Edit:
This bug seems to occur only when using static IP with Wi-Fi. There is also a potential devel level fix for the problem described in the most recent comments on bug thread https://bugs.maemo.org/show_bug.cgi?id=6615

If you can confirm the bug only occurs with static IP, please post it on comments of the bug. Also if you are knowledgeable enough to use devel level fix, do post the results.

---

Suffering from battery time of around 6 hours or less on low usage? Then you might benefit from the following info.

Some people have problem with N900 and Wi-Fi. For some reason N900 doesn't recover from having been connected to Wi-Fi and will keep eating loads of battery even after closing the connection. Only restart will help. Simply searching for Wi-Fi without making actual connection doesn't trigger the bug for me. (The phone searches for Wi-Fi whenever you manually start any internet connection.)

So, do not use Wi-Fi unless you want your battery to die fast. Or reboot after using it. This is until a real fix comes for the problem.

With this fix I managed to get my idle battery time to increase over 30 times! (From bellow 7 hours to over 9 days.)

Note: This is about IDLE usage. Everything you do will use battery of course and the real usage battery time will vary depending on what you do. This is just about the bottomline battery usage.


It seems some people do not have this problem with Wi-Fi usage on their N900. It may be related to encryption method used or the Wi-Fi router connected to. In any case, some circumstances trigger a bug in N900 causing excessive battery drainage even after having disconnected from Wi-Fi. And the issue is fully reproducible. So, if you get it once with given Wi-Fi connection saved, you will get it every time unless you change some settings elsewhere than on the phone. That is, until the real bug has been fixed.

There is also a bug report made about the problem. If you have the issue, do vote and add more info the the bug report:
https://bugs.maemo.org/show_bug.cgi?id=6615


I had first experimented with various other things and have found out that normally there is no battery drainage problem with following things: Calendar widget, media player widget, foreca weather widget (only manual update), location widget and alarm clock. Also keeping terminal window open on the back doesn't drain battery noticably.

Wi-Fi tests were done with following settings (no problems prior to activating Wi-Fi connection for first time after reboot):
- Ask when needing connection.
- Dual GSM mode (uses 3G here at all times, 3.5 G when there is data activity - signal is a bit weak but doesn't matter).
- Widgets: Calendar, media player, foreca weather (only manual update) and location widget (on other than the viewed virtual desktop)
- Software: X terminal on background; Custom programs installed: OpenSSH client.

I did some experiments about battery usage on different cases. Prior to testing I rebooted the phone and kept it at idle for 20 minutes to let the battery voltage stabilize after the booting load. Results:
Idle without having used Wi-Fi: 54 mAh usage in 9.3 hours -> 5.8 mA.
Idle with 3G internet connection idle: 32 mAh usage in 1.9 hours -> 16.8 mA.
Idle after connecting and manually disconnecting from Wi-Fi (10 mW): 183 mAh in 1.0 hours -> 183 mA.

The difference is huge. The fully charged new battery charge is 1267 mAh, so idle 5.8 mA usage gives about 9.1 days, idle with 3G internet connection idle 16.8 mA usage gives about 3.1 days and idle after connecting and manually disconnecting from Wi-Fi 183 mA usage gives 6.9 hours. That is over 30 times the increase from idle to Wi-Fi bug triggered idle. And still over 10 times the increase from idle with 3G internet connection idle to Wi-Fi bug triggered idle.

The test times are high enough for accurate results and as such, it can be safely said that the Wi-Fi bug makes a really huge difference to the battery time.

I also tried to test having the Wi-Fi connection on but idle at 10 mW for an hour, but the phone had automatically disconnected it at some point. The resulting usage from part idling Wi-Fi 10 mW connection and rest pure idle with battery drain bug activated was about same as the 183 mAh an hour from previous test. Thus, there is no way to avoid the bug opening a Wi-Fi connection triggers as of now. At least unless some other connection setting than ask when needed keeps the connection alive.

If you want to test out these things yourself, here are some hints to use. On terminal, you can get good information about battery status with command "lshal|grep bat". It will include some unnecessary information as well. To only get current mAh remaining use the following alias:
alias b="lshal|g ng.c"

And then to check battery level I use "uptime;b" to get both current time as well as current mAh remaining. Then keep the terminal on background to see both old values as well as new values. Do note you can scroll the terminal text to see backlog.

Edit: I have moved advanced stuff about battery information into a separate guide. See it for more information:
http://talk.maemo.org/showthread.php?p=421349

Do post your own experiments to see how common this Wi-Fi bug is. And if there are any other significant bugs related to excessive battery usage.

--
Note to mods: I am aware my last posting with preliminary tests a day ago was moved into general battery time suggestions thread. But this about a critical bug that alone makes the battery to die very fast. As such I feel there really is the need for separate thread. I did send a message to the mod who moved my previous posting, but as there has been no reply and the matter critical, so I decided I just have to let people know about it now that my thorough tests are finished. I apologize for acting alone.

will test. sounds good

Well, as someone's signature said (sorry, i forget whose): "in theory practice and theory are the same. In practice they are not"
In theory, you have a point. However the 30-fold increase in battery life is hard to believe. If it turned out to be true, someone on nokia is going to at least get a pay cut (maybe a 100% one, if you get what i mean) for missing this
What is needed here is practical testing: don't use wi-fi at all and see how much the battery lasts. I'm betting no more than 3-4 days, since that's what my dumb sony ericsson gets (admittedly on a smaller battery, but not much)

I bet there's a way of just turning the wifi radio off. Be it a kill switch in /sys/bus/pci/drivers/ or somewhere. I doubt you need to reboot the device just to do this...

This 30-fold increase is about IDLE time. Ie. no usage at all whatsoever. That is just the base bottomline battery usage on which everything you do will pile up. So, yes on moderate usage it won't last more than few days.

This is about a clear bug that is triggered by using Wi-Fi. I have heard of some people who do use Wi-Fi with battery time of 2 days on low usage. So, the bug is not affecting everyone. But for those that it is affecting, it is a very critical extra drainage for the battery.

i took my n900 off charge at 8:30am, and connected to my company wifi at 9:30am.

i only have a 2G connection, but even so my battery is still at 80% by 14:30pm, despite being continuously connected to the wifi for five hours now.

From my one day of usage I have found WiFi has a nasty habit of crashing if you are getting a poor signal. It will still claiming to see the WiFi network (at least sometimes) but fail to connect to it.

This may be connected as its quite likely its crashed it will not enter power saving mode properly. Impossible for me to say from only a days use though, the battery is expected to only last 7 hours or so on its first couple of charges.

I'm constantly on wifi and my battery life is excellent. 3g is a killer mind you.

Lithium batteries shouldn't experience significant battery capacity changes during first few charges. There may be something small, but nothing significant.

As for Wi-Fi crashing, there may be partial crashing from the bug that causes the Wi-Fi not to be able to enter power saving state. But there isn't any complete crashing as I can make a new Wi-Fi connection and data will go just fine, even after triggering the Wi-Fi bug. And I did manual disconnection from the Wi-Fi and the drainage still existed, so something is clearly wrong on the phone side.

Yes, I know some people aren't experiencing the problems with Wi-Fi. As for 3G, yes it uses battery like there is no tomorrow if you do transfer a lot of data. But the usage is very low if nothing is being transferred.