Function of the three way switch for my inverter

On the AC sub DB for my system there is a three position switch.

Top position is labelled “Solar” and this is the position that would usually be selected.

Middle position turns off all power to the property - no grid, no solar. The inverter and battery stay on. (OK, OK… the state of the inverter and battery does not change)

Bottom position is labelled “Grid”.

OK… so yesterday morning we had load shedding schedule from 4:00 to 6:30. I woke up about 5:45 as I had to get dressed and shaved earlier than usual as I would be commuting to Centurion that morning. I decided to have a cuppa. I walked into the kitchen, hit the light switch and … nothing. What?

I took a peek at the inverter. All the lights one would expect to see on were on and in their usual state. Except for the wifi indicator which was flashing - because it couldn’t connect to the wifi in the house.

So it must be earth leakage. What a nuisance at this time of the morning. Only it wasn’t. Every switch on the DB was where it should be.

So now I go and check the sub DBs for the solar system and I find that this three way switch has been left in the “grid” position. Hmmmm… throw that up to “solar” and everything in the house starts switching on. How that switch got to be in that position is a mystery to me, but that’s not really why I’m posting this. Suffice to say that I didn’t change this switch from “solar” to “grid” sometime the previous evening.

Here’s yesterday’s graph from the Goodwe portal

You can see everything going quiet at 4 am. Until then, the system had been behaving as expected. Loads (yellow line) are serviced by the battery, and the SOC (green line) is slowly dropping. At 4:00 the loads drop to almost nothing (actually about 50W, which I assume is self consumption by the inverter). Then just before 6:00 the loads come up again. That’s when I put that three way switch back into the top (“solar”) position.

So now my question: What is the “grid” position on that three way switch for? As I’ve said, the property was running off of battery until the load shedding started. If look at the graph for the day before (when I had no need to touch that switch) the system is working usually: house running off battery until solar starts coming on line, then solar takes over the loads and starts charging the battery. Once the sun goes down, the battery takes over.

So everything ran just as usual until there was a load shed. Then despite the inverter being on and the battery having better than 50% SOC, no power is supplied to the property.

I had thought that “grid” position switched the property to grid and bypassed the inverter, but clearly it does something else. But I’m not sure what.

Sounds super weird. Only thing I thought of was many years ago when we arrived on the farm early evening on a weekend (I was still in school) and in the kitchen my parents found that the refrigerator was turned off at the wall (the switch was easily accessible), while the condensor grid on the back was still warm to the touch, meaning it was running until recently.

Months later they found out that one of the farm workers was regularly breaking into buildings, but he threatened the others so nobody was brave enough to report him until he’d left on his own accord. What happened was he was in the house snooping around, saw the lights of the approaching vehicle, tried to turn off the light, accidentally turned off the refrigerator instead (switch was on the opposite side of the doorway), and then managed to get away without being caught. For that entire time, my father thought he must have absentmindedly turned it off…

An interesting tale, but I am not stealing from my own fridge (I just open it up and take what I want. After all, I paid for all of it).

So am I right that the bottom position of that switch should cause the inverter to be bypassed and allow the essential loads to be powered from grid?

I now remember some more details.

There was also a rip in one of the window screens (farm, flies can be an issue). My parents thought it was an animal.

The farm worked who finally reported it showed my father how it was done. He would tap a small hole in the glass of one of those small top windows, inside the putty, just large enough to list the handle inside. Then he would hook the screen open, and the handle of the larger window. That would yield access to the main bedroom. Leaving was easy enough too, you push the window close so the handle falls into place by gravity. Nobody is the wiser.

Indeed, I would expect that if it is wired like mine, and if that inverter is not somehow pushing energy to the non-essential side, that with the switch in the GRID position, the inverter effectively does nothing and the house is essentially powered by Eskom. Unless there is a super weird gotcha somewhere, that switch was thrown at 4AM.

In our area we have a team of robbers known as the “roof rats”. They gain access through the roof, disconnect your power (or they do it at the kiosk outside, if accessible). Then they wait for the alarm battery to die… which most does within a few hours since nobody maintains that thing until it dies anyway. Then they spend hours properly ransacking the house looking for specific items.

They hit my neighbour accross. Spend over 2 hours going through the house. Stole almost nothing, just a few mobile phones. They were looking for something specific. My personal guess is fire arms. My neighbour has a nice 4x4. He can be mistaken for a hunter… in reality he’s more of an outdoors guy.

(Also, racing. He has a Lotus 7 replica he uses for racing).

Makes perfect sense.

Top, your inverter is connected to the entire house, grid-tied, will work when there is LS.
Off is off, bar inverter, to work on shiite so stuff must be switched off.
Bottom for when the inverter is off for maintenance or whatever, your DBs are fed 100% by Eskom, no solar at all.

[ voluntarily deleted by author. Gives away house/car breaking techniques]

Mmmm, I may have had something along those lines … and I NREALY switched off the alarm.

Next morning, turned out to be a ma-moerse-spider sitting over a PIR …

BUT, fact is, I was seriously tempted to switch off the house alarm.

The move on the car, that is clever.

When I got burgled (or house robbed) it was clear they were looking for specific things. There was lots of stuff they could have taken but did not. And the things they took they handled very carefully and wrapped up to protect them.

I remember they took my car keys, started my car, then left it standing with the engine running. If you don’t want your car stolen, get a Hyundai diesel sedan.

Of course they are always looking for specific things. Things that are easily fenced, mobile phones (always need those for future jobs), tablet computers, jewelry, laptops, flat screen televisions, fire arms, and cash. Cash is king of course. But when they spend over two hours in the house and leave the jewelry and laptops behind… you know they were sent to the shops for something specific.

I don’t suppose there is a way that switch might have dropped with gravity? I have a Hager SFT240, which is quite stiff. No way does it switch without someone doing it. In fact, if you slam it through from solar to grid, many loads don’t even notice the break. They are that good.

I just checked. It is a Hager SFT240.

But that still wouldn’t explain what happened. With that switch in the down position the house was still drawing from the battery and inverter. That shouldn’t have happened. Draw from the grid should have been equal to the loads - but was not.

Does a DB door cover the switch and if not, do you by chance have a cat or an open window where a cat could enter?

We have a nuisance cat in the neighbourhood that gets up to all kinds of mischief

1 Like

Are you sure it was down?

Not sure if it’s possible with your inverter, but what the switch does is move the loads before the inverter (or between the grid and inverter to be more precise). If your inverter has a CT / meter that it uses to calculate the draw from the grid to keep it at a certain level (like 20W) then that should still work and power the loads.

In my case (Victron with a ET112 grid meter that I flipped the switch on) you can see the load is now all in front of the inverter, but getting powered by the inverter:

But the moment the grid goes down the meter goes down and the inverter will only be able to power loads after it.

When is the last time you had loadshedding? I’m just going to assume that the switch was flipped as a part of all your tests done for the CoJ stamp of approval and never switched back.

Edit:
On another note I wanted to be fancy and show off the GUI-v2 via VRM, but it doesn’t show it the same as the old UI and gives the impression that the loads are after the inverter:

Old:
image

New:

Yes. But I will make a controlled test tomorrow.

There’s been quite a bit of activity around that system recently.

  • City Power came to test it for SSEG registration purposes. They definitely threw that switch.
  • There was an electrician here to put in a socket fed directly from grid. All pre-existing sockets inside the main house were on the essential loads side of the DB, and I needed one connected directly to the grid for using the CIU supplied with the pre-paid meter. He ran a cable out of the sub DB that includes the 3-way switch
  • The installer was here updating the firmware on the FreedomWon battery. But I don’t recall anything turning off whilst they were here (I had asked them to let me know if they were planning to, as I wanted to save all my work first).

I restored power to the property by moving that switch from down, through middle (all off), and then to up. That was all I did. The inverter wasn’t rebooting - there’s a warning light that flashes when it starts up, and that was not flashing. No breakers down. No earth leakage trip. I just manipulated that switch.

Good question. We’ve been scheduled for load sheds, but I was suspecting that a number of them never actually happened for whatever reason.

COJ were here on March 6th

My logs show loss of grid power at

  • 7th 2:12 to 4:29 (I wouldn’t have known anything about this)
  • 7th at 10:00 to 12:36 (I would have noticed this, SEMS shows the system working and backing up loads)
  • 8th 8:10 to 10:09 (I would have noticed this)
  • 14th at 04:08 to 6:30
  • 14th at 20:31 to 22:02 (by which time the switch was back in it’s usual position).

The logs also show a loss of battery communications on the 11th. That was the upgrade of the battery firmware.

It’s interesting what I can piece together from various logs and apps. The electrician I referred to was actually here on Feb 21st, much earlier than I’d remembered. So it seems unlikely to have been him, since there were losses of grid power but no loss of power to the house since he was here. And City Power performed their safety tests since he was here and pronounced themselves satsified.

So something happened between 8th and the 14th.

I’m beginning to think I have misunderstood the function of that switch. Though something in my tiny, unskilled mind doesn’t add up. OK… with the switch in that position, the inverter might still be able to send power to the backed up circuits. But then why does that fail when the grid goes down? And why didn’t it fail when the grid went down on the 8th?

That switch must have been moved somehow.

This stands out to me. I’m assuming the firmware update would have required shutting down / bypassing the inverter to be on the safe side (I’ve never updated battery firmware, but that’s what I’d do) so most likely the switch was flipped to ensure the loads don’t go down during that time, but then wasn’t switched back. That’s similar to what I do when updating the inverter firmware.

They won’t be ‘backed up’ if the switch is in the grid position (my pedanticness), but yes, they should be able to still get power. If you make use of a grid meter / CT coil the inverter should be able to feed through the inverter’s AC IN to keep the grid setpoint (like 20W) and will be able to power the loads (until the grid goes down)

When the grid goes down the AC IN of the inverter goes down, so if the loads are between the grid and the inverter they’ll go down as well.

I’m just going to assume the switch was only moved to the grid position between then and now.

Definitely, but look at the bright side. It’s not a valve that takes two people to open / close and it didn’t take 10 days to find…

1 Like

So it starts to look like this happened on the 11th.

LOL

1 Like

The transfer switch is probably after the Grid feed to the inverter so the inverter will always get Grid power when needed and doesn’t get influenced by the transfer switch.

The 3 way switch bypasses the backup side of the inverter. So in theory you can bypass it when it fails.

When the switch is set to “grid” it’s like everything is on non-essentials.

1 Like