paulus

The improvement in battery capacity indicates that, to some degree, sulfation had been reversed. My hypothesis for the root cause of the pre-mature battery failure is that, in the absence of a battery balancer within a bank of batteries, the weak battery will keep getting undercharged and will progressively sulfate and get weaker and weaker. This will eventually spiral out of control and lead to the destruction of the whole bank of batteries! I added the HA02 battery balancer, then reconnected and powered-up the inverter. I repeated the capacity tests with the inverter + battery bank at an average inverter load of 450W (the current draw from the batteries was on average around 10A). The runtime had improved to 8h 40min! Not too shabby! (refer to the graph below for moredetails):  

 
8hours 40min runtime @ 10A vs a theoretical maximum 20h @ 10A indicates a battery bank health/capacity of 43%  
I also installed a shunt and an energy meter (ill share closeup photos in a future post) and recorded that the batteries had in fact delivered 3896Wh during the above test. A crude estimate of the (theoretical) maximum stored energy in a 48v 200Ah battery bank = 48*200 = 9600Wh. Interestingly, the measured capacity vs theoretical capacity also indicate a battery bank health/capacity of 40%! Not a bad improvement for batteries that are technically kaput!
 
I wouldn't trust them to carry me through the night, but they should be fine for 2h or 4h of load shedding!

So after desulfating the batteries one by one, I repeated the individual battery capacity tests. The following graphs show the performance:
Before desulfation:    29-30 Nov 2020
After desulfation:       01-06 Dec 2020

All 4 batteries achieved a runtime of more than 2.5hours with the same 50A continuous current draw!
Taking a closer look at Battery #1   Before & After:
29 Nov 2020 - Runtime         14m30s - 12Ah (8% capacity relative to the datasheet 150Ah)
01 Dec 2020 - Runtime     2h36m10s - 130Ah (86.6% capacity relative to the datasheet 150Ah)
 
Needless to say, I was really chuffed with these results!
I couldnt wait to power up the inverter and repeat the system test.
I was also exhausted from the old-school "paper, pencil & stopwatch" method of data capturing. This required urgent action to automate the testing and also to add datalogging (I was planning to run more tests...more on that in the next part)