Battery Drain

[Twiitcher]

So I noticed strange behavior of my battery. With the battery out of the unit overnight, no loss of charge. But inserted in the Pandora even if it's turned off completely, I get noticeable drainage in a few hours. I believe I lost about 30% charge overnight (@6-7 hours) Is this normal?

 

[tsh]

If everything is working properly, battery drain over night when switched off should be negligable (1% or so)

 

[WizardStan]

I haven't been able to reproduce this, but I remember there was someone who was able to get their Pandora in some kind of "zombie" state where they would initiate shutdown and then close the lid. It would be off, but not quite really off, or something; pretty much everything was off, but something had hung just enough to keep something going even though it looked like it was off.


Is this what you are doing? Closing the lid before it fully shuts down? If so, you may also be affected by whatever is going on (some good timing and a rare race condition, probably) and should try waiting until the Pandora is completely shut down (upper left corner blinks a cursor for a few seconds even after it looks like it has shut down) before closing the lid. It'll take a few days to confirm, but if that makes a difference and your battery doesn't drain so much overnight, then we can start an experiment of lid closing times to maybe finally figure out exactly what is going on.

 

[renegadechic]

it depends on the battery for me. its not a huge loss of charge but it does happen. i may have just mistreated my second battery though. it went from 1hr with 10% charge to 0% and shutting down without any warning so i think its just a bit duff. my other battery is fine though in comparison. i still get a good few hours out of it though

 

[tsh]

I have had some odd end of charge shutdowns too, and once saw the charge % jump from 15 to 30 when I unpugged too. Just tried now and it is stable at 30% going from from flat charge to discharge, so not repeatable. I keep meaning to add more battery logging but get distracted easily.

 

[joseLuís]

My unit always looses charge when it's off, and the battery is inside. And I can't figure out how (if) I'm aligning it bad. :huh:

 

[Twiitcher]

I verified last night, turned the unit completely off with battery at 100%. 9 hours later, powered it up, it was down to 88%.. so 12 percent battery loss with unit off. Everyone else getting this much drain?

 

[Twiitcher]

9 hours later, i get like 1%


try removing the battery and inserting, maybe it will behave differently

I reseated the battery before the test. I'll have a closer look at the contacts and such, they appear to line up fine and it seems to seat nicely. Thanks

 

[Twiitcher]

I'm still losing about 1% an hour with the unit completely turned off. The issue isn't the battery as out of the unit it does not drain at all overnight. I've reseated the battery multiple times....


I'm debating if this is worthy of a return, basically I have to constantly pay attention to battery life as it will totally lose it's charge in 4 days even if turned off. I'm out of ideas as to how else to test / solve this issue myself.

 

[Tor]

I'm debating if this is worthy of a return, basically I have to constantly pay attention to battery life as it will totally lose it's charge in 4 days even if turned off. I'm out of ideas as to how else to test / solve this issue myself.

Emptying a 4000mAh battery in 4 days when the unit is off is very bad. Most mobile devices drain a little bit of battery even when off, but the amount you indicate is way too much. My old Nokia N800 with a 1300 mAh replacement battery (the original was 1500) won't drain the battery in 4 days even when _on_. Off it could probably handle weeks (I never keep it off that long so I'm not sure). With the original fresh 1500mAh battery the N800 could stay on for almost ten days, when _on_. When off it uses a little battery too, because an alarm can power it on.


So there's something wrong if the Pandora sucks enough battery when off to notice at all, with that huge battery it has. I haven't got mine yet so I can't investigate.

 

[tsh]

To summarise what i remember people posting about this over the past few months. Several people have thought they had a problem with battery discharge whilst turned off, but seem to have silently resolved the issue (either with a replacement battery, or changing their usage).


From a few different reports, you should expect 1-2% loss when turned off for 24 hours. Noone has offered any results from a longer time 'off'.


If you see 10% battery loss in 24 hours, something is wrong (battery, pandora, charger, user). Troubleshooting this is probably complex, you would need to measure current from the battery whilst 'off' and that is not trivial to arrange. The only easy thing to do is try removing the battery for 24 hours to confirm that it is itself holding charge, and you have done that already. To some extent, it might just be an irritation and not a big inconvenience, so it might be hard to tell if it is more than a couple of people who have this problem.

 

[Twiitcher]

Just bumping...same issue, I've Pm'd Craig and waiting info on RMA for a replacement, barring any other new ideas form here.

 

[levi]

Is there any sort of log that the Pandora OS writes to to determine if it's doing something? Looking at my Fedora box here, the most active files in /var/log are cron and messages - the latter mainly containing ntpd sync messages, though I guess if ntpd and crond have shut down there might not be that much activity.


But presumably it is doing something if it's losing 10% of 4000mAh (what's that, 400mAh over ~8 hours - 50mA). I'm not able to find out how much juice something like the Pandora should take, but something's active (I just don't know if it's the CPU or not).

 

[WizardStan]

But presumably it is doing something if it's losing 10% of 4000mAh (what's that, 400mAh over ~8 hours - 50mA).

If it is is off then it is doing nothing, which is what makes the battery drain so mysterious.

 

[levi]

If it is is off then it is doing nothing, which is what makes the battery drain so mysterious.

True. I'm working on the theory that it's clearly doing something, if the battery drain is less when the battery is removed.


Of course, one possibility I haven't accounted for is that it's the battery manager recalibrating. I found the battery gauge on the ubuntu install on my netbook to behave a little erratically until I'd drained the battery twice (with the machine on for the duration).

 

[tsh]

Is there any sort of log that the Pandora OS writes to to determine if it's doing something? Looking at my Fedora box here, the most active files in /var/log are cron and messages - the latter mainly containing ntpd sync messages, though I guess if ntpd and crond have shut down there might not be that much activity.


But presumably it is doing something if it's losing 10% of 4000mAh (what's that, 400mAh over ~8 hours - 50mA). I'm not able to find out how much juice something like the Pandora should take, but something's active (I just don't know if it's the CPU or not).

I don't think this will help you. if the OS is running, the battery will drain 100% in 30 hours, even if the OS is not doing any work. You are seeing less than that, so much if the system is shut down. Maybe there is a fault with the charging circuit, or a software problem which is leaving the charging logic in a bad state. Clearly some things remain on, for the real time clock, but this should be very small (and from a different battery). No log files are likely to help (and the default if for log files to be held in RAM, so lost on a reboot)

 

[Tor]

If a running (but idle) OS can drain 4000mAh in 30 hours then the OS isn't working very well yet.. like not powering off any of the non-used parts. Comparing to other similar devices it should be possible to get 550 hours or more out of an idling Pandora OS. Or in any case an order of magnitude more than 30 hours. Obviously when turned off it should last much much longer than even that. That been said, I concur with everything else that tsh wrote about possible causes.


Edit: To clarify, what I said about 'should' above is what we should expect with a fully working power management feature in the Pandora OS, not what we should currently expect (as I understand it we don't actually have the PM part optimized yet. I haven't got my Pandora still so I can't test how good it is myself). When powered off, though, it should last long long long.. if it doesn't, there's something not working right.

 

[levi]

The Wiki has some estimates for expected battery life in various modes. It looks to me like there's still some room for improvement in low power mode scores (though it looks like it hasn't been updated for hotfix4 yet), but the expectation is for somewhere under 30 hours runtime in this mode at the moment.


It also notes that slowing the CPU clock to 14Mhz doesn't do much to reduce current drain - only turning off the radios does that. That seems odd to me.


Anyway, the issue here seems to be with a shut down Pandora, not a sleeping/low power Pandora. It shouldn't be running the CPU at all in that mode (as tsh says).

 

[tsh]

It also notes that slowing the CPU clock to 14Mhz doesn't do much to reduce current drain - only turning off the radios does that. That seems odd to me.

Observations based on my pandora. See the talk page for more detail, and if you have a unit, please contribute - it would be good to know the spread of performance.

 

[Twiitcher]

Bumping again in the hope I can get a response (official) about an RMA for my bsttery draining unit. Already PM'd Craig, no response there.