Earth leakage sometimes trip during switchover to inverter during load shedding


My earth leakage sometimes trip during load shedding, when the inverter kicks in.
For the last 6 times that we experienced load shedding it only happened twice.

I have a Deye 8kw inverter installed with 2x5kw HinaESS batteries. It was installed by a qualified electrician.
My DB board is configured as follows.
Main Switch, Geyser, Stove, Sub-DB (inverter) on the left of the earth leakage, lights and plugs on the right of the earth leakage.
Before the installation my lights weren’t on the earth leakage and only my plugs were.
Before the inverter was installed, I never had this problem during or after load shedding.

I tried to simulate the problem by switching on all my lights and trip the main switch (simulate load shedding). I even experimented by adding a 60W bulb to my external light, possibly to assess if it would contribute to the particular issue. I have done this multiple times but I’m unable to trip the earth leakage.
The installer had another look at the installation and they are confident that the DB board is correct. They even suggested that I replace my earth leakage, but wouldn’t it trip all the time if it was faulty (even without my simulations)?

Wouldn’t a light that is not correctly earthed, trip all the time?
I’m not sure what to do next, especially because the installer says everything is fine.

Did your installer install a Neutral Earthing relay?

What brand Earth Leakage do you have?

1 Like

What you are referring to is known as “nuisance tripping”.

An Earth leakage doesn’t trip at exactly 30mA. It MUST trip at 30mA, but it may trip before then, and most of them will trip somewhere between 20mA and 25mA. It should not however trip before 15mA.

If you have a leak that is around 15mA to 20mA, you will begin to get nuisance tripping. Sometimes it trips, sometimes not.

During changeover (connecting or disconnecting from the grid), there is always some extra disturbances, little’s bits of charge escaping to ground, so this nuisance tripping is more prevalent during changeover: Either when the grid is lost, or when the grid returns.

However you look at it, the basic answer remains the same. You have a leak. You have to find it.

There are two kinds of leaks, good ones, and bad ones.

A “good” leak may be caused by an appliance with some sort of EMI filter in it. Typically modern electronic appliances. Computer power supplies and the like. They all leak a small amount to earth. There is also always a little bit of leakage just in the wiring, but it should be very small. Add enough good small leaks together, and they put you over the limit.

A “bad” leak is where you have an exposed wire of some kind, moisture, that sort of thing. Even a bad leak could leak little enough that the RCD doesn’t trip, but because it increases the leakage, it puts you within that nuisance tripping bracket.

Generally I recommend a two-part approach. First, eliminate the bad kind of leak. For this, you need to hire a sparky, and he needs to put an insulation tester on your wiring. Make sure you don’t have a leak in the installation itself.

For the good kind of leak, my advice is to measure it. Your sparky will need a clamp meter that is capable of reading this low. By clamping it over the wires entering the RCD, the differential can be measured. If it is close to 15mA, and unplugging appliances don’t drop it significantly (eg, you find one appliance with a very large leak), then you need to split the house over two RCDs.

There isn’t an easy answer unfortunately. It comes down to finding the leak, and then deciding how to deal with it.

Hello Jaco

Please see attach screenshots.
Main DB

Sub DB

I would strongly suggest you though that Cheapie out and replace it with a CBI.

Thanks Jaco.
Can you please be more specific which one. Unfortunately I don’t have any experience or knowledge of these things.

The QA17C

Thank you. Just to confirm, replace the Main DB one with the QA17C.

That is correct. The cheaper ones like yours, Swan, lear and a few more Minirail RCD’s is problematic. Even more so the cheaper din rail RCD’s. The QA17 series hardly ever give issues on a solar system and when it does, there is a definite reason for it.

1 Like

Agree with Jaco. Chuck the Schenker, put in a CBI QA17C. Just note, it is about three times the price. For good reason too.

That is a general note on quality. It may not solve the issue. If it fails to solve the issue, then you at least know it is not because of a crap RCD.

(I’m actually using a Lear RCD on my EV charger… they really are cheap, overload plus residual protection in a device costing R300, makes you think).

I’d chuck the ACDC changeover as well and put in a Hager SF263, but that has nothing to do with your present issues :slight_smile:

If you are comfortable using a multimeter there is a quick test you can do. Open the main bord, switch off all the breakers, set meter to Ohms, (Resistor) measure the resistance between earth and neutral, should the reading increase its a sign of a leakage or capacitance in the circuit. Insulation breakdown as well as moisture in the circuit acts the same as a capacitor and as it charges up, the resistance reading will increase.


Thank you.

I have a concern that it might be my outside lights (4 lights) causing the issue. Last night at 22:00 when load shedding kicked in, the load was minimal (already in bed :slight_smile: ) and the outside lights were on and the earth leakage tripped.
But with that said, I have had load shedding at night with outside, inside lights on, tumble dryer and TV on, with the earth leakage not tripping when load shedding kicked in.

If it is, an electrician should be able to test the circuit with an insulation tester. If there is a leakage on that circuit, he will see it right away.

Yup. That’s your typical capacitor check, it starts at a low resistance and then increases as it charges. But in my experience, for smaller capacitances, your average electronic DMM is not going to be fast enough to catch it. The old needle style was quite nice, you’d see the meter swing up and then slowly drop down, but even then, a 220nF capacitor (typically used in capacitove dropper PSUs) will leak enough to trip an RCD, but may not show up on a DMM. You may get lucky… or not :slight_smile:

The intermittency is also quite literally timing-related. Depending on exactly WHERE on the AC sine wave the trip happens, the amount of current leaking through a well-intentioned capacitor (meant to suppress noise) will be different.

1 Like

Thank you Jaco and Izak.

I’ll get someone out to have a look. Was just afraid that it might be the inverter itself that is causing the issue.

I test mine with a DMM. Shows insulation breakdowns as well as moisture

The inverter can play a role, but normally it will just amplify and already existing issue. Opening and closing of a neutral earthing relay for example will allow the caps in a surge arrestors device to discharge and charge back up the moment it closes again or the moment it reconnects to the grid. To many surge arrestors circuits in your home can also contribute to the problem.


You have a decent DMM though :slight_smile:

Only work with Flukes. The one in my kit is the 287.

1 Like

My Brymen meter also seems pretty fast enough to pick up an effective capacitance of 300nF, but 2.2nF it won’t pick up. I’m not much of an expert in these meters (@_a_a_a ?), but I believe the kind of shunt capacitor you’d use to filter HF noise from a typical SMPS, is on the smaller end. If you have enough of them in a house to trip an RCD, a good meter will probably pick them up.

Edit: Also tested a 22nF, can pick that one up. So… totally concur. A resistance test should pick up shunt caps in the installation.