Well, a copy/paste of a powertop run would be nice

Thx for the answers
. powertop pasted
http://paste.org/pastebin/view/21570

After this, and while I am preparing to flash the device, I've uninstalled the next apps:
- qtirreco (I didn't need it, not related but just in case, I note it)
- Wifi Switcher

And I still have (that I will uninstall if the situation does not improve)
- Queen BeeCon Widget
- Personal IP address
- ConnectNow internet connection switch

. After this, powertop again, pasted
http://paste.org/pastebin/view/21571

my 900 is getting connection of wlan but no web page is opening pls help me

this is epic fail from you. because error margins are huge when we are talking time spans not measured by clock but in ones head.

e: btw I am getting 8+ hours 3g constantly on, online in facebook and gtalk 24/7 plus couple hours surfing in the internet...

Okey. Thank you again for these logs!

----
Just to be sure and make measurement repeatable & correct. I have to ask?
- You closed all other apps
- Opened xterm
- changed to root and executed powertop
- immidiately minimized xterm and went to desktop and locked it (with switch or double click power, it doesn't matter, but screen has to be blank)
- Left it on stable place (so that accel doesn't make any noise to measurements)
- after 1 minute opened it and took logs to pastebin?
----

Maemos own wiki gives pretty good measurement for pretty low activity state:
http://wiki.maemo.org/N900_Software_...ement#Powertop
I bet that on those measurements there was no apps running.

So i repeated powertop on my own N900 and to my amazement total wakeups (last lines in powertop) were around 1000-800 in idle state. Rebooted my device and let it settle for 3-4 minutes so that everything is up and running. Made couple of runs of powertop and now i´m starting to get ~400 wakeups and most of time in C4 cpu state, which is pretty good and looks like my system is currently working as it should.

My setup currently
-2g connected
-no wlan connected or wlan search on
-no mail checkup
-no IM running in background (from your powertop log it looks like you have skype running?)
-No modified kernel(you have it, at least it looks like it?)
-Calendar widget, 2 desktops
-no bluetooth
-no gps

Next i´m going to turn stuff on and see what happens. On your system there is quite a much irq activity going on. I would recommend to reboot and do as i did and measure again. I´m sorry but this is only way to really learn your system. If this feel uncomfortable and stupid please just stop and use what ever feels best for you :P

Yes, exactly that way (powertop modus operandi).
Skype running? mmm, it could be, did I forget of unable it? mmm,
Kernel modified? Yes, and charged with the ideal conf. (850-500 MHz if I'm not wrong). Based on related posts, battery shouldn't notice this overclocking (but who knows).

So, I'm looking forward to see your results after you enable similar apps. Meanwhile, I'll do more tests ... but indeed, I'll rather re-flash the device and do the tests, and then install the apps and compare. But I need to be sure that with a flashed device my times are as the ones that are said here.

BTW, I'm not feeling unconfortable. I appreciate very much people giving their time trying to help me. I guess my first posts were coming from a frustration state, that I think I wouldn't have if meego hadn't appeared, and in consequence we would be in PR1.3 version.

I think that the first confussion I did is to think that these forums are from nokia people. Since these aren't, you obsiously don't have any need to hear so many complaints.

Having said that, let's see if I can reproduce the battery times that some of you seem to enjoy

PD: isn't any script that can predicts the time of battery based on the consume during the last minute? As it does my netbook, If I activate the wifi and increase the brightness of the display, I can see how the expected battery time decreases in the prediction.
Does something similar exist in linux?

As for cpu use, petrelli's powertop figures look fine. I'd also be concerned about skype though, especially if you're using 3g or wifi without maximum powersaving.

The way 3g works causes a single packet of data to keep the radio active for many seconds. (operator dependent)


Also one issue I used to have was with the "Automatically switch to wifi when available" option, or something like that. It would sometimes cause wifi to eat huge amounts of power when outside wifi range and using 2g/3g..
if wlan0/wmaster0 appears in '/sbin/ifconfig' output when on 2g/3g, I'd try do 'ifconfig wlan0 down' to cut off wifi.

I removed my skype account pretty quickly after I found out how much it and msn drains battery, jabber uses probably a tenth of the powers those two require. even a standard sip uses a lot less battery...

depends on the sip provider. Mine sends keepalives every 10 secs, so that'd kill battery on 3g pretty damn fast

Fast additional info, with yesterday reboot, kernel ideal loaded: 2 powertops (2 measures) at least

- wifi, internet, gtalk and skype disabled -> 400, 589 wakeups

- internet (3G) enabled -> 1068, 807, 52000 (I saw an update alert, ignored), 1004, 553, 463, 892, 719 wakeups

Let's see what I have in settings
+ connect automatically -> any connection -> 10 minutes (interval)
. Is this active even when the device is already connected?
+ exchange -> every 30 min during 8->16:00, perhaps too much, but one connection each 30 min is not so dramatic, isn't it? I change to every 4h.
+ GPS, enabled. Umm, I though it was disabled. Now I remember that I re-enabled it when I saw the the iphone didn't need to turn it off to have decent battery. However, I see the GPS working only when any map app is active, is it true? In any case, I disable it.

-> 567, 490, 482, 863, 497 wakeups

- gtalks and skype ONLINE -> 1461, 3398, 839, 603, 999, 796, 1353 wakeups

Notes: At the beginning, I disabled gtalk and skype, some days ago I started to just putting the IM in OFFLINE rather than disable the accounts. I understand that the effect is the same, but perhaps there are some differences (?)

Well, I don't know if these values agree with the poor battery time. I will re-flash the device and see what happens then. But I will install some apps because I need them for next days, so I'll be unable to do "clean" tests ...