Doing a capacity test now.
Running on a Maynuo M9711, which is a cheap chinese clone of a HP electronic load.
I have an arduino in the back that hooks up to the RS232-port of the load and then uses modbus-protocol to control the load settings and measure what the voltage is on the input, which is the battery in this case.
It then echoes back the voltage and a timestamp to a serial terminal on a PC.
Current test cycle is as follows:
- Constant current load @ 3.2A
- Load ON for 15 minutes
- Load OFF for 5 minutes
- Repeat until battery hits 41.4V (more on that later)
The three values seen on the serial terminal are voltages sampled at different points. The first one is the voltage at the end of the 15 minute ON period, but before the load is turned off (under-load voltage), next voltage is the voltage about a second after the load has turned OFF (immediate release voltage) and the last one is the voltage at the end of the 5 minute rest period (resting voltage).
Okay, the DC load is a max 160W load, so it doesn’t really stress the battery so in this case all the voltage are going to be pretty near each other, but for example in the case of my 10S6P pack there is a bit of a difference between these voltages, as is shown in the graph below. This helps to get a little bit better picture of what the SoC is based on the resting voltage.
Now here’s some data I dug up on the battery pack, because I needed to figure out what the discharge cut-off voltage was before starting the capacity test.
The 12 cells are manufactured by Samsung SDI and are meant for electric vehicle use. They also upgrade their design and chemistry over time, but keep the same cells dimensions so car manufacturers don’t need to completely redesign their battery modules, but can just upgrade the cells to a higher capacity one.
This also explains the little confusion I had as I had looked at the 4,15kWh and 5,3kWh i3-modules, as they had the same dimensions based on their product pages. I even contacted the shop about this and they confirmed that the dimensions were correct, but now knowing this upgrade thing happening to the cells on the manufacturer side it makes sense, especially as I found this little sheet that reported the different cells based on their manufacturing date. This would also mean that the 4,15kWh i3-module uses the 94Ah cells, which then lowers the energy capacity to the 4kWh mark.
But boy oh boy, if they are gonna release those 150Ah and 180Ah cells sometime in the future 
I was still missing the datasheet for the cells, which specify the cut-off for both charging and discharging for the cells and no matter how hard I looked I couldn’t find the datasheet for the 120Ah cells in the module I have, but I did find the datasheet for the 94Ah cells, so I guess I’m gonna go with that info.
So it looks like operating voltage is specified at 2.7-4.15V but looks like effective SoC is 0% at ~3.4V, so I then decided to stop at 3.45V at the cells, so this then lead to the 41.4V termination voltage for the capacity test.
EDIT: it’s gonna take like 2 days to get the full discharge done on the battery