Hi,

I have a rebranded Voltronic Axpert MAX 7200 with 2x Pylontech US3000C which has been running for a few months now.
I notice how a software-triggered switch from SUB to SBU source priority, effectively resulting in moving from Grid Mode to Solar/Battery Mode, will sometimes result in the tripping of one of the RCBOs on the output side.
The tripping RCBO has a cheap type-AC RCD with 30mA current threshold.
In particular it's the RCD part of the combined device so presumably due to a detected earth leakage.

However, I haven't been able to reproduce it reliably so it's hard to tell if any corrective action will result in an improvement.
What I've recently realized though is that I've mostly seen this happen at night.
As far as I can tell, there are two peculiar conditions to this scenario in SUB:
1. There is no PV power. This means the inverter is essentially not feeding any power until it switches to SBU (where it will have to make a sudden spike in power production).
2. There will be lights (mostly neon and LEDs with their SMPS'es) turned on in the house

I've been reading the thread where @BritishRacingGreen and @Coulomb shared their reverse engineering efforts on such machines (for which I'm extremely grateful BTW) and came to realize how during that kind of switching the output neutral will get disconnected from the input (grid N) and bound to local earthing. I understand there might be some parasitic effects during this switchover but I'm nowhere near understanding the whole physics here.

Any idea what might be causing this issue and how to solve it?

1 hour ago, iurly said:

Hi,

I have a rebranded Voltronic Axpert MAX 7200 with 2x Pylontech US3000C which has been running for a few months now.
I notice how a software-triggered switch from SUB to SBU source priority, effectively resulting in moving from Grid Mode to Solar/Battery Mode, will sometimes result in the tripping of one of the RCBOs on the output side.
The tripping RCBO has a cheap type-AC RCD with 30mA current threshold.
In particular it's the RCD part of the combined device so presumably due to a detected earth leakage.

However, I haven't been able to reproduce it reliably so it's hard to tell if any corrective action will result in an improvement.
What I've recently realized though is that I've mostly seen this happen at night.
As far as I can tell, there are two peculiar conditions to this scenario in SUB:
1. There is no PV power. This means the inverter is essentially not feeding any power until it switches to SBU (where it will have to make a sudden spike in power production).
2. There will be lights (mostly neon and LEDs with their SMPS'es) turned on in the house

I've been reading the thread where @BritishRacingGreen and @Coulomb shared their reverse engineering efforts on such machines (for which I'm extremely grateful BTW) and came to realize how during that kind of switching the output neutral will get disconnected from the input (grid N) and bound to local earthing. I understand there might be some parasitic effects during this switchover but I'm nowhere near understanding the whole physics here.

Any idea what might be causing this issue and how to solve it?

Hi @iurly welcome to the forum. Its always nice if someone gives a short but concise description of a failure mode and its context.

I myself have just learned on my MAX that there is a different behaviour when changing from SUB to SBU and vice versa than I thought!  After a while the relays click, suggesting the output neatral is disconnected from grid neatral and bonded to earth locally, even with grid available (SUB  - > SBU). This is interesting as I thought the grid will still pass thru to entertain ac blending. 

So this transition from output neatral bonding might indeed cause metastable conditions as you describe. In my case the transition is seamless, ie the output RCD remains happy. 

Maybe, just maybe, you have a bonding issue whereby the output neatral somehow loses its bonding inside the machine. 

You can verify this by measuring  the ac voltage between output neatral and earth while transitioning takes places. It should be 0V, with maybe a very small fuctutation while transitioning. But if there is a high voltage (say 90vac) when in SBU, then there is a problem.

 

 

Edited April 2, 20232 yr by BritishRacingGreen

1 hour ago, BritishRacingGreen said:

I myself have just learned on my MAX that there is a different behaviour when changing from SUB to SBU and vice versa than I thought!  After a while the relays click, suggesting the output neatral is disconnected from grid neatral and bonded to earth locally, even with grid available (SUB  - > SBU). This is interesting as I thought the grid will still pass thru to entertain ac blending. 

Actually my initial thought was quite the opposite. I was expecting the grid to be always isolated from output (even with AC blending!) resulting in an IT system. I suppose that would be totally impossible for a transformerless inverter.

Then I assumed I would at least get an IT system when running in island mode (with AC input entirely disconnected), and was quite surprised to realize my RCD and my phase detector were still working in island mode. I came to suspect the only way to achieve that would be by connecting the output neutral to local ground and your diagrams finally confirmed this thought.

1 hour ago, BritishRacingGreen said:

You can verify this by measuring  the ac voltage between output neatral and earth while transitioning takes places. It should be 0V, with maybe a very small fuctutation while transitioning. But if there is a high voltage (say 90vac) when in SBU, then there is a problem.

Not quite sure I understand your point.

If you're suggesting I should take a reading while transitiong, I'm not sure how much I can trust my cheap multimeter.

If you're suggesting I should look at the steady value, I've done this on day 1 and I had a puzzling (or reassuring?) 0V.

Is it possible I do have a problem with the wiring (e.g. small leakages or suboptimal earthing) that only gets exhacerbated by the inverter transition?

One other thing worth noticing is I've only seen this happen when AC blending is in place but essentially idle (i.e. no PV production, as I understand it's not possible to have AC blending from batteries), AND the switch is triggered by software. I've never seen this happening when the inverter automatically switches to island mode when battery SoC goes above 30%.

Same thing when at night (and that's a different story) the inverter decides to switch to grid mode after load exceeds 2kW and then switches back to island mode once the load goes back under this empirical threshold of 2kW.

So I also wonder if it's got something to do with the relative timing of how the relays switch when the transition is triggered by software (solar assistant in my case).

4 hours ago, iurly said:

Is it possible I do have a problem with the wiring (e.g. small leakages or suboptimal earthing) that only gets exacerbated by the inverter transition?

I think so. If you end up with multiple connections from neutral to earth, then some of the neutral current flows through the earth conductors, meaning that neutral and live no longer completely cancel, and that trips the RCD.

These things can be devilish to find.

6 hours ago, Coulomb said:

I think so. If you end up with multiple connections from neutral to earth, then some of the neutral current flows through the earth conductors, meaning that neutral and live no longer completely cancel, and that trips the RCD.

These things can be devilish to find.

Uhm... so you think there might be some leakage from neutral to earth? I suppose not a straight connection as in that case the RCD would trip continuously, whereas in my case it only happens during transitions.

I was more thinking of some leakage between line and earth. So perhaps next thing I'm going to try would be to disconnect the output neutral from the RCD (using just grid as the power source) and see if that continuously trips the RCD.

This has happened to me already with a submersible pump -- when I accidentally disconnected the neutral from the pump, the RCD would trip. A few weeks later, the pump was dead. I suppose it was not that single disconnection that broke it, but it was on its way to death already... Please do correct me if I'm entirely wrong here.

18 hours ago, iurly said:

Uhm... so you think there might be some leakage from neutral to earth?

It's more likely that you have more than one connection from neutral to earth. That would be why it triggers when you are in battery mode, as the inverter provides the extra connection (supposed to be the only connection) from neutral to earth.

2 minutes ago, Coulomb said:

It's more likely that you have more than one connection from neutral to earth. That would be why it triggers when you are in battery mode, as the inverter provides the extra connection (supposed to be the only connection) from neutral to earth.

Please notice the triggering is a one-off during transition (and only under certain circumstances). After switching to battery mode, I can just rearm the RCD and everything works smoothly afterwards.

So a permanent connection between neutral and earth is very unlikely imho.

7 minutes ago, iurly said:

So a permanent connection between neutral and earth is very unlikely imho.

OK, fair point.

The installation of a 30mA RCD before any inverter is not recommend,since nearly every On or Off grid is not isolated.

(This cause continues small currents to earth )

Anything before the Inverter it just makes no sense for user protection, its just to prevent fire if any isolation fails.

 

The normal way to install hybrid inverter is:

This is for the input of the Inverter:

300mA( 100mA is also ok on small ones) RCD before the Inverter .

At the output there you can use the 30mA RCD  for user protection.

 

 

If you got a standard On grid :

 

Take a separate cable to the Inverter with overcurrent protection and RCD with 100mA or 300mA .( At this wire no plug or connection to any other thing )

 

 

 

By the way:

RCD typ AC is not recommend for Hightech inverter since there is a chance of pulsed and dc current to earth.

19 minutes ago, Georg594 said:

The installation of a 30mA RCD before any inverter is not recommend,since nearly every On or Off grid is not isolated.

(This cause continues small currents to earth )

I have one (actually, two) 300mA RCDs from the meter to inverter IN.

If I get this right, the first one (just next to the energy meter) was required by local regulations since there's a separate grid-tie inverter also bound to the same grid connection.

19 minutes ago, Georg594 said:

Anything before the Inverter it just makes no sense for user protection, its just to prevent fire if any isolation fails.

Are you suggesting a plain MCB should do the trick here? In the back of my mind I kind of had a gut feeling an RCD would in some way protect the equipment in case of a thunderstorm, but I guess that's just a myth, is that right? I did install a type-2 SPD right before the inverter so perhaps that's enough?

19 minutes ago, Georg594 said:

The normal way to install hybrid inverter is:

This is for the input of the Inverter:

300mA( 100mA is also ok on small ones) RCD before the Inverter .

OK so I guess we're covered here, unless having two is a problem.

I'm not quite sure about the type but I suspect they're both type-AC.

Notice I've also seen these two RCDs trip when switching mode (and when they do, they BOTH trip) from time to time, though not as frequently as the one between inverter output and the load.

19 minutes ago, Georg594 said:

At the output there you can use the 30mA RCD  for user protection.

Which is what I have.

19 minutes ago, Georg594 said:

If you got a standard On grid :

Take a separate cable to the Inverter with overcurrent protection and RCD with 100mA or 300mA .( At this wire no plug or connection to any other thing )

Mine is a grid-assisted (there's also a grid-tie but that's wired just like you say).

19 minutes ago, Georg594 said:

By the way:

RCD typ AC is not recommend for Hightech inverter since there is a chance of pulsed and dc current to earth.

What do you mean by Hightech inverter? Transformerless?

I've been reading all over the places about RCDs but I haven't been able to fully understand the implications of using a different type.

Would a higher-class RCD (e.g. type A, type F or B ) provide extra resilience or extra protection, or both?

My understanding was that (by definition) higher class RCDs would be able to detect additional type of faults (and thereby provide extra protection) and that would suggest that would make them more likely to trip.

Other places I have also been reading they additionally provide extra resilience to noise thereby suggesting it should make tripping less likely (i.e. less false positives).

So any explanation would be much appreciated here 🙂 .

Thanks!

On 2023/04/03 at 6:21 PM, Georg594 said:

The normal way to install hybrid inverter is:

This is for the input of the Inverter:

300mA( 100mA is also ok on small ones) RCD before the Inverter .

At the output there you can use the 30mA RCD  for user protection.

Where can one find a 300mA RCD for a Samite DB?

Did you ever solve this issue? I had this for a few months then it went away - now its back 9 months later.

It only happens when system goes from battery to grid/charging and sometimes in reverse.

 

On 2024/06/14 at 11:55 AM, Sysadmin said:

Did you ever solve this issue? I had this for a few months then it went away - now its back 9 months later.

It only happens when system goes from battery to grid/charging and sometimes in reverse.

 

No, I could not figure it out. I actually had two different issues:

1) The RCBO on the output side would occasionally trip upon on-grid/off-grid transitions. This issue was quite rare and it seems to have gone away at some point. I did replace the RCBO for other reasons, so that might have something to do with it.

2) I had two cascading RCBO on the input side which would also trip under the same circumstances. I replaced the first one with a plain MCB, but the second one would still trip from time to time. So I installed an automatic rearming device (called BTicino Stop & Go, that was about a year ago), and that did the trick for me in that I don't have to worry about it anymore. I had the intention of monitoring the situation so to see if there was any recurring pattern, but the rearming is so quick (something around 1 second or less) that solar assistant doesn't even detect/record the grid outage. So I have no historical data 😞

13 hours ago, iurly said:

No, I could not figure it out. I actually had two different issues:

1) The RCBO on the output side would occasionally trip upon on-grid/off-grid transitions. This issue was quite rare and it seems to have gone away at some point. I did replace the RCBO for other reasons, so that might have something to do with it.

2) I had two cascading RCBO on the input side which would also trip under the same circumstances. I replaced the first one with a plain MCB, but the second one would still trip from time to time. So I installed an automatic rearming device (called BTicino Stop & Go, that was about a year ago), and that did the trick for me in that I don't have to worry about it anymore. I had the intention of monitoring the situation so to see if there was any recurring pattern, but the rearming is so quick (something around 1 second or less) that solar assistant doesn't even detect/record the grid outage. So I have no historical data 😞

Replacing a RCBO with a auto resetting Mcb(BTicino Stop & Go) means you now have no residual current protection and electrical shock to humans are a possibility! This may have solved your tripping issue but you are now exposed to earth fault currents that to my understanding would not be detected by the BTicino Stop & Go. RCBO combines the functionality of an MCB and RCD/RCCB. When there is a current leakage, the RCBO trips the entire circuit. Consequently, internal magnetic/thermal circuit breaker components can trip the electronic device when the circuit is overloaded. Not sure which country you are residing but in SA this device would not be legal if not used in conjunction with a RCD.