**** BEGIN LOGGING AT Mon Apr 09 03:00:03 2018
Apr 09 06:17:14 <DocScrutinizer05> http://wstaw.org/m/2018/04/09/plasma-desktopG22448.png
Apr 09 06:18:35 <DocScrutinizer05> http://wstaw.org/m/2018/04/09/plasma-desktopd22448.png
Apr 09 11:15:48 <antrik> I wonder, how much control does the OS have over battery charging in a smartphone? would it be possible for a custom OS to limit the charging voltage, or is that hard-coded in the battery controller?
Apr 09 11:17:53 <PaulFertser> antrik: there's even an android app that does that on many popular phones.
Apr 09 11:18:11 <PaulFertser> So apparently at least some drivers allow that.
Apr 09 11:18:32 <PaulFertser> I wouldn't expect a smart battery controller to hardcode maximum voltage.
Apr 09 14:03:52 <PaulFertser> antrik: https://github.com/sriharshaarangi/BatteryChargeLimit
Apr 09 15:36:26 <antrik> PaulFertser: that only mentions charge level it seems?... not really the same
Apr 09 15:37:20 <antrik> better than nothing for sure -- but I considered it kinda given that the charge level can be limited. limiting voltage is the tricky bit
Apr 09 16:42:57 <PaulFertser> antrik: apparently you can poll the voltage level and stop at any time.
Apr 09 17:52:15 <bencoh> PaulFertser: charging at high voltage X and stopping charge when reaching Y isn't really the same as charging at Y, though
Apr 09 17:54:06 <bencoh> but assuming driver doesn't exposes charge voltage (it probably doesn't) you can always poke with i2cget/set in may cases
Apr 09 17:54:30 <bencoh> -s
Apr 09 17:57:04 <PaulFertser> bencoh: you can't "charge at high voltage", Li charging ICs always charge by CC until reaching the target voltage and then they should just stop charging.
Apr 09 17:57:54 <PaulFertser> bencoh: batteries usually have a chip inside and it measures current and voltage and that data is usually exposed via /sys/power_supply/
Apr 09 17:59:34 <bencoh> PaulFertser: sorry, the "high voltage" thing was misleading. but say, 4.3 instead of 4.2 (for instance)
Apr 09 17:59:46 <bencoh> (or more, even)
Apr 09 18:00:29 <bencoh> and no, you don't really stop "once target voltage is reached"
Apr 09 18:01:11 <bencoh> usually you monitor current as well, and eventually drop once it falls under threshold
Apr 09 18:01:42 <PaulFertser> bencoh: no, it can't be 4.3 because it's constant current so the voltage is slowly raising till the threshold. And then it stops.
Apr 09 18:01:47 <bencoh> (it's dependent on charger implem, though. but that's quite common nowadays)
Apr 09 18:02:09 <PaulFertser> bencoh: I digged PMU driver and process for Openmoko a lot. 
Apr 09 18:02:21 <bencoh> PaulFertser: same for other products ;)
Apr 09 18:03:11 <bencoh> bottom line is: it's highly implem dependent I guess :)
Apr 09 18:04:28 <bencoh> (note: I was surprised as well when I saw some actually charged their battery at 4.3 or more ... but it does happen, even with reknowned phone makers)
Apr 09 18:32:08 <PaulFertser> bencoh: I highly doubt anyone would charge standard Li at 4.3V, really doubt.
Apr 09 18:36:22 <bencoh> well ...
Apr 09 18:36:33 <bencoh> that's what I thought then
Apr 09 18:36:47 <bencoh> then I checked with a multimeter
Apr 09 18:40:37 <DocScrutinizer05> well, all that pretty much depends on particular hw platform
Apr 09 18:42:06 <DocScrutinizer05> there are 4.2 and 4.3 LiIon chemistries, there are smart charger chips and simplistic 100% sw-controlled  chargers. And LiIon is CC/CV charged
Apr 09 18:43:05 <DocScrutinizer05> when reaching 4.2V the charger limits voltage instead current, but charging is stopped only after current at 4.2V trips a low threshold
Apr 09 18:43:48 <DocScrutinizer05> very few chargers allow limiting voltage to say 3.9V
Apr 09 18:44:40 <DocScrutinizer05> then OTOH that's not needed, more releavnt for battery lifespan is to avoid deep discharges
Apr 09 18:58:57 <DocScrutinizer05> http://neo900.org/stuff/block-diagrams/neo900/neo900.html  -> http://www.ti.com/lit/ds/symlink/bq24297.pdf   http://wstaw.org/m/2018/04/09/plasma-desktopq22448.png
Apr 09 23:20:24 <antrik> bencoh: my current phone charges up to 4.5 V (!?!) according to the official applet
Apr 09 23:22:39 <antrik> DocScrutinizer05: unfortunately, it's not that simple. research shows that LiIon batteries actually suffer much more from approaching 100% charge state than approaching the low cutoff (somewhere around 20%)
Apr 09 23:24:16 <DocScrutinizer05> umm, my last 10 years RL experience suggest otherwise
Apr 09 23:24:59 <antrik> PaulFertser: modern cells are more robust, so they can survive a reasonable number of cycles even when charging at a voltage that would kill older cells in a few dozen cycles... but of course that doesn't change the fact that they would survive longer with a lower voltage
Apr 09 23:25:18 <DocScrutinizer05> my N900 floats at 4.2V all the time, since several years, and the batteries seem to cope with that quite fine
Apr 09 23:27:34 <antrik> DocScrutinizer05: well, it actually depends on the exact conditions. when reaching a very low charge level, you have to reduce current both during discharge and during charge aftewards to avoid stress... consumer electronics probably don't bother with that (unlike more demanding applications, such as electric vehicles...)
Apr 09 23:29:58 <DocScrutinizer05> my experience is: keep floating at 4.2V - years without significant impact to capacity; discharge to 0% (3.2V) - noticeable impact even for a single cycle; discharge to 0% at store like that for 2 weeks - dead cell
Apr 09 23:30:51 <DocScrutinizer05> slightly generalized
Apr 09 23:30:52 <antrik> DocScrutinizer05: well, if you really keep it floating, without ever discharging, the level is indeed less significant
Apr 09 23:31:25 <antrik> (older cells used to age significantly even when floating, especially at higher temperatures; but newer / higher quality cells seem to be less affected)
Apr 09 23:31:29 <DocScrutinizer05> ooh, there you have a point indeed
Apr 09 23:32:04 <DocScrutinizer05> the 100%-90%-100% cycle may be a lot of stress for the cell
Apr 09 23:32:11 <DocScrutinizer05> I can't tell
Apr 09 23:33:11 <antrik> yeah, the research I have seen suggests 100%-90%-100% is really rough on LiIon cells
Apr 09 23:33:12 <DocScrutinizer05> anyway with Neo900 chager chip you *can* choose CV 
Apr 09 23:33:56 <DocScrutinizer05> with the usual dirt cheap charging circuitry in many smartphones, you maybe can't
Apr 09 23:34:37 <antrik> Tesla vehicles charge only to a pre-defined level by default for that reason, and only charge fully when you explicitly tell it before a long trip...
Apr 09 23:35:43 <antrik> DocScrutinizer05: oh... I assumed it's all more or less the same
Apr 09 23:41:14 <antrik> PaulFertser: if the FreeRunner really didn't limit the charging voltage before reaching full charge, that would certainly explain the medicore battery life :-P
Apr 09 23:41:23 <DocScrutinizer05> no way, there's a shitload of different approaches, from smart autonomous chargers like the BQ24297 used in Neo900, to APE software GPIO controlled FET gates to linearly modulate the charge current with massive efficiency loss of linear regulator
Apr 09 23:42:19 <antrik> PaulFertser: all description I have ever read clearly state that LiIon is charged with constant current until reaching the maximum voltage, after which the rest is done with constant voltage while the current drops asymptotically
Apr 09 23:42:47 <DocScrutinizer05> N810 is of the latter class, so a hiccup in software may explode your battery, that's why Nokia was ULTRA.OMGNOWAY about opening up the sourcecode of BME
Apr 09 23:43:37 <antrik> it's also true for my current smartphone -- though admittedly that one lets the voltage get scarily high before limiting, and stops charging long before current nears zero... which is certainly good for charging time, but not for battery life
Apr 09 23:43:39 <DocScrutinizer05> CC/CV, yes
Apr 09 23:43:59 <DocScrutinizer05> all LiIon are charged CC/CV
Apr 09 23:44:14 <antrik> (nevertheless, it *does* have a constant voltage phase -- just an unusually short one...)
Apr 09 23:44:24 <DocScrutinizer05> with a current thrshold and a safety timer for charge-end detection
Apr 09 23:59:56 <DocScrutinizer05> ~batteryfaq
Apr 09 23:59:56 <apt> i guess batteryfaq is http://wiki.openmoko.org/wiki/Battery_Questions_and_Answers
Apr 10 00:00:35 <DocScrutinizer05> pabs3: openmoko server is severely fscked up
Apr 10 00:01:45 <DocScrutinizer05> I really don't get it, maybe needs a MySQL full reset
Apr 10 00:02:24 <DocScrutinizer05> http://wiki.openmoko.org/wiki/Battery_Questions_and_Answers still trying to load
Apr 10 00:02:39 <DocScrutinizer05> I'm sure eventually it will time out
Apr 10 00:05:24 <DocScrutinizer05> pabs3: http://wstaw.org/m/2018/04/10/plasma-desktopD22448.png
Apr 10 00:06:12 <DocScrutinizer05> still "contacting wiki.openmoko.org, waiting for reply"
Apr 10 00:06:43 <DocScrutinizer05> pabs3: please kill that planet stuff, it aggregates spam aiui
Apr 10 00:07:29 <DocScrutinizer05> ok, >>Connection to Server Closed Unexpectedly<< timeout
Apr 10 00:08:37 <DocScrutinizer05> many active (using CPU) /usr/bin/python -S /var/lib/mailman/scripts/driver subscribe
Apr 10 02:00:41 <DocScrutinizer05> pabs3: http://wstaw.org/m/2018/04/10/plasma-desktopu22448.png
**** ENDING LOGGING AT Tue Apr 10 03:00:02 2018