RCBO energised despite N-E short... until the power was cycled...

[happysteve]

The bit where I cut to the chase is in bold below if you can't be bothered with the back story.

Changing MCB in main house DB for garage supply, so turned off at main switch.

Contactum metal DB (about 4 years old), 2x MCBs for existing feeds to sub boards, the rest RCBOs (full height, single pole, type AC). TN-C-S, bonding good.

On turning the main switch back on, one of the RCBOs tripped, and would not reset. Kitchen sockets.

Kitchen re-done 5 years ago, new cables all the way back to the board. Ring continuity fine, N-E fault of <1Ω (measured at DB). Not MΩ, not even kΩ: Ω.

Traced to between 2 sockets: local R1+R2 between the two sockets = 0.15Ω; N-E = 0.09Ω from one, 0.06Ω from the other. So, dead short, 3/5 of the cable length between the two. Exact location and cause TBC as I forgot my X-ray specs and this is not a can of worms I wanted to open on a Friday.

RCBO (with circuit disconnected) tests out fine.

Question is... how was the circuit ever working with that fault there? And, am I right in thinking that if there had been a power cut since the fault occurred, it would have tripped then (and not been able to be reset)? Is there a way that a dead short, N-E, can "blind" an RCBO (in the same way flooding it with DC can)?

Was really, really annoying, especially on POETS day. All I did was turn the power off then on again, the fault makes itself apparent, and I have to deal with it. Client was understanding, particularly when shown fault readings and after explaining everything, but just as confused as I am about why it was working at all.

 

[ImpededLoop]

I guess it's possible that the Rcbo had temporarily failed so it didn't trip originally. Perhaps cycling the power has made it work again - a bit like when they sometimes don't trip when testing first time round after not having been tested for years, but loosen up after they've been operated.

Sounds like you've done a great job of testing and working out what/where the fault is.

 

[Dave OCD]

Is the earthing arrangement TT ? Often with a high Ze the RCD/RCBO doesn't see a N/E fault unless there's a fair load on the circuit.

 

[Lucien Nunes]

We don't know the nature of the N-E fault. If the conductors are just barely making contact, especially if one of them is steel (e.g. insulation cut through on steel back box or nail through cable) it is quite possible that it has never made very good contact due to oxidation / tarnish film on the metal surface. Then, when you opened the DP main switch (breaking the real neutral connection) there might have been a sudden current surge through the fault that broke down the insulative boundary and made it a proper low resistance fault.

 

[7029 dave]

Have you left it de energised as the ring continuity has been broken or temp down rate to 20A

 

[westward10]

A connection between neutral and earth will not necessarily trip an rcd as there will not necessarily be an imbalance. As above this can often be the case with TT due to the high resistance ground return, the majority current will just carry on down the neutral as it is the least resistive path. It can also occur in TN systems where the majority current will just keep to the neutral.

 

[Lucien Nunes]

I think the point the OP was making was that the fault was obviously already present, but had never previously caused a trip. Then, after doing a job on an unrelated circuit, it started making a nuisance of itself without any conditions apparently having changed. My theory being that it was the act of opening the DP main switch that causes a transient that made the fault more 'solid.' Sort of like a wireless telegraph coherer.

 

[happysteve]

Is the earthing arrangement TT ? Often with a high Ze the RCD/RCBO doesn't see a N/E fault unless there's a fair load on the circuit.

Thanks, good suggestion. Though in this case it's TN-C-S (Ze = 0.11Ω)

Have you left it de energised as the ring continuity has been broken or temp down rate to 20A

Would've if I could've... the only B20 RCBO I had with me would not have fitted the busbar, and nothing suitable/spare in the board to double up on, so it's still on the 32A until I go back on Wednesday, by which time I'll have a proper replacement of the correct manufacturer. Did a risk assessment on leaving energised with inadequate OCPD (32A) having audited the appliances and where the ring was split, deemed this safer than trailing extension leads across the kitchen to power essential loads. Fortunately they have a separate utility room so washing m/c and dryer not on the circuit... dishwasher, fridge and freezer on one side, kettle on the other. Client is aware of the risk of overloading until then, and basically knows what's what.

We don't know the nature of the N-E fault. If the conductors are just barely making contact, especially if one of them is steel (e.g. insulation cut through on steel back box or nail through cable) it is quite possible that it has never made very good contact due to oxidation / tarnish film on the metal surface. Then, when you opened the DP main switch (breaking the real neutral connection) there might have been a sudden current surge through the fault that broke down the insulative boundary and made it a proper low resistance fault.

I think the point the OP was making was that the fault was obviously already present, but had never previously caused a trip. Then, after doing a job on an unrelated circuit, it started making a nuisance of itself without any conditions apparently having changed. My theory being that it was the act of opening the DP main switch that causes a transient that made the fault more 'solid.' Sort of like a wireless telegraph coherer.

Thank you, Lucien! This is exactly the sort of thing I'm grateful someone suggested. My previous line of thinking was... is a certain type of fault (N-E short) treated differently if it occurs (a) while the device is already energised, and (b) if it's already present when the device is manually reset. Your suggestion is that the act of turning off the main switch changed the nature of the fault, by making a poor contact N-E into a really, really good contact.

I will let everyone know what I find... it may not be on Wednesday (when I go back to temporarily change the RCBO) as there is other urgent work at the property that needs doing first.

Thanks to everyone who posted

 

[pc1966]

What is puzzling is how that circuit ever passed IR testing in the original case. Maybe that was skipped, or maybe the buried cable has been damaged in the interim period (nailed by some addition to the install, or trapped years ago and gradually crushed until PVC gave out, etc).

My own suspicion is the RCBO had stuck, as probably they never have never tested them in the years since it was done. Why would they? Most folks see the modern CU as the old fuse box, it just sits there until something goes BANG! and then they look at getting someone in to sort it.

 

[newfutile]

Usually a neutral to earth fault will trip the (RCD) part when current is drawn on the circuit, however i have also had a case where the same fault caused the rcd to be untrippable either by test button or by any current from my rcd tester. After the fault was repaired the rcd worked fine !