RCB nuisance trips

[Huge]

We're getting nuisance trips on our main house RCB (which only covers power and not lighting). Once it trips, you can reset it.

We had about 10 in one day a few weeks ago, which I traced to the oven, which we've stopped using (we have 2, and the kitchen is scheduled for replacement, so no big deal. I assume the element has gone leaky), but we had one yesterday (reason unknown) and one this morning when I turned on the computer "rack" in the study, which contains an elderly Compaq PC (and its 15" LCD monitor) running my firewall, a Netgear network switch, an Arcam Alpha 9 amp (which powers the audio in the study) and the HP LJ5M printer (which was likely not switched on anyway - I forget). I switched on the Olson strip which provides power to this lot and the RCD tripped immediately. I reset it and everything came back just fine.

Can anyone suggest how on Earth I go about diagnosing this?

[Syd Rumpo]

Disconnect the earth wire from the socket strip plug and connect a DVM on AC amps between this wire and the pin it was connected to. Look for earth current as you plug in various appliances. IIRC the RCB should trip at 30mA, so look for something below that.

With luck, you just have one or more leaky suppressor capacitors.

Cheers

[Harry Bloomfield]

Huge a écrit :

They can trip due to leakage between L to E as well as leakage between N to E - so switching something off, will only usually isolate the L.

Oven isolators should isolate both poles.

Most computer equipment includes filters on the mains input, which will leak to E, if enough are on a circuit the leakage can be enough to cause RCB's/RCD's to trip. If there is already some leakage on another appliance, the computer equipment might push it over the trip threshold.

Finding the culprit is a matter of leaving one thing off at a time, but keep in mind that more than one item might have leakage = or checking the leakage with a meggar, after making sure everything with filters at their input have been unplugged.

A meggar should be able to help you find an intermitant leakage to earth.

[Andrew Gabriel]

The right way to do this is to pass the live and neutral conductors through a current transformer, and measure the differential current, rather than disconnecting the earth (which might not be the leakage path anyway, and is not safe, most particularly when you have suspected earth leakage).

However, current transformers themselves present some hazards, so this isn't easy to do safely. (Don't know if any clamp meters are accurate enough to measure 10mA or so, but that would be another way.)

[Syd Rumpo]

Well, it's a quick and easy test which won't give a false positive.

I'm guessing that Huge has a meter but not a CT, he is aware of shock hazards, and if he /was/ careless enough to get a shock I'm sure he has the vocabulary to deal with it.

Cheers

[Syd Rumpo]

But you are right, it is bad advice to give on a public forum.

Cheers

[John Rumm]

[ARW]

Possibly not a fault. A RCD when working correctly is allowed to trip at anything more that half it's rated trip current.

Leaky class I appliances can soon build this up to 16mA.

Not that such info helps you.

Whats the CU? It might be possible to get some of the MCBs that are RCD protected changed to RCBOs and reduce the leaks on the RCD.

[Andy Wade]

Yes, there are a number of 'leakage clamp' meters on the market now - e.g. the Fluke 360 which has 1 uA resolution on the 30 mA range. (I have one.)

Very useful for investigating RCD tripping problems by checking appliance leakages, etc, but it won't catch events that are purely transient in nature. However if the' background' leakage is found to be much more than 10 mA you're in a too close for comfort situation, where a transient might trip the RCD. (A 30 mA RCD can trip anywhere between 25 & 30 mA.)

[Andy Wade]

Bzzt. Typo: ... anywhere between 15 & 30 mA.

[John Rumm]

One way of squeezing a bit more resolution out of a "normal" clamp meter is to make up a test extension lead with a long length of the outer flex sleeve removed - just leaving the individual insulated wires exposed. You can then loop multiple turns through the clamp. It won't get you to

1uA sensitivity, but can often get down to a few mA.