Hey dr_frost_dk, if You - just like me - don't like bme, to say at least, consider installing advanced power monitor paqckage. In last version, you can choose from 4 reading modes - stock bme, bq27200, stock bme alternative, and bq27200 alternative. I think you will like it, cause both alternative modes are based on Your discharge graph algorithm (I suggested it to 412b).

advanced power monitor replaces stock status menu battery indicator, and also small icon in status arena, so, it's basically all you need to forget of bme @ userside (at least on parts clearly visible to user).

If You mix it with fuel gauge chip calibration (and, after that, use bq27200 readings), You get ~100% accurate battery percentage reports. Options based on your graph also works pretty well, except for huge spikes (~20% from high usage to almost-standby - some capacity "guessing" base on few voltage readings divided would be great), and overall wrong readings on 3650 mV and lower (at least for dual Scuds - all the time on such "low" voltage, it thinks that battery is almost empty, still, we have few hundred of mAh remain).

Long story short, I love this program so much, that I'll probably re-compile kp48 to contain "dangerous path" allowing bme and bq27200 to live together... Unless someone write better solution for getting bq readings (unfortunately, beyond my skill, as opposite to "trivial" .diff attaching and compiling)

Maybe ill try it, but for now my battery meter is fine for me, if i get the time i could look in to making some average points so it doesn't spike so much but instead "guesses" what current voltage state is.

Would be fun it the N900 could do a REAL TIME readout (50fps) of current voltage, then you can talk about spikes

Read time readings would be... Well, battery drain overkill, not to mention waking CPU

As for applets, don't get me wrong - Your idea is very good. It's just that Advanced Power Monitor disables vanilla monitor counterparts (thus they don't eat cycles/power), replacing it with own ones, + don't use widget space/resources, + contain many, easily switchable reading modes, including Your algorithm (also switchable between calculating on bme or bq27200 voltage readings).

Of course, one of best joys of FOSS is Freedom of Choice, so I'm not "pushing" you to switch

Right now im enjoying a fresh flashed phone, only advanced thing i have installed is 'Enhanced Linux Kernel for power users' just so i could OC to 700 for that little extra speed.

id any of u bought batteries from the battery test thread.do they really supply them or its spam.how can i b sure.i live in bangladesh.my sister lives in UK.is der ny shop in UK dat my sister can go and buy it from

Not easy to read.

are you saying that this thread is based on spam from the suppliers on page 1???????
Because i assure you like many other TMO users that this is no spam, i have used alot of time making this and alot of money, and not to mention the time used to check background on suppliers and dealing with the worst of them.....

So please say that you don't think all the work i did is SPAM!!!

EDIT: Also take any page and you will see some users talk about some of the battery's they bought on my recommendations from my tests.
DID YOU EVEN READ ANYTHING ON POST 1-3!!!!
Really pissed of right now if i understand you correctly.

AND if it was spam don't you think that moderators would have removed this thread mounts ago!!!!!!!!

@dr_frost_dk: trisha02 is just asking whether the suppliers are "spam" (which makes sense in a way considering that some of them they did sell bogus batteries). i doubt it's your hard work that is being described as such.
there's also concern since i believe the sister in the uk will be ordering the batteries and then shipping it to bangladesh.

@trisha02: there is no need to be writing in txt speak or sms language in an online forum. plain english will do. i can assure you (like many others here) that the batteries (specially the scud) are real as i've bought one myself ... dr_frost_dk has done all the hard work for us.

Sorry for off-topic, but trying to make Trisha consider using human language is quite pointless. I've tried looong time ago, and it seems that she/he/it only writes, without even reading answers

[/off-topic]

As for improving battery meter algorithm - I noticed, that You've set 3300 mAh as 0%. No big deal, but it's wrong (and result in mine being notoriously, for example, on "-1% capacity" ). In fact, BME turns phone on @ about 3200 mAh (in fact, little lower).

By "physical" limits, (with bme stopped) our bellowed N900 "dies" from out of power @ ~2850 mV. So I think setting 3000 mV as 0% is sane thing to do (by the way, bq fuel gauge got 3000 mV as 0% indeed - at this point device should be turned off unconditionally, but, for some reasons, it doesn't work on N900 bq implementation. Even thought EDVF is seto to 3000).

In my world that should equal ban, whats the point of us making alot of work in our areas and then some douche cant even take the time to read the basic (in this case post 1-3)

Back to battery:
Yes i have also seen my N900 go down to about 3200mV when it shuts off, and the battery tests are done from 4230mV to 3000mV, but most often i saw the N900 turn off at 3300mv and thats why i put 0% at 3300mv, but yes the real place would be 3000mV, only "problem" is that a 1320mAh battery will only have a maximum of 30mAh usable power from 3300mV to 3000mV hence why the N900 shuts off at between 3200mV-3300mV.

But again for full capacity tests it should be done all the way down to 3000mV.

As promised, after two complete discharge cycle until the 3200mV range, capacity of a 6 months hard-hammered-dual-japod is 2948 mA (twice the same reading!)