Battery incorrectly reported as critcally low at when almost full

This bug seems to be related to other simalar issues such as Bug #393008
I notice that the power manager applet doesnt recognise the AC has been unplugged. After running like that for an hour or so shut the machine down. Restarted later in the day and the indicator now correctly reported battery power and correct battery level however this soon changed to telling me the battery was almost empty (7% or so) but still with more than 30 min run time remaining.

On closer inspection running $ devkit-power --monitor-detail
The battery was incorrectly reported as

energy-full: 88.7883 Wh (or there abouts)
energy-full-design: 26.64 Wh

I didn't get chance to copy this as my machine shut down seconds after I noticed this. The machine is now on charge and looking at this again I notice the charge rate is extremely high which cannot be correct

  History (charge):
    1258072896 50.999 charging
    1258072865 50.609 charging
    1258072834 50.268 charging
    1258072803 49.683 charging
  History (rate):
    1258072896 715.928 charging
    1258072865 715.950 charging
    1258072834 715.850 charging
    1258072803 716.771 charging

The charge graphs are also all over the place. see screen shot at http://www.flickr.com/photos/bshephard/4098866663/

Bugs in Devicekit-power seem not to be handled by launchpad but I don't see any way to record the bug upstream either. Documentation also seems to be incomplete

Re-assigning to devicekit-power since it's reporting bad data and g-p-m is just copying it.

Invalidating upstream task, at least until we can confirm that it's not an Ubuntu-specific bug.

Actually, could you please execute the following command, as it will automatically gather debugging information, in a terminal:

apport-collect 473552

This will help us to find and resolve the problem. Bear in mind that you may need to install the python-launchpadlib package from the universe repository. Additionally, when prompted to give apport-collect permissions for Launchpad you will need to give it at least the ability to "Change Non-Private" data as it will be adding information to your bug report. Thanks in advance!

When reporting bugs in the future please use apport, either via the appropriate application's "Help -> Report a Problem" menu or using 'ubuntu-bug' and the name of the package affected. You can learn more about this functionality at https://wiki.ubuntu.com/ReportingBugs.

Note: this may be a dupe of Bug #410604.

Hello David

I'll do this as soon as I can but my aspire one is currently out of action and awaiting parts which seem to be on a slow boat from China. As soon as i get it back up and running again I'll sort this out.

Kind Regards
Ben

I am also using an Acer Aspire One A110 and Ubuntu 9.10 and I am having the same problem. However, I don't think it is a problem with Ubuntu because this problem was present on my Windows XP installation, too, before I installed Ubuntu. Unless XP and Ubuntu share some apparent flaw? Could this issue be rooted in the BIOS?

Oddly, I've been having this issue since November as well.

Thanks, it is just  a nuisance and doesn't really other me now. I did buy a new battery and even with that installed I get the same message.

It occurs on my eeepc900 and not with every distro.

Frank James
-

-- On Fri, 22/1/10, Buteo lineatus <email address hidden> wrote:

From: Buteo lineatus <email address hidden>
Subject: [Bug 473552] Re: Battery incorrectly reported as critcally low at when almost full
To: <email address hidden>
Date: Friday, 22 January, 2010, 18:20

Oddly, I've been having this issue since November as well.

--
Battery incorrectly reported as critcally low at when almost full
https://bugs.launchpad.net/bugs/473552
You received this bug notification because you are a direct subscriber
of the bug.

Could it be that https://bugs.launchpad.net/ubuntu/+source/gnome-power-manager/+bug/516023 is dublicate of this one?

Architecture: i386
DistroRelease: Ubuntu 9.10
LiveMediaBuild: Ubuntu 9.10 "Karmic Koala" - Release i386 (20091028.5)
Package: devicekit-power 011-1ubuntu1
PackageArchitecture: i386
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
Uname: Linux 2.6.31-14-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

Hi Ben,

Thanks for the update. Just to clarify, the root problem seems to be that devkit-power is reporting erroneous data, causing these low battery warnings and forced shutdowns?

It also still does not detect when you unplug the AC?

Thanks.

Hello David

Yes it seems like devkit-power reporting bad data is what's at fault. I didn't check the AC when unplugged because when I managed to rebuild the machine with the parts that eventually arrived I built it with 9.04 because 9.10 was unusable. I ran the above with live 9.10 so with no updates etc and just booted into 9.10 to run apport-collect and went back to my 9.04 install.

Also it might be worth while mentioning that I ran a live version of one of the 10.04 alpha releases and let the battery drain and I didn't see any errors but that was just a quick check and not a full test run of 10.04. I figured it could be down to the fact that HAL is in the process of being replaced with devkit-power but the process hasn't fully completed and it could just be an odd ball issue with my hardware and the implementation of hal / devkit-power.

I think I had the same problem on my other laptop too, A lenovo 3000 N100. sadly while I was investigating this the battery died it's death and now refuses to charge at all but I've given up with that one as I don't tend to use it much except for distro hopping and such.

If you want me to run apport-collect in any of the various states (both power and install such as with updates or without) let me know. It might take a bit if messing about installing to an external USB drive to get it going without ditching my working setup but I'll have time to play around with it more now.

Kind Regards
Ben

Great, thanks for the reply. I'm going to confirm this bug since you've provided more than enough information to confirm that it is, in fact, a bug.

If you can test further with Lucid Alpha 3 (released Feb 25) that would be most helpful. If the bug is still present then we'll just have to hope a developer gets time to look at this. :)

Actually, on second thought, if you can confirm that this is still broken in Lucid Alpha 3 (please provide devkit-power output), I'll create an upstream bug report for this.

I can confirm my case (https://bugs.launchpad.net/ubuntu/+source/gnome-power-manager/+bug/516023 ) in Lucid Alpha 3.

My case is: When AC is unplug I immediately see "battery critically low" then the hibernate/suspend action is happend according the settings in power management.

The same situation like was described here http://mail.gnome.org/archives/gnome-power-manager-list/2010-January/msg00001.html

So: devkit-power (or upower) reports enormous energy-rate for first time after cord is unplugged. GPM reacts on this first event not checking if situation changes after one second though then devkit-power starts to send right reports.

Plugging cord again not prevent to suspend or hibernate, but it is perhaps another bug in upstream https://bugzilla.gnome.org/show_bug.cgi?id=594739

How I can provide devkit-power output it seems that it is now upower.

But it seems upower --monitor-detail not fully implement devkit-power output

Will be gpm --verbose enough?

AC cable was unplugged then after about 5 sec. plugged back then Ctrl-C to prevent susbend

I file new bug https://bugs.launchpad.net/ubuntu/+source/upower/+bug/531190 because apport-collect is rejecting to post here:

Package devicekit-power not installed and no hook available, ignoring

Hi Konstantin, thanks for the information. Are you sure your issue is actually the same as Ben's? He never said anything about the problem only being immediately after unplugging.

Can you confirm that the issues are different, Ben?

Hi Ben,

No, I'm not sure. In my issue upower send "wrong data" just when unplugg or plug power cord.

Maybe it is not upower issue, something wrong with hardware or it is some transient process.

But gnome-power-manager reacts to this "transient process" with message and hibernating.

BTW. yesterday I install latest xfce4-power-manager. In version 0.9.90 it has upower in Depend

At first I was trying to set no action on critical battery. But in fact I have no problem with critical battery when unplugging with xfpm instead of gpm though gnome-power-statistics still show me power surges

In fact I asked here https://bugs.launchpad.net/ubuntu/+source/devicekit-power/+bug/473552/comments/9 about bug 516023 because that one is the same like my issue.

And I already marked 516023 like dublicate of this bug. But if you decide it is not I realise that I must change it back.

What you guess?

Hello David / Konstantin

i've downloaded Lucid Alpha 3 but unfortunatly I wasn't about to get any information what so ever out of upower --monitor-detail as even when unplugging and plugging back in it returned no output what so ever. Also the battery charge applet didn't give any options when clicked on. There was no way to view detailed battery status or any of the graphs except for the icon it's self which with the battery half charged changed to full when the charger was unplugged. this went back to the correct level and charging with the charger plugged back in. I noticed the power applet is no longer a dedicated power applet but a unified system status applet (but I don't remember it's exact name as I'm now back in 9.04). I didn't get chance to run apport-collect last night when I did this as it was getting pretty late so I'm going to do that now.

Ok apport-collect doesn't work in lucid because devkit-power is not installed although from what I've read it is but has just been renamed upower. Would I need to create another bug for that? I'm pretty new to Launchpad and I've not quiet got my head around all this yet.

From what I can see it looks like the same issue from 9.10 in lucid alpha 3 but I can't seem to find any way to read the output of upower to confirm that.

I'm running Lucid now so I can confirm the power applet is now called Indicator Applet 0.3.2 and is described as "An applet to hold all of the system indicators." I'm currently waiting for my machine to charge and then I'll let it discharge and see if I can tease any more information from it.

Hi Ben, thanks for your patience. Unfortunately, I can't get Lucid Alpha 3 working on my laptop at all right now because of another bug so I can't help you much with the upower stuff. As far as I can tell, apport won't actually give very good information in Lucid anyway because they haven't added the right hooks.

In other news, it sounds like this bug might be a dupe of bug #421985. Could you test the patched kernel given there <http://people.canonical.com/~smb/bug421985/> and see if it fixes this?

Hello David

I decided to bite the bullet and upgrade my aspire one to 9:10 so that I can do as much testing as possible on this. I upgraded and let it charge, run flat, and charge again with no patch applied. I kept the power history graph open so I could watch for any blips and at about 90% on the second charge it dropped to 20% suddenly. so I installed the patch and rebooted. When it came back up with the charger still plugged in it suddenly decided that the battery was discharging and that the battery was critically low at 92% so I pulled the plug out for a second then put it back in and that corrected it. It is still reporting the charge rate at 720+ watts. I'll let it run for a while and see if anything else happens.

I had to abandon testing on lucid alpha because i just couldn't get any information out of it at all but I'll try again before the release with the beta or something

Unplugged power but power applet still remained in the charging state. rebooted the machine and now this has corrected it's self. currently all the battery stats look normal but I'll give it a few charge cycles and report any other occourances.

Hi Ben,

Cool, thanks for all your help!

Just to make sure I understand you correctly, the weird reporting in charging/discharging went away after you rebooted once, right? It's not a recurring problem?

So the patch seems to fix the issue without creating any more?

Thanks again,
David

Hello David

Sorry I should have been a bit clearer.
When I posted the last comment I'd installed the patch and rebooted and waited for the battery to charge the last 10 percent or so.
It was still showing a charge rate of 720.0 odd watts.

After the patch install and reboot with the charger still in the indicator showed discharging even though the machine was still plugged in. I've never seen this behaviour before applying the patch. I unplugged for a second and plugged back in and it changed to charging.

When it did reach 100% I unplugged the charger but the indicator remained in the charging state. This time rather than reboot I plugged the charger back in and pulled it out again quickly. this time the charger going off line was properly detected and it's been discharging since (40% as I write)

As for the incorrect charge / discharge rate that is still incorrect when charging but so far as I can tell has been fine so far when discharging. I think this is what was suddenly getting corrupted causing the battery to show as critically low when it's 80% charged.

If it's all of a sudden reporting that the machine is pulling hundreds of watts from a battery with a design charge of 26.6 Wh it would work out the life if the battery to be in the minutes if not seconds depending on how far out of whack the discharge rate was being reported. Of course the high rate when charging is less of an issue because the power manager is aware that the machine is on AC power and won't shut down anyway.

The discharge issue being an intermittent issue there is no way of knowing if this is now fixed without using the machine for a week or so. I've managed to go for up to a week before without the machine shutting down on me but other times it's been four or five times in an evening.

Certainly the patch hasn't fixed everything but I can live with the bugs I've experienced so far since installing it.

It does seem like something is confusing mW for Watts somewhere a long the line.

I noticed somewhere the other day there was a bios update for the aspire one AOA 110-ab so I installed that but there was no changelog with it so I don't know what it was supposed to do but I've not noticed any difference since applying it but just thought it was worth while mentioning.

When I've tested this properly on 9.10 I'm also going to test on Fedora 12 and 13 alpha just to see how they behave too.

Thanks for your help
Ben

Think I just noticed another bug since i applied that patch. I didnt notice earlier because I had headphones on listening to music but my fan is running constantly even with acerfand installed. it doesn't feel hot at all but the the temperature reading is as follows

bshephard@bs-netbook:~$ cat /sys/class/thermal/thermal_zone0/temp
51000

Never seen that one before. It's usually between 50 and 80 and I'd have thought if something less than an inch from my chest was that hot I'd know about it by now :)

OK, I've linked to your comment in bug #421985 regarding the patch feedback. If you could monitor that issue and report any further requested information that would be great.

The patched kernel is actually rather old so it's possible it reverted some fixes that had been in the unpatched one you were running. (I'd suspect at least the thermal sensor is due to this.)

I've now moved to Fedora 12 and the install has revealed the cause of my fan control problem. The fan running constantly and possibly the high temperature value is down the the acer aspire one BIOS upgrade which breaks all third party fan control apps including in windows.

As for the original power management bug I've noticed that the AC being unplugged after a charge wasn't properly detected just as in Ubuntu. I'm going to run this for a while and keep an eye on the rate of charge / discharge. The charge rate was also 720 watts when I checked earlier so I'm looking to see if it suddenly reports the battery being critically low again. It seems the power history graph is not installed by default on fedora so I can't check that. I've not found any way of of installing this on fedora as yet.

I'm having the same issue with my Aspire One and the Lucid RC. It wouldn't be a problem except that when the power manager reports the wrong percentage of power (believing that my battery can hold 270 watts of power when it only holds about 48) the computer automatically shuts down (that's the setting I have for low battery in the power settings preferences). The power manager just seems to start thinking that my battery can hold much more than it can. I get "your battery may be broken" messages as well.

I'm also having this problem, and by this date it has not been fixed yet!.

Hi,

I just installed Ubuntu 11.10 and the issue occurred to me. I was on battery power and suddenly a pop up came up saying laptop battery critically low and need to suspend. After I clicked ok the laptop suspended. Though the battery capacity was 70% at that time. My laptop is Compaq Presario CQ40 616TU. I am using ubuntu since Lucid Lynx time and never had this issue before.

I tried the work around, though the suspend issue did not occur again. But the battery status now shows erroneous data for time. Sometimes it says 1.40 and suddenly 0.58 and again 1.22 like this. Every thing I observed on battery power no AC adapter connected.

Please let me know if you need any further information.

Thanks
Bishnu

The work around I tried is mentioned in
Bug #531190