Single core and earth lighting fault

[Richard Measures]

Could the better and wiser assist with an odd one...?

I've just finished a big CU swap on a 1950's house that uses single core & earth lighting circuits. All but one of these are fine, but the faulty one is pretty funky;

There are 3 accessories on this circuit in 3 separate rooms. The first room/switch/pendant works fine. The later 2 have the same issue - with no lamp in the fittings all is well, but put a lamp into either with the switch on and 3 things happen:
- The bulb glows very briefly (30mS?)
- The RCD covering this half of the CU trips
- The 6amp type B MCB covering the circuit trips

If this was just an RCD issue I'd be confident of a neutral/borrowed neutral issue, but how can a bulb (which doesn't blow/fail during this) cause an MCB to trip?

It's got me going round in circles so any clarity and advice is most welcome.

Thanks

 

[Charlie_]

Single lives looping through the switches and neutrals looping through the pendants?

 

[Richard Measures]

Yes Charlie, live looped through switches (2-way in the second room) and neutrals/earths looped via pendants - bonkers system IMHO

 

[westward10]

Assuming the protective device is not faulty only one thing will make a circuit breaker trip, excess current either through overcurrent or fault current.

 

[Dustydazzler]

Bad batch of bulbs ???

 

[Richard Measures]

Its always possible I suppose but I've tried 3, 2 of which are CFs

 

[Lucien Nunes]

Ah good, a proper problem at last. Let's see if we've got the symptoms right...

1. If you close either or both switches without lamps fitted, nothing ever trips.
2. If you install either or both lamps with the switches open, nothing ever trips.
3. If you install a lamp and close its associated switch, the RCD and MCB always trip immediately.

To trip the MCB and RCD with a single fault more or less requires it to be a hard L-E fault. The double trip could occur with a variety of multiple fault scenarios though, including L-N and N-E on this circuit, or L-N on this circuit and N-E on another on the same RCD, if the N-E fault is of relatively high resistance so that it only causes a trip when the short-circuit current from this circuit is present. Do you have good IR results for all circuits?

However, as well as the principle of Occam's razor, which suggests the simplest possible cause as the most likely, we also have the evidence that the lamp flashes momentarily. This also indicates an L-CPC fault. If L-N were shorted say at the lampholder, the potential across the lamp for the time it takes the breakers to trip would be zero, and it could not light. During that time, most of the 230V supply appears across the cable resistance, approx half of it in the line conductor from CU to fault and the rest in the neutral likewise. Obviously if they are different lengths, the 230V will divide proportionally, but none of it can appear across the lamp. If instead the fault is L-CPC, electrically near one terminal of the lamp, and other side of the lamp is connected to the unfaulted neutral, then the voltage drop along the CPC (approx 115V) appears across the lamp until one or other breaker clears, which is probably sufficient to make it light. An L-E fault not to the CPC but to a low-resistance earth like a bonded pipe, would not cause the lamp to light because it would represent a relatively smaller fraction of the fault loop impedance.

This gives us two pieces of evidence pointing at L-CPC faults at or near the lamps, so we should look for things these fittings have in common that could cause this. Are the lampholders the same? Are they metal and/or earthed? Perhaps they short internally when a lamp is installed? E.g. pressing the contact plungers causes the internal terminals move and contact each other / cable / lampholder body. How about internal visuals and / or an IR test L & N -> E with the lamps fitted.

I'm reluctant to suggest a faulty MCB but we do have to keep in mind the possibiity that it is not a genuine MCB trip event.

 

[telectrix]

i suspect 2 faults:
1. open circuit N somewhere between the problem lights and DB
2. a short circuit N-E. closer to the light/s.
with the bulb/s out, there is no voltage on the N. it is floating.
add a bulb, and because the N is O/C then the N will have 230V on it.

looks good in my head, but hope you can make sense of above. some simple testing should find the faults.

 

[Richard Measures]

Thanks guys. I was leaning towards the L-E fault having spent a while ensuring no borrowed neutrals are present.
I like the potential explanation of the brief lamp lighting

My plan (tomorrow) is to:
- Check the MCB
- Inspect both pendants visually and test for correct local continuities
- Identify and test resistance of individual wires between fittings and from CU
- Confirm continuity and separation of L-N-E from CU

 

[Lucien Nunes]

Did both of these work before you did the CU change? If so, what did you do that might have triggered the fault(s) ? Remove the lamps for testing?

 

[littlespark]

This was a CU change? So it worked before on just fuses? Or was there mcb before?

IR test these rooms, with switches in all possible configurations.

It’s going to be a bugger to put right if there’s an inaccessible joint box

 

[Dustydazzler]

Sounds like a job for #onedayrewire

 

[Richard Measures]

Well, found it but the fix isn't as easy as I would have liked.
There's a short between N and E between room 1 (that worked) and room 2.
Temp replacing this N sorts everything. Room 2 to room 3 is fine.
I'll need to lift floorboards and replace the N wire, or find the place where a floorboard nail goes through both
As this was a non-RCD system previously with large wired fuses, I'm guessing this fault has been present since construction, and only revealed by the modern RCD....

 

[Deleted member 9648]

Maybe I'm missing something, but surely your dead tests during the board change would have flagged a N-E fault before energising?

 

[Tony Reidy]

Could the better and wiser assist with an odd one...?

I've just finished a big CU swap on a 1950's house that uses single core & earth lighting circuits. All but one of these are fine, but the faulty one is pretty funky;

There are 3 accessories on this circuit in 3 separate rooms. The first room/switch/pendant works fine. The later 2 have the same issue - with no lamp in the fittings all is well, but put a lamp into either with the switch on and 3 things happen:
- The bulb glows very briefly (30mS?)
- The RCD covering this half of the CU trips
- The 6amp type B MCB covering the circuit trips

If this was just an RCD issue I'd be confident of a neutral/borrowed neutral issue, but how can a bulb (which doesn't blow/fail during this) cause an MCB to trip?

It's got me going round in circles so any clarity and advice is most welcome.

Thanks

Just a thought could be N + L reversed between the first and second pendant