Yet another charging issue - stop charge at 50%

MNT Pocket Reform
1

Something seems to be wrong with my Pocket Reform from August 30th.
I have a USB-C power supply that charges perfectly with -1A to -1.17A and 8.3V, as the OLED display shows me.
Charging in DE does not seem to be possible.
At around 42% the charge level drops to around -0.9A, at around 53% to around -0.06A.
I have already installed the update from August 26th.
Is this a hardware problem?
There is also a similar power drain problem when it is not plugged into the socket

1 Like
2

Try letting the batteries drain all the way then charge it again. Citing this post Charging is extremely slow - #4 by minute

As for draining while idle I’ve been following this Battery drain issues/questions

3

Thanks for your help…
After It was running down for a while, it runs this Minecraft and a Youtube Livestream for more then 2 hours on 1%… until the battery was empty
I charged now to 100% and will see what the drain is saying in 8 - 9 hours.

1 Like
4

So I’ve been having a similar thing and I tried letting it drain down and charge it back up a few times but all that’s done is let it charge back up to 46% instead of 38% and the odd thing is that I’m only getting like an hour and change of battery life total so I’m not even getting the “it stays at 2% for a million years” like above, it’s shutting down as soon as it ticks down to 2.

1 Like
5

OK so there is probably a different issue now, which is upower (?) thinking it has to shut down the system when the battery is reportedly near 0%. We need to disable this temporarily, otherwise it’s not possible to calibrate the gauge. As far as I know, the upower service is responsible for this. Could you try:

sudo systemctl stop upower

And then let it drain, does that work?

6

Oh, hah, yeah one command and 5.5 hours of continual use later and it drained down to a hard shutdown and now it’s charging properly on up again.

Thanks!

1 Like
7

@minute could the firmware maybe be changed such that userspace would think that the battery was much fuller than it actually is instead of thinking it is much more empty than it actually is? This is also an issue for the classic reform (remember the battery runtime benchmark by jeff geerling) where upower likes to shut down the device before lpc learned where zero percent actually is.

2 Likes
8

I would rather propose that we disable upower by default and people can enable it when they need it.