**** BEGIN LOGGING AT Sat Apr 28 03:00:00 2012 Apr 28 04:20:21 guys Apr 28 04:20:27 tell me if I made the right decision Apr 28 04:20:32 my girlfriend started playing minecraft Apr 28 04:20:52 so im like, what will get you to *not* play this game Apr 28 04:20:55 she says coin dozer Apr 28 04:21:08 so I get out the N900, install nitdroid and coin dozer Apr 28 04:21:27 i dont know which is worse Apr 28 04:27:14 Sazpaimon__ why not play minecraft? Apr 28 04:27:36 challenge her to build a computer there Apr 28 04:28:42 because minecraft is stupid Apr 28 04:29:12 no it's not Apr 28 04:29:25 sounds like you haven't played it! :p Apr 28 04:29:40 why not play with your girlfriend in multiplayer? Apr 28 04:36:00 I have played it Apr 28 04:36:02 its stupid Apr 28 04:39:08 Sazpaimon in comparison to what? Apr 28 04:40:34 it doesnt compare Apr 28 04:40:39 its in a league of its own Apr 28 04:40:42 yeah, it doesn't :) Apr 28 04:40:46 a league i want no part of Apr 28 04:40:52 it's more like painting than playing some "game" Apr 28 04:41:11 this is coming from someone that runs a minecraft server Apr 28 04:41:18 i just dont like it Apr 28 04:41:31 what sort of server do you run? Apr 28 04:41:53 just a vanilla server Apr 28 04:42:02 I didn't really bother much with it either before I found there was a technic pack and a multiplayer version too Apr 28 04:42:19 so it was great fun building machines with a couple of friends Apr 28 04:42:22 :) Apr 28 04:43:09 but anyway, it is more like a sandbox where you have to make you own fun, and some of it can be pretty intimidating for sure heh Apr 28 04:43:43 to me it's very impressive from a simulation perspective and social interaction Apr 28 04:45:52 on a side note a microsim fits well on an N900 Apr 28 04:46:09 once you align it correctly Apr 28 04:50:47 heh, I made a ghetto solar usb charger an hour ago, it's charging a tiny phone right now. anyone tried a solar panel with n900? Apr 28 05:02:37 joga: so long as your voltage is between 4.5 and 6V it'll attempt to charge Apr 28 05:02:51 So long as the voltage never goes over 20V, you won't ruin anything. Apr 28 05:03:46 robbiethe1st yeah, these have a decent 7.5V or so and I have 3 of them in parallel (with a 5v voltage regulator), it's the first time I just actually test them to charge something Apr 28 05:04:35 N900's TI charging chip is surprisingly tolerant; if the voltage droops too much however, it will slow down/stop charging... Apr 28 05:23:03 robbiethe1st ok Apr 28 05:23:30 I'll see how the little phone does first if I go to sleep and leave it there Apr 28 05:23:48 sun's shining here already Apr 28 09:51:19 when using gst-launch for streaming video out of n900, is there a way to focus the camera? Apr 28 09:52:18 I guess I could do it at least with a custom software that activates the gstreamer pipeline.. maybe even python, direct-on-device? Apr 28 10:00:24 at least if I'm able to acces the v4l2 device directly (can I get the fd handle from gstreamer? can I reopen the device? at least I could find the open fd from proc :P) Apr 28 10:00:33 ..I should be able to set the focus to 'auto' Apr 28 10:01:10 oh, it gives device-fd, so it should work alright. next question, is there an app that already does this?-) Apr 28 10:03:35 hi Apr 28 10:09:24 guys, recently i bought a nokia phone to my GF. for my surprise, the battery inside was the same (BL5J) as in n900, but 1430 mAh, instead of 1320, so we swap the batteries. The problem is, that lshal never shows more than about 1200 mAh charge, capacity is correctly reported (1415). Anything I can do to fully charge the battery? Apr 28 10:14:58 hi there, got a question. Apr 28 10:15:25 is it possible to continuous capture images with the n900? Apr 28 10:15:38 maybe over the shell? Apr 28 10:16:28 used to do the same with my canon eos Apr 28 10:21:05 anyone? Apr 28 10:30:19 phone-control has a option to take pics but ime they dont come out well Apr 28 10:46:47 freemangordon: your question is incorrect Apr 28 10:47:00 :) why? Apr 28 10:47:22 DocScrutinizer ^^^ Apr 28 10:47:34 lshal never shows more than VS I can do to fully charge the battery Apr 28 10:47:46 figure :-D Apr 28 10:48:29 well, ok. However, it seems bme really does not want to charge that over 80%. Apr 28 10:48:50 what makes you think so? Apr 28 10:49:03 that is what lshal shows Apr 28 10:49:13 Charging is OK, the meter isn't Apr 28 10:49:36 hmm, why is that? Apr 28 10:49:40 I thought he already noticed, after I critisized his Q, ShadowJK :-D Apr 28 10:50:09 bme does not know what current for how long? really? Apr 28 10:50:29 it's somewhat hardcoded for 1200mAh-ish battery :P Apr 28 10:51:21 And it has some voltage curve matching stuff going on too, mixed with the current integration stuff Apr 28 10:51:57 so, if i understand what you say, the battrey is fully charged, but i have 7 bars? And that is ok? Apr 28 10:51:59 get bq27k-detail to get an idea Apr 28 10:52:32 DocScrutinizer, your i2c script would do the better job? Apr 28 10:52:34 If you get green light with bme charging, it's more than full Apr 28 10:52:39 or bq27200.sh Apr 28 10:53:00 the one in wiki? Apr 28 10:53:37 (except if you insert charger when battery is close to full (80-90 percent or so), green light might come on right away, but it's still charging Apr 28 10:54:07 The charging, green light, and bme battery meter all operate independently.. Apr 28 10:54:22 ok, thanks a lot. Apr 28 10:54:36 and bars - percentage mapping isn't linear :) Apr 28 10:55:37 yeah, i know that, but lshal showing 80% and battery meter showing 7 bars 15 minutes after green led made me curious WTF is going on Apr 28 10:56:20 freemangordon: yes, the one on wiki (I guess) Apr 28 10:56:29 ok, thanks Apr 28 10:57:03 Also it's designed to show "0" relatively early on, so that device runs long enough to find charger cable and more Apr 28 10:57:23 yeah, know that too Apr 28 10:57:35 The hardware charge meter has nice accuracy, but it's 0 point is kinda low Apr 28 10:57:52 And default parameters in it are really wrong Apr 28 10:58:12 yeah, it needs calibration Apr 28 10:58:22 but it self-learns and adjusts those, however under bme control the required conditions for self-learning are never met Apr 28 10:58:50 DocScrutinizer, what has happened with your project (together with pali) for bme replacement? Apr 28 10:59:15 bbl, food Apr 28 10:59:26 bb Apr 28 10:59:27 I dunno about what pali does, I never had any such project actually Apr 28 10:59:48 but you wrote some specs iirc Apr 28 10:59:54 sure Apr 28 11:00:02 I had created driver Apr 28 11:00:10 which pali generously disregarded Apr 28 11:00:15 :D Apr 28 11:00:31 so, hof far are we from getting bme replaced? Apr 28 11:00:46 ready for testing Apr 28 11:00:54 wow, where? Apr 28 11:01:03 I wrote info somewhere in TMO usb host mode thread Apr 28 11:01:23 but that deserves its own thread aiui Apr 28 11:02:14 freemangordon: btw all I wrote was a 50% done API spec for a bq24150 driver Apr 28 11:02:28 nothing at all regarding bme replacement Apr 28 11:02:37 DocScrutinizer, my driver has interface as you described in spec Apr 28 11:02:51 ohh? never realized Apr 28 11:02:57 :-D Apr 28 11:03:00 I exported all settings via sysfs Apr 28 11:03:18 all what was written in bq datasheet Apr 28 11:03:41 so, now we just need a wrapper to i2c for bme? Apr 28 11:03:44 (no idea if TI bq chip has some undocumented registers..) Apr 28 11:03:54 freemangordon: nope Apr 28 11:03:56 no, we need to remove BME :D Apr 28 11:04:03 ouch Apr 28 11:04:12 freemangordon: with pali's (my) sysfs API you don't use i2c anymore Apr 28 11:04:24 yeah, but bme uses it afaik Apr 28 11:04:32 Pali: it has no undocumented registers Apr 28 11:04:44 but kernel driver doing all what is doing BME... Apr 28 11:04:56 and i am not sure we have all the knowlegde needed to integrate in the same way bme does Apr 28 11:04:56 though bme is writing to register #46 or sth Apr 28 11:05:00 X-P Apr 28 11:05:14 do we? Apr 28 11:05:27 so then my driver does not doing what BME :D Apr 28 11:05:38 forget that Apr 28 11:05:57 we neither need that nor do we want or like to do what bme does Apr 28 11:06:02 freemangordon, you need to replace hald-addon-bme and libbmeipc library Apr 28 11:06:17 for both exists open source implementations Apr 28 11:06:23 hmm Apr 28 11:06:48 you need only rewrite both programs to use sysfs interface from bq2415x and bq27x00 driver Apr 28 11:06:50 here? http://gitorious.org/mer-toggles/hald-addon-bme Apr 28 11:06:54 yes Apr 28 11:07:06 it is simple addon for hal Apr 28 11:07:09 nothing more Apr 28 11:07:12 hmm, let me see Apr 28 11:07:17 also check lshal -u on real n900 Apr 28 11:07:23 that version is for n8x0 Apr 28 11:07:41 didn't understand that Apr 28 11:07:47 you need to export BME properties Apr 28 11:07:58 you can see it with command: lshal Apr 28 11:08:05 aah, ok Apr 28 11:08:05 bme is like controlling your washing machine by hand via switching between max heating (95°C) and cold to keep the water @ 60° for the laundry, then manually stepping thru drain water, rinse, tumble, whatnot Apr 28 11:08:45 our bq24150 is the washing machine that perfectly knows to do all that without you(/bme) forcing that manually on it Apr 28 11:09:07 and libbmeipc is library which communicate with BME Apr 28 11:09:31 so needs to rewrite to use again /sys/power/bq... instead some BME socket interface Apr 28 11:09:49 well, why not do it then? Apr 28 11:10:07 no time :D Apr 28 11:10:24 aah, yeah :D Apr 28 11:11:15 freemangordon: basically it's still non-comparible to a stock system Apr 28 11:11:21 freemangordon, here is that thread: http://talk.maemo.org/showpost.php?p=1156614&postcount=973 Apr 28 11:11:27 thanks to friggin bq27200 driver Apr 28 11:12:09 I.E. on a system set up to run that sys/bq interface youz can't use original bme anymore Apr 28 11:12:30 you cannot use both BME and bq :D Apr 28 11:12:46 concurrently obviously not Apr 28 11:14:27 but tanks to the way sysfs and modules work, it isn't even an alternative at boot time (unless you make sure bq27200.ko isn't modprobed when you plan to boot to the bme version of system) Apr 28 11:15:48 almost same applies to different versions of hald-addon-bme and libbmeipc Apr 28 11:15:49 And why should I want that? i.e. having HW charger, why should I want to use those spaghetti called bme? Apr 28 11:15:50 DocScrutinizer, both modules (bq2415x_charger.ko and bq27x00_battery.ko) will be always blacklisted in kernel-power Apr 28 11:16:19 freemangordon: no idea :-D Apr 28 11:16:40 why the should we care about so-called "bme" version? Apr 28 11:16:47 I will look at LKML what problems have kernel devs for integrating bq2415x module into upstream Apr 28 11:16:57 ...and try to fix that Apr 28 11:17:02 Pali, it is stable in KP50? Apr 28 11:17:03 good plan Apr 28 11:17:18 I tested it and worked fine with pre v50 Apr 28 11:17:23 ok Apr 28 11:17:30 binary .ko file on my page is for pre v50 Apr 28 11:18:21 current version does not support autoselect changer current limit Apr 28 11:18:40 Pali: I got 3 weeks holiday now, *maybe* I could invest some time into this development Apr 28 11:19:06 but in upstream kernel is isp1704 driver which provide inromations about connected charger type Apr 28 11:19:08 and I have 4 days, lets do something about it Apr 28 11:20:00 Pali: this upstream isp1704 driver for _sure_ conflicts with our monolitic kernel USB driver Apr 28 11:20:12 yes I know Apr 28 11:20:33 but code for isp driver is copied from musb 2.6.28 Apr 28 11:21:00 I will look if it is simple to use our isp code in musb for that too Apr 28 11:21:13 ok, I am going to port hald-addon-bme Apr 28 11:22:08 Pali: check hostmode kernel for a printk dmesg string like "some bastard tried to probe for charger" Apr 28 11:22:11 OWWTE Apr 28 11:22:23 s/WW/WT/ Apr 28 11:22:23 DocScrutinizer meant: OWTTE Apr 28 11:23:54 doing D+/- short probing in isp1707/4 breaks any actual data transmission over USB that may run concurrently Apr 28 11:24:40 that'S the root cause why we need to remove bme for h-e-n Apr 28 11:27:10 that and of course the boostmode we operate bq24150 that conflicts with bme chargemode. Anyway we enable vboost and that gets detected by bme via vbus sysnode. bme thinks vbus == charger/host attached, and queries sysnode for fastcharger detect which in turn causes kernel to switch isp1707 to D+- short detect which breaks our USB hostmode setup Apr 28 11:29:12 * ShadowJK doesn't use bme Apr 28 11:29:53 so if your bme 'replacement' does the same (read fastcharger detection sysnode), it's as incompatible to h-e-n as bme_original Apr 28 11:32:12 since we use booston script to enable bq24150 boostmode via direct I2C, there's no way for kernel (yet) to know vbus is not applied by external charger/USB-host but rather endogenic, and thus kernel refusing to mess with 1707 to detect fastcharger when somebody reading that sysnode Apr 28 11:32:26 it doing upstream isp kernel driver Apr 28 11:33:43 my driver has 2 modes: one auto (prepaired mode for isp or something else which select charger current limit automatically) and manual mode when automode is disabled Apr 28 11:33:47 of course your/our bq24150.ko might know it's driving the chip into boostmode so "the kernel" now could know it's endogenic and thus tell musb kernel driver to reject D+- probing Apr 28 11:34:22 so when you activate hostmode, driver changed to manual and then do not listen for any auto changes Apr 28 11:36:13 :nod: not wrapping my head around it completely, but on first touch it sounds about correct way Apr 28 11:42:32 next will be to implement simple battery charging into u-boot Apr 28 11:47:26 well, that's extremely simple Apr 28 11:48:19 bare bones script like my PoC script will do Apr 28 11:48:33 max USB current = 500mA Apr 28 11:50:19 freemangordon: not wiki, http://maemo.cloud-7.de/maemo5/usr/local/sbin/ Apr 28 11:50:45 i found a post from estel with link to it Apr 28 11:51:05 Pali: (uboot) http://maemo.cloud-7.de/maemo5/patches_n_tools/charge21_500mA-USBlimit__UNTESTED.sh Apr 28 11:51:38 though that's for sure way too sophisticated and "large" Apr 28 11:52:02 it is same as charge21.sh? Apr 28 11:53:19 no, it's "patched" for max 500mA USB Apr 28 11:53:29 of course you can patch http://talk.maemo.org/showthread.php?p=658278#post658278 same way Apr 28 11:57:33 ok, 500mA limit is good for usb charger and also wallchanger Apr 28 11:58:15 freemangordon, I updated u-boot deb packages & source code on gitorious Apr 28 11:58:41 great Apr 28 11:58:47 # Register 0x1 Apr 28 11:58:49 # 8: 00 = 100, 01 = 500, 10 = 800mA Apr 28 11:58:50 # 4: 11 = no limit Apr 28 11:59:01 i2cset -y 2 0x6b 0x01 0x48; Apr 28 11:59:10 3 problems: mkimage not worked, syntax error in debian/rules and bad size (only half) of RAM in u-boot rx51 port Apr 28 11:59:16 fixed Apr 28 11:59:54 DocScrutinizer, 800mA is for what type of changer? Apr 28 12:00:10 along with cursor garbadge and eMMC -11 ? Apr 28 12:00:18 *-110 Apr 28 12:00:28 same version as yesterday Apr 28 12:00:31 0x48 = 0b0100 1000 = [(8,4) = 01 = 500)] Apr 28 12:00:32 ok Apr 28 12:00:44 only that 3 small fixes Apr 28 12:01:09 Pali: 800 is for chager that can deliver max 800 ;-) Apr 28 12:01:46 of course, but how BME detect this charger type? Apr 28 12:01:56 not at all Apr 28 12:02:36 unless USB ENUM would advertise such strange host that allegedly can deliver 800mA Apr 28 12:02:45 I think it's not possible Apr 28 12:02:54 In isp upstream kernel code is only 3 options: none, usb charger (500mA) and wallchanger (no limit) Apr 28 12:03:06 :nod: Apr 28 12:03:16 about what we can detect Apr 28 12:03:36 where "none"=100mA Apr 28 12:04:00 = 0b00xx xxxx Apr 28 12:04:24 in register #01 Apr 28 12:04:30 it is safe in u-boot enable permanent charging (e.g. when charger is not connected too) with limit 500mA? Apr 28 12:04:41 yes, mostly Apr 28 12:05:08 ok Apr 28 12:05:08 as safe as any random USB fan/lamp/rocketlauncher/whatever gadget Apr 28 12:06:06 USB mugwarmer even worse - they use up to 1000+mA without any "safety check" Apr 28 12:06:27 that's how you fry one out of 5000 laptops Apr 28 12:06:47 not nowadays anymore, I guess Apr 28 12:07:57 per definition USB has to be short circuit resilient Apr 28 12:08:57 one *some* PCs windows actively detects obercurrent/short on USB and powers down the USB in case of error Apr 28 12:10:25 since that doesn't improve your life in any way, it's getting rarely done like that nowadays Apr 28 12:11:13 nowadays each PC USB host shall have an electronic fuse that triggers on a safe threshold @ >=500mA Apr 28 12:11:27 and OS should keep sticky fingers out of that Apr 28 12:12:05 most PC have a threshold of >=1000mA Apr 28 12:13:51 I.E. they use only one electronic fuse with 1000 or even 2000mA and feed the vbus rail to 2 or 3 or 4 USB ports in parallel Apr 28 12:15:12 so you could either use 4 compliant 500mA loads on 4 USB ports, or one spec-breaking mugwarmer 1500mA+ on one USB port while not using the other 3 ports Apr 28 12:16:08 or s/mugwarmer/UMTS-stick, 3.5" HDD, whatever/ Apr 28 12:16:48 Nokia been extremely careful not to break USB spects with N900 Apr 28 12:17:19 regarding the max allowed current under any phase of ENUM/detection Apr 28 12:17:44 I think it's overly prudent Apr 28 12:19:59 but then, IANAL and maybe they have their very own history of paying refund for destroyed laptop USB ports, like they have with exploding batteries Apr 28 12:20:35 if any of the chinese USB-acme-gadget (mugwarmer) manufacturers was as prudent, they couldn'T sell a single unit Apr 28 12:24:01 My 3.5G USB dongle came with a Y-cable to split the vbus load on two USB ports of PC. They obviously had less scruple to violate specs (basically all 2G/3G dongles do) Apr 28 12:24:59 either that or they violate GSM/UMTS specs Apr 28 13:10:35 freemangordon, I uploaded my last version of bq2415x_charger driver and .ko binary for kp v50: http://atrey.karlin.mff.cuni.cz/~pali/bq2415x/ Apr 28 13:18:11 Pali, thanks Apr 28 13:23:56 freemangordon, is booting from sd card via u-boot working? Apr 28 13:30:37 Meanwhile: Portable Battery-Powered X-ray Machine Guarantees You'll Always Find the Best Kinder Surprise Toys Apr 28 13:39:18 ping NIN101 Apr 28 13:40:50 RST38h: awesome Apr 28 13:41:45 RST38h: reminds me on the x-ray machines they had in shoe shops here, back in the 50s and into the 60s, though in the 60s they already were forbidden Apr 28 13:44:44 I found one of those when I was dunno maybe 4 years old, and of course immediately tested how it works - my mum got *really* upset Apr 28 13:47:31 x-ray machines in shoeshops? wtf? Apr 28 13:47:59 DocScrutinizer, that explains everything Apr 28 13:48:04 lol j/k :p Apr 28 13:48:28 pedoskop : http://www.physikmuseum.at/index.php?aid=18 // http://www.spiegel.de/netzwelt/tech/0,1518,268071,00.html Apr 28 13:48:53 * DocScrutinizer aims a MTHEL at Termana general direction Apr 28 13:50:01 DocScrutinizer, but what's the point? Apr 28 13:52:25 I have a problem adding a target to scratchbox (to use gcc 4.6.2 from freemangordon). When trying to use fakeroot, it complains about a missing library, running apt-get upgrade gives me dependency errors, and running apt-get -f install gives me this on libc6: package architecture (i386) does not match system () Apr 28 13:52:46 Obviously, the architecture is wrong (should be armel). Anyone know how to fix that? Apr 28 13:55:18 sb-menu info on target: http://pastebin.com/sUESLT0C Apr 28 13:55:32 That shows arm as architecture :( Apr 28 13:59:41 Found the issue, SBOX_DPKG_INST_ARCH was not set. Apr 28 14:02:04 ErwinJunge: point of what? Apr 28 14:06:41 point of having an x-ray device in a shoestore Apr 28 14:14:21 well,isn't that obvious? x-ray baby feet with shoes on Apr 28 14:15:13 see if they fit Apr 28 14:15:26 fit where? Apr 28 14:15:41 *sigh* Apr 28 14:15:48 yep Apr 28 14:16:03 yep wut? Apr 28 14:16:23 go xray your kinder choclate eggs Apr 28 14:16:26 the feet already fit in those shoes, somehow Apr 28 14:16:46 Doc <-- obviously being playful Apr 28 14:16:46 Interesting definition of obvious there ;) I guess the harmfullness of x-rays wasn't know back then? Apr 28 14:17:26 ErwinJunge: I guess you noticed the URLs I posted above? Apr 28 14:18:28 Doc: Lumias have been jailbroken. Sadly, not the 800s with NOKIA bootloader though Apr 28 14:18:29 Yeah, but german isn't in my library of languages :) Apr 28 14:18:53 Or in this case Austrian I guess (does that matter?) Apr 28 14:21:48 ErwinJunge: especially for you: http://en.wikipedia.org/wiki/Shoe-fitting_fluoroscope Apr 28 14:25:05 Pedoscopes --> seriously? Sounds like something law enforcement would use to find sexual predators :) Apr 28 14:26:31 * DocScrutinizer hands ErwinJunge a glass of radon water Apr 28 14:26:38 Never knew these things existed. Thanks for the laughs :) Apr 28 14:29:45 Got to go, byebye all Apr 28 14:29:51 o/ Apr 28 14:37:05 damn...this funny small taiwanese phone I tested the solar thingy with is pretty remarkable in terms of power consumption...I charged it in direct sunlight for roughly two hours until it got a bit cloudy, and turning it on it showed 1 bar of battery...now it has a maintained a call for 36 minutes ._o Apr 28 14:41:11 this one :) http://mobile.softpedia.com/phones/i-mobile/i-mobile-310.shtml Apr 28 14:41:49 DocScrutinizer, can I vote for one extras-testing package version too times? Apr 28 14:42:10 *two or more times Apr 28 14:42:50 hm? Apr 28 14:42:54 or I found bug in maemo.org packages interface :D Apr 28 14:42:57 nope Apr 28 14:43:03 see http://maemo.org/packages/package_instance/view/fremantle_extras-testing_free_armel/kernel-power/2.6.28-10power50/ Apr 28 14:43:06 I CAN Apr 28 14:43:09 o.O Apr 28 14:43:24 you're special :-P Apr 28 14:43:46 I aways thought only mohammad was special Apr 28 14:44:07 til it turned out he's my father Apr 28 14:44:13 ~mohammadag Apr 28 14:44:14 methinks mohammadag is your father, or **special** Apr 28 14:44:28 :) Apr 28 14:45:16 for the geeks: http://www.fingers-welt.de/gallerie/eigen/elektro/roentgen/roentgen.htm Apr 28 14:51:53 what's the typical power consumption of a very low-end phone such as the i-mobile 310 I pasted above, or something similar during a call? Apr 28 14:52:42 (assuming decent environment etc) Apr 28 14:54:09 I haven't really benchmarked actual call times myself I think Apr 28 14:57:23 I guess that's sort of a dumb question though...this device just seems to require a lot less than advertised, but I dunno Apr 28 15:00:14 joga: always something between 50mW and 1W, depending on TX Powerlevel the BTS requests for the phone's transmitter Apr 28 15:01:09 DocScrutinizer I'm just very impressed it's been maintaining the call for an hour now, I was expecting it to perform maybe 10 minutes with the ghetto setup I had and would have been content with that too :) Apr 28 15:01:35 that's the reason why you never can get the call times the manufacturer claims: you never get the lowest TX PWRLVL in eal life Apr 28 15:02:15 but in this case I'd expect to get more than they claim, I should try it sometime.. Apr 28 15:02:59 then an even more funny and more commonly ignored detail is: while you're just "listening" your phone has no useful data to transmit, so will stop transmitting all together for 95%..99% of the time Apr 28 15:03:29 DocScrutinizer that's why I've been playing youtube videos and holding the other phone near my keyboard...both are in the same room now Apr 28 15:04:08 to sort of emulate an awkward conversation maybe :) Apr 28 15:04:25 good approach Apr 28 15:06:22 but this is pretty nice since I'll be going to some festival abroad for a few days next week and we'll be camping out, so at least I'm pretty confident I can use a phone if there's at least some sun :) Apr 28 15:07:51 freemangordon, see http://lists.denx.de/pipermail/u-boot/2012-April/123333.html Apr 28 15:10:00 ok...now it finally hung up, lasted for 1 hour 7 minutes Apr 28 15:10:27 the other phone was probably transmitting more though, but that's pretty good anyway Apr 28 15:15:14 btw my figure above been incorrect - poorly designed phones may burn up to 4..5W in bursts Apr 28 15:15:32 2W max TX power, *<50% efficiency Apr 28 15:31:35 DocScrutinizer since you probably know an answer (I'm less than an amateur with electronics)... right now I just have 3 panels in parallel and a voltage regulator before the usb port, but would it be possible to use a plain nokia battery (since I probably have a couple lying around) for the contraption instead of directly feeding it to the phone's battery? I'm not sure if they're the right kind of battery to use for such Apr 28 15:32:32 I was also thinking of a simple 6xaa case but can't find any junk with a big enough battery holder and should I make one myself it would probably suck anyway :) Apr 28 15:32:45 err, what? charge USB device from LiIon? doesn't work Apr 28 15:33:05 ok, that's what I basically wanted to hear Apr 28 15:33:07 :) Apr 28 15:33:15 I have some nimh but they're too tiny Apr 28 15:33:33 you need sth like 4V minimum to make it properly charge Apr 28 15:33:47 can you put those in series, if I have two Apr 28 15:34:01 LiIon has a voltage of nominal 3.7V which is too low Apr 28 15:34:27 USB doesn't "survivie" voltage >6V Apr 28 15:34:44 so 2 in series also is no option Apr 28 15:34:47 but my panels give about 7V currently Apr 28 15:34:58 (and it's 5.00V after the regulator) Apr 28 15:35:21 well a regulator can cope with whatever it's designed for Apr 28 15:35:39 I mean basically to charge the batteries first with solar and then feed that via regulator to phone or whatever Apr 28 15:35:43 for sure a 5V regulator can take 8V input Apr 28 15:36:11 it's just that the only decent battery I have around are some old nokia batteries I think Apr 28 15:36:18 you mustn't charge LiIon brute force! Apr 28 15:36:36 ok, I'll scratch the idea of using those for this :) Apr 28 15:36:56 you can do this with NiMH Apr 28 15:37:14 NOT with LiIon Apr 28 15:37:15 wish I had those as abundantly Apr 28 15:37:29 I was supposed to get a couple for this specific purpose but I've forgot Apr 28 15:38:14 4 AA cells in series are *somewhat* safe to use for USB powering Apr 28 15:39:15 voltage range should be 4.4V...6.4V, something USB *usually* can cope with Apr 28 15:39:53 how low can the voltage be that an n900 would charge? Apr 28 15:39:57 don't blame me when your USB emits blue magic smoke though Apr 28 15:40:19 as mentioned above, it needs ~4V Apr 28 15:40:34 ah, ok Apr 28 15:40:39 beter 4.5V Apr 28 15:41:12 but for all my experience it survives 7V easily Apr 28 15:41:33 so with 4 primary cells you're quite likely just fine Apr 28 15:42:31 at some point I should try and do this properly, ie. weave the panels into a backpack or whatever and put some adapters, battery and make it less embarrassing Apr 28 15:43:07 as 4* 1.5V(nominal) = 6V; 4* 1.6V(max)=6,4V; 4* 1.2V(real voltage)=4.8V Apr 28 15:43:28 DocScrutinizer thanks for all the info Apr 28 15:43:33 YW Apr 28 15:44:13 I'm pretty new to the electronics stuff but junk projects like these are fun to try, I also learn more albeit slowly :) Apr 28 15:46:05 I just probed a fresh alkaline cell: 1.67V Apr 28 16:00:24 freemangordon, that patch really fixing damaged framebuffer output Apr 28 16:00:47 I will rebase patches, create new build and sent patch series to u-boot mailinglist Apr 28 16:56:33 its there a way to rename my n900 user whitout any risc? Apr 28 17:11:48 my intuition says: no. Apr 28 17:12:23 My intuition sings the LumberJack song. Apr 28 17:15:15 DocScrutinizer, Pali, we have hald-addon-bme i386 compiled in SB and my IDA has hexrays :P Apr 28 17:15:30 I am almost ready with REing Apr 28 17:15:51 freemangordon, disassembler is not needed Apr 28 17:16:08 you do not need to know how hald-addon-bme is communicating with BME Apr 28 17:16:21 you only need to know hal interface for bme Apr 28 17:16:22 it is, the one you in gitorious does not initialize in the same way Apr 28 17:16:24 I know Apr 28 17:16:34 and hal interface is visible via lshal :-) Apr 28 17:16:40 but it handles dbus calls too Apr 28 17:16:52 com.nokia.bme.request Apr 28 17:17:05 status_info_req and timeleft_info_req Apr 28 17:17:07 better look at libbmeipc Apr 28 17:17:35 more closed apps are linked against libbmeipc Apr 28 17:17:47 well, let mi finish that one :P Apr 28 17:17:51 *me Apr 28 17:17:57 ok :-) Apr 28 17:18:06 anyway Apr 28 17:18:11 * freemangordon is afk Apr 28 17:19:04 * RST38h moos Apr 28 17:45:54 freemangordon, I published new build of u-boot with that cfb flush patch Apr 28 17:46:26 also now I send patch series (without bootmenu) to u-boot mailinglist Apr 28 17:46:56 my bootmenu was rejected, becuase it must be rewrite to new menu code (common/menu.c) Apr 28 18:42:02 Hey, anybody updated advanced-power-monitor today? It doesn't work anymore for me. Apr 28 18:42:47 It kind of runs, but the status bard power applet shows a gray bar instead of charge level Apr 28 18:51:19 When running from command line it prints "[F] BQ27200 Apr 28 18:51:19 [F] BQ27200 Alternative". [F] means failed? Apr 28 18:57:49 "lshal | grep bq27200" gives nothing... what could be the cause? Apr 28 19:14:08 LOL Apr 28 19:15:45 I CBA to check ftw this "monitor" does, but this msg sounds a lot like it's using the broken-badly bq27200.ko kernel module which got deprecated and hopefully completely removed from PowerKernel Apr 28 19:16:57 yep, it looks like it should use Nokia's BME Apr 28 19:17:13 but it also says "[L] Nokia's BME ... [U] Nokia's BME" Apr 28 19:17:22 as if it would load, then unload Apr 28 19:18:52 sure Apr 28 19:19:51 so which one is supposed to work? BME or bq27200? And yes, I have the power kernel Apr 28 19:20:01 but power manager worked until today Apr 28 19:20:54 hurray, just removed advanced-power-manager and all works now fine Apr 28 19:21:08 seems I have installed this thing some time just 4 lulz... Apr 28 19:21:32