| 00:23.11 | *** join/#card10badge aphotic (~aphotic@adsl-156.176.58.206.tellas.gr) |
| 00:27.07 | *** part/#card10badge aphotic (~aphotic@adsl-156.176.58.206.tellas.gr) |
| 00:29.56 | *** join/#card10badge aphotic (~aphotic@adsl-156.176.58.206.tellas.gr) |
| 00:30.17 | *** part/#card10badge aphotic (~aphotic@adsl-156.176.58.206.tellas.gr) |
| 00:55.24 | *** join/#card10badge maralorn (~maralorn@hera.m-0.eu) |
| 01:39.15 | *** join/#card10badge Rev_Illo (~revillo@unaffiliated/rev-illo/x-3122184) |
| 02:38.08 | *** join/#card10badge user_51 (~quassel@92.117.230.35) |
| 02:41.04 | *** join/#card10badge verin0x4 (~thelounge@x4dbf4fb6.dyn.telefonica.de) |
| 03:11.27 | *** join/#card10badge PyroPeter (~pyropeter@april-fools/2013/tenth/pyropeter) |
| 03:54.18 | *** join/#card10badge maralorn (~maralorn@hera.m-0.eu) |
| 04:19.13 | *** join/#card10badge DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 04:20.27 | *** join/#card10badge theAntimon (~matze@p54B09205.dip0.t-ipconnect.de) |
| 05:56.29 | *** join/#card10badge erlehmann (~erlehmann@46.114.35.169) |
| 08:04.35 | *** join/#card10badge Rev_Illo (~revillo@unaffiliated/rev-illo/x-3122184) |
| 09:43.27 | *** join/#card10badge shokinn (~quassel@static.157.8.69.159.clients.your-server.de) |
| 11:33.22 | *** join/#card10badge mmu_man (~revol@vaf26-2-82-244-111-82.fbx.proxad.net) |
| 12:15.58 | *** join/#card10badge mmu_man (~revol@vaf26-2-82-244-111-82.fbx.proxad.net) |
| 12:28.06 | DocScrutinizer05 | supersat: Twix: Stormwind_mobile: I tweaked the heater constants of gas sensor and got way better results in way shorter time from enabling sensor until it stabilizes (20s instead of 15min or longer) |
| 12:28.54 | DocScrutinizer05 | I found the sensor is very poor at sensing lighter gas |
| 12:31.02 | DocScrutinizer05 | at least when you compare that to e.g. cig smoke |
| 12:32.04 | DocScrutinizer05 | also seems air humidity (rh) has a massive impact on the gas sensor raw reading |
| 12:37.17 | DocScrutinizer05 | [Tuesday, 3 September 2019] [14:43:38 CEST] <DocScrutinizer05> BME680 epicardium driver doesn't really support Gas Sensor? My values for Gas Resistance are completely bogus, suspicion is the driver doesn't provide proper sensor element heating values [Tuesday, 3 September 2019] [15:12:26 CEST] <DocScrutinizer05> completely bogus means they increase over 30some minutes from 22000 to 79000 linearly and without any reaction to feeding |
| 12:37.18 | DocScrutinizer05 | concentrated propane to the sensor |
| 12:44.56 | DocScrutinizer05 | [Wednesday, 4 September 2019] [16:41:20 CEST] <DocScrutinizer05> https://git.card10.badge.events.ccc.de/card10/firmware/blob/master/epicardium/modules/bme680.c#L14 s/320/250/ L15 s/150/50/ ;;# Gas resistance value stabilizes around 31566 and a small shot of ethanol near cardio makes that drop to 18076 |
| 12:49.00 | Stormwind_mobile | Thanks for reporting! |
| 12:49.07 | DocScrutinizer05 | yw :-) |
| 12:49.30 | Stormwind_mobile | How much info did you gather so far about the mode of operation of that sensor? |
| 12:49.52 | Stormwind_mobile | Does the datasheet explain it sufficiently? |
| 12:50.24 | DocScrutinizer05 | well, I read the datasheets (not completely though) and did a number of real life tests like taking the card10 in gas sensor mode with me when shopping |
| 12:50.55 | Stormwind_mobile | The response seems to be dramatically dependent on the chemistry of the gas itself. I imagine that ionisation energy, polarity or something like that plays a role. |
| 12:51.06 | DocScrutinizer05 | the datasheet explains the operation of sensor in basics and explains the meaning of registers |
| 12:51.51 | Stormwind_mobile | Register values is one thing, *why* the values are what they are is another. |
| 12:52.47 | DocScrutinizer05 | it's a heated metaloxide(?) surface that changes resistance when reacting with certain gases (prolly mainly oxidable gases) |
| 12:53.05 | DocScrutinizer05 | the basics are in wikipedia |
| 12:54.07 | DocScrutinizer05 | what you get as sensor value is the raw resistance readout |
| 12:56.18 | DocScrutinizer05 | there's a closed blob library from Bosch that has arcane secret knowledge how to convert the operation parameters and resistor (and other, like rH readings) into a "air quality" value |
| 12:57.22 | DocScrutinizer05 | my tests seem to suggest that at least rH is a factor you need to use for compensation to adjust the raw resistance value |
| 13:03.50 | DocScrutinizer05 | what's quite puzzling is the sensor's capability to store 10 sets of operation parameters in an array, as if it would make sense to have the sensor operate at different temperatures and preheat times in quick sequence |
| 13:05.46 | DocScrutinizer05 | anyway that feature is completely unused in epicardium driver, despite an obscure comment >>/* Create a ramp heat waveform in 3 steps */ << that's evidently not related to what the code does |
| 13:06.00 | DocScrutinizer05 | https://git.card10.badge.events.ccc.de/card10/firmware/blob/master/epicardium/modules/bme680.c#L83 |
| 13:11.49 | DocScrutinizer05 | the datasheet at some point says "a preheat time of 30ms should usually suffice to allow sensor to reach its nominal temperature" OWTTE. My patch changes preheat time from 150ms to 50ms, and temperature from 320C to 250C |
| 13:13.00 | DocScrutinizer05 | and that made the difference between bogus readings and meaningful readings for me |
| 13:15.21 | pippin | DocScrutinizer05: the use of such a sensor (or larger similar ones) is apparently also documented in academic literature |
| 13:15.23 | DocScrutinizer05 | the driver API lacks a feature for app tp provide those sensor config parameters |
| 13:15.44 | DocScrutinizer05 | pippin: you got a link to share? |
| 13:16.10 | pippin | no I have third hand information only |
| 13:16.27 | DocScrutinizer05 | ahh thanks anyway, gonna have a google search for it |
| 13:16.57 | pippin | but apparently procedures can involve heating up to even higher temperature - to "clean the lab", before using lower temperatures in curves at intervals - and then comparing the resulting resistance plots to references |
| 13:17.19 | DocScrutinizer05 | I thought *exactly* that a few minutes ago :-D |
| 13:17.29 | DocScrutinizer05 | pyroclean |
| 13:18.10 | DocScrutinizer05 | the sensor has unbearable recovery times it seems |
| 13:18.31 | DocScrutinizer05 | a pyroclean cycle might fix that |
| 13:19.20 | DocScrutinizer05 | might also explain that parameters[10] |
| 13:21.08 | DocScrutinizer05 | pippin: many thanks, such input makes my day :-) |
| 13:23.49 | DocScrutinizer05 | I also wonder how much those heating+cooling cycles and the resulting air volume change helps with air exchange between ambient and sensor chanber through that really tiny hole in case |
| 13:23.58 | Stormwind_mobile | If i have an hour to spare, I can dig into the literature a bit. I have access to a lot of journals at work, and by coincidence, my working group does research on ceramics/metal oxide materials. |
| 13:24.21 | DocScrutinizer05 | :-D <3 |
| 13:24.22 | Stormwind_mobile | It's not the exact field, but reasonably close. |
| 13:25.08 | Stormwind_mobile | Would help if there was a bit of insight into the arcane magic that the Bosch firmware does, though. |
| 13:25.55 | Stormwind_mobile | I have the business card somewhere, of the Bosch employee that was at camp. Maybe there's a way to get some more info than the public datasheet. |
| 13:28.22 | DocScrutinizer05 | somebody owning a IDApro might want to disassemble the Bosch lib too ;-) |
| 13:29.32 | DocScrutinizer05 | can't be whitchcraft they do in there |
| 13:34.48 | DocScrutinizer05 | ok, my card10's battery is "dead" now, after weeks in suspend :-D |
| 14:26.39 | pippin | video of my in-progress anti-aliased vector canvas for microcontrollers : http://pippin.gimp.org/ctx/ctx2.mp4 card10 vs esp32 |
| 14:29.38 | pippin | a single core of <100mhz struggles to keep up with the 240mhz competiton :], but the performance is starting to approach acceptable for many uses :D |
| 14:40.31 | *** join/#card10badge davel (~davel@82-71-4-30.dsl.in-addr.zen.co.uk) |
| 15:00.49 | Stormwind_mobile | @DocScrutinizer05: sounds not too shabby. I wonder if anyone has done thorough investigation of the power consumption profile. I guess @schneider will have done some. |
| 15:03.29 | Stormwind_mobile | Is there an integrated battery gauge in the PMIC? Might be interesting to be able to estimate the amount of power that goes into the LiPo cell, and even more how much current is drawn. But this requires suitable circuitry. |
| 15:04.51 | Stormwind_mobile | Rahix: does epicardium count how much time the processor cores maybe also peripherals?) spends in what kind of sleep state? |
| 15:13.30 | Rahix | The PMIC does not have a battery gauge, unfortunately |
| 15:13.45 | Rahix | Right now, epicardium doesn't ever really sleep ... |
| 15:21.52 | Stormwind_mobile | The other Côte? |
| 15:21.57 | Stormwind_mobile | Core |
| 15:22.09 | Stormwind_mobile | Led backlight? |
| 15:22.16 | Stormwind_mobile | RGB LEDs? |
| 15:22.37 | Stormwind_mobile | Everything goes through a kind of OS, as I understand? |
| 15:39.52 | *** join/#card10badge Stormwind_mobile (~Stormwind@dslb-002-206-218-106.002.206.pools.vodafone-ip.de) |
| 17:04.10 | *** join/#card10badge DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 18:04.36 | *** join/#card10badge erlehmann_ (~erlehmann@46.114.35.169) |
| 18:09.11 | *** join/#card10badge DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 18:16.28 | *** join/#card10badge theAntimon (~matze@p54B09205.dip0.t-ipconnect.de) |
| 18:18.07 | *** join/#card10badge maxammann (~maxammann@v37187.1blu.de) |
| 18:55.28 | *** join/#card10badge theAntimon (~matze@p54B09205.dip0.t-ipconnect.de) |
| 19:06.11 | schneider | it does, but nothing is keeping track |
| 19:06.25 | schneider | standby current should be around 40 uA |
| 19:07.06 | schneider | but there are situations where this is not the case. namely if there was a sensor active when entering standby |
| 19:07.20 | schneider | (because the code does not yet turn all of them off when going to sleep) |
| 19:14.59 | Stormwind_mobile | <PROTECTED> |
| 19:17.01 | Stormwind_mobile | It's important to know when the LCD backlight is on, and the statuses of the LEDs, as well. Not sure how complicated that would become. I at least imagine it's theoretically easy to implement. But I imagine most of my ideas are theroretically easy... ^^° |
| 19:31.34 | schneider | i guess it depends on what you want to do with the value. It is possible to judge the energy consumption of a device that way, yes. But is it enough to create a semi accurate battery estimate? |
| 19:32.47 | schneider | i think showing a voltage based one is simply good enough. there are quite a few more interesting things to tackle (IMHO) |
| 19:35.33 | Stormwind_mobile | You know the system like few other people do, so I trust you on that. |
| 19:37.35 | Stormwind_mobile | A possibility to improve the accuracy of the charge state by voltage would be if someone characterized the discharge curve of the LiPo cell, or a bunch of those, preferably, to be able to get some statistics. |
| 19:37.54 | Stormwind_mobile | Or is there a datasheet of the cell that already gives that information? |
| 19:41.17 | schneider | i don't have a real datasheet of the cell |
| 19:41.30 | schneider | (But quite a few cells...) |
| 19:41.51 | Stormwind_mobile | I want to build a battery cycled since basically forever. |
| 19:42.22 | Stormwind_mobile | Well, since I worked that student job in micro LiPo battery integration |
| 19:43.39 | Stormwind_mobile | They have a 19" tower with expensive equipment that can do four point measurement, very granularly programmable charge/discharge cycles,... |
| 19:44.01 | Stormwind_mobile | With a terrible UI on a windows XP PC... |
| 20:33.26 | *** join/#card10badge DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 22:09.26 | *** join/#card10badge erlehmann_ (~erlehmann@46.114.33.159) |
| 22:31.20 | DocScrutinizer05 | lol |
| 22:32.45 | DocScrutinizer05 | >>[20 Oct 2019 15:34:48] <DocScrutinizer05> ok, my card10's battery is "dead" now, after weeks in suspend :-D<< didn't boot up anymore... BUT... battery wasn't dead enough to stop RTC which still is correct" (well, ~11 minutes fast now") |
| 22:33.17 | *** join/#card10badge Lorenzo (~lzmartini@fez.tardis.ed.ac.uk) |
| 22:34.19 | DocScrutinizer05 | strange enough, battery only took a 50mAh during charge, according to my USB meter |
| 22:35.08 | DocScrutinizer05 | thought the battery is supposed to be 100 (130?) mAh |
| 22:35.56 | DocScrutinizer05 | ok, 50mAh*5V |
| 22:36.37 | DocScrutinizer05 | meter says "255mWh" |
| 22:37.44 | DocScrutinizer05 | 100mAh * 3.7V (nominal) is a tad more than 255mWh |
| 22:48.04 | *** join/#card10badge theAntimon (~matze@p54B09205.dip0.t-ipconnect.de) |
| 23:35.44 | *** join/#card10badge erlehmann (~erlehmann@46.114.33.159) |
| 23:58.51 | *** join/#card10badge mmu_man (~revol@vaf26-2-82-244-111-82.fbx.proxad.net) |