Would TNC-S earth have any bearing on this? The spike is next to our house, we don’t have an earth from the street box.
Many many years ago, a friend of mine lived on a farm in Jeffrey’s area.
In panic mode he made me sell him one of my UPS’es. He was losing PCs, printers, and other electrical appliances in his house. The voltages, during LS, which should be zero, were like between i.e. 50v and 120v.
Made no sense.
Turns out, the new wind farm in the area, never disconnected from the grid during LS. Munic found that and sorted it.
Hence my interest is peaked when you guys mention that the circuits may not be as dead as expected, inverter sensing that. Maybe a neighbor right next to us, and on the same street box as us, doing something … maybe.
“TNC-S” and “spike next to our house” are different things though?
Apologies for my non-scientific information. 
Let me explain in more detail.
With the original Multigrid installation years ago, when Eskom comes back on after LS, the Multigrid will switch off on the spot. Each and every single time.
As we did the official tests, the sparkie found that the earth from the street box was nowhere near to specs. Asking the Munic to fix it, not a chance.
So they had to install a spike, that long rod thing, hammered into the ground, next to the house, and connect that to the earth of the DB. The Multigrid worked perfectly thereafter.
I was told, and the form reflects the same, that it is a TN-C-S connection now.
Test one concluded:
First Grid lost, was me doing a quick test.
The 2nd one was also me, off for one hour, with the same everything as this morning.
I think It may be time that spike got a boetie or two. Earthing conditions literally change with the weather.
Maybe wet that spike, and check the earth connection for corruption and tightness. If you do add spikes, right next to the existing one and very close to each other and connect all of them together right there, you want equal potential between them.
Groetnis
1 Like
with salt water…
Maybe 
because Salt is corrosive, and will kill plants nearby, so use sparingly.
Groetnis
Mmmm, there is a neighbor’s lawn there.
The next test has started. The breaker is off, waiting for LS to start. Will do it for a few LS schedules.
Sparkie will be on it after this Covid thing has left the house, at the least, he must test it all again, and leave me with all I need to continue the tests once he has left.
Well, maybe just make your relief spot after sundowners.
1 Like
As far as I know a spike right next or close to another will make a negligible difference and you will be wasting your time. Especially if you’ve got rock hard soil, got the T-shirt, tried to get spikes in at our new property for the fence energizer this past weekend, 60, 70cm after lots of water, lots of hammers, koefoet, lots of sweat and swearing, they even started to bend. Eventually we just grinded them off. If that earth gives issues I’ll have to get one of those SDS hammer drill attachments to try and get in new ones.
If your spike is 1.2m long, put the next 1.2m from the first, the third 1.2m away and so forth and then bond with moer expensive copper cable.
Just to keep things in line … earthing is only an issue if the sparkie finds it a problem. then he can fix it.
At this point in time, having removed any chances of any “foreign voltages” between the main DB / inverter for this LS cycle, the breaker is switched off.
Let’s see how this one rolls, this morning it took 5 minutes before the first error, now 42min in, no drama yet.
Perfect, worked as designed … if I switch the breaker off.
Will do that for a couple of LS events to make damn sure.
Think I need a solution to disconnect the DB from the street when Eskom goes off.
And keep it disconnected for 5 minutes after Eskom comes back on.
Or, as my dealer strongly recommends, I need to install a:
He says they are having endless relay issues with relays burning out on up to 300v from Eskom when it comes back on.
Titbit: How do you test the AC if you don’t want to open a DB?
Plug a computer power lead into a wall socket that is off due to LS.
Then using your tongue … ok, ok, maybe not … stick a multimeter into the power cord plug.
BUT … I realized, it does not happen all the time, see the first post. So unless one sits there the entire time, you are going to miss it.
As I understand it, it measures voltage across a closed switch (relay), and that really should be zero.
So either it is not zero, or the measurement is incorrect. It seems that being connected to a “dead” grid has something to do with it. I doubt it is the quality of the earth, although it probably wouldn’t hurt to have it checked. Even a floating Multi’s bonding relay test should succeed (but the larger relay test of course won’t), as long as the relay does in fact work.
I also doubt you’ll be able to see the problem with a normal multimeter. Well, maybe you’ll get lucky, but spurious problems like these… not holding my breath.
Edit: 10 bucks says it is a neighbour backfeeding noise with a generator…
Hold your bet for a moment. 
With the breaker off since 12am this morning, I got one again this morning …
My thinking is, it could be external to the house, yes, but with more than one cause.
@plonkster, other than the earth to be checked by a sparkie, still a few days to go due to a Covid outbreak in the house, can a surge protection device on the AC_Out contribute to this?
On top of now looking at devices connected onto AC_Out1, I’m also eyeing NoreRED running on the Large OS, straws come to mind.
Have you considered the possibility that the relay might actually be bad?
I did, do, consider that.
But for that it needs to go in, and with the influx of issues lately, parts shortages, not forgetting LS, it is going to become a rather serious exercise.
Wish there were loan units one could use whilst it goes in for service/repairs.
So before it goes in, I need to make 100% absolutely sure it is definitely the unit, and not something else.
While I generally think you should not be doing that, I cannot see that causing this particular issue.
I think you may have to consider that the relay is actually going bad. The firmware is doing its actual job. You may be able to get it swapped.