle test results.

 

[Murdoch]

Lighting circuits are NOT rings.

I draw a "layout" of the property, confirm which accessories are on the circuit, then go from there, making notes as I go.

Sometimes its the very last thing you look at, sometimes the first.

 

[poni]

the fault is always in the last place you look , as you have found the fault and so stop looking

 

[Andy™]

the fault is always in the last place you look , as you have found the fault and so stop looking

not always. there are some that will keep looking...

 

[brickers911]

That looks like a pinched cable so the damaged conductors flash over. Last one I investigated only got found when we worked out it tripped each time you stood on a particular floorboard.

I was trying even in the bathroom i couldnt hear it, too many years in a power station, and Black Sabbath/Saxon concerts

 

[brickers911]

That sort of insulation resistance is not likely to trip an mcb it won't develop enough fault current.

True, but it wasn't tripping until it did

 

[ProDave]

Sometimes an insulation test at 1000V will provoke it to flash over when it is not always flashing over at 240V

 

[mort2376]

As already said. Test at board, estimate halfway point, split there and retest. Repeat halving faulty side each time.