Joachim Breitner's Homepage

In June 2007, I bought a new 9-cell-battery for my ThinkPad notebook. Since then, I recorded, via a cron job, every five minutes these bits of information:

• time
• design capacity
• reported last full capacity
• state (charging, discharging, charged)
• (dis)charge rate
• remaining capacity.

I had planned to make some detailed analysis of this, but I haven’t gotten around to do more than this graph (warning, large), of which the following is a cut out. Since I now got a new laptop (with a new battery), I’ll share my records.

I find it interesting that the last-full-state sometimes jumps up, closer to the design state. This could be re-calibration. Or the battery, after a while, re-scales iits output to appear less degraded. This might be checked by comparing the speed of discharge and charge at various points in time.

If you want to do some analysis of your own, you can download the raw data (44354 data points). It has been anonymised by starting the time counter at zero, and shortening any gap of more than four hours to four hours. If you want to do some analysis in Haskell, you can take some code from my anonymizer script to parse the data. If you happen to find something interesting, appropriate crediting is appreciated.