Nuisance tripping of ringmain RCCB

[Daddy Boots]

We have 4 circuits with MCB's protected by a Siemens 80A 5 SM3 317-0 RCCB with 30mA rated residual current. Came back from holiday last year to a defrosted freezer and a tripped RCCB. Never found the source and it never re-occurred... until now.
After torrential downpours it tripped again. I was able to isolate the problem to the circuit that has a fused spur to the pond pump/uv light. The cable goes to switches mounted on the garden wall and then to a box on the ground with uv light and connection to the pump. Will it be moisture in the box on the ground that's the likely culprit ? What device can I install in the fused spur circuit so that trips rather than the 80A RCCB which takes out all of the ring circuits ?

 

[PrenticeBoyofDerry]

It is highly likely that the recent downpours have been responsible for many an RCCB to trip, so yes moisture ingress will be the culprit.
With regards to what device is best used to prevent the rest of the house tripping, I would assume better enclosures for the external accessories would be the simple answer. Installing a second inline RCD unit may solve it but there would be no guarantee which unit would trip first, taking the external circuit of the RCCB would also be an option and install it on it's own device. But I would at least have a double pole in-line switch between house electrics and pond feature, so it can be isolated across the two poles and then at least it will not prevent you prompt getting the electric back on.

 

[Daddy Boots]

Thanks. I have a spare 5A MCB in the CU that was intended for an additional lighting circuit that we won't be needing. It's quite a run from the CU to the current fused spur connection so another RCCB would be easier. Would a 10mA Rated residual current device not always trip before a 30mA ?
I don't suppose it's allowed to take a fused spur off a lighting circuit to feed the pond pump/UV ?

 

[RF Lighting]

It depends on the fault, but something like ingress of water, the RCD's will not discriminate and either one or both will trip.

Have you tried going round with the silicone and preventing the fault recurring, rather than spending all this time and effort trying to mask it?

 

[JohnW2]

It depends on the fault, but something like ingress of water, the RCD's will not discriminate and either one or both will trip.

Does not the effect of moisture or water ingress tend to arise gradually, such that the leakage current' will get to 10mA long before (as more moisture/water arrives on the scene) it gets to 30mA?

Kind Regards, John

 

[RF Lighting]

I've done plenty of ramp testing and investigating on moisture ingress RCD faults and they tend to either go short circuit or stay clear. It doesn't seem to be a gradual thing.

 

[JohnW2]

I've done plenty of ramp testing and investigating on moisture ingress RCD faults and they tend to either go short circuit or stay clear. It doesn't seem to be a gradual thing.

Fair enough - but that does very much surprise me, particularly with rain water, which tends to have a low conductance.

Of course, unless you actually undertake contrived/engineered experiments, you would not get much exposure to sub-30mA leaks, because they wouldn't usually come to anyone's attention.

Kind Regards, John

 

[RF Lighting]

My experience is probably slightly skewed as it's rare to see a 10mA RCD but it's very unusual to see a fault which has any real effect on a ramp test of a 30mA RCD.

 

[EFLImpudence]

but it's very unusual to see a fault which has any real effect on a ramp test of a 30mA RCD.

Are you sure?

I thought that was a way of determining on which circuit the leakage was (without the relevant clamp meter).

 

[RF Lighting]

Yes I've rarely seen any difference between the before and after ramp tests. It's always IR testing which reveals where the fault is.

 

[JohnW2]

My experience is probably slightly skewed as it's rare to see a 10mA RCD but it's very unusual to see a fault which has any real effect on a ramp test of a 30mA RCD.

That doesn't surprise me - as I said, if it's a 30mA RCD and the water/moisture ingress hasn't yet progressed to the point of causing a 30mA leak, there's no reason why you should even be on site, let alone doing a ramp test. That's why I strongly suspect that you rarely encounter a sub-30mA leak.

Kind REgards, John

 

[JohnW2]

Yes I've rarely seen any difference between the before and after ramp tests. It's always IR testing which reveals where the fault is.

Well, yes, but what sort of IRs do you see in that sort of situation? It would have to be as low as about 7.67 kΩ (0.0077 M&#937 to give you a 30 mA leak, or 23 kΩ (0.023 M&#937 to get a 10 mA leak - and I suspect that the IRs you usually see in the sort of situations you're talking about are probably a fair bit higher than that, aren't they?

Kind Regards, John

 

[securespark]

I find IR readings on a dodgy circuit with moisture ingress can be all over the place, but often I see them at almost, if not, zero.

 

[JohnW2]

I find IR readings on a dodgy circuit with moisture ingress can be all over the place, but often I see them at almost, if not, zero.

Same here, but what do you mean by "almost, if not, zero"? The lowest IR my 1652 can display is 0.1 MΩ, the 'next thing down' being zero. A 'zero' reading can therefore mean anything up to 50 kΩ, which would only result in a 4.6 mA leakage at 230V. The lowest possible non-zero reading (0.1 M&#937 could mean anything up to 150 kΩ (only 1.5 mA leak at 230V), and certainly at least 50 kΩ (only 4.6 mA leak). Our MFTs (and dedicated IRmeters) are fine for identifying seriously dodgy cables, but are next-to-useless for detailed investigation of 'leaks' which are only of the order required to (just) trip RCDs.

Kind Regards, John

 

[EFLImpudence]

I cannot see how 'my' ramp test cannot work.

Are we saying - not that it wouldn't/doesn't work but that it is unlikely to be useful because a circuit leakage of between the relevant values is unlikely?

However, appliances could have these leakages.
If an installation trips an RCD (30mA) then obviously a ramp test cannot be done but by disconnecting all loads the RCD would ramp test at 30mA.
I.e. a difference of 0mA and 30mA required to trip the RCD.

Were the leakage 20mA on one circuit and 10mA on another, then surely by disconnecting the 20mA leaking circuit the ramp test would read 10mA and by, instead, disconnecting the 10mA the ramp test would read 20mA.
By disconnecting both, then 30mA reading.

If there were a leakage of 20mA, how can the ramp test still read 30mA?
If it does, then isn't the ramp test faulty and/or useless?

Is this not the equivalent of 'leakage clamping' the tails and reading different values by disconnecting different circuits.