I have a solar panel (without any accumulator) and it works perfectly under constant sunlight. But once I tried to use it on backpack while I was walking in town and it looked like N900 discharged even more. I think the main problem is that display keeps turning on every time there is enough energy to charge, can I disable this behaviour? Is there something in mce.ini for that?

I tried solar charging with a 12V 1.5W (that has degraded to less than 1W) panel with DC-DC converter to 5V/USB. This setup is really weak, it won't charge if the N900 is being used at all, including if the backlight is on.

However my N900's display stays off until unlocked - it stays off when I plug in/disconnect a charger... In Settings/Display there's a "Display stays lit when charging" option, is that on?

I think he means the brief moments when the display turns on to show you the yellow "Battery is charging" banner. It does the same to tell you that "Battery is fully charged." Mine does that and I do not have that "stay lit while charging" box ticked.

I do not know if this is configurable but I have a faint recollection of this coming with something I've installed. Most likely CSSU or Advanced Power. It may not be present on stock devices.

1W is indeed not enough. It equals to only 200mA at 5V, which is just about enough to keep the N900 alive but not much else. It certainly won't give enough power to charge it while in use.

EDIT: I also find the yellow banner irritating but for a different reason. I often go to sleep with my N900 on the charger and under the pillow (so that the alarm wakes up only me and not whoever sleeps next to me). When the screen turns on to tell me it's fully charged, I may accidentally press something.

I think it is advanced power, as my N900 has not seen cssu yet. And I believe also that this is not seen on stock devices.

@pichlo
are you really fiddling all night long?
I think there is little chance to push something, but not impossible.

--
if it really is apmd, then you may check and possibly stop that behaviour by typing
sudo /sbin/initctl stop apmonitord

--
Just tested it and indeed the yellow pop up came from apmd. But the screen was unlocked and lightened up nevertheless So maybe this is really a stock feature?

Guilty as charged (pun intended ).

I think 1W is actually a lot. My N900 can easily reach 4 days with a single charge (1250mAh), so that's about 13mA (65mW @5V).

Obviously, with the screen on it needs much more. According to battery eye my average is 12mA (guess I'm not a power user

Ehm, reinob, please notice I said while in use

But hey, 4 days!??? When I literally force myself to not touch the phone at all, just leave it on doing nothing, I achieve ~2.5 days. Gone are the days of my old Palm phone that achieved 9 days easily with heavy daily use.

I almost exclusively use it with GPRS/EDGE (3G only when I need to download something big and quickly -- never . It's connected on 2G about 10hours per day, calling mbsync (previously offlineimap) every 20min.

At home I occasionally connect via WLAN and do some stuff, usually via ssh (so screen is off, and I can use a real keyboard).

Oops. yes you're right, the display does turn on and stay on for a minute (which is my timeout). II gave up on solar charging until I got a beefier solar cell.

The problem with the 1W supply is that it can't keep the voltage up at 5V, and can barely keep the voltage above the voltage needed to charge the battery. Ideally you need 2.5W or so to emulate a regular USB port, and if you can do the full 4+W of a wall charger, that would be great. Otherwise it would be best to take the battery out of the phone and charge directly from the solar cell to the battery - that way you don't have to deal with charger inefficiencies and parasitic draw. But yes theoretically at 1W you should be able to charge a BL-5J in just under 1 day if you have no parasitic draw and no conversion inefficiencies.

The DC-DC converter I was using with my solar cell was only around 70% efficient too, I think, so down to 0.7W at the USB port.

Hey guys not a lot of specifics in this thread. I've spent a fair bit of time monitoring the charging profile of the N900 and it's got a fairly sophistocated charging system to check for current capability and voltage stability.

A solar panel is a constant current source. To simplify this as much as possible the voltage collapses when presented with a load. Using a voltage regulator will not help this either unless it is specifically designed with a solar panel and a regulated output in mind. Long story short, don't bother..

Charging profile (from memory, but measured and confirmed in the past) looks at the voltage stability and current in the first second or so. If this is not good the current used is locked off at ~90mA. Looks like it is charging but it's not a useful charge in reality. In the first few seconds the charge then is locked off at 500mA for a number of seconds, then following again the same for 600mA then 700mA. This is an approx. look at the charging profile as a charger with the centre pins shorted..