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

Inverters
41

The surge protection I use is for the fridges, this UPS, I use the Surge protection ports: https://www.apc.com/za/en/product/BR900GI/powersaving-backups-pro-900va-230v/

The reason is when Eskom goes off/comes on, I can hear the fridge motors at times. UPS’es have brownouts, overvoltage, and other protections, not grid-tied inverters.

LS again, NodeRED is off, as is the surge protection UPS disconnected.

Nirvana. :slight_smile:

Pretty sure the suppliers would first want to check the unit and test it, before any swap-outs or fixes. But I will ask.

42

The problem would be that while you have it checked out, you will be in the same boat as normal South Africans that don’t have Solar power.
Ouch! You might want to time this to right after the Koeberg unit is back on line :smiley:

43

Nope, still, doing it … breaker is off, surge protection is disconnected, NodeRED is off.
image

Have asked the sparkie to come and test the earth, and anything else he can think of, ASAP.

I’m running out of what else to check.

44

Do your main breaker actually switch both Live and Neutral? Worth checking as you may still also be connected to the utility neutral earth? Just a thing to check…

Groetnis

45

Indeed, double pole 50a breaker.

46

I’m thinking the ground relay may in fact be failing. While in islanded mode, measure from neutral to earth at any plug point, note the voltage. One would expect this to be zero (or very close, it may be a few volts if there is a large load causing a drop on the neutral conductor), but during islanded operation you won’t be running large loads.

47

Just that I have seen some terrible (no pun) wiring inside DB’s. Still worthwhile to check that both wires are switched… even if it is a 2 pole in there.

Groetnis

48

You said your munic earth is lousy, so you got a spike.
This test apparently checks for voltage between Neutral and Earth whilst it is in inverter mode.
There should be no voltage, or else the test fails.
This begs the question which of those two earthing points is it checking to?
Have you bonded those two earths together, if you haven’t there may very well be a voltage between them.
I would ask the electrician to ensure the munic earth and the spike are electrically bonded.
And then I would ask him to make it is well within specs.

49

most definitive home test for hardware issue is probably disconnect/open breakers all AC-in and AC-out. If test fails then definitely hardware problem. Not foolproof with intermittent problem though.

50

Am I doing it right?
I took a computer power cable, which can fit the multimeter leads into it.
Set MM to AC.
Measuring the house plug, off due to LS: L/N N/E L/E are all zero, as expected.

Measuring a plug that is on via inverter, currently on:
L/N = 231v
N/E = 0v
L/E = 231v

Sparkie has been asked to come pronto, will give him your message.

So that would be like there is no inverter, AC_In and AC_Out are disconnected, basically MPPT’s and batteries powering the inverter only?

Could try that IF I’m lucky and it happens with the first try.

51

Feedback I got: As dit die relay is moet hy inkom, ons maak dit hier reg. Daar is nie uitruil eenhede nie en Victron gaan jou net terug verwys na ons toe.

52

Yes, that is what you want to see. That means it is correctly bonded. Although, we already know this issue is intermittent… you’d have to catch it in the act.

I don’t think that is going to matter. The bonding relay test should succeed even if the Multi is lying on the bench with nothing attached. It checks only that there is no voltage between N and T on its output, which there shouldn’t be if the relay properly “shorts” it out.

It is possible however that some noise coming in on a neutral or even on the earth could cause a “reading” when really there is nothing. I’ve seen that many times on cheaper generators, they essentially turn the wiring into a radio antenna and you get these ghost readings… but to read something like that, you may have to use something more sensitive. Like a scope. Or something that can be quite interesting, is holding one probe in your hand (turning yourself into an antenna), and then pushing the other one on the earth connection (not on live! obviously!) and seeing if you get a reading :slight_smile:

53

might not have to be on first try…depends if the multi will keep on checking for the grid? But obviously there are limits as to how long you will sit and wait for the problem to jump out.

54

Eskom back, when the inverter syncs, when it takes over, measured a split second of 6v difference between N/E. Which I bet is as expected.

House measures zero between N/E too.

55

Normal. During the relay test it has to open the bonding relay for a split second to make sure there isn’t another bond there (that should NOT be there).

56

Plonk, what are the chances that a UPS, connected to a plug next to the MPII, can cause similar “noise”, or “build up” kinda thing?

Nelius made a good comment to me over the phone, that reminded me that I had the same UPS for years far away from the inverter, on the same plug point origin at the inverter, till I moved it a while back to right close to the inverter. He suggests I remove the UPS’es, put it back where it was, and check again tonight.

I’m desperate enough to try this. :smile:
Are you serious though?

57

This is reminding me of the Electrician’s Mate Mike Nitro in Down Periscope :rofl:

58

Look boys, that was terrible! I’m gonna go outside, I’m gonna have a smoke. When I get back in, you better have the horn section figured out. I can’t work like this…

59

When Plonk posted that clip here, I had to watch the entire movie. Man, I hosed myself at times.

By the way, a suggestion was made yesterday, to “trap” the problem, as it is intermittent, and one wants to waste time sitting there for a long time, right?

Take off my shoes and socks, then:
Connect a wire from the N in the plug outlet, to one toe.
Then connect a wire from the E to my other toe.
The MOMENT I felt anything, then I will know there was a problem.

The wife said no. … you will not.
Can I use your toes dearie … she just walked away. No idea why. It is science, I shouted …

60

So far, so good.

Removed the UPSs, now 3 of them, and placed them strategically throughout the house. Literally at the end of the solar circuit, to protect computers, Wifi, alarm etc.
Also switched the entire main DB off, incl all the circuit breakers, to eliminate the best I can, any further issues from there, if any.

Will ponder on what to test at the 12am schedule, and then the 10am one tomorrow, based on what I see at the 12am one.

Bar the street effect that was speculated on, the 1st post, the really bad one, I have a suspicion the UPSs this close to the inverter, may have a residual effect too.

EDIT: My donner …

I may not have be electrically minded, but the picture I got, the upstream supply, Victron, could be affected when in close proximity, ito cable length, to a UPS, with Victron already having to deal with a host of factors, methinks.

Definitely happening:
And I’m so going to make it so that the DB disconnects L&N from the grid with LS, only re-connect 5min after Eskom has returned.

Either with this: 12-01036 - Tripconnect 1P | Clearline Store

Or cheaper 63a double pole “smart” breaker, one with a relay in it, using a Shelly to keep the relay off for 5 minutes after power returns.

Further Consideration:
The worry, and scare, to have your equipment fail, have a problem smack bam during bad LS, suppliers are backlogged with existing problems, parts are scarce, new units have weeks waiting list … we don’t have plans for that.