question

Jaco Reinecke avatar image
Jaco Reinecke asked

Error #8 Ground relay test failed during active load shedding

The twist here is that the error ONLY happens when the grid is OFF, during the 2-4 hours of load shedding.

Syncing to the grid works perfectly every time. Never a problem when the grid is on.

Equipment:
MPII 5kva, installed for a long time, never a problem.
Latest firmware on the inverter, MPPT's and the Cerbo.
Carlo Gavazzi, so not using AC_Out1.
Grid-tied.

Tests done:
Manually testing syncing syncs perfectly with the grid.
Also disconnecting the inverter, L&N from the DB during load-shedding, the error still appears.
Checked all loads.
No DB tripping either.

Inverter output during load-shedding event:
Live /Neutral = 231v
Neutral /Earth = 0v
Live /Earth = 231v

I repeat, never a problem when the grid is on, or when it syncs with the grid, just when the grid is off due to load-shedding, that the error randomly appears

See 2022-09-19 at 2:14 am, not a problem during that load shedding event.
The one starting on 2022-09-189 at 10:16 am, had the most errors ever.

1663678097962.png

Still getting errors, just not as frequent, since I am switching off the L&N between AC_In and DB.

1663679063344.png

The actual Error:

1663678108999.png

The volts on AC-Out1 drop for no reason, eliminated loads already. Just happens suddenly, and ONLY when the grid is off.

1663678119211.png

Batteries are normally charged, volts are perfect, BMS has no errors or warnings, checked all the battery connections.

Is this a unit problem/relay?

Does anyone have any more suggestions I can check, please?
This is not a normal error, no-one seems to have reported this ever that I could find.

To take the inverter in, during the current load-shedding woes, which will take weeks to check/repair assuming there is a fault, would be a complete disaster. Am also told it would be a repair, not a swap.

Multiplus-IIerrorRelayerror 11
1663678097962.png (154.3 KiB)
1663678108999.png (51.0 KiB)
1663678119211.png (155.3 KiB)
1663679063344.png (37.4 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

3 Answers
nickdb avatar image
nickdb answered ·

When the grid is off the relay should be closed and this is reporting open.

I had the opposite issue. When grid returned the relay test would fail as closed when it should be open. In this instance a neutral had broken and gone to ground, correcting this solved the problem - everything worked fine until that point.

If this worked perfectly before, and you haven't made changes like grid codes or having enabled ESS around the time this started, it could just be a bad relay.

Most commonly it seems to be wiring related.

3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Jaco Reinecke avatar image Jaco Reinecke commented ·

Thanks @nickdb

Having pulled just about all of it apart, someone today mentioned in passing how they had drama with a solar inverter and a UPS in a larger business. So not cheap stuff, and professionally installed.

That made me wonder. Can a UPS, and I tried 3 types, sine wave/modified sine wave/Online Double Conversion, when in close proximity (very short cable length) to a Victron inverter, cause harmonic distortion on the input side of the UPS, hence the output side of the Victron inverter?

Turns out it can: https://www.kohler-ups.co.uk/understanding-full-implications-harmonic-distortion/

I did move the UPS a while back to right next to the inverter, with no issues, until now that we got these severe load-shedding woes, 2.5 - 4.5 hours, up to 3 times a day.

So earlier tonight I moved the UPS back to the furthest point on the solar critical load's circuit, where it used to be, and tonight I had no errors.

Will keep on monitoring the next few LS events. Geez, I hope it is this easy!

Why use a UPS?
Solar inverters are pass-through devices, they do not have built-in protections like UPSs have. Victron switches over fast, but sometimes we can see the lights flickering worse, it is bad.

What I learned further:
With all the power woes here, all the installers/repair centers I spoke to today, are flooded with repair backlogs and/or part shortages, on top of rush new installs. Replacement units can take weeks to arrive due to shortages from more than a month of electricity woes, on top of the supply chain challenges.

Also learned that when the grid fails, disconnect the DB (L&N) completely from the grid. This mitigates neighbors feeding back power. And when the grid comes back, wait at least 5 minutes before syncing to the grid. 300v incoming from the grid in certain grid fault conditions is happening.

Will install protections and then ponder on a backup system for my backup system. :-)

0 Likes 0 ·
nickdb avatar image nickdb ♦ Jaco Reinecke commented ·
Great, who would have thought. At least it's stable again.

With stage 5/6 it's not fun if you can't depend on the system to work.

1 Like 1 ·
Jaco Reinecke avatar image Jaco Reinecke nickdb ♦ commented ·

Was perfectly stable for 3 sessions of load shedding ... till this morning 8 - 10h30 when we had LS again, it came back with a vengeance.

Someone suggested, "switch off the panels". Surprise, the error stopped!

But, only the last 30min, too short a period. So will check again, switching the panels off, this afternoon and tomorrow morning again.

1663759929522.png



0 Likes 0 ·
1663759929522.png (115.2 KiB)
Jaco Reinecke avatar image
Jaco Reinecke answered ·

Interesting to note, I go from this, in one LS event:

1664176810155.png


To this, over multiple LS events:

1664176793517.png



The difference is the time slots of the LS. The 8 am and 10 am ones are where I get errors, if I switch off the panels during those time slots, the errors go away. Where I have no errors the panels stayed on.

Found a changeover switch that was faulty by accident, and has since been bypassed. If it was/is related, need an 8 am or 10 am LS slot to compare apples with apples.


1664176793517.png (58.7 KiB)
1664176810155.png (115.2 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Jaco Reinecke avatar image
Jaco Reinecke answered ·

Got a test unit on loan and it seems to work with me switching off the DB board to simulate load-shedding. Of course, LS was stopped when the loan unit arrived, after shcedule changes from 8 am to later times. My unit can now go in for testing. I hope they find a problem, as this has been a trying few weeks leaving everyone stumped.

Quite a challenge when one's equipment develops an error during weeks of load-shedding, with a lot of repair centers flooded and replacement units in shortage due to increased sales, especially with Europe's growing energy challenges, parts shortage, the error not "trappable" nor consistent.

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.