RCD tripping. TT. where could be the fault? (FIRE ALARM SENSOR)

[JohnW2]

With what? I have been in a hurry and measured with a multi-meter and got open, but measured with an insulation tester showed nearly short circuit.

That certainly happens but it's hard to think of a way that it could happen the other way around - so, even if the 40Ω N-E measurement were at low voltage (e.g. with a multimeter), it's very unlikley that the true N-E resistance is not at least as low as 40Ω, perhaps even lower, isn't it?

Kind Regards, John

 

[EFLImpudence]

Indeed but, as I wrote, I think also even if there is load taking current on a a circuit protected by a different RCD (since some of the neutral current will then flow 'backwards' through the affected RCD and the N-E fault on one of its circuits.

You are correct. Have you discovered another thing that no one realised?

I presume that the shared current is so low that the RCDs don't generally trip.

I have drawn some diagrams and it would appear that any other circuit - RCCB, RCBO or just an MCB will have the same effect.

 

[JohnW2]

You are correct. Have you discovered another thing that no one realised?

I don't think I deserve any credit for this one. I'm pretty sure that I first came to realise it (a long time ago) when someone pointed it out to me - maybe even here

I presume that the shared current is so low that the RCDs don't generally trip.

That obviously "depends" on the circumstances - in particular the impedances of the 'normal return path' of the circuit without the fault and that of the path via the affected RCD ('backwards') and the N-E fault.

Don't forget that if the load on the intact circuit is appreciable, only a tiny proportion of it's neutral current has to go through the 'parallel path' (via the N-E fault) for it to be >30mA. If, say, that load is 13A, then only about 0.23% (at most) of the neutral current has to go via the 'affected' RCD to cause it to trip.

I have drawn some diagrams and it would appear that any other circuit - RCCB, RCBO or just an MCB will have the same effect.

Sure, the device feeding the other circuit is irrelevant. Indeed, even if, hypothetically, there were no OPD (e.g. circuit fed directly from CU Main Switch), the same would still happen.

Kind Regards, John

 

[EFLImpudence]

That obviously "depends" on the circumstances - in particular the impedances of the 'normal return path' of the circuit without the fault and that of the path via the affected RCD ('backwards') and the N-E fault.

Yes, but I don't remember anyone saying such a fault could be anywhere.
Only Bernard's diagram with both devices on the same Line busbar

Don't forget that if the load on the intact circuit is appreciable, only a tiny proportion of it's neutral current has to go through the 'parallel path' (via the N-E fault) for it to be >30mA. If, say, that load is 13A, then only about 0.23% (at most) of the neutral current has to go via the 'affected' RCD to cause it to trip.

Yes, again, but it is not common for two or more RCDs to trip - the one on the circuit with the fault and perhaps all the others.

 

[JohnW2]

Yes, but I don't remember anyone saying such a fault could be anywhere.

OK. As I said, I think someone one pointed it out to me - but I'm happy to take the credit

Only Bernard's diagram with both devices on the same Line busbar

It did, and his text said the same. That's why I jumped in with my comment that it could even be something on a 'different busbar' (different RCD)

Yes, again, but it is not common for two or more RCDs to trip - the one on the circuit with the fault and perhaps all the others.

That won't happen with a situation I described. The 'unaffected' RCD won't trip, because the L and N currents through it will be balanced - it's only the RCD protecting the circuit with the N-E fault.

I think the only common situation in which two RCDs will trip is nothing to do with faults but, rather, due to a 'borrowed neutral' situation.

Kind Regards, John

 

[EFLImpudence]

That won't happen with a situation I described. The 'unaffected' RCD won't trip, because the L and N currents through it will be balanced - it's only the RCD protecting the circuit with the N-E fault.

Oh. I thought that was the point you were making.

I see now I was mistaken.

 

[bernardgreen]

I think I have found why both RCDs can be tripped but only when the supply is a TT

The dotted green path is a sneak path to Ground

 

[JohnW2]

Oh. I thought that was the point you were making.

No, the point I was making was that the neutral current due to a load on the circuit on the 'other ('unaffected') RCD' splits upstream of that RCD, with the majority going through the normal return path to the tranny and a small amount going backwards through the 'affected' RCD, hence via N-E fault to earth - and since there is zero current through the L side of that 'affected' RCD (but the fault current through the N side) ,it will trip.

The other ('unaffected') RCD will not trip because equal L and N currents go through it, the 'splitting of the N current being upstream of that RCD.

I see now I was mistaken.

As above, I hope?

Kind regards, John

 

[JohnW2]

I think I have found why both RCDs can be tripped but only when the supply is a TT ... The dotted green path is a sneak path to Ground

I'm not so sure about that. The total current (normal blue path plus dotted green 'sneek path') through the N side of the lower RCD will surely be the same as the L side of the RCD (hence no trip), won't it?

Kind Regards, John

 

[bernardgreen]

There is no load on the top RCD so no load currents through that RCD. The only current through that RCD is the sneak current ( dotted green )

 

[EFLImpudence]

I'm not so sure about that. The total current (normal blue path plus dotted green 'sneek path') through the N side of the lower RCD will surely be the same as the L side of the RCD (hence no trip), won't it?

There is no dotted green sneak path through the lower RCD.

There is no load on the top RCD so no load currents through that RCD. The only current through that RCD is the sneak current ( dotted green )

The only sneak path is the black line along the top neutral to the fault.
The other neutral conductors carry normal load return current apart from the small amount missing upstream of the joint.

 

[JohnW2]

There is no load on the top RCD so no load currents through that RCD. The only current through that RCD is the sneak current ( dotted green )

Indeed - so the top one will trip. However, as I said, I see no reason why the L and N currents through the lower RCD should differ, since they are simply the currents going into and out of the load, so that one shouldn't trip.

Am I missing something?

Kind Regards, John

 

[JohnW2]

There is no dotted green sneak path through the lower RCD. ... The only sneak path is the black line along the top neutral to the fault.

That depends upon what you mean by the "sneak path".

A small proportion of the current going (from the load) through the N side of the lower RCD will ('upstream') go via the other RCD and the N-E fault to earth, the remainder (great majority) going through the normal return path back to the tranny. As discussed, that 'small proportion' may be enough to trip the top RCD, but I still see absolutely no reason why the bottom one should trip.

Kind Regards, John

 

[EFLImpudence]

That depends upon what you mean by the "sneak path".

I was using your wording.

A small proportion of the current going (from the load) through the N side of the lower RCD will ('upstream') go via the other RCD and the N-E fault to earth, the remainder (great majority) going through the normal return path back to the tranny.

Yes but the Neutral current from the load is not increased because there is a leak elsewhere - so no sneak current at the bottom RCD; just the load current.

As discussed, that 'small proportion' may be enough to trip the top RCD, but I still see absolutely no reason why the bottom one should trip.

Agreed, but that small proportion is part of the load return current to the joint which is then missing from the back to tranny return path upstream of the joint.

 

[mrmisior]

Thank you everyone for fantastic responses!
Kind and helpful

I have now removed all 8 double sockets leaving bare, exposed wires.
Tested all sockets with multimeter - fine, no shorts.
Mapped all wires so I have a diagram.
Found 1 spur to a double socket on the other side of wall. This is also fine, no shorts.
All wiring appears to be above plasterboard ceiling and two cables dropped to each socket within plaster.
Cables appear to be new-ish with no signs of deterioration.
Tested all wiring, all ends, with multimeter: continuity ~0.5 Ohm and no shorts between L/N/E
...except... for the first and last ends which both read 40 Ohm between N and E.

I was hoping to find the fault within the sockets or between them, but no luck.

Fault is somewhere between the consumer unit and either first or last socket point.

Either way, I am stumped. Cables are new. No rodents. No moisture. Nothing has been recently drilled. Nothing comes to mind.

Tomorrow, I will have the faulty section determined, after removing the ring wires from consumer unit and will take it from there.