Notice: Trying to access array offset on value of type bool in /srv/www/vhosts/ on line 649 FS#136 : rather high battery discharge while Pandora is powered off

OpenPandora Main OS

  • Status Closed
  • Percent Complete
  • Task Type Bug Report
  • Category Core
  • Operating System All
  • Severity High
  • Reported Version SuperZaxxon Beta 4
Attached to Project: OpenPandora Main OS
Opened by Nils Kneuper - 01.07.2010
Last edited by Michael Mrozek - 21.07.2010

FS#136 - rather high battery discharge while Pandora is powered off

The Pandora does seem to use too much energy while it is turned off. A testcase: The pandora was turned off over night (about 8 hours) and battery state dropped from 98% to 96%. When removing the battery from the pandora for the same amount of time, the state is not changed at all. This is based on the zaxxon firmware release with hotfix3, beta2. Problem: if you don't use the pandora for a while and forget to remove the battery you can cause a deep discharge which might damage the battery. This is also a problem if the Pandora does eg shutdown itself since power is out and you can't directly attach it to a charger. Possible causes: * Design fault in the board. * Too much powered on while the device *should* be completely powered off. The only things that should get power are the clock and the power button. This should basically not drain much. Regarding the clock: The PalmPre is based on the OMAP3, too. There you basically have no discharge when it is powered off, so it is unlikely that the fault is in the OMAP3 itself.
Closed by  Michael Mrozek
21.07.2010 00:55
Reason for closing:  Not a bug
Additional comments about closing:  See last comment.
tsh commented on 02.07.2010 07:45
Your test-case is too short, I don't believe battery state is trackable to that precision. I repeated this test with a 12 hour off period, and recorded 100->99, which is the same - anywhere between 0 and 2%. I would suggest a delta between two measurements of less than 5% is too noisy to extrapolate from. Also, it may be best to start the test at less than 100%, since the scale may saturate.
Nils Kneuper commented on 02.07.2010 09:49
Hmm, might be some wrong results (or battery monitor is acting up, no idea). I repeated the test over 12h. When I shut the pandora down current left was at 77%, after 12h at about the same temperature (something around 26°C), it is still at 77%. Is there a way to easily read out the amount of power left in the pandora? As in: which sysfs interface is used to get the power information?
Nils Kneuper commented on 02.07.2010 10:35
Okay, had a look through /sys and it looks like i found some of the relevant stuff. The information used is in "/devices/platform/i2c_omap.3/i2c-adapter/i2c-3/3-0055/power_supply/bq27500-0/". Just use 'cat' on the files there to get the readout. These are the relevant values: capacity -- remaining capacity in % (100=full, 0=empty) current_now -- discharging/loading rate in micro ampere (negative: discharging; positive: loading) present -- is a battery inserted? (1=yes; 0=no) status -- Discharging/Charging temp -- battery temperature in decidegree celsius (value of 250 = 25.0°C) time_to_empty_avg -- time left in secs based on drain on some average, not used for displaying in the GUI! time_to_empty_now -- time left in secs based on current drain, used to display in the xfce applet time_to_full_now -- time left for loading with the current usage, only relevant when loading, just an estimation, currently unused type -- should always read "Battery" since it is a battery... uevent -- empty voltage_now -- current voltage left in micro volt So the interesting data is most likely voltage_now before turning the pandora off and after turning it on again. Listed this information in the wiki:
tsh commented on 02.07.2010 12:47
My assertion is that battery readings of 96% and 98% are not significantly different. Particularly since there may be a small degree of relaxation when the battery current changes (recently charged cell reads higher than one which has been standing for an hour, recently discharged reads lower, by some mV). Good find on the status info. Note that there is no direct reading of watts, just voltage (which is a function of capacity, temp and charge) and capacity is probably temperature/history dependant too.
Nils Kneuper commented on 03.07.2010 09:17
Okay, I write a tiny script for checking the battery status and here are the results after a night off (running the script before and after): starting battery status Sat Jul 3 02:05:13 UTC 2010 voltage_now 4135000 capacity 97 done gathering battery status starting battery status Sat Jul 3 11:15:22 UTC 2010 voltage_now 4135000 capacity 97 done gathering battery status As you can see there is basically no difference, so the report is in fact invalid and was based on some strange things going on. Please mark it "invalid" and close it unless some new stuff comes up.
Nils Kneuper commented on 05.07.2010 10:40
Okay, behavior last night makes things appear *really* strange: starting battery status Sun Jul 4 22:29:11 UTC 2010 voltage_now 3778000 capacity 39 done gathering battery status starting battery status Mon Jul 5 11:29:30 UTC 2010 voltage_now 3780000 capacity 44 done gathering battery status As you can see, the battery gained 5% in its status while not plugged in and completely turned off. Voltage increased by 0.002V. The pandora was only powered up for some 2mins to start the battery_status script and shut down right afterwards. There should be no effect due to charging or heavy usage of the battery. So yeah, in general there seems to be no high battery discharge while turned off.


Available keyboard shortcuts


Task Details

Task Editing