--- Log opened Sat Nov 01 00:00:44 2025
2025-11-01T00:20:38 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-01T00:37:18 -!- ventZl [~ventZl@46-13-58-131.customers.tmcz.cz] has quit [Quit: Client closed]
2025-11-01T00:37:43 -!- ventZl [~ventZl@46-13-58-131.customers.tmcz.cz] has joined ##stm32
2025-11-01T00:55:20 -!- ventZl [~ventZl@46-13-58-131.customers.tmcz.cz] has quit [Quit: Client closed]
2025-11-01T01:21:44 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-01T01:37:54 < ds2> the ds18b20's are fine
2025-11-01T01:53:57 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-01T01:56:18 -!- jerrycash2 [~jerrycash@user/jerrycash] has quit [Ping timeout: 252 seconds]
2025-11-01T02:35:55 -!- ferdna_ [~ferdna@user/ferdna] has joined ##stm32
2025-11-01T02:44:48 -!- Netsplit *.net <-> *.split quits: c10ud, octorian, ice, ferdna, PaulFertser, karlp
2025-11-01T02:50:59 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-01T02:50:59 -!- Netsplit over, joins: PaulFertser, octorian, karlp
2025-11-01T02:52:03 -!- Netsplit over, joins: ice
2025-11-01T02:52:46 -!- fenugrec_ [~f@192.214.232.39] has joined ##stm32
2025-11-01T02:54:05 -!- Alexer- [~alexer@alexer.net] has joined ##stm32
2025-11-01T02:55:42 -!- benishor_ [~benishor@scene.ro] has joined ##stm32
2025-11-01T02:56:49 -!- splud [~noneya.bi@user/splud] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- benishor [~benishor@scene.ro] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- fenugrec [~f@192.214.232.39] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- Alexer [~alexer@alexer.net] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- blathijs [~matthijs@tika.stderr.nl] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:49 -!- ds2 [~ds2@69.54.141.155] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:50 -!- antto [~pewpew@antonsavov.net] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:51 -!- sugarbeet [~barbas@81.4.123.134] has quit [Ping timeout: 256 seconds]
2025-11-01T02:56:51 -!- sugarbeet [~barbas@81.4.123.134] has joined ##stm32
2025-11-01T02:56:51 -!- benishor_ is now known as benishor
2025-11-01T02:57:12 -!- blathijs [~matthijs@tika.stderr.nl] has joined ##stm32
2025-11-01T02:57:12 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-01T03:03:40 -!- server [~ds2@69.54.141.155] has joined ##stm32
2025-11-01T03:03:42 -!- antto [~pewpew@antonsavov.net] has joined ##stm32
2025-11-01T03:04:30 -!- splud [~noneya.bi@user/splud] has joined ##stm32
2025-11-01T04:40:34 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-01T04:47:41 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T05:26:03 -!- hexo_ [~hexo@user/hexo] has joined ##stm32
2025-11-01T05:26:27 -!- hexo_ is now known as hexo
2025-11-01T05:28:05 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-01T05:49:14 -!- _Posterdati_ [~posterdat@user/Posterdati] has quit [Ping timeout: 244 seconds]
2025-11-01T06:02:35 -!- _Posterdati_ [~posterdat@user/Posterdati] has joined ##stm32
2025-11-01T06:30:53 -!- ferdna_ [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-01T06:43:02 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-01T06:43:24 -!- emeb_mac [~emeb_mac@ip174-73-147-156.ph.ph.cox.net] has quit [Quit: Leaving.]
2025-11-01T06:50:40 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T07:23:51 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-01T07:29:48 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T07:34:55 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-01T07:37:49 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-01T07:40:46 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T07:53:18 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-01T08:52:14 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-01T09:06:52 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-01T10:39:21 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-01T11:52:23 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Quit: KVIrc 5.2.6 Quasar http://www.kvirc.net/]
2025-11-01T11:53:18 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-01T12:00:06 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-01T12:05:28 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Remote host closed the connection]
2025-11-01T12:05:58 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-01T12:07:09 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Remote host closed the connection]
2025-11-01T12:07:12 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T12:07:35 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-01T12:47:08 < BrainDamage> mawk: https://en.wikipedia.org/wiki/Four-terminal_sensing also called kelvin sensing
2025-11-01T13:04:44 -!- kovalevsky_ [~kovalevsk@181.168.239.174] has joined ##stm32
2025-11-01T13:05:35 -!- kovalevsky_ [~kovalevsk@181.168.239.174] has quit [Remote host closed the connection]
2025-11-01T13:16:23 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-01T13:45:40 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-01T14:00:56 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Quit: Segmentation fault]
2025-11-01T14:47:25 -!- teknix_ [~unknown@user/hsv] has quit [Ping timeout: 264 seconds]
2025-11-01T14:53:36 < jbo> moin
2025-11-01T14:59:28 < Steffanx> Gooday sir
2025-11-01T14:59:56 < Steffanx> What brings you here today?
2025-11-01T15:12:27 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-01T15:23:27 -!- kurfen [~kurfen@45.84.137.25] has joined ##stm32
2025-11-01T16:03:52 < jbo> sharing happyness with you
2025-11-01T16:09:32 < Steffanx> I see
2025-11-01T16:09:59 < Steffanx> Explains why you we're not sharing much atm :P
2025-11-01T16:10:53 < Steffanx> -you
2025-11-01T16:11:47 < jbo> lol
2025-11-01T16:11:59 < jbo> how is life?
2025-11-01T16:18:59 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-01T17:01:11 -!- teknix_ [~unknown@user/hsv] has quit [Ping timeout: 256 seconds]
2025-11-01T17:21:30 < Steffanx> Could be better could be worse 
2025-11-01T18:04:34 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-01T18:06:08 -!- Thorn [~Thorn@user/thorn] has quit [Read error: Connection reset by peer]
2025-11-01T18:07:21 -!- Thorn [~Thorn@user/thorn] has joined ##stm32
2025-11-01T18:32:42 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-01T18:38:54 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-01T18:54:19 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T18:56:45 -!- teknix_ [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-01T19:08:40 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-01T19:20:57 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-01T19:31:17 < jbo> aye
2025-11-01T19:44:57 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-01T20:01:51 -!- teknix_ [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-01T20:06:37 < Steffanx> And for mr jbo?
2025-11-01T20:38:29 < jbo> all good
2025-11-01T20:39:52 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-01T20:55:05 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-01T21:45:43 -!- akaWolf [~akaWolf@akawolf.org] has joined ##stm32
2025-11-01T22:20:48 < Steffanx> Awesome 
2025-11-01T22:22:05 < mawk> BrainDamage: but that's not for kilometers of wire is it
2025-11-01T22:22:10 < mawk> that's for local measurements
2025-11-01T22:23:57 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-01T22:35:51 < qyx> it is for whatever you need, it works
2025-11-01T22:43:53 < mawk> I see
2025-11-01T22:44:30 < mawk> I tried to draw a constant low current source but it turns out it's pretty hard to make it accurate over a range of load resistances
2025-11-01T22:45:55 < mawk> the typical 100k gain of an opamp ends up not being enough 
2025-11-01T22:55:04 < qyx> you don't use a constant current source
2025-11-01T23:18:47 < qyx> ok editing pcb in kicad with ~1600 pads, 850 vias and 2500 track segments is a little bit slow
2025-11-01T23:18:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-01T23:36:21 -!- teknix [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-01T23:40:07 -!- teknix_ [~unknown@user/hsv] has quit [Ping timeout: 256 seconds]
2025-11-01T23:41:31 < mawk> what do you do then
2025-11-01T23:41:36 < mawk> a wheatstone bridge?
2025-11-01T23:41:54 < mawk> this shows a constant current source https://en.wikipedia.org/wiki/Resistance_thermometer#Four-wire_configuration
2025-11-01T23:43:42 < qyx> did you see the link I pasted?
2025-11-01T23:44:20 < qyx> fig 1a
2025-11-01T23:47:05 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-01T23:48:51 < mawk> ah no
2025-11-01T23:49:02 < mawk> it's lost in the backlog
2025-11-01T23:49:09 < qyx> https://bin.jvnv.net/file/0V4dC/Screenshot_2025-11-01_22-48-52.png
2025-11-01T23:49:12 < mawk> thanks
2025-11-01T23:49:21 < qyx> https://www.analog.com/en/resources/technical-articles/highaccuracy-temperature-measurements.html
2025-11-01T23:51:28 < qyx> you need no precisiou voltage/current sources, you just do ratiometric measurement with a precision R_ref resistor, which is the only component in the system with tight tolerances
2025-11-01T23:52:26 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-01T23:52:29 < qyx> you are using the fact that whatever current you use to excite the sensor, it is the same along all series connected components, including the Pt100 itself and R_ref
2025-11-01T23:52:59 < qyx> so you can simply use a constant voltage with a series resistor to set the current (not precisely)
2025-11-01T23:53:04 < mawk> I see
2025-11-01T23:53:20 < qyx> and comparing the voltage of the Pt100 (sensed remotely with ~zero current) to the voltage on the reference resistor
2025-11-01T23:57:35 < mawk> nice
--- Day changed Sun Nov 02 2025
2025-11-02T00:07:54 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-02T00:09:13 -!- teknix [~unknown@user/hsv] has quit [Quit: leaving]
2025-11-02T00:09:31 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T00:11:24 -!- teknix [~unknown@user/hsv] has quit [Client Quit]
2025-11-02T00:13:07 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T00:13:51 -!- teknix [~unknown@user/hsv] has quit [Client Quit]
2025-11-02T00:15:31 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T00:18:35 -!- teknix [~unknown@user/hsv] has quit [Client Quit]
2025-11-02T00:19:30 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T00:31:35 -!- teknix [~unknown@user/hsv] has quit [Quit: leaving]
2025-11-02T00:33:08 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T00:35:38 < qyx> length match RMII and SDIO, tidy up silkscreen, another pcb done
2025-11-02T00:36:21 < qyx> I need to do another two before christmas to not be bored
2025-11-02T01:05:08 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 256 seconds]
2025-11-02T01:21:12 < mrec_> did anyone test the ethernet througput on the STM32H7 part?
2025-11-02T01:21:18 < mrec_> I wonder how fast it can go
2025-11-02T02:14:18 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-02T02:21:03 -!- noarb [~noarb@user/noarb] has joined ##stm32
2025-11-02T02:39:01 < zyp> measured by what metric? I expect the MAC itself to be able to saturate a 100M link
2025-11-02T02:42:33 < zyp> I mean, the MAC doesn't have its own packet memory, it just has a DMA engine, so it'll already have to be fast enough to keep up with 100M to not have over/underruns in the middle of a packet
2025-11-02T02:43:14 < zyp> and given that 100 Mb/s on a 32b wide bus, is just a bit over 3 MT/s, that's not really a challenge either
2025-11-02T02:43:49 < zyp> so «how fast can it go», is really just a question of how fast the firmware can prepare/consume packets
2025-11-02T03:05:12 -!- teknix [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-02T03:30:10 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T03:32:01 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-02T07:45:19 -!- qyx [~qyx@84.245.121.39] has quit [Ping timeout: 264 seconds]
2025-11-02T07:46:53 -!- qyx [~qyx@84.245.120.67] has joined ##stm32
2025-11-02T08:01:22 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-02T08:31:17 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-02T09:26:59 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-02T09:41:40 -!- NEYi [~NEYi@195.234.78.187] has quit [Read error: Connection reset by peer]
2025-11-02T10:26:09 < qyx> mrec_: stm32f107 was able to generate 70 mbit/s of UDP traffic iirc
2025-11-02T10:26:23 < qyx> with my lame code and lwip
2025-11-02T10:27:00 < qyx> so yes, as zyp says, it will definitely saturate 100mbit/s
2025-11-02T12:35:54 -!- nerozero [~nerozero@87.253.63.54] has quit [Read error: Connection reset by peer]
2025-11-02T12:51:38 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-02T13:47:52 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-02T15:34:10 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-02T16:45:04 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-02T17:22:09 -!- noarb [~noarb@user/noarb] has quit [Read error: Connection reset by peer]
2025-11-02T17:24:25 -!- noarb [~noarb@user/noarb] has joined ##stm32
2025-11-02T17:56:04 < mawk> should I use the uart to do the 1-wire talking
2025-11-02T17:56:09 < mawk> or bitbang it 
2025-11-02T17:56:38 < mawk> there's less delay stuff to worry about with the uart, so I can keep processing the radio stuff concurrently 
2025-11-02T17:56:52 < mawk> but it looks annoying with an external open drain driver required
2025-11-02T17:58:34 < mawk> why did they put 2 mosfets in figure 2 https://www.analog.com/en/resources/technical-articles/using-a-uart-to-implement-a-1wire-bus-master.html
2025-11-02T17:58:53 < mawk> just the one on the right seems enough 
2025-11-02T18:02:18 -!- teknix [~unknown@user/hsv] has quit [Ping timeout: 256 seconds]
2025-11-02T18:03:24 < mawk> ah to invert the signal
2025-11-02T18:19:24 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-02T18:28:48 < qyx> you most probably need nothing if using stm32
2025-11-02T18:29:05 < qyx> configure uart in half duplex modeopen drai 
2025-11-02T18:29:22 < qyx> and add an external 4k7 or so pullup
2025-11-02T18:29:53 < jpa-> mawk: you don't need external mosfets with STM32, you can just configure pin as open drain
2025-11-02T18:30:14 < mawk> I thought the pin is either alternate function or open drain but not both
2025-11-02T18:30:35 < mawk> at least on the f7
2025-11-02T18:31:39 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-02T18:32:19 < jpa-> otyper is separate from moder
2025-11-02T18:34:49 < mawk> ah right 
2025-11-02T18:36:02 < qyx> that was the case on f1
2025-11-02T20:00:22 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-02T20:02:32 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-02T20:02:45 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-02T20:13:12 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-02T20:29:05 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-02T20:31:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-02T21:28:44 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-02T21:30:50 -!- hexo [~hexo@user/hexo] has quit [Killed (copper.libera.chat (Nickname regained by services))]
2025-11-02T21:30:50 -!- hexo__ is now known as hexo
2025-11-02T21:30:54 -!- NEYi_ [~NEYi@195.234.78.187] has joined ##stm32
2025-11-02T21:31:56 -!- hexo [~hexo@user/hexo] has quit [Read error: Connection reset by peer]
2025-11-02T21:32:56 -!- hexo [~hexo@user/hexo] has joined ##stm32
2025-11-02T21:33:12 -!- sugarbeet [~barbas@81.4.123.134] has quit [Ping timeout: 256 seconds]
2025-11-02T21:34:13 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 256 seconds]
2025-11-02T21:37:18 -!- Netsplit *.net <-> *.split quits: teknix_, Thorn, c10ud_, Posterdati, karlp, PaulFertser, qyx, octorian
2025-11-02T21:37:45 -!- Netsplit over, joins: teknix_
2025-11-02T21:38:12 -!- Netsplit over, joins: qyx
2025-11-02T21:38:32 -!- teknix [~unknown@user/hsv] has quit [Quit: leaving]
2025-11-02T21:38:55 -!- sugarbee1 [~barbas@81.4.123.134] has joined ##stm32
2025-11-02T21:38:55 -!- Thorn [~Thorn@bl20-197-68.dsl.telepac.pt] has joined ##stm32
2025-11-02T21:38:55 -!- c10ud__ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-02T21:38:55 -!- Netsplit over, joins: PaulFertser, octorian, karlp
2025-11-02T21:39:08 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-02T21:43:32 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-02T21:53:53 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-02T21:58:03 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-02T22:00:51 -!- martinmoene__ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-02T22:07:57 < mrec_> that's great, I thought about using an FPGA first to interface the mitsubishi servos / spindle but the STM32H7 seems to be a good choice for this.
2025-11-02T22:08:59 < mrec_> I put a new battery into that CNC 3 months ago recently the power went down to 0V and the machine lost all the settings again, I think I have to make a hard cut as long as the controller is alive.
2025-11-02T22:10:06 < mrec_> if an older cnc loses all the settings you have to go to the controller vendor and they need to put in the options again it sucks.
2025-11-02T22:18:32 < qyx> can't you save the settings to a floppy disk?
2025-11-02T22:21:56 < qyx> so, what's next
2025-11-02T22:22:12 < qyx> I need to choose something from my backlog
2025-11-02T22:29:38 < zyp> mrec_, are you the same guy that's been talking about mitsubishi servos in the EO discord?
2025-11-02T22:32:10 -!- Livio_ [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-02T23:32:31 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Read error: Connection reset by peer]
2025-11-02T23:32:55 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-02T23:54:52 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 260 seconds]
--- Day changed Mon Nov 03 2025
2025-11-03T00:19:20 < mrec_> zyp: yes the machine works fine (here and there some quirks) but I'm looking forward :-)
2025-11-03T00:20:00 < mrec_> the options (features) of a cnc are usually locked by the vendor, operators cannot modify them
2025-11-03T00:20:50 < mrec_> Mitsubishi did quite advanced work prior 1998 (release date of the software), I'm still very impressed.
2025-11-03T00:30:34 < mrec_> I badly want 5 axis, the 1998 system only supports 3+2 axis
2025-11-03T00:48:18 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 260 seconds]
2025-11-03T00:50:36 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-03T00:55:49 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 264 seconds]
2025-11-03T00:57:11 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-03T01:04:49 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 264 seconds]
2025-11-03T01:08:49 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 246 seconds]
2025-11-03T01:17:10 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-03T01:17:13 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-03T01:26:47 -!- teknix [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-03T01:46:34 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-03T01:50:02 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-03T02:47:07 -!- _Posterdati_ [~posterdat@user/Posterdati] has quit [Ping timeout: 264 seconds]
2025-11-03T03:56:20 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-03T04:14:16 -!- _Posterdati_ [~posterdat@user/Posterdati] has joined ##stm32
2025-11-03T04:44:32 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-03T06:07:03 -!- akaWolf [~akaWolf@akawolf.org] has quit [Quit: Lost terminal]
2025-11-03T06:36:36 -!- NEYi_ [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-03T06:56:20 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-03T07:11:43 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-03T07:47:49 -!- noarb [~noarb@user/noarb] has quit [Ping timeout: 246 seconds]
2025-11-03T08:24:20 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-03T08:45:05 < mawk> 3+2 = 5
2025-11-03T08:47:55 < Steffanx> You sure mawk ?
2025-11-03T09:09:11 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-03T09:24:56 -!- _Posterdati_ [~posterdat@user/Posterdati] has quit [Quit: KVIrc 5.2.6 Quasar http://www.kvirc.net/]
2025-11-03T09:30:02 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-03T09:40:03 < qyx> looks like a political billboard
2025-11-03T09:55:42 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 252 seconds]
2025-11-03T09:59:52 < mrec_> 5 axis simultaneous, 3+2 is just you have A+B axis position them and process the part in a regular way with 3 axis
2025-11-03T10:05:34 < mrec_> the problem with that system is just it's old, I don't want to buy more outdated equipment to it. Everything that can be bought was lying on the shelf somewhere for 20 years.
2025-11-03T10:07:33 < mrec_> and I revealed the amplifier protocol like 90% already I have some good notes so linuxcnc would be a good target for it
2025-11-03T10:50:21 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-03T11:23:23 < mawk> I went for a blood sample and amazingly the nurse found a vein
2025-11-03T11:23:37 < mawk> after lots of pressing around my arms with her finger
2025-11-03T11:27:38 -!- teknix [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-03T11:30:07 < Steffanx> Ok
2025-11-03T11:31:01 < mawk> they don't appear to heal that fast
2025-11-03T11:31:09 < mawk> there's still a bunch of scars
2025-11-03T11:31:22 < mawk> I need to tattoo Steffanx 's face on top to conceal them
2025-11-03T11:39:47 < Steffanx> Hm
2025-11-03T11:40:48 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-03T11:44:06 < mawk> show me your best profile 
2025-11-03T11:51:43 < Steffanx> Will you subscribe to my profile mawk? It's only 3.99 euro bucks/month
2025-11-03T11:54:15 < tomeaton17> Steffanx: did you start an onlyfans?
2025-11-03T11:58:35 < Steffanx> Yeah daily pics of arm veins 
2025-11-03T11:59:27 < tomeaton17> whats up with all of this arm vein stuff
2025-11-03T12:07:02 < jpa-> mawk needs more MDMA streams to his peripherals
2025-11-03T12:10:57 < qyx> an external bypass maybe
2025-11-03T12:33:31 < mawk> of course Steffanx 
2025-11-03T12:33:50 < mawk> true gentlemen enjoy good veins tomeaton17 
2025-11-03T12:46:55 < BrainDamage> til mawk's a vampire
2025-11-03T12:47:27 -!- ilgrim [~ilgrim@xinu.me] has joined ##stm32
2025-11-03T13:12:02 -!- _Posterdati_ [~posterdat@user/Posterdati] has joined ##stm32
2025-11-03T13:50:34 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-03T14:45:57 -!- PhantomWork [~PhantomWo@user/phantom] has quit [Quit: Leaving]
2025-11-03T15:03:21  * qyx starting cube
2025-11-03T15:16:11 < qyx> why the hell does H7 have spi1, 2, 3, 4 and 6?
2025-11-03T15:16:14 < qyx> where is the 5th one?
2025-11-03T15:25:43 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-03T15:28:10 < jpa-> maybe it's on the higher pin count packages?
2025-11-03T15:35:05 < qyx> USART sync master mode can work as a SPI master, right?
2025-11-03T15:39:12 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-03T15:42:56 < jpa-> i'm not sure, it is weird
2025-11-03T15:45:36 < mawk> lots of chips have peripheral numbers that don't make a lot of sense, I assume they disable/bin peripherals according to which category of MCU it's going into
2025-11-03T15:45:45 < mawk> and the original template is the same
2025-11-03T15:54:15 -!- specing [~specing@user/specing] has quit [Ping timeout: 244 seconds]
2025-11-03T16:08:52 < qyx> https://community.st.com/t5/stm32-mcus-products/usart-in-spi-mode/td-p/377249/page/2
2025-11-03T16:09:01 < qyx> mr waclawek says it should work
2025-11-03T16:09:13 -!- specing [~specing@user/specing] has joined ##stm32
2025-11-03T16:21:25 < qyx> In this mode, the USART can be used to control bidirectional synchronous serial
2025-11-03T16:21:25 < qyx> communications in master mode. The CK pin is the output of the USART transmitter clock.
2025-11-03T16:21:28 < qyx> No clock pulses are sent to the CK pin during start bit and stop bit. Depending on the state
2025-11-03T16:21:31 < qyx> of the LBCL bit in the USART_CR2 register, clock pulses are, or are not, generated during
2025-11-03T16:21:34 < qyx> the last valid data bit (address mark). The CPOL bit in the USART_CR2 register is used to
2025-11-03T16:21:37 < qyx> select the clock polarity, and the CPHA bit in the USART_CR2 register is used to select the
2025-11-03T16:21:40 < qyx> phase of the external clock (see Figure 576, Figure 577 and Figure 578).
2025-11-03T16:21:43 < qyx> k sorry, why the hell it pastes newlines
2025-11-03T16:23:18 < qyx> it looks exactly like a spi master on the provided figures
2025-11-03T16:36:30 < mawk> apart from the useless start and stop bits
2025-11-03T16:36:33 < mawk> can you disable them?
2025-11-03T16:36:57 < mawk> or do spi peripherals not care about some lag in-between bytes
2025-11-03T16:37:12 < zyp> they don't
2025-11-03T16:38:04 < zyp> data is moved on the selected clock edge and as long as setup/hold times are respected, no other activity on the data lines matters
2025-11-03T16:40:20 < mawk> I see
2025-11-03T16:40:27 < mawk> you ran out of SPI peripherals qyx?
2025-11-03T16:43:09 < qyx> I need to connect 6 sensors with no FIFO
2025-11-03T16:43:17 < qyx> and they are specified for like 2 MHz max
2025-11-03T16:43:40 < qyx> and each is using about ~1 MHz of SPI bandwidth
2025-11-03T16:43:47 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-03T16:43:52 < qyx> they are clearly not designed to share a single SPI bus
2025-11-03T16:57:04 -!- turnip420 [~machinehu@109.202.195.26] has joined ##stm32
2025-11-03T16:58:20 < mawk> right
2025-11-03T16:58:36 -!- turnip420 [~machinehu@109.202.195.26] has left ##stm32 []
2025-11-03T17:06:18 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-03T17:30:31 -!- PaulFertser [paul@paulfertser.info] has quit [Read error: Connection reset by peer]
2025-11-03T17:30:40 -!- PaulFertser [paul@paulfertser.info] has joined ##stm32
2025-11-03T19:21:32 < ColdKeybo[a]rd> Damn RP2350 SDK is... interesting. When they only had RP2040 it still felt a bit hacky but then when they added RP2350A and RP2350B... It just makes me constantly question if the issue is in the FW or HW or SDK.
2025-11-03T19:22:01 < ColdKeybo[a]rd> Miss the good ol' days when I can just blame my poor FW :)
2025-11-03T19:22:02 < ventYl> yeah, they are creative
2025-11-03T19:22:24 < ventYl> toss memory protection into mix and you'll pull all your hair
2025-11-03T19:23:16 < ColdKeybo[a]rd> I could not read anything on a couple of GPIO for days... and then found an old thread on the forum that someone had a similar problem and they just suggested "Oh, undefine PICO_RP2350A and define it again as 0
2025-11-03T19:23:53 < ColdKeybo[a]rd> Because B version has more pins the SDK has to use a different path. But it just assumes you are using A all the time :)
2025-11-03T19:24:58 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-03T19:25:42 < ventYl> uh oh
2025-11-03T19:26:05 -!- flatmush [~benbrewer@149.71.27.16] has joined ##stm32
2025-11-03T20:02:54 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-03T20:04:10 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-03T20:05:48 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-03T20:22:44 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-03T20:32:07 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 244 seconds]
2025-11-03T21:15:25 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-03T21:46:56 < qyx> on H7, VDDUSB can be 0 when USB is not used
2025-11-03T21:47:02 < qyx> sounds weird but datasheet confirms
2025-11-03T21:55:49 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-03T22:04:15 < jpa-> sounds pretty normal, considering you can run rest of the part on 1.8V
2025-11-03T22:05:42 < jpa-> though the power sequencing for that is a bit fun, as VDDUSB must always be max. VDD+0.3V
2025-11-03T22:05:55 < jpa-> err, not always, until VDD is 1V that is :)
2025-11-03T22:15:30 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 256 seconds]
2025-11-03T22:26:09 -!- ice [~ice@loud.house] has quit [Quit: WeeChat 4.7.1]
2025-11-03T22:28:43 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 250 seconds]
2025-11-03T22:29:13 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-03T22:43:44 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-03T22:46:05 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 256 seconds]
2025-11-03T22:58:54 < qyx> also, H7 has 128k sectors
2025-11-03T22:58:58 < qyx> why on earth
2025-11-03T23:03:48 < mawk> what's wrong with that
2025-11-03T23:07:16 < qyx> it is unusable
2025-11-03T23:07:58 < qyx> filesystem on it is hard, settings on it is hard, nvm is hard, bootloader is hard, updating is hard
2025-11-03T23:08:33 < mawk> because it's not enough sectors you mean? and you have to erase lots of it
2025-11-03T23:08:47 < qyx> yes not enough erase granularity
2025-11-03T23:09:44 < mawk> iirc there's an application note about EEPROM emulation you could maybe take inspiration from to increase the granularity
2025-11-03T23:10:06 < mawk> I forgot how they do it
2025-11-03T23:21:32 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-03T23:50:54 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
--- Day changed Tue Nov 04 2025
2025-11-04T00:00:20 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-04T00:03:54 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-04T00:13:42 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 252 seconds]
2025-11-04T01:02:06 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 256 seconds]
2025-11-04T01:02:30 -!- jerrycash2 [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-04T01:04:18 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 260 seconds]
2025-11-04T01:04:39 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T01:05:07 -!- hackkitten [~hackkitte@94.31.119.144] has quit [Ping timeout: 264 seconds]
2025-11-04T01:09:28 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 256 seconds]
2025-11-04T01:11:16 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T01:18:46 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has joined ##stm32
2025-11-04T01:50:06 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-04T01:53:04 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has quit [Ping timeout: 244 seconds]
2025-11-04T01:54:31 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 244 seconds]
2025-11-04T01:54:54 -!- hackkitten [~hackkitte@94.31.119.144] has joined ##stm32
2025-11-04T01:59:28 -!- hackkitten [~hackkitte@94.31.119.144] has quit [Client Quit]
2025-11-04T01:59:39 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has joined ##stm32
2025-11-04T02:07:20 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-04T02:11:27 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-04T02:11:48 -!- hackkitten [~hackkitte@94.31.119.144] has joined ##stm32
2025-11-04T02:31:58 -!- teknix [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T03:28:01 < mrec_> while adding emergency stop feature to my controller I noticed there's a glitch when switching a timer to GPIO and pulling it low, did anyone notice that too?
2025-11-04T03:28:49 < mrec_> I'm registering the emergency stop and stop the PWM output (and also disable the differential output driver in hardware / not via STM32 logic)
2025-11-04T03:30:16 < mrec_> interestingly the glitch doesn't happen so often with timer4, timer3 which has 2 outputs linked it happens most of the time (and rarely doesn't output a logic high glitch for a short moment)
2025-11-04T04:10:13 -!- hackkitten [~hackkitte@94.31.119.144] has quit [Ping timeout: 264 seconds]
2025-11-04T04:10:32 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has joined ##stm32
2025-11-04T04:13:30 < mrec_> ah nah that's something else
2025-11-04T04:17:02 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-04T04:46:11 -!- ferdna [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-04T05:00:30 -!- Thorn_ [~Thorn@2001:8a0:dffd:1c00:e5:73fe:7923:da95] has joined ##stm32
2025-11-04T05:04:24 -!- sugarbeet [~barbas@81.4.123.134] has joined ##stm32
2025-11-04T05:07:39 -!- Netsplit *.net <-> *.split quits: c10ud__, Thorn, karlp, sugarbee1, hackkitten, octorian
2025-11-04T05:07:51 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-04T05:07:51 -!- Netsplit over, joins: hackkitten, octorian, karlp
2025-11-04T05:44:15 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-04T06:08:19 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-04T06:23:58 -!- hackkitten [~hackkitte@2a00:6020:ad3f:300::35e] has quit [Ping timeout: 256 seconds]
2025-11-04T06:35:26 -!- hackkitten [~hackkitte@94.31.119.144] has joined ##stm32
2025-11-04T07:23:11 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-04T07:45:37 -!- qyx [~qyx@84.245.120.67] has quit [Ping timeout: 246 seconds]
2025-11-04T07:45:42 -!- mra90 [~mra90@99.21.38.131] has joined ##stm32
2025-11-04T07:45:53 < mra90> STM32 website is always that bad or just today I can not register an account nor download any software as a guest?
2025-11-04T07:47:32 -!- qyx [~qyx@84.245.120.89] has joined ##stm32
2025-11-04T07:49:53 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-04T08:13:49 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-04T08:14:00 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-04T08:52:14 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-04T09:22:23 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-04T09:35:35 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-04T09:49:46 < qyx> so, placed a buydisplay order
2025-11-04T09:49:48 < qyx> let's see
2025-11-04T09:56:50 -!- mra90 [~mra90@99.21.38.131] has quit [Read error: Connection reset by peer]
2025-11-04T10:01:26 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-04T10:03:46 < c10ud_> i used them in the past
2025-11-04T10:03:51 < c10ud_> always delivered
2025-11-04T10:06:45 < qyx> now placing a jlcpcb order, much china today
2025-11-04T10:11:33 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 252 seconds]
2025-11-04T10:15:14 < Steffanx> Are you ok qyx?
2025-11-04T10:20:05 < qyx> partially
2025-11-04T10:27:24 < qyx> so much money
2025-11-04T10:27:33 < qyx> now placing a jlc3dp order
2025-11-04T10:28:19 < jpa-> separately?
2025-11-04T10:29:53 < qyx> no, combined
2025-11-04T10:29:57 < qyx> 130 eur
2025-11-04T10:30:11 < qyx> and 70 for displays, moneyed chinamen
2025-11-04T10:31:27 -!- Thorn_ [~Thorn@2001:8a0:dffd:1c00:e5:73fe:7923:da95] has quit [Quit: IceChat - It's what Cool People use]
2025-11-04T10:33:02 < jpa-> cheap
2025-11-04T11:02:48 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-04T11:03:42 < qyx> is importlib.resources.files(__package__) somewhat correct way of getting resource files directory?
2025-11-04T11:06:20 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-04T11:10:10 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-04T11:23:17 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-04T11:29:07 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-04T12:16:48 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-04T12:26:02 < qyx> zyp: is there any option for tool.pdm.version to tell it I am about to do a major prerelease? (that is, not a fix one?)
2025-11-04T12:29:20 < qyx> okay apparently there are a few options according to my AI friend
2025-11-04T12:29:22 < qyx> but idk if legit
2025-11-04T12:30:30 < qyx> it should understand conventional commits, so if the commit starts with feat:, it should bump the minor version instead of patch
2025-11-04T12:31:48 < qyx> ok feat: doesn't work, retarded AI friend
2025-11-04T12:34:47 < qyx> ok tagging with a prerelease version doesn't make pdm think it is a minor bump
2025-11-04T12:34:51 < qyx> it makes it crash instead
2025-11-04T12:35:12 < qyx>   File "/tmp/pdm-build-env-ngvnksew-shared/lib/python3.13/site-packages/pdm/backend/hooks/version/scm.py", line 265, in _bump_dev
2025-11-04T12:35:15 < qyx>     assert tail == "0", "own dev numbers are unsupported"
2025-11-04T12:35:21 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-04T12:39:58 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 246 seconds]
2025-11-04T12:56:30 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-04T12:58:21 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-04T13:02:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-04T13:15:48 < jbo> hello
2025-11-04T13:17:48 < Steffanx> Hiiiii
2025-11-04T13:17:53 < jbo> hello welcome!!!!
2025-11-04T13:22:05 < Steffanx> Gooday sir. Wish you a wonderful day today
2025-11-04T13:22:22 < jbo> same to you, thank you!!!!
2025-11-04T13:26:23 < Steffanx> Everything is awesome 👍🏿👍🏿
2025-11-04T13:27:34 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-04T13:27:47 < jbo> ah no, not that tune :D
2025-11-04T13:38:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-04T13:55:13 < jpa-> good morning door
2025-11-04T13:55:16 < jpa-> good morning jbo
2025-11-04T13:59:33 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-04T14:00:02 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-04T14:09:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-04T14:17:42 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-04T14:21:17 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-04T14:28:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 250 seconds]
2025-11-04T14:33:02 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T14:35:07 -!- specing [~specing@user/specing] has quit [Ping timeout: 264 seconds]
2025-11-04T14:45:59 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-04T14:47:25 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 264 seconds]
2025-11-04T14:50:09 -!- duude__- is now known as duude__
2025-11-04T15:05:21 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-04T15:14:01 -!- PhantomWork [~PhantomWo@user/phantom] has joined ##stm32
2025-11-04T15:14:22 < PhantomWork> Hi there, does anyone knows of a good tool to identify bugs, specially memory corruption?
2025-11-04T15:16:00 < PhantomWork> chatgpt and gemini are obsessed on one part of the code, which ain't bug, and their suggestion is what the code already do...   And I assume that this is why it can't find the real bug, because it have an obsession there...
2025-11-04T15:16:53 < jpa-> on PC or on STM32?
2025-11-04T15:16:56 < PhantomWork> (uart circular buffer using DMA, they both insist that doing "tail->end then start->head lead to a data loss)
2025-11-04T15:17:10 < PhantomWork> whatever work
2025-11-04T15:17:23 < jpa-> if you can run the code on PC, fil-c is pretty nice
2025-11-04T15:18:06 < PhantomWork> I suspect that the error is in my circular buffer code, that corrupt the memory elsewhere, including my eeprom in-memory copy
2025-11-04T15:18:28 < PhantomWork> don't think it can run on pc due to the DMA stuff...
2025-11-04T15:20:00 < PhantomWork> beside that, do you have some other idea?
2025-11-04T15:20:45 < jpa-> is the situation repeatable, i.e. is it always the same place that gets corrupt?
2025-11-04T15:25:43 < PhantomWork> I have some debug code, I can force an eeprom reread, and see the data on the screen (uart output). The eeprom data is indeed random.      Also, the uart output also have some corruption (missing data?)    so I assume both are related.    Also, the eeprom read code set the wrong flag, which is not even in the code...  so yeah, major issues
2025-11-04T15:27:48 < jpa-> if you can get a repeatable case, you can set a watchpoint
2025-11-04T15:27:55 < jpa-> though that will not do anything about dma
2025-11-04T15:28:28 < jpa-> but also, fix one thing at a time, if your uart driver is buggy, fix that first
2025-11-04T15:34:26 < PhantomWork> yeah   that is what I try to do...  but now, I found a "should be easy to catch" bug... so fixing this first...
2025-11-04T15:36:09 < karlp> are tehre any "modern" micros that have built in h-bridges?
2025-11-04T15:36:55 < qyx> that jpa/zyp thing with i2c
2025-11-04T15:38:14 < karlp> what?
2025-11-04T15:39:56 -!- drew [~drew@user/drew] has quit [Quit: Quit]
2025-11-04T15:42:01 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-04T15:45:25 < PhantomWork> And I found that simple bug...   had "+" instead of "-", copy-paste-edit error  :D
2025-11-04T15:50:26 -!- tomeaton17 [~tom@user/tomeaton17] has quit [Quit: ZNC 1.9.0+deb2ubuntu0.1~esm2 - https://znc.in]
2025-11-04T15:51:01 -!- tomeaton17 [~tom@user/tomeaton17] has joined ##stm32
2025-11-04T15:51:56 < karlp> hrm, infineon motix is actually reasonably priced, but far more than what I need.
2025-11-04T15:57:56 < mawk> PhantomWork: valgrind and 100% unit test coverage
2025-11-04T15:57:58 < mawk> on PC
2025-11-04T15:58:23 < mawk> valgrind tells you about any kind of uninitialized memory usage or off-by-one errors or overflow or underflow or any kind of flow you like
2025-11-04T15:58:24 < ventYl> PhantomWork: memory protection is a good tool to point out very coarse memory bugs
2025-11-04T15:58:52 < mawk> yeah without a MPU you can happily dereference NULL and it will work fine if the ITCM is enabled
2025-11-04T15:59:05 < mawk> on the cortex M7 and probably others
2025-11-04T15:59:30 < ventYl> mawk: valgrind can even find use after free, double free and similar stuff, if dynamic memory allocation is used
2025-11-04T15:59:38 < mawk> yeah
2025-11-04T16:00:11 < ventYl> memory isolation works on pretty much any Cortex-M except M0
2025-11-04T16:01:03 < ventYl> even STM32F103 can do it
2025-11-04T16:31:02 < PhantomWork> anyone here familliar with I2C eeprom? I have a weird bug.   At boot, cubemxide generated all the init stuff, then I start my own stuff, including an early eeprom read. The first 3 pages are corrupt. In my debug code, I call back the same function, the whole eeprom is OK.     does the I2C hardware take a while to init??? Or did I missed some init stuff? or what?
2025-11-04T16:32:12 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-04T16:33:54 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-04T16:42:11 < karlp> are you running ont he clocks  you expect at that point?
2025-11-04T16:43:12 < karlp> ventYl pimping their MPU OS again ;)
2025-11-04T16:46:26 < PhantomWork> ... I keep finding stupid mistakes that other bugs has hidden from being visible...  
2025-11-04T16:47:31 < PhantomWork> now the corruption is fixed, but... not the slow read after a write
2025-11-04T16:50:41 < jpa-> karlp: i don't know what qyx refers to, but i'd just use some DRVxxxx chip
2025-11-04T16:57:51 < mawk> why is this MDMA stuff so complicated
2025-11-04T16:57:59 < karlp> jpa-: yeah, that's what I'll almost definitely end up with :)
2025-11-04T16:58:17 < karlp> jpa-: was looking at this highly integrated christmas light controller though, was jealous of the single chip
2025-11-04T16:58:22 < karlp> https://www.nxp.com/products/MM908E624 is still active :)
2025-11-04T16:58:34 < karlp> though the infineon motix is probably significantly less pain...
2025-11-04T16:58:52 < karlp> and not meaningfully different in price.
2025-11-04T16:58:52 < tomeaton17> mawk: its not that bad until the next day when your jaw hurts
2025-11-04T17:06:09 < mawk> lol
2025-11-04T17:06:26 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-04T17:21:50 < mawk> I'm trying to use my H7 nucleo board as a coulomb counter
2025-11-04T17:22:02 < mawk> I already have the precision amplifier thing for the current, now I need to log it
2025-11-04T17:22:26 < mawk> half of the pins are bent and a few are broken but the board still appears to work, after bending back a couple necessary pins
2025-11-04T17:45:10 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T17:52:21 < mawk> seems to work first try, good thing there is cubemx to navigate all this DMAMUX/BDMA/DMA/MDMA bullshit
2025-11-04T18:25:01 < PhantomWork> Question guys, STM32F103, HAL and I2C... After an eeprom write, the read is very slow, is that a known issue and is there a work around?
2025-11-04T18:28:09 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-04T18:39:58 < karlp> you're really good at vague problem descriptions arent you :)
2025-11-04T18:40:26 < karlp> how long does the write take, how long does the read take, how are you measuring the times? how long do reads normally take, how do you measure those?
2025-11-04T18:40:36 < karlp> are you landing in timeouts inside hal?
2025-11-04T18:41:25 < PhantomWork> Let me add a write time counter in the code, but it's quick. After a reset, the whole eeprom take 8ms to read AND process the data. After a write that is now 423ms or so for reads
2025-11-04T18:41:31 < ventYl> karlp: forget people will find a connection between mentioning MPU and CreepyOS on their own
2025-11-04T18:47:49 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-04T18:48:32 < PhantomWork> 81ms write time, for the whole eeprom, 8ms read.     but sometime after the write it goes to 423ms...   reinitialising the i2c seems to help, but does not fix it fully, so it seems like it may be an hal issue?
2025-11-04T19:10:14 < jbo> karlp, ping
2025-11-04T19:12:18 < karlp> hrm?
2025-11-04T19:13:04 < jbo> karlp, I'm looking at ublox/trasna LARA-L6 at the moment. In their integration manual, they show layout examples where they simply run the RF antenna pin of the module directly to an SMA. but in other places they have a series AC coupling cap and some additional passives on there. what's up with that?
2025-11-04T19:13:14 < karlp> magic
2025-11-04T19:13:37 < karlp> some vendors ahve started packing more matching inside the SiP
2025-11-04T19:14:03 < karlp> but you can still "tune" it for your "special precious snowflakes" yourself
2025-11-04T19:14:21 < karlp> (to th ebest of my understanding, I'm not a rf guru, I just told you "you too can do rf!"
2025-11-04T19:14:51 < jbo> yeah, hence I am asking you specifically :p
2025-11-04T19:15:58 < karlp> I'm not looking at their manuals to see what else they say :9
2025-11-04T19:16:18 < jbo> not what I meant :p
2025-11-04T19:17:38 < jbo> just wanted to ask if it's common to have that nowdays
2025-11-04T19:17:54 < karlp> yeah, I've seen it on some wch parts, just yolo out that antenna baby
2025-11-04T19:18:02 < karlp> then even has registers for adjusting the caps they have internally
2025-11-04T19:19:01 < Steffanx> Didn't even know ublox sold their cellular division to tresna. I must have missed that news. 
2025-11-04T19:19:07 < Steffanx> I thought it was all done 
2025-11-04T19:19:12 < karlp> well, they said they were exiting
2025-11-04T19:19:23 < karlp> presumably someone was going to take it. but yeah, I ahdn't seen who'd taken it.
2025-11-04T19:19:35 < Steffanx> Why Richard and/or Stefan didn't tell me that, mawk?
2025-11-04T19:19:37 < karlp> is transna a spinout of ex staff? or some existing company they found to buy it?
2025-11-04T19:20:41 < Steffanx> Existing apparently 
2025-11-04T19:21:05 < jbo> ye
2025-11-04T19:21:27 < jbo> well screw this then, direct connection it is
2025-11-04T19:26:44 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-04T19:29:18 < mawk> PhantomWork do you still have your architecture with fake interrupts and critical sections everywhere?
2025-11-04T19:29:26 < mawk> that might slow things down
2025-11-04T19:29:34 < jbo> fake interrupts?
2025-11-04T19:30:05 < mawk> interrupts that do regular operations in them
2025-11-04T19:30:12 < mawk> without caring about timing 
2025-11-04T19:30:19 < jbo> sounds like fun
2025-11-04T19:30:48 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-04T19:31:01 < mawk> you can tick the freertos thing in cubemx to get a rtos PhantomWork 
2025-11-04T19:31:15 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-04T19:31:18 < mawk> or get the stm32_sequencer utility 
2025-11-04T19:31:31 < mawk> the latter gives you coopérative scheduling 
2025-11-04T19:32:08 < mawk> you can trigger events in interrupts and process events in tasks
2025-11-04T19:32:43 -!- _nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-04T19:33:43 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 240 seconds]
2025-11-04T19:34:28 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T19:35:52 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 256 seconds]
2025-11-04T19:40:16 -!- specing [~specing@user/specing] has joined ##stm32
2025-11-04T19:53:44 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 240 seconds]
2025-11-04T19:58:23 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-04T20:01:27 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 260 seconds]
2025-11-04T20:01:56 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-04T20:07:42 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-04T20:10:09 < PhantomWork> mawk: problem is not thing being slow, it's thing breaking.    If I press reset, and "read read read read read read" I get a consistant 8ms read and processing time. I hit "write", then "read read read" and now I get the 423ms, but not at every write, specially since I added the i2c reset... so, it seems like there is a bug somewhere
2025-11-04T20:10:57 < PhantomWork> mawk: for RTOS, I'ld need to learn it... and rewrite the code... which may happen, but for now, I need something that work for today...
2025-11-04T20:24:51 < qyx> PhantomWork: which eeprom is that
2025-11-04T20:26:10 < PhantomWork> M24C02
2025-11-04T20:43:28 < qyx> the first thing I would check is read and write/erase cycle time in the datasheet
2025-11-04T20:43:41 < qyx> the second one if you are not trying to communicate with the eeprom when it is busy
2025-11-04T20:44:12 < qyx> that would explain why the first read after a write fails, or takes longer for whatever reason (failed transaction, busy, etc.)
2025-11-04T20:44:30 < qyx> then, if you are debugging a hardware problem, scoping the communication is somewhat mandatory
2025-11-04T20:44:48 < qyx> and knowing what your hal does is definitely helpful
2025-11-04T20:47:56 < qyx> the datasheet confirms the memory doesn't do ACK when being busy
2025-11-04T20:48:18 < qyx> so if your read immediately follows a write transaction, it will fail as if the device doesn't respond
2025-11-04T20:53:06 < zyp> 14:36:09 < karlp> are tehre any "modern" micros that have built in h-bridges?
2025-11-04T20:53:17 < zyp> define what kind of h-bridge you're looking for
2025-11-04T20:54:05 < zyp> every push-pull gpio is a halfbridge, so for mA currents, any pair of GPIOs got you covered
2025-11-04T20:54:57 < zyp> for beefier stuff, I don't know how common it is to have the entire bridge built in, but there's plenty of stuff like stspin where gate drivers are
2025-11-04T20:56:02 < karlp> zyp: yeah, just needed "a bit more"
2025-11-04T20:56:15 < karlp> looking at christmas lights where ~80-100mA on batteries.
2025-11-04T20:56:33 < karlp> esp32 + half of a drv8833 works ok in quick test.
2025-11-04T20:57:07 < karlp> I sitll don't really get how the actual part got ~5.7Vp2p from the 4.5V supply though.
2025-11-04T21:01:37 < PhantomWork> qyx: my write function does have a ack wait
2025-11-04T21:02:54 < PhantomWork> .... or should....   need to check something....  but soldering stuff first...
2025-11-04T21:06:57 < PhantomWork> but you might be on something here. I noticed a corruption on the last page that took a few write to clear, I assumed a bug that I just fixed, but... maybe it don't wait enough and cause things to crash...
2025-11-04T21:07:14 < PhantomWork> my write function issue a read right after
2025-11-04T21:16:19 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-04T21:21:55 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T21:22:28 < mawk> it's really not that hard PhantomWork 
2025-11-04T21:22:34 < mawk> you can start with just one task
2025-11-04T21:22:40 < mawk> cube generates all the code for you 
2025-11-04T21:22:58 < mawk> just don't use malloc and don't use printf to print floats and you will have no surprises
2025-11-04T21:23:39 < PhantomWork> I don't use malloc, and printf for float was broken so I just avoided it
2025-11-04T21:24:12 < mawk> what is taking 500ms exactly 
2025-11-04T21:24:15 < mawk> how do you time it
2025-11-04T21:24:27 < mawk> you send an I²C command and it takes half a second to respond?
2025-11-04T21:24:32 -!- teknix_ [~unknown@user/hsv] has quit [Read error: Connection reset by peer]
2025-11-04T21:26:31 < PhantomWork> mawk: https://bpa.st/X72Q   simplified
2025-11-04T21:26:35 < mawk> why do you need a H bridge for that karlp 
2025-11-04T21:26:37 < mawk> it's AC lights?
2025-11-04T21:28:34 < mawk> btw it's called a watchdog refresh not reset, reset is what you want to avoid
2025-11-04T21:28:38 < jpa-> karlp: 4.5V supply driven in two sides with H-bridge is +-4.5V = 9Vp-p
2025-11-04T21:29:37 < mawk> PhantomWork the code doesn't seem correct to me
2025-11-04T21:29:53 < mawk> it should be (uint8_t*)&eepromData[page]
2025-11-04T21:29:54 < mawk> no?
2025-11-04T21:30:32 < mawk> or is it an array of points
2025-11-04T21:30:36 < mawk> pointers*
2025-11-04T21:31:32 < mawk> are the calls to HAL even succeeding? can you check the return codes
2025-11-04T21:35:33 < PhantomWork> mawk: the call does succede
2025-11-04T21:35:55 < PhantomWork> last time I checked it returned "HAL_OK"
2025-11-04T21:36:02 < PhantomWork> and the readout is fine
2025-11-04T21:36:08 < mawk> can you time before processing the data
2025-11-04T21:36:16 < mawk> what about eepromData what's the type
2025-11-04T21:36:31 < mawk> is it uint8_t [16 * EEPROMPAGES] ?
2025-11-04T21:36:56 < PhantomWork> yes
2025-11-04T21:37:06 < PhantomWork> err no
2025-11-04T21:37:27 < PhantomWork> [pages (16)][16]
2025-11-04T21:37:55 < mawk> so uint8_t eepromData[16][16]
2025-11-04T21:38:48 < PhantomWork> yes
2025-11-04T21:39:09 < mawk> and does it happen in an interrupt 
2025-11-04T21:39:44 < mawk> if you spend lots of time in a critical section in another interrupt then the others won't be processed
2025-11-04T21:40:24 < PhantomWork> volatile uint8_t.... [16][16] = {0};  then read
2025-11-04T21:40:39 < mawk> which stm32 is it again?
2025-11-04T21:40:51 < PhantomWork> 1 more connector to make, and I'll be avail...
2025-11-04T21:40:57 < mawk> did you try timing just after the read instead of after the bunch of processing 
2025-11-04T21:40:58 < PhantomWork> stm32f103
2025-11-04T21:41:08 < mawk> the dreaded bluepill
2025-11-04T21:43:13 < PhantomWork> https://imgur.com/a/tHBcasu
2025-11-04T21:43:25 < PhantomWork> that's the board... previous revision
2025-11-04T21:43:55 < PhantomWork> so not the bluepill
2025-11-04T21:50:28 < PhantomWork> And done
2025-11-04T21:51:38 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-04T22:01:06 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T22:09:47 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-04T22:10:57 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 260 seconds]
2025-11-04T22:11:17 < qyx> the board, what
2025-11-04T22:11:56 < PhantomWork> mawk: https://bpa.st/3YCA   the data readout from the eeprom, from my debug interface
2025-11-04T22:12:08 < PhantomWork> qyx: not a bluepill board
2025-11-04T22:12:51 < karlp> mawk: yeah, christmaslights / halloween lights have two strings of leds in reverse.
2025-11-04T22:12:55 -!- jerrycash2 [~jerrycash@user/jerrycash] has quit [Ping timeout: 264 seconds]
2025-11-04T22:13:16 < karlp> jpa-: I... can't see it.
2025-11-04T22:13:21 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T22:13:29  * karlp simulates
2025-11-04T22:17:50 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 245 seconds]
2025-11-04T22:17:55 < qyx> www wat
2025-11-04T22:18:00 < qyx> two strigs in reverse?
2025-11-04T22:18:34 < zyp> different colors?
2025-11-04T22:18:56 < qyx> or blinky shit
2025-11-04T22:19:04 < qyx> but I have never seen such thing
2025-11-04T22:19:09 < qyx> very interesting
2025-11-04T22:20:05 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T22:31:13 < karlp> even same colours.  leds are like ABABABABABA down the line, one polarity lights all the A leds, other polarity lighrts all the B leds, so you can https://bin.jvnv.net/file/3z5ir/20251103_212616.png and blink A leds threee times, then B leds 3 times...
2025-11-04T22:31:36 < karlp> I'm just staring at falstad and trying to figure out how I get 10V :) feeling very dumb
2025-11-04T22:31:43 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 240 seconds]
2025-11-04T22:34:25 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T22:34:59 < karlp> well.  https://tinyurl.com/2xhflemu
2025-11-04T22:35:09 < karlp> meh.  5v or 0 v, but still 10 v. 
2025-11-04T22:35:16 < karlp> time for a step away frrom the computer.
2025-11-04T22:36:07 < zyp> 10V where?
2025-11-04T22:36:15 < qyx> speaking of christmas lights
2025-11-04T22:36:33 < karlp> the peak to peak output is ~9.8V, not 5.
2025-11-04T22:36:35 < zyp> are you counting the 5V twice? since 5--5 = 10?
2025-11-04T22:36:37 < qyx> I have a classic led one, no reverse, powered by a 24 V SELV supply
2025-11-04T22:36:52 < karlp> I don't get in my head yet how switchign polarity of 5V gives me a 10V peak to peak.
2025-11-04T22:37:22 < qyx> some leds stopped working, some are *blinking*, some are dim, some super bright
2025-11-04T22:37:24 < zyp> because 5 - (-5) = 10
2025-11-04T22:37:34 < karlp> but like, in the real circuit, I have 4.2V battery, and it shows 5.7 or so across the output strips.
2025-11-04T22:37:54 < mawk> that's the average peak the peak right 
2025-11-04T22:37:59 < mawk> so it makes sense
2025-11-04T22:38:05 < karlp> yeah, but it has modes where it has only half of it.
2025-11-04T22:38:18 < zyp> what freq and what multimeter mode?
2025-11-04T22:38:34 < karlp> zyp: steady on state, dc voltage.
2025-11-04T22:38:38 < qyx> if there is a h-bridge at the output, maybe a doubler?
2025-11-04T22:38:52 < qyx> for.. some reason?
2025-11-04T22:39:02 < karlp> right, but even this guy https://wyliecoyoteuk.wordpress.com/2025/10/29/a-solution-to-annoying-led-string-flashing-part-2/ is just like "yeah, reverse polarity, 10V peak to peak"
2025-11-04T22:39:25 < zyp> peak to peak is a stupid way to measure
2025-11-04T22:39:31 < karlp> and that falstad shows 10v peak to peak as well, but I think this isn't real voltage it's just becuase it's comparing it to ground.
2025-11-04T22:39:35 < karlp> when ground is flipping
2025-11-04T22:40:13 < zyp> when you're measuring across the load, the reference point is moving
2025-11-04T22:41:05 < zyp> one half cycle it's measuring 5V relative to 0V
2025-11-04T22:41:14 < qyx> if it is treated as AC wave, simpky draw it and you see 10 Vpp
2025-11-04T22:41:20 < zyp> the other half cycle, it's measuring 0V relative to 5V
2025-11-04T22:41:24 < qyx> yes
2025-11-04T22:41:30 < qyx> but still only 5 V
2025-11-04T22:42:16 < zyp> yep, all absolute voltages are within the 0..5V range
2025-11-04T22:42:26 < zyp> but the relative voltage between two points can be twice that
2025-11-04T22:42:39 < zyp> I mean
2025-11-04T22:42:52 < zyp> the *difference* in relative voltage between two points can be twice that
2025-11-04T22:42:58 < mawk> not without filtering right 
2025-11-04T22:42:59 < mawk> ah
2025-11-04T22:44:05 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-04T22:45:03 < karlp> ok, but iwth a DMM across the output, that's referenced to one side only, so it should only show 5 or -5.
2025-11-04T22:45:13 < karlp> or "supply" or "minus supply"
2025-11-04T22:45:18 < karlp> not "bigger than supply"
2025-11-04T22:45:37 < zyp> no
2025-11-04T22:45:54 < karlp> unless there's something else in the actual christmas light IC itself (which could well be, to get high enough voltage for all types of leds)
2025-11-04T22:45:56 < qyx> I would say yes
2025-11-04T22:45:59 < mawk> karlp you put 5V amplitude in falstad
2025-11-04T22:46:02 < mawk> instead of 2.5V
2025-11-04T22:46:08 < mawk> for the square wave sources
2025-11-04T22:46:18 < zyp> karlp, with a DMM across the output, the reference is *moving*
2025-11-04T22:46:29 < mawk> it should be 2.5V DC component and 2.5V amplitude to get from 0V to 5V
2025-11-04T22:46:37 < zyp> the reference is your black probe, sometimes it's at 0V and sometimes it's at 5V
2025-11-04T22:46:39 < qyx> if the dmm is isolated properly it should not matter
2025-11-04T22:46:57 < karlp> yeah, but 0V on black, then 0V on read, should just show either +5 or -5.
2025-11-04T22:47:00 < karlp> not 9.
2025-11-04T22:47:14 < karlp> _iff it is just a plain h-bridge from the supply_
2025-11-04T22:47:16 < qyx> it simply can't measure voltage greater than the source is
2025-11-04T22:47:20 < zyp> karlp, yep
2025-11-04T22:47:25 < qyx> whatever happens
2025-11-04T22:47:43 < qyx> unless here is some sort of charge pump, doubler or other capacitor/diode shit
2025-11-04T22:47:54 < zyp> but that assumes your DMM is measuring correctly on a changing signal
2025-11-04T22:48:00 < karlp> yeah, so I have a) getting very confused by this web dude drawing peak to peak compared to _gound_ which is irrelevant reference point, and _also_ the actual chinesium IC having somethign that boosts
2025-11-04T22:48:13 < karlp> zyp: it's not changing in this point, I can select modes that ar ejsut steady on.
2025-11-04T22:48:29 < karlp> either A string or B string.
2025-11-04T22:48:41 < zyp> so do that, then use DMM to measure either side against ground?
2025-11-04T22:48:54 < zyp> instead of both sides against each other
2025-11-04T22:49:21 < zyp> if both sides are in the 0..5V range, your DMM is lying to you
2025-11-04T22:49:35 < zyp> and if they're not, there's your issue
2025-11-04T22:49:46 < qyx> or just get a scope, get a single ground, connect one probe across the i put and two probes on the outputs
2025-11-04T22:49:49 < qyx> and see
2025-11-04T22:49:57 < zyp> (or that)
2025-11-04T22:50:04 < qyx> the waveform on the output cannot be higher than the input
2025-11-04T22:50:41 < zyp> if the H-bridge is bootstrapped, theoretically the bootstraps could be leaking
2025-11-04T22:50:55 < qyx> and the dmm would need some RF voodoo or capacitive coupling or whatever to not float properly
2025-11-04T22:51:10 < qyx> it should just work
2025-11-04T22:51:29 < zyp> I don't think it's a matter of floating, if anything it's a matter of it not being pure DC and getting confused
2025-11-04T22:52:29 < qyx> wouldn't it be confused by showing less?
2025-11-04T22:53:01 < karlp> yeah, this is just make believe voltag because of looking t shit on the scope that was referenced to the battery negative
2025-11-04T22:53:46 < karlp> ok, I am not truly insane then. I could not figure out how you got extra volts.
2025-11-04T22:54:14 < qyx> how could an icelander not be insane
2025-11-04T22:54:53 < karlp> so, an esp32-c3pico and a cheap sot23 h bridge will fit in the case, for about 500 times the unit price :)
2025-11-04T22:55:11 < karlp> a china btle soc would be cheaper, but so much grosser to work with.
2025-11-04T22:55:50 -!- _nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Quit: Segmentation fault]
2025-11-04T22:55:58 < qyx> what sot23 h bridge
2025-11-04T22:56:05 < qyx> I need a 500 mA one
2025-11-04T22:57:02 < qyx> I don't like drv8601
2025-11-04T22:57:17 < qyx> it looks like a differential opamp
2025-11-04T22:57:24 < karlp> qyx: https://www.lcsc.com/category/1016.html and sort by availability/price
2025-11-04T22:57:33 < karlp> tc1118 looks like it should be fine.
2025-11-04T22:57:38 < qyx> nah lcsc
2025-11-04T22:57:53 < karlp> what's wrong with drv8601?
2025-11-04T22:58:25 < qyx> 30 kHz only
2025-11-04T22:59:19 < karlp> eh, more than enough for christmas lights :)
2025-11-04T23:00:48 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T23:01:02 < mawk> the falstad "voltemeter" doesn't care about ground, it uses the two points you give it; but the Vpp value is computed on the signal over time, so if the frequency is high enough that a full period is in the scope view it will tell you 10V
2025-11-04T23:01:04 < mawk> but it doesn't correspond to anything measurable at a given instant
2025-11-04T23:01:31 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-04T23:02:19 -!- teknix [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-04T23:03:20 < qyx> MP6513 looks more suitable
2025-11-04T23:04:01 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 246 seconds]
2025-11-04T23:07:03 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-04T23:19:21 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-04T23:23:14 -!- Livio_ [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-04T23:36:31 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 244 seconds]
2025-11-04T23:40:23 -!- Netsplit *.net <-> *.split quits: mrec_, soweli_iki, Sadale, zChris
2025-11-04T23:41:45 -!- Netsplit over, joins: zChris, soweli_iki, Sadale, mrec_
2025-11-04T23:44:32 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 240 seconds]
2025-11-04T23:47:30 -!- noarb [~noarb@user/noarb] has joined ##stm32
--- Day changed Wed Nov 05 2025
2025-11-05T00:01:51 -!- teknix [~unknown@user/hsv] has quit [Ping timeout: 250 seconds]
2025-11-05T00:05:21 < karlp> lol, I found a mp6513 on lcsc and was like.... wot?
2025-11-05T00:05:24 < karlp> it's not the same.
2025-11-05T00:05:27 < karlp> cmp6513 :)
2025-11-05T00:18:54 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-05T00:26:47 -!- teknix [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-05T01:09:35 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-05T01:10:00 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-05T01:12:13 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 255 seconds]
2025-11-05T01:15:00 < mawk> the H7 ADC can do 1.2Msps maximum with my package, but if I trigger it with a timer they say there will be some jitter around the trigger time
2025-11-05T01:15:03 < mawk> which doesn't sound good
2025-11-05T01:15:27 < mawk> I guess I can set it back down to 1Msps and give some extra nanoseconds for the clocks to synchronize
2025-11-05T01:16:00 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-05T01:17:04 < mawk> the reference manual is not extremely clear on how to remove that jitter, apparently adc_sclk must be twice the AHB clock for the synchronization to be perfect
2025-11-05T01:17:22 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-05T01:19:06 < mawk> so I set it up to take 917ns for each ADC conversion, and I trigger an acquisition every 1000ns, and then DMA is moving it around
2025-11-05T01:20:20 < mawk> I could skip the timer and set the ADC in continuous mode and have DMA copy the data and then it would be triggered exactly every 917ns right? maybe that's better
2025-11-05T01:20:24 < mawk> then I could even reach the 1.2Msps
2025-11-05T02:01:59 -!- drew [~drew@user/drew] has joined ##stm32
2025-11-05T02:10:08 -!- BrainDamage [~m-t6k752@user/BrainDamage] has quit [Ping timeout: 240 seconds]
2025-11-05T02:10:43 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-05T02:17:31 -!- BrainDamage [~m-t6k752@user/BrainDamage] has joined ##stm32
2025-11-05T03:44:02 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-05T03:51:38 -!- c10ud__ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-05T04:01:39 -!- Netsplit *.net <-> *.split quits: c10ud_, karlp, octorian
2025-11-05T04:07:13 -!- Netsplit over, joins: octorian, karlp
2025-11-05T06:59:41 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-05T07:15:28 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-05T07:28:40 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-05T08:05:22 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-05T08:23:45 < jpa-> karlp: i think you may be confusing between Vpk and Vpp :)
2025-11-05T09:09:18 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-05T09:29:29 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-05T09:35:39 -!- teknix [~unknown@user/hsv] has quit [Quit: leaving]
2025-11-05T09:57:43 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-05T10:05:28 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-05T10:06:19 -!- kurfen [~kurfen@45.84.137.25] has quit [Ping timeout: 264 seconds]
2025-11-05T10:12:38 -!- zChris [~zChris@user/zchris] has quit [Ping timeout: 244 seconds]
2025-11-05T10:20:48 -!- zChris [~zChris@user/zchris] has joined ##stm32
2025-11-05T11:01:59 -!- kurfen [~kurfen@45.84.137.24] has joined ##stm32
2025-11-05T11:13:36 -!- kurfen [~kurfen@45.84.137.24] has quit [Ping timeout: 252 seconds]
2025-11-05T11:14:03 < qyx> let's listen to some proper swiss metal and do some work
2025-11-05T11:19:18 -!- kurfen [~kurfen@45.84.137.25] has joined ##stm32
2025-11-05T11:23:57 -!- kurfen_ [~kurfen@45.84.137.24] has joined ##stm32
2025-11-05T11:24:37 -!- kurfen [~kurfen@45.84.137.25] has quit [Ping timeout: 264 seconds]
2025-11-05T11:37:48 -!- kurfen_ [~kurfen@45.84.137.24] has quit [Ping timeout: 252 seconds]
2025-11-05T11:41:21 < karlp> ok, so stack smashing protection is fun.
2025-11-05T11:41:29 < karlp> I hadn't clicked that it only triggers on function exit
2025-11-05T11:41:56 < karlp> so I was staring at the array memcpy that is at the end of the function, right where it stops, couldnt' see anything wrong with it.
2025-11-05T11:42:09 < karlp> finally realised it was just "somewhere in the function"
2025-11-05T11:42:13 < karlp> and found the nice bug :)
2025-11-05T11:44:25 < ventYl> yeah, the whole point is to prevent manipulating function return address
2025-11-05T11:45:28 < qyx> so what does it do, saves a copy on enter, checks on return?
2025-11-05T11:45:35 < qyx> a double-buffered stack?
2025-11-05T11:45:49 < ventYl> usually some kind of canary
2025-11-05T11:46:19 < ventYl> it is placed in front of return address and checked before returning from the function. if it doesn't match, the function calls abort instead of returning
2025-11-05T12:25:33 < mawk> yeah
2025-11-05T12:25:35 < mawk> just a single special word
2025-11-05T12:25:37 < mawk> for instance 
2025-11-05T12:26:55 < mawk> so how do I get maximum screaming fast speeds with the ADC, is continuous acquisition timed perfectly and I can rely on that ? each acquisition at 16 bits depths with the minimum sample time takes 10 ticks of the ADC clock, so I should get a new value each 10 ticks?
2025-11-05T12:47:21 < jpa-> continuous acquisition is timed perfectly
2025-11-05T12:59:04 < mercenary> trust, but verify. sample a triangle wave, look for discontinuities.
2025-11-05T13:08:21 < jpa-> and monitor the overrun status bits :)
2025-11-05T13:14:26 < zyp> karlp, which stack smashing protection? function return hooks, or the armv8m hardware checking?
2025-11-05T13:15:44 < zyp> my understanding was that the latter should check all sp-relative accesses
2025-11-05T13:16:20 < jpa-> does the hardware checking do more than just overflow protection?
2025-11-05T13:17:52 < zyp> oh, I'm mixing up
2025-11-05T13:18:06 < zyp> overflowing stack != overflowing stack frame
2025-11-05T13:18:08 < qyx> zyp the heterodyne
2025-11-05T13:20:22 < ventYl> armv8m hardware checking is just a convenient way how to save one memory protection region and still have stack protected
2025-11-05T13:20:40 < karlp> zyp: the gcc built in ones, -fstack-protector-strong.
2025-11-05T13:21:04 < karlp> it's built in to esp-idf, I'd just never fucked up and been caught with it before, so wasn't good at interpreting it's aborts.
2025-11-05T13:21:39 < karlp> also, my webbluetooth now works :)
2025-11-05T13:21:49 < karlp> now "just" more web design to make it prettier
2025-11-05T13:23:14 < ventYl> i'm being too paranoid for my own bad
2025-11-05T13:23:45 < ventYl> input checks "interface ID" parameter. if it is wrong, it returns error "interface out of range" copying interface ID in the response
2025-11-05T13:24:22 < ventYl> response processing code again checks "interface ID" of the response and as "out of range" interface ID is out of range.... it drops the response
2025-11-05T13:25:34 < zyp> ventYl, the issue with using MPU for stack protection is that you need a sort of guard area below each stack, and just restricting threads from touching each others stacks would be quite limiting
2025-11-05T13:26:08 < qyx> of yourse you protect each individual stack
2025-11-05T13:26:50 < zyp> allocating a buffer on stack and passing it into an API that results in it being filled by another thread isn't that uncommon
2025-11-05T13:27:14 < zyp> and that breaks down if the other thread can't access the stack the buffer was allocated on
2025-11-05T13:27:22 < ventYl> zyp: why do you need a guard area? MPU gives you hard limits?
2025-11-05T13:28:08 < zyp> AFAIK MPU can't distinguish between stack accesses and other accesses
2025-11-05T13:28:25 < ventYl> right. it cant.
2025-11-05T13:28:30 < ventYl> and there's nothing wrong at it
2025-11-05T13:28:45 < zyp> so to have any protection at all, you need an inaccessible area between accessible stack and other accessible areas
2025-11-05T13:29:16 < qyx> whats the point in protecting other areas if you already trashed your stack?
2025-11-05T13:29:49 < zyp> isolating the trash to a single thread?
2025-11-05T13:29:49 < qyx> ok you may miss a problem if you just jumb between stack and data/bss
2025-11-05T13:31:09 < ventYl> zyp: CMRX uses one MPU region for thread stack and other for data/bss. stack is always thread-private
2025-11-05T13:35:38 < ventYl> if you want to share stack-resident variable, you have to perform RPC call
2025-11-05T13:35:56 < jpa-> ventYl: but where do you allocate stacks from? what if you first run malloc() for a shared buffer, and then start a new thread that should access that buffer?
2025-11-05T13:36:33 < ventYl> jpa-: that's simple. you don't run malloc
2025-11-05T13:37:24 < jpa-> well what if you type     uint8_t my_shared_buffer[2048]; thread_t my_thread(1024); (or whatever syntax you have)
2025-11-05T13:37:39 < ventYl> yet, you can mark any variable as sharable using SHARED macro, then you can share this variable along the RPC call too
2025-11-05T13:38:09 < ventYl> it happens automagically
2025-11-05T13:39:00 < jpa-> that doesn't really explain anything
2025-11-05T13:39:17 < jpa-> either you have inaccessible area between stack and whatever is before that, or you don't
2025-11-05T13:39:35 < qyx> the problem with that is the developer's reasoning
2025-11-05T13:39:42 < ventYl> no, there's no inaccessible area
2025-11-05T13:39:54 < ventYl> https://talks.openalt.cz/media/openalt-2025/submissions/A8B9LP/resources/CMRX_talk_xdO1cLM.pdf
2025-11-05T13:39:59 < ventYl> ^^ page 10 - 12
2025-11-05T13:40:00 < jpa-> (that said, reserving e.g. 16 bytes per stack for that is fine; but on the other hand many stack overflows happen when someone does  char filename[256];  and then you need a much longer red area)
2025-11-05T13:40:05 < qyx> if you accept using a conceptually different approach to things you simply don't do such things like sharing a stack-allocated buffer
2025-11-05T13:41:02 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-05T13:41:10 < jpa-> ventYl: even if the "inaccessible area" is "stacks of other threads", it is still inaccessible and you need to verify nothing the thread *wants* to access gets put there
2025-11-05T13:41:59 < jpa-> qyx: but the problem exist for non-stack-allocated buffers too, unless the allocation system has something special in it
2025-11-05T13:42:11 < karlp> something somethign rust
2025-11-05T13:42:31 < ventYl> jpa-: well, this boils down to "don't share stack-allocated variables outside current thread, it won't work"
2025-11-05T13:42:50 < jpa-> ventYl: there can't always be stack before stack, there is something else before the first stack
2025-11-05T13:43:41 < ventYl> jpa-: ok and what's the problem with that?
2025-11-05T13:43:49 < jpa-> (tricks like allocating stacks always in beginning of RAM, or in a separate section like DTCM work when they suit the platform)
2025-11-05T13:44:06 < jpa-> ventYl: then you need to make sure the "stuff before first thread's stack" is not stuff that the first thread needs to access
2025-11-05T13:44:17 < ventYl> why?
2025-11-05T13:44:38 < jpa-> otherwise you need to have the first thread's MPU configuration allow access there, and then you no longer have stack overflow protection
2025-11-05T13:44:55 < ventYl> ah, ok
2025-11-05T13:45:19 < ventYl> now I understand what specific situation you are targeting
2025-11-05T13:45:49 < jpa-> but i admit MPU stack protection handles 90% of cases just fine
2025-11-05T13:46:26 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-05T13:47:20 < ventYl> while in theory this is a plausible scenario, in practice it is like 99,9999% impossible with more than 1 thread
2025-11-05T13:47:41 < ventYl> * more than 1 process
2025-11-05T13:48:29 -!- rkta [~rkta@user/rkta] has quit [Quit: sysupgrade]
2025-11-05T13:49:38 < jpa-> what makes it so impossible? it seems to me perfectly normal to have    my_adc_task.c    with     uint16_t adc_buffer[1024]; uint32_t thread_stack[...];  and they get allocated right after each other unless you specifically prepare for the situation with e.g. linker sections
2025-11-05T13:50:21 < ventYl> exactly the fact that automagic nature of CMRX lies in automatic preparation of linker sections :)
2025-11-05T13:50:50 < jpa-> that pdf you linked doesn't really explain how it works, can you elaborate?
2025-11-05T13:51:07 < jpa-> do you mean it collects stacks in separate section, or something fancier?
2025-11-05T13:51:25 < jpa-> you seem to put a lot of misplaced trust on this "automagic" :)
2025-11-05T13:51:30 < ventYl> that PDF are just slides for talk, explanation was verbal
2025-11-05T13:52:41 < ventYl> TL;DR version: a static library is considered a "process" if it is correctly decorated (that one OS_APPLICATION line + some CMake stuff)
2025-11-05T13:53:11 < ventYl> then all .data and .bss from within that static library is collected and properly aligned so it can be covered by MPU region
2025-11-05T13:54:01 < ventYl> threads are associated to processes and based on this they are granted access to .data and .bss
2025-11-05T13:54:20 < ventYl> stacks are allocated by kernel from memory which lies outside this area
2025-11-05T13:54:46 < ventYl> plausibility of this scenario is based on the fact that stacks also have quite large alignment, so they will be put at the beginning of kernel memory
2025-11-05T13:55:05 < ventYl> thus, there's non-zero possibility that last process' .bss will touch first allocatable stack
2025-11-05T13:55:08 < mawk> it's weird how the smaller packages of the H7 have a worse ADC
2025-11-05T13:55:20 < mawk> I guess they do binning and reserve the best for BGA maybe?
2025-11-05T13:55:39 < qyx> mawk: no, package limitations
2025-11-05T13:55:41 < ventYl> yet, if you have at least two processes, then first allocated stack will belong to process #1 and last allocated .bss will belong to process !#1
2025-11-05T13:56:13 < jpa-> ventYl: ah, ok; i agree that section arrangement makes it less likely
2025-11-05T13:56:27 < ventYl> moreover, the current stack allocation process sucks a lot as all stacks have to have same size
2025-11-05T13:56:32 < ventYl> so I have to add some kind of page allocator
2025-11-05T13:56:40 < ventYl> which will push stacks after all the kernel memory
2025-11-05T13:56:53 < mawk> like smaller buffers and higher analog path resistance in the smaller packages?
2025-11-05T13:57:08 < jpa-> so you allocate stacks dynamically but .data and .bss statically? why?
2025-11-05T13:57:38 < jpa-> mawk: yeah, exactly
2025-11-05T13:57:41 < ventYl> because it is possible to create thread via syscall, but it is not possible to create a process via syscall (yet)
2025-11-05T13:57:59 < jpa-> mawk: LQFP has long paths not only for the signal, but also for the ADC vref and vdd bypass caps
2025-11-05T13:58:21 < mawk> I see
2025-11-05T14:15:00 < mawk> I can lower the bit depth to get a higher sample rate but then I need to amplify the input 
2025-11-05T14:15:35 < mawk> I have 2mV/mA in input, but the signal can go up to like 3V at some point
2025-11-05T14:16:23 < mawk> maybe I can acquire both the scaled and unscaled signals concurrently on 2 ADCs, and select the one which isn't clipping 
2025-11-05T14:17:34 < zyp> 12:35:38 < ventYl> if you want to share stack-resident variable, you have to perform RPC call
2025-11-05T14:17:38 < zyp> that's a pretty big limitation
2025-11-05T14:17:58 < qyx> no, it is a concept
2025-11-05T14:18:10 < qyx> part of my concept is I don't have bss/data for services
2025-11-05T14:19:22 < qyx> which makes future work easier because every service is contained (it has an instance)
2025-11-05T14:22:27 < qyx> why the fuck is kicad now considering footprint centers snap points which cannot be turned off
2025-11-05T14:22:43 < qyx> it makes doing filled areas firtually impossible
2025-11-05T14:25:29 < ventYl> zyp: that's just a subtle hint that sharing a stack-allocated variable is not a wise idea
2025-11-05T14:26:15 < zyp> I'm inclined to disagree :)
2025-11-05T14:28:42 < zyp> that said, I've mostly got coros that work in that manner, so I have a coro-frame allocating stuff, passing a reference to another stack and waiting for the other thread to complete whatever
2025-11-05T14:28:43 < ventYl> well, either it is in some nested stack frame and then it can go out of scope at any time. then it is definitely not wise to share it outside of current call tree
2025-11-05T14:29:13 < zyp> another thread*
2025-11-05T14:29:15 < ventYl> or it was allocated in top-level frame and then, for all practical reasons it is functionally equal to static allocation
2025-11-05T14:30:47 < zyp> > … can go out of scope at any time.
2025-11-05T14:31:01 < zyp> that's why rust references comes with explicit lifetimes
2025-11-05T14:32:35 < ventYl> yeah, but borrow checker isn't and can't work across threads
2025-11-05T14:34:23 < PhantomWork> Do you guys have a trick to debug this: I have a DMA based circular buffer for the uart, I have an off by one error OR race condition OR electrical issue OR windows drop some data (using simply serial and putty). My simplification of the code seems to show that there is no off by one, my analysis seems to show no race possible, electrical I don't believe so, and I have a hard time believing that windows drop data?   so I'm at loss...
2025-11-05T14:37:17 < mercenary> What is the interrupt function called? Always remember, the two most difficult problems are cache coherency, naming things, and off-by-one errors.
2025-11-05T14:39:02 < zyp> ventYl, sure it can
2025-11-05T14:40:40 < mercenary> Simple debugging technique: add a few global counters. Increment each in different places in the code. After a run, see which one(s) lag behind by one.
2025-11-05T14:41:59 < qyx> I would just use printf
2025-11-05T14:46:37 < mawk> you can setup rtt pretty easily to get text debugging while you check your uart
2025-11-05T14:46:47 < mawk> and it won't affect the timing
2025-11-05T14:48:08 < jbo> hello Steffanx 
2025-11-05T14:49:00 < zyp> ventYl, I've made a wrapper for zephyr socket services, SocketService::register() takes a callback with explicit lifetime 'a, and returns a SocketRegistration with the same lifetime: https://paste.jvnv.net/view/n3BA9
2025-11-05T14:49:01 < ventYl> zyp: without Arc?
2025-11-05T14:49:37 < zyp> this enforces that the callback closure and anything it references must remain valid as long as the registration exists
2025-11-05T14:50:36 < karlp> mawk: the printf into the rtt can definintely affect timing ;)
2025-11-05T14:50:37 < zyp> and SocketRegistration::drop() enforces that the callback gets unregistered before it's dropped
2025-11-05T14:51:05 < Steffanx> Hello Mr jbo 
2025-11-05T14:51:57 < zyp> ventYl, Arc is just one way to safely share data between threads
2025-11-05T14:52:23 < mawk> well yeah karlp but that's the monstrous newlib printf's fault
2025-11-05T14:52:33 < mawk> there is a printf provided by segger that should be lighter
2025-11-05T14:55:29 < karlp> I'm pretty sure these gaps are actually caused by the rtt buffer readout blocking things too, but I've not 100% confirmed that yet.  https://bin.jvnv.net/file/C6baH.png
2025-11-05T14:55:44 < karlp> if I turn off the rtt prints, and just check the timestamps for gaps, they never happen.
2025-11-05T15:00:41 < qyx> whats that
2025-11-05T15:04:50 < qyx> zyp: A4/A3 on a bga qspi/hyper flash is NC? from top of your head?
2025-11-05T15:05:00 < qyx> how do you usually route the bypass cap
2025-11-05T15:05:06 < zyp> jbo would know
2025-11-05T15:05:15 < qyx> jbo: how do you do your cap
2025-11-05T15:05:34 < zyp> decoupling? on orbtrace I just put it on the back right where the vias came down
2025-11-05T15:05:42 < qyx> yes
2025-11-05T15:05:48 < zyp> orbtrace were double sided assembly anyway
2025-11-05T15:06:30 -!- rkta [~rkta@user/rkta] has joined ##stm32
2025-11-05T15:07:00 < zyp> if you're doing single sided assembly with BGA, just dump power and ground directly into inner planes and sprinkle decoupling around the BGA
2025-11-05T15:07:25 < zyp> it's generally a bad idea to route to cap first and then into plane
2025-11-05T15:07:40 < jbo> qyx, mine is double sided so 0402 directly on the corresponding pads
2025-11-05T15:08:18 < jbo> incidently, the H7 board I am currently doing I am considering doing caps on the same side as the BGA because the PCB doesn't require double sided assembly otherwise :/
2025-11-05T15:08:37 < qyx> yes that's exactly my though
2025-11-05T15:08:40 < qyx> t
2025-11-05T15:09:07 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-05T15:09:29 < jbo> this is the easy solution: https://bin.jvnv.net/file/kYQo0
2025-11-05T15:09:40 < jbo> but don't really feel like doing double sided on the H7 board just for that.
2025-11-05T15:11:11 < jbo> qyx, regarding your A3/A4 question: those are only needed for hyperflash, not for hyperram
2025-11-05T15:12:41 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-05T15:14:30 < qyx> oh yes lazy
2025-11-05T15:17:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-05T15:20:17 < jbo> qyx, did you check the AN regarding layout? they recommend some guard traces. I decided to not care.
2025-11-05T15:20:40 < jbo> AN211622
2025-11-05T15:25:04 < qyx> I am using qspi only
2025-11-05T15:25:18 < qyx> for now
2025-11-05T15:25:31 < qyx> but I'll check, I plan to do U5+hyper whatever
2025-11-05T15:25:33 < jbo> you could get a different package then to not deal with the decoupling capacitor issue
2025-11-05T15:27:05 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-05T15:29:42 < mawk> does BDMA on the H7 not have access to the regular RAM? I get a transfer error interrupt when I start the whole thing
2025-11-05T15:30:43 < zyp> isn't the regular RAM aliased at multiple addrs? maybe DMA can only use one of the aliases
2025-11-05T15:32:11 < mawk> maybe it can only access SRAM4 which is in the domain D3
2025-11-05T15:32:41 < mawk> well the regular DMA could access it fine, but now I have to use the basic DMA "BDMA" and with all the same settings I get the transfer error interrupt
2025-11-05T15:33:33 < mawk> because only ADC3 has access to the direct analog channels to maximize the sample rate, and ADC3 only talks to BDMA
2025-11-05T15:33:52 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-05T15:33:59 < zyp> which H7 is this?
2025-11-05T15:34:03 < mawk> apparently direct channels and fast channels have the same sample rate at 16 bits resolution so I guess I'll just revert to the normal ADC and normal DMA
2025-11-05T15:34:06 < mawk> H743ZI
2025-11-05T15:34:20 < mawk> description of ADC speeds: https://www.st.com/resource/en/application_note/an5354-getting-started-with-the-stm32h7-series-mcu-16bit-adc-stmicroelectronics.pdf
2025-11-05T15:34:21 < mawk> RM: https://www.st.com/resource/en/reference_manual/rm0433-stm32h742-stm32h743753-and-stm32h750-value-line-advanced-armbased-32bit-mcus-stmicroelectronics.pdf
2025-11-05T15:35:00 < zyp> yeah, page 104 says BDMA can only access SRAM4
2025-11-05T15:36:01 < mawk> ah right, thanks
2025-11-05T15:36:07 < zyp> and page 105 makes it obvious why
2025-11-05T15:37:59 < mawk> I'm not entirely sure what the limitation of BDMA are, apparently it halts the CPU when it does a transfer; but since the ADC runs at 12MHz and the CPU at 480MHz that shouldn't be a problem at all
2025-11-05T15:39:29 < zyp> but I don't see why you have to use BDMA; as far as I can see all DMA engines can access D3 domain, and DMAMUX 1 input 115 is ADC3, ref. page 694
2025-11-05T15:39:54 < mawk> I just let the Cube do it for me and it selected BDMA
2025-11-05T15:40:04 < zyp> don't trust the cube
2025-11-05T15:44:01 < mawk> ah yes much better
2025-11-05T15:44:03 < mawk> thanks
2025-11-05T15:44:54 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-05T15:46:14 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-05T15:48:22 < jbo> does anybody know of readily available microSD cards which support LV (low voltage signaling)?
2025-11-05T15:49:22 < jpa-> mawk: interleaving ADC1 & ADC2 get you faster than ADC3 with direct channels, i think
2025-11-05T15:51:03 < mawk> hmmm indeed
2025-11-05T15:51:38 < mawk> so theoretically I can do 2Msps
2025-11-05T15:55:03 < mawk> direct channel or fast channel is always the same speed for the LQFP144 packages anyway it seems, so it doesn't add anything
2025-11-05T15:55:10 < mawk> for 16 bit resolutionb
2025-11-05T16:08:22 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 256 seconds]
2025-11-05T16:10:28 < mawk> at least the HAL has support for it, but it's hidden in the adc_ex.c files and cubemx doesn't know about it
2025-11-05T16:14:34 -!- specing [~specing@user/specing] has quit [Ping timeout: 260 seconds]
2025-11-05T16:37:03 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-05T16:43:03 < mawk> the dutch are the jews of the north
2025-11-05T16:43:13 < mawk> dutchness could only be transmitted through the father
2025-11-05T16:43:23 < mawk> just like jewness is transmitted by the mother
2025-11-05T16:46:44 < mawk> why aren't ADC readings automatically referred to Vrefint like on nordic stuff
2025-11-05T16:47:26 < mawk> now I have to ensure VDDA didn't change that much since the start of reading, and it needs to last days
2025-11-05T17:10:17 -!- PhantomWork [~PhantomWo@user/phantom] has quit [Read error: Connection reset by peer]
2025-11-05T17:15:21 < jbo> what are you crafting?
2025-11-05T17:16:35 < mawk> a coulomb counter
2025-11-05T17:17:54 < jbo> with an STM32 ADC?
2025-11-05T17:17:56 < jbo> interesting
2025-11-05T17:18:12 < jbo> can they do diff measurements nowdays?
2025-11-05T17:18:14 < mawk> yes
2025-11-05T17:18:19 < jbo> ah cool
2025-11-05T17:18:25 < mawk> well
2025-11-05T17:18:32 < mawk> it can do differential measurements
2025-11-05T17:18:33 < jbo> pseudo diff I assume
2025-11-05T17:18:41 < mawk> but it wants the two inputs out of phase bla bla bla
2025-11-05T17:18:52 < jbo> but it's a single ADC with an analog mux, right? so your readings are "out of sync"
2025-11-05T17:19:22 < mawk> but I already have a precision difference amplifier to get the current as 2mV/mA referenced to ground so I don't need differential
2025-11-05T17:19:48 < mawk> what do you mean by out of sync?
2025-11-05T17:20:02 < mawk> it has 3 ADCs, and the first 2 can run concurrently, or interleaved
2025-11-05T17:20:03 < jbo> presumably you're taking a measurement across a shunt, right?
2025-11-05T17:20:08 < jbo> ooh, I see.
2025-11-05T17:20:15 < jbo> can you actually sync those up in hardware? 
2025-11-05T17:20:19 < mawk> yeah
2025-11-05T17:20:23 < jbo> that's neat.
2025-11-05T17:20:32 < jbo> I haven't been STM32 in a while. didn't get the memo about the new fancy stuff
2025-11-05T17:20:45 < mawk> but I need only 1 ADC if I assume VDDA didn't change, or if I suspect it might change I can take concurrent measurements of the signal and Vrefint at the same exact time
2025-11-05T17:20:57 < mawk> at half the sample rate
2025-11-05T17:21:06 < mawk> well not half, but slightly lower 
2025-11-05T17:21:23 < mawk> 1Msps instead of 1.2Msps, with the two concurrently running 
2025-11-05T17:22:02 < mawk> if you take the BGA packages and use lower bit depth than 16 bit you can do it way faster
2025-11-05T17:22:12 < jbo> I guess you can just slap a voltage reference on VDDA?
2025-11-05T17:22:15 < jbo> err, VREF
2025-11-05T17:22:30 < mawk> I suppose, it's a nucleo board so I would have to change some solder bridges 
2025-11-05T17:22:47 < jbo> I am/was dealing with a similar thing lately and my solution was just to use an ADS131M04 :p
2025-11-05T17:23:03 < mawk> it's powered by USB and then has a LDO apparently, so maybe it's already stable enough 
2025-11-05T17:23:05 < jbo> money shot: https://bin.jvnv.net/file/AoJOv.png
2025-11-05T17:23:19 < mawk> nice 
2025-11-05T17:24:56 < jbo> so you're mostly doing this as an exercise, right? you know that you can get dedicated coulomb counter chips?
2025-11-05T17:25:26 < mawk> yeah I convinced the boss to get the otii arc thing
2025-11-05T17:25:31 < mawk> but now he has to convince his boss to give the money
2025-11-05T17:25:50 < mawk> in the meantime I'm trying an artisanal solution to get some numbers
2025-11-05T17:25:53 < jbo> is that another joule scope?
2025-11-05T17:26:10 < mawk> it can emulate battery discharge curves too
2025-11-05T17:26:14 < jbo> oooh nice
2025-11-05T17:26:15 < mawk> and probably lots of other expensive stuff
2025-11-05T17:26:43 < jbo> sounds like you could just get a $5 chip on a PMOD or whatever and use that in the meantime?
2025-11-05T17:27:41 < mawk> it's kind of an off the record project, idk if the boss would allow buying new parts right now
2025-11-05T17:27:55 < mawk> I'm trying to do it with parts I already have, and as an exercise as well
2025-11-05T17:28:17 < mawk> the H7 board was taking dust in a drawer
2025-11-05T17:28:20 < jbo> alright, does sound like a nice side gig :)
2025-11-05T17:29:33 < mawk> I'm trying to measure power consumption of my custom low power firmware to convince the boss we can do better than a couple hours on battery
2025-11-05T17:30:23 < mawk> officially we promise 24h on a very new fully charged battery, and as soon as the battery is a bit old we wash our hands of that specification which is a bit dishonest 
2025-11-05T17:30:33 < mawk> so I'm trying to get 24h on an old battery and 48h on a new battery 
2025-11-05T17:30:54 < jbo> I hope they know how to appreciate that
2025-11-05T17:32:22 < mawk> I'd rather work on that than bullshit bug fixes
2025-11-05T17:32:36 < mawk> but if they end up not appreciating it I'll probably find work elsewhere 
2025-11-05T17:33:09 < mawk> it's a medical device so there is theoretically more stringent procedures than regular devices, but it's not like a pacemaker or anything critical and many of the procedures are self-inflicted
2025-11-05T17:33:24 < mawk> and result in us only doing bug fixes in the dumbest way possible
2025-11-05T17:36:51 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-05T17:39:11 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-05T17:40:25 -!- PhantomWork [~PhantomWo@user/phantom] has joined ##stm32
2025-11-05T17:45:40 < jbo> how come manufacturers never follow their own decoupling capacitor guidelines in their eval kits
2025-11-05T17:45:40 < jbo> ffs
2025-11-05T17:52:58 < zyp> «do as I say, not as I do»
2025-11-05T17:53:11 < jbo> sounds like something a parent would say
2025-11-05T18:33:25 < qyx> lol
2025-11-05T18:33:54 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-05T18:33:59 < BrainDamage> parent companies
2025-11-05T18:37:48 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-05T19:18:39 < mawk> just like with real parents, you do as they say and not they do until you're big enough to understand why you can do it
2025-11-05T19:19:24 < zyp> well put
2025-11-05T19:20:18 < PhantomWork> I'm still hunting that uart bug :(
2025-11-05T19:24:50 < mawk> is the baud rate especially low?
2025-11-05T19:42:32 < jpa-> jbo: STM32H7 has true differential ADC
2025-11-05T19:46:51 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-05T19:47:12 < qyx> I don't understand why
2025-11-05T19:47:21 < qyx> I would expect that from a G4
2025-11-05T19:47:49 < jbo> jpa-, cool
2025-11-05T19:48:21 < jpa-> qyx: well G4 has it too?
2025-11-05T19:48:59 < jpa-> but H7 has 16 bits
2025-11-05T19:50:21 < jpa-> mawk: uh, STM32H7 ADC readings are referred to Vref pin, which you can configure to come from Vrefint if you want
2025-11-05T19:51:41 < jpa-> it has the whole VREFBUF block that you can use to select the voltages
2025-11-05T19:52:56 < mawk> right
2025-11-05T19:54:18 < mawk> what do you mean by true differential? the ADC application note seems to want the two differential inputs out of phase
2025-11-05T19:54:57 < jpa-> AN5354? where does it want that?
2025-11-05T19:55:19 < jpa-> ah, page 9, you are just misunderstanding
2025-11-05T19:55:40 < jpa-> it is demonstrating how common mode voltage is defined/calculated
2025-11-05T19:56:24 < jpa-> because the differential input has somewhat limited common mode range (vref/2 +- 10%) for best accuracy
2025-11-05T19:57:12 < jpa-> actually quite limiting :)
2025-11-05T20:00:12 < jpa-> so yeah, i guess it is more for reducing noise level than making your life easy when reading shunts
2025-11-05T20:00:57 < jpa-> if you want real precision, you'd have a differential amplifier from the shunt to the ADC input
2025-11-05T20:01:19 < mawk> yeah I have that
2025-11-05T20:01:31 < mawk> 2mV/mA
2025-11-05T20:02:34 < mawk> with a 0.1% precision shunt
2025-11-05T20:03:12 < jpa-> yeah, so if that has a differential output, use that; if not, use single-ended
2025-11-05T20:05:40 < mawk> yeah it's just referenced to ground 
2025-11-05T20:12:43 < PhantomWork> mawk: was the baud rate question for me?  if so, 1Mbps, with an average of ~3.5kB/s output.
2025-11-05T20:15:29 -!- specing [~specing@user/specing] has joined ##stm32
2025-11-05T20:19:19 < mawk> yes
2025-11-05T20:20:53 < PhantomWork> I've updated the code, let me pastebin
2025-11-05T20:22:02 < mawk> try with something standard like 921600
2025-11-05T20:22:37 < ventYl> 90210
2025-11-05T20:24:12 < PhantomWork> https://bpa.st/2IWQ
2025-11-05T20:24:28 < PhantomWork> and... good call on that baud rate, will definitely check
2025-11-05T20:24:43 < PhantomWork> just checked the datasheet, and 1M isn't listed directly soooo...
2025-11-05T20:24:55 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-05T20:25:41 < jpa-> is this STM32? you can have whatever baudrate you want, just make sure the clock frequency divides neatly enough to it
2025-11-05T20:25:45 < mawk> it's more for the receiving device, it needs to be one of the standard supported baudrates
2025-11-05T20:25:53 < mawk> the stm32 doesn't care as much
2025-11-05T20:26:02 < jpa-> and yeah, obviously the baudrate has to be same on both ends
2025-11-05T20:27:39 < PhantomWork> jpa-: yes STM32, but also a CH340N, which may not handle 1M
2025-11-05T20:28:44 < PhantomWork> and.... waiting game...
2025-11-05T20:28:53 < jpa-> CH340N is garbage in general in my experience :)
2025-11-05T20:29:38 < jpa-> are you using RTS/CTS or without flow control?
2025-11-05T20:29:44 < PhantomWork> no flow
2025-11-05T20:30:21 < PhantomWork> (does STM32 even support hardware flow?)
2025-11-05T20:31:50 < PhantomWork> question about freertos... let's say that you have a task that do a busy loop, does that lock the other tasks?
2025-11-05T20:33:25 < PhantomWork> and    still corrupt
2025-11-05T20:34:33 < mawk> freertos has preemption
2025-11-05T20:34:35 < mawk> so no
2025-11-05T20:34:50 < mawk> of course stm32 has hw flow control
2025-11-05T20:35:02 < mawk> you can also do xon/xoff flow control by hand
2025-11-05T20:35:13 < mawk> but tricky
2025-11-05T20:36:55 < mawk> and did you disable flow control in your serial viewer?
2025-11-05T20:37:02 < mawk> that's pretty important
2025-11-05T20:37:26 < mawk> both hw and sw
2025-11-05T20:37:33 < PhantomWork> yeah no flow in the terminal
2025-11-05T20:38:15 < jpa-> PhantomWork: sure, STM32 has hardware flow control support
2025-11-05T20:38:18 < PhantomWork> I dropped down to 115200, that's BARELY fast enough as per my counters... so we'll see if that's better... the 921600 wasn't
2025-11-05T20:38:43 < jpa-> PhantomWork: so is the problem in STM32->PC direction or the opposite?
2025-11-05T20:39:38 < PhantomWork> jpa: STM32->PC at ~10.5kB/s average, and 1 key from pc to STM32  :D
2025-11-05T20:39:59 < PhantomWork> I have a few debug pages, I can select them with 0-9
2025-11-05T20:41:28 < jpa-> so what is the problem?
2025-11-05T20:41:52 < jpa-> ah, so dropped data?
2025-11-05T20:42:00 < PhantomWork> https://bpa.st/3YCA <== that   become this ==> https://bpa.st/6MTA
2025-11-05T20:42:16 < jpa-> yeah, CH340 buffers are filling up
2025-11-05T20:42:46 < jpa-> though, the effect does seem more like you could have STM32 DMA problem
2025-11-05T20:43:09 < PhantomWork> my code was sending it using standard interrupt before, without issue
2025-11-05T20:43:10 < jpa-> check with logic analyzer or a known-good UART chip that actually specifies how much buffer it has
2025-11-05T20:43:25 < PhantomWork> so yeah could be DMA issue or something, or it send too much at a time and something bottleneck
2025-11-05T20:44:02 < jpa-> to me it looks like you start writing new data to DMA buffer before previous one has finished sending
2025-11-05T20:45:58 < PhantomWork> which it shouln't, the tail is updated only after the transmission is completed, and there is a check for the head and tail so it does not collide... So shouln't be able to overwrite in the active area
2025-11-05T20:46:22 < jpa-> do you have a logic analyzer?
2025-11-05T20:49:04 < PhantomWork> no
2025-11-05T20:49:28 < qyx> first, trash ch340
2025-11-05T20:49:36 < qyx> get ftdi, cp2102 or similar
2025-11-05T20:49:39 < PhantomWork> but interessingly, at 115200, 13 minutes in, still no corruption
2025-11-05T20:49:54 < jpa-> bugs are often timing-related
2025-11-05T20:50:05 < jpa-> but i wouldn't be surprised at all if it is CH340 corrupting your data
2025-11-05T20:50:08 < qyx> then try sending in a loop without dma
2025-11-05T20:50:10 < jpa-> that chip is total garbage
2025-11-05T20:50:30 < qyx> 0x01 0x02 0x03.. etc to see if anythig is missing
2025-11-05T20:52:27 < PhantomWork> I think it is indeeed the CH340...
2025-11-05T20:52:39 < PhantomWork> it probably can't handle the bursts
2025-11-05T20:53:13 < PhantomWork> even if it claim to do 2Mbps
2025-11-05T20:53:29 < jpa-> https://github.com/felHR85/UsbSerial/issues/94  it has fun bugs :)
2025-11-05T20:54:19 < jpa-> buffer space: FT232R: 128 bytes; FT232H: 1024 bytes; CP2102: 576 bytes, CH340: doesn't specify, but online discussions report 32 bytes
2025-11-05T20:55:12 < PhantomWork> next time I spin the board, I'll swap for the cp
2025-11-05T20:55:45 < PhantomWork> and probably also a better LDO than what I use, maybe
2025-11-05T20:56:11 < PhantomWork> thanks guys     atleast 115200 seems to work
2025-11-05T20:56:27 < PhantomWork> so, by optimising the uart, I broke the ch340  lol
2025-11-05T20:57:31 < jpa-> it's pretty common with cheapie USB chips that they only have enough buffer when the PC is responding very fast
2025-11-05T20:57:42 < jpa-> if there is a slight delay on the PC side, it runs out of buffer
2025-11-05T20:58:09 < jpa-> why do you have usb-uart on board anyway? doesn't the stm32 have usb?
2025-11-05T21:04:17 < qyx> thats a forbidden question
2025-11-05T21:04:33 < qyx> because usb-uart bridge saves you 2 days og debuggin broken usb stacks
2025-11-05T21:05:44 < Steffanx> Only 2?
2025-11-05T21:06:30 < Steffanx> Oh 2 times 24 hours. So you mean 6 work days :P
2025-11-05T21:07:18 < PhantomWork> jpa-: I use CANbus, the F103 is usb OR can, can't do both
2025-11-05T21:09:04 < jpa-> 2008 wants it mcu back :)
2025-11-05T21:09:19 < ventYl> 2004
2025-11-05T21:09:42 < jpa-> qyx: it saves some time, but gets you snarky comments :)
2025-11-05T21:10:03 < jpa-> ventYl: i don't think there were any STM32 in 2004
2025-11-05T21:10:49 < jbo> did somebody say ch340?
2025-11-05T21:10:51 < jbo> fuck
2025-11-05T21:11:14 < ventYl> jpa-: hm, right. Cortex-M is from that year
2025-11-05T21:11:42 < Steffanx> Language jbo. Language 
2025-11-05T21:12:03 < jbo> apologies
2025-11-05T21:13:02 -!- server is now known as ds2
2025-11-05T21:13:11 -!- ds2 [~ds2@user/ds2] has changed host
2025-11-05T21:16:13 < mercenary> ch340. that chip that has a pin to switch buffer direction for RS485, and switches direction 'sometime, anywhere between 10 and 90% into the start bit'
2025-11-05T21:17:25 < jpa-> i wonder how many people on this channel have wasted at least an hour of their life with CH340 problems :)
2025-11-05T21:18:33 < jbo> :D
2025-11-05T21:19:11 < PhantomWork> in my case, it worked fine before I move from interrupt to DMA so I assumed the ch340 and windows side was fine... don't assume I guess
2025-11-05T21:19:42 < jpa-> also, buy a logic analyzer, it's great for solving problems like this
2025-11-05T21:20:39 < jbo> so is PhantomWork the new guy?
2025-11-05T21:20:41 < jbo> welcome \o/
2025-11-05T21:21:59 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 260 seconds]
2025-11-05T21:27:30 < mawk> the cdc acm example code generated by el cubo always worked first try for me
2025-11-05T21:27:32 < mawk> jbo the second french guy
2025-11-05T21:27:34 < mawk> I think 
2025-11-05T21:27:36 < mawk> do you live in the greatest city of the world PhantomWork ?
2025-11-05T21:27:38 < mawk> and why the f103
2025-11-05T21:28:27 < jbo> oh we're doing f103? that's sick. was my first STM32 too.
2025-11-05T21:28:43 < jbo> now that I am thinking of it... my CH340 pain was also with an F103 :D
2025-11-05T21:29:01 < jbo> I guess we all learn the CH340 pain early on.
2025-11-05T21:29:38 < mawk> the first stm32 I bought was at random, a stm32l1
2025-11-05T21:30:31 < Steffanx> It wasn't my first. It was the F100 or 101 here. Not such which of the 2. I think 100
2025-11-05T21:32:55 < qyx> "what the f103" is the new substitution for curse french words?
2025-11-05T21:34:21 < qyx> f103 was my first stm32 too
2025-11-05T21:34:26 < qyx> in 2011 or so
2025-11-05T21:40:24 < jbo> yeah I though it was 2009 but I was clearly wrong
2025-11-05T21:41:42 < PhantomWork> mawk: the F103, well, I first ordered a bluepill, did some proof of concept on it, then looked at the most common STM32 on jlcpcb, it happened to be the F103, and used ity
2025-11-05T21:41:54 < mawk> it's ooooools
2025-11-05T21:41:56 < mawk> old
2025-11-05T21:42:01 < PhantomWork> I'm not opposed to change
2025-11-05T21:42:20 < mawk> there's still time to change, maybe even with a pin compatible one if you want
2025-11-05T21:42:40 < mawk> so do you live in paris?
2025-11-05T21:42:50 < jbo> je parlais francais tres bien
2025-11-05T21:43:12 < mawk> mais plus maintenant ?
2025-11-05T21:43:21 < jbo> ca va mieux maintenant
2025-11-05T21:43:27 < mawk> :(
2025-11-05T21:43:45 < jbo> I figured that I lost my french mojo so I decided to hang out with more french people again.
2025-11-05T21:43:47 < PhantomWork> mawk: nope, montreal area
2025-11-05T21:43:53 < mawk> ohlala
2025-11-05T21:45:19 < PhantomWork> so   F407 maybe?
2025-11-05T21:45:30 < qyx> that's about 2014
2025-11-05T21:45:54 < qyx> oh no 2011
2025-11-05T21:46:13 < qyx> F1 was 2007 so already pretty old when I first touched it
2025-11-05T21:46:43 < mawk> depends what you're missing and what you would use PhantomWork 
2025-11-05T21:46:57 < mawk> if your thing runs on battery then maybe a low power line stm32 is useful 
2025-11-05T21:47:03 < qyx> PhantomWork: you either start looking in legit distributor stores and get decent stuff
2025-11-05T21:47:15 < qyx> or you are looking in china and get 10-20y old garbage
2025-11-05T21:47:29 < mawk> is lcsc china garbage 
2025-11-05T21:47:32 < PhantomWork> need CAN, IRDA and that's about it
2025-11-05T21:47:39 < mawk> infrared in 2025?
2025-11-05T21:47:47 < qyx> f103 in 2026 is garbage
2025-11-05T21:47:59 < mawk> qyx lives in the future
2025-11-05T21:48:03 < jpa-> 2 months left to redesign all f103 projects
2025-11-05T21:48:06 < qyx> personally I am refusing to help people with f103 since about 2000
2025-11-05T21:48:09 < qyx> sorry 2020
2025-11-05T21:48:21 < qyx> not worth the time
2025-11-05T21:48:27 < mawk> are they discontinuing it jpa- ?
2025-11-05T21:48:40 < mawk> the chinese have that market all ready to flood with the clones
2025-11-05T21:48:42 < jpa-> mawk: no but qyx has outlawed it
2025-11-05T21:48:45 < mawk> ah
2025-11-05T21:48:50 < mawk> lol
2025-11-05T21:49:04 < PhantomWork> so, which one do you suggest?
2025-11-05T21:49:15 < jbo> f407, the other oldschool memory - nice.
2025-11-05T21:49:29 < jbo> seems like everybody was doing f103 -> f407 back then :p
2025-11-05T21:49:30 < mawk> is it running on battery? is the uart fundamental? (in which case you could use usb instead)
2025-11-05T21:49:33 < qyx> at least a m4f one
2025-11-05T21:49:40 < jpa-> i suggest rp2350, because there are way too many stm32's to choose from :)
2025-11-05T21:49:47 < mawk> and is computing power needed? so floating point and stuff
2025-11-05T21:49:58 < mawk> that's illegal to say in ##stm32 jpa- 
2025-11-05T21:50:13 < jbo> PhantomWork, what are you crafting?
2025-11-05T21:50:27 < jpa-> mawk: says the person who said the c-word
2025-11-05T21:50:40 < mawk> :(
2025-11-05T21:50:54 < jbo> no matter what the c-word is, it's definitely less bad than the rp thing
2025-11-05T21:51:11 < qyx> c-word, rp-thing, irda, f103, where I am
2025-11-05T21:51:32 < jpa-> pico-sdk is like 10000% better than cube
2025-11-05T21:51:45 < jbo> how dare you compare anything to cube
2025-11-05T21:51:50 < jbo> that's like dividing by zero
2025-11-05T21:51:59 < mawk> I like cube, for old chips
2025-11-05T21:52:04 < jbo> what?
2025-11-05T21:52:17 < qyx> cube2 was pretty good
2025-11-05T21:52:18 < mawk> well it works perfectly fine for established chips
2025-11-05T21:52:29 < mawk> but not very well for the newest stuff like stm32wb
2025-11-05T21:55:42 < jbo> can't help but miss the old AVR 8-bit stuff sometimes.
2025-11-05T21:55:49 < jbo> AVR-Studio and all that... was a good experience.
2025-11-05T22:00:51 < PhantomWork> jbo: just some wireless comm for 14 steering buttons, with some canbus sensors bypass and stuff...
2025-11-05T22:02:07 < PhantomWork> not that much processing power required, but it transmit up to 800 CAN packets/sec, and process about 1500 incomming packets/sec
2025-11-05T22:02:39 < qyx> wireless steering buttons? for a tractor/machinery?
2025-11-05T22:02:47 < jbo> do you need FDCAN?
2025-11-05T22:03:09 < qyx> f103 has none
2025-11-05T22:03:50 < PhantomWork> no
2025-11-05T22:03:55 < PhantomWork> half is ok
2025-11-05T22:04:11 < jbo> qyx, true
2025-11-05T22:04:12 < PhantomWork> the F103 work fine right now
2025-11-05T22:04:51 < PhantomWork> and even with the very unoptimised code and massive debug output that is zero optimised, it still have 85+% idle power
2025-11-05T22:05:14 < jbo> why are we talking about different STM32 options then?
2025-11-05T22:05:17 < PhantomWork> and yes, I did implemented a cpu usage
2025-11-05T22:05:25 < PhantomWork> jbo: modernise it...
2025-11-05T22:05:34 < jbo> I assume F103 is ubiquitous enough to be available for the next decade or whatever.
2025-11-05T22:06:39 < PhantomWork> now... I just need to figure out a proper GPIO protection for 1MHz that tolerate short to +16V without frying the STM32
2025-11-05T22:07:26 < jbo> 1MHz? of/for what?
2025-11-05T22:07:28 < jbo> CAN?
2025-11-05T22:08:52 < PhantomWork> IRDA, the tranceiver is away, around... 3ft of wires...  
2025-11-05T22:09:22 < PhantomWork> so far, 1k on each side work, it do very little for the protection but mostly there as a test...
2025-11-05T22:10:34 < jbo> so you need something that tolerates a persistent +16V?
2025-11-05T22:10:48 < PhantomWork> yeah, for wiring damage or misconnection
2025-11-05T22:11:09 < jbo> sounds about right - good thinking. I have that conversation like every 2 months and people don't seem to get it.
2025-11-05T22:11:41 < jbo> last year I had three meetings with a customer where I explained why they need 48V portection on their CAN bus shit.
2025-11-05T22:12:04 < jbo> is this a cost sensitive application?
2025-11-05T22:12:05 < PhantomWork> the stm32 really don't like 12V, fried 1 so far, and fried 3 IRDA tranceiver, they absolutely does not tolerate abuse :(
2025-11-05T22:12:26 < PhantomWork> cost is always an issue, but space is a big one
2025-11-05T22:13:21 < jbo> you seem to know the game - so let me rephrase: Is cost a primary concern? :p
2025-11-05T22:13:32 < mawk> what about STM8 jbo 
2025-11-05T22:13:34 < mawk> did you try that
2025-11-05T22:13:40 < jbo> mawk, I actually did, jup.
2025-11-05T22:13:51 < mawk> stm32 wants 5V maximum
2025-11-05T22:13:53 < mawk> some pins are not tolerant and only want VDD
2025-11-05T22:13:57 < mawk> they can probably take higher voltage transients but not for very long
2025-11-05T22:14:04 < jbo> VDD+0.3V if I am not mistaken
2025-11-05T22:14:18 < PhantomWork> jbo: cost is always something I look for, but if it can be justified I will look at it for sure
2025-11-05T22:14:51 < jbo> PhantomWork, is this an industrial or medical application or something? or just "dicking around in my cousins backyard"?
2025-11-05T22:14:53 < qyx> I would use a $1 esp32 with wifi instead of f103+irda
2025-11-05T22:15:05 < mawk> be patriotic qyx 
2025-11-05T22:15:07 < mawk> stm32wb
2025-11-05T22:15:15 < PhantomWork> jbo: https://imgur.com/a/iTCWWK1 <=== 22mm diameter
2025-11-05T22:15:16 < qyx> wb is not wifi
2025-11-05T22:15:19 < PhantomWork> jbo: automotive
2025-11-05T22:15:20 < mawk> wl
2025-11-05T22:15:32 < jbo> qyx, what happened to ST67?
2025-11-05T22:15:32 < mawk> or whatever it is
2025-11-05T22:15:35 < mawk> they have one wifi one
2025-11-05T22:16:25 < mawk> or maybe it's with an extra external chip, I forgot
2025-11-05T22:16:44 < mawk> yeah it's another module
2025-11-05T22:16:56 < mawk> st67whatever
2025-11-05T22:17:50 < mawk> why do you have infrared in a car PhantomWork 
2025-11-05T22:19:08 < jbo> PhantomWork, I mean... you'd want the protection on the STM32 side anyway, and you presumably have more space there?
2025-11-05T22:22:39 < PhantomWork> jbo: need on both side
2025-11-05T22:24:26 < PhantomWork> mawk: there is not much ways to transfert data when both side can be disconnected, and exposed to the elements. We ended up having only 2 wires that we could pass, the Data? Data over power or through the plastic. Ended up with through the plastic. And standard RC is too slow, so IRDA is what I used. The other means, like NFC, is kinda too complex
2025-11-05T22:25:18 < mawk> both sides disconnected but still powered on right? so why was 2 wires not enough
2025-11-05T22:25:37 < mawk> or are you actually doing IRDA over wires
2025-11-05T22:27:54 < PhantomWork> the steering is removable for easy entry, the quick coupler comes with 2 pins, normally for the horn, reused it for power. there is 14 illuminated buttons on the steering, plus a led bar for RPM. IRDA is behind the quick coupler plate for the power
2025-11-05T22:28:33 < qyx> oh, single pair ethernet to the rescue!
2025-11-05T22:28:55 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 246 seconds]
2025-11-05T22:29:03 < qyx> I am building a tiny tractor and will use SPE for my steering wheel too
2025-11-05T22:29:06 < PhantomWork> can that also do power???
2025-11-05T22:29:10 < qyx> of course
2025-11-05T22:29:22 < PhantomWork> linky?
2025-11-05T22:29:27 < qyx> 10base-t1s
2025-11-05T22:29:31 < qyx> +PoDL
2025-11-05T22:29:42 < mawk> if the other side is low power there's also 1-wire
2025-11-05T22:29:56 < mawk> but you probably need more power
2025-11-05T22:30:11 < PhantomWork> mawk: no, 150-200mA iirc
2025-11-05T22:30:21 < mawk> that's a lot of milliamps
2025-11-05T22:31:08 < Steffanx> So is 1.5mA is some cases :P
2025-11-05T22:31:13 < PhantomWork> yup, I tough about 1 wire too, and it is actually something I considered, until the illumination request came up :/
2025-11-05T22:31:33 < mawk> wrong Steffanx 1.5 is basically 0
2025-11-05T22:31:47 < PhantomWork> and I'm pretty sure I saw a 1 wire 16 GPIO chip out there many years ago, so it was kinda perfect
2025-11-05T22:36:23 < ventYl> 1wire is crazy slow
2025-11-05T22:37:45 < PhantomWork> ventYl: 16.3kbps
2025-11-05T22:38:06 < ventYl> that is crazy slow, if you consider that device addresses are like 48 or even 64 bits
2025-11-05T22:38:22 < ventYl> and you have to address the device on each and every transmission
2025-11-05T22:38:50 < PhantomWork> PoDL is only 5W at 12V? that's low :(
2025-11-05T22:40:30 < ventYl> what is the exact problem you are trying to solve? too low pin count to push power and data over a connection?
2025-11-05T22:41:42 < PhantomWork> I only have 2 pins, longish wire, and a transparentish plastic 
2025-11-05T22:42:24 < PhantomWork> ow well, I guess I'll stay irda
2025-11-05T22:42:45 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-05T22:43:35 < ventYl> what limits you to only two pins? that PCB seems large enough to accommodate four of them
2025-11-05T22:44:12 < mawk>  you need 2.5W right PhantomWork ? 200mA at 12V
2025-11-05T22:44:27 < mawk> so why is that low
2025-11-05T22:46:39 < mawk> VREFBUF in "hold mode" is what I want if I can't disconnected the external VREF+ pin right? the VREF+ pin will be set to Hi-Z and the internal voltage reference will be used
2025-11-05T22:55:34 < qyx> PhantomWork: PoDL on T1S is not standardized yet, you can inject whatever you want
2025-11-05T22:55:43 < qyx> I am using 2 A at 28 V
2025-11-05T22:56:08 < PhantomWork> qyx: also I don't need to be compliant, but 2A? I'ld take it for sure
2025-11-05T22:57:08 < PhantomWork> qyx: do you happen to have a schem with parts? and do you need to have true ethernet data or you can pass arbitrary 1 and 0es?
2025-11-05T23:09:01 < karlp> qyx: how is it not standardized yet?!
2025-11-05T23:13:10 < PhantomWork> So, flow control on F103... not available if you use CAN
2025-11-05T23:13:21 < karlp> cute, esp32c3 errata has two erratas.  they have ~6 revisions of the silicon mentioned.  erratas affect _every single_ revision
2025-11-05T23:13:33 < karlp> PhantomWork: yeah, f103 is famous for "you can only have one"
2025-11-05T23:13:38 < karlp> like, no can+usb as well iirc.
2025-11-05T23:13:46 < karlp> why are you still using that dinosaur anyway?
2025-11-05T23:14:07 < PhantomWork> because it is what I got my hand on first and stayed on it    lol
2025-11-05T23:15:34 < karlp> re ch340, one thing is that there have been _many_ revisions of that over time...
2025-11-05T23:15:34 < qyx> karlp: it is for T1L
2025-11-05T23:15:37 < qyx> not for T1S
2025-11-05T23:15:50 < PhantomWork> and, to be fair, I don't see a major reason to upgrade, the others are bigger, more expensive, and, don't really offer a better functionality... howeverrrrr    now that the CH340M is found to be problematic, upgrading miiight be worth it
2025-11-05T23:16:37 < PhantomWork> will look into it for sure, but right now, I have zero concentration skill left today, and small headache, so.... I'll look tomorrow
2025-11-05T23:17:11 < karlp> oh man, that usbserial for android has it's own driver code, no wonder it has wonks.
2025-11-05T23:17:17 < PhantomWork> if I can have the usb-serial build in, that would fix things too
2025-11-05T23:17:25 < karlp> there's been tons of workarounds and fixes in various drivers.
2025-11-05T23:17:44 < karlp> (it would still never be my first choice, but it's far less fail prone with recent linux kernels than it used to be at least)
2025-11-05T23:19:42 < karlp> ch340 most famous pain for stm32ers was that the linux drivers didn't support parity for years, which stm32serial programming required ;)
2025-11-05T23:20:11 < PhantomWork> what do you think of STM32F407? good upgrade or need something better?
2025-11-05T23:20:55 < zyp> f407 is not bad, but kinda old
2025-11-05T23:22:23 < zyp> > don't really offer a better functionality...
2025-11-05T23:22:26 < zyp> yes they do
2025-11-05T23:22:42 < zyp> every single other stm32 offers better functionality than f1
2025-11-05T23:23:00 < ventYl> except L0 :>
2025-11-05T23:24:18 < karlp> (IMO G431 is ~cheaper, better, and edquivalently well sstocked on jlcpcb.com/parts, even if they have "more" f103s in stock.
2025-11-05T23:24:38 < zyp> l0 is not bad, given the choice between l0 and f1 I'd pick l0 if it could get the job done
2025-11-05T23:24:55 < zyp> but yeah
2025-11-05T23:25:16 < karlp> L4 and G4 are both available for ~$2
2025-11-05T23:25:22 < zyp> g431 is probably what I'd pick for some cheap generic midrange thing
2025-11-05T23:25:50 < karlp> eh, it's not _technically_ cheaper than f103, but... it is in the long run I promise ;)
2025-11-05T23:26:18 < zyp> if you care about cheap, go c0 or something :p
2025-11-05T23:26:27 < PhantomWork> hollymolly    that is big compared to the 103
2025-11-05T23:26:50 < zyp> what do you mean big?
2025-11-05T23:27:05 < zyp> physically or feature wise?
2025-11-05T23:27:06 < ventYl> phantomwork is working on something like racing steering wheel, right? probably not exactly something where $0.5 on BOM price will make a huge difference
2025-11-05T23:27:28 < zyp> yeah, for one offs price doesn't matter
2025-11-05T23:27:51 < PhantomWork> zyp: physically
2025-11-05T23:27:58 < PhantomWork> 4x the size of the 103
2025-11-05T23:28:01 < zyp> which one? g4?
2025-11-05T23:28:08 < PhantomWork> F407
2025-11-05T23:28:16 < zyp> you do realize that all of these comes in multiple package options
2025-11-05T23:28:31 < zyp> yeah, f407 is qfp64 or larger
2025-11-05T23:29:26 < zyp> anyway, if what you want is a f1 replacement, go with g431
2025-11-05T23:29:27 < karlp> I mean, you said you picked the most common stm32 on jlc, that's C8T6, g431 is CxTx as well...
2025-11-05T23:29:37 < zyp> yep
2025-11-05T23:29:40 < karlp> those ar ethe same, if you'r eordeing by "popular by stock on jlc"
2025-11-05T23:30:56 < PhantomWork> karlp: it was the 103 when I started it
2025-11-05T23:31:53 < zyp> I've been pushing stm32 for various work projects recently, so far we've got ongoing projects with h573, g071 and IIRC c051
2025-11-05T23:35:19 < PhantomWork> qyx: for the ethernet thing you talked about, do you need an ethernet phy on the stm? or is there a way to just talk serial over it?
2025-11-05T23:36:02 < zyp> t1s? you can get either just external phy, or phy+mac combo with spi interface
2025-11-05T23:38:17 < zyp> for t1s, I'd consider phy+mac combo, but for anything else ethernet I'd want a mcu with internal mac
2025-11-05T23:38:31 < zyp> in which h5 is a good candidate
2025-11-05T23:38:43 < ventYl> the intention is probably different here
2025-11-05T23:40:34 < PhantomWork> 100kbps is more than I need, 500kbps and it would be a dream (it open up some future expensive possibilities)
2025-11-05T23:46:27 < PhantomWork> I hate you guys, you make me want to change the µC, remove the ch340, add the T1S and all...
2025-11-05T23:46:40 < PhantomWork> :)
2025-11-05T23:46:57 < PhantomWork> it's already a somewhat tight board...  so   will be a pain
2025-11-05T23:46:58 < qyx> mawk: re h7 adc diff inputs, whats the outcome? I have ADC12_INN5 and INP5, does it have a pga?
2025-11-05T23:47:12 < qyx> or should I be less lazy and open the damned reference manual
2025-11-05T23:51:34 < Steffanx> Yes
2025-11-05T23:55:33 < PhantomWork> Well, thanks guys, really appreciated for the help!
2025-11-05T23:58:48 -!- Livio_ [~livio@user/livio] has quit [Quit: leaving]
--- Day changed Thu Nov 06 2025
2025-11-06T00:11:27 < karlp> heh, not sure this is going to fit real well: https://bin.jvnv.net/file/BOLKi.png
2025-11-06T00:12:14 < karlp> might be more feasible using the chip itself instead of the module...
2025-11-06T00:12:57 < qyx> karlp: are you doing BLE christmas lights?
2025-11-06T00:20:43 -!- ds2 [~ds2@user/ds2] has left ##stm32 []
2025-11-06T00:20:53 -!- ds2 [~ds2@user/ds2] has joined ##stm32
2025-11-06T00:24:50 < karlp> been considering it off and on for a few years, replacing the controller boards _in the existing housings_
2025-11-06T00:24:58 < karlp> because I have some "nice" waterproof ones
2025-11-06T00:25:09 < karlp> and I can't come close to making _that_ part at a sane price.
2025-11-06T00:26:26 < karlp> looks like it might work if I yolo certs and just plop the bare chip down
2025-11-06T00:26:32 < mawk> yes there is a PGA qyx 
2025-11-06T00:26:40 < karlp> ok. enough of this.
2025-11-06T00:26:54 < mawk> with gain 1, 2, 4, 8, 16
2025-11-06T00:27:15 < mawk> you need the diff input?
2025-11-06T00:27:31 < mawk> it needs to be biased around VREF/2 and not differ too much from it for maximal accuracy
2025-11-06T00:29:32 < mawk> but I already have a difference amplifier to amplify the shunt voltage drop so I don't need the diff input
2025-11-06T00:31:14 < mawk> vrefbuf being 2.5V maximum is not very good, that puts my input current at 1.25A maximum and I'm pretty sure I go above that at some point
2025-11-06T00:31:23 < mawk> I wish I could divide the signal with the PGA instead
2025-11-06T00:33:20 < mawk> I can use external resistors to do the dividing but then they need to have a pretty precise value
2025-11-06T00:36:29 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 250 seconds]
2025-11-06T00:41:59 < qyx> mawk: https://bin.jvnv.net/file/p8grN/Screenshot_2025-11-05_23-39-54.png
2025-11-06T00:42:08 < qyx> I just want to measure one pt1000
2025-11-06T00:42:13 < qyx> this lame method should be okay
2025-11-06T00:42:38 < qyx> even without any PGA it should provide some precision
2025-11-06T00:42:59 < qyx> and 0.6 mA measuring current
2025-11-06T00:43:27 < mawk> well if you don't care about super fast speeds with oversampling you should be more than ok
2025-11-06T00:43:31 < mawk> then you get 26 bits effectively
2025-11-06T00:44:21 < qyx> nah 1 Hz is enough
2025-11-06T00:48:13 < mawk> even 27 bits in differential mode
2025-11-06T00:49:16 < mawk> you can always take two measurements of TEMP_P and TEMP_N and subtract them if the differential mode doesn't work
2025-11-06T00:49:23 < mawk> you can take them at the same exact time with ADC1 and ADC2
2025-11-06T00:49:35 < qyx> yeah
2025-11-06T00:55:26 < mawk> so in your case you have TEMP_P - VDDA/2 = VDDA/2 - VDDA 1k/(2k+R), TEMP_N - VDDA/2 = -VDDA/2 + VDDA 1k/(2k+R)
2025-11-06T00:55:43 < mawk> so the two signals are properly differential when biased to VDDA/2 as they want
2025-11-06T00:55:44 < mawk> so it should work
2025-11-06T00:58:21 < qyx> ye
2025-11-06T00:58:22 < qyx> s
2025-11-06T00:59:27 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-06T00:59:47 < nomorekaki> pump sweats
2025-11-06T01:00:24 < qyx> work commited, tractor time
2025-11-06T01:00:31 < nomorekaki> finally tractor
2025-11-06T01:00:52 < nomorekaki> but is work ever done?
2025-11-06T01:01:28 -!- Phantom [~Phantom@user/phantom] has joined ##stm32
2025-11-06T01:01:46 < mawk> the common mode voltage should be within 10% of VREF/2, so I guess 1% 1KΩ resistors should do
2025-11-06T01:02:43 < Phantom> Question guys, would you use a 3.3v buck regulator + 5V buck regulator? OR a 5V buck and a 3.3V LDO off the 5V? I need both due to CAN..
2025-11-06T01:03:25 < zyp> no, I'd use a 3.3V CAN transceiver
2025-11-06T01:03:58 < mawk> the 5V is for CAN and 3.3V for the stm32?
2025-11-06T01:06:09 < Phantom> mawk: yeah
2025-11-06T01:06:35 < Phantom> zyp: can't, because then the bus side will not go to 5V as it is supposed to in this application, so will cause errors
2025-11-06T01:06:39 < qyx> I usually do vbus -> 5V buck and then 5V->3V3 another buck
2025-11-06T01:06:45 < qyx> and no, I don't use 3V3 can transceivers
2025-11-06T01:07:00 < mawk> I thought 3.3V can signalling is compatible with 5V
2025-11-06T01:07:27 < qyx> depends on the termination
2025-11-06T01:07:35 < Phantom> from what I read: it depend. It may, or may not.
2025-11-06T01:07:57 < qyx> if you use actively driven split termination, it is not
2025-11-06T01:08:16 < qyx> then the idle voltage is 2.5V
2025-11-06T01:08:37 < mawk> ah I see
2025-11-06T01:08:46 < mawk> if the 3.3V is exclusively for the stm32 and you're running out of room on the board a LDO seems fine to me, you say the CPU is not doing much so you can lower the core clock and make it consume less power
2025-11-06T01:09:11 < Phantom> power is not really an issue
2025-11-06T01:09:35 < qyx> a 3v3 buck is 10x10 mm max
2025-11-06T01:09:38 < Phantom> space kinda is, but I THINK I have enough space on the board for the bigger one
2025-11-06T01:10:48 < Phantom> https://i.imgur.com/wQDswOs.jpg  <=== bottom left, the 2 regs I have, 
2025-11-06T01:11:50 < mawk> well I mean power would be an issue with a LDO, it will heat up by 170mW for each 100mA
2025-11-06T01:15:07 < qyx> Phantom: uh? bottom left are two voltage regulators?
2025-11-06T01:15:41 < qyx> your layout is extremely suboptimal
2025-11-06T01:16:19 < Phantom> not that suboptimal I think
2025-11-06T01:16:23 < qyx> totally
2025-11-06T01:16:54 < qyx> I am not criticizing but you probably violated every single SMPS design rule
2025-11-06T01:17:41 < Phantom> like?
2025-11-06T01:18:11 < Phantom> imo, it is close to the reference design
2025-11-06T01:18:12 < nomorekaki> g'dam imgur link I almost couldn't escape scrolling
2025-11-06T01:18:21 < ventYl> :))
2025-11-06T01:19:04 < qyx> minimizing loops, very wide traces/polygons, small package sizes for MLCC caps (voltage derating), probably hitting saturation current of that inductor (too small for that value), etc.
2025-11-06T01:19:36 < qyx> you have to use massive amount of vias for proper thermal performance
2025-11-06T01:20:07 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 240 seconds]
2025-11-06T01:20:16 < qyx> feedback divider should be connected to the farmost capacitor (from the inductor)
2025-11-06T01:22:26 < Phantom> so, I did it not too bad then
2025-11-06T01:22:57 < qyx> https://bin.jvnv.net/file/9liMo/Screenshot_2025-11-06_00-22-12.png
2025-11-06T01:23:11 < qyx> this layout and vreg is 9x9 mm and can do 2 A on the output
2025-11-06T01:23:46 < Phantom> iirc this layout is rated 3A
2025-11-06T01:24:02 < qyx> definitely not
2025-11-06T01:24:40 < mawk> you should maybe add NRST to the programming port next time Phantom 
2025-11-06T01:24:53 < qyx> what inductor is that?
2025-11-06T01:24:57 < mawk> if you mess up the firmware or decide to enable readout protection for instance it will be annoying to reprogram
2025-11-06T01:25:05 < Phantom> 22µ 1A    so yeah not 3A....
2025-11-06T01:25:20 < qyx> which exact part number
2025-11-06T01:25:41 < Phantom> mawk: the stlink already work to remove the readout protection
2025-11-06T01:25:58 < Phantom> qyx: YNR4030-220M
2025-11-06T01:26:29 < mawk> Phantom: under reset yes
2025-11-06T01:26:48 < mawk> if you enable RDP 2 or RDP 1 with the debug port disabled, the stlink can't do anything until you connect under reset
2025-11-06T01:27:13 < Phantom> hmm
2025-11-06T01:27:18  * Phantom looks up the code
2025-11-06T01:27:21 < mawk> or even with a regular firmware if you disable the debug port
2025-11-06T01:27:54 < Phantom> what annoy me is that locking the chip crash it up
2025-11-06T01:28:05 < mawk> what do you mean 
2025-11-06T01:28:22 < mawk> you're supposed to do a full power on reset (power off, power on) after enabling RDP yeah
2025-11-06T01:29:05 < Phantom> level 1 is what I do in the code
2025-11-06T01:29:25 < Phantom> whick, thanks to make me notice that I should put that in it's own lib since I reuse the code
2025-11-06T01:29:45 < qyx> it doesn't matter for f103 anymore, it is already broken
2025-11-06T01:29:50 < mawk> lol
2025-11-06T01:30:30 < mawk> if you lock up the chip by disabling the debug port, you have to set up stm32cubeprogrammer or openocd to do a connect under reset, then you hold the reset button, start the connection on your PC and then quickly release it
2025-11-06T01:30:36 < mawk> or you just add NRST to the debug header and it does it for you
2025-11-06T01:30:44 < Phantom> I'll check tomorrow if I can fit the 407 on the board, it help that I'll eliminate the ch340
2025-11-06T01:31:07 < Phantom> yeah I'll add reset
2025-11-06T01:32:26 < mawk> is there even RDP on the F103? I can't find it in the RM
2025-11-06T01:32:32 < Phantom> yes
2025-11-06T01:32:55 < Phantom> and it was annoying at first, because I put it in the code while still debugging.... 
2025-11-06T01:33:17 < Phantom> I was quick to add a "if in debug mode, do not program the RDP" in the code
2025-11-06T01:34:57 < Phantom> https://i.imgur.com/HhOVnJI.png  <=== that plus a ground plane for the regulators
2025-11-06T01:35:22 < Phantom> (there is traces for the ground, incomplete, but still traces)
2025-11-06T01:41:09 < mawk> ah yeah the F103 is old enough that it doesn't have the more sophisticated readout protection
2025-11-06T01:42:27 < qyx> Phantom: the inductor *must* be literally touching the switching pin of the vreg
2025-11-06T01:42:36 < qyx> the closer the better
2025-11-06T01:44:17 < qyx> which regulator is it? there is probably not enough capacitance on the input and output too
2025-11-06T01:55:42 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-06T02:42:34 < Phantom> qyx: TPS5430DDAR
2025-11-06T02:45:38 < Phantom> qyx: https://i.imgur.com/cYc5wN0.png
2025-11-06T03:37:40 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-06T04:53:47 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-06T05:33:32 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-06T05:45:44 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-06T07:34:48 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-06T07:38:04 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-06T07:45:03 -!- qyx [~qyx@84.245.120.89] has quit [Ping timeout: 252 seconds]
2025-11-06T07:46:51 -!- qyx [~qyx@84.245.120.196] has joined ##stm32
2025-11-06T08:13:46 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-06T08:50:48 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-06T09:16:13 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 264 seconds]
2025-11-06T09:25:51 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-06T10:23:29 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-06T10:59:22 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-06T11:27:21 < karlp> wow, are those ultracuck symbols or something. heckin gross.
2025-11-06T11:27:44 -!- hackkitten [~hackkitte@94.31.119.144] has quit [Ping timeout: 240 seconds]
2025-11-06T11:27:44 -!- BrainDamage [~m-t6k752@user/BrainDamage] has quit [Remote host closed the connection]
2025-11-06T11:30:11 -!- BrainDamage [~m-t6k752@user/BrainDamage] has joined ##stm32
2025-11-06T11:32:24 < karlp> yuck, async buck :|
2025-11-06T11:32:37 < karlp> friends don't let friends use diodes... :)
2025-11-06T11:32:57 < qyx> whoa 2 mA Iq *non*-switching
2025-11-06T11:33:51 < karlp> don't worry, there's not enough vin coming in thought that feed trace to ever matter about...
2025-11-06T11:34:25 < karlp> yeah, Phantom  but seriously, having the L on the other side of the D and C makes  a  _bigggg_ swithcing loop
2025-11-06T11:35:07 < karlp> but I see what you mean about the ref layout :)
2025-11-06T11:35:23 < qyx> I was going to say, looking at the reference layout
2025-11-06T11:35:25 < karlp> the ref layout is like, "yeah, just put that L over there somewhere.
2025-11-06T11:35:26 < qyx> the hell is that
2025-11-06T11:35:27 < karlp> ffs.
2025-11-06T11:36:29 < karlp> oh f, it's in jlc basic :)
2025-11-06T11:36:41 < karlp> that's _a_ reason to use it I guess :)
2025-11-06T11:38:09 < karlp> Phantom: I do like seeing people design for the cehap parts though, big fan.
2025-11-06T11:38:34 < qyx> I hate that
2025-11-06T11:38:38 < qyx> we are not that poor
2025-11-06T11:38:52 < karlp> rememeber when you pick inductors though, that "heat rating" is the current at which it has risen _40C_
2025-11-06T11:39:05 < karlp> you may want to make that "bigger"
2025-11-06T11:39:30 < karlp> well, this one is +20, not +40, but "check your datasheetz"
2025-11-06T11:39:41 < qyx> definitely that 4x4 mm inductor with R=225 mOhm or so isn't gonna do 1 A
2025-11-06T11:39:54 < qyx> it would melt
2025-11-06T11:40:37 < jpa-> the reference layout is silly, and Phantom's layout is worse :)
2025-11-06T11:41:02 < jpa-> qyx: 0.22W in 4x4mm doesn't sound that bad
2025-11-06T11:41:48 < qyx> ouch 100 uF in 0805/1206 package
2025-11-06T11:42:33 < qyx> also 0.5V drop schottky
2025-11-06T11:44:12 < karlp> 100uF in 1206 is fine, cmon it's 2025.
2025-11-06T11:44:36 < qyx> for what, 20 uF?
2025-11-06T11:45:34 < karlp> does it need to be 22uH? that's abig inductor too, 
2025-11-06T11:45:53 < jpa-> 30V->5V at 500kHz does require somewhat big inductor
2025-11-06T11:46:22 < karlp> just running into the top of the range for 4040 sized parts.
2025-11-06T11:46:53 < jpa-> i wonder if that regulator is stable with big ceramic output capacitors.. the reference design is with tantalum and just mentions "The minimum ESR of the output capacitor must also be considered. For good phase margin, the ESR zero when
2025-11-06T11:46:54 < jpa-> the ESR is at a minimum must not be too far above the internal compensation poles at 24 kHz and 54 kHz."
2025-11-06T11:47:30 < qyx> karlp: no, if you pick a 100u 1206, x5r, 6v3, it definitely wouldn't work for 5 V
2025-11-06T11:47:49 < karlp> it just used tants because alu was no good and ceramics weren't valid when that part was invented in early 2000s?
2025-11-06T11:49:30 < karlp> I see your jlc basic push button switches too :)
2025-11-06T11:50:20 < karlp> ok, more than enough morning distractions...
2025-11-06T11:50:46 < jpa-> 100µF 1206 X5R looks to be 30µF at 5V, and 2 mohm ESR; so with 3x in parallel that would give 90µF and less than 1mohm, putting the ESR zero at 1.7 MHz
2025-11-06T11:50:53 < jpa-> i think that may be "too far above 54 kHz"
2025-11-06T11:52:03 < jpa-> assuming i'm calculating this stuff correctly :)
2025-11-06T11:52:59 < jpa-> i remember that SMPS controllers from that era were somewhat more picky about stability than the modern ones
2025-11-06T11:54:20 < qyx> yes l5973d didn't work with ceramics, it required alu electrolytics
2025-11-06T11:54:33 < qyx> very similar buck switcher and similar requirements
2025-11-06T11:57:42 -!- hackkitten [~hackkitte@94.31.119.144] has joined ##stm32
2025-11-06T12:02:05 < jpa-> https://www.ti.com/lit/an/slva237c/slva237c.pdf   TI has appnote for TPS5430, but weirdly they add external compensation even for aluminum cap :D
2025-11-06T12:03:03 < qyx> our motivation to be distracted from work is exceptional
2025-11-06T12:03:57 < jpa-> "The ESR zero, which could be as low as 1 kHz for aluminum capacitors and as high as 5 MHz for ceramic capacitors, impacts both the gain response and the phase response. This makes it difficult for the fixed internal compensation network to work with a broad range of various aluminum/ceramic output capacitors."    lol so without external compensation it only works with tantalums :)
2025-11-06T12:04:38 < qyx> jpa-: my son had a kid named "Tove" in their kid's magazine
2025-11-06T12:04:42 < jpa-> the ceramic feedback compensation network is one of the most complex ones i have ever seen :D
2025-11-06T12:04:45 < qyx> she looked a bit scandinavian
2025-11-06T12:04:51 < jpa-> qyx: jansson confirmed
2025-11-06T12:05:04 < jpa-> there are no other toves in scandinavia
2025-11-06T12:05:16 < karlp> lol
2025-11-06T12:05:19 < qyx> true, tove torvalds probably emigrated to usa
2025-11-06T12:05:53 < karlp> I have always been a bit dissapointed with the power supply options in jlc basic.
2025-11-06T12:06:35 < qyx> this is exactly the type of switcher which is present in every single china made SOHO gear
2025-11-06T12:06:42 < jpa-> https://jpa.kapsi.fi/stuff/pix/simplified_transfer_function.png  i'm happy they've simplified it
2025-11-06T12:07:06 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-06T12:07:17 < mawk> lol
2025-11-06T12:07:27 < qyx> lol
2025-11-06T12:07:54 < mawk> it's just a bit ugly but it looks pretty simple
2025-11-06T12:07:58 < mawk> two poles two zeroes
2025-11-06T12:08:39 < mawk> idk why they would write // for division and write literal × like in middle school 
2025-11-06T12:08:51 < mawk> ah // is parallel 
2025-11-06T12:08:52 < mawk> I suppose 
2025-11-06T12:08:53 < jpa-> i think // is "resistors in parallel"
2025-11-06T12:08:57 < mawk> yeah 
2025-11-06T12:09:14 < mawk> usually it's written ||
2025-11-06T12:09:16 < jpa-> the 1/(1/a + 1/b) operator we deserve, but never got
2025-11-06T12:09:41 < mawk> half the harmonic mean, 2H(a, b)
2025-11-06T12:09:47 < mawk> 1/2 *
2025-11-06T12:10:51 < qyx> jpa-: aurora (the singer) is how I picture finnish women, despite she being norwegian
2025-11-06T12:12:18 < mawk> my company is owned by the finns
2025-11-06T12:12:24 < mawk> they have the funniest english accent
2025-11-06T12:13:16 < tomeaton17> finland alcohol too expensive
2025-11-06T12:13:43 < qyx> I would recommend .cz then
2025-11-06T12:14:06 < tomeaton17> I am happy with .hu
2025-11-06T12:14:23 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-06T12:16:05 < jpa-> qyx: yeah, totally not finnish-looking
2025-11-06T12:16:37 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-06T12:21:27 < karlp> I had classic norwegians in the office yesterday.
2025-11-06T12:21:58 < karlp> said barely a few words, nodded a couple of times, didn't smile, thanked me very politely at the end.
2025-11-06T12:22:23 < karlp> only thing they coul dhav edone would have been turnning up on skis
2025-11-06T12:22:25 < jpa-> did they smell of fish? or you don't notice because icelanders do too?
2025-11-06T12:31:31  * karlp shrugs
2025-11-06T13:13:29 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-06T13:35:38 < karlp> meh, apple is teh suck.
2025-11-06T13:35:55 < karlp> no bt classic SPP in iOS, so you add btle.
2025-11-06T13:36:11 < karlp> now that you have btle, you think, "neat, I can use chrome to use webbluetooth" nope. not on iOS.
2025-11-06T13:37:23 < ventYl> isn't webbt something like unwanted child?
2025-11-06T13:39:08 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T13:39:31 < karlp> no, it's fucking awesome.
2025-11-06T13:39:43 < karlp> you want me to write a fucking app instead?
2025-11-06T13:39:46 < karlp> how is that better?
2025-11-06T13:43:16 < ventYl> well, both are bad, yet with app you at least don't have three-dimensional works in this setup martix, which is mostly full of zeroes
2025-11-06T13:45:02 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T13:46:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T13:47:27 < mercenary> webbt is another one of those 'If I wanted an operating system, I'd install an operating system, not a browser' things
2025-11-06T13:50:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T13:52:00 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 245 seconds]
2025-11-06T13:56:32 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T14:03:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T14:09:43 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-06T14:13:09 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-06T14:13:43 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Read error: Connection reset by peer]
2025-11-06T14:14:38 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T14:20:36 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-06T14:39:58 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T14:47:54 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-06T14:48:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T14:48:57 < PhantomWork> I see, you made "fun" of my buck regulator while I was asleep this morning! How dare are you to talk like that about it while I'm away!! :D      
2025-11-06T14:49:02 < PhantomWork> errr I mean, good morning!
2025-11-06T14:57:00 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 245 seconds]
2025-11-06T15:03:02 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T15:11:43 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T15:11:51 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T15:20:14 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Read error: Connection reset by peer]
2025-11-06T15:21:09 < jbo> mercenary, are you saying you're not running Microsoft Excel in your web browser????
2025-11-06T15:23:12 < qyx> any favorite SMD sound making device, aka beeper?
2025-11-06T15:23:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T15:25:11 < zyp> I've heard alu caps are good
2025-11-06T15:25:24 < zyp> just feed them enough voltage and you'll get an audible pop
2025-11-06T15:28:10 < jbo> qyx, board I'm gonna order today uses PKMCS0909E4000-R1
2025-11-06T15:29:17 < qyx> TODAY?
2025-11-06T15:29:22 < qyx> such fast work
2025-11-06T15:29:37 < zyp> ha
2025-11-06T15:29:54 < jbo> ask zyp... I have been going back and forth on this for months :p
2025-11-06T15:29:57 < qyx> I already ordered two this week :P
2025-11-06T15:30:32 < PhantomWork> zyp: yeah the visuo-acoustic AlCap are fun
2025-11-06T15:31:29 < qyx> jbo: how do you drive it?
2025-11-06T15:31:35 < qyx> directly off 3v3?
2025-11-06T15:31:52 < jbo> qyx, https://bin.jvnv.net/file/Va06u
2025-11-06T15:32:02 < jbo> keep in mind I haven't actually done it yet so... there be dragons I guess.
2025-11-06T15:33:10 < jbo> you can also get parts with integrated drivers but those are typically fixed frequency
2025-11-06T15:33:14 < qyx> hmm I already have 74HCT2G16 on the pcb, I may use it to drive it off 5V rail
2025-11-06T15:33:27 < qyx> with both channels parallel
2025-11-06T15:33:52 < qyx> I like your pkmc, thanks
2025-11-06T15:34:02 < qyx> saved me at least 1/3 hour
2025-11-06T15:34:04 < qyx> 1/2
2025-11-06T15:34:10 < jbo> <3
2025-11-06T15:34:23 < jbo> you want my kicad part too?
2025-11-06T15:34:29 < jbo> saves you another 10min :p
2025-11-06T15:37:37 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-06T15:37:47 < jbo> qyx, ^
2025-11-06T15:42:41 < PhantomWork> So, the conclusion is that my buck regulator is unsafe and must be replaced?
2025-11-06T15:42:51 < PhantomWork> was it that?
2025-11-06T15:42:52 < qyx> CUI CPT-9019 should be compatible
2025-11-06T15:42:57 < qyx> kicad has footprint for it
2025-11-06T15:43:13 < jbo> you use community libs? :s
2025-11-06T15:43:16 < qyx> unsafe is such a weird word
2025-11-06T15:43:27 < qyx> jbo: of course if I see they look ok
2025-11-06T15:43:31 < jbo> all bucks need to be made in rust
2025-11-06T15:43:47 < qyx> but for stm32 I never use kicad suppplied libs
2025-11-06T15:43:48 < qyx> they are ugly
2025-11-06T15:44:14 < jbo> qyx, yeah I "cherry-picked" in the past too but for a few years now I just create alls hit myself.
2025-11-06T15:44:35 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T15:45:47 < qyx> jbo: hey but CUI specifies it should be driven with 50% square wave push-pull
2025-11-06T15:45:56 < PhantomWork> I use easyeda
2025-11-06T15:46:04 < qyx> 3 Vpp nominal, up to 25 Vpp
2025-11-06T15:46:29 < qyx> and at 3 Vpp 4 kHz, max consumption is 1 mA
2025-11-06T15:46:46 < qyx> I will probably just use that buffer or a gpio directly
2025-11-06T15:48:58 < jbo> meh
2025-11-06T15:53:29 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-06T16:04:21 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T16:12:42 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-06T16:21:12 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-06T16:40:05 < PhantomWork> so, that 1 pair ethernet thing with poe, does anyone have a schematics and a part list? just realised that it miiight be problematic on the other part of the system...
2025-11-06T16:45:09 < mawk> a spark gap qyx
2025-11-06T16:45:12 < mawk> bzzzzzzz
2025-11-06T16:48:53 < jbo> :D
2025-11-06T17:06:49 -!- tomeaton17 [~tom@user/tomeaton17] has quit [Quit: ZNC 1.9.0+deb2ubuntu0.1~esm2 - https://znc.in]
2025-11-06T17:07:23 -!- tomeaton17 [~tom@user/tomeaton17] has joined ##stm32
2025-11-06T17:11:44 < qyx> jbo: https://bin.jvnv.net/file/3nPe6/Screenshot_2025-11-06_16-11-18.png
2025-11-06T17:35:18 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-06T17:43:12 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T17:48:41 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T17:48:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T17:54:05 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-06T18:02:07 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-06T18:03:21 < PhantomWork> the 407 have so much more periphericals... it's almost intimidating lol
2025-11-06T18:07:01 < ColdKeybo[a]rd> Is there a rationale/agreement on when to use set vs assert nomenclature?
2025-11-06T18:07:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T18:08:30 < ColdKeybo[a]rd> Sometimes for code readability it makes more sense to use `set_some_pin(true)` but I also see people forcing `assert_some_pin(true)` nomenclature and I'm not sure what is the reasoning here?
2025-11-06T18:09:03 < ColdKeybo[a]rd> To me both are very clear, but obviously at some point some on some newsletter disagreed :)
2025-11-06T18:11:11 < mercenary> set_x(); reset_x(); assert_x(true); assert_x(false);
2025-11-06T18:12:10 < mercenary> which one to use is probably a matter of personal taste, and 'stick to what the rest of the code uses'
2025-11-06T18:12:26 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-06T18:13:27 < ventYl> aren't they something like "retained for historic compatibility reasons?"
2025-11-06T18:14:37 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-06T18:16:31 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T18:17:08 < jpa-> PhantomWork: if F407 feels complex, wait till you end up using H7
2025-11-06T18:26:46 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 255 seconds]
2025-11-06T18:32:58 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T18:48:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-06T18:52:34 -!- fdarling [~forest@syn-096-033-104-194.biz.spectrum.com] has quit [Quit: Leaving]
2025-11-06T18:56:15 < PhantomWork> jpa-: the 103 only have 4 timers, the 407 has 14. I use 1.
2025-11-06T18:56:45 < ventYl> AVR8 should be enuff for the job
2025-11-06T18:56:58 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T18:57:29 < PhantomWork> ventYl: seriously, if I remove the debug code, yes
2025-11-06T19:00:05 < ventYl> AVR gave me PTSD
2025-11-06T19:00:33 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-06T19:02:41 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-06T19:02:43 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T19:07:25 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-06T19:12:01 < PhantomWork> beside power over single pair ethernet, do you have an idea for other simpler mean of doing data over 12V 2A power?
2025-11-06T19:18:31 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T19:19:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T19:24:07 < ventYl> simple way of doing data over 12V is K-line, but you probably can't easily draw enough parasitic current from it
2025-11-06T19:25:20 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-06T19:29:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 250 seconds]
2025-11-06T19:29:26 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T19:30:31 < Steffanx> AVR wasn't that bad ventYl . You gave yourself PTSD 
2025-11-06T19:30:45 < ventYl> the lack of debugger is bad enough
2025-11-06T19:31:59 < Steffanx> I had a debugger. DebugWire and/or JTAG
2025-11-06T19:32:19 < qyx> jtag? only atmega128 had that, no?
2025-11-06T19:32:33 < Steffanx> Some bigger ones yes
2025-11-06T19:32:59 < Steffanx> The thing that gave me sort of PTSD is the AVRDragon. It was too easy to break em
2025-11-06T19:34:26 < Steffanx> Also printf debugger ftw
2025-11-06T19:34:41 < ventYl> was there anything cheaper than avr-ice that cost like $130?
2025-11-06T19:35:18 < Steffanx> AVRDragon 
2025-11-06T19:35:38 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-06T19:35:41 < Steffanx> I think it was like 50
2025-11-06T19:36:17 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-06T19:37:07 < zyp> avrisp mkii was popular
2025-11-06T19:37:27 < ventYl> hm, that think knew debugwire?
2025-11-06T19:37:54 < zyp> I think that only did spi-style isp
2025-11-06T19:37:56 < ventYl> ah, avrisp, not usbasp
2025-11-06T19:38:24 < ventYl> flashing wasn't a problem, I had DIY ratty-soviet-electronics-look USBasp made of THT parts and universal solder board
2025-11-06T19:38:29 < ventYl> debugging was
2025-11-06T19:39:19 < ventYl> after gaining experience with Renesas E2 and learning that nucleo boards worth $12 have debugger on-board, I lost all interest into AVR
2025-11-06T19:40:01 < ventYl> which means that my DWH is INOP, because I refuse to fix bugs in its firmware to be able to configure new sensor addresses
2025-11-06T19:44:49 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-06T19:50:26 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T19:53:22 < Steffanx> Did you know ##stm32 started in #avr, ventYl ? Will you be OK?
2025-11-06T19:57:58 < ventYl> well, I started in #avr too... shall I have some problems with ##stm32 starting in avr, I'd probably have to jump under train for my own past first
2025-11-06T19:58:12 < Steffanx> I see..
2025-11-06T19:58:46 < qyx> you don't have it too far away, do yo
2025-11-06T19:58:47 < qyx> u
2025-11-06T19:58:57 < ventYl> some ~400 meters
2025-11-06T19:59:02 < ventYl> highway is another 300
2025-11-06T19:59:09 < qyx> so many possibilities
2025-11-06T19:59:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T20:04:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:09:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-06T20:10:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:16:06 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:17:02 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-06T20:17:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T20:30:08 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T20:30:11 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:39:16 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-06T20:40:01 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:45:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T20:45:48 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:53:16 < englishman> and ##gd32 started in ##stm32
2025-11-06T20:53:31 -!- kurfen [~kurfen@45.84.137.20] has joined ##stm32
2025-11-06T20:54:12 < ventYl> well, well, well. SBOMs are baking
2025-11-06T20:57:29 < ColdKeybo[a]rd> Did GigaDevices just clone STM devices or they are just using the same nomenclature?
2025-11-06T20:57:56 < ColdKeybo[a]rd> I have not yet used any GD but I see that f103 you can have in GD or STM flavor
2025-11-06T20:59:03 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T20:59:27 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-06T21:01:28 < jpa-> ColdKeybo[a]rd: they cloned it and changed the nomenclature
2025-11-06T21:01:47 < jpa-> if you look at refman, all the peripherals are the same but registers have different names
2025-11-06T21:05:04 < ColdKeybo[a]rd> So what would be the benefit of using GD over STM? Just cheaper clones because GD didn't have to develop it so they can offer lower pricing?
2025-11-06T21:05:34 < jpa-> well they *did* develop it from the specifications
2025-11-06T21:05:43 < jpa-> but yeah, mainly the price
2025-11-06T21:06:13 < jpa-> their SDK is also slightly cleaner than ST's, but because ST's SDK sucks so much that isn't hard to achieve
2025-11-06T21:06:33 < PhantomWork> also, afaik, GD are also less stable from what I read
2025-11-06T21:06:44 < jpa-> there are some speed differences too, but i'm not sure how much really
2025-11-06T21:07:18 < jpa-> PhantomWork: i haven't seen such problems with genuine GD32 parts
2025-11-06T21:07:36 < jpa-> ebay crap is full of all kinds of crap and qa-fail parts though
2025-11-06T21:07:56 < jpa-> (doesn't matter whether it reads "STM32" or "GD32" on top of the chip, it could be anything :)
2025-11-06T21:08:10 < ColdKeybo[a]rd> Sounds like the meme "same same, but different, but still the same" explains the difference between the two quite well :)
2025-11-06T21:09:02 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-06T21:12:49 < jpa-> someone was wondering a few days ago about non-continuous numbering of SPI peripherals on some STM32
2025-11-06T21:12:51 < jpa-> looks like GD32H7 numbers its timers as: 0, 1, 2, 3, 4, 5, 6, 7, 14, 15, 16, 22, 23, 30, 31, 40, 41, 42, 43, 44, 50, 51
2025-11-06T21:13:16 < jpa-> and no, they still have the classic ST style of randomly some timers being advanced control and others being basic etc.
2025-11-06T21:13:31 < zyp> nrf54l has some peripherals numbered 10, 20, 21, 22 and 30
2025-11-06T21:13:57 < zyp> but there the first digit is also a sane form of grouping
2025-11-06T21:14:17 -!- kurfen [~kurfen@45.84.137.20] has quit [Ping timeout: 260 seconds]
2025-11-06T21:15:06 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T21:15:22 < jpa-> i guess there is some kind of grouping here too.. the small numbers are as they would be on STM32H7, then 2x and 3x are "level 0 timer", 4x is "level 3 timer", 15 and 16 are "level 4 timer" and 5x is "basic timer"
2025-11-06T21:16:18 < jpa-> but really what i wanted to check is how their HRTIM is, but i think GD32H7 doesn't have that
2025-11-06T21:17:04 < jpa-> STM32H7 has that fun errata where they first specified full capability HRTIM, but then just dropped it from the datasheet and all of the high resolution settings are "reserved"
2025-11-06T21:17:30 < jpa-> it still works at them but apparently there is something *massively* wrong as they didn't just issue an errata but completely respecified it
2025-11-06T21:17:32 -!- kurfen [~kurfen@45.84.137.24] has joined ##stm32
2025-11-06T21:19:56 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-06T21:21:41 < PhantomWork> beside the power over single pair ethernet, do you guys have another simpler solution that don't use a huge magnetic?
2025-11-06T21:24:50 < jpa-> what counts as huge?
2025-11-06T21:25:38 < zyp> can't you do PoDL with just cap coupling?
2025-11-06T21:25:57 < jpa-> you still need magnetics to avoid the power supply decoupling from eating the signal
2025-11-06T21:26:19 < jpa-> but if you don't need isolation like in ethernet, the needs are less
2025-11-06T21:28:35 < jpa-> for milliampere levels of currents, you could use a diode or resistor instead, but if you need 2A then you definitely need an inductor
2025-11-06T21:28:48 < jpa-> and probably better to do 500mA @ 48V rather than 2A @ 12V
2025-11-06T21:33:35 < ventYl> I don' think he has the luxury of choosing on-board voltage
2025-11-06T21:35:41 < zyp> PhantomWork, how long distances are we talking here?
2025-11-06T21:36:50 < PhantomWork> 3-4ft of noisy environnement (car)
2025-11-06T21:36:53 < zyp> for short distances, the cost of overlaying power and data isn't really worthwhile vs just adding a power pair to the cabling
2025-11-06T21:37:23 < PhantomWork> I can not add wires, I have exactly 2 wires. no more, no less, no extra ground
2025-11-06T21:37:43 < zyp> but you get to choose what to run over them?
2025-11-06T21:38:31 < PhantomWork> yes, the quick coupling have 2 pins.... so it is power + optical (ir) or combined power + data
2025-11-06T21:39:02 < zyp> which quick coupling? who selected it?
2025-11-06T21:40:32 -!- kurfen [~kurfen@45.84.137.24] has quit [Ping timeout: 260 seconds]
2025-11-06T21:40:34 < PhantomWork> jpa-: https://imgur.com/a/PFqulq8   the underside of the big round can't really be used, except 20mm dia in the middle, which does include the 2 bolts
2025-11-06T21:41:02 < PhantomWork> zyp: the engineers did, and no real other solution
2025-11-06T21:41:42 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 260 seconds]
2025-11-06T21:43:39 < jpa-> a silly solution would be to put a full bridge rectifier and a H-bridge, and drive the data at +-12V :)
2025-11-06T21:43:56 < jpa-> but that incurs at least 1V drop and 2W loss in the rectifier
2025-11-06T21:43:58 < qyx> yeah like model trains
2025-11-06T21:44:04 < qyx> DCC
2025-11-06T21:44:08 < qyx> also one way only
2025-11-06T21:44:20 < jpa-> ah, that too
2025-11-06T21:44:38 < PhantomWork> jpa-: need bidirectional, and that is where that plan fail
2025-11-06T21:44:57 < jpa-> so are all the connectors on the edge channels that you need to handle?
2025-11-06T21:45:00 < qyx> just use bluetooth
2025-11-06T21:45:07 < qyx> like bikers do
2025-11-06T21:45:12 < qyx> for their shifters, etx.
2025-11-06T21:45:15 < PhantomWork> BT would then require that we FCC it
2025-11-06T21:45:43 < qyx> do BT at 0 dBm
2025-11-06T21:45:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T21:46:45 < zyp> okay, given those constraints, PoDL is probably the way to go
2025-11-06T21:47:27 < PhantomWork> I've seen power over RS485 too
2025-11-06T21:47:38 < PhantomWork> but that would require ground I think
2025-11-06T21:47:52 < jpa-> PhantomWork: can you clarify the amount of channels and what is the current draw per channel?
2025-11-06T21:49:01 < PhantomWork> 2 devices, 3ft between them via a quick coupler, 2 wires, 12V max 2A (more like 0.5A but want headroom for in the future)
2025-11-06T21:50:53 < jpa-> you could probably do it with just a series inductor on either end and capacitor to couple it into UART (with pull-up to set the DC level)
2025-11-06T21:51:04 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 255 seconds]
2025-11-06T21:51:47 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T21:53:03 < jpa-> actually best would be to have comparator with hysteresis for the RX side
2025-11-06T21:53:38 < jpa-> you would probably need to arrange for some robust ACK scheme to tolerate packet loss in case there are any load current spikes that corrupt transfers
2025-11-06T21:55:26 < ventYl> if you are nuts enough, you could use the protocol for room thermostats
2025-11-06T21:55:44 < ventYl> it is voltage + current mode, over just two wires
2025-11-06T21:55:50 < jpa-> you can prototype it quite easily, just put e.g. 22µH in series with V+ on both sides, couple with 100nF from GPIO output, send data and observe on scope at the opposite end
2025-11-06T21:57:29 < jpa-> for higher level protocol, you could e.g. have each side send checksumed packets right after they receive a valid packet from the other end, and if one side doesn't get a new packet for X milliseconds, it resends its previous packet
2025-11-06T21:57:42 < ventYl> powered end communicates via voltage fluctuation, unpowered end communicates via current fluctuation
2025-11-06T21:58:37 < jpa-> that works too, especially if you have more space on the powered end; the unpowered end would be pretty simple in that style of communication
2025-11-06T21:59:11 < ventYl> yeah, just some shunt driven by transistor controlled by UART
2025-11-06T21:59:19 < PhantomWork> ventYl: problem is that the current consumption vary quite alot, which make that not that possible
2025-11-06T22:00:14 < ventYl> PhantomWork: that's not a problem. you define a cap on power consumption of your device, e.g. 500mA. Anything below that is default bus state
2025-11-06T22:00:34 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T22:00:39 < ventYl> then your communication activates shunt which adds load higher than max. design consumption
2025-11-06T22:01:25 < ventYl> so if bus is in non-default state, the consumption is always larger than the max. design consumption of the device itself, otherwise it is always same or lower
2025-11-06T22:01:30 < jpa-> (and the current pulses can be very short)
2025-11-06T22:01:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 250 seconds]
2025-11-06T22:02:40 < PhantomWork> ventYl: that would be an excessive heat rejection, and would cause issues
2025-11-06T22:03:51 < ventYl> hm, OK, using 2A for communication, you could probably radio that directly to the pitlane
2025-11-06T22:04:48 < PhantomWork> but I did considered it... until there was a variable load added...
2025-11-06T22:04:51 < ventYl> well, whoever decidede those two wires did a terrible job
2025-11-06T22:05:26 < PhantomWork> more like a legacy issue and parts availability
2025-11-06T22:05:41 < ventYl> we are doing it like this for 15 years
2025-11-06T22:05:43 < ventYl> understood
2025-11-06T22:05:45 < PhantomWork> it's 2 wires, or we design something ourself from scratch
2025-11-06T22:12:37 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T22:12:48 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T22:12:50 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-06T22:13:10 < nomorekaki> hello steff
2025-11-06T22:15:39 < PhantomWork> and question: rs485.... is simply serial with a different transducer right? nothing special on the µC side right?
2025-11-06T22:16:03 < zyp> it needs a TXEN signal
2025-11-06T22:16:25 < zyp> hmm, I think DE is the usual name
2025-11-06T22:16:35 < PhantomWork> that is software controlled I guess
2025-11-06T22:16:54 < zyp> no, the newer STM32 UARTs can do it in hardware
2025-11-06T22:17:05 < zyp> not sure if f4 can though
2025-11-06T22:17:14 < ventYl> rs485 is using individual TX and RX signals, RS422 is one differential pair, IIRC
2025-11-06T22:17:23 < zyp> wrong
2025-11-06T22:17:35 < PhantomWork> ventYl: that is after the transducer
2025-11-06T22:18:14 < zyp> ventYl, when people use the terms like that, it's the other way around, but it's still incorrect, because those are different voltage/current specs
2025-11-06T22:18:39 < zyp> rs485 can be both single pair half duplex or dual pair full duplex
2025-11-06T22:19:02 < qyx> whats the difference between rs485 full diplex and rs422 then
2025-11-06T22:19:59 < zyp> IIRC the load specs are different, RS485 supports more devices on a bus
2025-11-06T22:20:59 < qyx> wat, wodering much that opening c&k datasheet actually opens a littelfuse one
2025-11-06T22:21:09 < qyx> then I realize there was some acquisition
2025-11-06T22:22:21 < zyp> don't you see that every day?
2025-11-06T22:23:20 < zyp> I looked at LT and micrel datasheets earlier today, despite them being acquired by analog and microchip respectively
2025-11-06T22:24:14 < zyp> and then you have infineon who add a «cypress is now infineon» page in front of all the cypress datasheets
2025-11-06T22:30:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T22:30:24 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 265 seconds]
2025-11-06T22:36:44 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T22:42:49 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-06T22:43:31 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-06T22:48:15 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-06T22:56:43 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T22:56:57 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-06T23:04:08 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-06T23:12:46 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T23:21:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T23:23:06 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T23:28:03 < mawk> how do I compute temperature from a NTC thermistor if I don't have the datasheet
2025-11-06T23:28:20 < mawk> it seems to be 10k at 25⁰C so maybe it's the same parameters as all the common 10k NTC
2025-11-06T23:30:04 < mawk> then I have to find the A B C coefficients for the whatever-whatever equation and compute it
2025-11-06T23:30:55 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-06T23:34:45 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-06T23:37:04 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T23:39:11 < qyx> beta is enough for most cases
2025-11-06T23:42:03 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-06T23:42:36 < mawk> right
2025-11-06T23:43:12 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-06T23:44:24 < mawk> I used lm324 as the difference amplifier, with not very precise resistors but it still seems somewhat accurate 
2025-11-06T23:44:42 < mawk> I compute 22.9⁰C and wall temperature is 22⁰C according to my thermostat
2025-11-06T23:45:03 < mawk> but it's close to the nominal 25⁰C so the A/B/C or β could be wrong
2025-11-06T23:45:09 < mawk> I need to try in ice water
2025-11-06T23:47:03 < karlp> do you have any reason to believe the wall thermostat is more accurate than your measurement?
2025-11-06T23:47:23 < qyx> haha
2025-11-06T23:47:53 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-06T23:49:31 < ventYl> wall thermostat may as well have some rather strong dampening
2025-11-06T23:52:43 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-06T23:53:31 < PhantomWork> hu ho.... -3C expected tonight... that mean I need to bring in the pressure washer, and disconnect the garden hose...
2025-11-06T23:53:42 < karlp> 17:39 <ventYl> after gaining experience with Renesas E2 and learning that nucleo boards worth $12 have debugger on-board, I lost all interest into AVR
2025-11-06T23:53:49 < karlp> 10000% this.
2025-11-06T23:58:01 < karlp> ventYl: that wall thermostat thing sounds a lot like mbus.
--- Day changed Fri Nov 07 2025
2025-11-07T00:00:30 < ventYl> karlp: its highly probable it will be the same or very similar thing
2025-11-07T00:01:15 < qyx> lonworks
2025-11-07T00:03:24 < mawk> maybe karlp 
2025-11-07T00:03:30 < mawk> it has a display to 0.5°C
2025-11-07T00:03:36 < mawk> so I would hope it's accurate to 0.5°C
2025-11-07T00:05:48 < karlp> lol
2025-11-07T00:06:09 < karlp> I thought you had learnt enough lessons by now :)
2025-11-07T00:06:25 < karlp> also lol, coilcraft sendign me ads for "low inductance power inductors"
2025-11-07T00:06:29 < karlp> like wot?
2025-11-07T00:08:20 < ventYl> the actual way how I learned about nucleos was STM sales reps and engineers coming over to $company around 2017 and giving us presentation on how awesome their microcontrollers are
2025-11-07T00:08:40 < mawk> I have a thermocouple but it's in my basement in the lead bucket
2025-11-07T00:08:43 < ventYl> and after like 1.5 hours, they asked what particular industry are we doing our business in
2025-11-07T00:08:44 < mawk> I have to decontaminate it
2025-11-07T00:09:01 < ventYl> we told them we do automotive and they were all like: O_o
2025-11-07T00:10:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T00:10:28 < mawk> with egg whites to detect remaining lead contamination
2025-11-07T00:16:08 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-07T00:27:49 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T00:29:21 < qyx> what is a cheap ST accelerometer which is not a pita to work with
2025-11-07T00:29:34 < qyx> don't be lazy and open st.com
2025-11-07T00:30:52 < qyx> ok, closing it
2025-11-07T00:31:55 < qyx> LIS2HH12 apparently but it is 2x2
2025-11-07T00:32:44 < qyx> oh LIS331 is still a thing
2025-11-07T00:32:46 < qyx> I have some
2025-11-07T00:33:47 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T00:42:07 < karlp> "The external RTC clock source enhances
2025-11-07T00:42:09 < karlp> timing accuracy and consequently decreases average power consumption, without impacting functionality
2025-11-07T00:42:11 < karlp> what a great line!
2025-11-07T00:48:25 -!- NEYi [~NEYi@195.234.78.187] has quit [Read error: Connection reset by peer]
2025-11-07T01:05:25 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 264 seconds]
2025-11-07T01:08:34 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T01:16:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-07T01:30:14 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-07T01:31:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T01:37:36 -!- BrainDamage [~m-t6k752@user/BrainDamage] has quit [Remote host closed the connection]
2025-11-07T01:39:44 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-07T01:40:56 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T01:44:50 -!- BrainDamage_ [~m-t6k752@user/BrainDamage] has joined ##stm32
2025-11-07T01:50:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T01:58:01 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-07T01:59:42 -!- BrainDamage_ is now known as BrainDamage
2025-11-07T02:05:01 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-07T02:06:29 < mawk> the H7 board suddenly doesn't work anymore with the same code as before, it hardfaults in random places, the ADC doesn't start, peripherals shut down by themselves
2025-11-07T02:06:32 < mawk> very peculiar
2025-11-07T02:07:05 < mawk> I blame ESD
2025-11-07T02:07:22 < mawk> I'd like to take a look at the master clock output from the stlink
2025-11-07T02:10:33 < nomorekaki> found new scifi I have not seen
2025-11-07T02:10:50 < nomorekaki> existenz (1999)
2025-11-07T02:19:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T02:23:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-07T02:24:48 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-07T02:30:39 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T02:32:14 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-07T02:35:48 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T02:46:54 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T02:58:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-07T02:59:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T03:04:53 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T03:04:58 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-07T03:05:50 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-07T03:08:54 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-07T03:11:43 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-07T03:12:02 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T03:12:12 < jbo> that is very good scifi nomorekaki
2025-11-07T03:17:42 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T03:28:27 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T03:37:17 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T03:40:33 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-07T03:47:14 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T03:47:42 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-07T03:53:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-07T04:07:04 < englishman> i liked it
2025-11-07T04:07:58 < englishman> filmed in canada
2025-11-07T04:15:10 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T04:19:39 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T04:49:46 -!- ferdna [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-07T05:05:24 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has quit [Ping timeout: 260 seconds]
2025-11-07T05:55:47 -!- HelloShitty [~psysc0rpi@2001:8a0:d043:1600:478:7dff:fe45:a43d] has joined ##stm32
2025-11-07T06:01:51 -!- HelloShitty [~psysc0rpi@2001:8a0:d043:1600:478:7dff:fe45:a43d] has quit [Remote host closed the connection]
2025-11-07T06:10:22 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-07T06:12:31 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has joined ##stm32
2025-11-07T06:43:42 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has quit [Ping timeout: 244 seconds]
2025-11-07T06:56:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T07:02:52 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T07:08:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T07:19:31 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-07T07:39:16 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T07:47:18 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-07T07:47:21 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T08:09:15 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T08:14:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-07T08:20:04 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-07T08:25:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T08:30:56 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-07T08:31:28 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-07T08:37:41 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T08:43:03 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T08:47:12 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-07T08:50:47 -!- c10ud__ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has quit [Ping timeout: 260 seconds]
2025-11-07T08:54:31 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T08:57:26 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-07T08:59:19 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T09:05:06 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T09:09:58 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-07T09:10:35 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Read error: Connection reset by peer]
2025-11-07T09:11:33 < Steffanx> Do you like Waterworld (1995), jbo ?
2025-11-07T09:16:00 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T09:21:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T09:22:34 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T09:25:45 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-07T09:34:10 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 246 seconds]
2025-11-07T10:19:54 < mawk> yes Steffanx 
2025-11-07T10:23:11 < Steffanx> I'm not sure about that. I think he never watch it.
2025-11-07T10:28:47 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 260 seconds]
2025-11-07T10:30:14 < mawk> idk I watched it
2025-11-07T10:33:47 < mawk> the H7 spontaneously breaking is really weird
2025-11-07T10:34:08 < mawk> there's no visible piece of wire shorting anything, no pints are bent far enough to short, it was just where I had left it the day before
2025-11-07T10:34:38 < mawk> but now I get peripherals shutting down back again after enabling their clocks in RCC, and busfaults about random memory addresses in reserved area that are nowhere to be found in the code
2025-11-07T10:35:01 < mawk> I tried with the 8MHz MCO from the stlink as a clock source, I tried with the HSI, the CSI, I tried with a lower clock speed, still all the same
2025-11-07T10:38:15 < mawk> and it seems like .bss is not being zero'd out either, but neither the startup.s file nor the linker file changed
2025-11-07T10:38:46 < mawk> oh
2025-11-07T10:38:53 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-07T10:38:55 < mawk> cubemx wrote the .bss section twice in the linker script
2025-11-07T10:38:57 < mawk> interesting
2025-11-07T10:39:20 < mawk> I'll never trust again anything cube-shaped
2025-11-07T10:45:42 < mawk> yes that was it
2025-11-07T10:45:48 < jpa-> regarding peripherals shutting down, maybe your decoupling is insufficient?
2025-11-07T10:46:49 < mawk> it was due to .bss not being zeroed out, now it seems all good
2025-11-07T10:46:57 < mawk> weird how that small thing did all that
2025-11-07T10:48:21 < ventYl> interesting, was it one-off, or it does emit .bss twice all the time?
2025-11-07T10:48:51 < mawk> I remember the same kind of bug a long time ago in cubemx where it would duplicate sections of the makefile, but it's the first time I see it in the linker script
2025-11-07T10:48:57 < mawk> it doesn't do it again when I generate it again
2025-11-07T10:50:46 < jpa-> maybe one of its magic headers "don't edit here" was slightly changed and it thought the whole section was missing?
2025-11-07T10:52:17 < mawk> there's no "USER CODE BEGIN" stuff in the linker script, so I assumed it regenerates the whole thing everytime
2025-11-07T11:02:23 < ventYl> at least CMakeLists are only generated once
2025-11-07T11:30:52 < mawk> I have to use this MDMA stuff if I want to interleave the ADC
2025-11-07T11:31:09 < mawk> I don't get why both master and slave couldn't fit in the common data register, it's 16 bits each and the register is 32 bits
2025-11-07T11:32:05 < mawk> apparently it's possible but the reference manual doesn't want me to
2025-11-07T11:32:46 < jpa-> isn't that exactly what the common data register does?
2025-11-07T11:33:33 < mawk> it's only valid if the slave conversion finishes before the new master conversion restarts
2025-11-07T11:33:55 < jpa-> ah
2025-11-07T11:33:59 < mawk> so if it's in one-shot mode it's fine, but in continuous mode if a master starts converting before the slave finishes then it's not good
2025-11-07T11:34:46 < mawk> maybe just the slave finishing before the master ends its new conversion is good enough
2025-11-07T11:36:21 < jpa-> https://i.imgur.com/10nPHqI.png figure 193 seems like you could read it from CDR (if you ignore the injected stuff in that diagram)
2025-11-07T11:36:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T11:37:20 < mawk> yeah if the timing is right like they show
2025-11-07T11:37:43 < jpa-> yeah, DMA needs to be able to read it quickly enough
2025-11-07T11:38:05 < mawk> that should be fine, the core frequency is 480MHz and the ADC runs at 12MHz
2025-11-07T11:38:09 < jpa-> yeah
2025-11-07T11:39:51 < mawk> currently each adc takes 1.5 ticks to sample and 8.5 to convert, and the slave starts 1.5 ticks after the master, so it should be fine
2025-11-07T11:40:01 < mawk> but I should maybe space the acquisitions evenly instead
2025-11-07T11:40:16 < jpa-> i don't see why you wouldn't space them evenly
2025-11-07T11:41:00 < mawk> yeah that's the plan, I just left 1.5 ticks of delay as the default for now
2025-11-07T11:42:55 < jpa-> also it really helps if you can spare a little bit more than the minimum sampling time.. i had TLV9064 as buffer with just 9 mm traces to LQFP100 STM32H7, and the sampling ripple was significantly higher than any other noise source
2025-11-07T11:43:33 < mawk> well I lowered the ADC frequency to what is in the application note to minimize that, but maybe I shouldn't entirely trust ST on that
2025-11-07T11:47:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-07T12:00:59 < mawk> idk what they mean by "it's mandatory to use the MDMA mode", how do I even configure that
2025-11-07T12:03:41 < mawk> I'll just ignore that
2025-11-07T12:05:17 < jpa-> for simultaneous dual modes at least, regular DMA1 worked fine for me
2025-11-07T12:07:42 < jpa-> i think they have mistypen "MDMA mode" when they really mean "set ADC_CCR DMADF bits to use the CDR register"
2025-11-07T12:08:43 < mawk> right
2025-11-07T12:08:46 < mawk> yeah it's not the only typo
2025-11-07T12:09:01 < mawk> right on the same page they wrote CCR register instead of CDR register
2025-11-07T12:10:46 < ventYl> Germans made something simple...
2025-11-07T12:10:57 < ventYl> ...third day spent dealing with something that used to be 10 clicks
2025-11-07T12:10:57 < ventYl> :>
2025-11-07T12:11:29 < mawk> if my input is already from a low impedance source does that help reduce the ringing?
2025-11-07T12:11:45 < mawk> maybe I can even use the builtin opamp, then there's less resistance overall as it's internally connected
2025-11-07T12:11:56 < mawk> in buffer mode
2025-11-07T12:14:55 < mawk> the buffer is too big to fit in DTCM, do I need special things to disable the cache or whatever?
2025-11-07T12:15:24 < mawk> I can also make the buffer smaller I guess
2025-11-07T12:18:00 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T12:26:41 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 250 seconds]
2025-11-07T12:29:56 < mawk> yeah I need to invalidate the cache indeed
2025-11-07T12:30:01 < mawk> let's move it to DTCM then
2025-11-07T12:30:43 < mawk> or I can setup the MPU to set the region as non cacheable
2025-11-07T12:35:28 < jpa-> or you can just invalidate it
2025-11-07T12:38:37 < jpa-> i'm not sure whether the built-in opamp would make things better or worse, but worth trying
2025-11-07T12:43:20 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-07T12:51:18 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T12:52:52 < mawk> so now theoretically I have 2Msps which is the maximum for this package
2025-11-07T12:53:23 -!- teknix [~unknown@user/hsv] has quit [Quit: leaving]
2025-11-07T12:53:42 < qyx> are you being speculative
2025-11-07T12:54:53 < jpa-> mawk: with interleaved conversion, you can also have the other ADC sampling for the whole duration the other is converting
2025-11-07T12:55:42 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-07T12:57:37 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-07T12:58:50 < mawk> qyx: idk maybe there will be ringing or whatever, but ST says this should be good
2025-11-07T12:59:08 < mawk> it's a bit hard to visualize the data, there's lots of it and just a UART going to the computer
2025-11-07T12:59:36 < mawk> right jpa- 
2025-11-07T12:59:50 < jpa-> the package limit is primarily a limit on the absolute sample time (in nanoseconds); so with two ADCs in interleaved mode, you can have higher sample time and higher samplerate
2025-11-07T13:02:51 < mawk> you mean have a sample rate inbetween the maximum for 1 adc and the maximum for 2 adcs to increase fidelity?
2025-11-07T13:06:18 < jpa-> hmm, ok, the application note suggests it is the vref settling time that limits with multiple ADCs
2025-11-07T13:06:31 < jpa-> though that says "When multiple ADCs convert simultaneously", not interleave
2025-11-07T13:08:05 < mawk> on page 16 they list the "simultaneous" and "interleaved" as having the same maximum sample rate for UFBGA176
2025-11-07T13:08:12 < mawk> so I assume it's the same for the other packages
2025-11-07T13:08:23 < jpa-> AN5354: "In 16-bit mode, UFBGA176 the sampling rate can reach up to 3.5 Msamples/s per ADC, meaning that system throughput using 2 ADCs can reach 7 Msamples/s. The user can also convert one channel using dual-interleaved mode with a sampling rate of 7 Msamples/s."   when the table lists the 3.5 MSps; and for LQFP100 it lists 1.8 MSps, so i'd interpret it as you can have 3.6 MSps in interleaved mode
2025-11-07T13:09:56 < mawk> it's LQFP144 on the nucleo-144 no? I vaguely tried counting the pins and it seemed it was 144
2025-11-07T13:10:10 < jpa-> ah, ok
2025-11-07T13:10:27 < jpa-> and if you have the long traces from nucleo to pin header etc, it will be crappy in every case
2025-11-07T13:11:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T13:12:10 < mawk> yeah it's not very close, the pin is on the bottom left of the chip then goes through a bunch of vias to the header on the top right
2025-11-07T13:12:41 < mawk> then the opamp might help
2025-11-07T13:13:03 < jpa-> true
2025-11-07T13:13:50 < mawk> and it's lucky, the opamp output happens to be on one of the fast channels of ADC1
2025-11-07T13:14:07 < mawk> all the direct channels are on ADC3 for some reason so I can't use them, but they're not faster than the fast channels here anyway
2025-11-07T13:16:13 < mawk> the opamp input is not a fast channel though, idk if that matters
2025-11-07T13:19:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T13:28:43 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-07T13:35:14 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-07T13:38:33 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-07T13:48:55 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 240 seconds]
2025-11-07T13:51:56 < mawk> uuuuuh
2025-11-07T13:52:01 < mawk> the LED is connected to the opamp output
2025-11-07T13:52:03 < mawk> annoying
2025-11-07T13:52:11 < mawk> I need to obliterate some solder bridges
2025-11-07T13:52:26 < mawk> let me get my oyster knife
2025-11-07T14:19:51 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-07T14:20:01 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-07T14:35:41 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-07T15:27:33 < Steffanx> Don't die mawk
2025-11-07T15:43:59 < jbo> Don't die Steffan
2025-11-07T16:12:16 < mawk> you don't like oysters Steffanx ?
2025-11-07T16:13:33 < Steffanx> I never had one. So I don't know
2025-11-07T16:13:40 < mawk> what
2025-11-07T16:13:44 < mawk> what are you waiting for
2025-11-07T16:14:15 < Steffanx> A moment to think: let's eat some oysters
2025-11-07T16:14:25 < Steffanx> I never had that moment yet
2025-11-07T16:27:34 < mawk> at Christmas 
2025-11-07T16:27:38 < mawk> it's Christmas food 
2025-11-07T16:27:40 < mawk> and snails too
2025-11-07T16:27:42 < Steffanx> Meh
2025-11-07T16:27:42 < mawk> you should try it
2025-11-07T16:27:53 < mawk> you can't say meh if you haven't tried it
2025-11-07T16:28:32 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-07T16:31:22 < Steffanx> I say "meh" to trying :P
2025-11-07T16:37:19 < mawk> is it realistic to try to transfer all that data in realtime jpa- 
2025-11-07T16:37:34 < mawk> it's 32Mbps
2025-11-07T16:37:48 < mawk> it should be very compressible but there's no time to do that
2025-11-07T16:42:55 < mawk> there's ethernet on the board 
2025-11-07T16:42:57 < mawk> with a PHY
2025-11-07T16:43:05 < mawk> but it looks super annoying to setup
2025-11-07T16:48:30 < mawk> what happens if I put a voltage that's above VREF but below AVDD on an ADC input, does it just max out? does it explode?
2025-11-07T16:58:16 < PhantomWork> should just max out
2025-11-07T17:08:15 < jpa-> yeah, i think it will max out, or at least no damage (it might not show exactly the maximum if it is too far outside the range)
2025-11-07T17:09:26 < jpa-> mawk: depending on board you might also have usb high speed
2025-11-07T17:09:40 < jpa-> there is plenty of time for compression too
2025-11-07T17:09:42 < mawk> I thought you needed a PHY
2025-11-07T17:10:02 < jpa-> which nucleo do you have?
2025-11-07T17:10:22 < mawk> well there's 30ms to process like 130KiB
2025-11-07T17:10:31 < mawk> NUCLEO-H743ZI
2025-11-07T17:10:43 < mawk> the obsolete one with the weird A/B usb port that can't power the boar
2025-11-07T17:11:26 < jpa-> i calculate that you have 100 cycles per sample or something, that's more than enough to calculate delta between samples and encode it as variable length integer
2025-11-07T17:12:09 < jpa-> if delta less than +-127, write one byte; otherwise write -128 and then 2 bytes sample value
2025-11-07T17:12:43 < jpa-> immediate almost 50% compression ratio with maybe 10 instructions per sample
2025-11-07T17:13:58 < jpa-> not sure it helps much though, because if you only have USB full speed, it'll be around 10 Mbps real throughput
2025-11-07T17:14:02 < mawk> this https://www.st.com/resource/en/user_manual/um1974-stm32-nucleo144-boards-mb1137-stmicroelectronics.pdf
2025-11-07T17:14:04 < mawk> right
2025-11-07T17:14:06 < mawk> I have 500 instructions per sample 
2025-11-07T17:14:08 < mawk> maximum
2025-11-07T17:14:42 < mawk> there's also 100M ethernet 
2025-11-07T17:15:12 < jpa-> sending raw ethernet packets out of that isn't so difficult
2025-11-07T17:15:19 < mawk> but I guess I can afford to drop a few samples, or rather only send the running total to the PC every N samples
2025-11-07T17:16:48 < jpa-> if you have e.g. some fx2 board or ft232h, you could dump the data to it
2025-11-07T17:17:48 < mawk> I might have one somewhere 
2025-11-07T17:21:34 < jbo> let's play "why is my shit not working" again
2025-11-07T17:21:37 < mawk> my temporal resolution is 500ns, and the board I'm measuring has a bunch of inductors and capacitors in the power path, I should probably be fine with not missing spikes in current
2025-11-07T17:22:34 < mawk> idk if I should add some filtering to average the input before it goes to the ADC
2025-11-07T17:23:46 < mawk> there are a bunch of SMPS on the board under measurement, so maybe it will add some sort of spikiness
2025-11-07T17:42:45 < mercenary> tell me again why you are re-inventing a picoscope?
2025-11-07T18:16:36 < jpa-> mawk: yes, you should put a filter that cuts out most stuff on frequencies over samplerate/2
2025-11-07T18:24:04 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 246 seconds]
2025-11-07T18:35:48 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has joined ##stm32
2025-11-07T18:38:08 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-07T19:08:36 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 252 seconds]
2025-11-07T19:10:32 < mawk> it's a coulomb counter mercenary not a scope
2025-11-07T19:10:45 < mawk> I asked bossman to buy an otii arc
2025-11-07T19:10:52 < mawk> and I'm having fun in the meantime 
2025-11-07T19:15:26 < mercenary> ah. for that there are off-the-shelf solutions also yes, besides a full-blown SMU for some project we had little modules keep track of battery state, with rs485 to read out the data.
2025-11-07T19:42:56 < mawk> you mean with battery voltage?
2025-11-07T19:43:04 < mawk> this is NiMH so voltage is kinda useless 
2025-11-07T19:46:01 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-07T19:46:08 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-07T19:47:22 < mawk> the only inductor I have is a giant through hole inductor, that will probably work for the filter 
2025-11-07T19:49:30 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 245 seconds]
2025-11-07T19:51:16 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-07T19:51:16 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-07T19:59:34 < mercenary> not just voltage, those modules had a series-shunt and proper coulomb-counter. Not very cheap from what I recall, but it was one of those 'the spec says you shall'
2025-11-07T20:02:28 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has joined ##stm32
2025-11-07T20:05:18 < mawk> I wonder what's the sample rate of a coulomb counter
2025-11-07T20:05:47 < mawk> maybe they use tricks like a current mirror charging a capacitor to theoretically never lose any data
2025-11-07T20:05:54 < mawk> I was thinking about a circuit like that
2025-11-07T20:05:58 < mawk> I need to try it
2025-11-07T20:06:08 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-07T20:14:41 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-07T20:14:41 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-07T20:14:59 < mercenary> sort of, if I understand https://www.analog.com/media/en/technical-documentation/data-sheets/ltc3337.pdf correctly.
2025-11-07T20:15:30 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-07T20:24:10 < mawk> hmm 
2025-11-07T20:24:23 < mawk> seems like the trick is to keep the battery crurent output either Ipeak or 0
2025-11-07T20:24:26 < mawk> using a supercap
2025-11-07T20:24:39 < mawk> then calculating the crurent is easy
2025-11-07T20:26:59 < mawk> smart
2025-11-07T20:28:24 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-07T20:56:01 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-07T20:56:33 < nomorekaki> movie recommend Steffanx: ExistenZ
2025-11-07T20:57:15 < nomorekaki> if you like that 90s dreamy dark scifi
2025-11-07T21:02:02 < Steffanx> Will I be ok?
2025-11-07T21:16:44 < nomorekaki> yes
2025-11-07T21:16:51 < nomorekaki> it's not scary
2025-11-07T21:18:14 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-07T21:18:14 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-07T21:18:15 < qyx> huh a cat on the window, but it looks like treated with sodium chlorite
2025-11-07T21:18:39 < qyx> I don't want a white-flaked cat here
2025-11-07T21:21:27 < qyx> sorry hypochlorite
2025-11-07T21:27:03 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-07T21:29:35 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-07T21:30:37 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 264 seconds]
2025-11-07T21:33:48 -!- duude__- is now known as duude__
2025-11-07T21:49:31 -!- flatmush [~benbrewer@149.71.27.16] has quit [Ping timeout: 246 seconds]
2025-11-07T21:58:43 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-07T21:58:43 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-07T22:01:18 -!- flatmush [~benbrewer@149.57.117.232] has joined ##stm32
2025-11-07T22:50:29 < mawk> poor kitty
2025-11-07T22:50:37 < mawk> let him in
2025-11-07T23:10:12 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-07T23:42:39 -!- NEYi [~NEYi@195.234.78.187] has quit [Read error: Connection reset by peer]
--- Day changed Sat Nov 08 2025
2025-11-08T00:14:58 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-08T00:25:43 -!- PlasmaHH [~PlasmaHH@user/plasmahh] has quit [Ping timeout: 240 seconds]
2025-11-08T00:42:57 < Steffanx> No
2025-11-08T01:17:56 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-08T01:17:56 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-08T01:34:42 -!- _Posterdati_ [~posterdat@user/Posterdati] has quit [Ping timeout: 252 seconds]
2025-11-08T01:43:28 < jbo> nomorekaki, did you enjoy existenz?
2025-11-08T01:43:34 < jbo> existenz is very good movie
2025-11-08T01:43:46 < jbo> hard time believing that Steffanx didn't see it yet :o
2025-11-08T01:45:57 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has quit [Read error: Connection reset by peer]
2025-11-08T01:53:50 -!- jmcgnh [~jmcgnh@wikipedia/jmcgnh] has joined ##stm32
2025-11-08T02:02:31 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-08T02:20:21 < nomorekaki> it was refreshing
2025-11-08T03:47:15 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-08T04:25:26 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Read error: Connection reset by peer]
2025-11-08T06:01:23 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-08T06:05:14 < ColdKeybo[a]rd> Damn RP2350 reads GPIO30 as low when DMM and scope show it as high...
2025-11-08T06:05:57 < ColdKeybo[a]rd> And I don't know if it's again some SDK BS where you have to define something to make "new" pins work properly...
2025-11-08T06:06:57 < ColdKeybo[a]rd> They really messed everything up with hard-coding everything for RP2040 and then when they released 2350 with higher pin-count, there are so many if-defs in the sdk :)
2025-11-08T06:23:37 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-08T06:40:38 < ds2> is it that complicated to read a GPIO that you need an SDK?
2025-11-08T06:58:04 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-08T06:59:41 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has joined ##stm32
2025-11-08T07:02:46 -!- nuxil [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 244 seconds]
2025-11-08T07:35:16 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-08T07:44:55 -!- qyx [~qyx@84.245.120.196] has quit [Ping timeout: 245 seconds]
2025-11-08T07:46:46 -!- ferdna [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-08T07:47:00 -!- qyx [~qyx@84.245.120.241] has joined ##stm32
2025-11-08T08:11:07 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-08T08:13:32 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-08T08:54:23 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-08T09:46:18 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-08T10:13:42 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-08T10:50:22 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-08T11:11:22 < Steffanx> There's a lot of movies I have not seen jbo 
2025-11-08T11:51:37 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-08T12:38:51 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-08T13:19:38 -!- ernest [~ernest@102.90.103.114] has joined ##stm32
2025-11-08T14:12:05 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-08T14:19:11 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-08T14:46:00 -!- ernest [~ernest@102.90.103.114] has quit [Quit: WeeChat 3.2]
2025-11-08T15:18:56 < mawk> jpa-: it's actually worse with the opamp
2025-11-08T15:19:07 < mawk> maybe it's because it's not rail to rail, I have to check
2025-11-08T15:20:51 < mawk> yeah it's not rail to rail
2025-11-08T15:20:55 < mawk> that's a problem
2025-11-08T16:20:04 < mawk> VDDA is all over the place on this nucleo board, maybe I shouldn't power it from USB
2025-11-08T16:46:53 < mawk> wtf is an injected channel
2025-11-08T16:47:15 < mawk> hmm
2025-11-08T16:47:18 < mawk> I think I see
2025-11-08T17:17:17 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-08T17:18:41 < Steffanx> Is seeing something you're not sure about mawk?
2025-11-08T17:19:51 < mawk> yes Steffanx 
2025-11-08T17:20:00 < mawk> take optical illusions for instance 
2025-11-08T17:20:12 < mawk> you think you see something but it's not there
2025-11-08T17:43:48 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Ping timeout: 252 seconds]
2025-11-08T17:46:29 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-08T18:56:25 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-08T20:27:03 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-08T20:36:37 -!- kurfen [~kurfen@45.84.137.20] has joined ##stm32
2025-11-08T20:42:00 -!- kurfen [~kurfen@45.84.137.20] has quit [Ping timeout: 252 seconds]
2025-11-08T20:48:34 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-08T20:49:20 < Steffanx> I see
2025-11-08T20:51:45 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-08T21:06:46 < jpa-> mawk: injected channels are when you want to measure something other in middle of a main measurement sequence
2025-11-08T21:06:53 < mawk> right 
2025-11-08T21:07:07 < mawk> I was thinking of using that to periodically remeasure VREFINT
2025-11-08T21:07:16 < mawk> but then the timing gets all messed up
2025-11-08T21:07:54 < jpa-> and why bother, can't you just use VREFINT as ADC VREF?
2025-11-08T21:08:08 < mawk> it's not high enough it's 2.5V max
2025-11-08T21:08:27 < jpa-> ah, you need to measure higher?
2025-11-08T21:08:29 < mawk> that corresponds to 1.25A current and it might go higher
2025-11-08T21:09:21 < jpa-> use a smaller shunt? :)
2025-11-08T21:09:37 < mawk> I need high precision at low current values (because the device spends most of its time sleeping) but also high range because peak currents might be high
2025-11-08T21:10:00 < mawk> it's a 0.1Ω shunt, but then the difference amplifier has a fixed gain of 20
2025-11-08T21:10:51 < mawk> I can maybe add a second shunt in parallel, I have a bunch of 0.1Ω±0.1% resistors
2025-11-08T21:12:18 < mawk> if the board power supply is stable enough it should be fine to measure it only once, I need to try with an actual lab PSU and not USB from a computer
2025-11-08T21:15:50 < mawk> I wonder why ST measures VREFCAL with only 16 bits, while you can get 26 bits with oversampling
2025-11-08T21:16:23 < mawk> I measure VREFINT with oversampling and divide it by 1024 to get the (floating point) value for VREFINT
2025-11-08T21:35:55 < jpa-> 16 bits is already 15ppm precision, while the thermal drift uncertainty is +-75ppm/°C
2025-11-08T21:37:47 < jpa-> and even high quality references have more than 20ppm drift per year
2025-11-08T21:41:56 < mawk> right
2025-11-08T21:42:25 < mawk> I wonder what happens if I measure VREFINT in parallel with ADC3 once in a while
2025-11-08T21:43:51 < mawk> how much samples get messed up
2025-11-08T21:44:34 < mawk> the application note says I should lower the sample rate to 1.4Msps if the three ADCs are running
2025-11-08T21:56:30 < qyx> why the hell are you sampling your coloumbs at >2 Msps?
2025-11-08T21:57:39 < qyx> that's 6 decimal degrees of magnitude more than some commercially available ICs do
2025-11-08T21:58:53 < qyx> bq35100 does 2-4 Hz effective sample rate iirc
2025-11-08T22:16:31 < mawk> the current can vary pretty quickly
2025-11-08T22:17:10 < mawk> the coulomb counters I've looked at use a totally different method that cannot miss spikes, by drawing a constant current from the battery and using a supercap to supply the load
2025-11-08T22:19:20 < mawk> I just chose the maximal resolution first and then just set the maximal sample rate
2025-11-08T22:19:39 < mawk> oversampling with a lower sample rate would do the same thing in the end as I'm just adding up data
2025-11-08T22:21:04 < mawk> bq35100 seems to be for super low power stuff, the datasheet lists 500μA as being "high power"
2025-11-08T22:43:56 < qyx> I would just filter it and sample at about 1 ksps
2025-11-08T22:51:27 < mawk> well I'll certainly downsample it before sending it to a PC for graphing, but nothing else is happening on the MCU so I can just sample at the max rate to avoid all the delay caused by a filter with low cutoff frequency
2025-11-08T22:51:43 < mawk> and it's for diagnostic purposes, like to see what happens when each part of the code is running, so good temporal resolution is good
2025-11-08T22:52:30 < mawk> it's not for long term monitoring
2025-11-08T23:48:54 < mawk> my "stm32l432kc" has markings that say "l432c6" which doesn't seem to exist
2025-11-08T23:50:31 < mawk> I guess C is for 256KiB flash and 6 is for -40°C to 85°C
--- Day changed Sun Nov 09 2025
2025-11-09T00:28:07 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-09T02:07:55 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-09T02:32:30 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Quit: Bye]
2025-11-09T02:36:34 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-09T03:44:57 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-09T04:11:03 -!- h4x0riz3d [~pewpew@antonsavov.net] has joined ##stm32
2025-11-09T04:13:29 -!- ilgrim2 [~ilgrim@xinu.me] has joined ##stm32
2025-11-09T04:14:07 -!- ThatDamnRanga_ [~ThatDamnR@UNASSIGNED-218-100-26-71.3cix.nzix.net] has joined ##stm32
2025-11-09T04:17:25 -!- Netsplit *.net <-> *.split quits: ilgrim, splud, ThatDamnRanga, ds2, antto
2025-11-09T06:58:06 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Read error: Connection reset by peer]
2025-11-09T07:31:18 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-09T07:57:07 -!- splud [~noneya.bi@user/splud] has joined ##stm32
2025-11-09T08:14:12 -!- ds2 [~ds2@69.54.141.155] has joined ##stm32
2025-11-09T08:29:36 -!- ferdna [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-09T09:23:31 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-09T09:40:52 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-09T09:41:07 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 264 seconds]
2025-11-09T09:42:45 -!- duude__- is now known as duude__
2025-11-09T10:22:18 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Read error: Connection reset by peer]
2025-11-09T10:23:38 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-09T10:39:46 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-09T10:46:10 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-09T11:27:43 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-09T11:42:22 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-09T12:17:47 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-09T12:33:22 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-09T14:17:45 -!- nuxil_ [~nuxil@telia-5908fd-87.connect.netcom.no] has quit [Ping timeout: 244 seconds]
2025-11-09T14:45:56 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-09T15:03:16 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-09T16:01:13 < mawk> kitten doesn't like spicy oil
2025-11-09T16:01:30 < mawk> that will teach him to steal in my plate when I'm not looking 
2025-11-09T16:18:03 < ventYl> aluminium foil tends to help
2025-11-09T16:32:49 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Quit: Bye]
2025-11-09T16:38:35  * qyx covering windows with alu foil
2025-11-09T16:45:52 < mawk> lol
2025-11-09T16:45:59 < mawk> why do you hate that poor cat
2025-11-09T16:47:09 < qyx> there is only one cat with the privilege to look through my windows
2025-11-09T18:26:47 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-09T19:14:51 < mawk> such feliphobia
2025-11-09T19:15:00 < mawk> inadmissible 
2025-11-09T19:47:50 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-09T20:01:08 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-09T20:41:04 < qyx> zyp: ADIN3310/ADIN6310
2025-11-09T20:41:12 < qyx> bye SJA1105
2025-11-09T20:41:37 < qyx> up to 4x SGMII
2025-11-09T20:41:45 < qyx> 3/6 port MAC switch
2025-11-09T20:41:58 < qyx> uhm 33 eur in singles
2025-11-09T20:43:05 < qyx> with SPI or ethernet host interface
2025-11-09T20:43:12 < qyx> the product I have been dreaming about
2025-11-09T20:46:56 < mercenary> but does it have a publicly available datasheet?
2025-11-09T20:47:05 < zyp> https://www.analog.com/media/en/technical-documentation/data-sheets/adin3310-6310.pdf
2025-11-09T20:47:15 < qyx> yes
2025-11-09T20:47:19 < qyx> yuite surprisingly
2025-11-09T20:47:48 < zyp> bga196 seems a bit overkill for a 3-port switch :p
2025-11-09T20:47:52 < mercenary> yay. that's a new experience
2025-11-09T20:48:04 < qyx> half of it is ground
2025-11-09T20:48:45 < zyp> hmm, also 256 for the 6-port
2025-11-09T20:49:19 < qyx> whoa 150 page software guide available too
2025-11-09T20:49:30 < qyx> and a official reference design
2025-11-09T20:50:45 < qyx> I need a T1L-T1L to T1S switch for field use, 4 ports basically
2025-11-09T20:51:00 < qyx> considering implementing it in software using two SPI MAC-PHYs
2025-11-09T20:51:12 < zyp> gigabit switch seems kinda overkill for that
2025-11-09T20:51:20 < qyx> ADIN1112 is actually a hw switch between the two ports
2025-11-09T20:51:32 < qyx> *ADIN2111
2025-11-09T20:51:52 < qyx> so I only need software switching between that and a T1S MAC-PHY
2025-11-09T20:52:41 < zyp> you don't even need switching, you can just do direct forwarding
2025-11-09T20:52:57 < zyp> since you already have a switch
2025-11-09T20:52:58 < qyx> most probably yes
2025-11-09T20:53:13 < qyx> the T1L backbone is switched to the branch in hardware
2025-11-09T20:53:36 < zyp> yeah, I mean you already have a three port switch, 
2025-11-09T20:54:01 < zyp> if it was *MII instead of SPI, you could just hook it up directly to a PHY
2025-11-09T20:56:07 < nomorekaki> Steffanx: should we watch some imdb 4.5/10 movies?
2025-11-09T20:59:55 < nomorekaki> Copolla's flopped Megalopolis
2025-11-09T21:05:08 < Phantom> what do you guys use to protect a GPIO from a short to 12V and ground?
2025-11-09T21:05:16 < Phantom> one that can do 1Mbps...
2025-11-09T21:06:30 < nomorekaki> is it intended to work as input or as output
2025-11-09T21:06:39 < nomorekaki> or both
2025-11-09T21:07:26 < nomorekaki> usually when you do 12v tolerant input output you use 2 GPIOs
2025-11-09T21:08:26 < nomorekaki> one being input and one output
2025-11-09T21:08:39 < nomorekaki> Phantom: 1Mbps?
2025-11-09T21:09:22 < nomorekaki> sounds like you could use one of those automotive tranceivers for one wired bus
2025-11-09T21:09:22 < Phantom> one of each direction actually, and yeah, data of ~1Mbps (actually around 640kHz)
2025-11-09T21:09:57 < Phantom> CAN tranceiver don't work, there is a DC component that cause it to disable itself
2025-11-09T21:10:17 < Phantom> same for LIN
2025-11-09T21:10:38 < nomorekaki> how much DC offset?
2025-11-09T21:11:06 < Phantom> not dc offset, but it idle low (busy) for too long, it cause the CAN tranceiver to go in protection mode
2025-11-09T21:13:28 < nomorekaki> can you make it not busy for too long
2025-11-09T21:13:57 < nomorekaki> because it gets complicated if you cannot use can or lin tranceiver
2025-11-09T21:15:29 < Phantom> I know...
2025-11-09T21:15:31 < nomorekaki> maybe some RSxxx tranceiver has suitable features?
2025-11-09T21:15:59 < Phantom> that is actually the last option I found that may be working
2025-11-09T21:16:26 < Phantom> like a max232, but it uses too much board space
2025-11-09T21:16:35 < nomorekaki> damn
2025-11-09T21:16:49 < nomorekaki> we have board space constraints as a bonus
2025-11-09T21:17:50 < nomorekaki> cant you get those RSxxx tranceivers in nifty TSSOP packages that are just 4x6mm or something
2025-11-09T21:17:50 < Phantom> https://i.imgur.com/VMcinY8.png <=== the IC there is no good, so I have that place. that board is 22mm diameter
2025-11-09T21:18:09 < Phantom> traces there are 0.3mm
2025-11-09T21:19:33 < nomorekaki> analog has some 3x4mm QFP
2025-11-09T21:20:07 < nomorekaki> https://www.analog.com/en/resources/technical-articles/tiny-rs232-transceivers-run-directly-from-batteries.html
2025-11-09T21:21:34 < Phantom> I'll take a look again at the 232 tranceivers, seems to be the best bet
2025-11-09T21:21:53 < nomorekaki> you need to browse a bit
2025-11-09T21:21:55 < Phantom> if I can fit the parts on the board, that's it
2025-11-09T21:22:38 < nomorekaki> if you have constraints you may not be able to use legacy part ids from 80s
2025-11-09T21:29:46 < Phantom> it'S annoying that there is not much protection IC out there
2025-11-09T21:33:52 < nomorekaki> given your board size constraints there is no option other than having a tranceiver chip
2025-11-09T21:35:20 < nomorekaki> single chip converting whatever bus voltage levels you need to comply with to logic levels
2025-11-09T21:35:35 < nomorekaki> and short protection
2025-11-09T21:35:39 < nomorekaki> for driver
2025-11-09T21:36:10 < nomorekaki> higher ESD rating probs also
2025-11-09T21:37:17 -!- fenugrec_ [~f@192.214.232.39] has left ##stm32 []
2025-11-09T21:38:33 < Phantom> ESD is actually not my concern, but miswiring is
2025-11-09T21:48:34 < nomorekaki> yes.. anyways a tranceiver IC is your protection IC what you are looking for and most minimal viable solution and only solution given the board contraint
2025-11-09T21:48:56 < nomorekaki> just browse datasheets
2025-11-09T21:49:24 < nomorekaki> *the most *the only
2025-11-09T22:06:53 < qyx> can/lin phy going into protection mode is a feature, not a bug
2025-11-09T22:07:28 < qyx> and a can IC definitely can be used by a 1 Mbaud UART
2025-11-09T22:07:36 < qyx> *CAN transceiver
2025-11-09T22:07:57 < qyx> because async UART has stop bits which are logic high, so reseting the protection
2025-11-09T22:14:23 -!- rgz [uid670983@user/rgz] has joined ##stm32
2025-11-09T22:15:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-09T22:29:09 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-09T22:45:14 < nomorekaki> Phantom: qyx has pro words for you
2025-11-09T22:46:07 < nomorekaki> it's almost pro o'clock
2025-11-09T22:47:05 < qyx> yes and also an empty head o'clock
2025-11-09T22:47:47 < nomorekaki> so unwarranted pro words start coming out
2025-11-09T23:21:32 < Phantom> qyx: that's the thing, it is for a remote IRDA tranceiver, which the TX idle LOW... While the Rx idle high... yeah... the Tx is inverted compared to Rx...   which made the CAN tranceiver not usable
2025-11-09T23:21:53 < qyx> that's sad
2025-11-09T23:22:10 < Phantom> Also, because of the delays in both transceivers, it is also unusable if I just invert the signal. I tried.
2025-11-09T23:22:52 < Phantom> This is why I'm investigating data over powerline, to get rid of the irda part, add the protection and all
2025-11-09T23:53:21 < nomorekaki> why IRDA
2025-11-09T23:53:36 < nomorekaki> I remember when IRDA was hot shit
2025-11-09T23:54:24 < nomorekaki> Nokia phones had that turquoise lens for IRDA
2025-11-09T23:55:14 < nomorekaki> trasmit data 115kilobits per second and more
2025-11-09T23:56:52 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
--- Day changed Mon Nov 10 2025
2025-11-10T00:03:50 -!- NEYi [~NEYi@195.234.78.187] has quit [Read error: Connection reset by peer]
2025-11-10T00:20:11 < Phantom> nomorekaki: due to some design limitations, I only have 2 pins for data + power, and a semi-transparent plastic around the 2 pins. I can't change the connection, so it is data over power, or power + light
2025-11-10T00:23:09 < nomorekaki> for data over DC you need CMC's and shit
2025-11-10T00:23:12 < nomorekaki> so light it is
2025-11-10T00:23:53 < Phantom> I checked some data over DC stuff. but they are low power or uses lots of board space due to the magnetics
2025-11-10T00:26:36 < Phantom> but also, IRDA is annoying due to the pulse width, it gives the equivalent of iirc 640kbps, which causes issues for most rs232 transceiver, which often stop at 115200...
2025-11-10T00:28:08 < nomorekaki> what you need that tranceiver for again
2025-11-10T00:28:21 < nomorekaki> also can you replace that with optics all the way
2025-11-10T00:28:39 -!- rgz [uid670983@user/rgz] has quit [Quit: Connection closed for inactivity]
2025-11-10T00:28:48 < Phantom> need to be disconnectable, and that's the issue
2025-11-10T00:29:10 < nomorekaki> what connector
2025-11-10T00:30:40 < Phantom> racing steering hub quick connector
2025-11-10T00:30:52 < nomorekaki> cool
2025-11-10T00:31:56 < Phantom> V1 used an remote control receiver, effective datarate is 1100bps and maximum 3 bytes followed by a pause
2025-11-10T00:32:09 < nomorekaki> why there is only 2 contacts?
2025-11-10T00:32:36 < nomorekaki> arent those racing hubs with 4 or more per anything automotive
2025-11-10T00:32:39 < nomorekaki> power and bus
2025-11-10T00:32:53 < Phantom> it used to be for the horn only
2025-11-10T00:33:23 < Phantom> then for the data they use an external spiral cable with connector
2025-11-10T00:33:48 < qyx> sounds like a clever idea
2025-11-10T00:33:59 < nomorekaki> can you use steering tube for optical path
2025-11-10T00:35:18 < Phantom> nope, the column do a T (where the steering is at the top)   so no direct path
2025-11-10T00:36:17 < nomorekaki> I dont still understand the need for rsXXX tranceivers if you have irda both side on steering wheel and steering tube
2025-11-10T00:36:30 < nomorekaki> you connect that directly to uart pins
2025-11-10T00:37:03 < Phantom> the problem is the wiring from the "box" to the coupler with IRDA
2025-11-10T00:37:32 < Phantom> that wiring can be messed up, and currently is directly connected to the STM32. Any short = dead chip
2025-11-10T00:37:44 < nomorekaki> that shouldn't be a problem
2025-11-10T00:37:50 < nomorekaki> why it is a problem
2025-11-10T00:37:56 < Phantom> and the IRDA transducer is even more sensitive
2025-11-10T00:38:08 < Phantom> so it need to have some form of protection
2025-11-10T00:38:31 < nomorekaki> but there should not be a posibility to mess up the wiring
2025-11-10T00:38:32 < Phantom> need to handle short to gnd and 12V
2025-11-10T00:38:51 < Phantom> sadly, there is... 
2025-11-10T00:39:32 < Phantom> and... some are idiot that start messing with the pinout when things don't work as it is supposed to be... 
2025-11-10T00:39:46 < Phantom> yes, it happened with V1
2025-11-10T00:39:50 < Phantom> twice
2025-11-10T00:40:34 < nomorekaki> but you dont have wire count constraint in nothing but the steering wheel coupling?
2025-11-10T00:41:04 < Phantom> the first one, he played with the pinout, and got it wrong. Fortunately he fried only the "inexpensive" side.    the second connected it wrong... fried everything...
2025-11-10T00:41:24 < nomorekaki> also just put MCU or inverter or something next to IRda
2025-11-10T00:41:33 < Phantom> no space
2025-11-10T00:41:53 < Phantom> I have a 22mm circle of space
2025-11-10T00:41:55 < nomorekaki> well fire that fool from the team
2025-11-10T00:42:32 < nomorekaki> :o
2025-11-10T00:42:44 < Phantom> https://i.imgur.com/I5aRhGy.png   https://i.imgur.com/LuTGees.png   current unprotected circuit for the coupler
2025-11-10T00:42:48 < nomorekaki> it's you isn't it?
2025-11-10T00:42:58 < nomorekaki> your boss?
2025-11-10T00:43:28 < Phantom> nomorekaki: one is a client, one is a mechanics that had paper + video + voice call   to connect it up...
2025-11-10T00:43:57 < nomorekaki> there shouldn't be wrong way to connect it
2025-11-10T00:44:10 < nomorekaki> without using violence
2025-11-10T00:44:50 < Phantom> there is at the vehicle side of the column, it ends up in a 8 pins connector (it also pick up 4 wires from 2 hall sensors)
2025-11-10T00:45:13 < nomorekaki> ah the vehicle side is the problem
2025-11-10T00:45:17 < Phantom> yeah
2025-11-10T00:45:32 < nomorekaki> I see it uses vehicle harness
2025-11-10T00:46:02 < Phantom> the problem is that there is another connector near it that is the same... with 12V all over...
2025-11-10T00:46:18 < qyx> oh poka yoke failed
2025-11-10T00:46:25 < nomorekaki> but if you use PLC you don't need to have magnetics inside that 22mm?
2025-11-10T00:46:30 < qyx> very sinful in automotive
2025-11-10T00:46:38 < Phantom> that is what the mechanics did... he ignored the instructions, disconnected too many things, mixed things up..
2025-11-10T00:46:53 < qyx> no, he did not
2025-11-10T00:47:10 < qyx> there shall not be two same, unkeyed, uncoded connectors
2025-11-10T00:47:22 < qyx> the whole automotive knows that
2025-11-10T00:47:53 < nomorekaki> if you use PLC is there a board size constraint or can you move everything out of that hub both sides?
2025-11-10T00:49:08 < nomorekaki> but it's horn wiring so
2025-11-10T00:49:22 < nomorekaki> not likelly to be ideal twisted pair
2025-11-10T00:49:30 < nomorekaki> one end might go to chassis
2025-11-10T00:50:06 < Phantom> PLC, the electronics on the 22mm board dissapear, in fact, the whole board dissapear, and 2 ring terminals is all what is needed.   On the steering side, the board is bigger, and miiigh have space. On the vehicle side, the board is single sided, so should have space, plus I can clean up stuff to make space
2025-11-10T00:50:23 < Phantom> the wiring on the column is new
2025-11-10T00:50:34 < Phantom> so not an issue up to the last 2" or so
2025-11-10T00:50:57 < Phantom> and on the steering side, it is bolted to the coupler connector
2025-11-10T00:51:04 < Phantom> so no real wire there
2025-11-10T00:51:34 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-10T00:58:20 < nomorekaki> anyways there is really small logic ICs for inverting shit
2025-11-10T00:58:55 < nomorekaki> size of grain of sugar or something
2025-11-10T01:00:00 < nomorekaki> mcus size of 3x3 or something
2025-11-10T01:00:16 < nomorekaki> 5V MCUs like AVR
2025-11-10T01:01:19 < Phantom> yeah, but I have not found any so far that tolerate short to 12V
2025-11-10T01:01:27 < nomorekaki> your regulator can be replaced by grain of sugar too
2025-11-10T01:01:46 < nomorekaki> Phantom: between RSxxx and irda
2025-11-10T01:01:52 < Phantom> the IRDA is power hungry, 300mA for the IR led, plus self consumption
2025-11-10T01:02:39 < qyx> it is power hungry for multiple reasons
2025-11-10T01:02:52 < nomorekaki> I have done 1.5amp switched regulator in same size board and had MCU in addition. altough it didn't have 2 holes in the middle
2025-11-10T01:03:05 < qyx> wide angle, daylight operation, long reach
2025-11-10T01:03:30 < qyx> for 10 cm reach in total darkness inside the tube, a common low power LED wohld work
2025-11-10T01:03:31 < nomorekaki> I think there are smaller IRDA tranceivers too
2025-11-10T01:03:44 < Phantom> the 2 holes in the middle is a royal pain...
2025-11-10T01:04:09 < Phantom> it really restrict the number of traces that can pass between them
2025-11-10T01:04:18 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-10T01:05:00 < qyx> and you don't even need to use the pulsed 3/16 modulation
2025-11-10T01:05:27 < Phantom> the 3/16 is done by the stm32
2025-11-10T01:05:43 < qyx> it wohld probably work directly attached to a uart output if your protocol was DC bias free
2025-11-10T01:06:46 < nomorekaki> Phantom: https://drive.google.com/file/d/1rf5F5IxeHpUr4P40-LzszazcBTcY8YUX/view?usp=sharing 18mm or something and one sided
2025-11-10T01:06:48 < qyx> I think you are not being adventurous enough
2025-11-10T01:06:59 < ventYl> Phantom: honestly, I feel that the QD gives you so much problems, you should remove it with hammer long ago
2025-11-10T01:07:01 < qyx> you are doing a fully proprietry thing
2025-11-10T01:07:08 < nomorekaki> 1.5A switching regulator adjustable with 0.5A steps and MCU
2025-11-10T01:07:17 < ventYl> some stuff used for years is simply not worth of keeping around anymore
2025-11-10T01:07:59 < qyx> also, I would definitely try half duplex comms using leds "only"
2025-11-10T01:08:04 < qyx> *only*
2025-11-10T01:08:11 < qyx> without photodiodes
2025-11-10T01:08:22 < qyx> also, you can cap-couple the comms
2025-11-10T01:08:44 < qyx> or do a nfc, that is, induction couple using small coils
2025-11-10T01:09:01 < nomorekaki> ventYl: what is QD?
2025-11-10T01:09:01 < qyx> neither of them require pcb to be i side the tube
2025-11-10T01:09:04 < Phantom> NFC require FCC stuff
2025-11-10T01:09:43 < qyx> you will not call it nfc
2025-11-10T01:09:45 < ventYl> nomorekaki: quick disconnect, the coupling they use
2025-11-10T01:09:54 < qyx> it is "communication over cojpled inductor"
2025-11-10T01:09:57 < ventYl> which was probably sourced in '70s
2025-11-10T01:10:08 < qyx> *coupled, sorry
2025-11-10T01:10:20 < Phantom> NFC = radio = FCC
2025-11-10T01:10:21 < nomorekaki> ventYl: it's inexpensive or something?
2025-11-10T01:10:42 < ventYl> nomorekaki: it's there. and automotive is all about "we are doing it like this for 15 years and so far it was working"
2025-11-10T01:11:14 < qyx> Phantom: quick google shows that nfc under 1w is exempt
2025-11-10T01:11:17 < nomorekaki> there are racing quick things that have like 9pin connectors in middle
2025-11-10T01:11:20 < qyx> how can you be that sure?
2025-11-10T01:11:21 < nomorekaki> or 4pin
2025-11-10T01:11:58 < nomorekaki> just make bluetooth steering Phantom
2025-11-10T01:12:12 < Phantom> nomorekaki: FCC
2025-11-10T01:12:25 < nomorekaki> with precertified modules
2025-11-10T01:13:31 < nomorekaki> if I have understood correctly you can get modules that you just use and it's all kosher
2025-11-10T01:13:45 < qyx> an induction coupling in a contained enclosure is at max an unintentionak radiator
2025-11-10T01:14:02 < qyx> otherwise all smps power supllies would need fcc too
2025-11-10T01:14:55 < Phantom> anyway, I'll dig it up this week, for now it's still sunday
2025-11-10T01:15:37 < nomorekaki> Phantom: look at my inspirational pic
2025-11-10T01:16:43 < Phantom> nomorekaki: yeah, but you don't have the 2 huge screws and 2 connectors to deal with
2025-11-10T01:18:14 < ventYl> nomorekaki: BTW, after more than decade, I have resurrected the project of board computer
2025-11-10T01:18:40 < ventYl> I couldn't find any information on how consumption is communicated by the ECU, so I scoped it and now I am a sad panda
2025-11-10T01:18:53 < nomorekaki> board?
2025-11-10T01:19:12 < ventYl> instrument cluster computer
2025-11-10T01:19:14 < ventYl> for car
2025-11-10T01:19:22 < nomorekaki> cluster
2025-11-10T01:19:36 < nomorekaki> so it's analog or what did you find?
2025-11-10T01:20:04 < ventYl> it's, kind-of PCM
2025-11-10T01:20:16 < ventYl> you have one pulse each ~150ms as a base
2025-11-10T01:20:31 < ventYl> and then more pulses in between these base pulses as the consumption increases
2025-11-10T01:21:03 < ventYl> you'll get two or three, separated ~10ms, then long nothing and then the cycle repeats 150ms after previous one started
2025-11-10T01:21:17 < nomorekaki> so we are in early.. 90s?
2025-11-10T01:21:27 < ventYl> more like mid-80s
2025-11-10T01:22:47 < ventYl> early '90s reportedly used to simply route the injector opening signal into cluster and cluster did the math
2025-11-10T01:23:00 < ventYl> this thing doesn't know how "much" injectors are opened
2025-11-10T01:27:23 < nomorekaki> why are you sad panda
2025-11-10T01:28:50 < ventYl> interwebs stated that this signal is similar to "vacuum" gauge even older cars had. so i'll have to do quite complicated math
2025-11-10T01:29:12 < ventYl> most importantly, i'll probably have to multiply consumption by tacho signal, which I don't currently have
2025-11-10T01:30:25 < nomorekaki> toyota in 90s had just green light that said economy or something
2025-11-10T01:33:23 < qyx> lol
2025-11-10T01:34:56 < nomorekaki> https://www.hagerty.com/media/lists/12-of-our-favorite-retro-digital-dashboards-volume-one/
2025-11-10T01:35:46 < ventYl> my commie car has just green stripe on tacho gauge between 2k and 3.2k RPM where engine is most efficient
2025-11-10T01:36:11 < nomorekaki> I think economy light was effectivelly the same
2025-11-10T01:36:15 < ventYl> yeah
2025-11-10T01:36:34 < nomorekaki> https://www.youtube.com/watch?v=bquL6HwayRU&t=31s this is from 80s btw
2025-11-10T01:38:10 < nomorekaki> real time gauges and shit in
2025-11-10T01:38:15 < ventYl> '80s gave us the first in-car touch interface
2025-11-10T01:38:34 < ventYl> it was so bad that after one or two years they thrown it out and replaced it with classic mechanical gauges
2025-11-10T01:40:45 < nomorekaki> if you look at that video some car factories struggled in hecking 2010s to get same responsiveness from their digital gauges
2025-11-10T01:41:29 < ventYl> i can't say i am surprised to read that :)
2025-11-10T01:41:50 < nomorekaki> and look at that boot time
2025-11-10T01:42:21 < nomorekaki> 1seconds to fully operational
2025-11-10T01:43:45 < qyx> I will do a tft lcd dash for mi tractor using stm32h7 and mipi
2025-11-10T01:43:52 < qyx> *my
2025-11-10T01:43:58 < qyx> in 2038
2025-11-10T01:44:05 < nomorekaki> first do the tractor
2025-11-10T01:44:18 < qyx> nooo
2025-11-10T01:44:26 < ventYl> this instrument cluster computer had some issues with latency
2025-11-10T01:44:28 < nomorekaki> then the decorations
2025-11-10T01:44:54 < ventYl> then I realized that arduino code I stolen for the I2C display was so bad it was redrawing screen four times for just one frame
2025-11-10T01:45:04 < ventYl> *stole
2025-11-10T01:45:15 < nomorekaki> typical .ino
2025-11-10T01:46:01 < nomorekaki> does random shit that worked just sufficiently for the original author
2025-11-10T01:46:13 < ventYl> yeah, qyx recommended some other IMBA library but it is roughly equal horrendeous and it also comes from the world of arduino
2025-11-10T01:47:36 < Phantom> … they just salted the road…    IT'S TOO EARLYYYYYYYYYY!!!! :(
2025-11-10T01:48:32 < nomorekaki> I guess it will happen here tonight too
2025-11-10T01:48:59 < nomorekaki> -4C outside or something and it rained before
2025-11-10T01:49:49 < nomorekaki> luckilly my car has misfire so I will cover toyota work van with salt if needed
2025-11-10T01:51:03 < nomorekaki> that is more expensive than my car but it desires to be utilized and abused to extent
2025-11-10T01:51:35 < ventYl> honda misfired too. turned out that cap contact in distributor decided it doesn't like close proximity of rotating contact
2025-11-10T01:51:45 < nomorekaki> when you jump on and off that seat you can make grunting work man noises
2025-11-10T01:52:27 < nomorekaki> I think I have plugpack failure
2025-11-10T01:53:08 < nomorekaki> I messed with the plugpack once and put it back and it worked after that
2025-11-10T01:53:18 < ventYl> plugpack? CoP?
2025-11-10T01:53:22 < nomorekaki> yes
2025-11-10T01:53:38 < ventYl> these things fail over time
2025-11-10T01:53:50 < nomorekaki> everything ignition related does
2025-11-10T01:54:22 < nomorekaki> it's all consumables
2025-11-10T01:55:19 < nomorekaki> I don't know if it helps to bake coil inside head
2025-11-10T01:55:29 < ventYl> most probably not
2025-11-10T01:55:55 < nomorekaki> it has done reasonably 280kkm
2025-11-10T01:56:03 < ventYl> if frame mounted, coils are usually immortal, switches can sometimes die
2025-11-10T01:56:38 < ventYl> bonus feature of such coils is that you can use them for self-defense
2025-11-10T01:56:51 < ventYl> if thrown or used as weight on stick
2025-11-10T01:57:02 < nomorekaki> those things rust and fall off at some point
2025-11-10T01:57:27 < nomorekaki> altough there is not much car around that coil at that point
2025-11-10T01:59:11 < nomorekaki> you still need to replace HV cables and plugcaps.. those must be cheap though
2025-11-10T02:00:10 < ventYl> I am nearing the stage where I simply buy a few meters of cable and raw caps and make cables on my own
2025-11-10T02:00:59 < ventYl> older cable sets had 5 different-length cables, so everything fit perfectly. last set I bought had two different lengths, so now there's HT cable spaghetti under the bonnet
2025-11-10T02:01:53 < nomorekaki> like in the olden days
2025-11-10T02:02:27 < ventYl> except for motorbike I never actually did that
2025-11-10T02:02:55 < ventYl> also, despite this car is an old crap by now, it already has "carbon" HT wires
2025-11-10T02:03:29 < ventYl> i have no clue how feasible is to DIY those wires
2025-11-10T02:05:01 < nomorekaki> idk
2025-11-10T02:05:08 < nomorekaki> they sell parts though
2025-11-10T02:05:43 < ventYl> that's a problem of future-me
2025-11-10T02:05:46 < nomorekaki> just connect the parts
2025-11-10T03:54:27 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-10T04:21:32 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-10T04:53:18 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-10T04:53:36 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 252 seconds]
2025-11-10T04:55:16 -!- duude__- is now known as duude__
2025-11-10T05:13:46 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-10T05:14:07 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 240 seconds]
2025-11-10T05:15:40 -!- duude__- is now known as duude__
2025-11-10T06:34:05 -!- dobson [~dobson@static.38.6.217.95.clients.your-server.de] has quit [Ping timeout: 245 seconds]
2025-11-10T06:44:26 -!- dobson [~dobson@static.38.6.217.95.clients.your-server.de] has joined ##stm32
2025-11-10T06:59:31 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-10T07:44:55 -!- qyx [~qyx@84.245.120.241] has quit [Ping timeout: 240 seconds]
2025-11-10T07:47:04 -!- qyx [~qyx@84.245.120.40] has joined ##stm32
2025-11-10T08:14:15 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-10T08:38:46 -!- h4x0riz3d is now known as antto
2025-11-10T10:03:40 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 245 seconds]
2025-11-10T10:05:45 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-10T13:13:39 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-10T14:03:19 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-10T14:26:32 -!- nerozero [~nerozero@87.253.63.54] has quit [Remote host closed the connection]
2025-11-10T14:28:37 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-10T14:49:39 < qyx> oh also zyp AD3306 but only preview
2025-11-10T15:02:13 < zyp> what's that? just a single port SPI MAC-PHY?
2025-11-10T15:02:24 < zyp> how does it differ from other SPI MAC-PHYs?
2025-11-10T15:02:49 < qyx> there are not many, microchip has one and onsemi has one
2025-11-10T15:03:25 < qyx> and the onsemi one has no ptp/tsn support
2025-11-10T15:03:41 < qyx> (T1S I mean)
2025-11-10T15:04:04 < zyp> hmm, how useful is TSN on T1S?
2025-11-10T15:04:17 < qyx> I hope it works, I plan to use it
2025-11-10T15:04:24 < qyx> for sync acquisition
2025-11-10T15:04:53 < qyx> eg. distributed accelerometers along a T1L path
2025-11-10T15:05:28 < zyp> yeah, I guess sync accuracy requirements is orthogonal to latency requirements
2025-11-10T15:32:44 -!- PaulFertser [paul@paulfertser.info] has quit [Ping timeout: 256 seconds]
2025-11-10T15:34:28 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-10T15:51:47 < karlp> fucking, reflashing a bunhc of boards, they've all been coated, so the serialpins are gunked up.
2025-11-10T15:52:08 < karlp> the jtag port was somehow excluded from the coat, for... no reason, but its unpopuylated.  
2025-11-10T15:52:20 < karlp> can't hold a connector down, because it has series resistors, which are also unpopulated.
2025-11-10T15:52:26 < karlp> jank tastic.
2025-11-10T15:54:37 < qyx> why populate them, you are never gonna use jtag, are you?
2025-11-10T15:55:00 < karlp> no, I don't want to populate them,
2025-11-10T15:55:06 < karlp> I wanted to use _some_ method of reflashing them, 
2025-11-10T15:55:16 < karlp> either holding a jtag connector on the jtag pads, or using the serial pins
2025-11-10T15:55:24 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-10T15:55:31 < qyx> nah, was it in the accepted requirement list?
2025-11-10T16:05:36 < karlp> of course not.
2025-11-10T16:11:45 < zyp> just get some sharp needle tips that can penetrate through the gunk?
2025-11-10T16:15:26 < karlp> that's what it's going to be, but nont of that is getting done today.
2025-11-10T16:15:30 < karlp> anyway, just a fucking hassle
2025-11-10T16:17:52 < zyp> every day a new hassle :)
2025-11-10T16:49:52 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-10T16:59:29 -!- PaulFertser [paul@paulfertser.info] has joined ##stm32
2025-11-10T17:09:03 -!- ColdKeybo[a]rd [~ColdKeybo@user/coldkeyboard] has quit [Quit: ZNC - https://znc.in]
2025-11-10T17:12:51 -!- ColdKeyboard [~ColdKeybo@user/coldkeyboard] has joined ##stm32
2025-11-10T17:23:20 -!- ColdKeyboard [~ColdKeybo@user/coldkeyboard] has quit [Ping timeout: 240 seconds]
2025-11-10T17:52:06 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-10T18:19:52 -!- ColdKeyboard [~ColdKeybo@user/coldkeyboard] has joined ##stm32
2025-11-10T18:24:50 < karlp> actual solution: figure out why the ethernet based update wasn't working in the first place, and use that....
2025-11-10T18:25:57 -!- joel135 [uid136450@id-136450.hampstead.irccloud.com] has joined ##stm32
2025-11-10T18:59:33 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-10T19:50:04 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-10T20:34:35 -!- joel135 [uid136450@id-136450.hampstead.irccloud.com] has quit [Quit: Connection closed for inactivity]
2025-11-10T20:35:28 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 255 seconds]
2025-11-10T21:20:05 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-10T21:20:05 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-10T21:34:04 < PhantomWork> question guys, for 5V tolerant GPIO, if the pin is set to output, and you backfeed 5V into it, what happen? does it fry or backfeed to the power pin or just "work"?
2025-11-10T21:45:38 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-10T21:47:05 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-10T21:53:19 < jpa-> PhantomWork: an output pin set to high is basically connected to VDD inside the chip, because the FET is on
2025-11-10T21:53:38 < jpa-> so.. if you feed +5V to it through resistor, a bit of current will flow and if VDD load is higher than that, everything is ok
2025-11-10T21:53:58 < jpa-> if you feed +5V to it directly, it's similar to as if you had shorted +5V to your VDD
2025-11-10T21:54:26 < PhantomWork> that is what I tought, :(   thanks
2025-11-10T21:55:33 < zyp> if it were that simple, the pin wouldn't be 5V tolerant because the FET would leak in reverse
2025-11-10T21:55:46 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-10T21:57:44 < jpa-> zyp: hmm, indeed, i wonder what the actual FET implementation is vs. what big ST wants us to believe ( https://stm32world.com/images/2/24/5v_tolerant_pin.png )
2025-11-10T21:58:56 < zyp> I don't think that schematic is correct at all, the FT specs I've looked at says you can't inject positive current into FT pins at all
2025-11-10T22:00:09 < jpa-> yet it is in every stm32 refman
2025-11-10T22:06:50 < zyp> repeating lies doesn't make them true :)
2025-11-10T22:07:42 < jpa-> yep, big ST thinks we are fools
2025-11-10T22:07:46 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-10T22:09:14 < jpa-> but we see through all the cube nonsense and know that their real aim is to hide the two big secrets of the universe: 1) FT IO structure  2) CRC->IDR, why?
2025-11-10T22:10:20 < jpa-> if you don't hear from me after this message, you know why..
2025-11-10T22:11:34 < zyp> speaking of cube, I wanted to see if there's any chips with both UCPD and TSC, and apparently TSC isn't even a filter
2025-11-10T22:45:08 < qyx> zyp: was telling dongs the same today, TSC is not a dedicated peripheral
2025-11-10T22:45:51 < qyx> even for chips marketed as having a TSC it was implemented using the analog switch matrix and some timers
2025-11-10T22:46:13 < jbo> are we hating on ST again?
2025-11-10T22:48:52 < Steffanx> Yeah fuck the engineers who designed the timer
2025-11-10T22:48:55 < Steffanx> s
2025-11-10T22:49:24 < qyx> are you having a bad day Steffanx
2025-11-10T22:49:47 < Steffanx> Lol no
2025-11-10T22:50:20 < jbo> what happened to "language, please", Steffanx ?
2025-11-10T22:50:42 < Steffanx> Who the fuck cares about that, jbo?
2025-11-10T22:50:51 < jbo> what's going on?
2025-11-10T22:50:59 < Steffanx> Pizza?
2025-11-10T22:51:10 < qyx> oh stef's drunken
2025-11-10T22:51:20 < Steffanx> You wish
2025-11-10T22:51:20 < qyx> lol no qyx
2025-11-10T22:51:24 < Steffanx> I'm all good
2025-11-10T22:51:58  * jbo hugs Steffanx 
2025-11-10T22:52:00 < jbo> let it all out buddy
2025-11-10T22:52:08 < Steffanx> \o\
2025-11-10T22:52:20 < jbo>  /o/
2025-11-10T22:52:23 < qyx> also jpa lol'd irl at CRC->IDR
2025-11-10T22:56:21 < Steffanx> I'm glad I was able to confuse you guys once again.
2025-11-10T22:56:28 < Steffanx> Happy days ##stm32 
2025-11-10T23:04:53 < qyx> \o/
--- Day changed Tue Nov 11 2025
2025-11-11T00:00:42 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-11T00:06:29 < qyx> my brother in lay just ordered oneplus nord CE4
2025-11-11T00:06:38 < qyx> just checking it, 6.67" screen
2025-11-11T00:07:15  * qyx looking at his 8" tablet laying on the table
2025-11-11T00:07:15 < specing> huge
2025-11-11T00:07:49 < qyx> sorr lying
2025-11-11T00:15:06 < zyp> I bought a pixel 9 pro fold this summer, quite happy with it
2025-11-11T00:18:09 < jbo> still doing the pixel 6 thing here
2025-11-11T00:18:57 < specing> still doing the pixel 2 thing here
2025-11-11T00:21:02 < zyp> I mean, it's the whole folding thing that I'm quite happy with
2025-11-11T00:21:08 < zyp> that the rest is nice is kinda a given
2025-11-11T00:24:43 < qyx> still doing the sony ericsson here
2025-11-11T00:30:41 < qyx> thing
2025-11-11T00:33:46 < zyp> I've been thinking I would kinda like a tablet, except it'd just be another device that's big enough I wouldn't bother bringing it with me so I actually have it available when I'd like to use it
2025-11-11T00:34:12 < zyp> so I decided to try something I could just fold up and pocket
2025-11-11T01:07:25 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-11T02:03:36 < mawk> my father had the bad idea to have a heart attack today
2025-11-11T02:03:42 < mawk> I had to spend the day at the hospital 
2025-11-11T02:04:05 < qyx> no gud
2025-11-11T02:05:33 < mawk> he's already joking about smoking again
2025-11-11T02:13:15 < jbo> sorry to hear that mawk. is he "alright"?
2025-11-11T02:21:44 < Phantom> mawk: so, he's getting a personal embedded tazer?
2025-11-11T03:32:11 -!- martinmoene [~martinmoe@132.229.46.129] has quit [Read error: Connection reset by peer]
2025-11-11T04:30:20 -!- tabemann [~tabemann@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has quit [Remote host closed the connection]
2025-11-11T04:30:54 -!- tabemann [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-11T06:19:15 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-11T06:26:24 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-11T07:27:05 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-11T07:27:45 -!- joel135 [uid136450@id-136450.hampstead.irccloud.com] has joined ##stm32
2025-11-11T08:24:03 -!- martinmoene [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-11T10:03:07 -!- martinmoene [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-11T10:48:04 < mawk> no Phantom it was a infarctus (a plumbing problem) not the heart stopping (an electrical problem)
2025-11-11T10:48:19 < mawk> so he got little springs in the coronary arteries to keep them openm
2025-11-11T10:48:32 < mawk> yeah jbo 
2025-11-11T10:49:13 < ventYl> maybe he'll end up like my neighbour. he had stent installed before vein could clog completely.
2025-11-11T10:49:30 < mawk> yeah he got 2 stents 
2025-11-11T10:49:38 < ventYl> after coming back from hospital, he lit another cigarette which made him to pass out almost instantly
2025-11-11T10:49:41 < mawk> lol
2025-11-11T10:49:56 < mawk> that's what the doctor told my father, even just one ciggie could be fatal
2025-11-11T10:50:00 < mawk> hopefully he will stop smoking now
2025-11-11T10:50:12 < ventYl> he did not die, just lost conciusness
2025-11-11T10:50:15 < mawk> sadly the doctor didn't tell him to stop the beer
2025-11-11T10:50:36 < ventYl> thus he stopped smoking 5 years ago
2025-11-11T10:50:52 < mawk> ah nice
2025-11-11T10:51:17 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Quit: KVIrc 5.2.6 Quasar http://www.kvirc.net/]
2025-11-11T10:51:37 < mawk> is STM32U3 the lowest power class run-mode mcu ST has? they say it's the "best-in-class in low power"
2025-11-11T10:51:46 < mawk> but apparently it's not the lowest power in stop mode that they have
2025-11-11T10:51:57 < mawk> I guess that depends how much your application is sleeping or running
2025-11-11T10:52:16 < qyx> U5 was lower than U3 iirc
2025-11-11T10:52:55 < Steffanx> Every low power MCU is "best in class"
2025-11-11T10:54:12 < mawk> lol
2025-11-11T10:54:24 < mawk> U3 seems lower in run mode, especially "9.5 μA/MHz Run mode @ 3.3 V (While(1) SMPS step-down converter mode)" whatever that means
2025-11-11T10:54:50 < qyx> that means Vcore is supplied by an internal SMPS instead of a LDO
2025-11-11T10:54:52 < mawk> while a random U5 I selected is slightly higher
2025-11-11T10:54:59 < mawk> no I mean the "while(1)"
2025-11-11T10:55:04 < mawk> the other run modes are also listed as SMPS
2025-11-11T10:55:41 < qyx> oh I read that as "while being in the smps mode" with a (1) footnote :D
2025-11-11T10:55:53 < mawk> ah lol
2025-11-11T10:56:47 < ventYl> mawk: PIC16!!!!!1
2025-11-11T10:56:53 < ventYl> at least that's what linkedin says
2025-11-11T10:58:36 < mawk> not doing 32 bits in 2025 is a capital offense
2025-11-11T10:59:39 < ventYl> you don't understand it. 32-bits in embedded is bloat. C++ is an offense to the art, C is barely tolerable. real pros do all their work in assembly
2025-11-11T10:59:40 < mawk> While(1) appears to be a benchmark like coremark and dhrystone
2025-11-11T11:00:39 < mawk> or a type of code; if we go by the name it's just the CPU doing for (;;) {} and then measure the power consumption
2025-11-11T11:00:44 < zyp> as in while(1) {} ? :p
2025-11-11T11:00:48 < zyp> yeah
2025-11-11T11:01:13 < zyp> «this cpu can busywait very efficiently»
2025-11-11T11:01:53 < mawk> for (;;) is much prettier than while (1)
2025-11-11T11:02:27 < zyp> loop {} is where all the rage is at nowadays
2025-11-11T11:02:33 < mawk> yeah so no additional caching or prefetch or all that
2025-11-11T11:02:46 < mawk> is that a real thing? that sounds like arduinese
2025-11-11T11:02:50 < ventYl> such code will even keep ALU dormant
2025-11-11T11:02:57 < ventYl> most probably
2025-11-11T11:03:10 < zyp> https://doc.rust-lang.org/rust-by-example/flow_control/loop.html
2025-11-11T11:03:51 < mawk> I see
2025-11-11T11:04:41 < mawk> what happens if you accidentally enable VREFBUF in a package where VDDA and VREF+ are connected together, does it explode?
2025-11-11T11:04:53 < mawk> I might have done this
2025-11-11T11:05:04 -!- martinmoene [~martinmoe@132.229.46.129] has joined ##stm32
2025-11-11T11:05:57 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-11T11:06:22 < mercenary> ask hexacube, iirc he has done that. I suspect you will have no more internal reference.
2025-11-11T11:06:57 < mawk> at least the VREFINT thing still works, to calibrate the adc
2025-11-11T11:10:36 -!- martinmoene_ [~martinmoe@132.229.46.129] has joined ##stm32
2025-11-11T11:13:06 -!- travisb_ [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-11T11:14:40 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Ping timeout: 256 seconds]
2025-11-11T11:14:40 -!- martinmoene [~martinmoe@132.229.46.129] has quit [Ping timeout: 256 seconds]
2025-11-11T11:14:40 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-11T11:14:40 -!- tabemann [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has quit [Ping timeout: 256 seconds]
2025-11-11T11:25:52 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-11T11:27:06 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-11T11:50:58 -!- PaulFertser [paul@paulfertser.info] has quit [Read error: Connection reset by peer]
2025-11-11T11:55:09 -!- PaulFertser [paul@paulfertser.info] has joined ##stm32
2025-11-11T12:09:24 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-11T12:11:39 < mawk> stm32l432kc has 20-bit oversampling with the ADC but then it force-truncates it to 16 bits
2025-11-11T12:11:41 < mawk> weird
2025-11-11T12:28:48 < jpa-> you're not really getting more than 16 useful bits out of oversampling a 12-bit ADC anyway
2025-11-11T12:30:12 < jpa-> at such high oversampling the linearity of the ADC really starts to limit things
2025-11-11T12:38:54 < mawk> right
2025-11-11T12:39:06 < mawk> but the H7 can do 26 or even 27 bits oversampling out of 16 bits
2025-11-11T12:39:09 < mawk> are those useless as well?
2025-11-11T13:32:57 < mawk> half of the pins on the nucleo-32 are not usable because of the default tarduino solder bridge configuration
2025-11-11T13:40:31 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-11T13:53:16 < jpa-> mawk: yes, probably H7 oversampling is useful to about 20 bits (oversample by 256x and right-shift 4 bits)
2025-11-11T13:53:27 < mawk> right
2025-11-11T13:54:20 < mawk> on a package that doesn't accept a normal HSE crystal but just wants a single pin master clock input (like from a stlink), what kind of oscillator do you put on the board to do that
2025-11-11T13:54:39 < qyx> the specified one
2025-11-11T13:54:46 < jpa-> anything
2025-11-11T13:55:17 < mawk> if on other chips it wants 2 pins for the oscillator, what do you do with the second pin
2025-11-11T13:55:27 < jpa-> no, not a crystal
2025-11-11T13:55:30 < jpa-> an oscillator
2025-11-11T13:55:44 < jpa-> they are usually 4 pins with one pin either adjust or enable
2025-11-11T13:55:45 < qyx> on other chips you set it to a bypass mode
2025-11-11T13:56:11 < qyx> you want to use a single partno for chips with 2 and 1 pins for clock input?
2025-11-11T13:56:30 < nomorekaki> https://www.youtube.com/watch?v=veL46JDH7IE some mecha anime irl
2025-11-11T13:56:43 < mawk> no I'm just wondering, I always saw CK_IN and CK_OUT for the HSE but the L432KC has only CK_IN
2025-11-11T13:57:08 < qyx> so just use an oscillator with cmos output (no a sine wave one)
2025-11-11T13:57:11 < qyx> as jpa- says
2025-11-11T13:57:15 < mawk> I see
2025-11-11T13:57:20 < mawk> thanks
2025-11-11T13:57:34 < qyx> but it is usually more power hungry
2025-11-11T13:58:40 < jpa-> for really low power, it would make more sense to put in the 32 kHz crystal for which the STM32 has the oscillator, and use CRS to trim the internal oscillator
2025-11-11T13:58:47 < mawk> so like this https://ww1.microchip.com/downloads/aemDocuments/documents/TCG/ProductDocuments/DataSheets/VCC4A-CMOS-Crystal-Oscillator-DS20006749.pdf
2025-11-11T13:59:06 < nomorekaki> how do you migrate your android phones?
2025-11-11T13:59:10 < jpa-> mawk: yeah
2025-11-11T13:59:11 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-11T13:59:24 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-11T13:59:28 < mawk> there's a migration feature builtin nomorekaki 
2025-11-11T13:59:53 < nomorekaki> does it have good results?
2025-11-11T14:00:21 < mawk> depends on what you care about
2025-11-11T14:00:30 < mawk> if it's photos and all that yeah
2025-11-11T14:00:42 < mawk> it will install the same applications, but I don't think it migrates application data
2025-11-11T14:01:02 < nomorekaki> phone numbers, all txt, all whatsapp including media, all telegram, all media
2025-11-11T14:01:12 < mawk> whatsapp and telegram are migrated separately
2025-11-11T14:01:12 < nomorekaki> authenticators are a plus
2025-11-11T14:01:18 < jpa-> it has worked for me
2025-11-11T14:01:24 < mawk> whatsapp is backed up to google drive, telegram I don't remember
2025-11-11T14:01:40 < mawk> if you're coming from an iphone it's a bit trickier to get whatsapp data but it still works
2025-11-11T14:01:45 < nomorekaki> I don't backup to drive
2025-11-11T14:02:05 < nomorekaki> and at this point I wont there is like 30gigs of whatsapp
2025-11-11T14:02:43 < mawk> you can also transfer from phone to phone by proximity with whatsapp apparently
2025-11-11T14:02:43 < nomorekaki> whatsapp is basically my journal
2025-11-11T14:02:51 < mawk> https://faq.whatsapp.com/209942271778103/?cms_platform=android
2025-11-11T14:05:08 < nomorekaki> is there some feature that just transfers files?
2025-11-11T14:05:13 < nomorekaki> quickly
2025-11-11T14:05:24 < nomorekaki> thats how I like to do the migration
2025-11-11T14:05:54 < nomorekaki> make everything into files then move them and import phone numbers and txt and shit
2025-11-11T14:07:07 < nomorekaki> I think I used SD for that
2025-11-11T14:07:23 < specing> adb backup ?
2025-11-11T14:07:35 < nomorekaki> I think I have used that too
2025-11-11T14:07:39 < mawk> HSI seems to have better characteristics than MSI
2025-11-11T14:07:50 < nomorekaki> my new phone has USB2.0 !
2025-11-11T14:08:11 < specing> I use seedvault on a newer phone, but it's a pain to setup server-side and either works or it doesen't (and if it doesen't you have to browse logcat)
2025-11-11T14:08:17 < mawk> if the phone is rooted maybe nomorekaki 
2025-11-11T14:09:10 < nomorekaki> is it viable to transfer 128gigs using usb2.0 and adb?
2025-11-11T14:10:01 < mawk> idk if it works in modern times
2025-11-11T14:10:04 < nomorekaki> I think there was only one actually fast transfer method via usb and it might have been adb
2025-11-11T14:10:05 < mawk> with the encryption stuff
2025-11-11T14:10:23 < qyx> nomorekaki: cx file explorer for example
2025-11-11T14:10:31 < qyx> you can do sshfs
2025-11-11T14:11:00 < qyx> also mtp works
2025-11-11T14:11:23 < qyx> also wifi direct may be actually faster than usb 2.0
2025-11-11T14:11:37 < qyx> but if the phone has 802.11ac/ax it probably also have usb3 of some sorts
2025-11-11T14:12:17 < qyx> anyway, soldering time
2025-11-11T14:12:55 < zyp> the trick is to not have anything on your phone that you care about that's not automatically cloud synced
2025-11-11T14:13:23 < nomorekaki> trick is to have all important stuff in phone not backed up in any way
2025-11-11T14:14:33 < mercenary> the trick is to not have important stuff on a phone
2025-11-11T14:14:40 < zyp> last I switched phones, pretty much everything went automatically, just had to reauthenticate some apps
2025-11-11T14:15:21 < mawk> the whatsapp link I gave uses wifi direct to transfer everything
2025-11-11T14:15:27 < mawk> I don't think you can just adb whatsapp data
2025-11-11T14:15:33 < mawk> it's all "secure"
2025-11-11T14:16:26 < zyp> I don't use whatsapp, but can't it cloud sync like any other chat app?
2025-11-11T14:16:55 < mawk> yeah it can
2025-11-11T14:17:00 < mawk> but nomorekaki is allergic to cloud
2025-11-11T14:17:05 < zyp> pretty much all chat apps I use do and can be logged into from multiple devices
2025-11-11T14:17:22 < zyp> so new phone just means logging into from yet another location
2025-11-11T14:17:22 < mawk> telegram can be fully logged in from another device, but not whatsapp it bridges to your phone
2025-11-11T14:17:30 < mawk> as far as I understand it
2025-11-11T14:17:36 < zyp> that's stupid
2025-11-11T14:17:47 < mawk> it's to ensure the E2E encryption 
2025-11-11T14:17:51 < mawk> maybe they figured out a better way since then
2025-11-11T14:18:17 < nomorekaki> I have done local copy of whatsapp and it works every time if you do it in right order
2025-11-11T14:18:34 < zyp> LINE also has stupid restrictions on logins, IIRC you can only be logged in from one mobile device and one computer at once
2025-11-11T14:18:46 < nomorekaki> it says "hey there is a local backup. want to use it now?"
2025-11-11T14:19:35 < mawk> when did you do that the last time nomorekaki 
2025-11-11T14:19:40 < mawk> the whatsapp website doesn't mention it
2025-11-11T14:19:47 < mawk> and when I did it last time like a year ago I had to use google drive
2025-11-11T14:19:58 < nomorekaki> like 2020
2025-11-11T14:20:00 < zyp> you can have E2E with cloud sync, just need to transfer the private key from device to device
2025-11-11T14:20:05 < BrainDamage> whatsapp does have a backup feature
2025-11-11T14:20:47 < mawk> but local?
2025-11-11T14:21:09 < BrainDamage> I never used it, so I don't know :P
2025-11-11T14:21:55 < zyp> speaking of chat apps; which do you like, which do you dislike?
2025-11-11T14:21:55 < mawk> yeah it's just google drive
2025-11-11T14:22:00 < mawk> no option for local backup in whatsapp
2025-11-11T14:22:05 < mawk> I like signal and telegram
2025-11-11T14:22:25 < mawk> also I used wickr to talk with drug dealers and my mistress
2025-11-11T14:22:30 < zyp> IMO telegram is the worst shit that people have persuaded me to use
2025-11-11T14:22:43 < mawk> I guess I like telegram for the people on it, not really the app itself
2025-11-11T14:22:55 < mawk> it's fake encryption unless you use "secure chat" that only work on mobile and doesn't support group chats
2025-11-11T14:22:58 < BrainDamage> I find them all equally terrible
2025-11-11T14:23:50 < mawk> also telegram keeps sending propaganda push notifications about pavel durov's arrest in france
2025-11-11T14:23:56 < mawk> and how the french government is terrible
2025-11-11T14:24:37 < BrainDamage> I have a matrix - telegram bridge, and I have setup filters for a bit of the nonsense
2025-11-11T14:24:58 < BrainDamage> matrix ofc terrible too
2025-11-11T14:25:28 -!- martinmoene_ [~martinmoe@132.229.46.129] has quit [Quit: Leaving]
2025-11-11T14:25:50 < ventYl> these modern slackoids are mostly all brain-damaged
2025-11-11T14:25:53 < zyp> I rank matrix (or element really, I guess) middle of the tree
2025-11-11T14:26:38 < mawk> what's the top
2025-11-11T14:26:40 < mawk> IRC?
2025-11-11T14:27:07 < BrainDamage> my only use for matrix is as a convergence point for various chat protocols
2025-11-11T14:27:10 < ventYl> element sucks the least (out of those I've been using)
2025-11-11T14:27:15 < ventYl> discord the most
2025-11-11T14:27:27 < BrainDamage> irc, telegram, whatsapp, discord, signal
2025-11-11T14:27:44 < zyp> UI wise, nothing that I've tried beats discord
2025-11-11T14:28:08 < mawk> I hate all the upsell shit with discord
2025-11-11T14:28:21 < mawk> constantly trying to get you to spend for useless cosmetics
2025-11-11T14:28:22 -!- martinmoene [~martinmoe@132.229.46.129] has joined ##stm32
2025-11-11T14:28:22 < zyp> I mean, discord has its own reasons to be shit, but it has the best UI out of the bunch
2025-11-11T14:28:26 < BrainDamage> and running my own private server, I get to see how much idiotic and resource wasteful the federtion costs ... even if you don't partecipate in the federation
2025-11-11T14:28:35 < ventYl> to me discord seems like complete chaos. I am on ~three servers with about a dozen of open channels. whenever there's a discord notification I can't find the damn place it comes from
2025-11-11T14:28:42 < mawk> would you have to ban hateful servers if you federated BrainDamage 
2025-11-11T14:29:07 < mawk> on the top right there is a totally unintuitive icon that's labeled "inbox" ventYl and you can find where it comes from
2025-11-11T14:29:15 < mawk> or the bell icon on mobile at the bottom
2025-11-11T14:29:26 < BrainDamage> my resource costs are crazy as it is, I am not interested in spending 80GB/month DB storage to federate
2025-11-11T14:29:54 < ventYl> "Must be your lucky day, there's a new update" is another Discord "feature" I love.
2025-11-11T14:30:08 < mawk> their update screen is kinda funny
2025-11-11T14:30:26 < ventYl> you can't close it
2025-11-11T14:30:38 < mawk> yeah I just use the web version
2025-11-11T14:30:53 < mawk> it never has to update and it works the exact same; then if you install it as a PWA it's integrated in your window manager
2025-11-11T14:31:20 < zyp> I just have a browser windows with tabs for each of the shits
2025-11-11T14:31:44 < zyp> discord, element, slack, messenger, telegram
2025-11-11T14:31:49 < ventYl> mawk: you have just improved my quality of life
2025-11-11T14:32:01 < mawk> nice
2025-11-11T14:32:20 < ventYl> that inbox icon, which looks like RJ45 socket went completely unnoticed, because despite there are unread notifications... it just sits there
2025-11-11T14:32:26 < BrainDamage> yeah, wheneverI use on my computer telegram, whatsapp or discord is always the web version
2025-11-11T14:32:26 < mawk> lol
2025-11-11T14:32:32 < zyp> I've almost considered trying out irccloud for irc just so I can stick it in the same browser window as the other shits
2025-11-11T14:32:51 < BrainDamage> the main annoyance of whatsapp web is that isdoesn't support calls
2025-11-11T14:32:54 < mawk> I had a setup like that once zyp , with kiwiirc connecting to my bouncer
2025-11-11T14:32:59 < mawk> but it's a bit ugly
2025-11-11T14:33:21 < mawk> yeah you need the desktop version BrainDamage 
2025-11-11T14:33:21 < ventYl> zyp: there's 32 gigs of RAM in this machine, so let's put some use of it
2025-11-11T14:34:15 < mawk> I have 63.999996185302734 GiB of RAM
2025-11-11T14:34:30 < mawk> there's one page missing
2025-11-11T14:34:41 < zyp> ventYl, you say that like that's a lot
2025-11-11T14:36:01 < ventYl> zyp: unless I do make -j0 on some large project, I rarely feel that it is limiting me
2025-11-11T14:36:20 < mawk> it's mainly debug symbols that use up all the RAM on large projects
2025-11-11T14:36:27 < mawk> like building LLVM/clang
2025-11-11T14:36:39 < mawk> and LTO stuff I gues
2025-11-11T14:36:44 < ventYl> I killed the whole machine by running make -j0 on linux kernel twice
2025-11-11T14:37:18 < mawk> -j0 -j0 ?
2025-11-11T14:37:28 < ventYl> it also has 32 cores, so it took several dozens of build jobs to make the CPU busy, then the machine ran out of RAM and OOM killer took the whole desktop down
2025-11-11T14:37:34 < zyp> amount of ram is not supposed to be limiting either
2025-11-11T14:38:01 < ventYl> mawk: "fork bomb mode"
2025-11-11T14:38:03 < BrainDamage> you can kill any machine with -j
2025-11-11T14:38:20 < mawk> -j$(($(nproc)+2))
2025-11-11T14:38:32 < BrainDamage> it'll spaw as many jobs as the kernel lets you to
2025-11-11T14:38:40 < mawk> there's superstition folk wisdom about doing nproc or nproc + 1 or even nproc + 2
2025-11-11T14:38:44 < mawk> idk which one is correct
2025-11-11T14:38:56 < BrainDamage> in my tests, increasing nproc +1 or +2 takes longer than nproc
2025-11-11T14:38:59 < zyp> I'm just saying that I'd expect any decent modern computer to have at least 32GB today
2025-11-11T14:39:39 < ventYl> once, I retrofitted distcc into build process and was stealing test engineer's CPU time with distcc running with nice +19
2025-11-11T14:40:44 < zyp> the computer I got in 2010 had 32GB RAM, so 15 years later that doesn't seem so impressive anymore :)
2025-11-11T14:41:15 < zyp> actually, the computer I just ordered for my wife the other day also comes with 32GB
2025-11-11T14:41:30 < ventYl> honestly I think that 16GB would probably be still sufficient as I rarely see more than ~10 GB "used"
2025-11-11T14:41:52 < ventYl> yet maybe I'll buy 64GB and move these sticks into HTPC
2025-11-11T14:42:00 < ventYl> they are pretty shitty spec anyway
2025-11-11T14:42:12 < zyp> I think my laptop has 16, that rarely matters
2025-11-11T14:42:22 < zyp> but then I don't use my laptop for heavy stuff anyway
2025-11-11T14:42:57 < ventYl> what's heavy stuff? building 15 C sources that pretend to be a firmware? 486 can do that.
2025-11-11T14:43:08 < ventYl> for PCB drawing, there's qyx
2025-11-11T14:43:36 < ventYl> and he probably runs KiCAD on Abacus or similar kind of underpowered device just because it can be powered by solar clock.
2025-11-11T14:44:04 < zyp> I consider «build yocto» as among the heaviest stuff I do, wouldn't want less than 64GB on something tasked with that
2025-11-11T14:44:52 < ventYl> ok, that's heavy task
2025-11-11T14:45:10 < zyp> last I had to build yocto I went and bought a decommisioned rack server, dual cpu thing with 36C/72T in total and 384GB RAM
2025-11-11T14:45:28 < zyp> best bang for the buck I found on the secondhand market :p
2025-11-11T14:45:47 < ventYl> this computer was originally put together from secondhand market parts
2025-11-11T14:45:58 < ventYl> except mainboard and PSU
2025-11-11T14:46:34 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-11T14:47:16 < zyp> the rack box takes like 30-40 minutes for a minimal yocto image from scratch, which is a decent improvement when anything I've used before spent hours to do the same
2025-11-11T14:50:39 < zyp> of course, yocto does incremental builds, so you're not building from scratch all the time, but I hate fragile stateful temporaries, and I like being able to just wipe state and start fresh without it costing me hours
2025-11-11T14:51:37 < ventYl> I'd probably build release builds from scratch anyway
2025-11-11T14:53:06 < ventYl> damn, I just spent like... three hours determining why code won't build. for one surplus comma
2025-11-11T14:53:50 < zyp> exactly, you want to be sure that a release builds clean, and doesn't either rely on or gets fucked up by some unclean state
2025-11-11T14:56:21 < ventYl> my releases are built by automation server and released automatically
2025-11-11T14:57:03 < ventYl> with unit tests and clang-tidy running before packaging, any of them failing causing the build to be aborted altogether
2025-11-11T14:57:55 < mawk> I also have hardware tests in the pipeline
2025-11-11T14:58:14 < mawk> I have a raspberry pi as a gitlab runner that flashes the firmware on all hardware revisions and checks that it boots fine
2025-11-11T14:58:28 < ventYl> I am not there yet
2025-11-11T14:58:29 < mawk> and optionally run a bunch of tests using relays to simulate various faults and button presses
2025-11-11T14:58:49 < ventYl> there is a bunch of hardware tests for the kernel that theoretically can run on any supported target
2025-11-11T14:58:55 < ventYl> but there's no automation for this yet
2025-11-11T14:59:21 < mawk> I made some python thingie that uses the gpio pins to flash the firmware and then watch for serial output and send commands and stuff
2025-11-11T15:01:21 < ventYl> that part is already automated. I am using GDB + openocd to flash and run tests. as it only concerns the kernel, there's little need to touch GPIOs or anything.
2025-11-11T15:02:12 < ventYl> so the test itself is usually a small userspace application that does something and triggers one of two breakpoints. triggering one means test succeeded, triggering other means it failed. triggering any other breakpoint also means that the test failed.
2025-11-11T15:02:58 < ventYl> as most of the function testing is covered by unit tests, this checks if the SW still behaves well on any given HW, so I call them HW-SW integration tests.
2025-11-11T15:04:33 < zyp> yeah, automated testing is so nice
2025-11-11T15:05:19 < ventYl> just, anything running on HW is painfully slow
2025-11-11T15:05:44 < zyp> I made this thing when I added nrf54l support to blackmagic: https://github.com/zyp/gdb_autotest
2025-11-11T15:06:45 < zyp> because I got tired of manually running commands and keeping track of what state the target was in
2025-11-11T15:06:55 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-11T15:07:13 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-11T15:08:47 < zyp> especially when adding benchmarks into the sequence: https://github.com/blackmagic-debug/blackmagic/pull/2009#issuecomment-2512835288
2025-11-11T15:09:53 < ventYl> https://github.com/ventZl/cmrx/blob/master/testsuite/debug.gdb
2025-11-11T15:10:02 < ventYl> probably very similar approach, but driven from within GDB itself
2025-11-11T15:10:57 < zyp> in my case I were testing the blackmagic driver, so testing that it could find, attach to and flash the device
2025-11-11T15:11:36 < ventYl> OK, so GDB was used just as an interface
2025-11-11T15:11:46 < zyp> well, blackmagic is a gdbserver
2025-11-11T15:11:51 < mawk> do you use loading to RAM when developing?
2025-11-11T15:11:58 < zyp> no, it's not worthwhile
2025-11-11T15:12:04 < mawk> cubeide always generates this ram only file but I never used it
2025-11-11T15:13:47 < zyp> the primary advantage of loading to RAM is that you eliminate flash wear when iterating, but when flash is typically rated for 10-100k cycles on a modern MCU, flash wear is not a real enough concern to justify the disadvantages
2025-11-11T15:14:05 < mawk> yeah
2025-11-11T15:15:12 < ventYl> I'd only ever consider it if the application itself would do extensive flash access
2025-11-11T15:16:06 < zyp> a typical MCU will have more flash than ram, so you could only fit a small firmware that way, compared to what would fit in flash
2025-11-11T15:16:50 < zyp> so you'll run out of resources way faster
2025-11-11T15:18:07 < ventYl> my firmware is usually smaller than 80kB so it would fit into pretty much anything reasonable person would buy these days
2025-11-11T15:22:58 < karlp> meh, my web bluetooth works "great" on my machine, but on a windows machine it seems to only get one reply at a time.
2025-11-11T15:24:25 < karlp> ok, it' sonly sending one request.
2025-11-11T15:24:32 < karlp> I bet this is with/without confirmation
2025-11-11T15:33:11 < ventYl> lessons learned: stop using fucking feature branches
2025-11-11T15:33:18 < ventYl> I'll have to merge this shit file by file
2025-11-11T15:37:46 < zyp> BLE is pretty bad about what features are available through different platform APIs
2025-11-11T15:40:34 < qyx> ventYl: you are holding it wrong
2025-11-11T15:46:04 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-11T16:28:15 < ventYl> BTW, has anyone tried using QEMU STM32 emulation?
2025-11-11T16:33:54 < mawk> I tried
2025-11-11T16:34:06 < mawk> without success 
2025-11-11T16:34:26 < mawk> you should probably just try to get it running on real hardware 
2025-11-11T16:34:44 < mawk> with all the quirks and erratas it's not like a x86 with decades of adjustment 
2025-11-11T16:34:57 < mawk> also most peripherals are missing
2025-11-11T16:35:13 < ventYl> that's not a problem at all for my usecase
2025-11-11T16:38:46 < mawk> my father is probably coming out of the cardiac clinic tomorrow 
2025-11-11T17:01:08 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-11T17:03:22 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-11T17:12:39 < PhantomWork> qyx: re: buck regulator disaster. you said that I used one of the worse chip or something like that? Have a better one on your mind that is avail at jlcpcb?
2025-11-11T17:16:26 < zyp> any synchronous one should be better :p
2025-11-11T17:17:07 < zyp> I don't remember what specs you wanted?
2025-11-11T17:27:17 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-11T17:31:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T17:33:10 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-11T17:39:07 < PhantomWork> zyp: Vin: 7-24V, Vout: 5.0 1A and 3.3 150mA??, or maybe just 5 with a LDO to 3.3      I was about to calculate the current
2025-11-11T17:39:46 < PhantomWork> I currently have 3 projects on my desk using it. Copy/paste is nice
2025-11-11T17:43:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-11T17:45:12 < zyp> PhantomWork, what about something like this? https://www.ti.com/lit/ds/symlink/tlvm23615.pdf
2025-11-11T17:45:37 < zyp> I've considering using one of those (maybe not exact that part, but same family) for some project I had in mind
2025-11-11T17:45:45 < zyp> considered*
2025-11-11T17:46:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T17:49:08 < PhantomWork> nice chip for sure, but basically no stock at jlcpcb, and last time I tried to have them source a simple PIC they couln't for some reasons... so trying to not risk it
2025-11-11T17:52:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T17:53:39 < zyp> it's 2025, nobody cares about PICs anymore :p
2025-11-11T17:53:57 < zyp> I've had jlc source a bunch of stuff without issue before
2025-11-11T17:55:33 < karlp> us too, sure, use their parts first, but if you need a part, buy it into your global sourcing libs
2025-11-11T17:56:29 < zyp> they've got 10 in stock at the moment, which should mean they can easily get more, and if they for some reason can't, they can get them from TI directly through the global parts sourcing thing
2025-11-11T17:57:07 < zyp> so for this particular part I wouldn't worry about the availability
2025-11-11T18:02:34 < PhantomWork> what about the STM32F103, think it will be avail for atleast 10-15 years still? or I should upgrade?
2025-11-11T18:04:44 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-11T18:09:52 < qyx> PhantomWork: for 24 -> 5 I am using lmr36006, max25231, ap63356/7
2025-11-11T18:10:19 < PhantomWork> qyx: 7-24, which make it wide input for sure
2025-11-11T18:10:45 < qyx> also lmr36015, lmr33620
2025-11-11T18:11:34 -!- upgrdman [~upgrdman@12.69.225.50] has joined ##stm32
2025-11-11T18:12:02 < qyx> lmr36006 is my favorite one, zero issues, acoustically quiet even in low power mode, pretty cheap
2025-11-11T18:12:22 < qyx> ap63357 specifically sings a lot, but otherwise works
2025-11-11T18:13:57  * karlp needs to update https://github.com/karlp/zypsnips/blob/master/regs-stepdown.txt for qyx
2025-11-11T18:15:30 < karlp> qyx:lmr36006 which? there's _quite_ a price range differenge
2025-11-11T18:16:32 < PhantomWork> will check them
2025-11-11T18:17:04 < PhantomWork> for now... I need a shrink ray gun.. so I can add those 5 parts in "that" area where they don't fit...
2025-11-11T18:19:41 < qyx> karlp: wat, there are two freq versions and maybe a Q1 one
2025-11-11T18:29:24 < karlp> yeah, which one do you use?
2025-11-11T18:33:41 < qyx> karlp: crnxt
2025-11-11T18:45:45 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-11T19:33:19 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T19:43:18 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-11T19:43:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-11T19:43:53 < nomorekaki> so Quick share didn't pair my old and new phone
2025-11-11T19:44:36 < nomorekaki> old phone has SD but it's not usable because of encryption, new phone doesn't have SD at all
2025-11-11T19:44:42 < nomorekaki> new phone has usb2.0
2025-11-11T19:48:14 < nomorekaki> let's see asus phone clone
2025-11-11T19:49:08 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T19:51:44 < zyp> what phones are you buying that's still usb2 only?
2025-11-11T19:54:09 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-11T19:57:25 < nomorekaki> zenfone 10
2025-11-11T20:00:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T20:05:26 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-11T20:06:38 < nomorekaki> I think I will need to trust third party to do the trick
2025-11-11T20:13:19 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-11T20:26:05 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T20:37:02 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T20:37:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T20:42:54 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T21:24:10 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T21:26:13 < PhantomWork> is LM5164 a good regulator?
2025-11-11T21:36:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-11T21:36:54 -!- upgrdman [~upgrdman@12.69.225.50] has quit [Remote host closed the connection]
2025-11-11T21:37:02 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T21:37:15 -!- upgrdman [~upgrdman@12.69.225.50] has joined ##stm32
2025-11-11T21:37:23 < qyx> 1
2025-11-11T21:37:28 < qyx> so8? meh
2025-11-11T21:40:27 -!- upgrdman [~upgrdman@12.69.225.50] has quit [Remote host closed the connection]
2025-11-11T21:40:48 -!- upgrdman [~upgrdman@12.69.225.50] has joined ##stm32
2025-11-11T21:42:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T21:43:47 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T21:50:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T21:53:19 < mawk> 15 diodes in series that's the most efficient regulator
2025-11-11T21:58:23 < qyx> buydisplay displays arrived \o/
2025-11-11T21:59:22 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T22:07:19 -!- upgrdman [~upgrdman@12.69.225.50] has quit [Ping timeout: 256 seconds]
2025-11-11T22:20:21 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T22:40:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T22:45:18 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-11T22:56:26 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T23:06:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-11T23:17:26 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-11T23:18:17 < mawk> can I use DMA to update the DMA peripheral
2025-11-11T23:22:04 < ventYl> self-modifying DMA?
2025-11-11T23:22:10 < mawk> yeah
2025-11-11T23:22:14 < mawk> like a train of DMAs
2025-11-11T23:22:18 < mawk> I think the MDMA peripheral can do that
2025-11-11T23:22:30 < mawk> but maybe I can do it unofficially with regular DMA
2025-11-11T23:23:03 < mawk> like a 100Hz timer triggering a DMA request to feed PWM data to a timer, and then a 1Hz timer triggering another DMA request to change the source address for the first request
2025-11-11T23:23:36 < ventYl> release train
2025-11-11T23:24:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-11T23:52:53 -!- upgrdman [~upgrdman@12.69.225.50] has joined ##stm32
2025-11-11T23:56:25 < specing> MDMA peripheral.. lol
--- Day changed Wed Nov 12 2025
2025-11-12T00:02:19 < mawk> lol it's the real name
2025-11-12T00:02:24 < zyp> mawk's favorite peripheral
2025-11-12T00:02:38 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-12T00:25:13 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-12T00:27:53 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T00:31:51 -!- ferdna [~ferdna@user/ferdna] has joined ##stm32
2025-11-12T00:33:40 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T00:33:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T00:38:37 -!- HelloShi1ty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T00:41:19 -!- upgrdman [~upgrdman@12.69.225.50] has quit [Quit: Leaving]
2025-11-12T01:04:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T01:09:48 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T01:20:37 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-12T01:42:15 < qyx> zyp: how do you test your backplanes?
2025-11-12T01:42:43 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T01:42:45 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T01:42:54 < qyx> I am considering doing some ATE muxing card for measuring resistance of all possible (valid and invalid) combinations between all slots
2025-11-12T01:44:16 < qyx> and on the same topic, a more elaborate math problem for mawk, I have 10 slots, each having 12 pins going to a backplane power rail
2025-11-12T01:45:13 < qyx> is it possible to derive resistance between each pin (of 120 total) and the power rail *without* touching the power rail?
2025-11-12T01:45:25 < qyx> that is, only measuring between pins
2025-11-12T01:45:43 < Phantom> now... do I move my project to the F407... or I stay on F103....  F407 is a tad more expensive, but I eliminate the usb-serial chip...
2025-11-12T01:47:13 < zyp> qyx, my current backplanes are passive, so there's not much to test :)
2025-11-12T01:48:02 < qyx> I had a bad solder job in the past
2025-11-12T01:48:10 < qyx> on a passive backplane..
2025-11-12T01:48:32 < specing> Phantom: wdym, F103 has USB
2025-11-12T01:48:53 < Phantom> specing: it has usb OR can, not both
2025-11-12T01:49:03 < Phantom> which annoy me
2025-11-12T01:50:03 < zyp> I guess if I were making many enough to justify making a test for them, I'd have something I could plug into one slot that could read and write all pins, and a simple board that I could plug into the other slots that would loop back half the pins into the other half
2025-11-12T01:50:31 < zyp> using non-adjacent pins so that it doesn't mask two adjacent pins being shorted together
2025-11-12T01:50:55 < specing> Phantom: really? I never used either, but I also don't see a mention in datasheet that it's one or the other
2025-11-12T01:50:57 < qyx> yeah I am considering something similar, two identical cards with 1:n muxes
2025-11-12T01:51:04 < qyx> specing: really
2025-11-12T01:51:14 < zyp> orbtrace testjig uses the non-adjacent loopback trick to test IOs
2025-11-12T01:51:34 < qyx> zyp: and using an external dmm in a 4 wire mote
2025-11-12T01:51:35 < qyx> mode
2025-11-12T01:51:36 < Phantom> specing: yeah, it's not quite clear, but when you see the sentence, it is clear...   both uses the same memory space...
2025-11-12T01:52:08 < zyp> qyx, ohming out everything?
2025-11-12T01:52:12 < qyx> oh a loopback, let's make the math even harder!
2025-11-12T01:52:24 < qyx> yeah, would probably take ages
2025-11-12T01:52:33 < zyp> is it worthwhile?
2025-11-12T01:52:48 < qyx> idk it is easy and even if it runs one hour, I don't care
2025-11-12T01:53:03 < qyx> I am not doing 1000s daily
2025-11-12T01:53:34 < zyp> https://github.com/orbcode/orbtrace_testsuite/blob/main/bin/test-io-loopback.py
2025-11-12T01:54:22 < qyx> but that looks too easy
2025-11-12T01:56:21 < specing> Phantom: well, that sucks. Why didn't they just cut away a section of normal SRAM for each (as in, you can use it for other stuff otherwise)
2025-11-12T01:56:37 < qyx> if I had to use analog muxes, would I isolate them and bias them in a way the dmm is in the middle of VCC/VEE?
2025-11-12T01:56:52 < Phantom> probably because it's an old chip, and didn't expected people to use both or something
2025-11-12T01:56:58 < Phantom> or
2025-11-12T01:57:02 < zyp> qyx, maybe use relays if you worry about that :p
2025-11-12T01:57:05 < Phantom> to force to buy the bigger one
2025-11-12T01:57:24 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T01:57:29 < qyx> zyp: I have 8 connectors with 24 pins each..
2025-11-12T01:57:31 < qyx> at max
2025-11-12T01:57:40 < zyp> you can get DMMs with scan cards
2025-11-12T01:58:03 < qyx> I would need to steal it
2025-11-12T01:58:07 < Phantom> so, beside the F407, is there a better one at around the same price?
2025-11-12T01:58:23 < qyx> Phantom: what is "around the same price"?
2025-11-12T01:58:44 < Phantom> aka not twice the price
2025-11-12T01:58:55 < qyx> are you joking
2025-11-12T01:59:15 < qyx> you are considering china prices where f103 is maybe $1.50 and f407 is $3 or so?
2025-11-12T01:59:24 < qyx> and you want to go cheaper?
2025-11-12T01:59:27 < Phantom> no
2025-11-12T01:59:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T02:00:22 < Phantom> just wondering if there is a better choice than the 407 without going in the overly expensive for no reason ones
2025-11-12T02:00:53 < Phantom> to be fair, they kinda all look like the same black box at first glance   :D
2025-11-12T02:01:36 < zyp> did you miss where multiple people recommended stm32g431 the other day?
2025-11-12T02:01:53 < Phantom> I guess I did miss it
2025-11-12T02:02:17 < zyp> g431 would be my pick for a «newer replacement for f1» part
2025-11-12T02:02:32 < qyx> yes same here
2025-11-12T02:02:52 < qyx> I am using g491 regularly, which is a bit more advanced
2025-11-12T02:03:37 < qyx> STM32G431RBT6 is even cheaper on lcsc
2025-11-12T02:03:47 < zyp> I'm not sure I'd pick any of the f-families for a project today
2025-11-12T02:05:05 < qyx> ok I'll probably try the analog mux approach and see what it does
2025-11-12T02:05:11 < Phantom> I'll check that more in detail tomorow
2025-11-12T02:05:13 < zyp> I'd probably pick g0 over f0, g4 over f1/f3, h5 over f4 and h7 over f7
2025-11-12T02:05:37 < qyx> U5 over L4/L5
2025-11-12T02:05:41 < qyx> U3 over L0
2025-11-12T02:05:51 < qyx> Lx is getting old too
2025-11-12T02:05:52 < zyp> isn't there u0 as well?
2025-11-12T02:05:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T02:06:07 < qyx> oh yes
2025-11-12T02:06:08 < Phantom> thanks
2025-11-12T02:06:19 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-12T02:07:11 < zyp> c0 over f0 as well
2025-11-12T02:07:34 < Phantom> hmmm what is the smallest modern one that have ADC + DAC?
2025-11-12T02:07:45 < zyp> small in what sense?
2025-11-12T02:07:51 < Phantom> small footprint
2025-11-12T02:08:09 < zyp> I assume you don't want wlcsp?
2025-11-12T02:08:42 < Phantom> I won't solder it, so whatever jlcpcb can assemble it's fine I guess
2025-11-12T02:09:18 < zyp> g031 or g051 in qfn28 is 4x4 mm
2025-11-12T02:09:56 < zyp> it also comes in wlcsp20
2025-11-12T02:10:47 < zyp> that's 1.94x2.4mm
2025-11-12T02:10:55 < Phantom> G031 don't have DAC
2025-11-12T02:11:18 < zyp> g051 does
2025-11-12T02:11:49 < Phantom> interessing
2025-11-12T02:13:10 < Phantom> bit pricy
2025-11-12T02:13:45 < Phantom> but, might be worth it
2025-11-12T02:14:03 < zyp> if you go up to 5x5mm qfn32, you get a lot more options
2025-11-12T02:14:10 < zyp> including the g431 mentioned above
2025-11-12T02:14:45 < Phantom> will all check that tomorrow more in detail for sure
2025-11-12T02:14:58 < Phantom> but, it is very interessing
2025-11-12T02:15:31 < zyp> another bonus with these newer parts is that they only have a single pair of supply pins in the small packages, so they need less decoupling around them as well
2025-11-12T02:16:57 < Phantom> I have an issue where I need to adapt the signal from an accelerator pedal to whatever the ECU want, the problem is that the pedal have 2 signals, 1 is inverted... I need both linear, and... the offset is not quite right... so the inversio don't work, so I need some pot, or do it somehow else... and I think a pair of microcontrollers might be the way to go. That or I use some digipot
2025-11-12T02:18:03 < Phantom> the offset error was about 0.05v, it upset the ECU :/
2025-11-12T02:26:10 < Phantom> you know what... the bigger chip might be needed finally lol
2025-11-12T02:37:19 < qyx> g4 has two dac channels
2025-11-12T02:37:36 < qyx> and you know you can use an opamp to invert a signal
2025-11-12T02:38:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T02:39:05 < Phantom> yes, but there is a slight offset
2025-11-12T02:39:24 < Phantom> which make the inverted signal not the same as the other channel
2025-11-12T02:40:38 < Phantom> but if I use an I2C isolator, and some I2C digipot, and some opamp in differential mode... it miiight work for an auto-adjust
2025-11-12T02:45:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T02:46:25 < qyx> nah ethernet again, switches again
2025-11-12T02:46:59 < Phantom> it's annoying to have different grounds....
2025-11-12T02:58:28 < nomorekaki> migration tool did partial jub
2025-11-12T02:58:50 < nomorekaki> copied camera reel from local storage not from SD
2025-11-12T02:59:35 < nomorekaki> copied applications without their user files
2025-11-12T03:00:05 < nomorekaki> couldn't copy application user files like 30gigs of whatsapp
2025-11-12T03:01:18 < nomorekaki> copied all phone calls and txt though nice.. but that's like 2minute job with SMS restore app
2025-11-12T03:02:13 < nomorekaki> if I just get the same transfer speed than with that migration tool it's fine
2025-11-12T03:02:36 < nomorekaki> 17gigs was like 5min
2025-11-12T03:22:22 < nomorekaki> maybe cx file explorer
2025-11-12T03:23:51 < Phantom> oh man   this new stm32 open up so many possibility     I just need to find some proper differential opamp with beyond the rail inputs...   …  maybe I don't even need?
2025-11-12T03:31:01 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-12T03:32:35 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T03:40:20 < qyx> do we have a "common footprint" gigabit magjack?
2025-11-12T03:41:05 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T03:42:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T03:58:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T04:06:20 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T04:15:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T04:16:31 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T04:25:49 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-12T04:31:20 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-12T04:38:55 -!- ferdna [~ferdna@user/ferdna] has quit [Quit: Leaving]
2025-11-12T04:46:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T04:54:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T04:57:30 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T05:02:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T05:13:03 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T05:20:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T05:42:44 -!- krjst [~krjst@2a0a:4cc0:2000:789a:b827:c6ff:fed6:bb48] has quit [Quit: bye]
2025-11-12T05:43:11 -!- krjst [~krjst@v2202504264396328666.nicesrv.de] has joined ##stm32
2025-11-12T05:52:56 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T06:02:46 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T06:28:07 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T06:36:11 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T06:51:30 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-12T06:53:13 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T07:05:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T07:10:28 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T07:10:48 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-12T07:23:47 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T07:28:44 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T07:38:32 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T07:45:43 -!- qyx [~qyx@84.245.120.40] has quit [Ping timeout: 250 seconds]
2025-11-12T07:47:43 -!- qyx [~qyx@84.245.121.75] has joined ##stm32
2025-11-12T07:53:34 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T08:00:37 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-12T08:25:58 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-12T08:40:18 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T08:58:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T08:59:15 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T09:09:06 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-12T09:12:07 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-12T09:31:30 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T09:47:21 < mawk> interesting problem qyx
2025-11-12T09:48:40 < mawk> how is it connected? each pin is individually connected by some possibly unique resistance to the power rail, and you can only probe the pins?
2025-11-12T09:49:36 < mawk> so we have R1, R2, …, R120 and we can measure each Ri+Rj with i≠j
2025-11-12T09:50:31 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T09:52:20 < mawk> yeah it's possible 
2025-11-12T09:52:23 < zyp> qyx, voltage or current mode?
2025-11-12T09:52:26 < mawk> as long as you have at least 3 pins
2025-11-12T09:56:49 < mawk> we can take the combinations r1,2; r2,3; … r119,120; r118,120 for instance, then it's just a big matrix
2025-11-12T10:12:54 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-12T10:15:23 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-12T10:16:25 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-12T10:17:08 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T10:25:37 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-12T10:40:55 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T10:54:25 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-12T11:08:16 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-12T11:09:27 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T11:20:39 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T11:51:50 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T12:02:01 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T12:12:40 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T12:22:39 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-12T12:23:33 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T12:38:47 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has joined ##stm32
2025-11-12T12:45:06 -!- HelloShitty [~psysc0rpi@bl18-185-162.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T12:50:20 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T12:58:28 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-12T13:00:18 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 260 seconds]
2025-11-12T13:01:55 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T13:09:08 -!- HelloShi1ty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T13:10:01 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T13:16:16 -!- HelloShi1ty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T13:16:30 < qyx> mawk: exactly
2025-11-12T13:17:56 < qyx> so for 3 pins with resistances a, b, c and an arbitrary power rail resistance p, I can measure a+p+b, a+p+c, b+p+c
2025-11-12T13:18:01 < qyx> but I have 4 unknowns
2025-11-12T13:18:45 < qyx> Ii can set x=0 because it is negligible and then it is solvable
2025-11-12T13:19:15 < karlp> qyx: I've not been tistic enough on gigabit so far, 100M has always been more than enough, so I've not done any research on compat footprints sorry
2025-11-12T13:19:33 < qyx> or if I had 4 pins, I can do the x too
2025-11-12T13:19:50 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T13:20:16 < qyx> zyp: lan9646, center taps with capacitors to gnd
2025-11-12T13:20:23 < qyx> 100n
2025-11-12T13:20:53 < karlp> PoE or not? (that matters too right?)
2025-11-12T13:21:13 < zyp> center taps to individual caps is voltage mode, center taps to 3.3V is current mode
2025-11-12T13:21:16 < qyx> bez poe
2025-11-12T13:21:19 < qyx> *without poe
2025-11-12T13:21:22 < zyp> IME the latter is way more common, you need the former
2025-11-12T13:21:47 < qyx> RJMG from amphenol looks somewhat cheap
2025-11-12T13:21:50 < qyx> I mean, not overly expensive
2025-11-12T13:22:23 < qyx> RJMG201K21120FR more specifically
2025-11-12T13:22:38 < zyp> 100M magjacks typically just brings out the two center taps so you can do what you want, but gigabit magjacks don't want to have enough pins for that, so they terminate internally
2025-11-12T13:22:58 < qyx> apparently it can do both, it has center taps on one pin and caps on another pin (gnd)
2025-11-12T13:23:16 < zyp> that sounds like current mode
2025-11-12T13:23:31 < zyp> yup
2025-11-12T13:23:44 < zyp> https://cdn.amphenol-cs.com/media/wysiwyg/files/drawing/rjmg201k2xx20xr.pdf <- schematic here is clearly a current mode magjack
2025-11-12T13:24:27 < zyp> voltage mode center taps should not be connected together like that, they can be at different potentials and must be DC-isolated
2025-11-12T13:24:45 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T13:25:39 < qyx> what, hm
2025-11-12T13:25:58 < zyp> the JLC option is X01AQ038AB2AD
2025-11-12T13:26:16 < zyp> I've used that before, had to look through pages of current mode magjacks before I found it
2025-11-12T13:27:26 < zyp> I think that was the only available gigabit voltage mode magjack on jlc when I looked
2025-11-12T13:27:40 < zyp> and there's no filtering on it, so I had to check all the datasheets
2025-11-12T13:30:06 < qyx> if lan9646 has voltage mode phys, why is there an output current setting resistor?
2025-11-12T13:30:37 < karlp> to determine the voltage based on the current ;)
2025-11-12T13:30:41 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T13:30:52 < zyp> voltage vs current mode is how the output drivers work
2025-11-12T13:31:07 < zyp> voltage mode drivers are H-bridges, i.e. push pull
2025-11-12T13:32:06 < zyp> current mode drivers are low side only, gets high side voltage from the center tap, droves only one half at the time and lets the magnetics mirror it on the other half
2025-11-12T13:32:17 < zyp> drives*
2025-11-12T13:32:55 < zyp> AIUI voltage mode drivers are better and more energy efficient, current mode drivers are cheaper
2025-11-12T13:34:12 < qyx> ok so let's direct my energy into finding a suitable transformer then
2025-11-12T13:34:23 < qyx> I have a bag full of RJHSE jacks
2025-11-12T13:34:39 < zyp> lan9646 datasheet also confirms, ref. page 199: https://ww1.microchip.com/downloads/aemDocuments/documents/UNG/ProductDocuments/DataSheets/LAN9646-Data-Sheet-DS00005175.pdf
2025-11-12T13:34:42 < zyp> The LAN9646 PHY port design incorporates voltage-mode transmit drivers and on-chip terminations. With the voltagemode implementation, the transmit drivers supply the common-mode voltages to the four differential pairs. Therefore, the transformer center tap pins on the LAN9646 chip side should not be connected to any power supply source on the board; rather, the center tap pins should be separated from one 
2025-11-12T13:34:48 < zyp> another and connected through separate 0.1µF common-mode capacitors to ground. Separation is required because the common-mode voltage could be different between the differential pairs, depending on the connected speed mode.
2025-11-12T13:35:28 < qyx> I was searching this document just 2 minutes ago a found nothing
2025-11-12T13:36:22 < qyx> I was even reading that section
2025-11-12T13:37:06 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T13:37:13 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T13:38:13 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T13:39:14 < zyp> once you've ran into it, you know to look for it
2025-11-12T13:39:39 < zyp> protip: if you want gigabit PoE, forget magjacks
2025-11-12T13:43:57 < qyx> hmm
2025-11-12T13:44:09 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-12T13:44:56 < zyp> I like those little single pair transformers, haven't used them yet, but I will next time I need something I don't already know of a suitable magjack for
2025-11-12T13:45:35 < qyx> yeah exactly my thoughs right now
2025-11-12T13:45:44 < qyx> + discrete CM chokes
2025-11-12T13:49:53 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T14:09:32 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T14:11:38 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T14:28:31 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-12T14:43:48 < PhantomWork> So, G431 don't have enough pin for the new crazy idea I got... :/
2025-11-12T14:45:11 < zyp> you know you can get g431 in 100-pin packages as well?
2025-11-12T14:46:48 < PhantomWork> but jlcpcb don't stock them, and global sourcing don't always work, so.... trying to have something they currently stock to atleast have the few next orders less painfull
2025-11-12T14:47:01 < zyp> how many pins do you need?
2025-11-12T14:47:11 < karlp> so you're going to make something that sucks, because you can't get global sourcing to work?!
2025-11-12T14:47:21 < karlp> file a bug on global sourcing, it's not that hard...
2025-11-12T14:48:02 < zyp> what do you mean «don't stock them»? did you even look? https://jlcpcb.com/partdetail/STMicroelectronics-STM32G431VBT6/C529359
2025-11-12T14:48:59 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T14:49:25 < qyx> also, where are you located?
2025-11-12T14:49:45 < qyx> why do you even consider lcsc as the only source?
2025-11-12T14:50:13 < zyp> presumably to have jlc do assembly without dicking with global sourcing
2025-11-12T14:50:15 < PhantomWork> karlp: I contacted the agent, they replied that they couln't source it.
2025-11-12T14:50:58 < PhantomWork> I think I hit a lazy person there
2025-11-12T14:52:00 < PhantomWork> as to how many pins, I currently have 48, so next up should work
2025-11-12T14:52:04 < PhantomWork> so 64-100
2025-11-12T14:52:27 < zyp> 64 is also stocked: https://jlcpcb.com/partdetail/STMicroelectronics-STM32G431RBT6/C431633
2025-11-12T14:53:46 < PhantomWork> yeah just found it via another search.... I hate mornings :D
2025-11-12T14:54:39 < qyx> looks like he is gonna make millions of devices
2025-11-12T14:57:53 < PhantomWork> nahh, but around an hundred a year, and I might drop it in 2-3 projects
2025-11-12T14:58:48 < PhantomWork> I use a stm32F103 in 3 current projects, 1 will definitely get the upgrade, the second half for it probably too. Third... also probably
2025-11-12T14:59:24 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T14:59:51 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T15:03:14 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-12T15:04:29 < PhantomWork> in stmcubeide, is there a way to upgrade the µC from it? Or I need to create a new project and move the code?
2025-11-12T15:06:28 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T15:16:58 < karlp> hecking, love coming back to a project, plug the board in, openocd find the target, gdb load...... fails to flash... wot
2025-11-12T15:24:16 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T15:36:55 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-12T15:37:33 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T15:37:59 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-12T15:42:31 < ventYl> 99% of time
2025-11-12T15:42:55 < ventYl> how is the shit connected to the debugger
2025-11-12T15:42:58 < ventYl> *which* debugger?
2025-11-12T15:43:02 < ventYl> where's the adapter
2025-11-12T15:43:30 < qyx> where is damned usb-c cable again
2025-11-12T15:44:02 < ventYl> this one got fixed by me buying 10 of them and having stock available. then occasionally I do garbage collection and restore the stock
2025-11-12T15:44:22 < ventYl> then, why can't openocd connect to the target?
2025-11-12T15:44:41 < ventYl> finally, the code won't build with quite a mysterious error
2025-11-12T15:46:25 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T15:48:26 < karlp> in parting shots at my webbluetooth I've had to shelve, UUIDs passed in as service filters are _required_ to be in lowercase hex.  uppercase hex will not be accepted, even with a warning.  You do at least get an info log that you're request was ignored due to a malformed uuid  :)
2025-11-12T15:49:10 < qyx> I would ban uppercase
2025-11-12T15:49:22 < qyx> all hexa should be lowercase my law
2025-11-12T15:49:42 < ventYl> and yesterday, I ended up pulling 4 months old feature branch commit-by-commit
2025-11-12T15:49:55 < ventYl> I wasn't able to understand why the code ended up merged the way it did
2025-11-12T15:50:23 < karlp> qyx: by all means emit it, but why would you hard ban it by spec...
2025-11-12T15:50:51 < qyx> by law!
2025-11-12T15:51:06 < karlp> fuckin fuckin got a new mouse, scrolls randomly just like the old one.
2025-11-12T15:51:16 < karlp> one of those days for switching projects...
2025-11-12T15:51:41 < ventYl> steelseries?
2025-11-12T15:52:46 < zyp> my work mouse has some gunk in the encoder wheel, some regions scrolls the wrong way
2025-11-12T15:53:08 < zyp> it's been pissing me off, but not enough to figure out how I open it without breaking it so I can clean it out
2025-11-12T15:53:48 < ventYl> same here. the sensor is mechanical so it can't even be opened and cleaned
2025-11-12T16:27:24 -!- specing [~specing@user/specing] has quit [Ping timeout: 260 seconds]
2025-11-12T16:34:03 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T16:50:36 -!- specing [~specing@user/specing] has joined ##stm32
2025-11-12T16:51:24 < karlp> ventYl: logi m705.  same one I have at home, which isfine, and same as old one, which was fine for ~12 months or so.
2025-11-12T16:51:54 < karlp> IT guy who came by with a replacement though, said they're seeing mice in general just barely make 6 months these days, even 3dconnexion space mice are dying within a year.
2025-11-12T16:54:27 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-12T17:01:45 < ventYl> karlp: meh. i bought some upmarket mouse because I got tired of microswitched being worn out, now wheel sensor died
2025-11-12T17:02:06 < ventYl> on 2nd mouse in row in like... 18 months
2025-11-12T17:14:36 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T17:14:53 < srk> can't beat mx518
2025-11-12T17:19:47 < srk> mx vertical is also pretty good
2025-11-12T17:21:03 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T17:22:40 < machinehum> til all debian packages are compiled natively
2025-11-12T17:23:18 < machinehum> Learn something new every day
2025-11-12T17:31:56 < zyp> I'm still using a mx510 here
2025-11-12T17:32:12 < zyp> IIRC bought in 2004
2025-11-12T17:58:10 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T18:02:50 < karlp> (I've actually just noticed that pressing the "what's this button for?" on top of the mouse puts it into clicky wheel mode, which works just fine, so I don't hav eto fuck around with figuring out how to tune the freewheeling "smooth" things)
2025-11-12T18:03:12 < karlp> I wonder if my one at home is in clicky mode and I just never made the association
2025-11-12T18:06:16 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-12T18:09:01 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-12T18:16:09 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T18:25:45 < PhantomWork> does anyone knows where I can find the default pin states for the G431 under reset and bootloader? The F103 surprised me with some pin that default high
2025-11-12T18:26:27 < PhantomWork> it happened to be the horn that is connected there, was quite a surprise when I pressed reset and it went HOOOOOOONk
2025-11-12T18:45:22 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T19:10:14 < mawk> the hardware chip select of ST is kinda useless
2025-11-12T19:10:18 < mawk> for SPI
2025-11-12T19:11:07 < mawk> at least if it worked with DMA, pulling it down at the beginning of the transaction, that would be useful
2025-11-12T19:13:01 < mawk> in the "reset" values of the corresponding GPIO registers PhantomWork 
2025-11-12T19:14:19 < mawk> but pins default high sounds dubious
2025-11-12T19:14:26 < mawk> pins are floating input by default
2025-11-12T19:14:32 < mawk> no pullup or pulldown 
2025-11-12T19:14:49 < PhantomWork> there is atleast one that is pullup while in reset
2025-11-12T19:14:54 < PhantomWork> on the f103
2025-11-12T19:15:29 < PhantomWork> I forgot which one, I swapped pins when I last did the board due to physical restrictions 
2025-11-12T19:15:31 < mawk> are you sure it's not the BOOT0 pin you're looking at?
2025-11-12T19:17:14 < PhantomWork> no, boot0 have a switch on it, boot1 a resistor, so no... I think it's one of the jtag pins
2025-11-12T19:18:11 < mawk> yeah JTAG pins are also pullup on reset
2025-11-12T19:18:47 < mawk> or pull down depends which one
2025-11-12T19:19:06 < mawk> but it's not a problem just set the pins back to reset mode
2025-11-12T19:19:12 < mawk> you don't have to avoid them
2025-11-12T19:19:21 < PhantomWork> it could have been any others and I wouln't have noticed or cared. but that one was the spooky one :D
2025-11-12T19:19:33 < mawk> lol
2025-11-12T19:19:54 < PhantomWork> oh it ain't really an issue, just that I pressed reset and HOOOONK    I wasn't prepared   lol
2025-11-12T19:20:25 < PhantomWork> now.... to learn about digipot and find i2c isolator... the joy...
2025-11-12T19:20:42 < mawk> why do you want to isolate
2025-11-12T19:21:24 < PhantomWork> the grounds ain't shared, and a wiring issue could shoot one of the 3 grounds to 12V
2025-11-12T19:21:33 < mawk> ah
2025-11-12T19:21:45 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T19:22:58 < PhantomWork> main circuit is vehicle ground with ignition power, second is sensor ground1 with sensor power1, while the third is sensor ground2 and sensor power2...
2025-11-12T19:26:59 < PhantomWork> food time
2025-11-12T19:28:40 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T19:34:40 < mawk> what is powering the sensors
2025-11-12T19:36:02 < mawk> use pairs of horns and mics for the I²C isolation 
2025-11-12T19:36:15 < mawk> makes debugging easy
2025-11-12T19:50:39 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-12T19:56:37 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-12T19:59:22 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has joined ##stm32
2025-11-12T20:00:42 < PhantomWork> mawk: the ECU... actually I misremembered, it's 2 dedicated power lines, but still, 2 different ground and 2 different VCC
2025-11-12T20:07:43 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-12T20:37:54 -!- splud [~noneya.bi@user/splud] has quit [Remote host closed the connection]
2025-11-12T20:39:54 -!- splud [~noneya.bi@user/splud] has joined ##stm32
2025-11-12T20:42:23 < mawk> what happens if you join them
2025-11-12T20:54:16 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-12T20:54:45 < nomorekaki> Steffanx: movie
2025-11-12T20:56:27 < Steffanx> Soldier of Orange 1977
2025-11-12T20:56:57 < nomorekaki> ok
2025-11-12T20:58:27 < Steffanx> Or do you want something else? 
2025-11-12T20:58:35 < nomorekaki> dont have VPN to watch that
2025-11-12T20:58:56 < Steffanx> Hm
2025-11-12T20:59:55 < nomorekaki> it's prime video and some watch online sites I will never open
2025-11-12T21:00:01 < Steffanx> Karate kid legends
2025-11-12T21:01:55 < nomorekaki> maybe
2025-11-12T21:02:18 < Steffanx> Lol idk
2025-11-12T21:02:33 < Steffanx> I havent seen a lot of movies lately
2025-11-12T21:03:16 < nomorekaki> https://www.imdb.com/title/tt0112431
2025-11-12T21:03:47 < Steffanx> Oh dear 
2025-11-12T21:07:27 < nomorekaki> Demolition was good
2025-11-12T21:23:46 < nomorekaki> Novocaine (Basic guy become tough guy to save a girl movie) ,
2025-11-12T21:23:47 < nomorekaki> Sisu (Tough guy action movie, over the top violence, gore),
2025-11-12T21:23:47 < nomorekaki> The Intern (Drama. De Niro, Hathaway) ,
2025-11-12T21:23:48 < nomorekaki> Flypaper (heist comedy, mystery),
2025-11-12T21:23:48 < nomorekaki> Burn after reading (Farce. Clooney, Pitt)
2025-11-12T21:30:10 < antto> nomorekaki, have you watched Raw?
2025-11-12T21:30:14 < antto> or what was it called
2025-11-12T21:31:07 < nomorekaki> "A young woman, studying to be a vet, develops a craving for human flesh."
2025-11-12T21:31:14 < antto> yeah, that one
2025-11-12T21:31:32 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-12T21:31:40 < nomorekaki> not my genre
2025-11-12T21:32:12 < antto> undefined reference to "my genre"
2025-11-12T21:35:02 < nomorekaki> okay Steffanx imdb 8.5 https://www.imdb.com/title/tt0482571/
2025-11-12T21:35:19 < nomorekaki> 2 Magicians battleing it out
2025-11-12T21:35:30 < antto> the prestige is very good
2025-11-12T21:36:25 < antto> it's one of those that requires you to watch it several times ;P~
2025-11-12T21:36:56 < nomorekaki> saw it once in 2008 or so
2025-11-12T21:37:06 < nomorekaki> still a bit recent to see it again
2025-11-12T21:37:59 < antto> i guess that's because Nolan
2025-11-12T21:38:07 < antto> Tenet is also like that
2025-11-12T21:38:09 < nomorekaki> It might have had classical nolan style following one timeline from multiple perspectives
2025-11-12T21:38:26 < nomorekaki> Tenet was only that
2025-11-12T21:38:44 < antto> tenet is exclusively a time-mind-f*ck
2025-11-12T21:40:38 < nomorekaki> it's a bit show off
2025-11-12T21:42:29 < antto> i recently watched some movie about some US(?) soldiers going to .de during one of the world wars and they find some Lab where they brew a secret sauce/serum that resurects dead soldiers (but with Funky Side Effects) o_O
2025-11-12T21:44:32 < nomorekaki> okay. name?
2025-11-12T21:44:39 < antto> i don't remember ;P~
2025-11-12T21:45:01 < nomorekaki> It's sounds distantly interesting
2025-11-12T21:45:40 < antto> it was interesting because i didn't expect it, perhaps i shouldn't have said about the sauce
2025-11-12T21:46:37 < antto> i'm not actually sure if they went to .de or .fr or so
2025-11-12T21:47:12 < Steffanx> .bg for sure 
2025-11-12T21:47:28 < antto> nope
2025-11-12T21:51:06 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-12T21:55:49 < antto> nomorekaki, i think i found it: https://www.imdb.com/title/tt4530422/?ref_=sr_t_41
2025-11-12T21:57:45 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-12T21:59:12 < nomorekaki> ok I watch that
2025-11-12T21:59:33 < nomorekaki> it's available
2025-11-12T21:59:35 < nomorekaki> thanks
2025-11-12T22:07:47 -!- martinmoene_ [~Martin@210-212-21-31.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-12T22:34:48 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-12T22:44:52 -!- ice [~ice@loud.house] has joined ##stm32
2025-11-12T22:49:15 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T22:54:02 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-12T23:12:53 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-12T23:48:16 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-12T23:50:45 < nomorekaki> antto: nice movie
--- Day changed Thu Nov 13 2025
2025-11-13T00:01:54 < karlp> here's a design I've not nseen before: https://www.aliexpress.com/item/1005006120590300.html  runs the strip on DC like normal, but also sends the 220AC along the line as well to provide an extension point...
2025-11-13T00:04:22 < qyx> when I see those EU non-conformant plugs/sockets..
2025-11-13T00:04:31 < karlp> it's pretty great :)
2025-11-13T00:04:48 < karlp> doing some "market research" on my stupid ideas...
2025-11-13T00:05:15 < karlp> don't want to pull apart my niec actually functional and in use led strips, need to buy some sacrificial ones.
2025-11-13T00:05:28 < qyx> maybe the idea alone is but there are pretty strict requirements for extension cords and also for flexible appliance cords in the EU
2025-11-13T00:05:39 < karlp> I suspect the controller is the same, just want to see what sort of offline power supply they use..
2025-11-13T00:06:00 < qyx> that's exactly the kind of shit when I am questioning whether CE should be a self-certification process
2025-11-13T00:06:14 < karlp> what's the alternative?
2025-11-13T00:06:28 < karlp> anyway, can't solve that here or now :)
2025-11-13T00:06:36 < qyx> of course you can't
2025-11-13T00:06:45 < qyx> alternative is to use properly sized conductors
2025-11-13T00:07:00 < qyx> proper sockets/cords even for class II devices (including PE)
2025-11-13T00:07:01 < qyx> etc.
2025-11-13T00:07:03 < karlp> but eyah. I'm not buying 220v extensions that happen to have fairly lights along them, with super thin wires.
2025-11-13T00:07:34 < karlp> I want to find the right china source for teh common "waterproof" (mostly) 2 terminal screw connector used on most of these outdoor lights.
2025-11-13T00:07:54 < qyx> which one?
2025-11-13T00:08:00 < karlp> "the one" :)
2025-11-13T00:08:07 < qyx> the screw-nut type thing?
2025-11-13T00:08:09 < karlp> yeah.
2025-11-13T00:08:19 < karlp> there's a couple I know, 
2025-11-13T00:08:27 < karlp> but never really found them easily.
2025-11-13T00:08:37 < karlp> never gone hard on alibaba/taobao either though so..
2025-11-13T00:08:39 < karlp> anyway
2025-11-13T00:08:53 < qyx> https://www.pcboard.ca/image/cache/catalog/products/cables-jacks-usb/2-pin-water-proof/2-pin-waterproof-set-800x800.jpg
2025-11-13T00:08:59 < qyx> this?
2025-11-13T00:09:33 < zyp> the ones I've used are more D-shape keyed
2025-11-13T00:09:34 < karlp> yeah, that's one of them, guess it's not so hard to find anymore and I should look harder...
2025-11-13T00:09:52 < karlp> I've also gone ones that are two differently sized flat blades
2025-11-13T00:10:00 < karlp> rather than plastic notch keying.
2025-11-13T00:10:01 < ventYl> i have these, or very similar on my solar DWH system sensors
2025-11-13T00:10:30 < karlp> where on pcboard to you actually get that connector
2025-11-13T00:10:32 < ventYl> yet the ones I've used are not exacly rated for transfering (a lot of) power
2025-11-13T00:10:42 < qyx> karlp: 404
2025-11-13T00:11:06 < karlp> https://www.pcboard.ca/2-pin-waterproof-connector-set
2025-11-13T00:11:11 < karlp> nicec price....
2025-11-13T00:11:29 < qyx> too expensive?
2025-11-13T00:11:53 < karlp> no :)
2025-11-13T00:12:09 < karlp> still, need the connectors themselves 
2025-11-13T00:12:24 < qyx> what's the goal? to get cheap waterproof connectors for your lights?
2025-11-13T00:12:28 < qyx> or be compatible with china?
2025-11-13T00:12:58 < karlp> goal is keep existing strips, possibly existing power supplies for ac shit, but definitely existing battery boxes
2025-11-13T00:13:07 < karlp> ac shits and this two pin connector is "stretch goal" shit.
2025-11-13T00:13:26 < karlp> original basic plan is replace the tiny board in battery boxed led strips with a "better" (more epxensive) one..
2025-11-13T00:13:43 < qyx> I would do weipu sp13 if I had to choose a cheap alternative
2025-11-13T00:14:16 < qyx> or techno thb 381 for a proper european alternative
2025-11-13T00:14:34 < karlp> weipu is _quality_ china
2025-11-13T00:14:39 < karlp> I want _basic acceptable_ china.
2025-11-13T00:15:12 < qyx> weipu is the baseline for me and even they leak and oxidize over time
2025-11-13T00:15:15 < karlp> I want XKB, or jing of the flying lol,
2025-11-13T00:15:27 < karlp> yeah, but you want permanent outdoors
2025-11-13T00:15:34 < karlp> I want seasonal outdoors
2025-11-13T00:16:30 < qyx> is lighting your outdoors a thing in iceland like in the great america?
2025-11-13T00:16:43 < karlp> hrm, somewhat.
2025-11-13T00:16:46 < karlp> I do it more for halloween
2025-11-13T00:17:05 < qyx> you are a bit late
2025-11-13T00:17:05 < karlp> but I'm just hinking, if I'm doign this, it's a _feasible_ market that's bigger than _me_ for it.
2025-11-13T00:20:24 < qyx> I am pretty sure I had 2pin jsp ph connectors somewhere
2025-11-13T00:32:58 < qyx> found them at the wrong place
2025-11-13T00:33:10 < qyx> there are so many wrong places in this room since I did some reorganizing
2025-11-13T00:39:27 < karlp> zyp: one of the ones in our kitchen window we put up for the winter has those d-shaped ones, not the notch keys :)
2025-11-13T00:44:52 < mawk> the LSE on my nucleo-32 board doesn't want to start, but I didn't change any solder bridges and the thing is clearly there
2025-11-13T00:46:04 < mawk> and I didn't connect anything to the associated pins
2025-11-13T00:47:02 < mawk> but if I disable "CRS SYNC" then it works again
2025-11-13T00:49:45 < mawk> yeah no it's just when that's disabled cube doesn't configure LSE
2025-11-13T00:49:48 < mawk> mysterious
2025-11-13T00:54:07 < Phantom> I2C... this isolator show 1k on each side... shouln't that be 4.7k??
2025-11-13T00:55:39 < zyp> should in what sense?
2025-11-13T00:56:01 < qyx> i2c pullup may be anything
2025-11-13T00:56:13 < Phantom> but isn't the standard 4.7k?
2025-11-13T00:56:17 < qyx> no?
2025-11-13T00:56:20 < zyp> there's no standard
2025-11-13T00:56:34 < zyp> i2c pullup strength is a tradeoff between speed and power consumption
2025-11-13T00:56:54 < qyx> and output driver strength
2025-11-13T00:58:33 < zyp> you've got pretty strong pullups when that starts to matter :p
2025-11-13T00:58:42 < Phantom> I'll read the datasheet more in detail tomorrow, I kept being interrupted so maybe I missed something
2025-11-13T00:59:08 < mawk> yeah it works on another board so the LSE seems kaput
2025-11-13T00:59:24 < qyx> LSE on stm32 randomly fails for me too
2025-11-13T00:59:39 < mawk> even in high drive mode and giving it 10s to start it doesn't work
2025-11-13T00:59:43 < Phantom> also, it seems like there is no way to have the full speed of the stm32g431?  when USB is enabled
2025-11-13T00:59:56 < mawk> why
2025-11-13T01:00:10 < mawk> because cubemx shows some red in the clock config?
2025-11-13T01:00:33 < mawk> it's either its own oscillator or its own PLL output right
2025-11-13T01:00:40 < Phantom> usb need 48MHz, there is no possible dividers that satisfy this and the 320MHz limit of the pll
2025-11-13T01:01:21 < mawk> the PLL has multiple outputs, and USB can use HSI48 as well no?
2025-11-13T01:01:25 < mawk> if it's in the chip
2025-11-13T01:01:26 < Phantom> unless I use the internal RC for the usb, which I have doubts about the reliability, but maybe it just work?
2025-11-13T01:01:45 < qyx> it is PLL'd from the USB itself
2025-11-13T01:01:55 < mawk> you can use HSI48 and enable CRS to trim it using the LSE, or using the USB start of frame 
2025-11-13T01:02:16 < mawk> see "CRS SYNC" in the RCC peripheral
2025-11-13T01:02:29 < qyx> also whether your MCU runs at 170 MHz or 168 MHz doesn't really matter
2025-11-13T01:02:56 < Phantom> 170 or 144
2025-11-13T01:03:03 < Phantom> but yeah, won't matter
2025-11-13T01:03:05 < qyx> oh 320 MHz limit
2025-11-13T01:03:11 < qyx> idk I am running at 128 MHz
2025-11-13T01:03:12 < zyp> HSI48 exists specificly for USB
2025-11-13T01:03:27 < Phantom> I'm at 72 on the F103, and I'm 90% idle anyway
2025-11-13T01:03:40 < zyp> so you can expect it to work (when used with CRS as intended)
2025-11-13T01:24:57 < qyx> let's smoke some pcbs
2025-11-13T01:29:29 < qyx> no smoke today
2025-11-13T01:52:59 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-13T02:01:53 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-13T02:23:34 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-13T03:44:38 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-13T04:19:48 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-13T05:23:18 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-13T06:58:05 -!- travisb__ [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-13T07:03:09 -!- kst [~krjst@2a0a:4cc0:2000:789a:b827:c6ff:fed6:bb48] has joined ##stm32
2025-11-13T07:04:10 -!- Netsplit *.net <-> *.split quits: octorian, travisb_, karlp, Posterdati, PaulFertser
2025-11-13T07:05:04 -!- splud [~noneya.bi@user/splud] has quit [Ping timeout: 256 seconds]
2025-11-13T07:05:05 -!- krjst [~krjst@v2202504264396328666.nicesrv.de] has quit [Ping timeout: 256 seconds]
2025-11-13T07:05:39 -!- Netsplit over, joins: PaulFertser, octorian, karlp
2025-11-13T07:12:54 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-13T07:13:05 -!- splud [~noneya.bi@user/splud] has joined ##stm32
2025-11-13T07:33:03 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-13T07:38:39 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Quit: bitmask]
2025-11-13T08:04:51 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-13T08:48:50 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-13T09:16:31 < qyx> there is no christmas tree on TME
2025-11-13T09:35:05 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-13T09:35:46 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-13T09:45:02 < jpa-> real pros order their trees from digikey
2025-11-13T09:47:00 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-13T09:47:11 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-13T10:33:28 < mawk> how would I synchronize 2 DMA requests, knowing that I can't trigger them with the same timer
2025-11-13T10:33:47 < mawk> I suppose make one timer a slave of the other? and if they use the same timer clock then they should never fall apart
2025-11-13T10:33:49 < qyx> synchronize timers
2025-11-13T10:34:13 < mawk> on the H7 you can synchronize the DMAs directly but not on this L4
2025-11-13T10:35:18 < qyx> yes, or try to generate trgo on update on the master one; and set trgi to cause reset on the slave one
2025-11-13T10:35:38 < mawk> ah reset sets the counter to zero then
2025-11-13T10:35:39 < qyx> but I usually sync them only once
2025-11-13T10:36:02 < qyx> beware there are some delays on some combinations
2025-11-13T10:36:20 < qyx> but I don't remember
2025-11-13T10:36:27 < mawk> there can be at most 7 spi clock cycles of delay so that's maybe fine
2025-11-13T10:37:08 < qyx> if not you can use a third timer as a master one to sync multiple slaves (with the same possible delay)
2025-11-13T10:37:16 < mawk> right
2025-11-13T10:39:26 < jpa-> considering two DMA requests will never occur simultaneously on same DMA controller, you are just as fine by starting the both timers inside a no-interrupts block
2025-11-13T10:40:01 < mawk> even if one is on DMA1 and the other on DMA2? but idk if that's possible with this chip 
2025-11-13T10:40:25 < jpa-> with separate DMA controllers they at least occur closer to each other :)
2025-11-13T10:40:50 < jpa-> still depends on bus access though
2025-11-13T10:41:07 < jpa-> if they access the same peripheral there will be at least 1 clock cycle difference
2025-11-13T10:42:33 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-13T10:45:03 < jpa-> i think DMA2 on L4 only supports TIM6/TIM7, which do not have interconnection with any other timers
2025-11-13T10:50:25 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-13T10:51:23 -!- martinmoene [~martinmoe@132.229.46.129] has quit [Read error: Connection reset by peer]
2025-11-13T10:53:35 < mawk> right
2025-11-13T10:53:45 < mawk> yeah there's few timer interconnections
2025-11-13T10:53:53 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-13T10:57:17 < jpa-> i guess you could counteract the delay by setting one timer to CNT=1 and one to CNT=0 and starting them from inline asm
2025-11-13T10:58:22 < jpa-> but all this only matters if the peripherals the DMA controllers are accessing are on different buses, and you may still get a few cycles of jitter due to CPU accesses
2025-11-13T10:58:31 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-13T11:00:13 < mawk> well as long as the delay is reasonable it's fine, I have to set a GPIO output within the first 7 bits of a SPI byte transmission 
2025-11-13T11:00:32 < mawk> if the delay stays constant then it's fine, if it drifts then not so
2025-11-13T11:00:58 < jpa-> timers clocked from the system clock of course stay in sync if their period is the same
2025-11-13T11:01:09 < mawk> right
2025-11-13T11:01:46 < jpa-> depending on what you are doing, the SPI NSS pin already kinda does that
2025-11-13T11:02:39 < mawk> the GPIO pin is basically a 9th bit for the SPI byte, it might change from byte to byte
2025-11-13T11:03:05 < mawk> the "data/command" pin of a screen controller
2025-11-13T11:03:25 < jpa-> ah
2025-11-13T11:03:36 < mawk> it's possible to actually use 9-bit SPI instead of a separate pin but then I need to change a tiny solder bridge with my humongous iron
2025-11-13T11:05:25 < mawk> it would be nice to be able to control NSS within each byte with SPI, you can either leave it always on during the transmission or pulse it between bytes, but nothing in-between 
2025-11-13T11:05:45 < jpa-> you could trigger the GPIO DMA from SPI RX DMA request
2025-11-13T11:06:36 < mawk> ah good idea
2025-11-13T11:07:07 < mawk> the SPI is transmit only, I can put it in bidirectional mode and then not set up the alternate function on the pin? and it shouldn't complain
2025-11-13T11:07:23 < jpa-> yeah, you can
2025-11-13T11:07:43 < jpa-> though with L4 SPI having a FIFO, i'm not sure what it does if you never read from it
2025-11-13T11:07:54 < mawk> qyx btw LSE was not starting because I was drawing too much current on a GPIO; I fixed that and it works again
2025-11-13T11:07:58 < mawk> I suppose overrun
2025-11-13T11:08:18 < jpa-> status register has no overrun bit, which makes me wonder
2025-11-13T11:08:23 < mawk> ah right 
2025-11-13T11:08:41 < jpa-> no wait, it does :D
2025-11-13T11:09:10 < mawk> yeah there is an overrun flag
2025-11-13T11:09:51 < mawk> they don't say if the error stops the peripheral or not
2025-11-13T11:10:26 < jpa-> "When an overrun condition occurs, the newly received value does not overwrite the
2025-11-13T11:10:26 < jpa-> previous one in the RXFIFO. The newly received value is discarded and all data transmitted
2025-11-13T11:10:27 < jpa-> subsequently is lost. "
2025-11-13T11:10:31 < jpa-> sounds like it doesn't stop
2025-11-13T11:11:41 < mawk> yeah
2025-11-13T11:16:03 < mawk> the screen I have is kinda ugly and there a pair of dead pixel rows
2025-11-13T11:16:43 < mawk> what kind of small screen (like 5×3cm) with a wide viewing angle can I use
2025-11-13T11:16:51 < mawk> maybe eink stuff
2025-11-13T11:17:30 < mawk> I have a 3-color eink screen from ages ago but the refresh rate is about 30mHz
2025-11-13T11:19:56 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has quit [Read error: Connection reset by peer]
2025-11-13T11:20:35 < jpa-> oleds are nice, color einks are still slow
2025-11-13T11:21:39 < ventYl> oleds are tiny
2025-11-13T11:21:44 < ventYl> at least those commonly available
2025-11-13T11:22:51 < mawk> what I have is oled
2025-11-13T11:22:56 < mawk> but the pixels are huge
2025-11-13T11:23:06 < mawk> maybe a more modern one is better 
2025-11-13T11:23:22 < mawk> also the contrast is bad but maybe I need to give it more current
2025-11-13T11:24:46 < mawk> if it's to display temperature maybe the eink is good, it doesn't change that often
2025-11-13T11:24:54 < mawk> but I stored it improperly for like 10 years 
2025-11-13T11:25:09 < mawk> it's supposed to be blanked and stored screen facing down, and I did the opposite of that
2025-11-13T11:33:12 < qyx> mawk: which gpio, one of those low power ones?
2025-11-13T11:33:28 < mawk> idk I have the choice
2025-11-13T11:33:32 < mawk> does it matter?
2025-11-13T11:33:48 < mawk> the cables are short
2025-11-13T11:33:59 < mawk> ah you mean for LSE
2025-11-13T11:35:15 < mawk> yeah a GPIO of the L4, let me see which one
2025-11-13T11:38:43 < mawk> I'm not sure what's happening with the pin on the nucleo-32 board but I was using one of the pins that should be "left floating" because of the default I²C configuration 
2025-11-13T11:39:09 < mawk> so likely the I²C pullups were sinking too much current into ground through the pins
2025-11-13T11:39:22 < mawk> I can't find the values of the pullups
2025-11-13T11:45:33 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-13T12:06:57 -!- c10ud [~c10ud@user/c10ud] has joined ##stm32
2025-11-13T12:10:16 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-13T12:17:06 < mawk> what kind of font engine do you use to display stuff
2025-11-13T12:17:19 < zyp> I've heard good stuff about mcufont
2025-11-13T12:17:20 < mawk> I feel like just a bunch of bitmap fixed width characters is enough 
2025-11-13T12:17:28 < mawk> ah that sounds interesting 
2025-11-13T12:21:35 < ventYl> I gutted some arduino library and mixed it with my own AVR-era code for displaying text
2025-11-13T12:21:57 < ventYl> it is dead stupid, but it works reasonably well
2025-11-13T12:22:40 < ventYl> it could be vastly simplified by reconfiguring SSD1306 to actually use linear FB
2025-11-13T12:58:04 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T13:03:27 < mawk> what do you mean by linear FB
2025-11-13T13:03:41 < mawk> by default it just scans left to right and top to bottom right
2025-11-13T13:05:20 < ventYl> SSD1306 has two modes: paged and linear. paged scans stripes of 8 lines, linear is... well, linear
2025-11-13T13:05:24 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-13T13:05:29 < ventYl> I guess they all have both modes
2025-11-13T13:06:08 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T13:06:27 < zyp> nah, only the monochrome ones do, rgb displays are always at least a full byte per pixel
2025-11-13T13:11:43 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Remote host closed the connection]
2025-11-13T13:11:58 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T13:19:34 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-13T13:26:49 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T13:31:44 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-13T13:40:11 < mawk> can I use DMA with the bit band
2025-11-13T13:52:35 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T14:19:45 < mawk> the scanning speed of the screen is not great, I can see it flickering when I move my eyes
2025-11-13T14:21:09 < mawk> ah I don't need the bit band there's already something for bitwise access in the GPIO peripheral
2025-11-13T14:29:00 < ventYl> well, you'll see this unless refresh rate becomes over 100Hz
2025-11-13T14:29:41 < ventYl> last week we did PWM LED dimming with students on the IoT class, it started by question: how fast can you blink a LED before you perceive it as a steady light instead of blinking
2025-11-13T14:29:50 < ventYl> answer was ~82 Hz, unless it is moving
2025-11-13T14:30:46 < mawk> idk I never saw this before
2025-11-13T14:30:57 < mawk> my tv is 60Hz and it doesn't do that
2025-11-13T14:31:26 < BrainDamage> the answer is not a single number, eg if you look at it off center your brain becomes much more sensitive to flicker
2025-11-13T14:31:38 < ventYl> both eyes and CCDs work on the "photon bucket" principle, integrating photons over time. if you move your sight far enough, you'll see it flickering
2025-11-13T14:31:43 < ventYl> is the TV OLED too?
2025-11-13T14:31:51 < mawk> it does it when my eyes move, like during the ocular saccade it takes an instant snapshot
2025-11-13T14:32:04 < mawk> maybe, I bought it like 6-7 years ago
2025-11-13T14:32:10 < BrainDamage> lcd refresh rate is different too
2025-11-13T14:32:35 < BrainDamage> lcd have capacitor-transistor that holds the value between frame changes
2025-11-13T14:32:37 < ventYl> do you use it often? if yes, then it is highly probably isn't OLED because it would develop ghosting already
2025-11-13T14:32:38 < mawk> also it depends on the capacitors that are next to the pixels right? it's fine to not refresh it often if it can keep its charge
2025-11-13T14:32:48 < BrainDamage> exactly
2025-11-13T14:33:31 < BrainDamage> for crt the exp decay of the phosphor was way way stronger
2025-11-13T14:33:42 < BrainDamage> hence why 60Hz on a lcd is not 60Hz on a crt
2025-11-13T14:33:44 < mawk> like in old scopes
2025-11-13T14:33:52 < ventYl> LCDs also have some inertia in crystals
2025-11-13T14:34:03 < mawk> and eink has infinite inertia
2025-11-13T14:34:22 < ventYl> so dimming isn't instant, while LED shuts down the very moment charge runs out
2025-11-13T14:34:46 < BrainDamage> you need to integrate the curve of the light intensity persistance of the system to get a fair comparison
2025-11-13T14:35:18 < BrainDamage> ventYl: depends, the led is instant, but the yellow phosphor in white leds can be slower
2025-11-13T14:35:33 < BrainDamage> so you can get a hue shift
2025-11-13T14:35:58 < mawk> are there white leds in oled displays? not r g b ?
2025-11-13T14:36:07 < ventYl> do yellow leds also use luminiscence to generate colour?
2025-11-13T14:36:31 < BrainDamage> mawk: to my knowledge,, it's rgb in on oled
2025-11-13T15:00:20 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-13T15:02:25 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 264 seconds]
2025-11-13T15:24:28 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T15:30:18 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-13T15:36:35 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T15:44:25 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 240 seconds]
2025-11-13T15:49:30 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T15:54:10 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 244 seconds]
2025-11-13T16:15:06 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T16:35:38 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-13T16:36:08 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T17:01:10 -!- specing [~specing@user/specing] has quit [Ping timeout: 265 seconds]
2025-11-13T17:21:04 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 244 seconds]
2025-11-13T17:21:08 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-13T17:24:57 -!- duude__ [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-13T17:25:30 -!- duude__- [~duude__@user/duude/x-4676560] has quit [Ping timeout: 256 seconds]
2025-11-13T17:26:28 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-13T17:28:19 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-13T17:33:38 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-13T17:49:22 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-13T18:00:37 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-13T18:10:38 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-13T18:24:59 < PhantomWork> on STM32G431, what is the maximum power consumption that the chip itself can use? Somehow I have doubt that 41mA is the right amount...
2025-11-13T18:30:00 < qyx> there is a table in the datasheet, sum the core, flash and peripheral consumption for your settings/speed
2025-11-13T18:42:36 < mawk> there is a tab in cubemx to calculate it PhantomWork 
2025-11-13T18:42:40 < mawk> you select which peripherals you use
2025-11-13T18:44:28 < mawk> the datasheet says 20-24mA for the core itself at 150MHz 25°C in run mode
2025-11-13T18:45:27 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-13T18:46:25 < mawk> so about 52mA with all the peripherals enabled and all pins disabled
2025-11-13T18:49:36 < mawk> and it doesn't count everything either, I'd just use the magic tool
2025-11-13T19:10:43 < PhantomWork> chatgpt said closer to 80mA, so I'll budget 100mA... anyway it is just to figure out the regulator I need to use. I'm planning to get rid of the 3.3V switcher and use a 5V instead with a 3.3V LDO
2025-11-13T19:11:47 < qyx> why not cascade the switchers
2025-11-13T19:12:13 < qyx> POL switchers are like 5x5 mm including the inductor
2025-11-13T19:13:05 < qyx> you can get eg. tps6282533, it only needs two caps and a 1210 sized inductor
2025-11-13T19:13:20 < qyx> if you don't need 2A, you can go with 0805 inductor
2025-11-13T19:14:16 < PhantomWork> I'll tackle the regulators next, for now, I need to figure out this digipot thing... meh
2025-11-13T19:14:51 < PhantomWork> atleast now it work in excel, so, not too bad
2025-11-13T19:24:54 < mawk> why on earth would you ask chatgpt for that
2025-11-13T19:25:19 < mawk> use the tool that's made for that that will give you an accurate result and not a guess
2025-11-13T19:25:45 < mawk> it depends entirely on which peripherals you use and which pins you have enabled
2025-11-13T19:25:49 < mawk> and on the core frequency
2025-11-13T19:25:58 < mawk> there's not just one easy convenient number that a clanker can guess
2025-11-13T19:42:59 -!- travisb_ [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-13T19:43:20 -!- joel135_ [uid136450@id-136450.hampstead.irccloud.com] has joined ##stm32
2025-11-13T19:50:49 -!- Netsplit *.net <-> *.split quits: joel135, travisb__
2025-11-13T19:50:49 -!- joel135_ is now known as joel135
2025-11-13T19:57:24 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-13T20:14:55 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 256 seconds]
2025-11-13T20:20:23 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-13T20:38:13 < karlp> PhantomWork: when you get there, https://jlcpcb.com/partdetail/TexasInstruments-TPS62826DMQR/C2073319 is on jlc parts, it's the same POL qy was talking about, but it will cose youtwo extra resistors, as that's the adjustable one...
2025-11-13T20:40:51 < PhantomWork> karlp: thanks, will check it
2025-11-13T20:41:13 < PhantomWork> ah for 5V to 3.3V
2025-11-13T20:41:33 < PhantomWork> do you have one for 30+V down to 5V?
2025-11-13T20:44:24 < karlp> qyx was sayign to use a POL buck, not a POL LDO...
2025-11-13T20:44:55 < karlp> I was just saying "for jlc, use _this_ variant of that POL, so it's super well stocked, and 30c, instead barely stocked and $2, when they're the "same" parts
2025-11-13T20:48:29 -!- specing [~specing@user/specing] has joined ##stm32
2025-11-13T20:59:33 -!- c10ud [~c10ud@user/c10ud] has quit [Read error: Connection reset by peer]
2025-11-13T21:16:25 < PhantomWork> karlp: yeah, I know, I have 2 regulators on the board, and I plan to replace both
2025-11-13T21:19:49 < PhantomWork> ... who is the imbecile that made that pin layout???? USART3 rx and tx pins are on the opposite side of the chip !
2025-11-13T21:29:43 < specing> it probably makes sense if you mount the chip across the connector leads
2025-11-13T21:30:16 < qyx> PhantomWork: you can definitely change that
2025-11-13T21:47:25 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-13T21:47:37 < bitmask> new glenn launch 2 in 10 min
2025-11-13T21:52:44 < PhantomWork> maybe it's because I defined all ADC in... let's try to disable some
2025-11-13T21:57:27 < PhantomWork> nope....the pin layout is really bad
2025-11-13T22:00:45 < PhantomWork> I2C_3 is sane, _2... 19 pins between both pins...  _1 is 8 between...
2025-11-13T22:07:16 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 255 seconds]
2025-11-13T22:12:01 < mawk> in cube PhantomWork ?
2025-11-13T22:12:17 < mawk> hold control and hold click on the USART3_RX
2025-11-13T22:12:22 < mawk> alternative pins will flash
2025-11-13T22:12:26 < mawk> and do the same with TX
2025-11-13T22:12:31 < mawk> then you can select the best pair
2025-11-13T22:12:52 < mawk> instead of spending an hour in the datasheet
2025-11-13T22:24:12 < qyx> I2C on G3 is insane?
2025-11-13T22:24:15 < qyx> *G4
2025-11-13T22:26:28 < qyx> I2C2 is PA8/9, neighboring pins
2025-11-13T22:26:50 < qyx> I2C1 is PA13/14/15, all neighboring pins
2025-11-13T22:26:51 < PhantomWork> mawk: control don't work...  weird
2025-11-13T22:27:09 < qyx> also on PB7/8, neighboring pins
2025-11-13T22:29:11 < PhantomWork> or maybe there is a problem with my cubeide?
2025-11-13T22:31:34 < PhantomWork> ah    need to hold down control...
2025-11-13T22:31:37 < PhantomWork> and mouse
2025-11-13T22:31:43 < PhantomWork> not just control-click
2025-11-13T22:39:59 < PhantomWork> so I2C1 conflict with swd.. 
2025-11-13T22:41:06 < PhantomWork> ... and cubemxide did a stupidity with I2C2 and put the pin on another side for no reason  o.O
2025-11-13T22:44:40 < PhantomWork> thanks for the tip
2025-11-13T22:44:54 < PhantomWork> I've been able to "deconflict" and remap
2025-11-13T22:49:12 < mawk> cubemx just selects the first pin it can choose with that alternate function 
2025-11-13T22:49:20 < mawk> it doesn't know you want them close
2025-11-13T22:50:28 -!- artok [~azo@88-112-154-28.elisa-laajakaista.fi] has quit [Quit: upgrade time for alpine]
2025-11-13T22:50:28 < mawk> you can disable swd if you really need to, then you can't debug but you can still flash if you route the reset pin to the debug header
2025-11-13T22:53:53 < PhantomWork> yeah, but now it's fine, I have a sane layout
2025-11-13T22:54:57 < PhantomWork> I still hesitate between using 1, 2 or 3 I2C bus... I have an eeprom and 2x digipot, each digipot are on the other side of a I2C isolator
2025-11-13T22:58:17 < mawk> just 1 seems fine no?
2025-11-13T22:58:36 < mawk> it's made for that
2025-11-13T23:01:55 < PhantomWork> yeah, will just check for address collision, but my guess is: it will just work(tm)
2025-11-13T23:04:51 < mawk> I²C slaves pretty often have an address selection pin to address just that
2025-11-13T23:04:56 < mawk> pun intended 
2025-11-13T23:06:50 < ventYl> you can go full blaupunkt, throw stm32 out of the window as it is too new and use mc68k-based MCU. it has shit ton of I2C
2025-11-13T23:06:54 < ventYl> like.... maybe 20
2025-11-13T23:07:32 < ventYl> there you can use i2c even for talking between your threads, because even if you connected peripherals of nearby PCB to your MCU, you'd still have some spare I2Cs
2025-11-13T23:07:49 < bitmask> wow, they landed it on the second attempt, impressive
2025-11-13T23:18:30 < qyx> I lost my wire cutters this week, somewhere here in my nerd cave/lab
2025-11-13T23:18:38 < qyx> I wasn't able to find them since monday
2025-11-13T23:19:08 < qyx> today, wifeqyx came and found it in like 30 seconds
2025-11-13T23:19:36 < ventYl> selective blindness
2025-11-13T23:19:41 < ventYl> you should defrag your brain
2025-11-13T23:19:53 < qyx> "don't look in a box with capacitors, they are surely not there"
2025-11-13T23:20:02 < qyx> "don't look in a rack with lifepo4 batteries.."
2025-11-13T23:20:11 < qyx> "don't look, hey, that's a freezer.. why"
2025-11-13T23:20:31 < qyx> "don't look in the internet rack, why would .. oh, really?"
2025-11-13T23:22:20 < zyp> my kid apparently lost her homework somewhere, didn't wanna go to bed because she were searching for it
2025-11-13T23:23:57 < qyx> I hope she found it in the freezer
2025-11-13T23:24:33 < zyp> I made her go to bed without finding it, don't think she checked the freezer :p
2025-11-13T23:31:59 < jbo> how does one loose homework?
2025-11-13T23:32:36 < jbo> wouldn't it be in the last place she put it?
2025-11-13T23:32:57  * PhantomWork is annoyed...
2025-11-13T23:33:36 < PhantomWork> can rx = boot0
2025-11-13T23:34:00 < PhantomWork> so... it could get stuck in bootloader because it received data at the same time it powered up?
2025-11-13T23:36:10 < karlp> ok, enough of jlc and kicad and vite and ts and shit.
2025-11-13T23:36:19 < karlp> too easy to "just finish this one thing" WFH
2025-11-13T23:36:56 < karlp> PhantomWork: it's not like i2c is differential or anything, who cares where they go.
2025-11-13T23:37:02 < jbo> karlp, I recently discovered that I am really good at bikeshedding since I am working on DIY/hobby projects again.
2025-11-13T23:37:10 < karlp> yolo, 1mhz, woooo, maybe 4 on a realllllllly good day with special parts...
2025-11-13T23:37:52 < jbo> also I blame zyp
2025-11-13T23:43:00 < qyx> that's ok, never blame qyx
2025-11-13T23:44:52 < jbo> never
2025-11-13T23:45:24 < jbo> lol, first time I see a section in a TI datasheed named: "Dos and Do Nots"
2025-11-13T23:47:26 < ventYl> .not
2025-11-13T23:57:01 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-13T23:59:36 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
--- Day changed Fri Nov 14 2025
2025-11-14T00:07:31 < zyp> jbo, turns out it was where I put it
2025-11-14T00:08:11 < zyp> I moved it out of reach of the little kid the other day so she wouldn't rip it up
2025-11-14T00:08:22 < zyp> only remembered I did so once I found it where I put it :p
2025-11-14T00:09:34 < zyp> in any case, since somebody didn't think about their homework before bedtime today, they get to wake up early so they can finish it before school tomorrow :p
2025-11-14T00:09:55 < jbo> so TL;DR being: stuff is generally where you last left it. no idea how you guys can be looking for stuff all the time.
2025-11-14T00:10:06 < jbo> zyp, lol
2025-11-14T00:10:18 < jbo> are you punishing her for you forgetting that you moved it? :p
2025-11-14T00:10:53 < zyp> it's not a punishment
2025-11-14T00:11:13 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 256 seconds]
2025-11-14T00:11:48 < zyp> it's a lesson in «when you have to do something, you shouldn't wait until bedtime before you start»
2025-11-14T00:12:05 < jbo> sounds good - I'll remind you next time.
2025-11-14T00:12:13 < zyp> :)
2025-11-14T00:15:12 < zyp> and yeah, I'm generally good at remembering where I leave my stuff, works well as long as nobody else moves it in the mean time
2025-11-14T00:16:13 < zyp> but I'm not trying to keep track of where everybody else's stuff is
2025-11-14T00:17:58 < jbo> "I'm not trying to keep track of where everybody else's stuff is while I am the guy who actually moved it" - checks out.
2025-11-14T00:19:12 < zyp> when you have kids, you're picking up mess and moving stuff out of small kids' reach all the time
2025-11-14T00:19:49 < karlp> everybody loves a danish landmine.
2025-11-14T00:20:04 < jbo> wtf is a danish landmine
2025-11-14T00:20:13 < karlp> brightly coloured pointy abs....
2025-11-14T00:20:17 < karlp> (lego)
2025-11-14T00:20:21 < jbo> aah
2025-11-14T00:21:34 < zyp> we don't actually have a lot of those
2025-11-14T00:22:08 < jbo> because you always move them elsewhere and then forget about it? :p
2025-11-14T00:22:37 < zyp> but right now there's like three or four open boxes of bracelet beads in the middle of the floor in her room
2025-11-14T00:23:24 < zyp> gonna be fun cleaning that up if somebody trips on them
2025-11-14T00:25:47 < zyp> five boxes actually: https://photos.app.goo.gl/kuFp4VyeanUdBqSW8
2025-11-14T00:28:03 < BrainDamage> https://www.youtube.com/watch?v=Oo91xho1EH4
2025-11-14T00:28:46 < jbo> BrainDamage :D
2025-11-14T00:29:07 < jbo> I was never allowed to leave stuff on the floor
2025-11-14T00:35:29 < Steffanx> Poor jbo
2025-11-14T00:36:35 < jbo> trying to figure that out right now. maybe I was living a lie all my life.
2025-11-14T00:42:30 < ventYl> you are living in a simulation, don't bother
2025-11-14T00:42:54 < jbo> I only accept that if you /nick fentyl again :<
2025-11-14T00:44:14 -!- ventYl is now known as sim_fentYl
2025-11-14T00:47:42 < jbo> I partly accept
2025-11-14T00:49:04 < Phantom> HOME! yay
2025-11-14T01:15:37 -!- qyx [~qyx@84.245.121.75] has quit [Ping timeout: 264 seconds]
2025-11-14T01:17:21 -!- qyx [~qyx@84.245.120.43] has joined ##stm32
2025-11-14T01:35:15 < Phantom> question guys. So, CANbus Rx goes to BOOT0, so there is a good chance that the µC start up while there is data on the bus, therefore bringing it into bootloader mode. The alternative pin for it is USB, which can not be relocated. Putting the CAN tranciever into standby... let it pass wake up signal, which can put the µC into bootloader mode... So... What do you suggest? 
2025-11-14T01:41:35 < zyp> is this for g4? do you need boot0?
2025-11-14T01:41:50 < zyp> IIRC you can disable boot0 in option bits if you're not using it
2025-11-14T01:41:54 < zyp> so, just do that
2025-11-14T01:42:39 < zyp> this is a feature on the newer stm32 families, older didn't have it
2025-11-14T01:42:46 < zyp> and I believe g4 is new enough to have it
2025-11-14T01:43:20 < mawk> but the builtin pull-up remains
2025-11-14T01:43:22 < mawk> right
2025-11-14T01:43:37 < jbo> zyp just called me old :<
2025-11-14T01:44:26 < zyp> you don't have a boot0 option bit?
2025-11-14T01:44:53 < jbo> I only did g4 once, like 2 years ago :<
2025-11-14T01:45:01 < mawk> there's no alternative pin for the CAN peripheral Phantom ?
2025-11-14T01:45:50 < mawk> sim_fentYl on your ssd1306 display does the "color contrast" command do anything? I can't get it to do anything at all no matter the values
2025-11-14T01:46:02 < mawk> it's supposed to control the maximum current going into the LED for each color
2025-11-14T01:46:12 < mawk> I set it to 0 to see and nothing changes
2025-11-14T01:46:50 < sim_fentYl> mawk: yes, it dims LED *slightly* yet still clearly noticeably by the naked eye
2025-11-14T01:47:10 < sim_fentYl> my OLED is white-only though
2025-11-14T01:47:30 < mawk> all the other commands work but not this one
2025-11-14T01:47:34 < jbo> mawk, I seem to remember some 256/101 shit: https://git.ugfx.io/uGFX/ugfx/src/branch/master/drivers/gdisp/SSD1306/gdisp_lld_SSD1306.c
2025-11-14T01:48:47 < mawk> it's different from the "master contrast" command which works fir'
2025-11-14T01:48:49 < mawk> fine*
2025-11-14T01:48:49 < sim_fentYl> mawk: supposedly the command I'm using is 0x81, 0x8F
2025-11-14T01:49:35 < sim_fentYl> teaching POSIX threads doing something they weren't designed to do is hard, heading to bed
2025-11-14T01:49:35 < mawk> ah but you have monochrome, then I guess it doesn't apply to your code
2025-11-14T01:56:14 < mawk> ah you're not using the command jbo
2025-11-14T01:56:17 < mawk> I guess it's useless 
2025-11-14T01:56:57 < mawk> btw there is a deeper sleep mode than the one implemented in this code, by switching the regulator off
2025-11-14T01:59:54 < Phantom> mawk: the alternative for CAN is the USB port, and there is no alternative for the USB port
2025-11-14T02:00:03 < mawk> ah
2025-11-14T02:00:11 < mawk> then disable boot0 in the option bytes
2025-11-14T02:00:25 < mawk> it just needs to be done once, you can do it when flashing the board
2025-11-14T02:00:31 < Phantom> that can be done? interessing
2025-11-14T02:00:44 < Phantom> I'll add it to the init code then
2025-11-14T02:00:58 < mawk> well you can do it in the code if you want
2025-11-14T02:01:07 < mawk> but make sure to not set it again if it's already set
2025-11-14T02:01:13 < mawk> or just do it when flashing the board
2025-11-14T02:01:40 < mawk> if you use stm32cubeprogrammer there is a GUI to set the option bytes
2025-11-14T02:01:57 < mawk> including other stuff like write protection or read protection or watchdog etc 
2025-11-14T02:02:04 < Phantom> I need to think of when I'll mass produce it...
2025-11-14T02:02:11 < Phantom> so don't want to forget it
2025-11-14T02:02:23 < Phantom> or... maybe I can add a tristate buffer on the rx side
2025-11-14T02:02:37 < Phantom> which is boot0
2025-11-14T02:02:38 < mawk> just set the option bytes
2025-11-14T02:02:46 < mawk> you don't need extra hardware
2025-11-14T02:03:09 < zyp> option bytes might even come with it disabled by default
2025-11-14T02:03:11 < Phantom> and if I want to usb flash, I could always add an in-code option for that
2025-11-14T02:03:57 < Phantom> or check for another chip than the g431  :D
2025-11-14T02:04:17 < mawk> you can jump to the ST bootloader from the code if you want
2025-11-14T02:04:38 < mawk> it also does it automatically when the flash is empty
2025-11-14T02:05:04 < zyp> nah, boot0 on PB8 is enabled by default
2025-11-14T02:06:04 < zyp> the bit controlling it is named nSWBOOT0, documented in RM0440 rev9 page 201
2025-11-14T02:06:32 < Phantom> I miss the good ol' atmega for their datasheet
2025-11-14T02:06:44 < Phantom> one massive complete document with proper examples for everything
2025-11-14T02:06:57 < zyp> that's RM0440
2025-11-14T02:07:38 < zyp> the atmega parts were just smaller, so there were less to document :p
2025-11-14T02:08:16 < Phantom> sure, but no other datasheet was as complete
2025-11-14T02:08:40 < jbo> +1 for atmega datasheets
2025-11-14T02:09:38 < jbo> yes, it was much, much less complexity. but they were still extremely good datasheets in my opinion
2025-11-14T02:10:58 < zyp> sure, I just don't think stm32 documentation is any less good
2025-11-14T02:11:36 < zyp> there's plenty of vendors with actually bad documentation
2025-11-14T02:11:57 < Phantom> *cough*pic*couch*
2025-11-14T02:11:59 < zyp> which reminds me I should check if infineon released register level documentation for fx3g2 yet
2025-11-14T02:12:44 < zyp> pic is hardly relevant today :)
2025-11-14T02:15:50 < Phantom> that's a project I need to revisit, one of our module use a PIC... and is currently coded in assembly... and don't work well at all. I will try to see if I can convert it to a STM32, and code it myself, in C. Will be maintainable now, and way easier to fix it
2025-11-14T02:17:38 < Phantom> I'll have to find the smallest physically chip  :D    need like 2 interrupts input, 1 standard gpio, 2 output PWM and 1 standard gpio... shouln't be hard...
2025-11-14T02:17:45 < Phantom> but that is next year
2025-11-14T02:26:04 < zyp> the qfn28 I mentioned the other day is probably the smallest that doesn't become a pain to assemble
2025-11-14T02:26:33 < zyp> there's also some in so8, but I believe the footprint for those is actually larger than the qfn28
2025-11-14T02:28:16 < zyp> yeah, so8 is 4.9x6mm, qfn28 is 4x4
2025-11-14T02:28:34 < zyp> oh, c0 is available in a 3x3mm qfn20, that's even smaller
2025-11-14T02:28:48 < zyp> as well as a 1.70x1.42mm wlcsp12
2025-11-14T02:29:01 < zyp> that is probably the absolute smallest stm32 you can get
2025-11-14T02:30:29 < Phantom> I'll look into it next year, literally
2025-11-14T02:30:35 < zyp> the wlcsp12 is 0.35mm pitch, triangular grid
2025-11-14T02:30:43 < Phantom> I have this one to finish, then another project to work on, then that one...
2025-11-14T02:31:08 < zyp> not sure you can get at the inner balls without laser vias
2025-11-14T02:31:33 < zyp> but there's only two of them, so if you only need the outer 10, it's probably still doable on a default spec board
2025-11-14T02:55:16 -!- mid-kid [~mid-kid@2a01:7c8:aac8:1e8:5054:ff:fe5e:cd48] has quit [Ping timeout: 256 seconds]
2025-11-14T02:57:11 -!- mid-kid [~mid-kid@2a01:7c8:aac8:1e8:5054:ff:fe5e:cd48] has joined ##stm32
2025-11-14T04:31:02 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-14T07:25:40 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-14T07:32:45 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-14T07:34:07 < qyx> nah even 0.4 mm wlcsp is pain to solder
2025-11-14T07:40:59 < Phantom> I'm still interessing into data over dc power, so if someone have a proper schem and all  :D
2025-11-14T08:05:09 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-14T08:10:14 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-14T08:23:16 < jpa-> Phantom: eh, you have to do *something* yourself too, you already got plenty of ideas the previous time
2025-11-14T08:24:51 < Phantom> jpa-: yeah, but beside the power over ethernet, which is too complex for my need, I don't recall to have seen any true solution being proposed (or I missed it). True as more than a theory, but with actual parts, or atleast something that I can find parts...
2025-11-14T08:25:46 < Phantom> I've seen lots of theory, very little practical example, except some expensive ones or 100% proprietary
2025-11-14T08:27:30 < Phantom> it is also annoying that if I would do it unidirectional, that the power is at the wrong side, so I can't "h-bridge" it
2025-11-14T08:32:36 < Phantom> I was looking at the STM32Hsomething.... and I saw 640k and like 570MHZ? with all the periphericals, including HDMI...  man that would have been a super fast computer many many years ago   :D
2025-11-14T08:42:00 < jpa-> Phantom: part selection and schematic design is obviously your task
2025-11-14T08:42:42 < Phantom> jpa-: part selection when you have an idea of what you can use yes. But I have no idea of what to use
2025-11-14T08:44:00 < jpa-> yeah, the task may be beyond your skill level, you probably need to prototype etc. to learn more before you can accomplish it
2025-11-14T08:44:50 < Phantom> exactly. And it does not seems easy at all, because nobody seems to have a working schematics
2025-11-14T08:45:25 < Phantom> all I read is theorical and "work on the bench with selected data with some corruption"
2025-11-14T08:45:39 < jpa-> that's what EE is about, coming up with the schematic and not just copying someone else's
2025-11-14T08:46:10 < Phantom> yes, but no. EE is also about making typical application schematics
2025-11-14T08:46:20 < Phantom> which seems to be totally missing
2025-11-14T08:46:35 < jpa-> easy parts you copy, but when you can't copy, you should be able to design from theory..
2025-11-14T08:50:49 < zyp> if you just want something that works, go copy a PoDL reference design
2025-11-14T08:51:08 < Phantom> zyp: that's the thing, I couln't find one
2025-11-14T08:51:56 < zyp> «I haven't searched and I'm all out of ideas»
2025-11-14T08:51:57 < zyp> https://www.google.com/search?q=podl+reference+design
2025-11-14T08:53:39 < sim_fentYl> actually, doing (online) research is about one thing LLMs don't suck in the embedded domain that much
2025-11-14T08:54:50 < Phantom> I'll check that again tomorrow
2025-11-14T09:01:15 < Phantom> to bed, I go, nite
2025-11-14T09:10:21 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-14T09:15:46 < qyx> sim_fentYl: they do
2025-11-14T09:15:56 < qyx> it was okay at the beginning
2025-11-14T09:16:07 < qyx> but now they impose some moderation and are unusable
2025-11-14T09:16:29 < qyx> "yes there exist some, do you want me to make a table?" yes of course
2025-11-14T09:16:45 < qyx> "i omited some resukts, do you want to include these?"
2025-11-14T09:16:59 < qyx> "they need another web search, may I perform I"
2025-11-14T09:17:04 < qyx> yes you lazy fuck
2025-11-14T09:17:11 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-14T09:17:29 < qyx> also android keyboard is being worse over time
2025-11-14T09:17:50 < sim_fentYl> well, I am not expecting it to get clean and correct result, just do the dirty job of getting the skeleton knowledge
2025-11-14T09:18:23 < sim_fentYl> as google got poisoned by copywriters and SEO, so it is basically unusable for most non-product of searching
2025-11-14T09:47:31 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-14T09:50:11 -!- c10ud [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-14T09:50:11 -!- c10ud [~c10ud@user/c10ud] has changed host
2025-11-14T10:08:21 -!- drew [~drew@user/drew] has left ##stm32 [WeeChat 4.1.1]
2025-11-14T10:45:25 -!- martinmoene [~martinmoe@132.229.46.129] has joined ##stm32
2025-11-14T11:20:45 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 265 seconds]
2025-11-14T13:54:44 < qyx> so, i need a connector for a ATE board
2025-11-14T13:55:04 < qyx> it switches some 12V things at 50A max and also has some basic DIO + analol input
2025-11-14T13:55:14 < qyx> what connector would I use?
2025-11-14T13:55:21  * qyx is eyeballing a DSUB
2025-11-14T14:32:23 < PhantomWork> GOOOOOOD MOOOORNIIIING!
2025-11-14T14:52:36 < karlp> ATE?
2025-11-14T15:01:55 < PhantomWork> zyp: is PoDL only for ethernet? The google results I get is only about that
2025-11-14T15:02:14 < karlp> it's the name used for how power is done in t1s and t1l.
2025-11-14T15:07:07 < PhantomWork> so that explain why I only find ethernet related stuff with that search/term... And some vague reference about power over CAN/RS485, which seems to be non-existant
2025-11-14T15:07:27 < PhantomWork> ethernet is too complex at multiple levels...
2025-11-14T15:07:47 < qyx> power over can/rs485 exists and works
2025-11-14T15:07:49 < qyx> to some degree
2025-11-14T15:08:46 < qyx> and yes, podl is ethernet's alternative to poe, which is ethernet's old alternative to plc
2025-11-14T15:08:57 < qyx> karlp: automatic test equipment
2025-11-14T15:11:41 < jpa-> PhantomWork: you should sleep more than 5 hours
2025-11-14T15:12:23 < PhantomWork> jpa-: I should, but I blame sleep issues since forever for this. I sleep when I'm tired, which is far from ideal...
2025-11-14T15:13:15 < PhantomWork> if I go to sleep when I'm not sleepy enough, I just don't sleep... so... 5h of sleep is better than 8 of turning over in the bed not sleeping
2025-11-14T15:13:39 < qyx> it is called circadian rhythm disorder
2025-11-14T15:14:32 < karlp> which is willfully inflicted on us by the forced changes to led street lighting, to keep us angry and easier to control.....
2025-11-14T15:14:45 < karlp> (conspiracy I saw on twitter yesterday)
2025-11-14T15:14:54 < qyx> sounds about right
2025-11-14T15:14:54 < jpa-> pull the foil hat down so that it covers your eyes too
2025-11-14T15:15:07 < karlp> jpa-: lol
2025-11-14T15:16:32 < qyx> ok I just discovered h-bridge is actually 4 transistors and not 2
2025-11-14T15:17:15 < PhantomWork> karlp: except, for me, it just make no sense, as I have a black roll down shade plus blackout curtains  :D  so your conspiracy theory don't work for me!
2025-11-14T15:17:16 < karlp> 2 is a half bridge
2025-11-14T15:17:23 < karlp> you dont' get a H with just half of it ;)
2025-11-14T15:17:37 < PhantomWork> I had a coworker that....
2025-11-14T15:17:45 < PhantomWork> ... said that...
2025-11-14T15:17:52 < qyx> I am fully aware of what half and full/H is, it just .. idk
2025-11-14T15:17:53 < PhantomWork> covid... 5G...
2025-11-14T15:17:58 < PhantomWork> and...
2025-11-14T15:18:10 < PhantomWork> bird ain't real, that it is a governement spy robot...
2025-11-14T15:18:27 < qyx> customer asking "oh yeah it is just 2 fets for reverse polarity protection and then 2 fets for the bridge.. that should be all"
2025-11-14T15:19:57 < qyx> now the question is whether I need to cool those mosfets if I only need one 50 A, 1 second pulse every minute or so
2025-11-14T15:20:16 < qyx> time to do some proper math
2025-11-14T15:27:14 < karlp> huh, zyp  just ran into a fun one with laks select mcu...
2025-11-14T15:27:24 < karlp> mem matching with numerical only doesnt' seem to work.
2025-11-14T15:27:36 < karlp> I had to quote numbers, to make them a string.
2025-11-14T15:27:57 < karlp> I suspect stm32s with 6 and 8 don't actually select. (would need a different tree with to try it out
2025-11-14T15:28:09 < karlp> anyway, papercut fixed, onwards...
2025-11-14T15:29:35 < qyx> karlp: was it you grumbling about BiSn not being avail in <1 mm?
2025-11-14T15:29:45 < karlp> no, I have no desire for bismuth 
2025-11-14T15:29:58 < karlp> smells like a jadew/ventyl thing?
2025-11-14T15:30:11 < qyx> idk how jadew or ventyl smells
2025-11-14T15:30:16 < karlp> like bismuth.
2025-11-14T15:31:40 < zyp> karlp, makes sense, sounds plausible it's broken on other parts, I've rarely used parts smaller than b
2025-11-14T15:32:19 < karlp> yeah, htat's what I suspected :)
2025-11-14T15:33:58 < jbo> karlp, what's your experience with laks? I still want to check it out some day (tm)
2025-11-14T15:34:25 < zyp> https://github.com/zyp/laks/blob/main/build/scons_tools/tool_platform_spec.py#L59 <- here's the line to fix, I'll accept a PR that coerces the left side to str if it's not None
2025-11-14T15:34:54 < karlp> jbo: I use it for experiments, and I'd like to use it more.
2025-11-14T15:35:13 < karlp> it's "easy" to get up and running with nice clean startup and then out of your way.
2025-11-14T15:35:19 < karlp> very easy to composit a system from pieces.
2025-11-14T15:35:32 < karlp> I woulud love it more if it wasn't ~abandonware ;)
2025-11-14T15:35:48 < jbo> yeah that would be my current issue too :D
2025-11-14T15:36:04 < karlp> well, now it's maintained by me I guess :)
2025-11-14T15:36:10 < zyp> :D
2025-11-14T15:36:39 < karlp> bu tyeah, all I do with it is use a nice clean shell that lets me skip opaque vendor startup messes
2025-11-14T15:37:05 < karlp> scons is still.... love/hate.
2025-11-14T15:37:24 < karlp> build variants still shits me, so finicky to get the # and object wrapping "just right"
2025-11-14T15:37:33 < zyp> you don't have to use scons, you know :p
2025-11-14T15:37:44 < karlp> sure, I can jinja it all myself.
2025-11-14T15:37:48 < karlp> no, not right now.
2025-11-14T15:38:07 < karlp> I mean, it mostly all "just works" 
2025-11-14T15:38:09 < zyp> no, I mean, there's a branch using erect instead of scons
2025-11-14T15:38:18 < karlp> oh right, even _less_ used code :)
2025-11-14T15:38:23 < zyp> :D
2025-11-14T15:38:25 < jbo> ahahahaaha
2025-11-14T15:38:26 < karlp> yeah, I vagueluy remember you doing that, but.....
2025-11-14T15:38:26 < zyp> https://github.com/zyp/laks/tree/dev_modules
2025-11-14T15:38:44 < jbo> and zyp has the audacity to say that I am used to like weird shit
2025-11-14T15:38:47 < karlp> like, seriouly, I don't need yaks on my yaks.  I already have yaks on my yaks on my yaks
2025-11-14T15:38:55 < jbo> s/used/known
2025-11-14T15:39:30 < karlp> zyp: personally I feel that "https://github.com/zyp/laks/blob/dev_modules/build.py" isnt' sufficient documentation for a user to start with it.
2025-11-14T15:41:30 < jbo> karlp, is the coro stuff feeling nice?
2025-11-14T15:41:43 < jbo> I was considering just stealing that
2025-11-14T15:41:55 < zyp> C++20 modules also still had maturity issues last I touched them
2025-11-14T15:41:57 < karlp> jbo: oh, I don't do any interesting c++ wank.
2025-11-14T15:42:02 < karlp> I just treat it as nicer C.
2025-11-14T15:42:16 < zyp> issues as in compiler crashing
2025-11-14T15:43:13 < jbo> karlp, oki
2025-11-14T15:43:19 < jbo> zyp, was the homework found this morning?
2025-11-14T15:43:32 < zyp> all good
2025-11-14T15:44:36 < karlp> oooh, pip install erect works...
2025-11-14T15:44:52 < jbo> pp install erect
2025-11-14T15:45:03 < karlp> so now instead of scons I have _different_ weird python :)
2025-11-14T15:45:03 < jbo> sure - checks out
2025-11-14T15:45:16 < karlp> anyway. back to the task at hand, it being almost 2pm already
2025-11-14T15:48:09 < zyp> erect is my «I'm tired of waiting for c++20 module support in scons, I'll just attempt my own thing that also fixes the other things I don't like about scons while I'm at it»
2025-11-14T15:59:02 < karlp> well, there's a milestone at least.
2025-11-14T15:59:17 < karlp> laks entry/startup, mcux "main" andclock config.
2025-11-14T15:59:28 < karlp> laks irq for usb just passing down to the mcux one.
2025-11-14T15:59:31 < karlp> and it still enumerates!
2025-11-14T15:59:52 < karlp> now... let's start pulling out the clock and board shit again in this direction and see if we can keep it working.
2025-11-14T16:00:06 < karlp> and we can finally graft k70 into mcux via laks.
2025-11-14T16:00:11 < karlp> I can smell the success.
2025-11-14T16:00:18 < karlp> it tastes like bismuth or something...
2025-11-14T16:00:37 < jbo> what are these bismuth references?
2025-11-14T16:04:54 < qyx> mount: /mnt/mmc1: unknown filesystem type 'exfat'.
2025-11-14T16:04:59 < qyx> meh
2025-11-14T16:07:08 < qyx> let's copy the whole image then
2025-11-14T16:11:20 < qyx> what the hell is systemd-resolved
2025-11-14T16:12:22 < qyx> but there is no resolvectl
2025-11-14T16:12:55 < jbo> qyx, did you end up doing PKMCS0909E4000-R1 beep boop?
2025-11-14T16:13:23 < jbo> looks like murata feels like no longer doing that
2025-11-14T16:15:31 < qyx> I did
2025-11-14T16:15:36 < qyx> although I ordered a compatible one
2025-11-14T16:16:18 < qyx> PKMCS0909E4000-R1, CPT-9019A-SMT
2025-11-14T16:17:47 < qyx> ok no way that copying 256GB microsd over 100mbit ethernet is gonna take 6 hours at least
2025-11-14T16:29:46 -!- ice [~ice@loud.house] has quit [Ping timeout: 255 seconds]
2025-11-14T16:34:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-14T16:36:05 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-14T16:40:40 < jbo> what is the "standard" voltage rating for 0402 100nf x7r mlccs?
2025-11-14T16:43:18 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-14T16:43:34 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-14T16:44:35 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 256 seconds]
2025-11-14T16:47:12 < karlp> well... now the whole laks clock setup works. with the code copy pastad from last time I tried this.
2025-11-14T16:47:17 -!- duude__ [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-14T16:47:33 < karlp> obviously there's some tiny difference somewhere, which I've not found, but "worked around"
2025-11-14T16:48:11 -!- duude__- [~duude__@user/duude/x-4676560] has quit [Ping timeout: 250 seconds]
2025-11-14T17:02:04 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-14T17:24:12 < jpa-> jbo: 25 or 50V, though you probably get quite a bit of derating above 20VDC
2025-11-14T17:56:48 < qyx> 50 V is a bit high for 0402
2025-11-14T17:56:54 < qyx> I would say even above the package rating
2025-11-14T17:57:49 < qyx> today I was somewhat curious if my 0603 100n survives 30V
2025-11-14T18:18:04 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-14T18:27:46 < PhantomWork> Almost time to go home already, yay!     atleast now the servers will shutdown in case of a blackout, which may happen this weekend
2025-11-14T18:28:06 < PhantomWork> 'bout time to set up nut completly... been 4 months since the new servers lol
2025-11-14T19:02:51 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-14T19:53:01 -!- ice [~ice@loud.house] has joined ##stm32
2025-11-14T20:08:55 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-14T21:35:05 < mawk> jbo do you know paul bismuth?
2025-11-14T21:39:29 < qyx> hudat
2025-11-14T21:47:17 < mawk> it's the alias of Nicolas Sarkozy that he used when committing his illegal shit
2025-11-14T21:47:23 < mawk> and why he went to jail
2025-11-14T21:47:31 < mawk> he actually stole the name from a guy he knew
2025-11-14T21:49:07 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-14T23:04:20 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-14T23:36:31 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
--- Day changed Sat Nov 15 2025
2025-11-15T00:07:27 < Steffanx> Your fav. President for sure mawk 
2025-11-15T00:07:54 < qyx> any of you pros can recommend a serious and working SD card reader?
2025-11-15T00:08:22 < qyx> unbelievable the shit keeps breaking
2025-11-15T00:10:30 < ds2> how is it breaking?
2025-11-15T00:10:44 < qyx> connectors losing contacts
2025-11-15T00:10:54 < ds2> USB side or SD side?
2025-11-15T00:10:54 < qyx> both usb and sd/microsd
2025-11-15T00:11:22 < qyx> they are all chink quality, I bet the usb cable is made of steel
2025-11-15T00:11:33 < ds2> wow... never had that happen and I use the cheapest possible ones i can
2025-11-15T00:11:35 < qyx> because there so much voltage drop it is being reset
2025-11-15T00:11:44 < ds2> Oh you have a tail on it?
2025-11-15T00:11:55 < ds2> (vs a dongle style)
2025-11-15T00:11:57 < qyx> one with, one without, and 3 other hubs with the same problem
2025-11-15T00:12:11 < qyx> consumer low quality sh*t
2025-11-15T00:12:35 < ds2> same here.... mine are packaged as a "free with purchase of cards"
2025-11-15T00:13:08 < qyx> raidsonic icybox should be ok
2025-11-15T00:13:12 < qyx> is it?
2025-11-15T00:13:15 < qyx> 15 eur for a SD reader
2025-11-15T00:13:20 < ds2> the plastics have broken on mine and they are often painfully slow but electrically, they work
2025-11-15T00:13:31 < qyx> https://www.alza.sk/icy-box-ib-cr301-u3-s-usb-3-0-d12522639.htm
2025-11-15T00:14:18 < ds2> why not ones without the tail?
2025-11-15T00:14:43 < qyx> one more connector to break
2025-11-15T00:15:04 < ds2> https://www.amazon.com/Reader-Adapter-Camera-Memory-Wansurs/dp/B0B9QZ4W4Y/
2025-11-15T00:15:06 < ds2> like that
2025-11-15T00:17:40 < qyx> hm idk
2025-11-15T00:18:26 < qyx> got another one working, doing 7MB/s
2025-11-15T00:27:57 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-15T00:37:07 -!- qyx_ [~qyx@84.245.120.43] has joined ##stm32
2025-11-15T00:37:33 -!- HelloShi1ty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-15T00:39:01 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 264 seconds]
2025-11-15T00:39:30 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-15T00:41:13 -!- Netsplit *.net <-> *.split quits: octorian, karlp, mid-kid, HelloShitty, PaulFertser, qyx
2025-11-15T00:44:22 -!- mid-kid [~mid-kid@2a01:7c8:aac8:1e8:5054:ff:fe5e:cd48] has joined ##stm32
2025-11-15T00:44:22 -!- Netsplit over, joins: PaulFertser, octorian, karlp
2025-11-15T00:55:40 -!- sim_fentYl is now known as ventYl
2025-11-15T05:01:35 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-15T05:44:24 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-15T05:44:31 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Read error: Connection reset by peer]
2025-11-15T06:50:46 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-15T07:11:20 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-15T07:44:54 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-15T08:05:49 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-15T08:41:59 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-15T09:06:32 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-15T10:40:02 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 244 seconds]
2025-11-15T10:46:42 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-15T11:14:35 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-15T12:32:44 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-15T12:39:08 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-15T12:50:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-15T14:42:25 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-15T14:44:29 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-15T16:02:01 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-15T16:50:05 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-15T16:54:08 < jbo> ventYl, NOOOOOOO
2025-11-15T17:02:44 < mawk> have you used this thing https://bitfontcreator--grayscale-software-informer-com.cdn.ampproject.org/v/s/bitfontcreator-grayscale.software.informer.com/amp/?amp_gsa=1&amp_js_v=a9&usqp=mq331AQIUAKwASCAAgM%3D#amp_ct=1763218921802&amp_tf=Source%C2%A0%3A%20%251%24s&aoh=17632188945760&referrer=https%3A%2F%2Fwww.google.com&ampshare=https%3A%2F%2Fbitfontcr
2025-11-15T17:02:44 < mawk> eator-grayscale.software.informer.com%2F
2025-11-15T17:02:50 < mawk> wow that's a nice url
2025-11-15T17:03:14 < mawk> bitfontcreator-grayscale.software.informer.com
2025-11-15T17:03:17 < mawk> should be better
2025-11-15T17:04:30 < Steffanx> Did you jbo?
2025-11-15T17:18:30 < mawk> what did jbo do
2025-11-15T17:18:33 < mawk> is it bad?
2025-11-15T17:18:35 < mawk> cancel jbo
2025-11-15T17:37:04 < Steffanx> Use it.
2025-11-15T17:37:09 < Steffanx> He's our gfx guy
2025-11-15T18:02:52 < mawk> I tried to find a low power LCD but it's all like 2-3mA consumption which is a lot
2025-11-15T18:02:59 < mawk> maybe I should do eink
2025-11-15T18:05:29 < BrainDamage> sharp memory lcd
2025-11-15T18:06:18 < BrainDamage> they are retroreflective lcd with an extra transistor to hold state ala tft, consumption in the order of uA
2025-11-15T18:11:41 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-15T18:16:20 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-15T18:43:17 -!- qyx_ is now known as qyx
2025-11-15T18:44:16 < qyx> mawk: there are definitely some in the sub-mA range
2025-11-15T18:44:22 < qyx> also as bd says
2025-11-15T18:44:40 < qyx> eink is power hungry during updates
2025-11-15T18:44:42 < mawk> yeah but that gets hard to obtain in single qty no?
2025-11-15T18:44:48 < mawk> and then it's not simply SPI
2025-11-15T18:45:12 < qyx> there are definitely some at buydisplay.com
2025-11-15T18:45:25 < qyx> I have bought some in the past
2025-11-15T18:45:37 < mawk> I found this but it's like 2-3mA https://www.crystalfontz.com/product/cfag12864u3nfh-128x64-low-power-graphic-monochrome-lcd
2025-11-15T18:45:58 < mawk> ah thanks let me look
2025-11-15T18:47:39 < qyx> eg ERC12864-20 specifies 450 uA max
2025-11-15T18:47:42 < mawk> eink might be fine, I don't expect a lot of updates, like maybe once an hour or so
2025-11-15T18:47:47 < qyx> excl. backlight of course
2025-11-15T18:47:52 < mawk> right
2025-11-15T18:48:52 < mawk> ah it wants 9V
2025-11-15T18:49:42 < mawk> then it need a charge pump I guess 
2025-11-15T18:52:05 < qyx> it is internal
2025-11-15T18:52:44 < Steffanx> 2-3mA is only twice 1-1.5mA mawk
2025-11-15T18:54:28 < qyx> mawk: see the interfacing zip, there is a schematic pdf inside
2025-11-15T18:54:33 < qyx> including spi
2025-11-15T18:55:01 < qyx> ok not including, it is spi *only*
2025-11-15T18:55:26 < qyx> it needs 3v3 and 4 caps only
2025-11-15T18:58:05 < mawk> ah nice
2025-11-15T18:58:11 < mawk> wrong Steffanx 
2025-11-15T18:58:20 < mawk> 1-1.5mA is like 0mA
2025-11-15T18:58:28 < mawk> but 2-3mA is like 20A
2025-11-15T18:58:33 < mawk> that's how maths works 
2025-11-15T18:58:42 < Steffanx> Aah that's why
2025-11-15T18:58:53 < qyx> nah 3 mA is like 3.1416 
2025-11-15T19:03:27 < qyx> BrainDamage: any place to get sharps besides digikey?
2025-11-15T19:05:48 < BrainDamage> it's been 10y since I used them in bulk, once upon a time mouser had them, nowdays it seems they only have dev kits
2025-11-15T19:08:29 < karlp> supply current for llcd: 450uA, nice.
2025-11-15T19:08:35 < karlp> supply current for backlight: 70mA.
2025-11-15T19:08:36 < karlp> ok.....
2025-11-15T19:11:28 < qyx> they are transflective, I am using it with much lower currents
2025-11-15T19:15:30 < BrainDamage> the transreflective lcd can be used just fine indoors with ambient illumination, it's the stuff that was in wristwatches and calculators until a short time ago, and still is there in cheaper models
2025-11-15T19:16:09 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-15T19:28:16  * qyx upgrades his debianz
2025-11-15T19:28:24 < Steffanx> Don't die
2025-11-15T19:28:41 < qyx> nah on my mobile 11" laptop only
2025-11-15T19:52:04 < karlp> hrm, there's a bunch of china xx116 sot23 single channel dc motor drivers
2025-11-15T19:52:14 < karlp> anyone know what that's all a clone of?
2025-11-15T20:02:10 < karlp> ok, little demo with the drv8833 module I got on ali some years back, a chopped up led strip and micropython works well.
2025-11-15T20:02:14 < karlp> can blink and fade and shit
2025-11-15T20:12:37 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-15T20:15:26 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-15T20:21:41 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-15T20:43:04 < Steffanx> Xmas toys karlp ?
2025-11-15T20:45:58 < karlp> no, been thinking about better control for halloween lights for years, finally playing with it.
2025-11-15T20:46:12 < karlp> went and looked at what's at the hardware store for chrristmas lights and it's all just random tuya shit.
2025-11-15T20:46:31 < karlp> I'm going to take over the world with my l33t prodcuts.
2025-11-15T20:46:49 < karlp> offering more expensive, more power hungry solutions that do the same thing, but "you can modify them!"
2025-11-15T20:47:05 < karlp> instead of being fixed with the couple of patterns and rough timer loops they come with standard.
2025-11-15T21:23:53 < mawk> it's too late for halloween
2025-11-15T21:40:43 < qyx> hrackarske banky
2025-11-15T21:40:46 < qyx> sry
2025-11-15T22:00:11 < mawk> hrcjrskrhckzj 
2025-11-15T22:03:32 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-15T22:04:10 < mawk> I stole a eink price display tag from the supermarket
2025-11-15T22:04:13 < mawk> idk if I can make it work
2025-11-15T22:05:42 < qyx> you will go to jail
2025-11-15T22:05:53 < mawk> there is a uart port on it
2025-11-15T22:05:54 < qyx> they will find you
2025-11-15T22:06:01 < mawk> :(
2025-11-15T22:28:26 < qyx> whoever uses python's kconfiglib, why the hell there are no 'menuconfig', 'defconfig', etc. commands anymore
2025-11-15T23:17:59  * qyx is removing 10y old legacy code
2025-11-15T23:34:27 < mawk> I took a linux console font and wrote some quick python to extract the glyphs I need
2025-11-15T23:34:55 < mawk> I can even do the funny characters like ┓ to draw boxes
2025-11-15T23:35:37 < qyx> that's great, we all tend to write the same piece of code all over again
2025-11-15T23:36:08 < mawk> lol
2025-11-15T23:36:24 < mawk> there's no pleasure in reusing someone else's code
2025-11-15T23:36:55 < mawk> I took the DejaVu font it's pretty
2025-11-15T23:37:05 < qyx> nah no worky in practice
2025-11-15T23:37:10 < qyx> been there, done that
2025-11-15T23:37:17 < qyx> what size? 1 bit only?
2025-11-15T23:37:36 < mawk> yeah it's just a bitmap there's no antialiasing, 1 bit per pixel
2025-11-15T23:37:43 < qyx> any pictures?
2025-11-15T23:37:54 < qyx> w is pretty shit for example
2025-11-15T23:38:16 < qyx> sorry Stef
2025-11-15T23:38:17 -!- fdarling [~forest@syn-096-033-104-194.biz.spectrum.com] has joined ##stm32
2025-11-15T23:38:19 < qyx> w doesn't look good
2025-11-15T23:39:33 < mawk> https://i.imgur.com/kNRtRDd.png
2025-11-15T23:39:50 < mawk> it's a bit ugly because it's zoomed but at the proper size it's fine
2025-11-15T23:40:08 < mawk> I like the roboto font from google too but I'm too lazy to convert it to the right format
2025-11-15T23:41:26 < qyx> mawk I know how dejavu fonts look like
2025-11-15T23:41:39 < qyx> I wanted to know how it looks on the lcd or drawn after your conversion
2025-11-15T23:41:57 < zyp> mawk, what brand/type?
2025-11-15T23:42:05 < mawk> ah I'm not there yet, I'm extracting the glyphs I want
2025-11-15T23:42:26 < qyx> I ended up using the nokia 3310 font
2025-11-15T23:42:28 < mawk> zyp: hanshow XL3N, apparently updated remotely with zigbee
2025-11-15T23:42:35 < qyx> it is super pro
2025-11-15T23:42:36 < mawk> and takes 4 lithium coin cells in parallel
2025-11-15T23:44:40 < Phantom> in parallel? that is unusual
2025-11-15T23:45:03 < mawk> I assume it's in parallel at least, it's 2 pairs of 2 cells, the cells being in parallel in each pair
2025-11-15T23:45:17 < zyp> heh
2025-11-15T23:45:21 < Phantom> maybe 2S2P?
2025-11-15T23:45:29 < mawk> you can't swap out the cells they're tack welded to the holder, so it better last long on battery
2025-11-15T23:45:36 < zyp> the ones I've got has a single prismatic primary cell
2025-11-15T23:45:56 < mawk> or spot welded
2025-11-15T23:46:09 < zyp> cell is soldered to the board and then enclosure is welded or glued shut
2025-11-15T23:46:55 < mawk> I took it when it was still in service mode so the screen shows the zigbee mac address or something, and apparently some youtube guys have reverse engineered it
2025-11-15T23:47:41 < Steffanx> Albert will not be happy mr mawk
2025-11-15T23:47:53 < mawk> yeah it's in parallel
2025-11-15T23:48:19 < mawk> it had slipped under a shelf Steffanx they had lost it anyway
2025-11-15T23:49:26 < Phantom> parallel also make sense, but it's unusual to see that. usually they just use a bigger cell
2025-11-15T23:49:32 < mawk> it's to make it flat
2025-11-15T23:50:28 < mawk> I see a debug header that says GND, RST, TX, VCC, RX, SWS; idk what SWS is
2025-11-15T23:50:54 < zyp> serial wire shit
2025-11-15T23:51:05 < Phantom> also, I hate bad drivers... https://i.imgur.com/gVYxAhb.jpg
2025-11-15T23:51:10 < qyx> single wire debug, but actually single this time
2025-11-15T23:51:25 < mawk> "single wire slave" apparently
2025-11-15T23:51:43 < zyp> here's the ones I've got: https://photos.app.goo.gl/6SK9KRUcusmosmdE7
2025-11-15T23:52:02 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-15T23:52:20 < qyx> the whole right rear door is bent?
2025-11-15T23:52:57 < mawk> I found custom firmware for it https://github.com/atc1441/ATC_TLSR_Paper
2025-11-15T23:53:18 < mawk> nice zyp 
2025-11-15T23:53:34 < qyx> oh is it rechargeable?
2025-11-15T23:53:36 < Phantom> qyx: no, just the bumper
2025-11-15T23:54:11 < qyx> I can see some highly risky knife job
2025-11-15T23:54:19 < Phantom> cracked bumper, broken brake light assy, broken reverse light assy, plus whatever is behind that is not visible
2025-11-15T23:54:25 < zyp> qyx, no, primary cell
2025-11-15T23:54:31 < zyp> first time I saw a prismatic one
2025-11-15T23:54:34 < qyx> whoa what what
2025-11-15T23:54:38 < qyx> me wanna for my shit
2025-11-15T23:55:00 < zyp> these things are specced for 8 years lifetime on a non-replaceable battery
2025-11-15T23:55:11 < qyx> a 58x50x10 mm one
2025-11-15T23:55:21 < qyx> if I may choose
2025-11-15T23:56:37 < zyp> I've got enough of the protocol reversed that I can send them arbitrary images, the video in the album is a replay of a command I captured to one tag, replayed to another
2025-11-15T23:56:40 < Phantom> https://i.imgur.com/cZp4qqE.jpg  <=== once I replaced the bumper in place, had to remove the light for that, but easy
2025-11-15T23:56:41 < qyx> actually if I stack 6 of yours I get 54x50x9 which is pretty close
2025-11-15T23:56:55 < zyp> and the last photo is me adding text to it
2025-11-15T23:57:44 < qyx> the pouch cell is apparantly made by Wuhan Fanso
2025-11-15T23:57:50 < zyp> it's pretty nice hardware, my idea is to label all my euroboxes with them, hence that first photo
2025-11-15T23:57:52 < qyx> covid included
2025-11-15T23:59:00 < zyp> the vendor subscription is pretty pricy, and I don't want the vendor system anyway :p
2025-11-15T23:59:18 < mawk> the firmware I found for mine has implemented BLE, which could be pretty useful
2025-11-15T23:59:27 < zyp> mine are BLE-based
2025-11-15T23:59:30 < mawk> it's also pretty large
2025-11-15T23:59:32 < mawk> ah nice
--- Day changed Sun Nov 16 2025
2025-11-16T00:00:27 < qyx> https://en.actec.dk/batteries/disposable-batteries/lithium-cylindrical-disposable-batteries/li-mno2-lithium-batterier/ultra-thin-lithium-batteries/3v-lithium-batteri-3000mah-ultra-tynd
2025-11-16T00:00:31 < qyx> 1000 mA max
2025-11-16T00:00:37 < qyx> that's maybe the reason they used it
2025-11-16T00:00:51 < zyp> the security in the protocol is pretty shit, if I were malicious I could go rewrite every tag in a store using these
2025-11-16T00:00:51 < qyx> 2 A Ipk, much more than a 2032 coin cell mawk has
2025-11-16T00:00:56 < mawk> that's not very cylindrical
2025-11-16T00:01:18 < qyx> he said prismatic
2025-11-16T00:01:33 < mawk> in the url it says cylindrical
2025-11-16T00:01:56 < zyp> could just have an app on my phone with a background service that listens for beacons for these tags, and when detected does a rewrite
2025-11-16T00:02:13 < zyp> then just put the phone in my pocket and take a walk through a store
2025-11-16T00:02:28 < mawk> mine is 4 2450 in parallel, capacity is something like 2400mAh total if I look at random 2450 cells
2025-11-16T00:02:43 < mawk> then lower the price and have the store honor it
2025-11-16T00:02:53 < zyp> :p
2025-11-16T00:03:16 < qyx> if I use 12 pcs of yours 2450, I get 7200 mAh
2025-11-16T00:03:28 < qyx> if I use 3 of the pouch ones, I am at 9000 mAh
2025-11-16T00:03:43 < zyp> I don't know of any stores using this brand of tags though
2025-11-16T00:03:55 < qyx> and in the same form factor I can use 3x AA LiSOCl2, which is about 8100 mAh
2025-11-16T00:06:40 < zyp> anyway, I'm a bit undecided if I want to stick with the vendor protocol, or replace the firmware
2025-11-16T00:07:26 < qyx> it may be a bit risky to disclose the vulnerale protocol if you want to make it public
2025-11-16T00:07:55 < qyx> and with a custom protocol you are not vendor locked
2025-11-16T00:08:19 < zyp> ideally I would have liked to reverse engineer the bootloader protocol so I could just rewrite the firmware without opening the enclosure, but I haven't done that
2025-11-16T00:08:36 < zyp> and if images are properly signed, that might not even be possible
2025-11-16T00:09:36 < zyp> so alternatively I've considered just milling a slot in the back of the enclosure so I can get pogos to the SWD pads
2025-11-16T00:10:06 < zyp> with a fixture on the cnc router at the hobby workshop, that should be a fairly quick operation even if I buy a ton of these
2025-11-16T00:16:26 < mawk> mine has the programming port available without breaking it open
2025-11-16T00:18:28 < mawk> but I can't find it on aliexpress
2025-11-16T00:47:46 -!- Livio_ [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-16T01:17:43 -!- qyx [~qyx@84.245.120.43] has quit [Ping timeout: 264 seconds]
2025-11-16T01:19:12 -!- qyx [~qyx@84.245.121.177] has joined ##stm32
2025-11-16T05:45:11 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-16T07:19:02 -!- jerrycash2 [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-16T07:22:31 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 264 seconds]
2025-11-16T07:27:37 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-16T08:06:37 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 264 seconds]
2025-11-16T09:02:04 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-16T10:48:48 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-16T11:05:51 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-16T11:13:34 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Quit: KVIrc 5.2.6 Quasar http://www.kvirc.net/]
2025-11-16T11:16:38 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-16T11:28:09 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-16T12:08:48 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-16T14:05:49 -!- Ice_Cube [~PC@176.194.180.156] has joined ##stm32
2025-11-16T14:19:15 -!- Ice_Cube [~PC@176.194.180.156] has quit [Remote host closed the connection]
2025-11-16T14:33:22 < qyx> warning: initializer element is not a constant expression [-Wpedantic]
2025-11-16T14:33:26 < qyx> what the hell is that
2025-11-16T14:33:43 < qyx> I tried #define LED_SEQ_SET 0x02
2025-11-16T14:33:48 < qyx> also #define LED_SEQ_SET 0x02UL
2025-11-16T14:33:58 < qyx> also enum {LED_SEQ_SET = 0x02};
2025-11-16T14:34:21 < ventYl> are you initializing something that's const by pointer to something else that's const?
2025-11-16T14:36:13 < jpa-> time to create a minimal example, it's probably caused by something else than just that bit of code
2025-11-16T14:38:45 -!- HelloShi1ty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 252 seconds]
2025-11-16T14:40:45 < qyx> https://godbolt.org/z/x3oj6vhj7
2025-11-16T14:40:48 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has joined ##stm32
2025-11-16T14:42:44 < qyx> wait what
2025-11-16T14:43:07 < qyx> this https://godbolt.org/z/Mz1Y17od9
2025-11-16T14:43:12 < qyx> with the right compiler and option
2025-11-16T14:48:55 < karlp> how do I unfuck the layout you did in the last one
2025-11-16T14:49:44 < karlp> ahh, just drag anddrop
2025-11-16T14:51:48 < karlp> are you running into something the seq_rgb overflowing on the shifts?
2025-11-16T14:52:22 < karlp> yeah, with brightness set to 40 it's ok.
2025-11-16T14:52:35 < karlp> 255 is too much, I think you're doing unsigned 32bit overflow or something
2025-11-16T14:52:51 < qyx> no, open the compiler output window
2025-11-16T14:52:53 < karlp> yeah, brightness is now max 254
2025-11-16T14:52:58 < qyx> I am not that far
2025-11-16T14:53:16 < qyx> i am hitting
2025-11-16T14:53:18 < qyx> <source>:12:21: warning: initializer element is not a constant expression [-Wpedantic] 12 | #define LED_SEQ_SET 0x2
2025-11-16T14:53:20 < karlp> I know,
2025-11-16T14:53:28 < karlp> but seriously, make brightness 254
2025-11-16T14:53:34 < karlp> and you 'll seee it all go away
2025-11-16T14:53:44 < qyx> wat
2025-11-16T14:53:47 < karlp> I removed the LED_SEQ_SET from the static const, to see what else it didn't like.
2025-11-16T14:53:55 < karlp> and then you see it start to get upset about the rgb
2025-11-16T14:54:09 < karlp> you'redoign undefined left shift of something
2025-11-16T14:55:26 < karlp> something is wild though, that really makes the diags all fucked.
2025-11-16T14:55:50 < qyx> oh I have to do 255UL
2025-11-16T14:56:28 < karlp> just "u" should be enough, I guess.
2025-11-16T14:56:36 < qyx> also 0UL
2025-11-16T14:56:44 < karlp> those diags are erally unhelpful though :)
2025-11-16T14:57:01 < karlp> this will teach you for using macros
2025-11-16T14:57:43 < qyx> same with enums
2025-11-16T14:57:50 < qyx> they are signed and generated the same warning
2025-11-16T14:58:05 < karlp> it's the SEQ macros,
2025-11-16T14:58:08 < karlp> not the types
2025-11-16T14:58:21 < qyx> yes 'u' works too
2025-11-16T14:58:46 < qyx> I mean I replaced SEQ macros with enums and the outcome is the same
2025-11-16T14:59:07 < qyx> because all the interweb is don't use macros, use enum, be cool
2025-11-16T15:00:10 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-16T15:00:47 < karlp> it's the function macros, like SEQ_RGB that are the problem, not the enum vs define for the ids.
2025-11-16T15:02:52 < karlp> needs constexpr or more c++ or something....
2025-11-16T16:18:46 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-16T17:04:56 -!- zChris [~zChris@user/zchris] has quit [Quit: ZNC 1.8.2 - https://znc.in]
2025-11-16T17:06:16 -!- zChris [~zChris@user/zchris] has joined ##stm32
2025-11-16T17:09:37 -!- fdarling [~forest@syn-096-033-104-194.biz.spectrum.com] has quit [Quit: Leaving]
2025-11-16T17:09:53 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-16T17:29:20 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-16T17:40:01 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-16T18:13:44 < nohit> when using global sourcing with jlcpcb, what do you put in the LCSC field for those parts when you generate BOM file
2025-11-16T18:34:35 < jpa-> i think if you preorder, you get a private lcsc number
2025-11-16T18:34:53 < jpa-> otherwise i'd put the manufacturer part number
2025-11-16T18:44:09 < Steffanx> nohit jlcpcb support tends to be helpful for questions like this. At least in my experience ;)
2025-11-16T19:12:49 < nohit> ok thanks
2025-11-16T19:49:53 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-16T19:49:53 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 265 seconds]
2025-11-16T19:56:06 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-16T20:07:33 < qyx> 	if (color == 0xffffffu) {
2025-11-16T21:03:05 -!- akaWolf [~akaWolf@akawolf.org] has joined ##stm32
2025-11-16T21:06:00 < antto> racism in conditional statement detected
2025-11-16T21:11:53 < specing> LOL
2025-11-16T21:38:55 -!- artok [~azo@88-112-154-28.elisa-laajakaista.fi] has joined ##stm32
2025-11-16T22:08:11 -!- travisb_ [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has quit [Remote host closed the connection]
2025-11-16T22:10:52 -!- travisb [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-16T22:33:19 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 240 seconds]
2025-11-16T23:03:25 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-16T23:07:16 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-16T23:11:42 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-16T23:21:11 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-16T23:25:21 -!- Livio_ [~livio@user/livio] has quit [Remote host closed the connection]
--- Day changed Mon Nov 17 2025
2025-11-17T00:10:02 -!- hexo [~hexo@user/hexo] has quit [Remote host closed the connection]
2025-11-17T00:21:41 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-17T00:37:24 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-17T04:03:19 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-17T04:53:05 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 250 seconds]
2025-11-17T04:53:18 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-17T04:55:13 -!- duude__- is now known as duude__
2025-11-17T05:13:32 -!- duude__ [~duude__@user/duude/x-4676560] has quit [Ping timeout: 256 seconds]
2025-11-17T05:13:36 -!- duude__- [~duude__@user/duude/x-4676560] has joined ##stm32
2025-11-17T05:15:30 -!- duude__- is now known as duude__
2025-11-17T07:26:27 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-17T07:50:50 -!- qyx_ [~qyx@84.245.121.177] has joined ##stm32
2025-11-17T07:52:53 -!- akaWolf1 [~akaWolf@akawolf.org] has joined ##stm32
2025-11-17T07:56:41 -!- Posterdati [~Posterdat@user/Posterdati] has quit [Ping timeout: 256 seconds]
2025-11-17T07:56:41 -!- qyx [~qyx@84.245.121.177] has quit [Ping timeout: 256 seconds]
2025-11-17T07:56:41 -!- akaWolf [~akaWolf@akawolf.org] has quit [Ping timeout: 256 seconds]
2025-11-17T07:56:41 -!- jerrycash2 [~jerrycash@user/jerrycash] has quit [Ping timeout: 256 seconds]
2025-11-17T07:56:56 -!- akaWolf1 is now known as akaWolf
2025-11-17T08:03:22 -!- Posterdati [~Posterdat@user/Posterdati] has joined ##stm32
2025-11-17T08:31:06 -!- Lt_264_user45 [~losefka@139.94.83.71] has joined ##stm32
2025-11-17T08:31:41 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-17T08:46:48 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-17T08:53:24 -!- Lt_264_user45 [~losefka@139.94.83.71] has quit [Ping timeout: 256 seconds]
2025-11-17T09:20:14 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-17T09:45:01 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-17T10:01:10 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 245 seconds]
2025-11-17T10:33:32 -!- qyx_ is now known as qyx
2025-11-17T10:53:09 < mawk> is it useful to put a RC damper next to the ADC pins for my 20m long thermistor cable? or should that have been done next to the thermistor
2025-11-17T10:53:27 < mawk> because apparently it's acting like an antenna or something because the temperature I read keeps jumping around ±1⁰C
2025-11-17T10:54:48 < mawk> it was only jumping around ±0.1⁰C when reading the thermistor directly next to it
2025-11-17T10:57:16 < mawk> if I was able to read all three pins simultaneously it would probably help, I have the current sense resistor and the two pins for the voltage drop across the thermistor; but I have to read them like 20μs apart as there's only one ADC
2025-11-17T10:59:54 < qyx> yes you need a pretty massive filter for a 20m analol cable
2025-11-17T11:01:54 < qyx> this is insane, weekend of weird code
2025-11-17T11:02:40 < qyx> I have a parameter r ge25519_p1p1 *r, of type which is a struct and contains bignum25519 x, y, z, t;
2025-11-17T11:02:53 < qyx> then they do 	bignum25519 *rb = (bignum25519 *)r;
2025-11-17T11:02:58 < qyx> like, what
2025-11-17T11:03:17 < qyx> then they are using it like
2025-11-17T11:03:18 < qyx> 	curve25519_add(rb[2+signbit], rb[2+signbit], c); /* z for +, t for - */
2025-11-17T11:03:23 < qyx> who the hell writes such code
2025-11-17T11:03:26 < qyx> I know, djb
2025-11-17T11:03:42 < qyx> then I open some quality zyp's code to make the world better
2025-11-17T11:04:01 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 250 seconds]
2025-11-17T11:04:36 < qyx> and see enum Type { PushPull, OpenDrain, };
2025-11-17T11:04:53 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-17T11:07:56 < qyx> and later.. if(t) {g->OTYPER |= 1 << n;
2025-11-17T11:08:50 < qyx> how the hell could zyp write an enum without explicitly specified values and yet depends on them
2025-11-17T11:16:32 < jpa-> mawk: it's useful to put a capacitor right next to ADC pins, a resistor or ferrite can help too
2025-11-17T11:17:00 < jpa-> but i'd say a capacitor is the most important bit
2025-11-17T11:17:05 < mawk> if it's the first element of the struct it's a bit cavalier but it's fine
2025-11-17T11:17:59 < mawk> jpa- does that affect the impedance seen by the ADC? I guess so
2025-11-17T11:19:40 < jpa-> yes, the ADC needs to charge a 4pF capacitor during the sampling time, so it is faster to get the charge from a local capacitor
2025-11-17T11:20:07 < jpa-> another effect is that any capacitive coupling to your cable is attenuated in relation to the parallel capacitor
2025-11-17T11:20:18 < mawk> right 
2025-11-17T11:20:49 < mawk> yeah currently I had to put the maximum sample time because of the high impedance seen by the ADC
2025-11-17T11:20:50 < qyx> mawk: it is not, if it is not specified, I am free to move the items around without making the code break
2025-11-17T11:20:57 < qyx> the same goes for the ed25519 sources
2025-11-17T11:21:02 < mawk> yeah
2025-11-17T11:21:31 < qyx> casting like a god
2025-11-17T11:21:33 < qyx> *dog
2025-11-17T11:21:38 < mawk> lol
2025-11-17T11:21:51 < mawk> the compiler doesn't complain?
2025-11-17T11:21:57 < mawk> maybe it needs -pedantic
2025-11-17T11:22:00 < qyx> of course it does but who cares
2025-11-17T11:22:07 < qyx> and that's a reference impl
2025-11-17T11:22:21 < qyx> I hate mathematicians
2025-11-17T11:22:29 < mawk> lol
2025-11-17T11:22:48 < karlp> qyx: who would go and modify the enum thouh? :)
2025-11-17T11:22:50 < jpa-> they could have written it as    struct { union { bignum25519 parts[4]; struct { bignum25519 x, y, z, t; } } } and qyx would be happy
2025-11-17T11:22:52 < mawk> some mathematics enjoyers write beautiful code
2025-11-17T11:23:23 < mawk> anonymous structs and unions are awesome 
2025-11-17T11:24:24 < qyx> jpa-: yes I would definitely be and the compiler too
2025-11-17T11:24:59 < jpa-> and zyp could have just put the = 0 and = 1
2025-11-17T11:25:06 < qyx> is it that hard to write  if (sign) { access z; } else { access t; }
2025-11-17T11:25:13 < qyx> yes
2025-11-17T11:25:23 < jpa-> maybe they wanted it branchless
2025-11-17T11:25:31 < jpa-> crypto people are sometimes afraid of branches
2025-11-17T11:26:50 < qyx> so they should invent a better way to do it
2025-11-17T11:26:53 < BrainDamage> qyx: comparisons have to be done in constant time to avoid leaking information
2025-11-17T11:27:40 < qyx> I know what a timing attack is but that's not an excuse
2025-11-17T11:27:44 < zyp> qyx, sounds like you want rust
2025-11-17T11:27:56 < jpa-> why is it not an excuse?
2025-11-17T11:28:17 < zyp> in rust, if you want to use the numeric enum values, you have to specify them explicitly :)
2025-11-17T11:28:22 < qyx> excuse for writing unreadable and broken code
2025-11-17T11:28:35 < qyx> because it could be done as you said
2025-11-17T11:28:40 < jpa-> is it actually broken? i would assume it passes test cases?
2025-11-17T11:28:48 < jpa-> and if you go and break it, test cases show it
2025-11-17T11:28:54 < qyx> it is broken, because it doesn§t pass -Werror
2025-11-17T11:28:54 < jpa-> so it doesn't seem that big of a deal
2025-11-17T11:29:25 < BrainDamage> mawk: if you haven't done so already, clamp ferrite beads onthe cable near both ends, that will kill common mode which is a significant portion of the induced rfi
2025-11-17T11:29:25 < qyx> also -pedantic, because it has semicolons and shit spread around the code
2025-11-17T11:30:01 < qyx> zyp: ocpd helps too, but yes, thats exactly the case and a possible fix
2025-11-17T11:30:23 < qyx> jpa-: also the code is not const-correct
2025-11-17T11:30:39 < karlp> worksonmymachine.pro
2025-11-17T11:30:39 < qyx> full of warnings cast discards 'const' qualifier blah blah
2025-11-17T11:30:54 < qyx> the ed25519 I mean
2025-11-17T11:30:57 < karlp> don't question the crypto bois, they will just sneer at you and tell you're wrong
2025-11-17T11:31:00 < mawk> nice BrainDamage , I think I have one at least 
2025-11-17T11:31:31 < qyx> also, in other news, things broke on monday morning completely
2025-11-17T11:31:40 < qyx> good start of the week
2025-11-17T11:31:49 < jpa-> i still wish i'd have some nice way to make nanopb fully const-correct
2025-11-17T11:32:03 < jpa-> i want to use common function for iterating fields, but sometimes i have a const struct and sometimes non-const
2025-11-17T11:33:11 < qyx> oh and also this casting code doesn't even consider the compiler is free to add padding between the elements
2025-11-17T11:35:37 < jpa-> kinda free, but when would it actually pad bignum25519 x,y,z,t; and not pad bignum25519 xyzt[4]?
2025-11-17T11:36:48 < jpa-> language lawyers often try to make C into something it is not (a well defined and really safe language)
2025-11-17T11:37:46 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-17T11:41:31 < qyx> warning: "unreachable" redefined
2025-11-17T11:41:32 < qyx> another one
2025-11-17T11:41:48 < jpa-> isn't redefining variables perfectly legal C?
2025-11-17T11:44:28 < karlp> yeah, building a big old project with -werror is something only project maintainers should ever do.
2025-11-17T11:48:38 < qyx> I guess I have no nerves for this
2025-11-17T11:52:55 < qyx> then some other guy uses FLT16_MANT_DIG > 0 to check if FLT16_MANT_DIG is defined...
2025-11-17T11:53:11 < qyx> warning: "FLT16_MANT_DIG" is not defined, evaluates to 0 [-Wundef]
2025-11-17T11:53:16 < qyx> oh there is a comment regarding this
2025-11-17T11:53:41 < qyx> /* Check for FLT16_MANT_DIG using integer comparison. Clang headers incorrectly * define this macro unconditionally when __STDC_WANT_IEC_60559_TYPES_EXT__ * is defined (regardless of actual support for _Float16).
2025-11-17T11:54:04 < qyx> so blame clang and broke the code for all other compilers
2025-11-17T11:54:50 < jpa-> could have done   #if defined(FLT16_MANT_DIG) && FLT16_MANT_DIG > 0
2025-11-17T12:05:49 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-17T12:27:24 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-17T12:39:06 < mawk> or check for clang
2025-11-17T12:53:58 < mawk> nrf54l15 looks sweet
2025-11-17T12:54:07 < mawk> there's a risc-v coprocessor 
2025-11-17T12:56:16 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-17T13:34:40 < zyp> mawk, the coprocessor is a bit ridiculous
2025-11-17T13:34:45 < mawk> I added a random choke-shaped object and 220nF across the thermistor, and a 100μF to ground on the sense resistor, and now it's much more stable BrainDamage 
2025-11-17T13:35:01 < mawk> but the choke and the 220pF do apparently nothing, the 100μF seems enough 
2025-11-17T13:35:03 < zyp> currently, the only official purpose of the coprocessor is bitbanging QSPI: https://docs.nordicsemi.com/bundle/ncs-latest/page/nrfxlib/softperipheral/doc/introduction.html
2025-11-17T13:35:09 < mawk> lol
2025-11-17T13:40:45  * karlp dances the success dance.
2025-11-17T13:41:02 < karlp> stubbed out just the bare minimum of mcux to get k70 support with working usb!
2025-11-17T13:41:05 < karlp> whee...
2025-11-17T13:41:33 < karlp> exactly as I foretold, the final thing was just switching target in laks from k64 to k70 and then flashing the target board instead of the frdm-k64 board and... it just worked. bang.
2025-11-17T13:41:36 < karlp> winning!
2025-11-17T13:53:46 < mawk> nice
2025-11-17T13:54:57 < ventYl> I could use some knowledge gathered while making that stub
2025-11-17T14:03:21 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-17T14:21:46 < karlp> in what part?
2025-11-17T14:23:42 < ventYl> I tried do it, but soon it seemed that mcux will eventually drag the whole SDK in or you have to stub quite a lot of stuff
2025-11-17T14:23:59 < ventYl> my goal was to just include startup and system files and CMSIS headers
2025-11-17T14:24:08 < karlp> there's nothign actually "private" in this code, let me dig up a url for you...
2025-11-17T14:24:10 < karlp> https://github.com/karlp/mcux-sdk-frdm-k64f/tree/k70-support
2025-11-17T14:24:24 < karlp> i've used laks for everything excep tthe usb middleware.
2025-11-17T14:24:43 < karlp> https://github.com/karlp/mcux-sdk-frdm-k64f/blob/k70-support/boards/frdmk64f/usb_examples/usb_host_hid_mouse_keyboard/bm/SConstruct
2025-11-17T14:25:17 < karlp> I used upstream fsl_common.h, but commented out fsl_common_arm.h incldues, as that has a lot of overlap with laks stuff
2025-11-17T14:25:35 < ventYl> meh. i am blind. mcux vs. nrfx
2025-11-17T14:25:49 < karlp> yeah, didn't think you were doing any kinetis shits...
2025-11-17T14:26:17 < ventYl> actually I was, but MKL81 has some NXP-specific MPU which is not compatible with Cortex-M MPU
2025-11-17T14:27:19 < ventYl> and last weekend on conference, guy from NXP was giving away links to supposedly internal website where you could order NXP devboards for free with words: "I shouldn't share it with you, but it could be find by using Google anyway."
2025-11-17T14:27:33 < ventYl> so I might order some NXPs
2025-11-17T14:28:16 < ventYl> and oh, CMRX has an intern who wants to add NXP support anyway
2025-11-17T15:38:34 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 244 seconds]
2025-11-17T15:44:57 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-17T15:45:04 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Read error: Connection reset by peer]
2025-11-17T15:55:25 < qyx> poor intern
2025-11-17T15:55:31 < qyx> dealin with nxp
2025-11-17T16:07:06 < ventYl> currently he is coming to terms with specifics of interacting with HALs from within memory protected contexts
2025-11-17T16:07:15 < ventYl> something, no HAL was ever designed to cope with
2025-11-17T16:12:50 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-17T17:00:18 < karlp> lots of non-communication :)
2025-11-17T17:00:26 < karlp> kinetis k64/k70 boots with MPU enabled.
2025-11-17T17:00:40 < karlp> every single example code automatically turns it off immeidately.
2025-11-17T17:01:02 < karlp> (that was a fun trap, usb peripheral didn't hve access to packet ram)
2025-11-17T17:16:13 < ventYl> Yeah, so these have NXP-specific MPU too. Cortex-M MPU does not affect peripherals
2025-11-17T17:16:22 < ventYl> I can't decide if designing this was a smart or stupid move.
2025-11-17T17:28:06 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-17T19:34:24 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-17T19:52:31 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-17T21:22:32 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 256 seconds]
2025-11-17T22:04:56 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 244 seconds]
2025-11-17T22:50:42 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-17T23:16:20 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-17T23:24:58 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-17T23:52:26 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
--- Day changed Tue Nov 18 2025
2025-11-18T00:39:27 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-18T00:46:55 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-18T00:50:30 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-18T00:55:33 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has joined ##stm32
2025-11-18T00:55:43 < nomorekaki> official night pump
2025-11-18T00:56:45 < nomorekaki> is crew ready?
2025-11-18T00:56:55 < qyx> partially
2025-11-18T00:57:40 < nomorekaki> how are projects
2025-11-18T00:58:07 < nomorekaki> are you scheduling chilling?
2025-11-18T01:17:49 -!- qyx [~qyx@84.245.121.177] has quit [Ping timeout: 246 seconds]
2025-11-18T01:19:52 -!- qyx [~qyx@84.245.121.157] has joined ##stm32
2025-11-18T01:20:03 < ventYl> i am printing
2025-11-18T01:28:31 -!- ventYl [~ventyl@adsl-dyn152.78-98-151.t-com.sk] has quit [Ping timeout: 240 seconds]
2025-11-18T01:30:25 -!- ventYl [~ventyl@adsl-dyn130.78-99-211.t-com.sk] has joined ##stm32
2025-11-18T01:33:01 < qyx> not enough electricity for that 3d printing gigafactory
2025-11-18T01:34:10 -!- hexo [~hexo@user/hexo] has joined ##stm32
2025-11-18T01:35:16 < ventYl> and also disconnecting, apparently
2025-11-18T05:14:47 -!- nomorekaki [~nomorekak@176-93-71-17.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-18T07:27:16 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-18T09:10:06 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-18T09:14:46 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-18T09:45:15 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-18T09:52:53 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-18T09:54:29 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-18T10:13:12 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-18T11:57:29 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-18T12:00:52 -!- HelloShitty [~psysc0rpi@bl18-184-235.dsl.telepac.pt] has quit [Ping timeout: 256 seconds]
2025-11-18T12:03:10 -!- HelloShitty [~psysc0rpi@bl18-184-247.dsl.telepac.pt] has joined ##stm32
2025-11-18T12:04:32 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-18T12:22:35 < karlp> fecking, rename a directory of example code, update the name in the scons file...
2025-11-18T12:22:42 < karlp> now it can't find some file in another component?!
2025-11-18T12:22:50 < karlp> so strange sometimes.
2025-11-18T12:23:23 < karlp> it's like something doesn't have a dep right, and it worked before because some ordering was different?
2025-11-18T12:44:44 < ventYl> fixed include path?
2025-11-18T13:10:32 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-18T13:17:30 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-18T13:34:15 < karlp> yeah, I figured it out, I didn't fix all the paths. my own footgun.
2025-11-18T13:34:37 < karlp> I'm tryign to clean up some splatter I made while hacking this back and forth.
2025-11-18T13:45:09 < ventYl> well, I like the "modern CMake" approach, where creating a dependency on a module (library, archive, whatever) may inject include paths and flags into dependent module
2025-11-18T13:46:33 < ventYl> the only aspect where no good solution exists is what include path to inject -> if #include <module/header.h> or #module <header.h> will work
2025-11-18T13:46:59 < ventYl> former generates additional work with renaming paths if you rename the module, latter creates possibility of header name collisions
2025-11-18T13:47:33 < Steffanx> Don't rename your libs :)
2025-11-18T13:58:02 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-18T14:03:31 < jpa-> https://i.imgur.com/5bPuxZk.png   what do these weird new symbols kicad puts on pin names mean?
2025-11-18T14:03:48 < jpa-> it doesn't appear to depend on pin type (bidirectional or other) but (maybe?) on pin name
2025-11-18T14:04:31 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-18T14:08:45 < jpa-> aaah, it has a fancy new "alternate assignment" function
2025-11-18T14:10:07 < jpa-> right-click the pin and you can select it to show up as e.g. "ADC1_IN0" instead of "PA0", so that when you go make a gpio mapping for the firmware, you have to look up each pin separately by pin number instead
2025-11-18T14:10:12 < karlp> fecking, ok.  on k64, mice, keyboards and the barcode scanner all work.
2025-11-18T14:10:24 < karlp> on k70, the mice and scanners work but... not the keyboards?!
2025-11-18T14:10:28 < karlp> I thought I was fucking done.
2025-11-18T14:10:55 < karlp> jpa-: you can use symbols that have long strings instead if you prefer.
2025-11-18T14:11:05 < karlp> this is being all things to all people.
2025-11-18T14:11:10 < jpa-> yeah
2025-11-18T14:11:30 < jpa-> i guess it would be nice for them to name the library symbol alternative functions as   PA0 (ADC1_IN0)  or something
2025-11-18T14:11:38 < karlp> I think that keeping "PnD" as a rpefix would have been more useufl thoug
2025-11-18T14:11:48 < karlp> yeah, and that seem slike worthwhile feedback...
2025-11-18T14:12:33 < karlp> now we just need orcad to let you re-arrange the pins on the symbol, like they're a heirarchical sheet,
2025-11-18T14:12:41 < karlp> so you can make it "neatest" for this particular project...
2025-11-18T14:13:04 < jpa-> i wonder if the kicad library feedback is now more pleasant.. a few years back the library maintenance seemed pretty annoying to try to contribute to
2025-11-18T14:14:46 < jpa-> https://gitlab.com/kicad/code/kicad/-/issues/17594   looks like it is already reported
2025-11-18T14:19:47 < jpa-> karlp: funnily in pcbnews you can select the "Allow free pads" in case you want to easily move footprint pads (why would you do that), but in eeschema there is no equivalent :)
2025-11-18T14:25:00 < ventYl> meh, did freecad become completely broken, or is it just me?
2025-11-18T14:27:22 < jpa-> just you
2025-11-18T14:27:26 < jpa-> freecad is better than ever
2025-11-18T14:29:36 < ventYl> i've drawn sketch, extuded it, OK. drawn another sketch, extruded it, random things in model went into hidden mode and I could not make 2nd extrusion to show up
2025-11-18T14:33:42 < jpa-> i think i have had something similar happen before.. not sure what causes it, but ok away from the task and then make sure the latest object in the tree is visible
2025-11-18T14:35:14 < ventYl> tried that, could not make it show up even after I jumped out of the task
2025-11-18T14:35:31 < jpa-> weird
2025-11-18T14:35:54 < ventYl> yeah, it used to be pretty worky before
2025-11-18T14:38:30 < karlp> jpa-: yeah, I never understood the "allow free pads" options, what fucking niche was that solving?!
2025-11-18T14:42:54 < jpa-> karlp: maybe it was meant for through-hole resistors or something, but still seems silly
2025-11-18T14:45:42 < karlp> apparently one? use case for it is, "I have a 10 terminal symbol, for a wire harness, and I solder each wire to pads directly, and i move the pads in my "10 pad footprint" freely to where I hafve space for those wires.
2025-11-18T14:45:50 < karlp> which sounds like a super janky way of working... but... ok
2025-11-18T14:46:20 < karlp> someone from allegro used it: https://forum.kicad.info/t/can-the-allow-free-pads-feature-apply-to-a-footprint/28756
2025-11-18T14:46:52 < karlp> "allow free pads" was added so it could be turned off, tostop people accidentally doing it,
2025-11-18T14:47:05 < karlp> so yeah, it should never have been in, now we have a dangling compat flag... whee
2025-11-18T14:47:53 < karlp> ok. now... keyboards not enumerating, but mice and scanners are... but only when I switch platform. what's likely to be the problem....
2025-11-18T15:24:02 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-18T15:24:33 < karlp> fecking, keypads and mice are both LS. I thought I had htis solved.
2025-11-18T15:44:23 < karlp> weirdly, on k64 with the beagle, I never see errors.
2025-11-18T15:44:39 < karlp> on k70, i see errors.... but the mice always seem to sort it out.
2025-11-18T15:44:46 < karlp> and FS devices are fine..
2025-11-18T15:44:51 < karlp> zero errors there.
2025-11-18T16:04:42 < karlp> n other news, 1920x1080 in 0.71" screen?: https://www.aliexpress.com/item/1005009794940505.html
2025-11-18T16:05:16 < ventYl> years ago where were toilet stall data projectors with very similar performance :D
2025-11-18T16:36:05 < BrainDamage> that's quite cute
2025-11-18T16:36:08 < mawk> I want to get 2500 ADC readings each 1μs apart, every 250ms; so I set up one timer in output compare mode so that the trigger output is only on 10% of the time, with a period of 250ms, and another timer as a slave in gated mode with an update even every 1μs which is then fed to the ADC
2025-11-18T16:36:13 < mawk> does that sound safe enough
2025-11-18T16:36:35 < mawk> maybe I could've used PWM mode for the first timer, but it seems like it would do the same thing
2025-11-18T16:41:27 < jpa-> sounds reasonable
2025-11-18T16:51:09 < zyp> I probably would just have used a free running trigger timer, set up a DMA job for 2500 samples, enable ADC, wait for DMA complete interrupt, disable ADC, wait for next interval, set up a new DMA job, and so on
2025-11-18T16:55:36 < jpa-> me too, but if exactly cycle-accurate 250 ms interval is needed, that gets harder
2025-11-18T17:02:20 -!- zChris_ [~zChris@user/zchris] has joined ##stm32
2025-11-18T17:08:26 -!- Netsplit *.net <-> *.split quits: octorian, mid-kid, PaulFertser, karlp, zChris
2025-11-18T17:11:01 -!- mid-kid [~mid-kid@2a01:7c8:aac8:1e8:5054:ff:fe5e:cd48] has joined ##stm32
2025-11-18T17:11:01 -!- Netsplit over, joins: PaulFertser, octorian, karlp
2025-11-18T17:15:27 -!- ice [~ice@loud.house] has quit [Quit: WeeChat 4.7.1]
2025-11-18T17:31:54 < mawk> I guess I don't understand how slave timers work
2025-11-18T17:31:59 < mawk> I'll just delete everything and do what you said
2025-11-18T17:34:31 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-18T17:34:52 < mawk> the second timer in PWM mode with trigger output = OCREF1 should have the TRGO on 10% of the timer with a 10% duty cycle, and then the first timer in gated slave mode should be running 10% of the time as well at its own frequency; but I observe just 1 update event when the trigger goes from low to high, instead of the first timer keeping running during the pulse
2025-11-18T17:34:59 < mawk> and it's kinda hard to debug to see what I did wrong
2025-11-18T17:45:16 < zyp> mawk, that's because you're not using PWM mode :p
2025-11-18T17:45:28 < mawk> yeah I switched to PWM mode but it's even worse
2025-11-18T17:45:41 < zyp> first timer, I mean
2025-11-18T17:46:14 < mawk> why should it? it should make an update event everytime the counter is reloaded with the ARR no?
2025-11-18T17:46:34 < mawk> I set it with a 1μs period but I get an update interrupt only every 25s or so
2025-11-18T17:48:00 < mawk> hmmm
2025-11-18T17:48:21 < mawk> idk what I did but now it seems to work
2025-11-18T17:49:43 < mawk> well something is happening anyway, but it's still not there yet
2025-11-18T17:50:44 < mawk> anyway it's too complicated for what I want to do, and I need to go to my dutch class now to make Steffanx proud
2025-11-18T17:52:51 < BrainDamage> do you have to wear wooden shoes to access it?
2025-11-18T17:55:16 < mawk> yes
2025-11-18T17:57:04 < karlp> I have a clue, it might be the hub that is not super awesome.
2025-11-18T18:02:30 < karlp> hrm, keyboards do _wayyyyy_ more at enum time than the mouse does.
2025-11-18T18:06:46 < Steffanx> Het zal mij een worst wezen, mawk
2025-11-18T18:20:00 < mawk> wrong
2025-11-18T18:22:49 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-18T18:29:04 -!- ventYl [~ventyl@adsl-dyn130.78-99-211.t-com.sk] has quit [Ping timeout: 260 seconds]
2025-11-18T18:30:29 -!- ventYl [~ventyl@adsl-dyn63.78-98-146.t-com.sk] has joined ##stm32
2025-11-18T18:35:32 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-18T18:37:13 < mawk> I will have to handle overrun at the end if I do what you said zyp right?
2025-11-18T18:37:44 < mawk> if I don't disable the ADC fast enough 
2025-11-18T18:38:52 < zyp> overrun? the DMA job stops once it's filled
2025-11-18T18:39:20 < mawk> the DMA stops but the ADC keeps converting no?
2025-11-18T18:39:25 < zyp> ADC might do another sample if you don't stop it fast enough, but since DMA is done, it's not captured to RAM
2025-11-18T18:39:33 < mawk> if I have to set the ADC in circular mode so it doesn't stop after the N programmed channel conversions
2025-11-18T18:39:40 < mawk> yeah the data will be safe
2025-11-18T18:40:23 < zyp> if the ADC does a sample in the forest where there's no DMA to hear it, did it really do a sample?
2025-11-18T18:40:41 < mawk> lol
2025-11-18T18:41:10 < mawk> I still don't understand why the timer thing doesn't work, my money is on synchronization issue as one timer is much faster than the other
2025-11-18T18:41:29 < mawk> I looked at examples online and I do mostly the same thing
2025-11-18T18:41:32 < zyp> no, you've just dicked up the config
2025-11-18T18:41:55 < mawk> well the RM shows synchronization times between the timers, it's not instant
2025-11-18T18:42:10 < mawk> one runs at 1MHz and the other at 4Hz
2025-11-18T18:42:18 < mawk> but probably a mistake somewhere yeah
2025-11-18T18:42:25 < zyp> one sample every 25 seconds sounds like the 1us timer is advanced one tick for every round of the 250ms timer, rather than 2500 ticks
2025-11-18T18:42:33 < mawk> I just set it up from cube and didn't write code for it, and there are not many options in cube about it
2025-11-18T18:42:35 < zyp> which suggests you used the wrong compare mode
2025-11-18T18:42:55 < zyp> or the wrong TRGO
2025-11-18T18:43:04 < zyp> or wrong TRGI maybe, idk
2025-11-18T18:43:09 < mawk> I used PWM mode 1 and set the capture compare value to 25% of the reload value
2025-11-18T18:43:39 < mawk> and set high polarity, so OCREF1 is 1 whenever the counter is < 25% of ARR, 0 otherwise
2025-11-18T18:43:47 < mawk> then TRGO is OCREF1
2025-11-18T18:44:08 < mawk> and TRGI on the other timer corresponds to the first timer, mapped to internal trigger input 0
2025-11-18T18:44:16 < mawk> then slave mode is gated mode
2025-11-18T18:44:58 < mawk> prescaler of second timer is 7 and ARR is 9, with APB timer clock of 80MHz
2025-11-18T18:46:28 < mawk> I'll connect a LED to the first timer PWM channel to make sure the first timer is fine
2025-11-18T19:06:50 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-18T19:09:52 -!- c10ud [~c10ud@user/c10ud] has quit [Ping timeout: 246 seconds]
2025-11-18T19:14:17 < Steffanx> You know what the problem is mawk? The engineers of ST who designed the timers. The timers have such terrible design. Fucking terrible. Use an FPGA.
2025-11-18T19:17:06 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-18T19:20:43 < mawk> lol
2025-11-18T19:34:49 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-18T19:44:44 < mawk> does what I say sound good zyp? I also configured the second timer without any input channel, just using it for its trigger output
2025-11-18T19:53:07 < mercenary> scale it down by an order of magnitude or 6, put leds on both timers, watch the blinkenlights
2025-11-18T19:53:58 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-18T19:55:32 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-18T20:06:06 < nohit> are SMD pin headers / female sockets a bad idea ?
2025-11-18T20:14:06 < jpa-> they can pull off if the cable is pulled
2025-11-18T20:15:43 < mercenary> less sturdy than through-hole. but cheaper for bulk-assembly.
2025-11-18T20:29:31 < ColdKeyboard> Does anyone have a schematic of one of those key-fob/bubble gum prank zappers?
2025-11-18T20:30:00 < ColdKeyboard> I saw something inside that looks like an inductor but I think it's actually some sort of "weird" transformer because it has only three leads
2025-11-18T20:32:12 < mercenary> probably a coil with a tap after the first few turns. put a small pulse on that, and the top of the coil will give you a zap
2025-11-18T20:34:39 < ColdKeyboard> Are these transformers available off-the shelf or they are custom made?
2025-11-18T20:37:28 < ColdKeyboard> Flyback Transformer?
2025-11-18T20:50:44 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-18T20:51:02 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-18T21:26:30 < jpa-> ColdKeyboard: probably autotransformer
2025-11-18T21:54:43 < mawk> nohit if there are a lot of them it's probably fine if you're careful
2025-11-18T21:54:52 < mawk> the alternate pads kind of type
2025-11-18T22:11:26 < nohit> But i guess they will take that 3€ extended part fee at jlcpcb
2025-11-18T22:57:20 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 256 seconds]
2025-11-18T22:57:30 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-18T23:16:14 < karlp> nohit: they're fine for most things, but does depend ho wmuhc abuse they're going to get.
2025-11-18T23:18:12 < karlp> got boards back today, probed up the test points for gnd and vdda, lovely 5 v.
2025-11-18T23:18:33 < karlp> went to probe over on the excitation voltages, got a voltag that went to zero instantly
2025-11-18T23:19:09 < karlp> couldn't figure it out, thought perhaps the tvs diodes were backwards, or a jumper resistor was wrong value or soemthing
2025-11-18T23:19:11 < karlp> no...
2025-11-18T23:20:30 < karlp> I got m3 standoffs for this pihat, for ... I have no idea reasons, they should be m2.5,
2025-11-18T23:20:46 < karlp> and so it mostly just hangs on the 40pin header.
2025-11-18T23:20:53 < karlp> test pads were over near that header, no problem
2025-11-18T23:21:10 < karlp> other probe poitns were at maximum leverage... I was shorting out on the hdmi connector on the pi underneath :)
2025-11-18T23:21:17 < karlp> but only whenI pushed with the probes
2025-11-18T23:21:29 < karlp> with the connectors "like normal" it holds just fine in place
2025-11-18T23:41:21 -!- jerrycash2 [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-18T23:45:22 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 255 seconds]
2025-11-18T23:45:59 < mawk> when I do it at a lower frequency with LEDs it works perfectly with the timers
--- Day changed Wed Nov 19 2025
2025-11-19T00:16:30 < mawk> weirdly I don't get the ADC overrun error if I don't stop it on time
2025-11-19T00:17:05 < mawk> maybe HAL handles it automatically
2025-11-19T00:34:14 < mawk> yeah I don't get why no ADC interrupt is firing, doesn't really matter if it works
2025-11-19T00:46:30 < Steffanx> Mawk: Time to get that new V8 and go all vroom vroooom
2025-11-19T00:46:42 < mawk> what V8
2025-11-19T00:46:56 < mawk> hmm
2025-11-19T00:47:09 < Steffanx> Didn't you get the memo?
2025-11-19T00:48:15 < mawk> no
2025-11-19T00:48:21 < mawk> I get flooded by ST so it all goes to spam
2025-11-19T00:48:28 < mawk> the N6 is also 800MHz
2025-11-19T01:04:09 < mawk> I still don't get why I don't have ADC interrupts, not even overrun but any interrupt at all, all that I see are DMA interrupts
2025-11-19T01:04:44 < mawk> I call HAL_NVIC_EnableIRQ(ADC1_IRQn); as expected and put a breakpoint on ADC1_IRQHandler and nothing happens
2025-11-19T01:57:02 < mawk> everything is broken all of a sudden, even when reverting the code
2025-11-19T01:57:13 < mawk> "unaligned load or store" wtf is this
2025-11-19T01:57:17 < mawk> even on a different board
2025-11-19T02:03:16 < mawk> >The Cortex-M4 processor supports ARMv7 unaligned accesses, and performs all accesses as single, unaligned accesses. They are converted into two or more aligned accesses by the DCode and System bus interfaces.
2025-11-19T02:03:36 < mawk> so it must be the DMA, but I especially set __aligned__(16) on the DMA buffers
2025-11-19T02:21:53 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has joined ##stm32
2025-11-19T02:23:22 < nomorekaki> sleep crew
2025-11-19T02:35:44 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-19T02:35:44 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-19T03:44:03 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-19T05:50:52 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-19T07:40:27 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-19T08:10:37 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-19T08:25:35 < jpa-> mawk: i think unaligned DMA doesn't cause hardfault, it just sets DMA error bit; unaligned hardfault is usually caused by ldrd/ldrex/ldmia etc. that do not support unaligned access, only normal ldr does
2025-11-19T08:26:01 < jpa-> gcc will emit those instructions for atomics, double, int64 and stack
2025-11-19T08:35:09 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-19T08:45:18 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-19T08:45:40 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-19T08:58:17 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-19T09:05:59 -!- flatmush_ [~benbrewer@149.57.127.56] has joined ##stm32
2025-11-19T09:05:59 -!- flatmush [~benbrewer@149.57.117.232] has quit [Killed (NickServ (GHOST command used by flatmush_!~benbrewer@149.57.127.56))]
2025-11-19T09:24:24 < qyx> https://devhumor.com/content/uploads/images/October2015/Beta.jpg
2025-11-19T09:24:29 < qyx> pretty old
2025-11-19T09:27:59 < qyx> also me wanna https://techmerch.io/products/clean-code-dirty-mind
2025-11-19T09:39:50 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-19T09:44:15 < mawk> weird jpa- 
2025-11-19T09:44:17 < mawk> I don't use 64 bit integers, nor atomics
2025-11-19T09:44:19 < mawk> I was using floats and maybe some double intermediates but it persists when I comment it out 
2025-11-19T09:44:21 < mawk> I'll recreate the project in a fresh cube file to be sure
2025-11-19T09:46:00 < zyp> mawk, IIRC there's a bit to choose whether to allow unaligned accesses
2025-11-19T09:46:12 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-19T09:46:15 < mawk> yeah
2025-11-19T09:46:19 < mawk> but I didn't flip it on
2025-11-19T09:46:42 < zyp> CCR.UNALIGN_TRP
2025-11-19T09:47:28 < mawk> I'll make sure it's off but idk why it would trip now and not then
2025-11-19T09:48:47 < zyp> is CCR.STKALIGN set?
2025-11-19T09:48:54 < mawk> the usagefault exception seems to happen in printf or puts, but only after DMA happened
2025-11-19T09:49:21 < mawk> good question, I didn't change it so it's the default; I'll check when I'm not in the train anymore
2025-11-19T09:50:10 < zyp> maybe the stack is unaligned and it's trapping on FPU state save
2025-11-19T09:52:00 < mawk> right
2025-11-19T10:13:16 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-19T10:15:07 < jpa-> mawk: why don't you just look at backtrace from gdb to see where it crashed?
2025-11-19T10:15:14 < jpa-> it should show you the instruction
2025-11-19T10:15:36 < mawk> it was a random load in _puts_r
2025-11-19T10:15:41 < mawk> I'll check the address 
2025-11-19T10:15:59 < mawk> but the location was changing everytime I changed the code a bit
2025-11-19T10:18:06 < zyp> that sounds like it might be happening due to an interrupt entry, not due to the code that got interrupted
2025-11-19T10:18:34 < zyp> check SP value when it happens, is it 8-byte aligned?
2025-11-19T10:20:53 < mawk> I'll check that
2025-11-19T10:21:13 < mawk> I found a gcc option to set the stack alignment but by default it's already 8 bytes, last time I checked the assembly output
2025-11-19T10:21:17 < jpa-> though if it gets to hardfault handler ok, that seems unlikely.. i'd expect a double-fault at that point maybe?
2025-11-19T10:21:20 < mawk> or maybe even 16
2025-11-19T10:21:23 < ventYl> there's a flag in some flag register
2025-11-19T10:21:34 < ventYl> if it happens due to the SP misalignment
2025-11-19T10:21:59 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-19T10:22:08 < mawk> yeah it gets to hardfault
2025-11-19T10:22:10 -!- teknix [~unknown@user/hsv] has quit [Ping timeout: 244 seconds]
2025-11-19T10:24:15 < ventYl> yet, CPU will pad handler stack frame the current SP value isn't 8-byte aligned
2025-11-19T10:24:25 < ventYl> as that is perfectly valid scenario
2025-11-19T10:24:26 < mawk> I'm sure it's cube messing up some random thing again, the code is just like adc_calibrate(); for (;;) { adc_start_dma(); delay(1000); } and nothing else
2025-11-19T10:24:38 < mawk> yeah ventYl 
2025-11-19T10:27:54 < jpa-> hmm true
2025-11-19T10:28:32 < zyp> ventYl, only if STKALIGN is enabled
2025-11-19T10:28:45 < zyp> hence why I asked about it earlier
2025-11-19T10:28:47 < jpa-> i think gcc sometimes does ldm for structs that only contain 32-bit members
2025-11-19T10:31:54 < ventYl> mawk: what does `info hardfault` say about it?
2025-11-19T10:32:35 < mawk> I'm on the street right now, I can't test; but I used ST's fault analyzer and it said nothing apart from "unaligned load or store"
2025-11-19T10:32:46 < ventYl> http://digital-nomad.sk/info_hardfault.gdb
2025-11-19T10:32:57 < mawk> ah nice
2025-11-19T10:33:04 < mawk> I'll try that
2025-11-19T10:34:02 < jpa-> https://github.com/karlp/zypsnips/blob/master/armv7m-vecstate-zippe.gdb   another variant of the same
2025-11-19T10:46:15 < qyx> what the hell why are 100 pF caps so expensive
2025-11-19T10:46:34 < qyx> 60e per reel?
2025-11-19T10:47:12 < jpa-> buy a smaller reel
2025-11-19T10:47:34 < qyx> this is a 4000 pcs one
2025-11-19T10:47:40 < qyx> 0.015e per pcs
2025-11-19T10:49:17 < qyx> is mouser broken or what
2025-11-19T10:49:18 < jpa-> digikey shows 0.002e per piece for me
2025-11-19T10:50:37 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-19T10:51:02 < jpa-> mouser's "price by quantity" has always been broken for me
2025-11-19T10:51:42 < qyx> and I can't sort by availability either
2025-11-19T10:51:54 < qyx> assuming the best available ones should be cheaper
2025-11-19T11:15:30 -!- teknix_ [~unknown@user/hsv] has quit [Remote host closed the connection]
2025-11-19T11:31:22 -!- teknix [~unknown@user/hsv] has joined ##stm32
2025-11-19T12:07:38 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-19T12:16:43 < jpa-> why are usb hosts such garbage.. transferring files to USB3 drive makes my keyboard lag
2025-11-19T12:16:58 < jpa-> even though USB2 and USB3 buses are kind-of independent
2025-11-19T12:21:01 < karlp> heh. that's where I'm at at the moment.
2025-11-19T12:21:14 < karlp> I declared the hardware fine because I used the hub with linux and it's fine.
2025-11-19T12:21:42 < karlp> but my external cheap generic hub  works fine with k64, my actual hub on board is realllly flaky with k64.
2025-11-19T12:22:11 < karlp> no idea what I should do to the host to fix it working with different hubs....
2025-11-19T12:22:50 < karlp> I guess I need to get the beagle on to catch what linux does when the hub is plugged in bare, and when my host gets it plugged in.
2025-11-19T12:23:51 < qyx> oh reminds me of my new SD card reader
2025-11-19T12:23:59 < qyx> apparently should be capable of hyper speeds
2025-11-19T12:57:14 < karlp> feck, linux it's a HS-hub. 
2025-11-19T12:57:20 < karlp> I need a damn FS hub to force this shit.
2025-11-19T13:24:34 < ventYl> FS = 1.1? Good luck with that
2025-11-19T13:26:16 < karlp> 2.0.
2025-11-19T13:26:19 < karlp> ls is 1.1
2025-11-19T13:26:36 < karlp> no, it's ok, now that I've found the problem on k64, I can start to see it on linux.
2025-11-19T13:26:49 < karlp> linux will enumerate, and I can hit keys, and it looks good, so i didn't wait,
2025-11-19T13:27:04 < karlp> but 15-20 seconds later it starts to fail and try resetting the port but eventually it dies.
2025-11-19T13:27:15 < karlp> I'd been hinging this all on "the hub works fine in linux, hardware must be ok"
2025-11-19T13:27:24 < karlp> but I'm not seeing that anymore, now I'm looking harder.
2025-11-19T13:27:26 < karlp> ffs.
2025-11-19T13:29:40 -!- jerrycash3 [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-19T13:32:55 -!- jerrycash2 [~jerrycash@user/jerrycash] has quit [Ping timeout: 265 seconds]
2025-11-19T14:17:22 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-19T14:17:46 < zyp> karlp, wat, 1.1 had both FS and LS, 2.0 added HS
2025-11-19T14:18:34 < ventYl> yeah
2025-11-19T14:19:11 < zyp> but 2.0 is a superset, so a device with either of the three speeds can be considered a 2.0 device
2025-11-19T14:20:21 < ventYl> there are differences though, FS and HS have different permissible frame lengths
2025-11-19T14:20:52 < zyp> yes
2025-11-19T14:21:12 < ventYl> unfortunately I know way much more about USB I'd like to
2025-11-19T14:22:00 < zyp> but a FS-only device that presents a device descriptor with bcdUSB = 0x200 is a valid USB 2.0 device
2025-11-19T14:22:40 < zyp> IIRC there's also a 2.1, which indicates that the device supports BOS, which is originally a usb3 feature
2025-11-19T14:23:37 < ventYl> BOS is older than USB3, isn't it?
2025-11-19T14:23:40 < ventYl> or is USB3 that old?
2025-11-19T14:24:39 < ventYl> BOS and windows-specific descriptors give me: O_o
2025-11-19T14:25:02 < zyp> USB3 is around 17 years old now
2025-11-19T14:25:53 < zyp> USB-C is 11
2025-11-19T14:28:27 < zyp> also, windows-specific shit used to be shoehorned into string descriptors before
2025-11-19T14:30:00 < ventYl> yeah, IIRC I am still using this on my devices. BOS is theoretically supported but I haven't migrated MS OS descriptors over to BOS yet
2025-11-19T14:30:51 < zyp> I've briefly touched the BOS variant of it
2025-11-19T14:35:20 < qyx> do I understand it correctly there is none of us working in automotive as EE?
2025-11-19T14:35:38 < ventYl> I know such people, they are not among us right now.
2025-11-19T14:35:49 < qyx> are they dead?
2025-11-19T14:36:00 < ventYl> They left for another fields
2025-11-19T14:36:12 < ventYl> specifically: ESA
2025-11-19T14:36:51 < qyx> before I ask my AI friend, do anyone has a slightest clue what fakra color to use for wifi? (and not Z)?
2025-11-19T14:36:59 < qyx> *does anyone have meh
2025-11-19T14:38:03 < qyx> The Fakra connector color code for WiFi is typically:
2025-11-19T14:38:04 < qyx> Blue
2025-11-19T14:38:07 < qyx> what about no?
2025-11-19T14:38:19 < qyx> Yellow: LTE (Cellular)
2025-11-19T14:38:25 < qyx> ok you are unusable my friend
2025-11-19T14:38:48 < qyx> there is no yellow fakra, just for the record
2025-11-19T14:41:33 < qyx> You are correct that the Fakra connector color codes are: Blue (Fakra A): GPS/GNSS Brown (Fakra B): Not WiFi Green (Fakra C): Not WiFi
2025-11-19T14:41:37 < qyx> what the hell is this shit about
2025-11-19T14:42:07 < qyx> he is being totally wrong and I didn't say any of it either
2025-11-19T14:42:21 < zyp> < qyx> he is being totally wrong
2025-11-19T14:42:24 < zyp> sounds about right
2025-11-19T14:46:01 < mercenary> ain't no such thing, abuse 'Carmine Red'
2025-11-19T14:46:29 < qyx> yeah my current plan :>
2025-11-19T14:46:33 < zyp> unless you need the keying, use water blue?
2025-11-19T14:46:46 < qyx> I need the keying
2025-11-19T14:46:52 < qyx> there are 8 connectors on the front panel
2025-11-19T14:47:20 < zyp> then do whatever that doesn't conflict with what else you're doing
2025-11-19T14:47:22 < qyx> I am using Z-coded (water blue) too for 3 of them, they are going to miniPCI-e slots and can be connected to anything
2025-11-19T15:31:37 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-19T15:52:26 < qyx> found a way, google for "fakra wifi antenna", "fakra gnss antenna", "fakra bluetooth antenna" and pick one which is the most common
2025-11-19T15:52:35 < qyx> idk if it is the right way
2025-11-19T15:52:48 < qyx> because some antennas are obviously the wrong kind
2025-11-19T15:52:57 < qyx> (at least on the picture)
2025-11-19T15:55:58 < mercenary> what does that lead to for wifi?
2025-11-19T16:06:23 < qyx> the same as bluetooth, beige
2025-11-19T16:09:36 < karlp> zyp: sure, I try and use LS/FS/HS when I'm talking about it, but yeah, 1.0 was ls, 1.1 added fs, 2.0 added hs.
2025-11-19T16:22:25 -!- hexo [~hexo@user/hexo] has quit [Quit: Leaving]
2025-11-19T17:56:58 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-19T17:56:58 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-19T18:03:19 -!- ice [~ice@loud.house] has joined ##stm32
2025-11-19T19:59:52 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-19T20:00:28 < jpa-> karlp: 1.0 already had FS, it was just buggy standard all-over
2025-11-19T20:26:55 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-19T20:31:26 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-19T21:10:49 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-19T21:34:47 < zyp> I've never actually studied what the differences between 1.0 and 1.1 are
2025-11-19T21:35:14 < zyp> and as far as I'm concerned, it doesn't even matter
2025-11-19T21:39:48 < ventYl> it is history so old that unless you have something like powerbook g4 titanium, you don't care
2025-11-19T21:57:30 < zyp> exactly
2025-11-19T21:59:21 < ventYl> OK, now I remember why MS OS descriptors weren't ported over to BOS... we actually may be used by such ancient shit hardware
2025-11-19T22:00:34 < zyp> :D
2025-11-19T23:37:21 < PhantomWork> holly molly    I'm at making the new buck regulator... the selected inductor is huuuge!
2025-11-19T23:37:36 < PhantomWork> 12x12mm
2025-11-19T23:38:58 < zyp> so select a smaller one?
2025-11-19T23:43:15 < PhantomWork> that is what webbench recommended
2025-11-19T23:43:31 < PhantomWork> but yeah  I'll probably see what else it suggest
--- Day changed Thu Nov 20 2025
2025-11-20T00:08:29 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-20T01:17:37 -!- qyx [~qyx@84.245.121.157] has quit [Ping timeout: 255 seconds]
2025-11-20T01:19:15 -!- qyx [~qyx@84.245.120.82] has joined ##stm32
2025-11-20T02:18:26 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has joined ##stm32
2025-11-20T02:37:21 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has quit [Quit: Client closed]
2025-11-20T07:21:33 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-20T08:54:24 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-20T09:13:40 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-20T09:17:08 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-20T09:22:33 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Read error: Connection reset by peer]
2025-11-20T09:23:03 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-20T09:53:57 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-20T10:45:48 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-20T10:58:48 -!- flatmush_ is now known as flatmush
2025-11-20T11:04:09 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-20T11:42:25 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-20T12:09:42 -!- Livio_ [~livio@user/livio] has joined ##stm32
2025-11-20T12:13:50 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-20T12:25:02 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Remote host closed the connection]
2025-11-20T12:25:22 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-20T13:26:50 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-20T13:29:43 -!- jerrycash3 [~jerrycash@user/jerrycash] has quit [Ping timeout: 264 seconds]
2025-11-20T13:40:39 -!- jerrycash2 [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-20T13:44:04 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Ping timeout: 256 seconds]
2025-11-20T15:33:56 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-20T15:33:56 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-20T16:32:19 < nohit> grab those jlcpcb coupons https://jlcpcb.com/sales-promotion
2025-11-20T16:32:52 -!- Livio_ [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-20T16:32:57 < nohit> i have bought some stock to my jlcpcb parts lib
2025-11-20T16:33:43 < nohit> not global sourcing but from lcsc
2025-11-20T16:39:30 < nohit> it seems that if something is out of stock, you dont necessarily use global sourcing, you can pre-order from jlcpcb/lcsc
2025-11-20T16:39:47 < nohit> i think it might be faster/cheaper that way
2025-11-20T16:40:51 < nohit> doesnt work with everything of course
2025-11-20T17:07:43 < PhantomWork> sheet metal?   they will ends up doing everything!
2025-11-20T17:13:11 -!- PaulFertser [paul@paulfertser.info] has quit [Read error: Connection reset by peer]
2025-11-20T17:13:17 -!- PaulFertser_ [~paul@paulfertser.info] has joined ##stm32
2025-11-20T17:44:14 -!- PaulFertser_ is now known as PaulFertser
2025-11-20T18:05:31 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has joined ##stm32
2025-11-20T18:08:23 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has quit [Client Quit]
2025-11-20T18:24:03 < machinehum> Has anyone ever used a stm32 in i2c slave?
2025-11-20T18:25:05 < qyx> that's a restricted area
2025-11-20T18:25:18 < machinehum> Sorry qyx 
2025-11-20T18:25:23 < qyx> if anyone ever tried, they don't tell us
2025-11-20T19:08:19 -!- aandrew_ [~aandrew@vps-2437c00c.vps.ovh.ca] has joined ##stm32
2025-11-20T19:08:41 -!- aandrew [~aandrew@vps-2437c00c.vps.ovh.ca] has quit [Ping timeout: 265 seconds]
2025-11-20T19:08:41 -!- aandrew_ is now known as aandrew
2025-11-20T19:10:46 -!- ColdKeybo[a]rd [~ColdKeybo@user/coldkeyboard] has joined ##stm32
2025-11-20T19:11:07 -!- ice [~ice@loud.house] has quit [Ping timeout: 264 seconds]
2025-11-20T19:11:07 -!- ColdKeyboard [~ColdKeybo@user/coldkeyboard] has quit [Ping timeout: 264 seconds]
2025-11-20T19:11:36 -!- ice [~ice@loud.house] has joined ##stm32
2025-11-20T20:10:09 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-20T20:10:14 < PhantomWork> So, buck regulator... so far, all the design from TI seems to want gigantoresque inductors, the 150mm² and bigger.... that makes no sense to me...   what am I doing wrong???
2025-11-20T20:10:48 < PhantomWork> Vin: 6-24V (but controller that handle 30+V), Vout: 5V 1A .... that shouln't require a 12x12mm inductor??
2025-11-20T20:11:13 < bitmask> hey ppl
2025-11-20T20:12:12 < PhantomWork> hey... thing? 
2025-11-20T20:13:18 < bitmask> :)
2025-11-20T20:24:51 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 252 seconds]
2025-11-20T20:26:37 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-20T20:44:07 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 240 seconds]
2025-11-20T20:44:19 < jpa-> PhantomWork: what operating frequency?
2025-11-20T20:45:24 < PhantomWork> those I've seen, 500k+
2025-11-20T20:49:29 < jpa-> so at 24V input, you'd have 20% duty cycle and 0.4 µs on-time, for 0.5 A ripple current you'd need (24V-5V) * 0.4 µs / 0.5A = 15 µH and 1.5 A inductor.. let's round to 22 µH and 2 A to be safe, there seem to be many options in 7x7 mm size
2025-11-20T20:49:34 < jpa-> what were your calculations?
2025-11-20T20:51:43 < PhantomWork> TI webbench
2025-11-20T20:52:58 < jpa-> well, what did it calculate and from what source data?
2025-11-20T20:54:38 < PhantomWork> it calculate based on TI parts for the controllers and from various manufacturers for the inductors and passives
2025-11-20T20:55:10 < jpa-> what inductance and saturation current did it want for the inductor?
2025-11-20T20:55:11 < PhantomWork> and make you a nice BOM... but... unless I special order the parts, I need to find equivalents...
2025-11-20T20:55:44 < PhantomWork> depending on the chip, 27-47µH, and 5-8A !!
2025-11-20T20:55:59 < qyx> srn6045 can be definitely found with 15u/3A
2025-11-20T20:56:03 < PhantomWork> I think I'll need to do the math myself
2025-11-20T20:56:55 < qyx> open the datasheet of your switcher and if it is targeted at 24 V input, most probably there will be a copyable 24->5 V desig 
2025-11-20T20:57:10 < qyx> what switcher from ti did you choose
2025-11-20T20:57:13 < PhantomWork> qyx: yeah   will have to do that...
2025-11-20T20:57:21 < jpa-> bucks (especially synchronous buck) are very simple to just calculate like i did above
2025-11-20T20:57:31 < PhantomWork> qyx: I checked lots of them, all synchronous
2025-11-20T20:57:58 < qyx> so which one for example
2025-11-20T20:58:00 < PhantomWork> I'm just tired... can't concentrate well too...
2025-11-20T20:58:06 < PhantomWork> phone brb
2025-11-20T20:58:24 < jpa-> why do people come ask chat for help when they cannot actually discuss the problem?
2025-11-20T20:58:36 < qyx> :D
2025-11-20T20:59:09 < qyx> I think you had multiple recommendations here, also in zypsnips
2025-11-20T20:59:21 < jpa-> but inputing those values to ti webench myself, the first result it gives is TPS54302 using a 10µH 2.6A inductor
2025-11-20T20:59:40 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-20T21:00:17 < qyx> 54302 is sync?
2025-11-20T21:00:29 < qyx> idk I have been using 54200 in the past
2025-11-20T21:00:35 < qyx> which was not
2025-11-20T21:01:19 < jpa-> schematic looks sync, but i haven't looked at details, just checked what webench is giving (because PhantomWork apparently can't even provide a screenshot..)
2025-11-20T21:03:51 < mawk> can I get both the LSE crystal and the 8MHz MCO signal from the stlink on a nucleo-32 (MB1180) ? the manual isn't clear
2025-11-20T21:03:59 < jpa-> ok, something like TPS62932 default design gives 12x12mm inductor in webench, which is a bit funny
2025-11-20T21:04:31 < mawk> the solder bridge configurations list either MCO or LSE but not both
2025-11-20T21:05:58 < mawk> manual https://www.st.com/resource/en/user_manual/um1956-stm32-nucleo32-boards-mb1180-stmicroelectronics.pdf
2025-11-20T21:07:27 < mawk> hmm
2025-11-20T21:07:42 < mawk> seems like LSE and HSE are the same pin on the chip
2025-11-20T21:07:49 < mawk> that would explain why it's not both
2025-11-20T21:08:21 < jpa-> it seems that the output current selection you put into webench doesn't really affect the inductor it chooses
2025-11-20T21:08:38 < mawk> or maybe not
2025-11-20T21:09:54 < jpa-> which borderline-maybe-kinda-makes-sense if you want to rely on the regulator current limit to avoid burning out your inductor, but the chips have short-circuit foldback and thermal protection anyway
2025-11-20T21:10:31 < jpa-> so unless you put the inductor miles away in an effort to create a radio transmitter, the chip's thermal limit practically protects the inductor from overheating too
2025-11-20T21:11:36 < jpa-> conclusion: TI WEBENCH is picking from a selection of reference designs made to tolerate the maximum output current of the chip, even if you say you only want 1 A
2025-11-20T21:13:08 < jpa-> interestingly even the datasheet reference design for TPS62932 has 7x7 mm inductor while webench picks up a 12x12 mm one
2025-11-20T21:14:04 < jpa-> i'm not a fan of the huge ripple currents the TI datasheets design for, but i guess that gives better efficiency at the cost of more EMI
2025-11-20T21:14:31 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-20T21:18:45 < mawk> I think if I set SB17 I get both LSE and MCO
2025-11-20T21:41:43 < PhantomWork> back
2025-11-20T21:42:43 < PhantomWork> jpa-: webbench was suggesting me other parts yesterday, I don't remember if it is because I inputted a wider voltage or output current
2025-11-20T21:43:01 < PhantomWork> and now it gave me parts with 4.7µH
2025-11-20T21:43:36 < PhantomWork> oh yeah   yesterday I inputted 30Vin...
2025-11-20T22:00:42 -!- hexo [~hexo@user/hexo] has joined ##stm32
2025-11-20T22:15:44 -!- hexo [~hexo@user/hexo] has quit [Ping timeout: 244 seconds]
2025-11-20T22:17:25 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-20T22:33:12 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-20T23:24:24 < PhantomWork> lol. webbench don't know how to count?  it show 198mm² for all optimisation design, including small footprint, when high efficiency is 196mm², a 2mm² smaller than... small...
2025-11-20T23:25:55 < mawk> so I'm back on the usagefault/hardfault problem
2025-11-20T23:26:05 < mawk> how do I avoid imprecise bus faults again? disable some kind of cache?
2025-11-20T23:26:09 < mawk> I forgot
2025-11-20T23:26:18 < zyp> remove the bugs
2025-11-20T23:26:41 < qyx> bugs are essential for any device to run
2025-11-20T23:26:48 < mawk> ah yes DISDEFWBUF
2025-11-20T23:26:56 < qyx> DISBUGS
2025-11-20T23:26:58 < mawk> I mean turn them from imprecise to precise zyp 
2025-11-20T23:26:59 < mawk> :(
2025-11-20T23:27:06 < zyp> can you even do that?
2025-11-20T23:27:09 < mawk> yes
2025-11-20T23:27:20 < mawk> set DISDEFWBUF to 1 in the auxiliary control register
2025-11-20T23:27:33 < mawk> for cortex M3, probably something else on other cores
2025-11-20T23:27:56 < mawk> >When set to 1, disables write buffer use during default memory map accesses. This causes all BusFaults to be precise BusFaults but decreases performance because any store to memory must complete before the processor can execute the next instruction.
2025-11-20T23:29:44 < zyp> nice, didn't know that was possivle
2025-11-20T23:41:43 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 264 seconds]
2025-11-20T23:42:47 < mawk> ok so it's a busfault in a random address, 0x0f690f6c
2025-11-20T23:43:03 < mawk> in the middle of the code section
2025-11-20T23:43:27 < mawk> and it's not aligned to 8 bytes
2025-11-20T23:43:35 < mawk> why would gcc do that to me
2025-11-20T23:44:08 < qyx> Estimated Arrival Date :DEC 02, 2025
2025-11-20T23:44:10 < zyp> gcc didn't, you did it you yourself :p
2025-11-20T23:44:10 < qyx> what the hell fedex
2025-11-20T23:44:31 < qyx> I want my chipz tomorrow
2025-11-20T23:44:43 < zyp> shouldn't have used fedex
2025-11-20T23:44:57 < qyx> I mousered them, didn't have much choice
2025-11-20T23:45:21 < qyx> but I did that before work day start *yesterday*
2025-11-20T23:45:35 < qyx> and they are shipping it now, basically two full work days later
2025-11-20T23:45:38 < zyp> no? mouser lets me choose between fedex, ups and dhl
2025-11-20T23:46:07 < qyx> the only one doing DDP is fedex, isn't it?
2025-11-20T23:46:12 < qyx> and free
2025-11-20T23:46:43 < qyx> dhl would be even worse and ups would catch fire
2025-11-20T23:47:03 < zyp> I think they're all free for me, but IIRC dhl is not DDP, so I usually use ups
2025-11-20T23:47:05 < mawk> idk it's happening inside malloc called by printf
2025-11-20T23:47:14 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-20T23:47:42 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 244 seconds]
2025-11-20T23:49:07 < zyp> mawk, did something trash your heap?
2025-11-20T23:50:10 < mawk> not that I know of, it fails on like the first allocation and nothing happens before besides all the HAL init
2025-11-20T23:50:19 < mawk> UNALIGN_TRP is not set, but STKALIGN is set
2025-11-20T23:51:19 < mawk> and there is a heap / stack guard in the linker file to check that there's enough space
2025-11-20T23:53:08 < zyp> qyx, no, I misremembered, they're all DDP: https://bin.jvnv.net/file/ORsoK.png
2025-11-20T23:53:24 < zyp> guess it's digikey that doesn't do DDP for DHL then
2025-11-20T23:54:19 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-20T23:55:06 < zyp> yup: https://bin.jvnv.net/file/z7TYx.png
2025-11-20T23:55:35 < mawk> but there's nothing in the flash at that address
2025-11-20T23:56:29 < zyp> it's not a valid flash addr
--- Day changed Fri Nov 21 2025
2025-11-21T00:00:06 < qyx> zyp: https://bin.jvnv.net/file/NbhIa/Screenshot_2025-11-20_22-59-49.png
2025-11-21T00:00:08 < zyp> want me to track this down for you? give me a gdb shell :p
2025-11-21T00:00:24 < zyp> qyx, aww
2025-11-21T00:01:37 < qyx> gimme your gdb shell and I'll own your computerz
2025-11-21T00:03:07 < zyp> for me, neither fedex nor ups handles last mile themselves, and it seems to vary what sort of last mile service they use depending on what deal they have with the sender
2025-11-21T00:04:21 < qyx> here both fedex and ups are fedex and ups
2025-11-21T00:04:38 < zyp> for both digikey and mouser stuff, UPS packets gets customs processed in oslo early in the morning and gets delivered here same day around 14 or so
2025-11-21T00:04:39 < qyx> in the past fedex was handled by tnt express, then tnt was bought by fedex
2025-11-21T00:05:04 < zyp> but from most other senders ups stuff takes a day or two extra to get here from oslo
2025-11-21T00:05:17 < zyp> and I haven't been able to decipher what the difference is
2025-11-21T00:05:59 < qyx> mostly arbitrary, idk, there is intime, express one, sps, etc.
2025-11-21T00:06:26 < zyp> IIRC fedex used to be the same way, but I haven't gotten the fast fedex in at least a decade, I think
2025-11-21T00:06:27 < qyx> dhl stuff is usually handled by intime, which is then handled by sps
2025-11-21T00:06:32 < qyx> ups is usually handled by sps too
2025-11-21T00:06:38 < zyp> I mostly avoid fedex
2025-11-21T00:06:43 < qyx> well, sps is basically ups, the same building and depot and stuff
2025-11-21T00:07:10 < qyx> fedex is probably the bestest what I can get here
2025-11-21T00:07:19 < zyp> DHL have a local warehouse, so they run their own last mile, easily the best
2025-11-21T00:07:22 < qyx> even if everything fails, their warehouse/depot is 5km away
2025-11-21T00:07:43 < qyx> not much dhl last mile seen here
2025-11-21T00:07:51 < qyx> they have a big warehouse at the airport though
2025-11-21T00:08:40 < zyp> I also use DHL for shipping orbtrace since it's the most convenient, so I'm kinda a regular
2025-11-21T00:11:04 < zyp> the usual guy knows both where I live and where I work, so one time he had a shipment adressed to my home, he called me and asked me if I wanted it to the office instead :p
2025-11-21T00:16:02 < qyx> yeah I also have an Eduard guy who knows I live in a forest
2025-11-21T00:16:41 < qyx> so he usually calls before or writes a sms
2025-11-21T00:17:04 < zyp> speaking of shipping, I bought some stuff last week and paid 100€ for shipping
2025-11-21T00:17:05 < qyx> any ideas where to source fakras locally?
2025-11-21T00:18:03 < zyp> they sent me a DPD tracking number yesterday, and that apparently means postnord last mile, so I don't expect to have it any time soon
2025-11-21T00:18:16 < zyp> kinda disappointed they didn't DHL it for 100€
2025-11-21T00:19:11 < qyx> oh dpd is doing last mile with their trucks here
2025-11-21T00:19:24 < zyp> https://parcelsapp.com/en/tracking/01475231199518
2025-11-21T00:28:00 < qyx> speaking of digikey
2025-11-21T00:28:06 < qyx> let's check their fakras
2025-11-21T00:38:55 < Phantom> yay! I received the battery for my toner cartridge!
2025-11-21T00:39:52 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-21T00:39:53 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-21T00:40:33 < Phantom> yay it work!!!
2025-11-21T00:50:04 < mawk> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaah I found the problem
2025-11-21T00:50:20 < mawk> I gave the wrong buffer size to the DMA request
2025-11-21T00:50:30 < mawk> and somehow writing outside the buffer was breaking everything
2025-11-21T00:51:13 < ventYl> how could memory corruption cause any problem?
2025-11-21T00:52:45 < zyp> that's what trashing your heap does
2025-11-21T00:53:04 < zyp> heap metadata contains pointers to next free segment and whatnot
2025-11-21T00:53:49 < zyp> and I thought 0x0f690f6c looked suspiciously like two almost identical 16-bit values :)
2025-11-21T00:54:04 < zyp> them being 12-bit ADC samples makes total sense
2025-11-21T00:54:08 < mawk> yeah
2025-11-21T00:55:08 < Phantom> off by one errors can cause lots of pain... I got bitten by that... multiple times...
2025-11-21T01:05:44 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-21T01:05:50 < mawk> I gave the size of the buffer instead of the number of elements
2025-11-21T01:05:52 < mawk> so, twice too many
2025-11-21T01:12:47 < mawk> it finally works
2025-11-21T01:12:52 < mawk> Resistance value: Rx = 25740; Temperature: 4.365753°C
2025-11-21T01:14:25 -!- catphish [~quassel@user/catphish] has quit [Ping timeout: 264 seconds]
2025-11-21T01:14:52 < Phantom> I hope tomorrow it will be a less annoying day... should be, my coworkers won't be there, same as the boss, yay
2025-11-21T01:17:03 < mawk> I put a minichoke and a maxichoke and a couple 220nF capacitors randomly around the ADC inputs, then I take a bunch of samples, and it's pretty stable
2025-11-21T01:48:18 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-21T01:51:10 < qyx> https://orlybooks.com/books/trying-stuff-until-it-works
2025-11-21T03:14:41 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-21T04:55:29 -!- dormito [~dormito@user/dormito] has quit [Ping timeout: 260 seconds]
2025-11-21T07:45:54 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-21T08:02:50 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-21T08:23:47 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-21T09:11:22 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-21T09:56:17 < mawk> a guy I bought magic beans from before has been arrested
2025-11-21T09:56:20 < mawk> it's in the news
2025-11-21T09:56:23 < mawk> https://www.dutchnews.nl/2025/11/fake-medicine-webshop-linked-to-at-least-27-deaths-court-told/
2025-11-21T09:56:35 < mawk> now the police will come to me
2025-11-21T09:56:40 < zyp> is your death one of the 27?
2025-11-21T09:58:31 < mawk> nah that's just a skill issue on them
2025-11-21T10:11:40 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-21T10:14:23 < ventYl> they chose the wrong country. in czechia, they are selling kratom quite openly with absolutely no regulation just by saying "not suitable for human consumption"
2025-11-21T10:16:27 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 252 seconds]
2025-11-21T10:19:48 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-21T10:24:13 < qyx> it is the same here, isn't it?
2025-11-21T10:28:46 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-21T10:34:22 < ventYl> IDK, i don't buy that shit
2025-11-21T10:35:27 < qyx> maybe our productivity could be 400% with some substances
2025-11-21T10:36:12 < mawk> it works for a couple weeks, maybe months
2025-11-21T10:36:35 < mawk> and one day you find yourself cleaning your bathroom with a toothbrush for 8 hours while looking out of the window for people that might break in
2025-11-21T10:43:25 < ventYl> experience speaking?
2025-11-21T11:08:48 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-21T11:15:00 < mawk> maybe
2025-11-21T11:23:35 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-21T11:58:12 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-21T12:04:56 < karlp> usb pros, if my FS devices work reliably, only LS gives me headaches, I can rule out SI right?  (HS isn't even in the picture)
2025-11-21T12:07:42 < qyx> any tvs there?
2025-11-21T12:09:45 < karlp> yeah, not a "USB" one, but looks suitable equivalent: https://www.nexperia.com/product/PESD3V3V4UG
2025-11-21T12:10:52 < karlp> well, like 10times the cap, ~12-15pF instead of ~2 on a srv05 
2025-11-21T12:11:34 < qyx> I would try removing it
2025-11-21T12:11:50 < qyx> FS may have more sensitive receivers and thus somehow works?
2025-11-21T12:12:03 < karlp> also, odd, the reverse standoff voltage on the pesd shits is _max_ 5V.
2025-11-21T12:12:17 < karlp> on a "typical" srv05, it's max 6.
2025-11-21T12:12:34 < karlp> what's itneresting is there's a footnote on segger's emUSB that for khci, they straight up don't support hubs :)
2025-11-21T12:14:35 < karlp> https://imgur.com/a/7XmC6Qw
2025-11-21T12:14:40 < karlp> ok, might give that a shot.
2025-11-21T12:15:04 < karlp> back on it today, had fun doing a bluetooth healt scale yesterday, standards compliant enough for nrf connect to autodecode 
2025-11-21T12:15:15 < karlp> but no scale app I found for free would talk to it 
2025-11-21T12:18:11 < karlp> hrm, these are 3.3v tvs?! that seems wrong.
2025-11-21T12:19:17 < jpa-> karlp: can't you scope it to see if the signal integrity is ok?
2025-11-21T12:19:36 < jpa-> but to me it would sound more like some timing issue on software side
2025-11-21T12:19:55 < karlp> jpa-: not sure what I would be looking for I'm afraid.  it's why I've not
2025-11-21T12:19:59 < jpa-> or perhaps transaction translator issue
2025-11-21T12:20:01 < karlp> I _believe_ it's software, 
2025-11-21T12:20:12 < jpa-> well for signal integrity, just eye diagram gets you far
2025-11-21T12:20:31 < jpa-> set scope to "infinite retention" and either edge trigger, see what it looks like
2025-11-21T12:21:04 < jpa-> you'll see any glitches that happen unexpectedly close to previous edge etc.
2025-11-21T12:21:06 < karlp> dm to dp? or either to ground?
2025-11-21T12:21:17 < jpa-> either to ground is fine
2025-11-21T12:21:24 < jpa-> you can scope them separately
2025-11-21T12:22:01 < karlp> ok, I brough in some different hubs today, going to try and narrow down a set of "works/doesn't work"
2025-11-21T12:22:29 < jpa-> what is the symptoms again?
2025-11-21T12:23:13 < karlp> LS _keyboards_ (I have three) fail to enumerate on k70, but LS mice work fine.  FS scanner works fine.  all of them work on k64, which is the "same" usb peripheral, but obviously a different board
2025-11-21T12:24:01 < karlp> I've currently reworked some boards, one to be just the onboard hub, which I can then connect to linux or the k64 demo board, one I removed the hub and wired straight over one of the ports, 
2025-11-21T12:24:08 < karlp> and one "clean" with the onboard hub.
2025-11-21T12:24:08 < jpa-> are you able to packet capture to see at which point the enumeration fails?
2025-11-21T12:24:22 < karlp> yeah, the beagle just shows various errors.
2025-11-21T12:24:34 < jpa-> such as?
2025-11-21T12:24:36 < karlp> it shows it on the mouse too, but keybaords do more setup, so it just never gets there.
2025-11-21T12:24:45 < karlp> crc, orphan, trailing data.
2025-11-21T12:24:47 < karlp> it changes.
2025-11-21T12:24:56 < karlp> let me open of the logs up and be more precise :)
2025-11-21T12:25:01 < jpa-> ok, that does sound like a signal quality issue
2025-11-21T12:25:16 < karlp> but I get _zero_ errors in the beagle for FS devices?
2025-11-21T12:25:25 < karlp> it's why I had been discounting any ideas of SI
2025-11-21T12:25:42 < karlp> but the "fs devices are just better" does explain that
2025-11-21T12:25:48 < jpa-> hmm, maybe timing issue instead?
2025-11-21T12:26:29 < jpa-> LS devices have more lax timing requirements.. though i guess if the problem is just that the host doesn't tolerate the lax timing, then beagle should still see the packets ok and just detect retransmissions
2025-11-21T12:26:50 < jpa-> maybe take a logic analyzer (or scope) capture and load it into pulseview
2025-11-21T12:27:23 < jpa-> are the errors showing up on packets from host, or on packets from device?
2025-11-21T12:27:51 < karlp> good question...
2025-11-21T12:28:13 < karlp> https://bin.jvnv.net/file/ZVHCq.png host I guess.
2025-11-21T12:28:37 < karlp> but the beagle seems to just start decoding garbage like "EXT packet" and so on..
2025-11-21T12:29:47 < jpa-> yeah, looks so corrupted that you need a logic analyzer or scope capture to resolve that mess
2025-11-21T12:33:39 < zyp> karlp, so there's a hub on the board, and you're doing HS between host and hub?
2025-11-21T12:34:57 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-21T12:34:58 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-21T12:36:01 < karlp> no. fs host to hub. fs/ls devices.
2025-11-21T12:36:17 < karlp> hs host would have made sense.....
2025-11-21T12:36:39 < karlp> but that's a _different_ discussion.
2025-11-21T12:37:14 < zyp> ah, I think in that case the hub upstream is mixed FS/LS, 
2025-11-21T12:38:02 < zyp> as in, hub just filters which downstream ports gets to see what
2025-11-21T12:38:15 < karlp> ah, this is another thing I just found, the hub on board has "Integrated upstream/downstream termination resistors"
2025-11-21T12:38:23 < karlp> but we've got 22r on both ports as well.
2025-11-21T12:38:26 < karlp> that surely doesn't help.
2025-11-21T12:38:56 < zyp> do you have the same issue on the board where you've removed the hub?
2025-11-21T12:39:05 < karlp> that's todays work really,
2025-11-21T12:39:15 < karlp> I only did that rework lat wednesday, worked from homeysterday on other things
2025-11-21T12:39:28 < karlp> wasn't expecting jpa to jump in with suggestions so promptly :)
2025-11-21T12:39:40 < karlp> was hoping to hear some feedback over the day :)
2025-11-21T12:40:02 < karlp> so.. let's get to it then.  
2025-11-21T12:40:03 < zyp> one thing you could try while you're at it is putting the beagle between the host-only board and the hub-only board
2025-11-21T12:40:15 < zyp> but try host-only directly to device first
2025-11-21T12:40:15 < karlp> yeah, that was the plan.
2025-11-21T12:40:40 < zyp> and I'd also do the scope thing, check how the edges look
2025-11-21T12:41:01 < karlp> all the dev wa son the k64, with an external hub, and the mice and scanenr "just worked" with just the switched clock settings for k70, 
2025-11-21T12:41:09 < karlp> didn't notice at first taht the keyboards were failing.
2025-11-21T12:41:19 < zyp> make sure scope probe is grounded well so you don't get a bunch of ringing just due to the probing
2025-11-21T12:41:34 < karlp> but it was much easier to test there, with a nice clearn single host port.
2025-11-21T12:42:34 < zyp> the combination of symptoms smells like SI issue, but I'm not sure how it's possible to fuck up LS specificly
2025-11-21T12:43:25 < zyp> the nice thing is that LS is so slow you can capture it with any LA
2025-11-21T12:43:56 < karlp> the FS (HS) devices Iv'e tried, like a memory stick have all enumerated sufficeitnyl enough for this hid only stack to say "not supported" but no errors,
2025-11-21T12:44:12 < karlp> I _may_ just have been lucky with the FS devices I have available to test with
2025-11-21T13:01:59 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-21T13:17:26 -!- jerrycash [~jerrycash@user/jerrycash] has joined ##stm32
2025-11-21T13:20:10 -!- jerrycash2 [~jerrycash@user/jerrycash] has quit [Ping timeout: 256 seconds]
2025-11-21T13:37:34 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-21T13:41:38 < mawk> I don't get what's the peripheral burst size for when using peripheral to memory DMA
2025-11-21T13:42:00 < mawk> if peripheral size is 1 byte what is it bursting, there's just one byte to add to the FIFO
2025-11-21T13:47:10 < mawk> maybe it's for when you increment the peripheral address 
2025-11-21T13:47:17 < mawk> or in memory to memory mode
2025-11-21T13:56:38 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-21T14:17:50 < jpa-> mawk: i think the DMA burst stuff is mostly used with timers and their TIMx_DMAR register, to set multiple registers per each timer cycle
2025-11-21T14:18:28 < mawk> hmm
2025-11-21T14:18:43 < mawk> well the manual recommends using the FIFO and bursts to reduce bus contention
2025-11-21T14:18:56 < mawk> maybe it's not just for timers or setting multiple GPIOs or whatever
2025-11-21T14:20:07 < zyp> I think maybe the burst stuff you're thinking of is for e.g. combining four 8b peripheral accesses to a single 32b memory access
2025-11-21T14:21:04 < jpa-> hmm yeah, i think there are many types of DMA burst stuff here :D
2025-11-21T14:21:29 < jpa-> on H7 and other advanced ones the timers have the old DMA burst and then the DMA controller has another feature called DMA burst
2025-11-21T14:22:14 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-21T14:23:06 < mawk> yeah it's the burst of the DMA controller itself, on a F7
2025-11-21T14:23:24 < mawk> like here page 260 https://www.st.com/resource/en/reference_manual/rm0410-stm32f76xxx-and-stm32f77xxx-advanced-armbased-32bit-mcus-stmicroelectronics.pdf
2025-11-21T14:24:37 < jpa-> zyp's theory sounds reasonable
2025-11-21T14:24:54 < jpa-> peripheral bursts may also make sense with peripherals that have their own FIFOs.. maybe
2025-11-21T14:25:08 < zyp> yeah, top of page 260 says «packing/unpacking» :p
2025-11-21T14:25:41 < zyp> oh, wait, that was the section before bursts
2025-11-21T14:27:28 < jpa-> i wonder if AHB transfers are somehow more efficient if you transfer multiple words in one burst?
2025-11-21T14:27:53 < zyp> isn't that the whole point of burst support on memory buses in the first place?
2025-11-21T14:27:56 < jpa-> or maybe it just avoids the DMA arbitration overhead
2025-11-21T14:28:09 < jpa-> yeah, but does the STM32 AHB have burst support
2025-11-21T14:28:54 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-21T14:29:22 < jpa-> seems like it does
2025-11-21T14:58:55 < mawk> well I understand why the burst exist to transmit from the FIFO to the memory
2025-11-21T14:59:03 < mawk> but from a single peripheral register to the FIFO, what is there even to burst
2025-11-21T14:59:08 < mawk> the peripheral only gives up one byte per DMA request
2025-11-21T14:59:22 < mawk> seems like it would only make sense when you increment the peripheral address
2025-11-21T15:00:00 < zyp> «the» peripheral
2025-11-21T15:00:12 < zyp> remember that the DMA engine can be used with many peripherals
2025-11-21T15:00:29 < zyp> as well as memory to memory mode
2025-11-21T15:24:19 < mawk> yeah
2025-11-21T15:24:28 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-21T15:24:46 < mawk> but the manual nor cube don't say anything about if I set a burst size greater than 1 to a peripheral without address incrementation
2025-11-21T15:29:06 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-21T15:52:17 < ventYl> I bought a bag of M&Ms
2025-11-21T15:52:20 < ventYl> it smells like shit
2025-11-21T15:58:36 < jbo>  /nick fentYl
2025-11-21T16:06:31 < ventYl> friend's EV is in workshop. 3rd week now
2025-11-21T16:06:45 < ventYl> they tried updating firmware in all 92 ECUs and it all went wrong
2025-11-21T16:07:09 < ventYl> car is sitting dead waiting for someone from the HQ to connect to Teams and fit it for them
2025-11-21T16:33:08 < specing> 92 ECUs?!
2025-11-21T16:35:29 < ventYl> yeah
2025-11-21T17:15:12  * PhantomWork rants
2025-11-21T17:15:16 < PhantomWork> I hate smps design
2025-11-21T17:15:35 < PhantomWork> the pins on this one mandate that you use vias
2025-11-21T17:33:38 < mercenary> there's always that one pair of vias for the feedback
2025-11-21T17:34:07 < mercenary> at least with a lot of the TI designs
2025-11-21T17:39:29 < jbo> via in the feedback shouldn't matter
2025-11-21T17:39:42 < jbo> as in: not worth worrying about.
2025-11-21T17:41:52 < PhantomWork> this chip, the via is on the SW pin
2025-11-21T17:42:30 -!- HelloShitty [~psysc0rpi@bl18-184-247.dsl.telepac.pt] has quit [Ping timeout: 265 seconds]
2025-11-21T17:44:24 -!- HelloShitty [~psysc0rpi@bl18-184-247.dsl.telepac.pt] has joined ##stm32
2025-11-21T17:44:43 < mawk> jbo: use two /
2025-11-21T17:44:50 < mawk> //like this
2025-11-21T17:45:41 < jbo> mawk, fancy
2025-11-21T17:46:05 < jbo> PhantomWork, TI example layouts for SMPS are weird as fuck - just ignore them.
2025-11-21T17:48:18 < PhantomWork> jbo: https://i.imgur.com/03EQQl9.png
2025-11-21T17:48:38 < jbo> yeah no fuck that
2025-11-21T17:48:51 < PhantomWork> the SW pin is painful
2025-11-21T17:49:03 < jbo> just put the inductor right next to it.
2025-11-21T17:50:51 < jbo> PhantomWork, https://bin.jvnv.net/file/RAcHp.png
2025-11-21T17:53:13 < jbo> PhantomWork, https://bin.jvnv.net/file/XkUGV.png
2025-11-21T17:53:49 < karlp> now I'm jus tconfused.  the beagle shows _zwero_ traffic when connected on the k64, either between k64 and hub or between hub and devices.  that definitely worked in days past.still happily shows data when connected on the k70.
2025-11-21T17:53:56 < karlp> that's clearly a "fuck fridays" thing
2025-11-21T17:54:19 < karlp> earlier I had it showing me errors on _everything_ when one was working and one wasn't.
2025-11-21T18:01:23 < PhantomWork> jbo: looking, thanks
2025-11-21T18:01:59 < PhantomWork> atleast this one should have a proper non-overloaded circuit compared to what I had, even if it take the same space...
2025-11-21T18:13:17 < PhantomWork> jbo: is it possible that the 6 bottom left vias are wrong? shouln't that be GND instead of 3V3?
2025-11-21T18:18:16 < PhantomWork> more of a sanity check than anything, just wondering if I missed something or if you made a mistake/different design
2025-11-21T18:26:21 < PhantomWork> https://i.imgur.com/mQqB24q.png   seems to be quite more sane now, green is the old smps board space, so a bit taller, but quite less wide. Considering that the next step is to make a 5 to 3.3V buck, that will replace the 6-24V to 3.3v buck, will be quite smaller in the end
2025-11-21T18:27:54 < PhantomWork> I could even add a 4th input cap (about 15ft of noisy 12v wire)
2025-11-21T18:27:58 < PhantomWork> thanks alot
2025-11-21T18:29:27 < PhantomWork> time to go home, short friday!
2025-11-21T19:08:03 < karlp> so... they should probably both fail to enumerate, and it's luck taht the mouse does? https://imgur.com/a/lvhpDRl
2025-11-21T19:08:21 < karlp> goin to look at the hub with it on monday, and remove the tvs...
2025-11-21T19:08:41 < karlp> this had no particular attention to the grounding of the scope prob either, just wanted to get a feel for it.
2025-11-21T19:16:26 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-21T19:22:57 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-21T19:30:40 -!- catphish [~quassel@user/catphish] has joined ##stm32
2025-11-21T19:37:23 < zyp> karlp, does the beagle see vbus? IME it doesn't bother attempting anything else before it sees vbus
2025-11-21T19:38:34 < zyp> but, uh, is that a mix of 3.3V and 5V IO?
2025-11-21T19:40:34 < zyp> kinda looks like somebody fucked up vddio for the USB pins somewhere
2025-11-21T19:43:13 < zyp> you get two levels since USB is bidirectional and you have different vddio on either end
2025-11-21T19:43:54 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-21T19:44:35 < zyp> and the issue might actually be something silly like your 5V input buffers being at marginal V_IH when fed 3.3V
2025-11-21T19:47:11 -!- amg_project [~spot@user/amg-project:52504] has joined ##stm32
2025-11-21T20:13:10 -!- amg_project [~spot@user/amg-project:52504] has quit [Quit: Leaving]
2025-11-21T20:15:23 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-21T20:19:26 -!- Netsplit *.net <-> *.split quits: mid-kid, octorian, karlp
2025-11-21T20:24:43 -!- mid-kid [~mid-kid@2a01:7c8:aac8:1e8:5054:ff:fe5e:cd48] has joined ##stm32
2025-11-21T20:24:44 -!- Netsplit over, joins: octorian, karlp
2025-11-21T20:43:59 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-21T20:47:34 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-21T20:53:59 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-21T20:59:42 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-21T21:59:13 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 260 seconds]
2025-11-21T22:17:55 < qyx> jbo: what a beautiful one
2025-11-21T22:31:52 -!- c10ud__ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-21T22:34:56 -!- haritzondo [~hrtz@140.228.70.141] has joined ##stm32
2025-11-21T22:35:25 -!- ventYl_ [~ventyl@adsl-dyn63.78-98-146.t-com.sk] has joined ##stm32
2025-11-21T22:35:40 -!- teknix_ [~unknown@user/hsv] has joined ##stm32
2025-11-21T22:37:11 < karlp> zyp: it gets vbus, yeah, I have no idea why it wasn't seeing things properly on the k64 this afternon.
2025-11-21T22:37:32 < karlp> the two levels were "fucking wot"? but out of time.
2025-11-21T22:37:40 < karlp> what 5V input buffers do you  mean?
2025-11-21T22:38:00 < zyp> I mean, it looks like you're feeding 5V into a rail that should have 3.3V
2025-11-21T22:38:23 < zyp> if you have some schematics, I can take a look
2025-11-21T22:38:33 < karlp> there's a 3.3V tvs, which I think is part of the problem, I want to remove that next, I think it should be either a proper usb designed one, or at least a 5V one.
2025-11-21T22:38:49 < karlp> I just couldn't find another board to abuse late this afternoon.
2025-11-21T22:39:57 < zyp> are you able to share schematics? I'm invested into this now :p
2025-11-21T22:40:21 < karlp> I can, was just logging into altium cloud shits
2025-11-21T22:40:33 < karlp> but I can only share the original sch, with the hub on it....
2025-11-21T22:41:21 < zyp> the one that doesn't work is what I'm interested in, MCU and hub
2025-11-21T22:41:27 -!- Netsplit *.net <-> *.split quits: haritz, c10ud_, ventYl, teknix
2025-11-21T22:41:35 < karlp> part A: the hub, the left USB0 is to the MCU pins: https://bin.jvnv.net/file/zGASn.png
2025-11-21T22:42:28 < karlp> part B: the downstream ports: https://bin.jvnv.net/file/J4POk.png
2025-11-21T22:42:39 < karlp> no idea why the TVS is not connector side.
2025-11-21T22:42:43 < karlp> no idea why it's 3v3.
2025-11-21T22:43:05 < karlp> no idea why the 22R resistors are there, hub advertises internal termination, and k70 recommends 33r anyway, not 22r.
2025-11-21T22:43:41 < zyp> where were the scope traces probed?
2025-11-21T22:44:06 < karlp> I did experiment with using either the 48MHZ oscillator that feeds the hub to also be the usb clock for the k70, or to use one of it's plls isntead, made no discernible difference.
2025-11-21T22:44:23 < karlp> probing was grounded shell of the usb conenctors to... either dm or dp. 
2025-11-21T22:44:33 < karlp> one tick, I know which one I was physically looking at
2025-11-21T22:44:42 < karlp> oh, but those scope traces are frrom the board where I lifted the hub off.
2025-11-21T22:44:49 < zyp> good
2025-11-21T22:44:49 < karlp> soo....
2025-11-21T22:44:55 < zyp> mcu schematic?
2025-11-21T22:45:14 < karlp> not much to see there.
2025-11-21T22:45:22 < karlp> usbDP and DM go to usbDP and DM...
2025-11-21T22:45:34 < karlp> there's pins for teh hub reset and shit, they work.....
2025-11-21T22:45:43 < zyp> that's not what I'll be looking for :)
2025-11-21T22:45:54 < karlp> then tell me what you're looking for .)
2025-11-21T22:46:28 < zyp> where the 5V is coming from
2025-11-21T22:47:02 < karlp> in the real design, there's a system 5V that goes to the port switches
2025-11-21T22:47:14 < zyp> not that
2025-11-21T22:47:37 < karlp> when I lifted teh hub off, I just took a 5V directly and soldered it to vbus on the connector
2025-11-21T22:47:54 < zyp> I mean the 5V in your scope traces
2025-11-21T22:48:27 < zyp> from what you've told me so far, it's coming from the mcu, I wanna see why the MCU is outputting 5V on DP/DM
2025-11-21T22:49:44 < karlp> I can't think wha tit would be, but I can't share that sheet in itðs entirety I don't think.
2025-11-21T22:50:00 < zyp> do you need me to sign a NDA? :)
2025-11-21T22:50:05  * karlp shrugs
2025-11-21T22:50:17 < karlp> probably not :)
2025-11-21T22:50:44 < zyp> then you can pretend I did and send it to me privately or email it or something
2025-11-21T22:50:55 < karlp> I would more suspect my rework being an issue.
2025-11-21T22:51:16 < karlp> the termination resistors are dangling stubs now, with thehub lifted off, the port power switches are still there, ... 
2025-11-21T22:51:27 < karlp> what's your best email?
2025-11-21T22:51:32 < zyp> zyp@jvnv.net
2025-11-21T22:57:32 < karlp> my plan for monday is to find another board, remove the tvs first, see how that goes, and also try probing a normal board on the lines from the mcu to the hub.
2025-11-21T22:57:59 < karlp> and maybe try scoping with a FS device as well.
2025-11-21T22:58:12 < karlp> and fingers crossed the beagle decides to work on the k64 again.
2025-11-21T22:58:18 < karlp> and I guess move my laptop into the lab.
2025-11-21T23:05:32 < zyp> ah, I love that NXP makes me log in to download the RM
2025-11-21T23:05:51 < zyp> and then I had to update the password as well
2025-11-21T23:07:12 < karlp> oh, I could have sent you that.
2025-11-21T23:07:16 < karlp> yeah, k70 is weird.
2025-11-21T23:07:25 < karlp> it got killed, nrnd, 
2025-11-21T23:07:38 < karlp> it's the only kinetis part not supported in mcuxpresso,
2025-11-21T23:07:50 < karlp> and then late last year ti came back to life, 2035 supply commitment.
2025-11-21T23:07:58 < karlp> so obviosuly there's some big customer using it....
2025-11-21T23:08:04 < karlp> no idea if they use usb though....
2025-11-21T23:08:09 < qyx> so, I was going to start working
2025-11-21T23:08:13 < qyx> it was one hour ago
2025-11-21T23:08:15 < qyx> thanks youtube
2025-11-21T23:15:08 < zyp> karlp, so according to the datasheet, the internal USB PHY is fed by the internal regulator that's supplied 5V through VREGIN
2025-11-21T23:15:48 < zyp> regulator has a pass-through mode, but it's autoswitching depending on input voltage, so there's no setting there to fuck up
2025-11-21T23:16:28 < zyp> but there's no decoupling on VOUT33, datasheet calls for 2.2uF nominal, 1.76uF minimal
2025-11-21T23:17:03 < zyp> I suspect lack of output capacitance makes regulator unstable and fucking up your PHY supply
2025-11-21T23:18:23 < zyp> since VOUT33 is on ball L1, you should be able to access it from the side and bodge on some capacitance, I suggest trying that
2025-11-21T23:20:14 < zyp> the capacitance requirement comes from page 66 in here: https://www.nxp.com/docs/en/data-sheet/K70P256M150SF3.pdf
--- Day changed Sat Nov 22 2025
2025-11-22T00:13:40 < karlp> that's a very good catch.
2025-11-22T00:14:46 < zyp> hope it helps and it's not yet another false lead
2025-11-22T00:15:09 < zyp> but it's the most plausible explanation I have given your scope traces
2025-11-22T00:15:22 < karlp> yeah, I really wasn't expecting those double levels 
2025-11-22T00:15:40 < karlp> well, it can wait til monday now. hope you can tolerate the wait.
2025-11-22T00:15:43 < zyp> the reasonable looking ones would be the replies from the device
2025-11-22T00:15:48 < zyp> :)
2025-11-22T00:15:56 < karlp> no fucking clue if I'm going to be able to do that rework anyway.
2025-11-22T00:16:20 < karlp> and we still just have to assume that fs works becuase fs devices have better transceivers in them?
2025-11-22T00:16:27 < zyp> just grab a pth cap, solder one leg to ground and poke the other at the ball and see if it starts behaving
2025-11-22T00:17:10 < zyp> FS might work because the regulator behaves better with a higher/more frequent load or something
2025-11-22T00:17:34 < karlp> fecking caps right up against the part there.
2025-11-22T00:17:39 < karlp> still fuck that, mondays' problem.
2025-11-22T00:18:05 -!- jerrycash [~jerrycash@user/jerrycash] has quit [Read error: Connection reset by peer]
2025-11-22T00:18:23 < karlp> https://bin.jvnv.net/file/1FILb.png
2025-11-22T00:18:54 < zyp> worst case you get to desolder the bga, bodge a magnet wire to the pad and put it back :)
2025-11-22T00:18:58 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-22T00:19:30 < zyp> looks accessible enogh
2025-11-22T00:20:41 < zyp> just throw a piece of tape over the 5V cap and you can even use it for support :)
2025-11-22T00:35:45 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 245 seconds]
2025-11-22T00:37:15 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 250 seconds]
2025-11-22T01:17:21 -!- qyx [~qyx@84.245.120.82] has quit [Ping timeout: 252 seconds]
2025-11-22T01:19:20 -!- qyx [~qyx@84.245.121.191] has joined ##stm32
2025-11-22T01:22:51 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 252 seconds]
2025-11-22T01:36:19 -!- haritzondo [~hrtz@140.228.70.141] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-22T01:56:52 -!- travisb is now known as tabemann
2025-11-22T01:58:10 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-22T01:59:52 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-22T02:22:40 -!- hexo [~hexo@user/hexo] has joined ##stm32
2025-11-22T05:33:20 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has joined ##stm32
2025-11-22T05:36:37 -!- c10ud__ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has quit [Ping timeout: 264 seconds]
2025-11-22T05:37:22 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-22T07:35:08 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-22T07:47:03 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-22T07:58:36 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-22T08:52:32 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-22T09:08:07 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-22T10:21:33 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 265 seconds]
2025-11-22T10:55:08 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-22T11:00:57 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-22T11:17:49 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-22T12:13:58 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-22T12:21:57 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-22T12:23:56 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-22T13:25:14 -!- polprog [~ath0@about/hackers/polprog] has quit [Remote host closed the connection]
2025-11-22T14:06:08 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-22T14:33:30 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-22T14:45:56 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-22T15:23:14 -!- polprog [~ath0@about/hackers/polprog] has joined ##stm32
2025-11-22T15:27:53 -!- polprog [~ath0@about/hackers/polprog] has quit [Remote host closed the connection]
2025-11-22T17:46:13 -!- ColdKeybo[a]rd [~ColdKeybo@user/coldkeyboard] has quit [Ping timeout: 264 seconds]
2025-11-22T17:47:01 -!- ColdKeyboard [~ColdKeybo@user/coldkeyboard] has joined ##stm32
2025-11-22T19:03:10 < jbo> qyx, my buzzer is buzzin'
2025-11-22T19:11:06 < Steffanx> I'm proud of you 🙌🏿
2025-11-22T19:12:44 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-22T19:19:35 < mawk> single precision floats suck
2025-11-22T19:19:57 < mawk> I spent an hour questioning my maths abilities while it was just float fuckery
2025-11-22T19:20:07 < mawk> it works with doubles but there's no hardware double support
2025-11-22T19:22:25 < Steffanx> I'm sorry mawk 
2025-11-22T19:34:02 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-22T19:34:02 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-22T20:44:16 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-22T20:55:27 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 252 seconds]
2025-11-22T21:24:36 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-22T21:54:51 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 252 seconds]
2025-11-22T21:55:41 < qyx> jbo: my buzzer is stuck in texas
2025-11-22T21:55:47 < qyx> oh sorry, in france
--- Day changed Sun Nov 23 2025
2025-11-23T00:03:37 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-23T00:44:31 < mawk> he's drinking wine and smoking a cigarette right now qyx 
2025-11-23T01:28:59 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has joined ##stm32
2025-11-23T01:40:47 < nomorekaki> original DJI ronins go for nothing nowdays
2025-11-23T01:45:04 < nomorekaki> like 50eur but no batteries
2025-11-23T01:50:04 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-23T02:12:36 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Quit: Bye]
2025-11-23T06:45:11 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-23T07:56:53 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-23T08:48:13 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-23T08:55:22 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-23T09:15:43 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 240 seconds]
2025-11-23T10:01:03 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-23T10:17:11 < Steffanx>  Is the buzzer a privileged white man, mawk??? 
2025-11-23T10:38:23 < Steffanx> (you probably have no clue what this is about. It's some dutch thing of a few years ago)
2025-11-23T10:43:44 < ventYl_> only dutch?
2025-11-23T10:48:54 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-23T11:00:28 < Steffanx> I don't know. It was a dutch guy who said some things 
2025-11-23T11:03:54 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-23T11:39:05 < ventYl_> Well, maybe these people weren't dutchmens, but male with white privilege used to be a rather common phrase a couple of years ago
2025-11-23T11:39:09 < ventYl_> in many countries
2025-11-23T11:39:35 < ventYl_> or variants, like white cis male scum
2025-11-23T12:40:22 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-23T14:11:31 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has joined ##stm32
2025-11-23T14:15:41 < jbo> why did I have to read that sentence three times until I noticed the 's' in "scum" ._.
2025-11-23T14:19:12 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-23T14:19:33 < ventYl_> ask your ophtalmologist
2025-11-23T15:53:20 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Ping timeout: 240 seconds]
2025-11-23T15:53:47 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has joined ##stm32
2025-11-23T16:37:31 -!- IanW_ [~IceChat9@jindivik.force9.co.uk] has quit [Quit: Bye]
2025-11-23T16:54:01 < Steffanx> You missed an h in that word ventYl_ 
2025-11-23T17:09:37 -!- nomorekaki [~nomorekak@37-219-167-227.nat.bb.dnainternet.fi] has quit [Ping timeout: 250 seconds]
2025-11-23T17:58:27 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-23T18:03:30 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-23T18:03:30 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-23T18:07:47 -!- polprog [~ath0@about/hackers/polprog] has joined ##stm32
2025-11-23T20:47:43 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-23T21:31:11 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-23T21:38:29 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-23T23:45:12 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
--- Day changed Mon Nov 24 2025
2025-11-24T01:18:30 -!- qyx [~qyx@84.245.121.191] has quit [Ping timeout: 256 seconds]
2025-11-24T01:20:11 -!- qyx [~qyx@84.245.120.70] has joined ##stm32
2025-11-24T01:28:25 -!- martinmoene_ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 246 seconds]
2025-11-24T04:00:29 -!- martinmoene_ [~martinmoe@132.229.46.129] has joined ##stm32
2025-11-24T04:04:18 -!- martinmoene [~martinmoe@132.229.46.129] has quit [Ping timeout: 260 seconds]
2025-11-24T07:42:00 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-24T08:17:29 -!- haritz [~hrtz@user/haritz] has quit [Remote host closed the connection]
2025-11-24T08:24:46 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-24T09:13:19 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-24T09:24:55 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-24T10:20:01 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 250 seconds]
2025-11-24T10:24:26 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-24T10:57:27 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-24T12:09:18 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-24T12:09:18 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-24T13:19:35 < qyx> no innovation today
2025-11-24T13:19:42 < qyx> innovation killed by dumb requests
2025-11-24T13:19:50  * qyx unpacks his mouser
2025-11-24T13:48:40 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-24T14:04:32  * jpa- mouses qyx a bit
2025-11-24T14:05:29 < zyp> be nice
2025-11-24T14:05:46 < polprog> slow day at work today
2025-11-24T14:17:21 < mawk> yummy jbo 
2025-11-24T14:17:41 < mawk> I have a mouser currently trying to eat my router
2025-11-24T14:18:47 < mawk> https://i.imgur.com/1lPcJOL.jpeg
2025-11-24T14:19:05 < mawk> is backup RAM slower than regular RAM?
2025-11-24T14:19:54 < mawk> the finns are currently giving a business presentation and idk why the guy is rambling about saunas
2025-11-24T14:19:58 < mawk> very finn of him
2025-11-24T14:20:07 < mawk> now he's rambling about santa claus
2025-11-24T14:20:11 < mawk> very important meeting
2025-11-24T14:22:34 < zyp> mawk, probably, but maybe not enough to matter
2025-11-24T14:23:41 < mawk> my colleagues did not know about the backup ram and kept using the external EEPROM to transmit data across reboots
2025-11-24T14:23:50 < mawk> we even have a backup battery to keep it and the RTC powered
2025-11-24T14:25:36 < qyx> I am using backup registers for that purpose
2025-11-24T14:25:44 < qyx> if rtc backup power is available
2025-11-24T14:26:50 < zyp> looks like it has an AHB hookup, separate from the APB hookup the rest of the backup domain has, so probably not a significant speed penalty
2025-11-24T14:27:17 < mawk> they're also using the RTC backup register but it's just 4 bytes so it's not enough
2025-11-24T14:27:29 < mawk> ah I see, thanks
2025-11-24T14:28:39 < mawk> if only the device had its lid tamper detection pin connected to a RTC tamper pin we could save a lot of state in the backup RAM and assume it's still valid if the device has not been opened
2025-11-24T14:28:44 < zyp> if it really mattered, I'd suggest measuring it
2025-11-24T14:29:32 < zyp> why do you need tamper detection?
2025-11-24T14:30:25 < mawk> it's a medical device, we need to know if the medicine inside has been stolen or replaced or whatever while the device was off
2025-11-24T14:31:39 < mawk> and send a snitch alert to the cloud
2025-11-24T14:31:40 < zyp> uh, so why don't you do that?
2025-11-24T14:32:11 < mawk> the lid microswitch wasn't wired to the RTC tamper pin, so we can't know what happens when the device is off
2025-11-24T14:32:16 < mawk> so we just re-analyze everything
2025-11-24T14:32:38 < mawk> the PCB design happened a long time ago and it's frozen, the boss doesn't want to hear about PCB changes
2025-11-24T14:32:55 < zyp> stuff like that is literally what the tamper pin is for
2025-11-24T14:32:59 < mawk> yeah
2025-11-24T14:37:30 < mawk> and there's a RTC tamper pin that's unused on the chip, a bit baffling why they didn't use it
2025-11-24T14:40:18 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-24T14:43:05 < polprog> ive set up a small source code indexer on http://vxworks.st
2025-11-24T14:43:18 < polprog> curious about what you think
2025-11-24T14:43:38 < polprog> for larger code bases ill try to set up bootlin Elixir later at some point
2025-11-24T14:55:08 < mawk> nice
2025-11-24T14:57:01 < karlp> well, I'm all out of ideas.
2025-11-24T14:57:24 < karlp> zyp: I got a wire onto that ball, and the levels are now stable and consistent, and it makes zero difference :|
2025-11-24T14:57:30 < karlp> I removed the tvs that was out of spec, 
2025-11-24T14:57:41 < karlp> I'v ebypassed the "extra" termination resistors, 
2025-11-24T14:58:10 < karlp> hrm, I need to get the beagle back on now, see if it has different errors now though
2025-11-24T14:58:39 < karlp> k70+mouse, k70+keyborad and k70scnaner are all fine, but the keyboards don't work through the hub.
2025-11-24T14:58:47 < karlp> but the mouse and scannera re 100% reliable.
2025-11-24T15:14:14 < karlp> yay, the beagle won't even decode the "successful" enumerations so it' obviosuly still pretty fucking borderline.
2025-11-24T15:31:32 < jpa-> have you figured out what makes it borderline? voltages or timing?
2025-11-24T15:33:02 < karlp> no idea.
2025-11-24T15:33:16 < karlp> out of ideas on how to narrow it down beyond, "yeah, hw just seems sus"
2025-11-24T15:41:24 < polprog> whats the device that doesnt enumerate?
2025-11-24T15:42:27 < karlp> three different keyboards
2025-11-24T15:43:32 < jpa-> karlp: have you tried to check the eye diagram with scope?
2025-11-24T15:45:44 < karlp> yah, but a) I don't have any sort of boundary condition, and b) it looks the "same" for working and non-working.
2025-11-24T15:46:16 < jpa-> ok, well b) was what i was interested about :)
2025-11-24T15:46:16 < karlp> and c) the keyboards work at root ports, just not past hubs, which should be re-sending everything and cleaning it up?
2025-11-24T15:46:47 < jpa-> hub could clean the timing or mess it up further
2025-11-24T15:47:19 < jpa-> does beagle decode ok without hub?
2025-11-24T15:48:23 < karlp> reports lots of errors still for the keyboard.
2025-11-24T15:49:04 < karlp> but it reports that on the mouse too, and that could just be my re-worked probed lines being less ideal.
2025-11-24T15:49:22 < jpa-> regarding boundary conditions for voltage, USB spec gives: Vlo <= 0.3V, Vhi >= 2.8V, cross-over voltage 1.3V - 2.0V
2025-11-24T15:49:47 < karlp> I think I would need to do a significantly better job probing to evaluate that.
2025-11-24T15:50:21 < karlp> there's a lot of ring from having it just clipped on and lots of stubs on the lines, this reworked board doesnt install confidence in SI muhc for me.
2025-11-24T15:50:41 < karlp> but... it still behaves consistently as always, and as the other boards.
2025-11-24T15:50:58 < karlp> mice and scanners and hubs work, even two deep, just not a keyboard after a hub
2025-11-24T15:51:03 < jpa-> boundary conditions for timing: rise & fall time: max 300 ns, bitrate: 1.4775 - 1.5225 Mbaud/s, jitter max +- 152 ns
2025-11-24T15:51:54 < jpa-> LS is so slow that i would expect to stay within the limits even with pretty crappy probing
2025-11-24T15:53:18 < jpa-> does the beagle decode ok when the keyboards are plugged into PC without hub?
2025-11-24T15:53:54 < jpa-> also, have you checked that the Vbus going to the keyboard is 5V and not something silly?
2025-11-24T15:55:45 < polprog> that would be weird, how much can a keyboard pull
2025-11-24T15:56:01 < karlp> yeah, actually, werd, no, the beagle is not doing good things on my desktop
2025-11-24T15:57:36 < karlp> looks like ~same errors in the beagle with the keyboard attache directly to my laptop
2025-11-24T15:58:40 < jpa-> polprog: well a stupid mistake like FET backwards could drop voltage at even just a few mA
2025-11-24T15:59:22 < polprog> true but then i would not expect it to work at all
2025-11-24T15:59:59 < jpa-> why? for example a FET backwards would give 4.3V that could be just borderline enough
2025-11-24T16:00:16 < polprog> hm
2025-11-24T16:00:30 < polprog> ok, that makes sense then it would give just a bit less when vbus is just a bit less
2025-11-24T16:23:18 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:1dfb:8281:7385:1ee7] has joined ##stm32
2025-11-24T17:08:14 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:1dfb:8281:7385:1ee7] has quit [Ping timeout: 256 seconds]
2025-11-24T17:29:33 < mawk> is it a good idea to periodically invert the screen to prevent OLED burn-in?
2025-11-24T17:31:29 < jpa-> is it monochrome or RGB display?
2025-11-24T17:51:40 < karlp> these are very unequal crossoves right? what would cause that? https://imgur.com/a/A3Rhd61
2025-11-24T17:54:36 < mawk> RGB
2025-11-24T17:55:27 < mawk> I also have color bleeding on the rows where pixels are brighter, but I think I can fix that by increasing the discharge period and the discharge voltage
2025-11-24T18:02:02 < mawk> the seller says it's a 64k color screen while it's actually 262k, idk why they would do that
2025-11-24T18:09:18 < ventYl_> they probably didn't realize
2025-11-24T18:13:07 < mawk> the controller does 64k colors by default but it definitely supports 262k too
2025-11-24T18:18:52 < jpa-> karlp: why is the pink voltage going so high? https://i.imgur.com/qqayj7r.jpeg
2025-11-24T18:19:30 < jpa-> usb spec is 3.6V max
2025-11-24T18:20:19 < jpa-> but yeah.. looks funky
2025-11-24T18:23:46 < jpa-> ah, i think part of it may be your trigger level being so high?
2025-11-24T18:23:55 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-24T18:25:20 < jpa-> i guess you could do only rising edge trigger to avoid that issue, otherwise if you have trigger level not in middle, the rising and falling edges from different captures overlap funnily
2025-11-24T18:26:14 < jpa-> the "doesn't work" case for some reason only appears to have JKJKJK and no actual data where same line state would stay for more than 1 bit
2025-11-24T18:38:10 < karlp> you're right, the unqueual ws jsut rom trigger voltage being not in the middle correctly., didn't think of that.
2025-11-24T18:38:59 < karlp> pink is just what the devices are giving me? that's just probing one of the lins to ground on the downstream port, yellow was k70 to hub, pink is hub port 1 to device.
2025-11-24T18:40:59 < karlp> no idea, yeah, 4.3V seems high doens't it.
2025-11-24T18:42:08 < karlp> only the LS devices have high voltage,
2025-11-24T18:42:11 < karlp> the FS device is 3.3v
2025-11-24T18:44:24 < jpa-> what kind of hub was this again?
2025-11-24T18:45:39 < jpa-> it seems you only get the high voltage with the hub?
2025-11-24T18:49:51 < jpa-> ah, another fun part is that depending on hold-off time on the scope, you may only get the host side that way.. because the host packet always comes before the reply
2025-11-24T18:51:15 < jpa-> but if things are that bad, you don't really need to mess with eye diagram at all, just get a regular single capture and look at it :)
2025-11-24T18:57:51 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-24T19:00:11 < jpa-> for reference, this is what i get with a usb ls keyboard and a random USB HS hub https://jpa.kapsi.fi/stuff/pix/usb_ls_signals.png
2025-11-24T19:02:28 < jpa-> hm, and practically the same with USB FS hub (so no TT)
2025-11-24T19:22:35 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-24T19:22:47 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-24T20:01:09 < karlp> Ive tried a couple of different hubs, two generations of cypress ones, plus something else I had lying around.
2025-11-24T20:10:33 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-24T20:42:46 < zyp> karlp, so adding decoupling fixed the weird voltage on the host port, but now you have another weird voltage on the hub port?
2025-11-24T21:00:17 < zyp> karlp, what's the pink trace in the two first pics?
2025-11-24T21:45:12 < PhantomWork> if I have a high capacitance on the usb Vbus on my device, I need to put something there to avoid overloading the host right? What is it?
2025-11-24T21:49:54 < jpa-> PhantomWork: inrush current limiter
2025-11-24T21:50:50 < PhantomWork> any recommendation on a particular type?
2025-11-24T21:51:00 < jpa-> i've used TPS2552
2025-11-24T21:53:38 < PhantomWork> looks like a perfect diode
2025-11-24T21:53:47 < PhantomWork> so  no need to even have the diode then!
2025-11-24T22:06:38 < PhantomWork> thanks, this now look more sane for sure!
2025-11-24T22:09:55 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-24T22:17:30 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-24T22:32:12 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-24T22:38:26 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-24T22:42:48 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-24T23:00:48 < karlp> zyp: in this album: https://imgur.com/a/A3Rhd61 the first two pics, the pink trace was actaully on the vbus line, don't read anything into that.
2025-11-24T23:01:52 < karlp> zyp: well, I only ave this high voltage on the port between the hub and teh LS devices now...
2025-11-24T23:02:46 < karlp> I think I need to make a new spy cable to better sub out hubs and still inspect them.
2025-11-24T23:03:20 < karlp> removing the suspect tvs, shorting out the extra termination resistors and fixing teh decouplign really felt like it should have been enough to unfuck this.
2025-11-24T23:03:42 < karlp> starting to think segger knows something where they refuse to support hubs with their paid host driver for KHCI.
2025-11-24T23:09:18 < karlp> just ... kinda feels like it really might just not ever work.
--- Day changed Tue Nov 25 2025
2025-11-25T00:40:43 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 255 seconds]
2025-11-25T00:45:16 < qyx> one more lm66200 to find and I can call this proto completed
2025-11-25T00:45:42 < qyx> but lm66200 is hiding well
2025-11-25T01:03:31 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-25T01:06:13 < Phantom> it'S nice, I basically been able to fit both 5V and 3.3V reg into the same area as one older (also broken) version, plus a proper usb power switch!
2025-11-25T01:06:41 < Phantom> a bit more tinkering and that section will be done for good yay
2025-11-25T01:10:56 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-25T02:40:24 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-25T02:44:51 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-25T03:59:36 -!- PhantomWork [~PhantomWo@user/phantom] has quit [Read error: Connection reset by peer]
2025-11-25T05:24:26 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-25T05:26:42 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-25T05:42:12 -!- c10ud [~c10ud@user/c10ud] has joined ##stm32
2025-11-25T05:42:26 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-25T05:43:20 -!- c10ud_ [~c10ud@host-87-2-122-85.retail.telecomitalia.it] has quit [Ping timeout: 240 seconds]
2025-11-25T08:11:35 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-25T09:14:35 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has joined ##stm32
2025-11-25T09:27:38 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-25T09:44:26 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-25T09:44:37 -!- Livio [~livio@user/livio] has quit [Client Quit]
2025-11-25T10:18:55 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-25T10:56:44 < qyx> what the hell, 100base-T1 requires master/slave setting
2025-11-25T11:25:39 < qyx> so I am failing at reading phy registers
2025-11-25T11:43:39 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-25T11:43:39 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-25T12:59:02 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has joined ##stm32
2025-11-25T13:18:40 -!- artok [~azo@88-112-154-28.elisa-laajakaista.fi] has quit [Ping timeout: 244 seconds]
2025-11-25T13:20:48 -!- artok [~azo@91-153-163-37.elisa-laajakaista.fi] has joined ##stm32
2025-11-25T13:24:56 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has quit [Remote host closed the connection]
2025-11-25T13:28:03 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has joined ##stm32
2025-11-25T13:32:27 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has quit [Max SendQ exceeded]
2025-11-25T13:33:39 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has joined ##stm32
2025-11-25T13:57:10 -!- rob_w [~bob@host-82-135-31-73.customer.m-online.net] has quit [Remote host closed the connection]
2025-11-25T14:25:13 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-25T15:16:45 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has quit [Remote host closed the connection]
2025-11-25T15:34:52 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has joined ##stm32
2025-11-25T16:16:46 < karlp> fuck me, this rigol rp2200 150 mhz scope is _garbage_ at usb.
2025-11-25T16:17:03 < karlp> dug out the correct matching siglent one that cam ewith it, and now both channels so "same" data :)
2025-11-25T17:07:12 -!- Livio [~livio@user/livio] has quit [Remote host closed the connection]
2025-11-25T17:10:55 < qyx> [   60.834794] fec 42890000.ethernet end1: Link is Up - 100Mbps/Full - flow control off
2025-11-25T17:11:02 < qyx> half a day it took
2025-11-25T17:22:16 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-25T17:25:51 < qyx> the link is flapping between two DP83TC813 which may be caused by my farmer-quality cabling
2025-11-25T17:26:11 < qyx> DP83TC813 to a 10base-t1 switch works though
2025-11-25T17:26:25 < qyx> if i set DP to slave, I get no-carrier, when I set to master, I get link up
2025-11-25T17:26:42 < qyx> RMII between the switch and imx9 seems to be working, I have link up
2025-11-25T17:26:52 < qyx> but the switch is not switching
2025-11-25T17:26:57 < qyx> I wonder what the default config is
2025-11-25T17:43:12 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-25T17:45:44 -!- ilgrim2 [~ilgrim@xinu.me] has quit [Ping timeout: 240 seconds]
2025-11-25T17:47:40 -!- ilgrim [~ilgrim@xinu.me] has joined ##stm32
2025-11-25T18:20:02 < karlp> is this your own switch, or did you buy a fancy cots t1s switch?
2025-11-25T18:20:35 < karlp> also, I'v ebeen meaning to ask, for your debian imx9x builds, what are you buildind it with? buildroot? or something of your own? you're allergic to yocto iirc?
2025-11-25T18:26:31 < karlp> Steffanx: what do you think his means in badly translated dutch? " Industry standard shields for extention with Bluetooth, WiFi, Ethernet based" ?
2025-11-25T18:26:35 < karlp> shields?
2025-11-25T18:26:42 < qyx> vanilla imx-linux
2025-11-25T18:27:09 < karlp> does the imx-linux tree include RFS creation shits? not just kernel?
2025-11-25T18:27:14 < qyx> and debootstrapped debian or official prebuilt .tgz openwrt rootfs
2025-11-25T18:27:47 < karlp> those... are very different things, but ok...
2025-11-25T18:27:55 < qyx> yes, I am using both
2025-11-25T18:28:13 < qyx> debian for dataloggers, openwrt for network stuff
2025-11-25T18:28:34 < qyx> although I don't have openwrt for imx93 in field yet, only imx6
2025-11-25T18:31:16 < karlp> oh, they might actually mean arduino shields, "Hardware slot for arduino uno shields."
2025-11-25T18:31:22 < karlp> thjat's.... impressive
2025-11-25T18:47:56 < mawk> where's the dutch
2025-11-25T18:48:14 < mawk> do you like buildroot qyx ?
2025-11-25T18:53:16 < karlp> mawk: there' sno dutch, reading a dutch company website and seeing odd english
2025-11-25T18:53:33 < mawk> ah right
2025-11-25T19:07:53 < karlp> fucking podman just absolutely fucking murdered people with this migration of db backends: https://discussion.fedoraproject.org/t/podman-5-7-boltdb-to-sqlite-migration/171172/7
2025-11-25T19:07:58 < karlp> "just reset and make it all again"
2025-11-25T19:24:43 < Steffanx> Is it WPL Karlp?  I found the page, but even the dutch page makes no sense
2025-11-25T19:28:28 < Steffanx> Almost like they wrote it in shitty English and translated that to dutch, word for word.
2025-11-25T19:54:25 < qyx> karlp: do you have any experience with the DSA switching thing in linux?
2025-11-25T19:59:55 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-25T20:42:25 -!- PhantomWork [~PhantomWo@user/phantom] has joined ##stm32
2025-11-25T21:00:01 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 264 seconds]
2025-11-25T21:18:30 < mawk> using ip qyx ?
2025-11-25T21:18:34 < mawk> I mean the ip command
2025-11-25T22:17:02 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has joined ##stm32
2025-11-25T22:34:52 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-25T23:02:16 < mawk> when I look at the current waveform of our F7 during sleep mode there is a weird negative (relative to the average) current spike every 17 ms
2025-11-25T23:02:19 < mawk> what could it be
2025-11-25T23:02:39 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-25T23:02:47 < mawk> something from the buck converters? but why the weird period then
--- Day changed Wed Nov 26 2025
2025-11-26T00:24:56 -!- bitmask [~bitmask@ool-43525d25.dyn.optonline.net] has quit [Ping timeout: 240 seconds]
2025-11-26T00:43:21 < zyp> mawk, what currents are we talking? got a plot to share?
2025-11-26T00:44:07 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-26T00:46:04 < mawk> like a mA
2025-11-26T00:46:08 < mawk> iirc
2025-11-26T00:46:23 < mawk> I'll make a plot when I'm back at werk
2025-11-26T00:46:52 < zyp> okay, so not very deep sleep
2025-11-26T00:47:30 < zyp> are you measuring with or without debugger attached?
2025-11-26T00:50:18 < qyx> mawk: no, at the devicetree level
2025-11-26T00:50:34 < qyx> making lan9370 behave
2025-11-26T00:59:14 < qyx>                 spi0 = &lpspi4;
2025-11-26T00:59:14 < qyx>                 spi1 = &lpspi3;
2025-11-26T00:59:14 < qyx>                 spi2 = &lpspi1;
2025-11-26T00:59:18 < qyx> I hate this life
2025-11-26T00:59:56 < qyx> can I name it spisw, spimpci, etc.?
2025-11-26T01:04:34 < karlp> Steffanx: yeah, wpl
2025-11-26T01:04:48 < karlp> qyx: no dsa experience sorry
2025-11-26T01:05:02 < karlp> it's "the new thing" with "infinite flexibility" 
2025-11-26T01:05:21 < karlp> Steffanx: yeah, I sort of almost got the feeling it was a dutch frront for a shenzen operation....
2025-11-26T01:06:00 < Phantom> this circuit is annoying... I need ground isolation between two parts as several volts can differ between both grounds... Yet need to communicate.... and take measures (ADC)... and make adjustement (digipot)... So... best solution seems to be... I2C an ADC... :/
2025-11-26T01:18:04 -!- qyx [~qyx@84.245.120.70] has quit [Ping timeout: 246 seconds]
2025-11-26T01:20:04 -!- qyx [~qyx@84.245.121.36] has joined ##stm32
2025-11-26T02:01:03 < zyp> have you considered just doing a difference amplifier?
2025-11-26T02:03:25 < zyp> hmm, might be inconvenient if measuring below your reference ground
2025-11-26T02:05:05 < zyp> I still don't have the full picture of what you're doing, but it kinda sounds like you might have the MCU on the wrong side of the isolation barrier
2025-11-26T02:08:15 < ventYl_> so, writing an operating system is not enough
2025-11-26T02:08:35 < ventYl_> i wrote a port that essentially emulates a MMU-less system inside Linux process and runs my RTOS inside it
2025-11-26T02:08:56 < ventYl_> yet, this was still not enough, so I added a peripheral emulation mechanism
2025-11-26T02:09:17 < ventYl_> now, now I can debug an application, an operating system, a CPU and a bunch of peripherals, all at the same time
2025-11-26T02:16:19 < Phantom> zyp: yeah, that.s the issue, the STM side ground will be higher than the circuit to be measured ground... so standard diff amp don't work, so I'ld need to add and remove an offset, that would then cause issues to figure out how much to precisely remove, or to add a circuit to create a negative rail...
2025-11-26T02:17:38 < Phantom> and no, because actually it is 2 circuits with different ground (total 3 grounds) to be measured, and 99% of the work need to be done on that side, so nope
2025-11-26T02:19:54 < Phantom> seriously, simplest, and possibly cheapest, is to use the 85 cents I2C ADC, plus it is a 4 channels 16 bits. ok, it is a slow one (800sps!) but plently fast
2025-11-26T02:31:20 -!- ThatDamnRanga_ [~ThatDamnR@UNASSIGNED-218-100-26-71.3cix.nzix.net] has quit [Ping timeout: 240 seconds]
2025-11-26T02:34:03 < qyx> I can't even get SPI working
2025-11-26T02:36:38 -!- ThatDamnRanga [~ThatDamnR@UNASSIGNED-218-100-26-71.3cix.nzix.net] has joined ##stm32
2025-11-26T02:37:40 < Phantom> I got I2C to work on F103, but weirdly it is not reliable on the LDO from USB, but work fine from the "broken" buck regulator
2025-11-26T02:39:27 < Phantom> now I will upgrade to G431, with not only an eeprom, but also 2x I2C isolators + 2x I2C digipot + 2x I2C ADC
2025-11-26T05:21:47 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-26T05:43:47 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-26T06:14:40 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-26T06:31:31 < ds2> is your pull ups too small?
2025-11-26T06:32:28 < Phantom> ?
2025-11-26T06:34:31 < Phantom> oh  the issue... 
2025-11-26T06:34:42 < Phantom> 4.7k on 3.3V
2025-11-26T06:35:06 < Phantom> so maybe, but 1 device only and about 1-1.5" of trace
2025-11-26T07:31:32 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-26T07:40:21 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-26T08:16:40 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-26T08:23:57 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-26T08:27:49 < jpa-> Phantom: if you need an isolation barrier, SPI is usually easier to pass through isolators than I2C
2025-11-26T08:28:35 < jpa-> but differential amps do remove ground offset without problem
2025-11-26T08:28:48 < jpa-> you just need one with wide enough common mode range
2025-11-26T08:47:43 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 240 seconds]
2025-11-26T08:59:20 < qyx> is it morning again
2025-11-26T09:06:09 < jpa-> yes, blabbing will now continue until innovations improve
2025-11-26T09:24:58 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-26T09:41:30 < qyx> We are pleased to inform you that Advent International will complete its acquisition of u-blox Holding AG today
2025-11-26T09:41:33 < qyx> wat
2025-11-26T09:54:31 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-26T09:54:39 < BrainDamage> with those messages, I always feel like they are the ones pleased - and none else
2025-11-26T09:59:22 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:a142:6824:e115:e246] has quit [Read error: Connection reset by peer]
2025-11-26T10:11:32 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-26T10:21:30 < mawk> I thought it was already acquired by trasna qyx 
2025-11-26T10:21:40 < mawk> ah maybe trasna only got the cellular stuff
2025-11-26T10:21:45 < mawk> and u-blox is only GPS remaining
2025-11-26T10:35:34 < mawk> <ds2> is your pull ups too small?
2025-11-26T10:35:39 < mawk> it's not the size that matters
2025-11-26T10:35:46 < mawk> it's the way you use it
2025-11-26T11:11:51 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-26T11:25:24 -!- Livio [~livio@user/livio] has quit [Quit: leaving]
2025-11-26T11:39:04 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-26T11:39:04 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-26T12:00:29 -!- tomeaton17 [~tom@user/tomeaton17] has quit [Quit: ZNC 1.9.0+deb2ubuntu0.1~esm2 - https://znc.in]
2025-11-26T12:01:05 -!- tomeaton17 [~tom@user/tomeaton17] has joined ##stm32
2025-11-26T13:40:13 -!- aandrew [~aandrew@vps-2437c00c.vps.ovh.ca] has quit [Ping timeout: 264 seconds]
2025-11-26T13:40:49 -!- aandrew [~aandrew@vps-2437c00c.vps.ovh.ca] has joined ##stm32
2025-11-26T16:52:00 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-26T16:57:01 -!- yukam [~yukam@user/yukam] has quit [Ping timeout: 264 seconds]
2025-11-26T17:07:25 -!- yukam [~yukam@user/yukam] has joined ##stm32
2025-11-26T17:38:42 < mrec_> can anyone recommend an stm32h7 lwip/freertos example that doesn't suck and just works?
2025-11-26T18:36:15 < mrec_> I wonder what is so difficult to write a documentation which just works. All the memory allocations aren't rocket sience yet they still fuck up everything
2025-11-26T18:52:25 < ventYl_> portability is a curse word in the world of embedded
2025-11-26T18:52:35 < ventYl_> everyone is: here we integrated our stuff for you
2025-11-26T18:52:59 < ventYl_> oh, you don't sport the MCU x RTOS combo they created an integration? what a shame
2025-11-26T19:00:45 < mrec_> stm32cubeide just copies the stuff together but it won't work afterwards because it's missing all the memory settings (at least with the H7 part)
2025-11-26T19:01:26 < ventYl_> aren't memory settings an integrator's duty to configure, after all?
2025-11-26T19:01:59 < mrec_> if it would be properly documented how things play together certainly.
2025-11-26T19:02:37 < mrec_> St is changing options shoveling them from A to B ... lwip is renaming section names
2025-11-26T19:02:50 < mrec_> (among different versions)
2025-11-26T19:05:52 < mrec_> if properly documented the basics can be done within 30 minutes.. yet I'm googling and finding tons of discussions about it
2025-11-26T19:06:09 < mrec_> and hours have passed thanks to that nonsense.
2025-11-26T19:15:55 -!- PhantomWork2 [~PhantomWo@modemcable022.175-37-24.static.videotron.ca] has joined ##stm32
2025-11-26T19:16:15 < mrec_> anyway I'm sure I'm almost there.
2025-11-26T19:18:59 -!- haritz [~hrtz@user/haritz] has quit [Ping timeout: 240 seconds]
2025-11-26T19:18:59 -!- artok [~azo@91-153-163-37.elisa-laajakaista.fi] has quit [Ping timeout: 240 seconds]
2025-11-26T19:19:00 -!- PhantomWork [~PhantomWo@user/phantom] has quit [Ping timeout: 240 seconds]
2025-11-26T19:19:32 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-26T19:21:11 < ventYl_> you are always welcome to improve the docs :)
2025-11-26T19:21:15 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-26T19:21:15 < ventYl_> lwip is an opensource project
2025-11-26T19:22:44 -!- yukam [~yukam@user/yukam] has quit [Ping timeout: 240 seconds]
2025-11-26T19:26:26 < qyx> lwip should habe been dead long before 2020
2025-11-26T19:26:47 < ventYl_> what else to use?
2025-11-26T19:26:48 < qyx> together with fatfs and freertos
2025-11-26T19:27:20 < qyx> thats exactly the problem, the whole embedded world is depending on a couple outdated projects
2025-11-26T19:27:48 < qyx> and we are lazy to come up with some more current and usable alternatives
2025-11-26T19:28:08 < ventYl_> nah, i already wrote OS, I am not going to write TCP stack from scratch
2025-11-26T19:28:10 < qyx> cyclonetcp looks good for examplebut it is gpl
2025-11-26T19:28:13 < ventYl_> someone else should do it
2025-11-26T19:28:41 < specing> qyx: you can make your firmware gpl as well :)
2025-11-26T19:29:15 < qyx> yes it already is, it is not problematic from my pov
2025-11-26T19:29:23 < specing> CycloneTCP is available either as open source (GPLv2 license) or under a royalty-free commercial license (non-GPL license).
2025-11-26T19:29:33 < qyx> but random people just get randomly offended by anything gpl
2025-11-26T19:29:43 < mrec_> look at libc it's also garbage, it's the pure definition of spaghetti code
2025-11-26T19:30:00 < mrec_> but better don't look at it you'd be surprised about the bugs inside
2025-11-26T19:30:26 < mrec_> looking for 5 minutes at it and I already found a DOS issue which is able to lock an entire system
2025-11-26T19:30:32 -!- artok [~azo@91-153-163-37.elisa-laajakaista.fi] has joined ##stm32
2025-11-26T19:30:33 < mrec_> and that's the current glibc
2025-11-26T19:35:54 -!- yukam [~yukam@user/yukam] has joined ##stm32
2025-11-26T19:36:55 < specing> qyx: sure, they want gratis source :)
2025-11-26T19:49:40 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-26T19:51:39 < mawk> mrec_ cubemx generated code for freertos just works
2025-11-26T19:52:17 < mawk> you've got to select the right heap type and mark newlib as reentrant in the settings
2025-11-26T19:56:59 < mrec_> all done and it doesn't work
2025-11-26T19:57:08 < mrec_> enabled interrupts, increased the stack size, etc
2025-11-26T19:58:11 < mawk> is it a nucleo-h7?
2025-11-26T19:58:49 < mawk> and what code did you write, just like a task that blinks a LED or something?
2025-11-26T19:59:40 < mrec_> I'm taking one of the St examples now to compare, I need to free some memory on that notebook
2025-11-26T20:00:03 < mrec_> surprisingly so far the flash script looked the same
2025-11-26T20:01:51 < mrec_> I'm sure I will figure it out soon I just need to clean up some parts around it
2025-11-26T20:08:11 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-26T20:12:27 < mawk> the LD script shouldn't really be affected by having freertos or not
2025-11-26T20:19:20 < mrec_> must be a good sign even their example doesn't work with the stm32h723
2025-11-26T20:19:35 < mrec_> and I have two of those boards now (I recovered one this morning)
2025-11-26T20:26:22 -!- PhantomWork2 [~PhantomWo@user/phantom] has changed host
2025-11-26T20:34:02 < mrec_> ok... that gave me the indicator that the cable might be bad
2025-11-26T20:34:15 < mrec_> I tested my project and the sample project both work properly
2025-11-26T20:34:53 < mrec_> the leds on the ethernet board just looked fine I shouldn't have trusted that
2025-11-26T20:35:08 < mrec_> I just picked a random cable from a box here :-(
2025-11-26T20:42:54 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 260 seconds]
2025-11-26T21:53:34 < Steffanx> Lol mrec_ 
2025-11-26T22:19:11 < mawk> actual bugs in cubemx or the HAL sometimes exist but when you're trying to do something simple and it seemingly doesn't work it's likely the problem is not because of ST
2025-11-26T22:31:42 < Steffanx> EXCEPT when it comes to their Ethernet stuff
2025-11-26T22:32:13 < Steffanx> There's a topic on their forum where someone listed most/all of the problems
2025-11-26T23:33:20 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
--- Day changed Thu Nov 27 2025
2025-11-27T00:11:06 < mawk> I won't stand for such blatant STphobia
2025-11-27T00:13:49 < qyx> 18:12 < mawk> the LD script shouldn't really be affected by having freertos or not
2025-11-27T00:13:57 < qyx> yes it is kinda affected because of their memory allocators
2025-11-27T00:14:20 < qyx> some examples over there don't include the right symbols defined in the ld
2025-11-27T00:14:56 < qyx> also there are some sbrk implementations in the wild using different shit
2025-11-27T00:15:30 < qyx> ok time to finish this board
2025-11-27T00:19:55 < mawk> I mean the ld file generated by cubemx
2025-11-27T00:20:07 < mawk> even if you don't use freertos it generates the heap and stack symbols
2025-11-27T00:56:39 < Phantom> jpa-: quiestion: why do you say that I2S is easier for an isolation barrier? 
2025-11-27T00:57:00 < qyx> he said SPI
2025-11-27T00:57:12 < qyx> and because it only has unidirectional signals
2025-11-27T00:57:29 < Phantom> oh    same same
2025-11-27T00:57:31 < Phantom> ah
2025-11-27T00:57:39 < Phantom> yeah that can help yeah
2025-11-27T00:58:10 < qyx> there are i2c isolators but are more rare
2025-11-27T00:58:19  * Phantom wonders what is more common... I2C digipot/ADC    or SPI...
2025-11-27T00:58:27 < Phantom> yeah using ISOsomething on the board
2025-11-27T01:00:17 < qyx> nah anyone did LED output multiplexing with LEDs of different colors?
2025-11-27T01:00:41 < qyx> is it doable with variable duty cycle for each color? (resistors are the same for all LEDs)
2025-11-27T01:00:57 < Phantom> I'm happy, the 2 buck regulators I put should be better than what was there. Plus it alloweds me to eliminate the LDO, plus have proper usb power management, so yeah, not bad
2025-11-27T01:01:09 < Phantom> qyx: that is what those controllers do: PWM
2025-11-27T01:01:31 < Phantom> 3 PWM channels, with resistors for each leds
2025-11-27T01:01:57 < Phantom> if you can, try to have 1kHz    often controllers use like 100Hz, and it flicker
2025-11-27T01:02:27 < qyx> but the question was about multiplexing, not pwming indicidual LEDs with individual resistors
2025-11-27T01:02:43 < qyx> I have a bunch of LEDs on the front panel and want to connect them in a matrix
2025-11-27T01:02:48 < qyx> hm
2025-11-27T01:03:00 < qyx> maybe I should stop reinventing things and just use some i2c LEd driver
2025-11-27T01:03:00 < Phantom> leds are leds
2025-11-27T01:03:11 < Phantom> or adressable ones
2025-11-27T01:03:26 < qyx> ok let's stay on the ground
2025-11-27T01:03:36 < Phantom> too bad that "neopixels" ain't natively supported by STM32..
2025-11-27T01:03:48 < qyx> they are
2025-11-27T01:03:54 < qyx> I mean, what means natively?
2025-11-27T01:04:01 < Phantom> in hardware
2025-11-27T01:04:16 < Phantom> you need to bitbang using DMA...
2025-11-27T01:04:22 < Phantom> quite wastefull
2025-11-27T01:04:26 < qyx> no, just use spi
2025-11-27T01:06:24 < Phantom> last time I checked, SPI wasn't usable reliably for neopixel due to timing issues
2025-11-27T01:06:49 < qyx> neopixel and derivatives are not usable reliably themselves
2025-11-27T01:07:06 < qyx> even the derivatives and clones have a bit different timing
2025-11-27T01:07:17 < qyx> (of the "same" chip)
2025-11-27T01:12:18 < zyp> depending on the chip, uart might be a better fit than spi
2025-11-27T01:12:26 < zyp> that's what I did last time I did adressable leds on stm32
2025-11-27T01:17:41 < ventYl_> qyx: Turris does it. They had to implement gamma correction.
2025-11-27T01:23:01 < qyx> too much code for a led
2025-11-27T01:33:47 < ventYl_> they wasted like 70% of CPU time doing gamma correction :>
2025-11-27T01:49:22 -!- yukam [~yukam@user/yukam] has quit [Ping timeout: 256 seconds]
2025-11-27T01:55:20 < qyx> I already have an offgrid house in the woods
2025-11-27T01:55:28 < qyx> sorry bad window :>
2025-11-27T01:57:33 -!- yukam [~yukam@user/yukam] has joined ##stm32
2025-11-27T02:14:20 < Phantom> … Those wire nuts with a screw on the side... No wonder why they dissapeared. That one was loose... grrr
2025-11-27T02:14:26 < Phantom> wago to the rescue!
2025-11-27T02:47:25 < Phantom> 4 wires + 1 to the outlet... I don't think an outlet box have enough cuin for all that :D
2025-11-27T04:08:06 < mrec_> hm no I looked at the wrong position yesterday. FreeRTOS+LWIP does not work with the STM32H723Z nucleo board (I checked with two boards and even tried to load the example from ST Micro)
2025-11-27T04:09:35 < mrec_> Did anyone try ethernet with the nucleo-h723z board?
2025-11-27T04:58:30 -!- PaulFertser [~paul@paulfertser.info] has quit [Ping timeout: 252 seconds]
2025-11-27T05:18:29 < mrec_> can anyone recommend an STM32 Nucleo board which is confirmed to work with ethernet?
2025-11-27T05:18:57 < mrec_> the STM32H723Z definitely does not the way it is with the latest stm32 cube nor did any examples work for me
2025-11-27T05:30:09 -!- PaulFertser [paul@paulfertser.info] has joined ##stm32
2025-11-27T05:35:40 < mrec_> seems like I found one shaky example which responds and submits some wrong ICMP data, that's a start at least.
2025-11-27T05:44:57 < mrec_> disabling the dcache fixed the problem okay now I can fix that and have a working example
2025-11-27T05:50:10 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-27T05:58:27 < mrec_> 64 bytes from 192.168.10.1: icmp_seq=231 ttl=255 time=1009 ms
2025-11-27T05:58:27 < mrec_> wrong data byte #54 should be 0x36 but was 0x62
2025-11-27T05:58:27 < mrec_> #16	10 11 12 13 14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20 21 22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f 
2025-11-27T05:58:30 < mrec_> #48	30 31 32 33 34 35 62 0 
2025-11-27T05:58:35 < mrec_> that's the dcache problem
2025-11-27T05:58:51 < mrec_> 64 bytes from 192.168.10.1: icmp_seq=4 ttl=255 time=0.385 ms
2025-11-27T05:58:51 < mrec_> 64 bytes from 192.168.10.1: icmp_seq=5 ttl=255 time=0.297 ms
2025-11-27T05:59:01 < mrec_> this is the result when it's fixed
2025-11-27T06:08:36 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-27T07:34:05 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-27T08:02:23 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-27T08:50:23 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-27T08:56:50 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-27T09:03:58 < jpa-> well this is interesting; after 9 years of use, a MLCC shorted and blew a fuse I didn't know I had.. https://jpa.kapsi.fi/stuff/pix/blown_mlcc.jpg
2025-11-27T09:06:29 < qyx> was it stresesd?
2025-11-27T09:07:36 < qyx> oh I see the fuse too
2025-11-27T09:08:50 < jpa-> ebike battery.. so high vibration environment, but on the other hand 1.6 mm PCB mounted to a metal support
2025-11-27T09:09:42 < qyx> missing strain relief? you are biking from -40°C to +40°C
2025-11-27T09:09:54 < qyx> mechanical stress from the bonding?
2025-11-27T09:09:57 < jpa-> what kind of strain relief?
2025-11-27T09:10:03 < qyx> on the pcb mount
2025-11-27T09:10:54 < jpa-> well the screws have about 0.5 mm sideways space in the holes, but other than that, no strain relief
2025-11-27T09:11:13 < jpa-> three screws so it shouldn't be able to bend the PCB
2025-11-27T09:16:56 < Steffanx> From brown out to burn out issues, awesome:D
2025-11-27T09:18:04 < jpa-> i wouldn't be surprised if the two were related
2025-11-27T09:19:33 < jpa-> i think if it hadn't burned that trace, it would have kept operating just fine.. the battery has plenty of power to remove the shorted capacitor
2025-11-27T09:20:02 < jpa-> BMS cuts out at 30A
2025-11-27T09:22:03 -!- PhantomWork3 [~PhantomWo@modemcable022.175-37-24.static.videotron.ca] has joined ##stm32
2025-11-27T09:24:42 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-27T09:25:49 -!- PhantomWork2 [~PhantomWo@user/phantom] has quit [Ping timeout: 264 seconds]
2025-11-27T10:10:47 < BrainDamage> at least replacing the "fuse" won't require tinfoil here
2025-11-27T11:46:34 < mrec_> as soon as I update to the latest bsp ethernet stops to work with the STM32H723zg
2025-11-27T11:51:54 < mrec_> great all fixed now small modifications are required
2025-11-27T11:54:12  * karlp is excited about phantom's "let's use i2c for _everything_ plan" :)
2025-11-27T13:09:04 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-27T13:29:05 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-27T15:05:08 < qyx> Reply #9 on: October 24, 2020
2025-11-27T15:05:12 < qyx> I use mikroPascal compiler and debugger, CodeGrip programmer, STM32F107VE processor.
2025-11-27T15:05:15 < qyx> poor soul
2025-11-27T15:21:36 < qyx> is Aisler joking
2025-11-27T15:21:51 < qyx> they sent a newsletter "oh we are adding nice colors!"
2025-11-27T15:22:27 < qyx> now trying, classic green is 51.20e, other colors are 519e
2025-11-27T15:23:45 -!- PhantomWork3 [~PhantomWo@modemcable022.175-37-24.static.videotron.ca] has quit [Read error: Connection reset by peer]
2025-11-27T16:04:13 < jbo> lol wut
2025-11-27T16:04:31 < jbo> when they sent the "oh we do >4L now" I wanted to check DR but there were none
2025-11-27T16:04:59 < jbo> looks like there are still none.
2025-11-27T16:17:37 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-27T16:23:01 -!- mercenary [~mercenary@user/mercenary] has quit [Ping timeout: 255 seconds]
2025-11-27T16:25:21 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-27T16:37:51 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-27T16:55:12 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-27T16:55:12 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-27T17:27:11 < karlp> fuckin, https://www.sensodroid.com/en/p/wireless-bluetooth-digital-scale-sensodroid-s5000 is using developer svc and char uuids.
2025-11-27T17:27:24 < karlp> fuck btsig, fuck using private 128bit uuids, just ship it with dev mode.
2025-11-27T17:28:04 < qyx> is that bad
2025-11-27T17:28:18 < karlp> just makes a bit of a mockery of everyone else trying to do the "rightthing"
2025-11-27T17:28:19 < tomeaton17> damn g4 doesnt have can bootloader
2025-11-27T17:28:33 < karlp> I implemented teh bluetooth sig "weight scale service"
2025-11-27T17:28:42 < karlp> and... none of the scale apps show my scale as supported
2025-11-27T17:28:48 < qyx> tomeaton17: and.. is that bad
2025-11-27T17:28:54 < karlp> read teh docs and tehy all support like 30 different private specs
2025-11-27T17:29:03 < qyx> karlp: interesting
2025-11-27T17:29:11 < qyx> so whats the sense of standard services then
2025-11-27T17:29:15 < karlp> re-read teh weight scale _profile_ and it said "device info service" is required, add that, still shows up as non-suppoted.
2025-11-27T17:31:25 < karlp> no idea why openscale doesn't show my scale. https://github.com/oliexdev/openScale/blob/master/android_app/app/src/main/java/com/health/openscale/core/bluetooth/scales/StandardWeightProfileHandler.kt#L98-L103 looks like it just needs to have the service.
2025-11-27T17:31:36 < karlp> nrf connect works, it understands it as a standard service.
2025-11-27T17:31:51 < karlp> meh, so much for bonus thursday fun features.
2025-11-27T17:33:19 < tomeaton17> I need the amount of adc channels which the g4 has
2025-11-27T17:33:36 < karlp> tomeaton17: just pick a bigger package?
2025-11-27T17:34:01 < karlp> qyx: to be fair, the WSS bt sig spec is pretty shit.
2025-11-27T17:34:10 < karlp> 5g finest resolution on offer :|
2025-11-27T17:34:21 < karlp> who would wilfully write a spec like that?!
2025-11-27T17:35:01 < karlp> bt itself has all the methods for presenting numbers with descriptors but.... noooooo write a fixed set of number formats into the spec itself.
2025-11-27T17:42:04 < karlp> qyx: theres't not even a "vendor" 16bit range, you can _buy_ one of the 16bit values from 0xfe00 to ffff, and there's a list of them,
2025-11-27T17:42:16 < karlp> but apparently people just squat the high ones, as bt sig is selling frrom the bottom up...
2025-11-27T17:46:20  * karlp pokes an esp-iot-soltion demo app
2025-11-27T17:46:40 < tomeaton17> karlp: i am already using 144, I think only g4 has enough channels 
2025-11-27T17:47:19 < karlp> sounds like fun then :)
2025-11-27T17:47:54 < karlp> back in the day I could get enough external adc cahnnels on L1 in qfp48 thjat would have been qfp64 on f2/f4 but if you're already at qfp144, then sounds... pleasant times :)
2025-11-27T17:48:22 < tomeaton17> yeah the hardware guys are cooking up something interesting for sure
2025-11-27T17:54:13 < karlp> nice. the scale apps I've got don't recognise the esp-iot-solution demo either.
2025-11-27T17:59:45 -!- mercenary [~mercenary@user/mercenary] has joined ##stm32
2025-11-27T18:11:13 < mrec_> does anyone actually use ethernet with an stm32 part?
2025-11-27T18:11:29 < mrec_> reading through the forum is scary ... ST doesn't seem to know what they are doing
2025-11-27T18:12:03 < mrec_> and there's a difference between 'works' for a minute and works reliable
2025-11-27T18:12:09 < srk> yes, but I implemented driver and minimal stack from scratch
2025-11-27T18:12:35 < srk> as the cube generated stuff works exactly like you describe (in some versions of cube :))
2025-11-27T18:13:35 < mrec_> I would like to stick with cube and fix the rest manually, currently I'm at the stage it works but not for long
2025-11-27T18:14:03 < mrec_> yesterday it didn't work at all so at least I'm one step forward today
2025-11-27T18:16:34 < mrec_> https://community.st.com/t5/stm32cubemx-mcus/lwip-v6-5-0-stmh32h735-disco-https-github-com-stm32-hotspot/m-p/91721
2025-11-27T18:16:52 < mrec_> ST employee:  An MPU region (cacheable) will be added to manage RX pool section.
2025-11-27T18:17:08 < mrec_> a user: All of the normal memory is cacheable by default. So what is the purpose for a adding a cacheable region? There isn't - it's just another complete nonsense from you... It actually needs to be a non-cacheable region.
2025-11-27T18:19:25 < srk> eh, h7 peculiarities
2025-11-27T18:19:46 < srk> I'm using f7 where I fortunately don't have to deal with this stuff
2025-11-27T18:25:35 < mrec_> probably no one is buying this chip so they don't care anymore
2025-11-27T18:31:39 < karlp> just use ai to measure weights by taking a photo! :) https://play.google.com/store/apps/details?id=com.gonder.scaleal
2025-11-27T18:39:47 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-27T18:47:36 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-27T18:54:05 < mrec_> First Tx Descriptro Address is the same as Rx Buffers Address in ST's holy network example...
2025-11-27T19:04:06 < mrec_> hah that way St can ensure high compatibility. It just doesn't work among any St h7 parts :D
2025-11-27T19:06:19 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 260 seconds]
2025-11-27T19:38:36 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-27T20:11:35 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 240 seconds]
2025-11-27T21:01:24 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 256 seconds]
2025-11-27T21:02:24 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-27T21:43:20 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-27T21:53:21 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-27T22:47:25 < ColdKeyboard> If you have WiFi antenna on the board and have to use DCDC... is ~1MHz or 2MHz switching frequency better for RF?
2025-11-27T22:47:41 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-27T22:48:15 < ColdKeyboard> Or it does not really matter since it's 1-2MHz and WiFi is 2.4GHz?
2025-11-27T22:54:44 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-27T23:16:45 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-27T23:23:38 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
--- Day changed Fri Nov 28 2025
2025-11-28T00:06:11 < ds2> higher frequencies are easier to filter out but higher frequencies would need lower order harmonics to land in the wifi band (or any other sensitive band depending on the rcvr design)
2025-11-28T00:15:54 -!- NEYi_ [~NEYi@195.234.78.187] has joined ##stm32
2025-11-28T00:19:20 -!- NEYi [~NEYi@195.234.78.187] has quit [Ping timeout: 240 seconds]
2025-11-28T00:25:20 -!- NEYi_ [~NEYi@195.234.78.187] has quit [Read error: Connection reset by peer]
2025-11-28T00:59:55 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-28T01:17:10 -!- qyx [~qyx@84.245.121.36] has quit [Ping timeout: 246 seconds]
2025-11-28T01:19:14 -!- qyx [~qyx@84.245.120.224] has joined ##stm32
2025-11-28T01:56:00 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 244 seconds]
2025-11-28T02:32:16 < qyx> hey jpa- some true nordic sailing for you in case you are struggling https://youtu.be/Cf3Ro6sstQg?si=pdeK6HepnFTYPIzF&t=275
2025-11-28T02:57:26 -!- remcycles [~remcycles@c-24-19-134-216.hsd1.wa.comcast.net] has quit [Ping timeout: 265 seconds]
2025-11-28T03:21:32 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-28T03:39:00 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-28T04:20:35 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-28T04:26:45 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-28T04:30:29 -!- flatmush [~benbrewer@149.57.127.56] has quit [Ping timeout: 244 seconds]
2025-11-28T04:31:08 -!- flatmush [~benbrewer@149.71.24.176] has joined ##stm32
2025-11-28T06:08:40 < Phantom> Quite some progress on my circuit at work yay, all the main parts are in place, now I just need to move a few easy traces, and wire up a small part, then finish to connect the microcontroller, yay
2025-11-28T06:08:57 < Phantom> it's annoying that there is power on all 4 sides
2025-11-28T06:30:06 -!- sugarbeet [~barbas@81.4.123.134] has quit [Ping timeout: 265 seconds]
2025-11-28T06:39:22 -!- sugarbeet [~barbas@81.4.123.134] has joined ##stm32
2025-11-28T07:39:02 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-28T08:00:40 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-28T08:01:00 < ds2> power planes!
2025-11-28T08:02:47 < Phantom> kinda, top layer is a ground plane, but the bottom layer there will be a 3.3V square under the STM32, so I can just via to the multiple pins, and it look not too atrocious
2025-11-28T08:22:20 < jpa-> if you are doing analog stuff, i'd really recommend 4-layer board.. it cuts down on the noise
2025-11-28T08:25:20 < Phantom> the STM32 only have a low precision channel used... I need like 0.5V on 12V signal precision. My tests on the F103 show that it had like 30mV noise I think, with an offset of about -0.3V, which is the input diode, so, close enough :D
2025-11-28T08:25:59 < Phantom> and the rest... well, it will be on a 16 bits low speed ADC on I2C... Should be fine, I'll just sample it a few times probably
2025-11-28T08:26:10 < jpa-> ok yeah, 5 ENOB is not much :)
2025-11-28T08:27:43 < Phantom> it will be better than trying to adjust it with a multimeter anyway
2025-11-28T08:28:36 < Phantom> so, I finally decided to split the I2C stuff into 2 busses, one dedicated to an eeprom, the other dedicated to the digipots+adc
2025-11-28T08:28:46 < Phantom> and... I just realised that I made a copy-paste error
2025-11-28T08:38:11 -!- hexo [~hexo@user/hexo] has quit [Ping timeout: 265 seconds]
2025-11-28T09:08:21 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-28T09:10:27 -!- hexo [~hexo@user/hexo] has joined ##stm32
2025-11-28T09:40:09 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 260 seconds]
2025-11-28T10:05:30 < qyx> hey brainstormers, any wild ideas how to measure 3d displacement (let's say +-20 mm in x/y/z)?
2025-11-28T10:05:42 < qyx> cameras are not feasible
2025-11-28T10:06:15 < qyx> offset only, rotation/angle is not required
2025-11-28T10:06:25 < qyx> imagine measuring two concrete blocks
2025-11-28T10:06:28 < mercenary> laser and ToF chip?
2025-11-28T10:07:02 < qyx> I mean 20 mm is the full scale range, not precision/accuracy/resolution
2025-11-28T10:08:06 < mercenary> which leads to the question, 'what precision do you need?'
2025-11-28T10:09:00 < qyx> < 0.1 mm, accuracy below 1 mm
2025-11-28T10:09:36 < mercenary> and contactless?
2025-11-28T10:09:42 < qyx> so 8 bits, but I would go for 12-16
2025-11-28T10:09:43 < qyx> yeah
2025-11-28T10:10:38 < qyx> my current idea is some sort of 3d arrangement of LVDT transformers on a cube faces with a piece of ferromagnetic core (a sphere) in the middle
2025-11-28T10:13:05 < mercenary> light, sound or field. ultrasound won't get you that precision, so light or field. Which leaves you with laser (ToF or interferometry), or inductive or capacitive sensing. How close to the thing can sensors be?
2025-11-28T10:15:20 < qyx> they are mounted directly on the sensed durface/s, so 0
2025-11-28T10:15:33 < zyp> qyx, what makes cameras infeasible?
2025-11-28T10:15:40 < qyx> zyp: price and dust
2025-11-28T10:16:08 < zyp> wouldn't dust apply to lasers as well?
2025-11-28T10:16:24 < qyx> yeah but a bit less
2025-11-28T10:16:36 < qyx> so capacitive or inductive
2025-11-28T10:16:58 < qyx> or RF interferometry
2025-11-28T10:17:10 < zyp> how about a magnet surrounded by an array of magnetometers?
2025-11-28T10:17:43 < qyx> interesting idea, I like it
2025-11-28T10:18:07 < zyp> we've got a project at work where we're doing single axis contactless position sensing with a magnet and a row of magnetometers
2025-11-28T10:18:42 < qyx> how would you handle multi axes wrt. magnet poles?
2025-11-28T10:18:57 < qyx> wih math?
2025-11-28T10:19:21 < zyp> yeah, meth it out
2025-11-28T10:20:00 < qyx> this approach would work for rotation in addition
2025-11-28T10:21:25 < zyp> it'll probably need some calibration both for the magnet and the installation size
2025-11-28T10:23:26 < qyx> those magnetometers wohld probably be mounted on some sort of hollow tube or a cube with the magnet insideon a stick or so
2025-11-28T10:24:11 < qyx> maybe a springy stick to prevent sensor damage when it goes out of range
2025-11-28T10:24:15 < zyp> hmm, actually, with an array of magnetometers around the magnet, you can probably just average all the measurements and subtract that to remove the earth's magnetic field and other external influences
2025-11-28T10:26:27 < qyx> nah for a small magnet in a <50 mm distance the earth field would be negligble
2025-11-28T10:27:29 < zyp> AIUI it was significant enough in our project that we have to calibrate it out
2025-11-28T10:28:32 < zyp> but we're using single axis magnetometers placed in a line
2025-11-28T10:31:09 < zyp> and to be honest I'm not sure we're measuring the optimal axis for what we're doing :)
2025-11-28T10:33:39 < qyx> were you able to properly liearize the outpit?
2025-11-28T10:34:51 < qyx> hm inductive or capacitive can have simpler sensors (electrodes) but require more electronics
2025-11-28T10:35:06 < qyx> whereas for magnetos I only need some i2c muxes at max
2025-11-28T10:40:16 < jpa-> TOF sensors?
2025-11-28T10:40:41 < jpa-> ah, already suggested
2025-11-28T10:41:42 < jpa-> https://hackaday.io/project/187172-os3m-mouse   this uses inductive sensors, i built one but it won't really handle +-20mm
2025-11-28T10:43:11 < zyp> qyx, I didn't work on that part of it, but IIRC they achieved sub-mm accuracy
2025-11-28T10:43:34 < zyp> I only did some infrastructure code for that project
2025-11-28T10:45:14 < zyp> and yeah, the sensor board on this project was a long board with 16 magnetometers hooked to two 8ch i2c muxes
2025-11-28T10:45:43 < jpa-> the cube arrangement could work with capacitive sensing: have PCBs in cube-corner shape, with sense electrode on one side and shield electrode on opposite (outside); then have a metal cube in the middle and connected with flexible wire to the COM electrode
2025-11-28T10:45:55 < qyx> jpa-: ldcxxxx I am using right now for other project
2025-11-28T10:46:11 < zyp> the driver cheats and enables all channels to broadcast a «start oneshot measurement» to all sensors, and then reads out the results one by one
2025-11-28T10:47:57 < jpa-> with some 3D printed flexure arrangement you could probably measure it all with just magnetic encoder sensors
2025-11-28T10:52:29 -!- goodvibrations32 [~user@user/goodvibrations32] has joined ##stm32
2025-11-28T10:55:34 < qyx> nah I want contactless and it can be in a corrosive env
2025-11-28T10:56:23 < jpa-> air moisture level variations could be a problem for capacitive sensing
2025-11-28T10:57:05 < zyp> speaking of 3d prints, I've been wanting to do some labelling on 3d prints, and the achievable resolution with multicolor prints isn't all that impressive
2025-11-28T10:57:18 < zyp> so I picked up a laser engraver and have been playing with it: https://bin.jvnv.net/file/w6TZZ.jpg
2025-11-28T10:57:30 < jpa-> it is
2025-11-28T10:57:43 < zyp> bottommost is borderline
2025-11-28T10:58:20 < jpa-> does the engraver have some safety features or is it one of those "put it on the table and hope you have your goggles on" thingies?
2025-11-28T10:58:47 < zyp> it's an xtool f2, so fully enclosed
2025-11-28T10:59:05 < jpa-> oh, pricey
2025-11-28T10:59:46 < qyx> is it a fdm print?
2025-11-28T10:59:50 < zyp> they did a preorder discount a few weeks ago
2025-11-28T10:59:58 < zyp> yeah
2025-11-28T11:00:43 < zyp> here's some more photos: https://photos.app.goo.gl/iWdu7XfcuNuAnfVh8
2025-11-28T11:01:59 < zyp> I missed totally with the parameters for the test matrix on the stainless :p
2025-11-28T11:02:50 < jpa-> https://github.com/ocaml/ocaml/pull/14369   haha, such vibe coding
2025-11-28T11:02:51 < zyp> on the last print, the text sizes are respectively 4pt, 3pt and 2pt; the lines to the right of the text are millimeters
2025-11-28T11:09:26 < zyp> jpa-, shit like that makes it very understandable why many projects just adopt a «no AI generated code» policy
2025-11-28T11:13:16 < mawk> what's the difference between USART and SPI? apart from the chip select thing
2025-11-28T11:17:43 < mawk> lol jpa- 
2025-11-28T11:17:44 < mawk> >My work was just directing, shaping, cajoling and reviewing.
2025-11-28T11:17:51 < mawk> even the guy's answer reek of AI
2025-11-28T11:17:54 < zyp> mawk, start/stop bits
2025-11-28T11:18:05 < mawk> he probably wrote them himself but his brain is broken by the AI
2025-11-28T11:18:14 < mawk> I thought those were gone in synchronous mode zyp
2025-11-28T11:18:22 < mawk> but maybe not
2025-11-28T11:18:39 < zyp> usually not, but in some implementations they can be turned off, and then you effectively have SPI
2025-11-28T11:18:40 < jpa-> synchronous USART is a weirdo, but STM32's implementation of it is similar to SPI
2025-11-28T11:19:03 < jpa-> some other manufacturers have totally different way to do synchronous USART
2025-11-28T11:19:53 < zyp> SPI is just bits shifted with a clock, UART is just bits shifted with start/stop bits for framing and oversampling for async reception
2025-11-28T11:20:23 < zyp> sync USART is UART plus clock signal, and if you remove the start/stop bits you're left with SPI
2025-11-28T11:21:08 < zyp> oh, and some implementations of sync USART does start/stop bits on the line, but doesn't clock them
2025-11-28T11:21:30 < zyp> in which case the same bytes can be received by both SPI and UART receivers with no change in signal
2025-11-28T11:22:52 < zyp> it's also not uncommon for microcontrollers to have a single peripheral with both SPI and UART modes (and sometimes I2C)
2025-11-28T11:23:09 < zyp> nrf does that, I think some TI parts does too
2025-11-28T11:24:25 < mawk> right
2025-11-28T11:24:58 < mawk> some weirdly enthusiastic ST presentation is describing synchronous USART mode as plain SPI https://www.st.com/resource/en/product_training/STM32G0-Peripheral-USART-interface-USART.pdf
2025-11-28T11:25:15 < mawk> maybe they intend to replace the SPI peripherals with more copies of USART completely 
2025-11-28T11:26:56 < qyx> yes sync uart is spi if you adhere to the contract that any non-clocked data is ignored and clock doesn't need to be continuous
2025-11-28T11:27:01 < qyx> which is true for spi
2025-11-28T11:27:22 < qyx> but usart peripheral can't do slave
2025-11-28T11:28:06 -!- jbo [~jbo@user/tct] has quit [Quit: ZNC 1.9.1 - https://znc.in]
2025-11-28T11:28:27 -!- jbo [~jbo@user/tct] has joined ##stm32
2025-11-28T11:32:39 < mawk> they say it can do slave on the G0 at least, but it could be a different kind of USART than on the other stm32s
2025-11-28T11:33:01 < mawk> as it handles chip select
2025-11-28T11:40:31 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-28T11:40:31 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-28T11:40:42 -!- travisb_ [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has joined ##stm32
2025-11-28T11:41:17 -!- splud [~noneya.bi@user/splud] has quit [Ping timeout: 244 seconds]
2025-11-28T11:41:23 -!- tabemann [~travisb@172-13-49-137.lightspeed.milwwi.sbcglobal.net] has quit [Read error: Connection reset by peer]
2025-11-28T11:41:45 -!- splud [~noneya.bi@user/splud] has joined ##stm32
2025-11-28T11:50:39 < karlp> zyp: that engrave "is this size readable" thing,what's the colour from? was that two colour 3d print, but the top layer was just black to cover and be exposed by the engraver?
2025-11-28T11:54:39 < zyp> no, it's a fully black print, AIUI the laser makes the plastic foam up or something
2025-11-28T11:54:40 < ventYl_> I am out of office and will get back to you on Monday, 1st of December.
2025-11-28T11:54:41 < ventYl_> Your emails will not be read or forwarded.
2025-11-28T11:54:50 < ventYl_> tell me you are German without telling me that you are German
2025-11-28T11:55:24 < zyp> I hate automated out of office replies
2025-11-28T11:57:44 < ventYl_> better than nothing
2025-11-28T11:58:24 < zyp> most of the time it's just noise from somebody that's CCed on a mail nobody is expecting them to reply to anyway
2025-11-28T11:58:43 < karlp> avr/ti xmc both have combo uart/spi/ periphs.
2025-11-28T11:59:02 < karlp> not ti, infineon xmc
2025-11-28T11:59:47 < karlp> what are your plans for the engraver? that's a nice toy....
2025-11-28T12:00:05 < karlp> doing your bit to keep up .no appearances.... :)
2025-11-28T12:00:19 < zyp> hmm?
2025-11-28T12:10:41 < zyp> I generally don't get a lot of work emails in the first place, so I've never felt the need to get away from them while on vacation, so I glance over anything I get
2025-11-28T12:11:42 < zyp> and if anybody expects a reply from me, I'll either manually tell them I'll look at it when I'm back or just handle it while on vacation, depending on how interesting it is :p
2025-11-28T12:19:14 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-28T12:20:15 -!- sugarbeet [~barbas@81.4.123.134] has quit [Ping timeout: 240 seconds]
2025-11-28T12:21:23 -!- sugarbeet [~barbas@81.4.123.134] has joined ##stm32
2025-11-28T12:37:38 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-28T13:08:44 < Steffanx> You should make that your auto reply. 
2025-11-28T13:21:12 < karlp> lol
2025-11-28T13:47:16 -!- System_Error [~SystemErr@user/systemerror] has quit [Ping timeout: 272 seconds]
2025-11-28T14:06:41 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-28T15:59:06 < tomeaton17> if work wants you to have emails installed on your phone they can provide a phone. especially with this intune bs
2025-11-28T16:18:20 < mawk> tomeaton17: on my phone there is a "work profile" separate from the rest
2025-11-28T16:18:27 < mawk> that's where I installed the intune crap and they can't see the rest
2025-11-28T16:18:41 < mawk> I asked for a work phone but they said no
2025-11-28T16:27:13 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-28T16:46:29 < jpa-> meh, i already fixed this scons problem last week by randomly poking around, and now it is back
2025-11-28T16:46:30 < jpa-> scons: *** Found dependency cycle(s): Internal Error: no cycle found for node ..... (<SCons.Node.FS.File object at 0x224f8280>) in state failed
2025-11-28T16:50:05 < zyp> better switch to erect :p
2025-11-28T16:59:05 < Steffanx> Mawk: If work wanted me to run their intune crap they would have to give me a phone. If not than fuck them.
2025-11-28T17:03:51 < zyp> what's intune?
2025-11-28T17:22:19 < tomeaton17> mawk: but are you required to have work stuff on your phone?
2025-11-28T17:38:51 -!- travisb_ is now known as tabemann
2025-11-28T18:30:37 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 264 seconds]
2025-11-28T19:07:22 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:8b9:7ebe:aac3:5190] has joined ##stm32
2025-11-28T19:25:35 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 240 seconds]
2025-11-28T19:40:23 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:8b9:7ebe:aac3:5190] has quit [Remote host closed the connection]
2025-11-28T19:41:12 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:8b9:7ebe:aac3:5190] has joined ##stm32
2025-11-28T19:47:18 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-28T20:12:01 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-28T20:34:11 < karlp> mawk: and how do these "profiles" work?
2025-11-28T20:34:39 < mawk> it separates the application and the data
2025-11-28T20:34:42 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-28T20:34:49 < mawk> so they can't see what you have installed or your files
2025-11-28T20:35:07 < mawk> they can check stuff like which wifi you're connected to, the security level of your phone
2025-11-28T20:35:48 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:8b9:7ebe:aac3:5190] has quit [Ping timeout: 260 seconds]
2025-11-28T20:36:10 < mawk> it's like a different user with its own set of a applications and files
2025-11-28T20:36:55 < mawk> no tomeaton17 but since I'm always late I need to have teams on my phone to attend standup meetings
2025-11-28T20:55:08 < Steffanx> What phone and OS does Mr mawk run?
2025-11-28T21:31:49 < mawk> pixel 9 pro
2025-11-28T21:32:06 < zyp> does it fold?
2025-11-28T21:34:26 < mawk> no
2025-11-28T21:34:33 < mawk> folding phones are stupid
2025-11-28T21:38:04 < zyp> no, it's great
2025-11-28T21:43:49 < mawk> why
2025-11-28T21:44:00 < mawk> it's fragile and mega expensive and just a gimmick
2025-11-28T21:45:23 < zyp> I'll agree with your second point, decent phones in general are ridiculously expensive nowadays
2025-11-28T21:45:53 < zyp> but I got a pretty nice deal on mine
2025-11-28T21:46:11 < zyp> and fragility doesn't seem to be an issue
2025-11-28T21:47:04 < zyp> but I'm still curious to see if it'll hold up for >5 years like my previous phone
2025-11-28T21:47:50 < zyp> and I also find it genuinely useful
2025-11-28T21:49:28 < ventYl_> my primary concern is size, i switched to smartphone only around 2018 and don't like the idea of carrying large paddle with me
2025-11-28T21:49:40 < zyp> that's the whole point
2025-11-28T21:49:45 < ventYl_> yet this is less of a concern now that phone screens are basically edge to edge
2025-11-28T21:50:26 < zyp> I've been considering getting a tablet for a while, but I always figured I wouldn't bother actually bringing it with me much
2025-11-28T21:51:20 < zyp> foldable phone gives me the convenience of a larger screen whenever I need it, and then it just folds up and goes in my pocket like any other phone would
2025-11-28T22:16:36 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:2432:ca90:632b:fbd6] has joined ##stm32
2025-11-28T22:26:57 < Steffanx> With grapheneos mawk?
2025-11-28T22:28:26 < ventYl_> I need lighter and less ancient notebook and then the need for larger phone screen will essentially cease to exist
2025-11-28T22:45:00 < englishman> what do you find it useful for zyp, and is that with stylus or without
2025-11-28T23:01:33 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-28T23:43:15 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-28T23:47:09 -!- goodvibrations32 [~user@user/goodvibrations32] has quit [Remote host closed the connection]
--- Day changed Sat Nov 29 2025
2025-11-29T00:38:13 < karlp> mawk: I understood the concept, I jut hadn't seen it before in android
2025-11-29T01:13:33 -!- akaWolf [~akaWolf@akawolf.org] has quit [Quit: Lost terminal]
2025-11-29T01:22:54 -!- akaWolf [~akaWolf@akawolf.org] has joined ##stm32
2025-11-29T01:24:18 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Quit: Bye]
2025-11-29T01:25:12 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-29T01:25:27 -!- yukam [~yukam@user/yukam] has quit [Ping timeout: 252 seconds]
2025-11-29T01:33:56 -!- yukam [~yukam@user/yukam] has joined ##stm32
2025-11-29T03:07:50 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 245 seconds]
2025-11-29T04:41:13 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-29T04:47:38 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-29T05:45:48 -!- remcycles [~remcycles@c-24-19-134-216.hsd1.wa.comcast.net] has joined ##stm32
2025-11-29T06:54:27 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-29T09:29:48 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-29T09:41:51 -!- CygniX [~CygniX@user/CygniX] has quit [Excess Flood]
2025-11-29T09:42:09 -!- CygniX [~CygniX@user/CygniX] has joined ##stm32
2025-11-29T11:43:15 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-29T11:43:28 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T11:50:13 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-29T12:11:29 < Steffanx> I'm still not sure what android mawk is running.. maybe I missed it
2025-11-29T12:13:57 < qyx> android 2.3
2025-11-29T12:55:41 < englishman> "vibe coding, where 2 engineers can now create the tech debt of at least 50 engineers”
2025-11-29T16:10:36 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-29T16:41:12 < qyx> regexps on embedded y/n?
2025-11-29T16:41:43 < qyx> I am going to use regex to filter log sources (together with the log level) to display logs
2025-11-29T16:41:59 < qyx> this looks usable https://github.com/kokke/tiny-regex-c/tree/master
2025-11-29T16:42:08 < qyx> although a bit ram hungry
2025-11-29T16:42:31 < qyx> 304 bytes
2025-11-29T16:44:05 < qyx> qyx@qyx5:~/git/tiny-regex-c$ arm-none-eabi-gcc -c re.c -mthumb -Os
2025-11-29T16:44:11 < qyx>    text	   data	    bss	    dec	    hex	filename
2025-11-29T16:44:11 < qyx>    1546	      0	    280	   1826	    722	re.o
2025-11-29T16:47:25 < qyx> remove re_print and it is 1184 bytes
2025-11-29T17:05:15 < ventYl_> regex is just a state machine
2025-11-29T17:05:35 < ventYl_> I'd expect that it is possible to compile a regex into declarative state machine or a piece of C code
2025-11-29T17:05:46 < ventYl_> and that must be tiny
2025-11-29T17:06:53 < qyx> I hope there are no static variables and it is thread safe
2025-11-29T17:06:57 < qyx> lazy to check
2025-11-29T17:07:14 < qyx> because re_compile returns a pointer
2025-11-29T17:08:30 < qyx> of course it is a static array
2025-11-29T17:08:43 < qyx> who the hell even writes a library like that
2025-11-29T17:12:05 < ventYl_> that shouldn't be a problem as it is constant input
2025-11-29T17:12:16 < ventYl_> regex evaluation only holds the current state
2025-11-29T17:12:23 < qyx> it definitely is
2025-11-29T17:12:30 < ventYl_> and cursor into input
2025-11-29T17:12:34 < qyx> re_compile compiles your constant regex string into a array
2025-11-29T17:12:47 < qyx> and then you use re_match or whatever to match your string agains that compiled array
2025-11-29T17:12:53 < qyx> and that *array* itself is static
2025-11-29T17:13:05 < qyx> so you can only use one single regex in your firmware
2025-11-29T17:13:08 < qyx> at te same time
2025-11-29T17:13:19 < qyx> https://github.com/kokke/tiny-regex-c/issues/3
2025-11-29T17:33:43 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 240 seconds]
2025-11-29T17:42:46 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T17:51:33 < mercenary> the power of regex has sufficient magic that you only need one. "(regex1)|(regex2)|(regex3)" and use whichever matchgroup you need
2025-11-29T17:52:05 < mercenary> efficiency? What efficiency?
2025-11-29T17:56:09 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 250 seconds]
2025-11-29T17:56:26 < ventYl_> regex is actually quite efficient
2025-11-29T18:00:38 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T18:04:18 < mercenary> it is, in most cases. unless you hit some strange backtracking case (looks at cloudflare)
2025-11-29T18:14:09 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 252 seconds]
2025-11-29T19:02:43 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-29T19:10:10 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-29T19:21:11 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-29T19:50:00 < mawk> no Steffanx it's regular android
2025-11-29T19:50:34 < mawk> zyp maybe the latest foldable phones improved the fragility 
2025-11-29T19:50:43 < mawk> but when I looked into it a while ago it was definitely very fragile
2025-11-29T19:51:00 < mawk> like breaking the screen if a little dust or sand got into the hinges kind of fragilr
2025-11-29T19:51:22 < mawk> or flexing it backwards a bit too hard and breaking 
2025-11-29T19:52:11 < zyp> idk, I'm not a gorilla, doesn't seem to be an issue for me
2025-11-29T19:52:40 < mawk> if you leave it unfolded it was easy to sit on it or have the cat sit on it or whatever
2025-11-29T19:52:52 < mawk> and even non-gorillas can live in an environment with dust and sand
2025-11-29T19:53:31 < mawk> and the inner flexible screen not being scratchproof at all also
2025-11-29T20:02:46 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-29T20:02:46 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-29T20:16:09 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T20:20:08 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-29T20:22:55 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 240 seconds]
2025-11-29T20:24:20 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T20:24:34 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-29T20:28:55 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 240 seconds]
2025-11-29T20:34:05 -!- System_Error [~SystemErr@user/systemerror] has quit [Remote host closed the connection]
2025-11-29T20:39:54 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T20:41:06 -!- System_Error [~SystemErr@user/systemerror] has joined ##stm32
2025-11-29T20:57:03 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 265 seconds]
2025-11-29T20:58:36 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T21:05:48 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 260 seconds]
2025-11-29T21:10:55 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T21:15:45 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 245 seconds]
2025-11-29T21:21:51 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T21:33:16 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-29T21:37:55 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 240 seconds]
2025-11-29T22:43:40 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 245 seconds]
2025-11-29T22:47:47 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T23:15:30 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 256 seconds]
2025-11-29T23:19:27 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has joined ##stm32
2025-11-29T23:23:54 -!- IanW_ [~IceChat9@arcanum.force9.co.uk] has quit [Ping timeout: 244 seconds]
2025-11-29T23:26:10 -!- IanW_ [~IceChat9@31.94.16.231] has joined ##stm32
2025-11-29T23:29:29 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-29T23:33:19 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 264 seconds]
--- Day changed Sun Nov 30 2025
2025-11-30T00:20:13 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 244 seconds]
2025-11-30T00:46:15 -!- IanW_ [~IceChat9@31.94.16.231] has quit [Quit: Bye]
2025-11-30T00:57:16 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 265 seconds]
2025-11-30T01:18:07 -!- qyx [~qyx@84.245.120.224] has quit [Ping timeout: 240 seconds]
2025-11-30T01:20:24 -!- qyx [~qyx@84.245.121.29] has joined ##stm32
2025-11-30T01:41:06 < mawk> I'm thinking about implementing profiling and coverage into the integration tests
2025-11-30T01:41:22 < mawk> then write the profile data to the SD card
2025-11-30T01:41:35 < mawk> I could even get GCC to write directly to the SD card all the data without having to have it in RAM, in case it's too big
2025-11-30T01:41:51 < mawk> using memory mapped I/O
2025-11-30T02:17:03 < qyx> why
2025-11-30T03:01:15 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-30T03:03:15 < mawk> to get coverage info
2025-11-30T03:03:20 < mawk> and profiling
2025-11-30T03:03:27 < mawk> obviously
2025-11-30T05:06:49 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-30T05:25:17 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-30T06:04:24 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:2432:ca90:632b:fbd6] has quit [Remote host closed the connection]
2025-11-30T06:14:59 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:f4dc:605d:3f6c:a6f4] has joined ##stm32
2025-11-30T06:24:24 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:f4dc:605d:3f6c:a6f4] has quit [Remote host closed the connection]
2025-11-30T06:27:29 -!- jfsimon [~jfsimon19@2a01:cb14:b9b:2000:f4dc:605d:3f6c:a6f4] has joined ##stm32
2025-11-30T07:11:05 -!- haritz [~hrtz@2a01:4b00:bc2e:7000:d5af:a266:ca31:5ef8] has joined ##stm32
2025-11-30T07:11:05 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-30T07:34:23 -!- nerozero [~nerozero@87.253.63.54] has joined ##stm32
2025-11-30T07:37:52 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-30T09:47:37 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-30T09:57:23 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-30T10:08:43 -!- haritz [~hrtz@user/haritz] has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
2025-11-30T10:11:00 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-30T10:38:36 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-30T12:03:37 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 264 seconds]
2025-11-30T12:20:39 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-30T13:11:48 -!- NEYi [~NEYi@195.234.78.187] has joined ##stm32
2025-11-30T13:39:37 -!- martinmoene [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-30T14:45:55 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has quit [Ping timeout: 264 seconds]
2025-11-30T14:46:34 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-30T16:55:50 -!- NEYi [~NEYi@195.234.78.187] has quit [Quit: Leaving]
2025-11-30T16:59:34 -!- Livio [~livio@user/livio] has joined ##stm32
2025-11-30T17:36:35 -!- haritz [~hrtz@140.228.70.141] has joined ##stm32
2025-11-30T17:36:35 -!- haritz [~hrtz@user/haritz] has changed host
2025-11-30T18:58:00 -!- Livio [~livio@user/livio] has quit [Ping timeout: 272 seconds]
2025-11-30T19:09:50 -!- Linux_Kerio [~Linux_Ker@chello085216213137.chello.sk] has joined ##stm32
2025-11-30T19:11:44 -!- infisc [uid692580@user/infisc] has joined ##stm32
2025-11-30T20:29:19 -!- nerozero [~nerozero@87.253.63.54] has quit [Ping timeout: 240 seconds]
2025-11-30T20:43:08 -!- hexo__ [~hexo@user/hexo] has joined ##stm32
2025-11-30T20:43:21 -!- nerozero [~nerozero@178.134.84.110] has joined ##stm32
2025-11-30T21:07:10 -!- IanW_ [~IceChat9@31.94.26.31] has joined ##stm32
2025-11-30T21:20:06 -!- leptonix [~leptonix@tiger.leptonix.net] has quit [Quit: leaving]
2025-11-30T21:23:11 -!- martinmoene__ [~Martin@205-20-132-5.ftth.glasoperator.nl] has joined ##stm32
2025-11-30T21:24:44 -!- nerozero [~nerozero@178.134.84.110] has quit [Ping timeout: 260 seconds]
2025-11-30T21:24:58 -!- remcycles_ [~remcycles@c-24-19-134-216.hsd1.wa.comcast.net] has joined ##stm32
2025-11-30T21:25:55 -!- sugarbeet [~barbas@81.4.123.134] has quit [Ping timeout: 240 seconds]
2025-11-30T21:26:15 -!- martinmoene [~Martin@205-20-132-5.ftth.glasoperator.nl] has quit [Ping timeout: 240 seconds]
2025-11-30T21:26:18 -!- sugarbeet [~barbas@81.4.123.134] has joined ##stm32
2025-11-30T21:27:15 -!- remcycles [~remcycles@c-24-19-134-216.hsd1.wa.comcast.net] has quit [Ping timeout: 240 seconds]
2025-11-30T21:31:28 -!- infisc [uid692580@user/infisc] has quit [Quit: Connection closed for inactivity]
2025-11-30T22:06:50 -!- hexo__ [~hexo@user/hexo] has quit [Ping timeout: 256 seconds]
2025-11-30T22:58:55 -!- remcycles_ is now known as remcycles
2025-11-30T23:48:17 -!- das_ventZl [~das_ventZ@ip-46.34.253.27.o2inet.sk] has joined ##stm32
--- Log closed Mon Dec 01 00:00:25 2025