Charging single cells was too slow, so I configured my LabJack U3 to measure two cells and doubled my charging speed. I also discovered that my cells are wildly out of balance. This isn't terribly surprising since they were manufactured in March and haven't been charged since.
- Cell 1832 required fewer amp-hours to reach end of charge (defined in this profile as 4.2V at 3A).
- Cell 1833 never got onto the exponential(?) voltage rise near end of charge.
- Cell 1832 took about 4Ah from the beginning of the end of charge voltage rise.
- Cell 1832 had a higher charge voltage throughout the whole charge, it is unknown if this can be used to estimate state of charge (the literature says no).
- Both cells relax quickly after charging is stopped. More than 24 hours later, 1832 is at 3.366V and 1833 is at 3.341V.
- My bench supply isn't well cooled, when run at 3A it overheats, the maximum output current falls somewhat as it heats up but not enough to protect itself. To be fair it's sold as a 2A supply but I'm running it at 3A. You can see the current fall at the beginning of the charge and then rise again after I stick a fan on it.
- Don't power the fan from the charging power supply itself as this messes with your measurements (I'm not sure why).
So we can conclude that cell 1833's state of charge was at least 4Ah less than 1832. We don't know why this is, most likely the cell's self-discharge rate is different, but it could also be differing state of charge when the cells left the factory. Differing self-discharge is easy to measure but you have to wait for the self-discharge to happen. Once a fully featured BMS is installed and regular charging is undertaken, maybe the source will be discovered.
I'm going to have to find some sensible way of storing this data so I can analyse the lifetime behaviour of my cells.
So far, I've fully charged 4 cells and partially charged 5 (>10Ah in without hitting end of charge), my goal of 30 by October 20th is easily achievable.
Stay tuned for White LED - Darlington balancers.