.

The amount of energy for the N800 to wake up and display that illuminated LCD message dozens (maybe hundreds) of times each year exceeds what the charger draws at idle plugged in continuously for an entire year.

This message actually wastes more electrical energy than it preserves.


.

Even a ballpark information would be useful. I remember when having ipaq 3870 it was like 90mA when idle with display off, 160 with display at full brightness, ~300mA at full brigtness with CPU running at 100%. Using audio added 20mA. Don't remember bluetooth. This was easy with ipaq since the battery charging chip provided discharge current directly (just as notebooks report it via ACPI).

You can fix CPU speed via tuning /sys/dev/ices/system/cpu/cpu0/cpufreq/scaling* files. Also absolute values does not matter much too. Even the relative ones when changing one thing only are interesting.

If you think you have time and other resources for it I can prepare some scenarios to minimize random noise. Sadly I don't have the tools so my way is to set the device to some mode, measure how long battery lasts and then divide battery capacity with time. A bit slow process :-)

ok, i will measure it.

I checked the files, most max at 400000, min at 165000. Which one file I can change to make the tablet go power conservative and very slow, or I have to change many of them? I will set them and measure over several days, I am curious too

files are
affected_cpus
cpuinfo_cur_freq
cpuinfo_max_freq
cpuinfo_min_freq
scaling_available_frequencies
....

N800/N810/700 Chinook/Diablo

bun

LEDs are bad for battery. I'm not sure how much a modern, small LED uses, but the old style 'big' LEDs would use some 20mA. If you run that LED for 30 days it would have used 14.4 ampere hours, which is enough to make a bad dent even in a car battery. I know someone with a solar-powered cottage who came to the cottage and found a flat battery, not because of the fact that there hadn't been any sun, but because the charger circuit had a nice big LED.

I once made the mistake of installing an application on my Palm T3 which would turn on its tiny LED if the battery level got too low, as a warning. Talk about stupid.. I had left the T3 at home over the weekend and on Monday it was flat and all my data and applications were wiped out (normally it would have lasted probably a week, at least, in that condition).

Yes, some of this has been d/w here, http://www.internettablettalk.com/fo...ad.php?t=12464. I think it would be interesting to make the tablet go
1) overclock
2) default
3) super slow

and measure the battery life and see whether it makes any 'practical' difference. And if it does, I will run my favorite progs and see if it survives under the reduce performance, I will stick with that when I am travelling up in plane or camping! I know Diablo is 'suppose' to have it built-in, I still want to measure it.

bun

set both min and max to your desired value out of scaling_available_frequencies (as root)
or you can set governor to one of scaling_available_governors

governors settings are conservative ondemand null powersave performance

are these in the order of decreasing battery use, i.e. conservative>ondemand>null>powersave>performance?

Just for the sake of asking, you mind to shed some light on their approximate power consumption, if you have done that in the past, i.e. on a scale of 1 to 100, pls scale them, or, is performance a step too, or it is just a term to link to the first 4? TIA,

my current setting is ondemand, is it the default too?


bun

By special request . I tried to put in the original message but it was too long so I shortened it down.

Enjoy:

https://bugs.maemo.org/show_bug.cgi?id=2285#c9

I wrote up a quick tutorial showing how to replace that message with your own.

i swear, your either a caffeine addict or have way to much time one your hands, qwerty12