RCD trip on power off

[Ddeclercq]

Hi Guys

 

For years I have had trouble with my RCD tripping when grid lower is turned off. This happens when the power is turned off and not when. It is restored. It also only happen if it is turned off on the grid, I have a breakout box with a 2 pole isolator before the main db board which houses the main breaker and RCD. Can turn the power of as many times as I like there and no problems. But every other time from the grid it trips.

 

I recently installed an Inverter and it still trips when. The grid goes down. However switching of in the breakout box just before loadshedding and no tripping

 

I have had the earthling on the installation tested as that seems to be the 1st port of call and it all comes back good. And from my limited troubleshooting if removing the grid removes the problem it is unlikely to be on my side? Also because it happens before power is restored and even when the inverter should keep everything running I can't see how it can be in rush current?

 

Any ideas on what it could be or how I can fix it would be greatly appreciatetd 

[___]

So to be clear, the RCD trips in the death-throws of a grid outage, just as the network de=-energises, but it remains on if you disconnect it yourself?

Do you have a 3-phase supply and 3-phase loads?

What kind of earthing system is in place, TN-S, or TN-C-S?

Man, I would love to put some kind of data-logger on that to capture the moment. Rcds trip because more energy is going through the one leg than comes back through the other one. This would imply that some sort of leak develops at precisely the moment the grid fails (but only if you are connected to it).

One possible way this could happen, is if a grid outage causes a momentary rise in voltage, causing a surge arrestor in an appliance to start conducting. This might shunt just enough current to earth to imbalance an RCD.

For example, here are some MOVs in my garage door opener that trips my RCD on power return (14d431k). There are three of them, one across live and neutral, and one each from L/N to earth.

This MOV has an RMS berakdown voltage of 275V. That's where it starts conducting.

So one possible solution that is not too expensive, is to put another surge arrestor BEFORE the RCD.

Disclaimer: I am not an electrician. I'm guessing. And I have some experience with these pesky little devices.

[Ddeclercq]

Hi, thanks a lot for the reply

Yes, as the grid de-energizes. If I disconnect from the grid, before they de-energize it then it remains on

No, only single phase

To be honest I am not sure how to tell when earthing system is in place, but I do think it is TN-S as I have separate conductors for L, N, and E but could be wrong here

Yep, I wish I could see the data too, as tripping when it powers on makes sense to me, but when it powers down just boggles my mind

I found some mention or surge arrestors before but I have had this happen when pretty much everything in the house was powered off. Except the Fridge and 1 light and it still happened. And it has been doing it before they even had surge arrestors at all

I do have 2 surge arrestors in the DB now as part of the Inveter install, one before and after the supply to the invertor, which are both before the SPD. Not sure if it would be worth moving them onto the SPD itself

I was also thinking of adding a RCD on the supply from the grid, the idea being that if I could get it to trip instead of the DB, at least the inverter would keep the important bits powered, so I would have something like this:
Grid -> 2P Breaker -> #RCD -> House Main Breaker -> Breaker to Invertor -> Invertor -> Breaker from Invertor -> *RCD -> Load
* Is the one that currently trips, was tripping even before the inverter installation

# New one I am thinking of adding

With House main breaker having the non-inverter loads such as the stove on it which is never on when we have loadshedding anyway. 

Was also thinking maybe a less sensitive RCD could possibly also help?

[___]

38 minutes ago, Ddeclercq said:

Was also thinking maybe a less sensitive RCD could possibly also help?

You're not allowed to install anything less sensitive than 30mA if it feeds sockets. You do however get 30mA RCDs that have some nuisance tripping immunity, but they are expensive. I'm talking 4k expensive for the cheapest one, and even then it is not guaranteed to solve the problem (cause you first need to identify the problem).

[Ddeclercq]

Ah, Thanks a lot for pointing that out, would have made a fool of myself trying to find one lol.

Best the local electricians has been able to recommend was to try a Merlin Gerin RCD as the ones I have might be too sensitive if I have a lot of electronic load in the house. 

They tested the earth again today and everything checks out fine, checked the resistance on earth all the way into the street supply cabinet and that also checks out. So at this point I cant think of anything else to try. Kinda hoping the builders put in a cheap RCD and me replacing it with the same brand/model didnt make any difference and maybe a better switch might work.

All that being said, I would love to really be able to analyze and really get to the bottom of the problem than just spending $$$ on the symptoms and hoping for the best

[NigelL]

I suggest that you verify the following two things:

1. The voltage between the incoming mains Neutral wire and Earth is close to zero. 
2. The voltage between inverter output Neutral and Earth is close to zero. Test this with Mains connected and also with Mains disconnected/Loadshedding.

If there is any significant Neutral-earth voltage, or change in voltage as the inverter kicks in, this could point to the source of the problem.

Edited December 11, 2019 by NigelL

[Ddeclercq]

Cool, let me test those 1st and see.

I was having this problem before I had the Inveter installed which is why I didn't immediately think it would be the problem, but I do think it could make the problem worse still?

If I do get a difference in voltage, what does that point too? Bad Inverter, or look for something connected in the house causing issues?

[phil.g00]

The fridge isn't necessarily on, even though constantly plugged in and powered. It could be the thermostat has switched out the weak part of the circuit.

A motor is an inductive load, that can create a high back emf as the magnetic field collapses. This can break down weak insulation.

Disconnect the fridge's earth wire  and leave the fridge door open, (so the compressor is running) when you're due to be load shed. See if it still happens.

[___]

One other thing I can think of, if you have a good multimeter that can hold the maximum value (eg a Fluke... if you can afford one), you could put that across the incoming line when the power fails and see what kind of peak it registers.

[Ddeclercq]

I can just completely plug it out, seeing as you have load shedding so often lately

My expectation was that the inverter would take over the load when the grid goes down so nothing strange should happen as it takes over the load? but that is just a layman's understanding

Also the fridge has been changed as the old one dies during the previous round of loadshedding, but the tripping issue has outlasted it sadly

[Ddeclercq]

nope, dont have such a good meter, but it might be a good excuse to get one  

[NigelL]

22 minutes ago, Ddeclercq said:

If I do get a difference in voltage, what does that point too? Bad Inverter, or look for something connected in the house causing issues?

A high Neutral-Earth voltage, on incoming mains, suggests a fault in the Protective Earth-Neutral (PEN) bond, or some faulty wiring between the incoming mains and the main DB.

A high Neutral-Earth voltage, on the Inverter output, points to a wiring problem between the Inverter and the backup circuits or an internal fault with the Earth Relay.

What brand Inverter are you using?

[phil.g00]

8 minutes ago, Ddeclercq said:

I can just completely plug it out, seeing as you have load shedding so often lately

Well, as this initiates the tripping, I wouldn't waste an opportunity rule things out systematically.

14 minutes ago, Ddeclercq said:

Also the fridge has been changed as the old one dies during the previous round of loadshedding, but the tripping issue has outlasted it sadly

If load-shedding knackered the first fridge, it could make another one faulty as well.

[Ddeclercq]

Ah ok, thanks. 
So on my Multimeter I get I get 0 between all Netural and earth, Main incoming/To Inveter, from Invete. Disconnecting from the grid changes that to 60V between all of the N and Earth

I am using a GoodWe ES 5038D inverter

[Ddeclercq]

• Report post
•  
•  

Posted 17 minutes ago

  32 minutes ago, Ddeclercq said:

I can just completely plug it out, seeing as you have load shedding so often lately

Well, as this initiates the tripping, I wouldn't waste an opportunity rule things out systematically.

  32 minutes ago, Ddeclercq said:

Also the fridge has been changed as the old one dies during the previous round of loadshedding, but the tripping issue has outlasted it sadly

If load-shedding knackered the first fridge, it could make another one faulty as well.

 

Good point, will try that then

Edited December 11, 2019 by Ddeclercq

[NigelL]

10 minutes ago, Ddeclercq said:

Disconnecting from the grid changes that to 60V between all of the N and Earth

I am using a GoodWe ES 5038D inverter

I am not familiar with the GoodWe inverters, but this does not look correct. I suspect that you will solve the tripping problem if you can work out why you have a high voltage between Earth and Neutral, when disconnected  from the grid.

[Ddeclercq]

My next test when the wife is done using power is to disconnect the inverter and check if it still happens. What concerns me is this is an issue I had before I had the inverter installed, so would be nice to rule in/out the inverter. 

Is it just me, or is it weird that it ends up being 1/4 of the voltage, would assume some sort of transformer would be involved in causing this then?

[NigelL]

This voltage is most likely due to the input-filter capacitance of some switching power-supplies connected to the output of the inverter. This also points to a possible mechanism which causes the tripping.

RCDs are designed to trip if there is more than a 30mA difference in current between Live and Neutral. There is a fair tolerance on this level, so one can get RCDs that will trip at say 20mA (your electrician should have a meter to measure the trip threshold). 

When the Neutral-Earth voltage jumps from 0V to 60V, as the grid is disconnected, this could result in an additional small current between Neutral and Earth - which may just push you over the limit for the RCD.

The tripping problem could also be a combination of factors - i.e. an overly sensitive RCD, a high residual earth leakage from your appliances (especially "surge-protector" plugs and anything with a heating element) and the problem of the floating Neutral on the output of your Inverter.

I recommend contacting someone with experience in wiring the Goodwe Inverter.

[Ddeclercq]

11 minutes ago, NigelL said:

This voltage is most likely due to the input-filter capacitance of some switching power-supplies connected to the output of the inverter. This also points to a possible mechanism which causes the tripping.

RCDs are designed to trip if there is more than a 30mA difference in current between Live and Neutral. There is a fair tolerance on this level, so one can get RCDs that will trip at say 20mA (your electrician should have a meter to measure the trip threshold). 

When the Neutral-Earth voltage jumps from 0V to 60V, as the grid is disconnected, this could result in an additional small current between Neutral and Earth - which may just push you over the limit for the RCD.

The tripping problem could also be a combination of factors - i.e. an overly sensitive RCD, a high residual earth leakage from your appliances (especially "surge-protector" plugs and anything with a heating element) and the problem of the floating Neutral on the output of your Inverter.

I recommend contacting someone with experience in wiring the Goodwe Inverter.

Thanks Nigel. I think you are correct around the inverter. Some more testing showed the inverter on, but grid disconnected ~20V between N and E. As I power on the loads one by one it gradually grows to ~40V. At any time if I reconnect the grid it goes back to 0V immediately. 

Will drop the supplier an email and see if I can get more info from them

[Ddeclercq]

29 minutes ago, NigelL said:

This voltage is most likely due to the input-filter capacitance of some switching power-supplies connected to the output of the inverter. This also points to a possible mechanism which causes the tripping.

RCDs are designed to trip if there is more than a 30mA difference in current between Live and Neutral. There is a fair tolerance on this level, so one can get RCDs that will trip at say 20mA (your electrician should have a meter to measure the trip threshold). 

When the Neutral-Earth voltage jumps from 0V to 60V, as the grid is disconnected, this could result in an additional small current between Neutral and Earth - which may just push you over the limit for the RCD.

The tripping problem could also be a combination of factors - i.e. an overly sensitive RCD, a high residual earth leakage from your appliances (especially "surge-protector" plugs and anything with a heating element) and the problem of the floating Neutral on the output of your Inverter.

I recommend contacting someone with experience in wiring the Goodwe Inverter.

On last question. So when the inverter was installed 2 SPDs was installed in the DB on the live wire both before and after the inverter. With surge-protector plugs being a general suspect for all these things, would it be worth removing all of them on the pugs in the house now. Am I right in saying with the SPDs in the DB they don't add any extra protection and in fact can now cause more trouble than they are worth?

[___]

1 hour ago, Ddeclercq said:

Disconnecting from the grid changes that to 60V between all of the N and Earth

That means the inverter does not bond the neutral and earth (it should). I don't know enough about the Goodwe to say how to fix it. On Victron units there is a bonding relay and the inverter creates a bond when the grid is disconnected.

[___]

Just now, Ddeclercq said:

Am I right in saying with the SPDs in the DB they don't add any extra protection and in fact can now cause more trouble than they are worth?

An SPD on the input side is a good thing... but it has to go BEFORE the RCD (because the create a small amount of earth leakage and there is no point in adding that on to your other standing losses). An SPD on the output is pointless, at least IMHO.

[Ddeclercq]

1 minute ago, plonkster said:

That means the inverter does not bond the neutral and earth (it should). I don't know enough about the Goodwe to say how to fix it. On Victron units there is a bonding relay and the inverter creates a bond when the grid is disconnected.

The GoodWe is nowhere near the same league as Victron so I suspect that is something they skimped on... Either way I have send the question trough and hope to get some sort of feedback from them on that one

[___]

Just now, Ddeclercq said:

something they skimped on

I doubt it. The Goodwe isn't THAT bad. Must be a configuration issue.

[Ddeclercq]

Just now, plonkster said:

An SPD on the input side is a good thing... but it has to go BEFORE the RCD (because the create a small amount of earth leakage and there is no point in adding that on to your other standing losses). An SPD on the output is pointless, at least IMHO.

The 1st one is before the RCD as it is on the live on its way to the inverter. The 2nd one is on the way back, back but also before the RCD so my assumption would be while its a complete waste, as both of them is before the RCD they should not really cause any hassles as far as the tripping is concerned? Our estate really does have a ton of requirements which IMHO is complete overkill, but that is not a battle I will win or can even fight