New to the forum, trying to learn about neutral-earth fault troubleshooting!

[jezcunningham]

Qualified in electronics, only did electrical work as a DIY, now retired and enjoying life. Except when the RCD trips for no obvious reason...
Cheers
Jez

 

[DPG]

Welcome to the forum!

 

[littlespark]

Welcome

We all get frustrated at RCD trips, but its doing it for as reason... you just need to find the reason.

Can be a faulty appliance, unplug everything, not just switch it off. More details of your problem please

 

[jezcunningham]

Well, you asked for details - sorry if this is too long...

I have suffered random RCD tripping since I moved into this house 8 years ago. It happens infrequently - about once every 3 months on average - and I have struggled to find any pattern to it. Perhaps it is worse when the weather has been wet – but this isn’t a hard and fast conclusion.

For the initial house insurance I needed an inspection certificate (and every 5 years) and I asked the first electrician to advise on the tripping. He did all the standard measurements, found nothing wrong (insulation-resistance-wise) and replaced the RCD. Didn’t seem to solve it (still tripped occasionally) so we’ve lived with it, assuming it might be a mouse in the loft or…

A year later we had the kitchen replaced and some changes to the consumer unit – new induction hob with new radial circuit and disconnection of old wiring, repositioned oven and microwave with new radial circuit. Time rolls on and it’s still tripping now and again.

Year 5 comes and a different electrician does the inspection; some low IR measurements but no appliances/lighting disconnected so not considered a fail. (I’ve done a census – I have a total of 95 devices permanently powered – but that’s a whole other thread…)
He notes that “RCD cannot be reset without turning off all the MCBs”. Advises that I probably have a neutral-earth fault, which is notoriously time-consuming to troubleshoot – he’ll do it for a daily rate but can’t estimate how many days it might take.

We live with it a while but in October it tripped 4 times in as many days. So I decide it’s a suitable task for lockdown and I can do some of the grunt work myself. I buy myself a basic insulation resistance tester, a current clamp ammeter that reads down to 1mA, and I make a 13A extension cable with the cores separated so I can check each appliance. I spend a lot of time checking the appliances – nothing shows anything other than 0mA – but I’ve yet to check lighting…

So I move on to leakage current at the meter tails – it’s fairly steady at a minimum of 11mA but jumping up to 18mA or even 29mA occasionally – and I flick off the MCBs one-by-one to see which circuit might be contributing most. Strangely there are two MCBs which each reduce the 11mA to zero – one of the ring mains (not the kitchen) and the radial circuit to the garage.

So now I do insulation resistance (just at 250V) on those two – the ring gives 87Mohms and the garage gives 14Mohm. Same result on both L-E and N-E of course as loads are still in circuit between L and N.

Advance 3-4 days and it’s been raining and I repeat the garage tests and now I see less than 1Mohm so I suspect damage to the underground armoured cable. I turn off the 2-pole isolator in the garage (so I’m only measuring the cable) and – at 500V this time - I get >200M N-E and 70M L-E. I turn to the other side of the isolator and measure the garag e (radial power and radial lighting) and find it is the lighting circuit measuring under 1Mohm – it’s very simple – two switches and two fluorescent strips. They’re rusty and probably >20 years old so I assume the ballasts are shot.

I’m now at the point that I expect to replace the fluorescent fittings with IP-rated LED battens but I’m not convinced I’ve really found the root cause. For example, why did the leakage current drop to zero when I flick off the ring MCB _or_ the garage MCB? And why does the wet weather affect the results? Maybe I’m just living life on the edge at 11mA and any old transient pushes the RCD over the limit?

And it's not tripped for 10 days now...

 

[plugsandsparks]

Couple of things, its great that you have such detailed results as this helps alot. It certainly looks like you are narrowing down the problem. If you are definitely clamping the L and N meter tails and you are getting 11mA and its jumping around, then i would suggest the tripping is real and the RCD is doing a good job.

Tracing faults that are caused by dampness are notoriously difficult to pinpoint and readings from 0 to 200Mohms on a IR depending on which way the wind is blowing is typical of a nasty gremlin. Sometimes a visual clue is there, e.g. condensation or any wetness visible.

Lets say there is a NE fault on the garage lights, it does not matter which MCB you flick, its the one thats pulling a real load that you will notice, it does not really matter which one.

If you can, turn off the 2 pole isolator in the garage, clamp the tails and turn on some big loads like cooker, oven, kettle etc you get the drift. Measure the leakage. Then turn garage back on and repeat.

Hopefully the gremlin will expose itself and you can exterminate it.. lol

Good Luck, P&S

 

[SJD]

Rather than having lots of circuits on one RCD, perhaps consider changing the consumer unit to an all-RCBO board.

Firstly, if there are various circuits contributing to the leakage and tripping, it is far less likely to occur.
Secondly, if a circuit trips, you can see which one it is, rather than wondering which of several it might be.
Thirdly, the inconvenience of a trip will be much less if just one circuit trips.

 

[jezcunningham]

Ah - good idea (plugs&sparks)... Now I know what I'm doing tomorrow...
Thanks
Jez

 

[jezcunningham]

Rather than having lots of circuits on one RCD, perhaps consider changing the consumer unit to an all-RCBO board.

Firstly, if there are various circuits contributing to the leakage and tripping, it is far less likely to occur.
Secondly, if a circuit trips, you can see which one it is, rather than wondering which of several it might be.
Thirdly, the inconvenience of a trip will be much less if just one circuit trips.

I did consider that - until i saw the price of an RCBO - the board has 18 breakers...

 

[DPG]

I did consider that - until i saw the price of an RCBO - the board has 18 breakers...

Perhaps a dual RCD board with the two highest leakage circuits on opposite sides?

 

[SJD]

I did consider that - until i saw the price of an RCBO - the board has 18 breakers...

Which brand(s) did you check?

For example. Fusebox RCBOs are approx. £10 each.
Some makes are a lot more expensive.

 

[jezcunningham]

Which brand(s) did you check?

For example. Fusebox RCBOs are approx. £10 each.
Some makes are a lot more expensive.

I have a Crabtree CU so I think I looked for something that would plug in directly replacing the MCBs - seemed to be about £30 each

 

[SJD]

I have a Crabtree CU so I think I looked for something that would plug in directly replacing the MCBs - seemed to be over £40 each

I'd ditch the Crabtree board and start again.

 

[plugsandsparks]

Let us know how you get on. Its a good job you did not mention any outside lights,,, eeek

 

[jezcunningham]

Perhaps a dual RCD board with the two highest leakage circuits on opposite sides?

I'm still hoping to find a fault that means I can keep the existing CU

 

[jezcunningham]

I'd ditch the Crabtree board and start again.

But the fault might still be there...

 

[jezcunningham]

Let us know how you get on. Its a good job you did not mention any outside lights,,, eeek

well... there are a couple or three, but as I mentioned I've not got around to the lighting circuits yet - they didn't show any leakage current

 

[SJD]

But the fault might still be there...

Yes the fault might still be there, but at the moment it seems you don't know how many faults you have, or where they might be. The whole point of an RCBO board is to make each circuit independent, a fault on one won't affect another. I'd say it is a worthwhile improvement irrespective.

That said, I've not seen details of the supply (TN-S, or TN-C-S, vs. TT) which might have some bearing on any consumer unit change.

 

[plugsandsparks]

If you are looking for an LE fault, its one process, for a NE fault it is another. The trick is to reliably get readings, its also the hard bit. Without getting into the difficult forum issues of safety etc etc. Your best bet is to provoke an NE fault by drawing lots of current, does not matter where from, i.e. kettle, immersion heater and all that. Then isolate L and N from suspecting areas. i.e. your double pole garage isolator. Otherwise its get a pro in to start to pull circuits out of the board to test, assuming the fault wants to play.

 

[nicebutdim]

I'd ditch the Crabtree board and start again.

Might not be a popular suggestion, but there are always plenty of Crabtree RCBOs at reasonable prices on ebay. Many seem to be sold by people in the trade.

Chances are it'd still be cheaper to replace the lot with another brand.

 

[jezcunningham]

Yes the fault might still be there, but at the moment it seems you don't know how many faults you have, or where they might be. The whole point of an RCBO board is to make each circuit independent, a fault on one won't affect another. I'd say it is a worthwhile improvement irrespective.

That said, I've not seen details of the supply (TN-S, or TN-C-S, vs. TT) which might have some bearing on any consumer unit change.

I’d love to just rip it out but to spend the best part of a grand to narrow it down for fault finding seems a bit drastic. It may come to that...
It’s T-N-CS by the way.