lighting intermittent fault. mcb trips fault

[tudge56]

Hi guys, wondering if anyone's got suggestions for an intermittent fault.

Ground floor lighting circuit, Aprox 10 lights, mostly basic pendants occasionally trip 6A MCB.

Customer says it happens once every 2-3 weeks, no noticeable link to bad weather or time of day etc.
Also reports it happening when turning on different lights, but not always instantly and once they think It tripped with nothing switched on at all. So I assume a supply fault rather than a switch line fault?

MCB also protected by new RCD (tested and working fine) which has never tripped....so fault must be live to neutral....am I right?

When I arrive to test, neutral to CPC insulation resistance is around 2Mohms (no lights switched on) however within about 2 mins of being disconnected, rises back up to 150M ohm. How can the fault be between neutral and earth without tripping RCD?? Most of the lamps the customer has in are evil incandescent filament lamps- could there be equal leakage from live and neutral and this isn't detected by the RCD??

Any advice on how I can proceed with trying to separate the circuit to track down the culprit? I only seem to have a few minutes from disconnecting before the readings rise back up and the fault disappears.

Many thanks.

 

[Pete999]

Whats the totol load with all the lights on?

 

[telectrix]

might sound arse about face, logically, but have you tried replacing the MCB? could be tired if there have been several lamp blows over the years. the fault current can sometimes be quite large.

 

[Davisonp]

I had one last week where a lighting circuit with 30 40W BC bulbs (old bulbs as they almost never blow) on a theatre style dimmer that was tripping every time this dimmer was turned up. It seemed to be an arcing problem but no obvious wiring fault, so I ended up taking every bulb out and re inserting them to ensure they all have a good connection, and three bulbs had the glass come away from the base as I took them out. It turns out that one bulb was arcing, but never burning out. Possibly air getting into a bulb. Replaced the three bulbs and all ok since.

This would not explain the IR reading or tripping without any lights on but worth considering.

 

[bigspark17]

Hi tudge, a couple of questions not ment as diggs just for more info on the installation and what youve tested.?
1- why are you IR testing with the switches off, unless all switches are in closed position you are not IR'ing the whole circuit,
2- what are youre other IR results? L-Earth, ( you should not disconnect the cpc when IR testing to earth)
3-What was the ze at cu and zs at end of line.?
These faults are a pain being so intermittent but persistance wins at the end of the day, and the fact you have sorted a fault possibly other "sparks" have failed too should give you an insentive to digg as deep as you can to find it. Good luck

 

[Wilko]

Hi - if I'm understanding correctly, the IR from N-CPC starts at 2meg and climbs to 150meg over 2mins? I recall the testers operate by placing 500V (or ...) on the output and measuring any current flow, so sounds like it's charging up a capacitor (within appliance? Smoke detector?). At 2meg the current flow would be too low to trigger the RCD in normal operation. Anyway, just some random thoughts after a few so it might be rubbish. Spot says "Good Hunting"

 

[tudge56]

Thanks guys, lots of helpful stuff.

In answer to your questions:
total load is aprox 700W (4x 60W incandescents, 2x 100W, 3x 50W 12V, 3x 11W CFL) so well below 6A supposedly.

The 12v light is an ancient thing with a humongous copper transformer supplying 3 halogens - could this be a contributor? I didn't think so as the tripping was happening without it being switched on.

I definitely had my suspicions about the incandescents running high without actually blowing - anyone know if this is possible? Still doesn't explain why MCB is tripping with nothing switched on (although, let's be honest, I take information from customers always with a pinch of salt - if I believed everything I was told, RCDs could rest themselves and lights function with no lamps in them )
I have replaced the incandescents with CFLs to see if that makes any difference but I'm not optimistic.

I did also replace the MCB, but again, I'm not holding my breath - it was a CU that I upgraded in the spring so everything brand new (and of course the tripping only started with the new board....always a head ache trying to explain to customers that the good old rewireables which worked perfectly are not actually better than these new nuisance ones!)

And I wasn't very clear with my test results - I did indeed test with all switches closed and CPC connected to earth (all readings in Mohms)
All switches OPEN, circuit freshly turned off:
L-N 299
L-CPC 150
N-CPC 2
All switches CLOSED, lamps in place, circuit freshly turned off:
L-N dead short obviously
L-CPC 2
N-CPC 2
All switches CLOSED, lamps taken out:
L-N 299
L-CPC 150
N-CPC 50
However this last one is tricky to judge as by the time I'd gone round and taken all the lamps out, I don't know if that higher reading was due to the lamps being disconnected or just due to the time passing after disconnecting!!
I replaced all the incandescent lamps with CFL, energised the circuit with all switches closed for a couple of minutes, then tested and got a reading of 60Mohms N-CPC so it hadn't dropped as dramatically....but presumably this could be explained because the load was much lower and therefore cables running cooler??

Finally Ze was 0.25, nice and normal. I didn't test Zs I'm afraid, just calculated it from R1+R2 which was 1.44. Is it worth measuring it directly do you think?

Thanks for all the help!

 

[Midwest]

Have you tried testing without the transformer connected? Have you done a global IR tests, to see if other circuits have got a fault?

 

[westward10]

An IR of 2 Meg ain't causing that breaker to trip. When you say it rises to 150 Meg is that with you leaving the 500V test applied to it. Got to agree with Tel and try the obvious, change the breaker.

 

[tudge56]

Yep, tests without the transformer connected remain the same.

I didn't do a test of all circuits at the same time, but looking back at my certs for upgrading the CU, all circuits were pretty healthy - just a kitchen socket circuit giving a N-CPC reading of 35Mohms. But a fault on a different circuit wouldn't make an MCB trip on my lights would it?

And Westard, I agree the reading is still too high to make a MCB trip, but cos it's an intermittent fault and typically all behaving fine when I appear, I'm rather clutching at straws....might not even be that reading that's causing the tripping! I'll just have to wait for the customer to call me back...I did tell him I might have to move in and become his lodger

Thanks as always for all the speedy replies and help

 

[Midwest]

Have you got two way lighting on the circuit?