stefan44

these hoselect 48v batteries are in fact 13 cell NMC chemistry and not lifepo4 as stated in the manuals, thus 4.2v per cell is fine(its in the high knee of the voltage spectrum but it is the "factory settings") to double their lifespan only charge them up to 4.10v and discharge to 3.4v

yes, 1000 time this. and it has a 600ma active balancer

there are plenty that sells for less than R4.40/W

the spark that you saw, might be the potential difference between the grid's earth and your house earth. in a TNC-s system the neatral and earth are joined at the grid side(substation, distribution), but it is split again at your house( that's why floating neutral and earth is basically the same thing. when your inverter is in island mode ( no grid present) then you should only be connected to the earthing system that you installed at your house and not the grid's earth, that's probably why the issue persists even after you flip the breaker after you disconnected the grid.

induction is a product of AC power, DC only momentarily induces a current in nearby conductors only the first moment the switch is flicked and the magnetic field builds around the powered conductor until it reaches its constant value after that it does not "induce" anymore as both the conductor and the magnetic field is stationary, as induction is a product of a changing magnetic field (faradys law), a conductor moving inside a magnetic field produces induced current(motors, generators) or a stationary conductor inside a moving magnetic field like AC conductors. this does not solve your problem, but at least it is not induction. I do not know what earthing system you use in your country tn-c, tn-s-c ect. but im fairly certain you have a floating neutral/ floating earth whatever you want to call it.

I also have a 16l geyser for a couple of years already, I also experience the same fluctuating gas usage. 48kg last 5 weeks, 19kg last 3 weeks and 9kg last 2 weeks, it depends on environmental factors alot, but in general the ratio between the cylinders remain the same, in summer you can add a week for every size cylinders lifespan. based on this usage patterns that I have observed over a couple of years, I only use 9kg cylinders on my gas geyser and stove as they are the most economical in my case.

if it was me, I would 1000 times rather get a second hand victron/meanwell inverter in the 250 to 500 VA(victron) watts(meanwell) have seen them being sold on facebook marketplace for less than R1000 and pair the inverter with a secondhand blue or smart solar 75/10, 75/15, 100/20 also below R1000 secondhand. and I am willing to bet that a secondhand victron or meanwell will outlast a brand new china mall inverter multiple times.

from 9:55 up to 20:17 and only lost 0.06v, you could go for days without sun. sometimes you just have to figure out the logic behind the board's parameters and your own system working.

at that price it will be a hard sell, recently I saw a lot of ads for 500 - 550w panels for R2300 -R2800( JA, Canadian, Jinko) per panel. it seems as if the prices crashed.

what you can try and I think it will work is to remove the connection from bms P- to regulator negative output, now connect your pv panel and test the regulator output (should get 12.6v) and only now connect the bms P- to the output negative of the regulator, now the board should be powered where are you going to connect the sonoffs?

yup seen this happen before with dc/dc regulators. my uneducated guess is that the BMS sees the low resistance of the regulator as a dead short and protects the cells as soon as you connect it.

it depends on the amp rating of the board, more mosfets = more amps these small cheap boards I ignore amp rating and work with 5amps per mosfet(one row for charging and 1 row for discharging) for longevity and a cooler running temp.

L_D Remove the DC/DC converter from the output of the bms and see if the bms powers on. what is this that you are making?