There are now two events planned:
No 1:
If there are no errors after 2h of LS, the last professionals will have to come back to check their work on the roof.
No 2: When there is LS again at the 8 am/10 am slot
Dude, Iām beginning to think youāve pretty much narrowed it down to an actual fault. I mean the idea that outside influences might have an effect is cool and all, but this equipment must be able to actually survive and deal with finicky setups tooā¦ your issue is now merely to survive until load shedding is sufficiently over that you can have it repaired.
Send a log of the alarms with it when you do, with a full explanation of the intermittent nature. What youāre hoping for is that someone goesā¦ oh well, hereās a new unitā¦ go away!
Dude, think a moment about that comment.
A new unit comes in ā¦ and does exactly the same thing?
That makes no sense at all.
Moving on.
Some guys on a WA group, others Iām in direct contact with, who wants to help, throwing some ideas around, made me want to try a new thing.
Switch the changeover switch and disconnect the Critical Loads DB from the main DB. Just try that, then switch on the panels again.
Well, blow me the fā¦k down ā¦
I could not turn it ā¦ and when I try and turn it, I can hear the inverter groaning some more.
The changeover switch is fused, stuck, fooked.
Well that might explain itā¦ Iād have to think how it does itā¦ but it might.
Get a Hager SFT240. Why people use anything else I donāt get. Qualityā¦ without costing an arm and a leg.
2 Likes
Thing is, you were beginning to talk about getting an installer back to look at the panelsā¦ the DC sideā¦ and I really think thatās barking up the wrongā¦ well roof.
Jy weet mos, hoe meer kope in een sak, hoe beter. Bly ons kon help.
1 Like
As opposed to us, who just want to see the world burn 
There are some suggestions made here yes, good ones, ones that were on the To-Do list ā¦ and a lot of silence too. 
Letās be honest, these kinds of problems are time-consuming and extremely elusive, one lines may help, ost of the time they do, and then not.
Unless one walks the talk with the person in dire straits, step for step, helping to eliminate possibilities, thinking together for hours on end over days, that goes far far beyond a post or two.
We all know this about forums, off-forum is where it happens.
My main goal here, is maybe that gem comes around, more importantly, as there was no record anywhere on Google of this type of problem, now there is. And the steps taken to find it.
Sparkie is on his way, will bypass that damn changeover, see if solves it going forward.
1 Like
In my book that is not a bad answerā¦
Anyway, glad you found something. Still not sure how that will cause the bonding relay test to fail. But then also, Iām not sure of some of the detail.
You were going to test whether the bonding relay fails even with the grid completely disconnected, by turning the breaker off before the time. What was the eventual outcome after several testsā¦ does that work around it, or not?
Because I can sort of see how a bad changeover switch can leave a neutral connectedā¦
But then, a changeover switch that wasnāt moved in months is probably fused in a working state, just not operable anymoreā¦
Unless it burned and carboned up slightlyā¦ tracked over to the other end?
I worry that youāre going to do thisā¦ have two successful load sheddings just to get your hopes up, and then boomā¦problem back.
Ask me how I knowā¦ Iām fighting a stupid bluetooth DAC issue with my car radio. Musical ear. I hear the music slow down and speed up, and I have now traced it to the car radio itself. Iāve tried many bluetooth settings, and they all do thatā¦ initially seems to work, and thenā¦ *insert sound of dragging record"ā¦ the only solution appears to be a replacement head unit. And I HATE touch screensā¦ which is all that is in the aftermarket. Yeah sorry about the OT
The switch is out, completely jammed.
Three updates.
- Not one error with the panels off. I repeat, not one error with the panels off during 8 am LS.
- Sparky next to me, Eskom comes back on, I switch the DB on, inverter gives the first clack.
- In that 65s break before it syncs, I switch on the panels, and the inverter immediately gave an error, can hear the fridge motors rattle.
So, panels off again. Inverter recovers, and syncs with Eskom.
Switched on the panels a 2nd time, and it all works again.
Then I told him, now listen to the inverter. I try to change over the changeover, you can hear the inverter taking strain, the lights dip, UPSāes beep, and if you try and turn the switch even harder, a weird noise emanates from the top of the inverter. Crackling sound.
Sparkie is stumped. It is impossible what Iām showing him. Makes absolutely no sense at all.
Bypass is in, now I need another 3-5 LS events, between 8 - 10 am. Yes, I need that, or this problem remains dormant.
And THEN a penny dropped!
The sparkies mentioning a previous situation and this ā¦
A few months back I woke up, the house is off, UPS in the room beeps something horrendous. Errors on the Online Double Conversion UPS at the inverter.
So I switched the changeover switch back to Eskom, and nothing happens. Switched all off. Open the switch ā¦ and a red wire is loose inside.
I fixed it after chatting with the sparkie.
That was the first time I got an error, at night, when everyone was asleep.
Until I can replicate the problem on demand, I will be wasting their and my time.
Getting closer ā¦
Letās see what transpires now that the changeover is removed.
Tested every single permutation I could think of, worked perfectly every time unless it is during the 8 am and 10 am slots. It makes absolutely NO sense whatsoever.
EDIT:
If schedules stay, I can check tomorrow morning, the 1st check, at 8 am.
Pretty sure it is still going to give errors though ā¦ at least I found the switch.
@TheTerribleTriplet you mention something about that the main box in the road isnāt locked, go drop your mains outside and go check. Same as LS then
the only one in all of South Africa looking forward to load shedding 
Mmm, that is an idea!
However, I must focus on doing one thing at a time. Control the variables they say.
If the error persists, after LS is over tomorrow, then I will go and switch the house off at the street in the box and see if that replicates the problem.
Jip, it is so wrong!
Cause if LS goes away now, specifically the 8 and 10 am slots, then the variables may change.
Or worse, the problem does not get identified and stays dormant, ready to blow the inverter up come the next round, or something as bad.
ā¦ am on a roll now man ā¦ on a roll.
my last go at throwing cr@p and seeing if something will stickā¦ but hey I am here to learn 
you went 6 panels, then 12, and since after last bad load shedding, added 3 more (my incomplete hypothesis is though dependent on not having any daylight load shedding since then since the problem should have come up then).
any case, inspired by @JacoDeJongh 's post, with the extra panels, possible surge arrestor leaking to earth when MMPT/s MPPT/s sweep for power point (the sweep possibly explaining the intermittent appearing nature). But I do not know how the neutral will come into the mix.
There, I tried. 
2 Likes
AHA!!! Another point to ponder on ā¦ check.
Background: The box was āspeccedā for 250 MPPT.
250/85, the main one, has 12 x 350w panels on it.
150/35 has 3 x 355w panels connected. Kinda the portable camping MPPT.
However ā¦ keep in mind, 8/10 (8 - 10 am only), not the afternoon LS slots, where the max power sits.
Or ā¦ is this because of the lower morning powers, the arrays being slightly NW, i.e. peak power afternoons.
Combiner Box:
1 - Spare
2,3,5 - 3 strings of 4 x 350w panels to the 250/85
4 - 1 string of 3 x 355w panels to the 150/35
6 - Goes to the 150/35
7 - Goes to the 250/85
If this is a problem ā¦ let me know.
Tests were done:
Did switch off the string for the 150/35, made no difference on the errors.
Only when 2,3,4 and 5 are off, do the errors go away.
EDIT: That green earth from the Citel, goes to a spike/pipe, down below the combiner. It may, or may not, be connected to the TNC-S earth outside the house. BUT, I did connect a wire from the spike back to a plug outletās E. Made no difference.
Ever since having a car blow a fuel pump fuse RIGHT on the first start after replacing the ignition harnessā¦ and then spending an hour trying to figure out where I messed up, re-checking the firing order even, checking spark, etc etcā¦ only to discover a blown fuse in another subsystem completelyā¦ ever since that day back in 2001, I have learned never to discount stupid dumb luck when it comes to debugging.
Correlation does not imply causation. Except when it does. This observation of mine may or may not imply that sometimes correlation may in fact imply causationā¦ or not. Left as an exercise to the reader.
IF the system errors again tomorrow morning, I will:
@Village_Idiot, you get the badge for providing that gem in my book.
It turns out, the new array was not on surge protection.
But that is not all ā¦ the Citel DX240-130DC, which was professionally installed years ago for the 250 MPPT, is adequate for the 150 MPPT, not really for the 250 MPPT.
So, come next week, I need to get another surge protection in place, for the 250 MPPT.
Now that I check EsP loadshedding schedules like a hawk, plan for the 8am slot, double triple confirm it before going to bed after 1am, only to wake up at 6am to see the schedules have changed, now LS only at 1600.
Can feel the pain of people who have to plan their lives, and businesses, around these moving targets.