@kresil It's not clear what the video is intended to imply. But since it presumably starts charging from empty cells, then that is one of the easiest cases, even for naive SOC algorithms, since they already know the exact SOC. Thus if they are matching to some model cell(s) (and the cell is similar) then the results will be very close. That's the trivial case for fuel gauge algorithms.
The harder (nontrivial) case is when you start charging at nonzero SOC (e.g. 30% or 40%), as is typical in normal use. Also one needs to test cells of widely varying capacities, and widely varying IR. That's where the sophisticated algorithms differentiate themselves from naive heuristics.
I was involved in doing similar tests for some professional level chargers, and they are extremely time consuming, so they are typically automated. This is not an easy task for a charger that has no means of automatic control. But perhaps it might be possible to jury rig some sort of automation.
The harder (nontrivial) case is when you start charging at nonzero SOC (e.g. 30% or 40%), as is typical in normal use. Also one needs to test cells of widely varying capacities, and widely varying IR. That's where the sophisticated algorithms differentiate themselves from naive heuristics.
I was involved in doing similar tests for some professional level chargers, and they are extremely time consuming, so they are typically automated. This is not an easy task for a charger that has no means of automatic control. But perhaps it might be possible to jury rig some sort of automation.
Last edited: