X210 fan and battery communications

[L29Ah]

Hi there, i just got the third batch™ X210 kit, and it seems like it cannot report battery current in Linux (0 via ACPI where X230 reported correct values) and neither report nor control the fan speed. The former makes me unable to optimize my battery life, and the latter results in the CPU frequency throttling kicking in earlier than the fan decides to run at its full speed. I've tried the x62 tools, but they don't work for X210. Could anyone do anything about these shortcomings?

EDIT: fixed in the EC firmware, see the EC hacking thread.

Also i really miss the original Thinkpad EC feature that allowed me to undercharge my battery; now i feel like i'll have to replace it every year due to rapid capacity loss caused by the cells being constantly heated while being fully charged, like on the other [censored] laptops.

[dmdsoftware]

Which kernel and distro are you using? I'm still using debian jessie with kernel 4.8 (need to upgrade badly). I'm getting battery read outs just fine. When I installed kernel 4.9, I was getting ACPI read error on bootup.

[L29Ah]

Which kernel and distro are you using? I'm still using debian jessie with kernel 4.8 (need to upgrade badly). I'm getting battery read outs just fine. When I installed kernel 4.9, I was getting ACPI read error on bootup.

Gentoo, 4.20.6. I get the other battery data just fine, but no current (/sys/class/power_supply/BAT0/current_now is always zero). I don't see any errors in dmesg, but here it is: https://bpaste.net/show/29375e46d82c

[L29Ah]

Not really an answer to my questions, but https://github.com/erpalma/throttled/ can adjust CPU temperature ceiling and display CPU power consumption (among other immensely nice things) that can be used to tune the battery performance of X210 and other newer Thinkpads.

[vladisslav2011]

it cannot report battery current in Linux

Yes. It looks, like average current readout is blocked in EC firmware for some reason. One may try to estimate current by dividing capacity change by time interval and average voltage, but in my case it gives inaccurate results. I have one non-genuine battery, that has incorrectly configured BMC/fuel gauge. It constantly tries to recalibrate and shows both negative and positive discharge rate spikes in 10%-0% charge range...

the CPU frequency throttling kicking in earlier than the fan decides to run at its full speed.

It is expected due to the way, it is implemented. Take a look here viewtopic.php?f=80&t=128156&p=833443#p833443

Also i really miss the original Thinkpad EC feature that allowed me to undercharge my battery;

Hmmm. In my case it constantly udercharges the battery. It reports battery as charged when it reaches 95% capacity and continues to slowly charge it in background. But as maximum charger voltage for 3s battery is fixed at ~12.55v, it does never charge it to 100%.

[mjg59]

https://forum.51nb.com/forum.php?mod=re ... d=32315397 suggests that they're working on the charging issue, but there's been no release yet. Nuvoton are, it turns out, unenthusiastic about giving out docs for these parts even to companies that have NDAs…

[vladisslav2011]

Battery "undercharging" issue is partially resolved.
It reports correct charge percentage now, but still undercharges for 1-2% due to lower voltage (12.5V instead of 12.6V). See "EC datasheet..." topic.
I don't think, that undercharge bug may be fully fixed in software due to the way, charging voltage switching is implemented in.
As voltage switching is done via GPIO97, it looks like there is star-like circuit of 3 resistors, connected to ground, charger out, EC GPIO97 and star center is connected to charger feedback pin, so driving GPIO97 high sources some current into feedback pin decreasing charger voltage and driving GPIO97 low sinks some current from feedback pin increasing charger voltage.
Maybe resistors are just not as precise as required...

[L29Ah]

Also i really miss the original Thinkpad EC feature that allowed me to undercharge my battery;

Hmmm. In my case it constantly udercharges the battery. It reports battery as charged when it reaches 95% capacity and continues to slowly charge it in background. But as maximum charger voltage for 3s battery is fixed at ~12.55v, it does never charge it to 100%.

I've actually meant i could limit the Lenovo's Thinkpad charge controller by a lower voltage when i don't need the max possible capacity so the battery would degrade much slower (LiIon cells lose capacity faster at >4V, especially at elevated temperatures observable in laptop batteries).

[jdhurst]

Just out of curiosity, can you find another hard drive for it, load Windows 7 or 10 on it and see if battery charging is OK in Windows.

I have an X230 spare machine (as per an above post here) with Windows 10 Insider (future Windows). It is on its second battery in 6 years and the current battery (not at all new) goes for about 3 hours still