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.