Earth leakage switch tripping when loadshedding starts

I have installed the Victron MultiPlusII 8kv and rececently and intermittently, when the power goes during loadshedding, the Earth leakage trips cutting electricity to the house. Does anyone know what the cause this might be?

If I had to guess, you have a standing leakage that is close to the point where the RCD starts tripping, and when the grid fails the disturbance caused by the Multi disconnecting the grid and then closing the ground bonding relay creates a transient that is just enough to push it over the limit.

Standing leakage: In every house there is a small amount of current that leaks to ground. This is due to actual leakage in conductors (usually very low), but also due to EMC and surge protection devices that is now fitted to many appliances. It is not uncommon to have a few milliamps of “normal” leakage, and this is known as the “standing leakage”.

In my house the standing leakage is 13mA. RCDs starts to trip around the 15mA mark. I actually measured mine and during switchover it would peak at 20mA.

The solution was to split my loads into two blocks, each with their own RCD, and now a much healthier 7mA leakage on each. Tripping resolved.

At a previous house I tried the same trick and I could not fix it. I ended up fitting an expensive “high immunity” RCD, and that solved it.

But also: it is usually a good idea to 1) check the RCD itself for proper working, using a tester (not just the button on the RCD), and 2) actually check your insulation as well, with an insulation tester. May need to get a sparky in for those jobs.

Thanks for the information. I will get my electrician onto it.

The cable leakage (resistance) to earth should be almost nil otherwise it’s time to rewire the house…


Just want to add to that, at the time, we both had the same machines, the same problem, but totally different causes.

In my case, the earth of the house was not properly connected outside the house, unbeknownst to the occupants over years of living here. When that was fixed, all worked.

Good idea. Let them test the earth of the house first, takes seconds.

So in summary, from cheapest to most expensive:

  1. Test the earth leakage, with something like this.
  2. Test the earth continuity, sparky will have a tool for that
  3. Test the insulation of every circuit, live and neutral to earth, sparky has the tool for that.
  4. Check the standing leakage (there is a trick for that that avoids expensive tooling), and if necessary split your circuits into two blocks, if it is over 10mA.
  5. Install an impact-resistant RCD.

The trick for testing standing leakage without the expensive clamp meter: Use the testing tool I linked in 1. Those have an adjustable test current, going down to 10mA. By first testing the RCD on its own (to determine where it trips), and then using this tool to preload the RCD with 10mA, and then by turning on individual circuits until it trips, you can employ a proverbial “difference method” to detect if you have more than 5mA leakage on a circuit.

I mention this, as many sparkies won’t be running around with an earth leakage clamp meter all over the place… :slight_smile:

Edit: I notice now the RCD tester I linked as example is rather on the expensive side. There are cheaper ones that are non-digital. That’s what I had in mind. It is also more likely that the sparky has a RCD tester in his bag than having a fancy earth leakage clamp tester.


In my case I also got lucky, I had just enough space in the DB to fit another RCD, by removing the ancient wide Heinemann RCD (1) and installing a narrower CBI QA17C. I even had a spare Neutral bar (2).


In my experience the CBI breakers aren’t too trigger-happy, and good quality, even if they are pushing R1k in price by now. When you have a Samite board, your’e somewhat limited in choice though.

I’m fairly sure that is the only way: an expensive clamp meter will most likely not give a good answer and the leakage current in a modern home is not 50Hz or linear. This is a deceptively difficult measurement to make…

1 Like

What is an impact-resistant RCD? I have heard of anti-nuisance tripping RCDs.

must be for rugby players :grin:

That’s for when you installed a R3000 RCD and your problem still doesn’t go away… ** HEADDESK **

Which is probably why the meter is so expensive, because to be any good it needs to be at least True RMS. Also, it was a good excuse for me to buy a good clamp meter at the same time… and I was truly fed up with the situation by that point.

Oh I probably screwed up the terminology. Not that there is agreed-upon terminology. It is essentially an RCD with some resistance to nuisance tripping. It will ignore those little transients above 20mA and only trip if the energy value exceeds something dangerous.

Now for the people still reading these essays… if you’re going to make a study of what RCD to buy, you are going to find out that RCDs come in types. Type-AC is the cheap one which is really no good for a modern home anymore. Type-A is what you should use at least.

But then you also get type-S (with a delay, not allowed for use in a house, for selectivity in larger installations).

You get type type F, which is a good one to have if you have deep pockets. And you get type B which is essentially unaffordable.

But the good news is that many manufacturers make transient/impact/surge/whatever resistant RCD’s within the type-A bracket.

Hager calls theirs “HI” (high immunity). ABB calls it “type A-APR”. Gewiss calls it “Impulse Resistant”, type IR. I had to look all of these up, because I know they exist, but I forget the names… hence me butchering the term. I was probably thinking of Gewiss’s name for it.


Thanks for the explanation, I will make sure that sparkie does the thorough testing.

A ‘quick’ test to confirm this would be to remove these devices before loadshedding kicks in. That’ll be things like the TV & Fridge safe devices, multiplugs with surge protection etc. Just unplug them from the wall.

That should be an easy test to see if it is in fact standing leakage.

I guess you should technically just remove the surge protector, but keep the TV etc. connected, but I don’t want to say that and then there’s a surge or something.

1 Like

Do you have any of these in your home?

If you do, please start by removing them?

The following can also play a role. Old Alarm transformers are also well known to cause this.

Basically every thing with build in surge protection will contribute to your current situation.

Lastly, it might just happen that your earth leakage is on its way out and that its sensitivity is not on 30ma anymore. I do respect @plonkster previous experience with the earth leakage and that buying an very expensive RCD fixed hie situation but before I even suggest that solution we try and eliminate all of the above first. 95% of the time its one of the devices above, 4% sensitivity issues and only 1% required that expensive RCD as we could not fine any other reasons.

Those devices are usually just a bunch of MOVs (metal oxide varistors), usually from both neutral and live to earth, sometimes between L and N. They do leak below the rated voltage, which is why some of the more expensive arrestors also have a gapped device in series, which does not leak. MOVs are also found in many appliances for protection (but also as warranty indicators). Fortunately they are usually only L-N. I would remove these devices and just have one of these on the incoming line (after the main MCB) 900454 | TYPE 2 10kA 340V 2P 1PH+N SPD | N - Safety : Surge Protection Devices (SPDs) | ElectroMechanica (this is the cheapest spark gap plus varistor arrestor that I know of)

Then the other cause of leakage is the caps in the input filter of most SMPSs. Those are not easily eliminated, and the only solution is a better RCD, or more RCDs with fewer circuits on each (which is probably why RCBOs are becoming popular in other parts of the world.)

where we cant remove the device we simply remove the MOV’s. Small side cutter = problem solved. Cutting out the surge protection has no effect on the operation of the rest of the circuit.

I read half as usual… I only saw now what you were saying, and i agree, removing the SMPS is not easy… and I agree.

So we do Power Stations. And as Jaco state above. We also experiance that surge protectors cause an overload on the setup. Almost immediately, even if you only plug in an 8w Led bulb.

I’m seriously considering one of these: