**** BEGIN LOGGING AT Fri Jan 30 03:00:00 2015
Jan 30 05:37:52 <Sicelo_> i will install cssu-t sometime soon
Jan 30 05:37:56 <Sicelo_> sorry
Jan 30 05:41:51 <freemangordon> Sicelo_: why sorry?
Jan 30 07:25:31 <zGrr> moin :)
Jan 30 09:46:56 <bencoh> yay, just received a new polarcell
Jan 30 09:47:20 <bencoh> now ... do I need to do anything to "calibrate" it ?
Jan 30 09:49:21 <kerio> you can calibrate the bq27200 chip in the n900 if you want
Jan 30 09:49:31 <kerio> but it's not something that's used by the standard BME
Jan 30 09:49:37 <kerio> so it won't be reflected by the battery applet, for instance
Jan 30 09:57:14 <bencoh> hmm
Jan 30 09:57:41 <bencoh> does it mean that bme has some hardcoded constants for nokia bl-5j ?
Jan 30 10:00:14 <kerio> nobody knows what bme does
Jan 30 10:00:17 <kerio> except that it's shit
Jan 30 10:00:33 <bencoh> :]
Jan 30 10:00:42 <Vajb> i thought that bme just uses that resistor value to estimate capacity
Jan 30 10:00:54 <bencoh> hmm
Jan 30 10:01:22 <Vajb> tho bme replacemnt seams to lean on same value...
Jan 30 10:01:31 <Vajb> bleh
Jan 30 10:01:33 <Vajb> seems
Jan 30 10:01:45 <kerio> you don't understand how bme replacement works then
Jan 30 10:01:56 <bencoh> how does it work ?
Jan 30 10:02:05 <Vajb> that is so true :)
Jan 30 10:02:18 <kerio> if bq27k is calibrated, it'll use all the values from it
Jan 30 10:02:28 <Vajb> but i can clearly see what my battery applet shows and what not
Jan 30 10:02:51 <kerio> otherwise it'll use the resistor value
Jan 30 10:03:07 <kerio> if rx51_battery is loaded
Jan 30 10:03:08 <Vajb> well it is calibrated but value is not even close to last measured discharge value
Jan 30 10:04:13 <kerio> then it's not calibrated
Jan 30 10:04:14 <Vajb> 1487 last measured and applet shows 1277
Jan 30 10:04:20 <kerio> CI?
Jan 30 10:04:23 <kerio> it's one of the flags
Jan 30 10:04:41 <Vajb> ci?
Jan 30 10:04:46 <kerio> it's one of the flags
Jan 30 10:04:50 <bencoh> looks like rx51_battery is loaded but unused according to lsmod
Jan 30 10:04:59 <bencoh> I'm using the original bme though
Jan 30 10:05:04 <bencoh> (cssu-s)
Jan 30 10:05:28 <Vajb> ci:1
Jan 30 10:05:32 <kerio> Vajb: so there you go
Jan 30 10:05:33 <bencoh> or does it just periodically check a /sys entry ?
Jan 30 10:05:34 <kerio> it's not calibrated
Jan 30 10:05:46 <Vajb> hmm how come?
Jan 30 10:05:54 <kerio> because the calibration was marked invalid due to aging
Jan 30 10:05:59 <Vajb> i ran bq five times
Jan 30 10:06:14 <Vajb> and this is quite new battery
Jan 30 10:06:19 <Vajb> two months maybe
Jan 30 10:06:31 <kerio> CI is set after 30 charge cycles
Jan 30 10:06:41 <Vajb> but ok ci is the flag to follow
Jan 30 10:06:41 <Vajb> guess it should be 0 then
Jan 30 10:06:52 <kerio> it's not that "it should be 0"
Jan 30 10:07:01 <kerio> it's that if CI is set, you can't trust the bq27k values
Jan 30 10:07:04 <kerio> that's what the flag means
Jan 30 10:07:12 <kerio> it's also set in case of extreme conditions, iirc
Jan 30 10:14:25 <bencoh> what I understand from this is that nokia fucked up with an onboard charge controler :p
Jan 30 10:14:44 <kerio> eeh
Jan 30 10:15:09 <kerio> bq27k is intended to be attached to the battery itself, yes
Jan 30 10:15:17 <bencoh> I hope that batteries are cheaper at least
Jan 30 10:15:31 <kerio> i don't think the cost of a chip is relevant
Jan 30 10:15:31 <bencoh> I mean, effectively cheaper, for customers (us)
Jan 30 10:15:48 <bencoh> then it was a stupid idea :)
Jan 30 10:15:59 <kerio> the cost of raw nand compared to the cost of a uSD is pretty much the same
Jan 30 10:16:09 <kerio> even though a uSD has some extra component
Jan 30 10:16:14 <kerio> *components
Jan 30 10:17:42 <bencoh> kerio: bottomline ... should I do anything regarding the new polarcell, apart from just swapping it with the old one ? :)
Jan 30 10:17:58 <kerio> recalibrate bq27k, if you want
Jan 30 10:18:12 <kerio> but it's hard to calibrate it if you're running bme_rx51
Jan 30 10:18:30 <bencoh> does bme remplacement run on cssu-s ?
Jan 30 10:19:48 <Maxdamantus> Why is it hard to calibrate if you're running bme_rx51? Because it shuts down the system?
Jan 30 10:25:57 <Vajb> kerio: so i need to run bq272k 30 times to calibrate or just load and discharge my battery normally 30 times? It shows now that i had 35 cycles since last calibration
Jan 30 10:26:05 <kerio> wut
Jan 30 10:26:14 <kerio> no, you just do the calibration once
Jan 30 10:26:19 <kerio> it'll last for 30 cycles
Jan 30 10:26:36 <Vajb> but it never showed the right value so i guess it was never succesfull
Jan 30 10:26:52 <kerio> huh
Jan 30 10:26:54 <kerio> that's weird
Jan 30 10:27:00 <kerio> are you sure you have the bme replacement installed?
Jan 30 10:27:04 <Maxdamantus> cat /sys/class/power_supply/bq27200-0/capacity
Jan 30 10:27:04 <kerio> and not some weird battery applet
Jan 30 10:27:17 <bencoh> 11:26 < kerio> no, you just do the calibration once
Jan 30 10:27:17 <bencoh> 11:26 < kerio> it'll last for 30 cycles
Jan 30 10:27:32 <Vajb> i mean it changed value everytime, but it was always 200mah less
Jan 30 10:27:45 <kerio> ah yes, that's another issue
Jan 30 10:28:09 <kerio> bq27k won't adjust too much
Jan 30 10:28:25 <kerio> so until you get two calibrations in a row with roughly the same value, you're not getting a correct value
Jan 30 10:28:34 <bencoh> bleh
Jan 30 10:28:47 <bencoh> it feels like black magic
Jan 30 10:28:52 <kerio> it's to avoid outliners
Jan 30 10:28:56 <kerio> because of weird issues
Jan 30 10:29:08 <kerio> but, again, it shows that bq27k was supposed to be attached to a constant power source
Jan 30 10:29:13 <kerio> because you lose the calibration when you switch battery
Jan 30 10:29:15 <bencoh> exactly
Jan 30 10:29:34 <Maxdamantus> Would it still actually be effective even if part of the battery?
Jan 30 10:29:45 <kerio> of course
Jan 30 10:29:53 <bencoh> it would be calibrated once and for all
Jan 30 10:30:05 <Maxdamantus> I don't think I ever let my N900 fully charge then fully discharge, until a few days ago.
Jan 30 10:30:07 <kerio> you just have to be mindful that every once in a while, you have to let it discharge completely
Jan 30 10:30:15 <Maxdamantus> for the almost-year I've had it.
Jan 30 10:30:28 <Maxdamantus> Yeah, so it's something users have to consciously do.
Jan 30 10:31:02 <bencoh> Maxdamantus: same here, but my old polarcell has never shown anything beyond 1230mAh
Jan 30 10:31:22 <bencoh> (I dunno how old it is)
Jan 30 10:31:24 <kerio> not all polarcells are equal i guess
Jan 30 10:31:31 <bencoh> prolly
Jan 30 10:31:37 <kerio> i bought a random-ass battery with no discernible brand
Jan 30 10:31:37 <Maxdamantus> and what does it mean by the 30 charge cycles?
Jan 30 10:31:41 <kerio> and i got a ~1500mAh thing
Jan 30 10:31:43 <bencoh> and I guess they age too
Jan 30 10:32:05 <Maxdamantus> what constitutes a charge cycle, if not letting it fully charge then fully discharge?
Jan 30 10:32:13 <Maxdamantus> (which should calibrate it)
Jan 30 10:32:24 <kerio> charging and discharging a total of LMD
Jan 30 10:32:46 <kerio> so if you discharge from 100% to 50% and then charge it back to 100% twice, it's a cycle
Jan 30 10:32:59 <Maxdamantus> Ah.
Jan 30 10:33:03 <kerio> or roughly a cycle i guess
Jan 30 10:33:12 <kerio> bq27k does some automatic lowering of LMD as time goes
Jan 30 10:33:35 <kerio> so your expectations are never too high
Jan 30 10:33:51 <kerio> it's a pretty nice chip :3
Jan 30 10:34:32 <bencoh> kerio: buying a random noname battery feels like placing a bet
Jan 30 10:34:33 <bencoh> "will it blend ?"
Jan 30 10:34:45 <Maxdamantus> I'm still sceptical about the fact that users would have to consciously calibrate it.
Jan 30 10:34:49 <kerio> it was suggested by vi____, iirc
Jan 30 10:35:10 <kerio> Maxdamantus: if you want really accurate measurements, it's even worse
Jan 30 10:35:14 <bencoh> vi____ ?
Jan 30 10:35:22 <kerio> because the faster you discharge it, the less charge it's going to show
Jan 30 10:35:30 <kerio> because it's not an ideal capacitor
Jan 30 10:35:41 <kerio> there's lots of chemical bullshit going on in a lithium battery
Jan 30 10:35:50 <bencoh> it's no bullshit :p
Jan 30 10:36:01 <Maxdamantus> I think I found the opposite.
Jan 30 10:36:11 <bencoh> (no, I'm no chemist)
Jan 30 10:36:38 <Maxdamantus> When I first tried to calibrate it a few days ago, I discharged it in about two hours, and it recorded the LMD as 1486 mAh.
Jan 30 10:38:01 <Maxdamantus> It's now down to 1177 mAh after another similarly fast one (~2h30) and one slow one (~20h)
Jan 30 10:39:17 <kerio> yeah, because your battery is roughly 1200 but the default was 1500
Jan 30 10:39:23 <kerio> so you need two or three calibrations to get an accurate value
Jan 30 10:39:42 <kerio> regardless of how fast you're discharging the battery :)
Jan 30 10:39:55 <Maxdamantus> The default where?
Jan 30 10:40:05 <Maxdamantus> provided by the chip manufacturer?
Jan 30 10:40:19 <Maxdamantus> or did the previous user use a higher-capacity battery?
Jan 30 10:40:19 <kerio> the default in bq27k, yes
Jan 30 10:40:23 <Maxdamantus> Mm.
Jan 30 10:40:30 <kerio> no no, the data is reset when you pull the battery
Jan 30 10:40:38 <kerio> istr it's not attached to the backup battery
Jan 30 10:40:47 <kerio> which is probably not working anyway
Jan 30 10:41:06 <Maxdamantus> Indeed.
Jan 30 10:42:00 <Maxdamantus> Hm. I though I'd already pulled the battery out (long enough for it to forget the time) since the first calibration.
Jan 30 10:43:58 <Maxdamantus> maybe I haven't.
Jan 30 10:44:43 <kerio> ooh actually i'm wrong
Jan 30 10:44:48 <kerio> the default is ILMD
Jan 30 10:44:52 <kerio> Initial Last Measured Discharge (ILMD): 2056.32 mAh
Jan 30 10:45:46 <Maxdamantus> on the N900?
Jan 30 10:45:56 <Maxdamantus> It looks like the ILMD is meant to be programmed in.
Jan 30 10:46:01 <kerio> yep
Jan 30 10:46:03 <kerio> but it wasn't programmed
Jan 30 10:46:17 <Maxdamantus> Heh, because they weren't expecting to use it.
Jan 30 10:46:18 <kerio> it's not particularly meaningful, the way it was implemented on the n900
Jan 30 10:46:34 <kerio> you program a batch of chips with the approximate charge your batteries are going to hold
Jan 30 10:53:52 <bencoh> 11:44 < kerio> Initial Last Measured Discharge (ILMD): 2056.32 mAh
Jan 30 10:53:58 <bencoh> wait, what ? 2056mah ?
Jan 30 10:54:10 <bencoh> where does that come from ?
Jan 30 10:55:44 * Maxdamantus wonders if it'd be better if the software let people specify what they think the battery's capacity is.
Jan 30 10:56:08 <bencoh> prolly not
Jan 30 10:56:46 <Maxdamantus> and just use that and `LMD - NAC`
Jan 30 10:57:02 <kerio> it's not linear
Jan 30 10:57:21 <kerio> it's used as a starting point for calculations
Jan 30 11:01:44 <Pali> initial value comes from EEPROM chip in n900
Jan 30 11:02:09 <Pali> programmed by Nokia
Jan 30 11:06:34 <kerio> why wasn't it set in a more sensible way, tho
Jan 30 11:45:42 <Pali> kerio: because nokia bme does not use ILMD, so it was probably programmed to default max value
Jan 30 11:46:19 <kerio> yeah but
Jan 30 11:46:22 <kerio> they know their BL-5J
Jan 30 11:46:56 <Pali> I think they did not set ILMD value and by default some factory settings is 0xFFFF or so
Jan 30 11:47:02 <kerio> oic
Jan 30 12:18:39 <Vajb> i guess bme replacement is now sorted. So maybe someone can explain how is it possible that my n900 still reboots daily with sw_rst even though i have r&d mode enabled with 3 wd flags
Jan 30 12:19:37 <Vajb> syslog doesn't show anything unusuall. Is there other ways to debug?
Jan 30 12:35:45 <Pali> lifeguard reset?
Jan 30 12:36:01 <Pali> hw reset?
Jan 30 12:48:38 <Vajb> enabled
Jan 30 12:48:48 <Vajb> also omap is enabled
Jan 30 13:56:10 <Pali> then some hw reset.. low battery? other hw problem?
Jan 30 14:13:14 <Vajb> battery is fresh
Jan 30 14:13:38 <Vajb> dunno about hw problems. Everything seems to work
Jan 30 14:14:05 <Vajb> also worth mentioning is that it has no sim just wlan connection always on
Jan 30 14:15:13 <Vajb> i thought first that maybe battery connection is the cause, but tightening them brought no difference
Jan 30 14:15:51 <Vajb> how to pinpoint hw problem?
Jan 30 15:10:42 <Pali> provide syslog
Jan 30 15:10:47 <Pali> maybe there is something
Jan 30 15:17:20 <Vajb> r there any specific lines to search for?
Jan 30 15:18:11 <Vajb> give my device a few days and it will produce syslog to look at.
Jan 30 15:18:25 <Vajb> should i post it to pastebin for example?
Jan 30 15:38:44 <Pali> there must be something which caused reboot
Jan 30 15:39:45 <Pali> either regular reboot (and rsyslogd daemon write to systlog that was stopped) or HW reboot (no indication about regular reboot)
Jan 30 16:01:22 <freemangordon> Pali: EEPROM?
Jan 30 16:01:45 <freemangordon> if it is EEPROM we can reprogram it
Jan 30 17:25:33 <Pali> freemangordon: cannot, it is connected to bq chip
Jan 30 17:25:50 <Pali> and reprogramming needs some voltage on hw level...
Jan 30 17:26:13 <Pali> ask shadowjk, he could know more...
Jan 30 18:15:22 <ShadowJK> iirc 40V on pins not exposed
Jan 30 20:29:17 <stryngs> Is it insane to upgrade dpkg itself?
Jan 30 20:29:22 <stryngs> Would it break my n900?
Jan 30 20:45:47 <stryngs> Nobody? =-)
Jan 30 20:46:02 <stryngs> Common folks, I'm trying to build a native package builder for the n900
Jan 30 20:46:12 <stryngs> This will be of great use to those who wish to build .debs directly on their n900
Jan 30 20:46:21 <stryngs> Just need a wee bit o help
Jan 30 20:53:17 <Sicelo_> dpkg did get updated in some CSSU version iirc.. anyway, since you do things your own way, i'd say take a backup, try what you want, and see...
Jan 30 20:53:44 <stryngs> true =)
Jan 30 20:53:53 <stryngs> It'll be neat though Sicelo, a native dh_make and such
Jan 30 20:54:02 <stryngs> I know people harp on scratchbox and all
Jan 30 20:54:18 <stryngs> i was missing dpkg-buildflags, so i needed to update =)
Jan 30 21:08:01 <merlin1991> stryngs: there should be a newer debhelper version in extras
Jan 30 21:08:30 <merlin1991> maemo.org/packages/view/debhelper7/
Jan 30 21:08:43 <merlin1991> dunno if that is enough to include what you want
Jan 30 21:22:23 <stryngs> looking for 8 =/
Jan 30 21:22:31 <stryngs> i'm playing the dependency hell game
Jan 30 21:22:34 <stryngs> but im almost htere!
Jan 30 21:22:42 <stryngs> and when done I'm gunna repo it up for the community !
Jan 30 21:22:46 <stryngs> Pretty exciting =)
Jan 30 21:23:08 <stryngs> Would upgrading perl be bad for the n900
Jan 30 21:30:41 <Maxdamantus> Mm, so now LMD is 1175 mAh, down from 1177. I guess that means it's found about the right value.
Jan 30 21:41:53 <stryngs> zomg, my head is gunna pop
Jan 30 21:41:59 <stryngs> i just want to say, I HATE scratchbox
Jan 30 21:43:21 <merlin1991> nah scratchbox is fine
Jan 30 21:43:24 <merlin1991> when you know the quirks ;9
Jan 30 21:43:48 <merlin1991> btw please make your new debhelper package in a way that it can coexist with the original one, because if not we are in build hell ...
**** ENDING LOGGING AT Sat Jan 31 02:59:58 2015