As far as experiments go, this is probably the longest and least exciting experiments of them all. It started years ago when I wanted to see just how long some randomly branded Chinese lithium-polymer cell based power bank would last, as another manufacturer’s unit failed within a handful of cycles. Fast forward, and this is the 16th iteration of post, representing cycles 751 to 800 from the beginning of the experiment, or 816 cycles accumulated since manufacture.
Isn’t it always the case – just when you thought nothing would happen, something inevitably does? This latest batch of data shows a marked change in the degradation behaviour, with the slope taking a severe downward turn. Values of close to 2800mAh have been recorded, with the peak sitting just low 3100mAh. At this point, it has potentially irreversably passed the 80% of initial capacity point and failed – somewhere in-between 750 and 800 cycles. Whether this is actually the case deserves some more investigation in terms of putting more cycles on the unit to see if the trend continues downward or not.
Initially, extrapolations pointed out to 1200-or-so-cycles. This doesn’t seem to be plausible now – however, a good reason for this may be the prolonged storage at full charge due to my PhD causing a step “dip” in capacity of about 200mAh. Without this, the cell may have made the extrapolated figure.
When plotted with reference to zero, the degradation is starting to affect the capacity noticeably, but there’s still a lot of capacity remaining. The continued degradation trend will be particularly interesting.
It seems that “out of nowhere”, a new regime of degradation has begun with a different gradient of curve. Whether it has irreversibly passed the 80% capacity mark is uncertain as the readings are +/- 100mAh or so and we do get some ups and downs, but it seems likely to be confirmed. The future degradation trend is of interest – whether it continues along with the new gradient or whether degradation will accelerate.
In other words, stay tuned for the next installment, which might see some more sharp drops in capacity (through to 850 cycles since commencement of experiment) – exactly what we’ve been waiting for.