RCD trip with a difference!

[pushrod]

Was at a relative's house (big and old) and they said they sometimes flicked a particular light switch and it caused the CU to trip. Said i would take a quick look!

It turns out it was a split load board about 6 years old with one rcd. All the light circuits were on the non rcd side and when it was tripping it was taking out the rcd and leaving the lights on!. Taking out a 6 amp mcb on the non rcd side isolated the light and there was nothing loose at the light fitting or the switch. They said it used to happen very regularly but was less frequent recently. Needless to say we couldn't get it to happen when i was there. The switch in question was a smiths time guard, that comes on at night, but other than that it appeared perfectly normal. Didn't have time for any tests.

Must admit i was pretty much at a loss why a circuit on one side should be taking out the rcd on the other side! Am i missing something obvious? Anyone any suggestions? cheers

 

[nickblake]

it could be a faulty RCD had the same sort of thing happen at my sister in laws she would switch on lights round her house and it would trip the RCD and her lights were on a different board i dont know why it happend i just did the installation was the same age too 6 years tested every thing and was all ok put in a new RCD and problem solved hasnt tripped since

 

[telectrix]

could that one light have a borrowed neutral. bit of a long shot

 

[rocker]

Yep I'm seconding this - the light will have a borrowed neutral off a socket circuit.

Could also be an insulation problem between 2 seperate circuits, I spose.

EDIT: Actually on re reading your post, probably not a borrowed neutral if it's only intermittent.

 

[telectrix]

just re-read OP and seen intermittent, so agree with rocker. i would check to see if wiring was notched in joists (with or without protection) and look for intermittent short between circuits.

 

[nickblake]

you could get a leakage clamp tester and see what the leakage is across the earths , i got one worth its weight in gold

 

[rocker]

Nick how do they work? Like what do you do, just clamp all the earths in that clamp bit???

 

[nickblake]

its taken me a while to get used to it but what you do is clamp it round the earth of that circuit and it gives you a reading in MA just like the current clamp meters but using the earth not the phase , so what i do is check each circuits earth and add them up i mean the majority of the times the readings are 0.02ma so you know there is very little leakage one place i did find it very handy was a schools it class the leakage there was 75ma its a cracking piece of kit for fault finding and also if you are doing a CDU upgrade or converting to RCD protection then you can calculate the circuit splits the only thing i wont do is give you the leakage on the main earth to the board the modle and make is a TIS TIS550 be great for the fault here as you could check the leakageTIS 550 Mini Leakage Clamp Meter

 

[ringer]

Why won't it give the leakage on the main earth to the board, Nick?
I got the Di-log DL6507 for a smidge over a ton, but not had the opportunity to put it to work on a real problem yet.

 

[WayneL]

Hi mate,

IR test N - E on RCD side of board - my money is on a neutral/earth fault on RCD side

 

[telectrix]

Hi mate,

IR test N - E on RCD side of board - my money is on a neutral/earth fault on RCD side

so why does the RCD trip only when light on non-RCD side is switched on?

 

[ringer]

I always thought a N-E fault would involve current flowing from L (via load) to N to E, but is this possible?....
A N-E fault exists on a RCD circuit, on another (non RCD) circuit there is major earth leakage. This earth leakage current feeds from the earth into the (RCD protected) neutral via the N-E fault, then the RCD trips?

 

[WayneL]

I always thought a N-E fault would involve current flowing from L (via load) to N to E, but is this possible?....
A N-E fault exists on a RCD circuit, on another (non RCD) circuit there is major earth leakage. This earth leakage current feeds from the earth into the (RCD protected) neutral via the N-E fault, then the RCD trips?

This is correct - and the RCD only notices when circuit is under load.
Also, the bigger the load, the more the leakage increases until - trip - which makes it an intermittent fault.

Now, think about this:

Follow the current.

You have a neutral/earth fault on the RCD side - say RFC.

You switch lights on (non-RCD side)

Current flows Line - Load - Neutral - Neutral bar............now it should then go through the main switch and down the supply neutral, to the transformer.

BUT..... there's another path - via the neutral feeding the RCD, up the neutral of the faulty RFC, and down to earth.

It will still register as an imbalance on the RCD and trip.....and it only needs 30mA to travel this path.

Just an idea......I might be wrong

 

[ringer]

I think you might be wrong I think I might be wrong too
If there is a N-E fault on the RCD circuit, then the RCD will trip in normal use well before any extraneous (to this circuit) current is able to trip the RCD.

 

[kung]

long shot but when the light is on try walking round the house every where ! without seeing how its wired could be loose floorboards nail intermittently bridging l-e on a cable when floorboard is walked on or even nipping one cable say line on poss a lighting circuit and through another cable say cpc of rfc that would trip rcd and leave lighting on etc or damp etc. a long shot i know but doesnt sound like an obvious fault and if thats whats causing it might not always show on an ir or continuity test.