| - chomwitt (QUIT: Remote host closed the connection) (~chomwitt@2a02:587:dc0f:ad00:3949:ab21:9834:ebc) | 00:00 | |
| mrus | alright cool, then let's go for RHEL | 00:40 |
|---|---|---|
| kfx | _E: well, you know how the old saying goes. resistance is futile. | 00:43 |
| _E | kfx: ironically so, since I think R53's melting shorted it to ground | 00:44 |
| _E | almost tempted to clean up the melted solder and reseat it | 01:30 |
| _E | it'd be interesting to see what still works -- if anything does | 01:33 |
| ex-parrot | _E: careful, I think that moved resistor might be the intentional board fix on early units? | 01:35 |
| ex-parrot | let me compare with mine | 01:35 |
| ex-parrot | yes, confirmed, that R53 should be moved like that | 01:36 |
| ex-parrot | and was likely done by minute's own hand during manufacturing :) | 01:36 |
| ex-parrot | it would be pretty amazing if the CPU usage alone melted solder | 01:37 |
| _E | ex-parrot: what, not even a little yellow wire for the trouble? | 01:39 |
| ex-parrot | it had to be disconnected and moved | 01:40 |
| ex-parrot | so I'm not even sure it could be done with a bodge wire | 01:41 |
| _E | all I know is a reliable IR thermometer clocked at least one part of the board at 210°F | 01:41 |
| ex-parrot | are you running from battery? | 01:41 |
| _E | it was on battery for a while, but I plugged it in and nothing good has happened since | 01:42 |
| ex-parrot | there is a known quirk with the balancing resistors where if e.g. a battery becomes disconnected during operation they can get very hot and burn up | 01:42 |
| ex-parrot | due to an undesirable current path | 01:42 |
| ex-parrot | I've never managed to get the CPU on my unit above about 70 degrees even working it very hard intentionally, though I suppose a hundred degrees might be possible | 01:42 |
| _E | the 210°F reading was, as best I can tell, the right side battery connector | 01:43 |
| ex-parrot | that's around where the balancing resistors are | 01:44 |
| ex-parrot | so I'd say it was a power issue | 01:44 |
| _E | various parts of the board near the LPC and the battery controller were in the mid-to-upper 150°s | 01:44 |
| ex-parrot | the balancing resistors are those bigger 4.7 ohm ones in a row near the board edge | 01:45 |
| ex-parrot | a few people have reported instances of them overheating and cooking in some situations | 01:45 |
| _E | I don't recall them being terribly hot, but I can check the resistance later | 01:46 |
| ex-parrot | minute finally identified why recently, basically if a battery is removed the charge / discharge current can be directed through one of them, which exceeds their rating | 01:46 |
| ex-parrot | (someone correct me if I'm wrong) | 01:46 |
| _E | the batteries were solidly connected during the incident | 01:46 |
| _E | at least physically | 01:46 |
| ex-parrot | yeah, hard to say really, I'm just guessing at the most likely thing to get hot over in that area of the board | 01:47 |
| _E | can the board run off just AC? | 01:47 |
| _E | no batteries present, just the power adapter | 01:47 |
| ex-parrot | yes but there may be conditions you have to fulfill first that I don't remember, so probably best to check with minute or someone more familiar with that operation | 01:48 |
| ex-parrot | is your LPC still going, e.g. you can hit circle and get the little menu? you could check the cell voltages there if so | 01:48 |
| _E | it wasn't at the time, but I could absolutely be convinced there was a thermal limit tripped | 01:49 |
| ex-parrot | I'm not sure there is much that would implement a thermal limit per se | 01:49 |
| _E | I would guess the LPC has an internal one? | 01:50 |
| _E | I saw mention of one on a datasheet recently, but I can't say if it was the LPC's | 01:50 |
| ex-parrot | the LPC is just a microcontroller | 01:50 |
| _E | must have been the iMX8's sheet then | 01:51 |
| _E | yeah, that was it | 01:52 |
| ex-parrot | yeah the iMX has a thermal sensor on it | 01:52 |
| _E | yeah, it was 5.4 in the manual | 01:53 |
| _E | that's definitely what I was looking at | 01:53 |
| - mtm (QUIT: Ping timeout: 246 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:02 | |
| mrus | wait, so that means we can't blame Feodra on melting solder and burning PCBs? | 02:16 |
| ex-parrot | I'm skeptical | 02:16 |
| mrus | :[ | 02:16 |
| _E | well, I took it somewhere less flammable | 02:24 |
| _E | it was back up to an unreasonably high temp almost as soon as I plugged the battery packs back in | 02:24 |
| _E | up past 100°F within 30s | 02:25 |
| _E | with the laptop off | 02:25 |
| ex-parrot | right, sounds like you've got a fault in the balancing circuitry | 02:26 |
| _E | by the time I got the plugs out it had already started emitting That Smell™ | 02:26 |
| ex-parrot | you can probably dribble some isopropyl alcohol on the board while cold | 02:26 |
| ex-parrot | to see which components are heating up | 02:26 |
| ex-parrot | (e.g. it'll evaporate fastest off the hot bits) | 02:27 |
| ex-parrot | are you using the stock batteries? | 02:27 |
| _E | yep | 02:27 |
| _E | straight from CS | 02:27 |
| ex-parrot | cool, that's a good start | 02:27 |
| ex-parrot | you could check that they are all delivering a reasoable voltage in to the cell holders | 02:28 |
| _E | sure, one sec | 02:29 |
| _E | right pack is all between 3.2-3.3 | 02:31 |
| ex-parrot | cool, in that case I'm out of good ideas for what might be heating up. I'd suggest try measuring it with some isopropyl alcohol if you can | 02:31 |
| _E | left pack is all 3.3 | 02:31 |
| _E | I have a thermocouple for my meter, but I can't exactly get the batteries unplugged quickly if something starts going very wrong | 02:32 |
| _E | unless someone has an idea for speedy battery removal | 02:36 |
| _E | correction: I found a pair of needle nose pliers | 02:39 |
| _E | which is probably as good as I can hope to get | 02:39 |
| _E | I would like the record to state that this is a bad idea | 02:39 |
| ex-parrot | yeah, don't pull out individual cells | 02:45 |
| ex-parrot | while the machine is on | 02:45 |
| ex-parrot | or you can burn the balancing circuitry | 02:45 |
| _E | ok no that's not a good idea, the batright connector was back up past 140°F within 10 seconds of being plugged back in | 02:45 |
| _E | for reference, the ambient temp is roughly 66°F at the moment | 02:46 |
| _E | ex-parrot: I'm only using the connectors, not messing with cells | 02:47 |
| ex-parrot | ok v good | 02:47 |
| ex-parrot | if the connector itself is getting hot | 02:47 |
| ex-parrot | I assume something is shorted | 02:47 |
| _E | no obvious debris on the board | 02:48 |
| _E | gonna take off the peripherals and try again | 02:49 |
| ex-parrot | _b | 02:51 |
| - alex4nder (QUIT: Ping timeout: 250 seconds) (~alexander@ip98-171-191-229.sb.sd.cox.net) | 02:55 | |
| _E | no change | 03:00 |
| _E | batright was up to 170°F before I got it out | 03:00 |
| + alex4nder (~alexander@ip98-171-191-229.sb.sd.cox.net) | 03:01 | |
| _E | the resistors were also hot, but not as... violently so | 03:01 |
| ex-parrot | weird | 03:01 |
| _E | I wonder if it happens with only one battery pack? | 03:01 |
| _E | ...it still happens with F2 gone | 03:06 |
| _E | that's somewhat unexpected | 03:09 |
| ex-parrot | that is weirder! | 03:10 |
| ex-parrot | maybe the connector itself is damaged in some way? | 03:10 |
| _E | could be? | 03:10 |
| _E | I don't know why it would have happened spontaneously like that | 03:10 |
| ex-parrot | as I've said, seems weird :P | 03:11 |
| ex-parrot | hopefully someone with more experience on that bit of the hardware will show up and have some more ideas about trbouleshooting it | 03:11 |
| _E | I'm gonna look over the schematics a bit | 03:11 |
| _E | unsurprisingly, that didn't help one bit | 03:39 |
| _E | though I am somewhat shocked at how little of a role F2 plays | 03:39 |
| _E | I thought it would have stood in the way of everything, battery control circuitry included | 03:40 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 04:08 | |
| + xktr (~xktr@108.62.5.153) | 04:17 | |
| - GNUmoon (QUIT: Ping timeout: 240 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 04:30 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 04:42 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:40) | 06:03 | |
| - adjtm (QUIT: Quit: Leaving) (~adjtm@53.red-83-56-237.dynamicip.rima-tde.net) | 10:08 | |
| + Christoph_ (~Christoph@p54bf6fc1.dip0.t-ipconnect.de) | 11:18 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 12:24 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 12:24 | |
| bluerise | minute: so you have cm4 adapters ready? :P | 12:54 |
| minute | bluerise: only 1 that i'm still testing... i can make you another one if you agree to struggle with DSI again with me | 12:55 |
| bluerise | I can agree to struggle with DSI again | 12:55 |
| minute | cool | 12:56 |
| minute | the rcm4 in the tweet was made independently by jacqueline (from my files) | 12:56 |
| minute | she already cloned the rest of reform independently | 12:57 |
| bluerise | Yeah, I remember she built her own reform already | 12:57 |
| minute | yeah | 12:57 |
| bluerise | with USB-C port | 12:57 |
| minute | yeah | 12:57 |
| bluerise | I found it really impressing. So much work, all those tiny solderings | 12:57 |
| minute | i feel like this is kind of a big deal, actually reproducible oshw (and complex one) | 12:57 |
| bluerise | Yeah! | 12:57 |
| bluerise | I have two other OpenBSD devs with an MNT Reform, one of them was already asking me about CM4 ;) | 12:58 |
| ex-parrot | \o/ | 12:59 |
| bluerise | https://twitter.com/jcs/status/1504148181977817089 | 12:59 |
| ex-parrot | oh jcs got his | 13:00 |
| ex-parrot | that's really good | 13:00 |
| bluerise | Yeah | 13:00 |
| bluerise | Love that guy | 13:00 |
| ex-parrot | me too | 13:00 |
| + chomwitt (~chomwitt@2a02:587:dc0f:ad00:bf65:8d32:22b6:5d0b) | 13:14 | |
| minute | ok, we'll assemble some more rcm4s next week then | 13:31 |
| - mtm (QUIT: Ping timeout: 240 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:02 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 16:09 | |
| - GNUmoon (QUIT: Ping timeout: 240 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:01 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:01 | |
| _E | mrus: I still plan on blaming Fedora until there's something better to blame | 18:05 |
| _E | ...like my kernel config | 18:05 |
| _E | minute: can the reform run without any batteries connected? | 18:34 |
| minute | _E: yes, no problem | 18:34 |
| _E | any special config needed compared to the default setup? | 18:35 |
| minute | _E: you can also power it with some usb-c powerbanks + an adapter | 18:35 |
| minute | _E: i don't believe a special config is needed | 18:35 |
| minute | we're doing a bunch of testing without any batteries for each laptop | 18:35 |
| _E | trying to troubleshoot some sort of power issue with a ~week old unit and I'd much rather do it without batteries in the mix | 18:35 |
| minute | _E: yes, good call | 18:41 |
| _E | well, that was a short-lived test | 18:42 |
| _E | I don't think the outlook is good for this board | 18:43 |
| _E | something in the neighborhood of L1 and U12 is getting very, very hot, very, very quickly | 18:51 |
| _E | ex-parrot: I have a sinking feeling that it's more than just the battery system | 18:52 |
| minute | _E: hmm, that's bad. something shorted perhaps | 18:53 |
| minute | _E: which country are you in? | 18:53 |
| _E | US | 18:54 |
| minute | ah, then it's a long haul to let me look at it | 18:54 |
| _E | similar symptoms during battery-only operation | 18:54 |
| _E | this is with absolutely nothing attached at this point | 18:55 |
| _E | besides the display cable, because it was a huge pain to get on | 18:55 |
| _E | minute: it's not exactly lab-grade equipment, but I do have a multimeter and some solder in case there are tests I can do myself | 18:56 |
| minute | _E: yeah, stop testing with batteries for now, i would say | 18:57 |
| minute | _E: would you be able to get a cheap lab power supply? that way you can limit the current | 18:57 |
| _E | maybe, but there would be quite a bit of emphasis on "cheap" | 18:58 |
| alex4nder | you can get cheap buck/boost converters with current limiting for < $15 | 19:03 |
| alex4nder | (e.g. off amazon) | 19:03 |
| minute | yeah but you can also get an ok lab power supply for EUR 35 or so | 19:10 |
| minute | ok, these are $50-$60 in the US it appears https://www.amazon.com/Adjustable-HANMATEK-HM305-Variable-Switching/dp/B0852JZQZR | 19:11 |
| _E | just tried ex-parrot's isopropyl alcohol trick | 19:21 |
| _E | it's looking like Q17 or something in it's very immediate vicinity | 19:21 |
| _E | at least, as the heat source | 19:22 |
| _E | and by that I mean it's getting hot enough for the alcohol to sizzle | 19:29 |
| _E | about 3-5s after having power applied | 19:29 |
| _E | no peripherals or removable components, entirely extracted from the chassis | 19:31 |
| mrus | _E: I fully support this | 19:44 |
| alex4nder | minute: hey, am I remembering right that you do compilation on a honeycomb lx2? how do you like it? | 19:45 |
| minute | alex4nder: i like it a lot, i also did 2 days of video editing on it | 19:47 |
| minute | but i did the silent noctua fan mod and it doesn't cool well enough | 19:48 |
| minute | so i ran into thermal shutdowns when a lot of cores were grinding | 19:48 |
| alex4nder | ok, good to know | 19:48 |
| minute | i am using an old radeonsi card with it | 19:48 |
| minute | also i'm using UEFI, debian, but with custom kernel build from boundary | 19:49 |
| minute | mainline is missing some smmu stuff which leads to pain | 19:49 |
| minute | sorry, not boundary, solidrun | 19:56 |
| alex4nder | sweet, using nvme as storage? | 19:56 |
| minute | alex4nder: yep | 19:56 |
| _E | mrus: you know, we could go further and blame poettering directly | 19:57 |
| alex4nder | minute: cool, did you go with ECC sodimms? it looks like it supports them | 19:57 |
| mrus | _E: you mean for crimes against humanity (systemd)? | 19:58 |
| minute | don't remember atm alex4nder | 19:58 |
| _E | mrus: why does no one remember he's also behind Avahi | 20:00 |
| mrus | I do remember, but avahi hasn't really become a crucial component of everyday Linux | 20:01 |
| mrus | it doesn't really impact people, no matter how good or bad it is | 20:01 |
| minute | _E, mrus: please refrain from personal attacks in this channel | 20:01 |
| mrus | minute: I'm obviously joking | 20:01 |
| _E | lennart started it | 20:02 |
| _E | /s | 20:02 |
| mrus | quote from Wikipedia: "In 2017, Poettering received the Pwnie Award for Lamest Vendor Response to vulnerabilities reported in systemd" | 20:03 |
| mrus | it's not even a personal attack in first place | 20:03 |
| mrus | it's raising awareness for the awards he received | 20:03 |
| _E | minute: I can't say for sure if I'll be able to get an appropriate power supply any time soon | 20:03 |
| _E | so if there are other tests that can be done in the interim | 20:04 |
| minute | _E: ok, i'm currently on vacation and sick, so i must not look at schematics etc right now, but can help from tuesday | 20:04 |
| minute | idk by heart what Q17 is | 20:05 |
| alex4nder | _E: where in the US are you located? | 20:05 |
| _E | alex4nder: VA | 20:05 |
| alex4nder | ah ok, I'm on the other side | 20:05 |
| _E | minute: I'm impressed at your ability to schedule your recovery from an illness | 20:05 |
| minute | _E: easy | 20:06 |
| minute | the rule is, i get sick as soon as i am on vacation | 20:06 |
| minute | and recover in time for the end of it | 20:06 |
| alex4nder | perfect | 20:06 |
| minute | perfectly tuned organism | 20:06 |
| kfx | _E: plenty of hackerspaces in virginia. might see if a nearby one can lend gear to a troubleshootiung effort | 20:12 |
| _E | not so much in my area | 20:13 |
| _E | only one I'm aware of is a few cities away, and of dubious open-ness | 20:13 |
| alex4nder | hmm, looks like ECC sodimms are not that expensive right now | 20:17 |
| + adjtm (~adjtm@53.red-83-56-237.dynamicip.rima-tde.net) | 21:16 | |
| - xktr (QUIT: Ping timeout: 240 seconds) (~xktr@108.62.5.153) | 21:20 | |
| + alexande1 (~alexander@wsip-184-186-225-187.sb.sd.cox.net) | 21:28 | |
| alexande1 | ok.. reform nvme debootstrapped from upstream debian.. | 21:38 |
| alexande1 | and of course I've got broken X11 fonts | 21:38 |
| * alexande1 -> al3x4nder | 21:42 | |
| minute | al3x4nder: x11? | 21:44 |
| al3x4nder | minute: yah, anything going through Xwayland is rendering fonts as blocked rectangles | 21:44 |
| al3x4nder | I've just derped something, but I'm not sure what yet | 21:44 |
| minute | al3x4nder: i know what that is | 21:45 |
| minute | al3x4nder: you need to set ETNA_MESA_DEBUG=nir in your profile | 21:45 |
| minute | kinda a running gag at this point | 21:45 |
| al3x4nder | ah ok | 21:45 |
| al3x4nder | ACTION tries that | 21:45 |
| - al3x4nder (QUIT: Quit: brb) (~alexander@wsip-184-186-225-187.sb.sd.cox.net) | 21:47 | |
| + al3x4nder (~alexander@wsip-184-186-225-187.sb.sd.cox.net) | 21:48 | |
| al3x4nder | minute: rad, thank you.. I would have been searching for that forever | 21:49 |
| minute | > It was on display in the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying "Beware of the Leopard" | 21:50 |
| al3x4nder | dude exactly | 21:50 |
| scops | hmmm if i want to replace the kernel for booting vom nvme, is it better to flash the emmc (V2-rescue is installed) or should i just replace initrd.img (Image on emmc) and dtb with the ones that comes with the sd card image? | 21:54 |
| scops | s/vom/of/ | 21:55 |
| scops | +f | 21:55 |
| minute | hmm, i don't really have much experience with testing the v3 stuff with initrd on emmc yet, so i don't want to give false advice | 22:05 |
| minute | maybe josch can weigh in | 22:06 |
| josch | scops: I'll test eMMC boot this weekend (we have an extra holiday due to easter) and will come back to you then. | 22:06 |
| josch | ACTION hasn't yet changed the dip switch under the heatsink | 22:06 |
| scops | check 🙂 atm i booted into a fresh done encrypted nvme v3 image with the old v2 kernel on emmc | 22:07 |
| scops | (and making a backup of my old emmc data just for the case) | 22:08 |
| minute | cool! | 22:09 |
| - al3x4nder (QUIT: Ping timeout: 246 seconds) (~alexander@wsip-184-186-225-187.sb.sd.cox.net) | 22:18 | |
| + MajorBiscuit (~MajorBisc@2a02-a461-129d-1-6d4c-38a4-18b7-4b48.fixed6.kpn.net) | 22:46 | |
| scops | hmmm another problem... i just flashed the lpc firmware. the keyboard can't wake up the lpc. any ideas? | 22:54 |
| scops | ok i think it was tools/lpcrc which isn't build with the main Makefile | 22:56 |
| - MajorBiscuit (QUIT: Ping timeout: 240 seconds) (~MajorBisc@2a02-a461-129d-1-6d4c-38a4-18b7-4b48.fixed6.kpn.net) | 23:07 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:19 | |
| scops | ah a suggestion; it would be nice if firmware updates doesn't require a whole compilation and >2gb compile tools (for the lpc firmware for example). providing the ready to use binaries like for the systemimages would be really nice, imho | 23:27 |
| vagrantc | scops: i put a little bit of work into packaging the firmware for debian: https://salsa.debian.org/vagrant/mnt-reform-firmware | 23:30 |
| vagrantc | but it has been a while since i looked at it ... would be really helpful if the different firmwares were in separate repositories | 23:31 |
| scops | yeah | 23:32 |
| vagrantc | but i actually used the firmware produced from that on my mnt/reform ... at least the lpc and maybe one other | 23:32 |
| vagrantc | or at least all the firmwares were in a separate repository | 23:33 |
| scops | i just updated my lpc and keyboard firmware (from official repo) since i haven't done it since i got my little black laptop 😉 (https://community.mnt.re/t/mnt-reform-recommended-lpc-and-keyboard-firmware-update-to-fix-battery-drain-in-off-state/559) | 23:34 |
| vagrantc | right, probably those are the two i updated | 23:36 |
| scops | ah and it wasn't noticed in the handbook that i had to install the package "gcc-arm-none-eabi" to compile the lpc firmware... and the tools/lpcrc stuff | 23:39 |
| scops | this should be added 🙂 | 23:39 |
| vagrantc | i prefer packaging to prose instructions, as it's easier to get right :) | 23:44 |
| vagrantc | and easy to test if it works or not (e.g. not making assumptions about what's already installed) | 23:46 |
| scops | Me too, but the handbook should also have all important informations in it, if they are relevant. On the lpc flashing part nothing is mentioned about the tools you need to compile the stuff. On the keyboard files at least the readme.md lists the tools needed | 23:49 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!