RealBlackStuff wrote: ↑
Sun Feb 10, 2019 5:50 am
FYI: if you want to use non-Lenovo batteries, you MUST do the battery-check EC firmware mod as described at the top of this forum.
I got myself an after-market battery manufactured in China with a company label "Irvine
" for USD 25. I connected it up, not sure of what to expect, but the battery worked just fine. The battery capacity is listed as 4400 MAH and the voltage as registered by linux varies from 11.6 to 10.3 volts over a cycle (as listed in Xubuntu's Xfce Power Manager).
When I took a look at the details of the battery(in Power Manager again), it's showing itself as a Sanyo 45N1023, clearly indicating that it contains one of the following:
1. A Sanyo board inside it.
2. Sanyo firmware on some random board.
3. A custom firmware on a random board that's programmed to identify itself as a `Sanyo` and possibly return other false data when queried as detailed in the `Funny s**t` section below
Funny s**t about the battery that I've noticed:
1. The battery displays a design capasity of 48.8 WH, but actual fully charged capasity keeps showing different values ranging from 82% of full capacity
and then back up to 92%
. Currently it's indicating a 88% charge.
2. At 100%, linux's power indicator gives battery left time greater than 12 hours consistently (varying from 15 hours to one instance of 21 hours plus change), but the last 6 cycles show that the battery doesn't last longer than 3 hours, whatever the workload.
3. The tail end drops reaaaaaallly quickly: Linux does an auto-shutdown when battery falls below 5% with the original Lenovo battery, but with this, it goes from 20% to power down so quickly that linux doesn't register anything beyond a popup for low battery.
4. The battery doesn't give a cycle count, which is really bugging me as I want to show the seller the details.
I've had the battery for 4 days now, and I've been using just it, running full charge and discharge cycles on it, hoping to get it to recalibrate itself, but I'm getting the feeling that the firmware is messing around with it's calibration data to report inflated charge values. i was hoping that it will firm down to a believable value close to reality, but it seems a faint hope.