Another RCD puzzler

[Luke]

Hi everyone,

I've got a "Steeple" split load consumer unit installed in 1997 with 2

• ring mains on the 30ma 80A RCD side. On the non-RCD side there are a number of other circuits, including an outside socket, protected by a "Powerbreaker" 30ma integral RCD socket.

We moved into the house in July. Ever since we moved in, we've been having nuisance RCD trips, seemingly following no pattern. Everything would be fine for a while, then all of a sudden the RCD would trip - usually when using the shower, or maybe switching the kettle on. Lately the problem seems to have been getting worse - it tripped during the night, a period of low activity, apart from the immersion heater which is on the non-RCD side anyway.

I thought it might be mouse activity biting through the cable, but then...

...this morning I went to switch on the vacuum cleaner, a double insulated device, i.e. no path to earth. Immediately the RCD tripped. The happens on the upstairs and downstairs rings. As I couldn't see how there could possible be an earth fault in the vacuum cleaner, I plugged it into the outside socket protected by its own RCD but NOT protected by the RCD in the consumer unit. It worked fine and did not trip the outside socket's own RCD.

Am I right to suspect the RCD itself in the CU rather than any sort of appliance or wiring fault? FYI the house was rewired in 1997 along with the CU replacement - I've had a lot of floorboards up following CH install and it all looks in good nick.

I am contemplating chancing =A330.00 on a new RCD and wanted to get the group's general opinion.

Thanks

Luke

[Dave Osborne]

Yes, you should definitely suspect the RCD. Also, you may have an insulation fault that is pushing the RCD very close to its trip level.

This means that the slightest imbalance or increase in humidity will set it off. The vacuum cleaner sets it off because of the inrush on starting (John Rumm might be along later to explain why).

You can buy a Hager 63A/30mA RCD from TLC for £16.80 +VAT

get your installation checked out at some point to see if you can find where the leakage current is coming from.

[ARWadsworth]

If the RCD is not at fault then this is almost certainly a N-E fault or possibly a L-E fault on the RCD side of the installation. I would suspect that a N-E fault is more likely than a faulty RCD. I would start by inspecting the back of all the sockets to check that a fixing screw is not touching a neutral cable.

[tony sayer]

In article , Luke scribeth thus

Sounds like you may have a Neutral to Earth short somewhere.

Switch off at the main switch and pull the fuses or trip the MCB's and do a high ohms test from the neutral of each circuit to Earth and you may well find it there.

Whilst at it there should be no ohms from Live to earth either if there is and you do find which circuit then trace it down from there .. easier said then dun;!...

[tony sayer]

Prolly thats causing the leakage to earth, via live or neutral to alter..

Blimey .. they've dropped in price since I last bought one;!..

[Tabby]

No, most likely is an electrically leaking appliance. Spend your notes on a multimeter first.

?title=3DMultimeterNT

[Andy Wade]

Hmm, I'd prefer to have quite a lot of ohms between line and earth :~)

[Andy Wade]

I've never known an RCD fail by becoming over-sensitive. I've known a few though that have become insensitive and failed to trip when they should. The test button is there for a purpose...

For this case a N-E short or leakage, which could be in an appliance or in the fixed wiring, is no. 1 suspect, as others have said. I'd put the RCD itself a long way down the list.

[tony sayer]

OooH!, you pedant;!. Woss wrong with no ohms /indicated/ on yer meter;?..

[Andy Wade]

Milliohms, megohms, what's the difference :-)

If my meter indicated no ohms I'd think it needs a new battery.

[John Rumm]

Covered here:

Also, get your installation checked out at some point to see if you can

Following some of the tests described here can find many causes:

[Dave Osborne]

Ah, I wasn't suggesting that the RCD becomes over sensitive, rather that is is and has always been over sensitive.

An RCD that trips at 17mA is likely to be noticeably more of a nuisance than an RCD that trips at 27mA.

I've known a

The test button is of limited facility. There's no substitute for testing with an RCD Tester. I've had RCD's that didn't trip on the test button but did trip on a "proper" fault and vice versa.

[Ronald Raygun]

I wouldn't have thought so. Surely if there is a leakage (sorry, residual) current of 17mA then there is something seriously wrong somewhere and it's worth finding it and fixing it.

[Andy Wade]

Fairynuff, but you shouldn't consider one that trips at 17 mA as over-sensitive - the standard allows them to trip anywhere between 15 and 30 mA. In my experience though they all tend to trip at around the

20 mA mark, ±2 or 3 mA.

Agreed, but it will (usually) alert you to a device that isn't going to trip at all. If the test resistor circuit has gone o/c you'll get a false negative result, which errs on the safe side.

Agreed, but, OOI, how many times have you found an RCD that failed on elongated trip times but did actually trip on the test?

[Dave Osborne]

Given that Hager (for example) sell RCDs that consistently trip at

27-28mA, I can (and I do) consider one that trips at 17mA as over sensitive.

To be clear, though I meant over sensitive in the sense of "10mA more sensitive than it needs to be" rather than "failing the 1/2 Idn test".

- the standard allows them to trip anywhere between 15

Actually, the standard says that they must not trip at 15mA and must trip at 30mA (within 200ms). It is entirely with the manufacturer to decide what actual trip level they set their product at.

In my experience though they all tend to trip at around the

Well, in my experience 'most' 30mA RCDs trip at 25-28mA.

I do temporary power distribution for outdoor events. I don't know exactly how many RCDs are in my care but it's in the region of 80~100 and of a variety of makes.

They all get fully tested at least twice (and often several times) a season.

I expect to get failures at the start of the season (bearing in mind that they may have over-wintered un-powered in an unheated shed, garage or container).

To answer your question, which I take to mean "how many times have I found an RCD that tripped at the required fault current, but *not* within the prescribed time" then I would say at least a couple each season.

Sometimes you get a duff batch and (usually after a season or two of being fine) most or all of them fail to survive over-wintering.

[js.b1]

A clamp meter on the CUs earth wire can be useful (mA range).

A 30mA RCD will tend to trip at 23-25mA. Several appliances with mains filters each leaking circa 1.5mA can create a "background" leakage of 15mA making it very easy for another appliance to push the total over the RCD trip threshold.

If the house does have a single RCD (TT supply or "whole house") then conversion to a dual RCD CU would make life a lot easier when the fault is found.

[tony sayer]

Well if there's no indicated 'omes then theres no problem;)..

[John Rumm]

It depends a bit on what you have on the circuit. Lots of IT kit and other electronic items for example will typically "leak" a small amount due to the mains input filtering. With enough items on the circuit the cumulative "normal" leakage can add up and take a lump out of your available trip threshold.

[The Natural Philosopher]

And more to the point, if there is any step surge or interference on the mains input, can transfer quite large currents resulting from the harmonics higher than 50Hz.

[Ronald Raygun]

I don't find either of these two answers particularly illuminating.

Would you guys mind explaining why mains input filtering, or harmonics, would "leak" current to earth? Surely *no* equipment should have *any* connection to earth other than from the enclosure, and there should be no current path to earth from any of the "gubbins". And why would interference harmonics cause a leak?

Doesn't in all circumstances "what goes in must come out" apply? In other words why should the live current ever (at any instant) differ from the neutral current (except in direction)?

Or is it that there is never a real leak, but the breaker is fooled into thinking there is, in which case what is the mechanism for fooling it?