5kva MPII ground relay test failure during LS, not when syncing

This is a new thing.

Note: It only happens when Eskom is in active LS mode.

The first time was weeks ago early morning when I became aware of it, as the UPS, which protects selected circuits, was beeping.

At the time, according to Victron LED definitions, it is “Charging from PV - Equalization”, as these LED’s are on: Absorption, Float, and Inverter On.

The errors: Note: Last night’s LS, no issue, this morning, issue.

On the inverter:

The AC_Out volts dropping at the time:

Is this because of LS i.e. something off-site causing the problem?
Or is it some circuit that is only affected because of LS?
Or is it the inverter?

Ps. I swapped UPS’es, as a process of elimination, was still there.

That #8 is significant, somewhere and I can’t remember where ( perhaps on the Victron site) it details in what sequence the various tests are done.
Each test attracts a number.
If you seek this documentation out it’ll give you a clue as to what is wrong.

Hence me asking here, as I have scrounged the web, answers range from a faulty inverter to a bonding issue on the premises … and an external bonding issue too.

I have a TNC-S connection, as the box in the street had a problem way back when the system was originally installed.

So, recently, as I said, it only happens with active LS.

And then, when Eskom returns, absolutely no problem, clickity-clack and all is back, perfection every time.

I think you may be confusing this with VE.Bus error 11, of which step 8 is an obvious one. Step 8 means your live and neutral is swapped or something is wrong with the earth connection.

Error 8 is specific to the relay test. It expected the bond to be gone… but it was still there. Or it expected the bond to be there, but it wasn’t. Navigate into that error to see which one it is: PE2 open or PE2 closed.

Older units (which I think yours are, before 2021) checks the relay by measuring the voltage. So it closes the relay and checks for 0V. Then opens it and checks the voltage again, although I am unsure what it is looking for,

Also see this: VE.Bus Error Codes [Victron Energy]


Bugger, missed the last step:

Error: PE2 Open: Comes from the main controller, which means that the relay is measured as being open while expected to be closed.

As I said, only happens during active LS, nowhere else. Anything I can do, check?

I don’t understand what this means. Do you mean it only happens during an outage that is part of a larger load shedding campaign (which seems weird), or do you mean it only happens during a grid outage, not when you artificially create an outage by throwing the breaker?

Specifics matter… :slight_smile:

And there I thought, by repeatedly stating specifically, “it only happens during an active LS event”, would mean “an outage that is part of a larger load shedding campaign”. :grinning:

Hence me asking, wondering, and as you say, it seems “weird”.

Even downloading the data, it also reflects that. Can see nothing amiss.

The confusion come with all those large words you are adding in there. ACTIVE / CAMPAIGN / LARGER. Everything is inactive during Loadshedding, Even during campaigns. I do like larger :rofl:

I would start by measuring earth neutral and the bond between them, during non LS and then after LS…


The relay test is only done after the grid is seen as “on”. The system wont do the relay test during load-shedding. Once the power is restored, the system will perform the relay test and only after all boxes has been ticked and its happy with all the results, will it reconnect.

What I read in this is its testing for a bond between Neutral and earth on its input (Its expecting the bond) and fails the test because its not seeing such bond.

No it doesn’t. While load shedding is active, the inverter is disconnected from the grid. It happens when power comes BACK after an outage, right?

So if I understand this correctly, then if the power returns after an “external outage”, then when the inverter does its relay test prior to accepting that new power, it goes into error 8. But if you trip the breaker to cause an “internal outage”, you cannot reproduce it. That is what I understand.

If the inverter actually raises error 8 while the grid is out… well that might be interesting… then you need to get in there and find out if there is some lingering voltage or something causing it to incorrectly do a relay test when the grid has not returned.

Software people need exact steps to reproduce it. Believe me, we already did the “straightforward how we think it will be used” test when we built the thing… :slight_smile:

Nope, if Eskom comes back on, it works perfectly. Also now checked, switched off the L&N to the inverter, waited a minute, then switch it back on. ±65s later, as Jaco posted, it synced perfectly with the grid.

When Eskom is off, THAT is when it happens.

Here is two of the errors in the first post. Note: Eskom is off.

Have you not got another power source that may be feeding the MP’s input? UPS, another battery bank?

1 Like

?? maybe this. but don’t know why it would change if was working all along… (happens only during dark hours?)

Yea, that bond testing. Have to open the DB for that … and know what to do, I’m not too comfortable with that idea. Call me chicken if you want.

What I can do:
Will switch Eskom feed to the inverter off, and wait maybe 120min and see what happens.
When Eskom goes off again, I will switch the L&N between the main DB and the inverter OFF. See if that works.

Something is causing this, so far the UPS has saved the day every time.

That is why I also think, see. And Eskom is not there to bond with.


Nope. Stock standard.
UPS is on AC_OUT1, powering some selected circuits, and I have tried 3 UPS’es already.

Good thought. That fun project has been sold on, inverter too.

LS is starting at 6pm. I’m going to switch off the breaker between DB and MPII, until after LS.

I think that will eliminate any other loads coming in from the DB side.

OK hang on… it appears the bonding relay isn’t only tested as part of the grid connection relay test, but also periodically while inverting. From the above mentioned document:

During inverter operation, disconnected from mains, the ground relay is continuously tested to be closed.

It does this by testing the voltage between T and N (or on models after 2021, using an auxiliary contact), and it seems that at times it detects a voltage there, concludes the bond has failed, and switches off.

Don’t know. You probably need to escalate this through some official channel.

Edit: But before you do… latest firmware on the Multi. Always try that first.

If I look at this:

This is how I would expect it to work:
If supply is lost those two AC1 input breakers should open. And if the inverter is to continue to supply ac within 20ms they should open fast.
That is whether it’s load-shedding or your own local supply CB tripping.
At which point the Neutral - E test should be done and a voltage is expected as it is now floating,
That should not alarm as it is an expected condition.
Then the PE CB should close earthing the Neutral and another test should be done confirming 0V.
That should not alarm as it is an expected condition. ( Check that that Earth is good).

It is interesting that this happens only for load-shedding and not when the local CB is operated.
Perhaps, during load-shedding, there is contact-racing happening as the supply decays as opposed to abruptly being switched off.
This may cause the test to happen before the AC1 input CBs are actually open and fool the logic.

Cerbo is on v2.90 and the rest are up to date:

Have sent it to my dealer. I smiled, he also missed the part that Eskom being off. It is not a normal error.

I have switched the DB/inverter breaker off. I will try that for as long as I can … food-prep time and all that.

Definitely breaker off from 6 pm till 8h30 during the official LS event. That should eliminate any foreign voltages from the main DB, to the inverter.

Note: That Eskom doesn’t actually switch off the L and the N, the N is still connected to Earth at your transformer until the MP2 AC1 contactors open.