It’s Boxing Day, and while a lot of people are outside partaking in the shopping frenzy of underwhelming discounts, I’m firmly sitting at home plodding away at some data analysis to deliver this last part of the power bank endurance test.
This whole project started because I had another Li-poly power bank which didn’t last very long, so I decided to test a different type of Li-poly power power bank with a relatively “unknown” pedigree to see how well it would fare with cycling. That was back in mid-to-late 2014, when I expected the unit would probably not make the expected 300-500 cycle life of most Li-type batteries.
Now, it’s late 2016, and the unit has reached its 1000th cycle since the commencement of the experiment, or 1016th since it was manufactured. Lets see how it fared.
The last 50 cycles appear not to have continued quite the same steepness of decline in capacity, instead, more plateauing out and maintaining the capacity at about 2400mAh or so. This was an unexpected deviation.
Plotted referenced to zero, it’s clear the power bank still has about 60% of its initial capacity remaining, and is far from being useless. No component failures occurred, and the whole assembly is visually still quite indistinguishable from a unit that’s seen light use.
By request, and because I also had a hypothesis earlier that the degradation may be more cell age (i.e. time since manufacture) related, I have plotted the measured capacity versus the test date and time. Ultimately, there are a few gaps in the graph, due to holidays and having the test equipment diverted to an investigation for my PhD, but ultimately, it seems it could be possible that cell age does have a role to play.
Where the cycle graph shows the “dwell time” resulted in a big jump in the cell capacity, visually on the date-time based chart, it appears that these gaps are part of a larger continuum that looks somewhat bi-phasic with two near-linear segments.
The same graph as above, referenced to a zero-scale shows it may actually be more of a curve rather than a bi-phasic linear function.
The rather unknown power bank survived 1000 cycles with no failure of any component at all. While the battery only has about 60% of its initial capacity, it’s not entirely useless, and its decline was fairly controlled. Based on the graph, it had eclipsed its 80% mark by ~760 cycles, which is still above and beyond expectations. Graphing the decline as a function of test date/time reveals that the decline may be somewhat time-related rather than cycle-related, although both are believed to have contributed.
Unfortunately because of time and logistical constraints, since the experiment ends here with this final installment, we will never know what will happen to this particular cell in the future. While the bigger trend appears to be a steepening decline in capacity, the latest cycles seem to show the cell holding more-or-less steady compared to the previous decline. Regardless, this one well outlived my expectations, and was very much uneventful in its decline.