**** BEGIN LOGGING AT Mon Oct 29 03:00:01 2018 Oct 29 03:00:33 nothing in amixer to enable or disable it Oct 29 03:01:16 can you simply measure whether the micbias is present Oct 29 03:01:36 yeah, I'll just check that now and also check what voltages are going in Oct 29 03:01:58 hmm, analog.com wiki claims the micbias voltage can be set via DT Oct 29 03:02:03 but I see no evidence of this in the driver in 4.14 Oct 29 03:02:46 the driver is quite a few files Oct 29 03:03:27 not really, everything of relevance is in adau1977.c Oct 29 03:04:00 adau1977.h is just a few definitions and adau1977-i2c.c is a miniscule wrapper Oct 29 03:04:27 not sure where that bit is then :) Oct 29 03:04:45 their default micbias is 8.5 V Oct 29 03:05:54 yeah, nothing in the latest mainline kernel either Oct 29 03:12:19 doesn't look that's on those pins Oct 29 03:13:17 while recording with arecord? Oct 29 03:16:29 hmm, the widgets and routes might actually be needed to get dapm to turn things on... this is a problem I've never encountered, but my codec driver doesn't support any sort of power mangement, the adau1977 driver does Oct 29 03:18:20 also => https://godandme.wordpress.com/2013/10/21/debugging-dapm-for-asoc-codec/ Oct 29 03:18:41 (except debugfs is of course already mounted at /sys/kernel/debug ) Oct 29 03:19:40 just checked again to see if it was different while running arecord, which I should have thought of :) Oct 29 03:19:50 It's still the same, voltages look sensible. Oct 29 03:20:04 i.e no bias. Oct 29 03:20:16 how is that "sensible" ? Oct 29 03:20:32 oh you mean you don't want micbias? Oct 29 03:20:48 well, I mean I don't want any bias. Oct 29 03:21:13 but of course if it's not enabling stuff then that might be related Oct 29 03:21:35 but also my voltages don't look like they clip no always or anything Oct 29 03:22:44 that debugfs stuff looks really useful though! Oct 29 03:22:45 thanks Oct 29 03:26:59 ah, micbias is merely an output and it's up to the hardware design to use it or not Oct 29 03:27:28 anyway, my guess would be: it's either a hardware problem or a software problem Oct 29 03:27:31 good luck with debugging! ;) Oct 29 03:28:54 I'd still try adding the stuff from https://pastebin.com/raw/yhRZ6HJ1 just to see if it makes any difference Oct 29 03:29:00 *still suggest Oct 29 03:48:01 looks handy, stuff looks okay on my scope... Oct 29 04:03:57 i'm pretty sure the problem will be this format specifier... Oct 29 04:06:14 "this format specifier" ? Oct 29 04:08:37 simple-audio-card,format = "dsp_a"; Oct 29 04:09:50 mmm though the supported DAI formats say both DSP_A and DSP_B are supported. Oct 29 04:11:34 yeah, it's a weird choice, but it looks like the codec and its driver support it Oct 29 04:11:52 both the codec and mcasp are quite flexible w.r.t. format used Oct 29 04:12:31 yeah, my only thought was that given anything apart from zero was full range then maybe it was reading bits the wrong way Oct 29 04:13:02 maybe record some data in raw format and hexdump it to see if there's a clue in that Oct 29 04:13:47 It was my understanding that the Windows driver issue had been resolved. Is that correct? Oct 29 04:14:23 that's my understand as well Oct 29 04:14:36 with the latest image, the beaglebone should work plug&play Oct 29 04:14:45 without driver installation Oct 29 04:16:25 Hmmm, I reformatted the SD card, Flashed a new image. The driver files still have the same date of 5-17-2018 3:13 PM Oct 29 04:17:00 the change was made in the usb device descriptors Oct 29 04:17:10 to make windows automatically use its own driver Oct 29 04:17:28 (like it used to before microsoft made some random changes) Oct 29 04:17:54 So should the driver files even be in the image? (if the do not help) Oct 29 04:18:01 mmm... there are some non-full scale or zero values in here, with hex edit. Just not much in between on and off.. Either very loud or or off. Oct 29 04:18:50 So should the driver files even be in the image? (if they do not help) Oct 29 04:20:04 RD: I'd say no, but maybe it's for some very old windows versions? Oct 29 04:20:05 dunno Oct 29 04:20:24 I use linux, so I'm used to it simply working without hassle :) Oct 29 04:23:13 Yeah. I heard about dual booting....... not sure where to begin. (My day job as a manufacturing engineer requires ERP systems and Solidworks .......Windows. Oct 29 04:23:45 or connect the beaglebone to an ethernet network instead of using usb networking Oct 29 04:26:48 RD: I'm using linux for windows subsystem on windows 10 and things work as far as compiling the kernel etc goes Oct 29 04:27:15 though make with j1 works but j > 1 does not. Oct 29 04:27:32 which means builds are 10 ish times slower. Oct 29 04:29:07 what goes wrong with -j >1 ? Oct 29 04:29:32 zmatt this is a Beaglebone blue. I do not see a CAT5 connection for ethernet. Oct 29 04:29:41 RD: oh Oct 29 04:29:46 RD: the blue has wifi Oct 29 04:29:51 Yes Oct 29 04:30:20 so in that case, connect to a wifi network and use that to reach the beaglebone Oct 29 04:30:58 suprothunderbolt: which windows build are you using? Oct 29 04:33:38 poking around for the Beaglebone password Oct 29 04:37:07 suprothunderbolt: if you're getting "Invalid argument" errors during compilation, it's a known bug that has been fixed nearly a year ago (build 16299.98 or later, or Insider build 17035 or later) Oct 29 04:38:05 I think I am connected Oct 29 04:43:03 zmatt: it doesn't find some of the includes if I use j8... Oct 29 04:43:14 it works fine on -j1 Oct 29 04:43:36 I'm using windows 10 on automatic update, I'm not sure what build... Oct 29 04:43:59 doesn't find? as in you specifically get "No such file or directory" errors? Oct 29 04:44:02 17134.345 Oct 29 04:44:19 yeah, no such file or dir Oct 29 04:46:15 weird Oct 29 04:47:25 is that in a volfs mount (e.g. somewhere in /home) or a drvfs mount (/mnt/c) Oct 29 04:48:07 since the latter is more prone to weirdness than the former Oct 29 04:49:28 in /mnt/c Oct 29 04:49:40 you might want to try your home directory instead Oct 29 04:49:54 ahh. didnt think of that Oct 29 04:51:18 of course, if that fixes it then you may want to file a bug report at https://github.com/Microsoft/WSL/issues Oct 29 05:08:27 Thank you zmatt Oct 29 07:16:45 got the audio codec working now by setting left_j mode. Only issue is the sample rate is now wrong. I.e. if I record something I need to slow it down by half to get it sounding right... Oct 29 07:27:52 o.O Oct 29 07:28:21 ;) Oct 29 07:29:01 so the driver is somehow misconfiguring the codec Oct 29 07:29:19 I glanced at the relevant code in the driver, I didn't check it in any detail Oct 29 07:29:21 yep Oct 29 07:29:24 what does it look like on the scope? Oct 29 07:29:45 I haven't looked at the i2s / tdm channels Oct 29 07:29:59 just the audio channels which all look correct Oct 29 07:30:11 well the samplerate would be the interesting thing in this case Oct 29 07:30:28 what about dsp_b ? Oct 29 07:30:33 that's the format I typically use Oct 29 07:30:48 if I record them and slow them down by half it sounds correct Oct 29 07:30:59 i tried dsp_a and dsp_b Oct 29 07:31:09 yeah so you're getting the wrong samplerate Oct 29 07:31:24 both just give me 100% full range values. Oct 29 07:31:37 though behaviour is differet for the different sample formats Oct 29 07:31:47 hm Oct 29 07:32:09 S32_LE is what works with left_j Oct 29 07:32:26 though 16 does too Oct 29 07:32:31 24 doesn't work with anything Oct 29 07:33:10 and S32_LE doesn't work with DSP_A or B at all, S16_LE with either is full range is the input is non-zero Oct 29 07:33:15 the driver should reject values that can't be supported with given master clock frequency though Oct 29 07:33:16 and zero otherwise Oct 29 07:33:26 okay so it sounds like the driver is pretty borked Oct 29 07:33:27 S32_LE is always full value on. Oct 29 07:33:46 these are all the same requested sample rate Oct 29 07:34:23 using the hw:0,0 to get around any alsa conversions with plughw Oct 29 07:34:45 yes, using hw: is a good idea when testing this sort of stuff Oct 29 07:35:22 you're requesting 48 kHz sample rate I presume? Oct 29 07:35:28 yes Oct 29 07:36:51 24 bit frame width definitely can't be supported with a 24.576 MHz master clock and 48 kHz samplerate Oct 29 07:37:00 so the driver should outright reject that Oct 29 07:38:19 I'm testing with arecord and it doesn't Oct 29 07:38:42 arecord -vvv -c1 -f S32_LE -r 48000 -D hw:0,0 /dev/null Oct 29 07:39:04 that's not 24-bit, that's 32-bit Oct 29 07:39:14 yeah, that's the one that works Oct 29 07:39:15 24-bit is S24_3LE or something like that Oct 29 07:39:29 S24_LE :) Oct 29 07:39:33 or wait, S24_LE too sorry Oct 29 07:39:56 format is a weird combination of on-wire format and in-memory format :/ Oct 29 07:41:37 i attempted adding dai-tdm-slot-num and I couldn't get that to even create the driver Oct 29 07:44:09 you're using a 4.14 kernel right? Oct 29 07:45:25 yes Oct 29 07:45:59 4.14.71 Oct 29 07:46:25 bone not TI Oct 29 07:58:21 mmm... I wonder if the codec gets the clock from different places depending on the mode? Oct 29 07:59:15 The ADAU1978 has a built -in analog PLL to provide a jitter -free master clock to the internal ADC. The PLL must be programmed for the appropriate input clock frequency. Oct 29 08:00:49 hah, 24 bit slot width is not supported at all when the codec is in master mode Oct 29 08:01:08 yeah, noticed that too Oct 29 08:01:37 so that explains the crazy results when you try 24 bit Oct 29 08:01:53 this is a bit confusing though Oct 29 08:02:28 like, why does BCLKRATE exist as a separate thing from SLOT_WIDTH Oct 29 08:03:46 I'm surprised that the config i'm using doesn't specify slot_width anywhere Oct 29 08:03:51 I feel it should... Oct 29 08:04:04 &mcasp has a property for it doesn't it? Oct 29 08:04:07 &mcasp0 Oct 29 08:04:17 don't think so Oct 29 08:04:25 if it does i'm not using it Oct 29 08:04:38 no tdm-slots = <4>; ? I thought I saw that Oct 29 08:04:49 yeah, but that's the number of slots Oct 29 08:04:56 not how many bits they are Oct 29 08:04:57 oh sorry, slot-width Oct 29 08:05:02 yeah Oct 29 08:05:04 doh, reading is hard apparently Oct 29 08:05:10 all good :) Oct 29 08:05:46 well there's no need to configure that in the DT since that's normally negotiated with both drivers (cpu and codec) to figure out what is supported Oct 29 08:05:58 at least that's how it's supposed to work :P Oct 29 08:06:03 :) Oct 29 08:06:20 I feel more comfortable sometimes with me telling it what works :) Oct 29 08:07:21 though I am wondering if this clock definition is correct too. Oct 29 08:07:39 hmm? Oct 29 08:09:12 zmatt: do you maybe have a pointer or docs on the pSLC mode that can be used on the BBB? Oct 29 08:09:59 LetoThe2nd: I have a patched version of mmc-utils with a command that performs the reconfiguration (with lot of checking) Oct 29 08:10:21 LetoThe2nd: ultimately the reference is the eMMC spec from JEDEC, which is not the clearest of documents Oct 29 08:10:23 just not sure from this that the MCLK is generated by the BB and then the PLL is locked to that to generate the bit clock] Oct 29 08:11:05 suprothunderbolt: the system clock definition looks correct to me Oct 29 08:11:21 okay Oct 29 08:11:45 so clock is right but width or format is wrong? Oct 29 08:11:53 that would be my guess. Oct 29 08:11:57 or just that the alsa layer is reporting it wrong Oct 29 08:11:58 you could try dumping the i2c registers Oct 29 08:11:59 zmatt: hmkay... ist that available somewhere? Oct 29 08:12:02 while recording audio Oct 29 08:13:07 LetoThe2nd: https://www.jedec.org/sites/default/files/docs/JESD84-B51.pdf (free registration required) Oct 29 08:14:05 zmatt: I'll attempt that tomorrow! Pretty close to getting this all working :) Time to run to the supermarket before it closes though... Oct 29 08:16:17 LetoThe2nd: since performing the steps manually with mainline mmc-utils is tricky and error-prone, and mistakes are a bad idea with one-time-programmable configuration, I added a command that performs sanity checks and then reconfigures eMMC into SLC mode with reliable-writes enabled. See the last commit of branch 'custom' of https://github.com/dutchanddutch/mmc-utils Oct 29 08:25:49 zmatt: great, thanks for sharing! (late reply because of emergency coffee refill!) Oct 29 09:53:29 zmatt: small follow-up question: is the patched mtd-utils micron-specific, or hsould the be emmc-generic? Oct 29 09:54:17 I don't think I've ever done it on a micron eMMC since the beaglebone has been using Kingston since ages, but it should be pretty generic Oct 29 09:54:48 the emmc spec is a bit vague about it still, since you're configuring the eMMC to be "enhanced" without clear indication of what that means, but in practice it means pSLC mode Oct 29 09:54:58 zmatt: i have one right in front of me with micron Oct 29 09:55:03 but yeah, still 2G Oct 29 09:55:13 as a sanity check, my code verifies that the maximum "enhanced" size is exactly half of the eMMC's capacity Oct 29 09:55:43 since reconfiguring MLC into SLC mode cuts capacity in half Oct 29 09:56:09 yeah.... well i'm musing SDINBDG4-8G Oct 29 09:56:19 for some new design, hence the investigations Oct 29 09:58:58 if in doubt you could ask the manufacturer Oct 29 09:59:20 I can probably also tell from an "./mmc extcsd read /dev/mmcblk1" dump if you have one attached Oct 29 10:00:55 heh nah fuess you got that wrong. we are a board manufacturer, and as i knew that it exists and you use it for the bbb, i think about reusing it for maybe some other thing Oct 29 10:01:02 s/fuess/guess/ Oct 29 10:01:47 @LetoThe2nd: lol, sed Oct 29 10:02:03 ? guess I got what wrong? Oct 29 10:02:51 zmatt: ah or sorry, i mixed up two posts. i read it as "the manufacturer can tell you from an mmc extcsd read" Oct 29 10:03:07 lol what Oct 29 10:03:15 the manufacturer doesn't need an mmc extcsd read Oct 29 10:03:19 I'd hope Oct 29 10:03:21 haha Oct 29 10:03:22 which would be myself. but you meant to ask the nand manufacturer, right? Oct 29 10:03:34 I mean the eMMC manufacturer Oct 29 10:03:45 can tell you what "enhanced" means for their eMMC Oct 29 10:03:50 yeah that way it makes sense Oct 29 10:04:43 or, if you have one attached and dump an "extcsd read", I can tell you whether it looks to me like the command I made should work Oct 29 10:05:03 though I'd basically just be doing exactly the same sanity checks that my command already does too Oct 29 10:05:40 ok, gotcha. nah no hardware nearby yet. more like crating a bom at the moment Oct 29 10:07:53 and is the BBBs DDR3 really just connected x16? looking at the schematic Oct 29 10:08:07 a single x16 chip yep Oct 29 10:08:57 was thinking about that too but i fear the performance hit Oct 29 10:09:25 the beaglebone enhanced uses two x8 chips 512MB each since (single-rank) 1GB DDR3 chips are too obscure/expensive Oct 29 10:09:47 would have expected a x32 connection Oct 29 10:09:57 the am335x doesn't support wider than x16 Oct 29 10:10:04 ah ok Oct 29 10:11:35 i want know Oct 29 10:11:42 hi Oct 29 10:12:11 i want to know installing gnuplot on beaglebone back Oct 29 10:12:43 but we don't use sudo command on beaglebone black Oct 29 10:12:46 same as on any other debian system Oct 29 10:12:52 eh yes you do Oct 29 10:13:38 i don't understand exactly Oct 29 10:14:00 unfortunately I have to go, time to head to work Oct 29 10:14:08 maybe someone else will help Oct 29 10:14:16 okay Oct 29 10:14:23 thank u Oct 29 10:14:25 so much Oct 29 10:15:07 @beagleboneblack Oct 29 10:16:29 answer please~ Oct 29 10:16:40 i want to know installing gnuplot on beaglebone back Oct 29 10:16:45 but we don't use sudo command on beaglebone black Oct 29 10:16:45 anwlstkd: it works just like on any other debian system Oct 29 10:16:50 apt-get install gnuplot Oct 29 10:16:55 is there somthing else?? Oct 29 10:17:33 sh : apt-get : command not found Oct 29 10:17:40 we got this error Oct 29 10:17:59 and you are conencted thorugh ssh? Oct 29 10:18:25 yes we already did Oct 29 10:18:58 then i am relatively sure that this should work Oct 29 10:19:18 and maybe our version is Amstrong Oct 29 10:19:24 not debian Oct 29 10:19:37 then make sure that your version is debian Oct 29 10:19:53 just for fun, throw sudo in....and paste the error if it happens. Oct 29 10:20:55 then... we can't use Amstrong version? Oct 29 10:21:09 just debian? Oct 29 10:21:26 anwlstkd: the angstrom images have been unsupported for several years. Oct 29 10:25:10 thank you Oct 29 12:45:49 Hello! I want to ask about the SoftAP0 mode of BBB Green Wireless. I restore it with backup image of Beaglebone Black Wireless. So kindly help to enable it's SoftAP0 mode. Thank You for any kind response Oct 29 14:51:17 m Oct 29 14:54:03 good morning, and happy integral day! ∫ Oct 29 14:58:17 * tbr is not well integrated Oct 29 22:35:05 Today when I rebooted BBx15 I got the gray-greenish screen with the authentication menu (greenish-gray screen with large @ sign at the left). It did not accept the default debian/temppwd, but accepted root/root credentials. Then it loaded some strange window environment with the greenish bird. I guess it is LXQt. Oct 29 22:35:50 The questions: 1) why did it happen; 2) how can I get back to the previous simple windows environment? Oct 29 22:36:08 o.O Oct 29 22:36:14 I am running the latest Debian image Oct 29 22:37:13 well, what did you do prior to rebooting? :P Oct 29 22:39:17 both the fact it didn't accept debian/temppwd but did accept root/root (which is normally not a valid login) are equally weird Oct 29 22:40:32 can you still log in via ssh or otherwise open up a terminal? Oct 29 22:45:14 dreamhiker: hello? Oct 29 22:46:45 Sorry. Oct 29 22:47:56 Confused. Login root with password root is valid and is described in the manual Oct 29 22:48:39 hmm? okay I could be wrong about that, I thought the root account normally didn't have a password set Oct 29 22:50:00 so the obvious question is: what did you do? did you install something, change some configuration? Oct 29 22:50:31 did nothing at the best of my knowledge. Oct 29 22:50:56 can you still log in via ssh? Oct 29 22:51:06 Moreover, even after reinstalling the image on the next reboot it gets into this greenish lxqt Oct 29 22:52:01 I have no idea what you're describing or what you're expecting instead, I run use a desktop environment on beaglebone nor beagleboard-x15 Oct 29 22:52:08 After login I can open the terminal Qterminal. I have not done ssh yet to BBx15 Oct 29 22:52:10 *I don't use Oct 29 22:52:51 you're saying the image is giving a different environment on first boot vs on every later boot? Oct 29 22:53:15 Allow me to reproduce ... Oct 29 22:56:11 - Reflashing eMMC Oct 29 23:02:05 - reflashing done. Removing the microcard an rebooting ... Oct 29 23:02:31 ... and ... Oct 29 23:03:48 - Rebooted into the previous blackish window environment with the doggy and beaggleboard.org screen Oct 29 23:04:33 - rebooting ... Oct 29 23:06:12 - Sorry, now it reboots back to the expected blackish window environment with the doggy and beaggleboard.org screen Oct 29 23:08:59 What do you mean you don't use the desktop environment? Did you do anything special to boot into the shell and not to start the desktop? Oct 29 23:09:12 I don't have a display attached Oct 29 23:09:30 nor x11 installed for that matter Oct 29 23:10:11 I just log in via ssh and work from the comfort of my laptop Oct 29 23:10:22 Very interesting. Oct 29 23:10:35 (like pretty much anyone who uses an embedded system for anyone) Oct 29 23:10:40 *for anything Oct 29 23:10:49 *a linux-based embedded system I should say Oct 29 23:12:06 Which environment do you use to develop C/C++ code? Oct 29 23:12:18 vim and a makefile Oct 29 23:12:19 meaning, IDE Oct 29 23:13:00 Thanks! Oct 29 23:13:23 Did you uninstall X11? Oct 29 23:13:54 I don't install it to begin with (I start with the console image, not the lxqt image) Oct 29 23:19:21 Oh, I see. You have started with Debian "9.5 2018-10-07 4GB SD IoT" instead of "Debian 9.5 2018-10-07 4GB SD LXQT" Oct 29 23:19:29 no Oct 29 23:19:36 ? Oct 29 23:19:50 what image did you use? Oct 29 23:20:21 the stretch-console image for x15 ( https://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Stretch_Snapshot_console ) Oct 29 23:21:43 Thanks! Oct 29 23:22:25 Curious, what is the word Stretch means? Oct 29 23:22:34 it's the name of the debian release Oct 29 23:25:19 I understand. But why do they call it stretch? Oct 29 23:25:40 Oh maybe it is a normal word for realeases? Oct 29 23:25:52 debian releases are named after characters from the Toy Story Oct 29 23:25:54 http://pixar.wikia.com/wiki/Stretch Oct 29 23:26:22 *the Toy Story movie Oct 29 23:26:46 Thanks! I wouldn't know that. Oct 29 23:27:23 Are you using 32GB microSD for the additional storage? Oct 29 23:28:52 ... storage needed for the development ... Oct 29 23:29:07 32 GB ? o.O what on earth would I need that for? Oct 29 23:29:24 I run from eMMC and have 2.3G of free space there Oct 29 23:30:12 are you trying to compile Google Chrome on your bbx15 or something? :P Oct 29 23:32:14 Definitely much smaller than that :) 2.3G will be sufficient :) Oct 29 23:34:16 Thanks a lot for sharing your experience! Oct 30 01:17:30 got an obscure query - anyone able to help identify a connector type? its for a serial debug header, but its not a type I've ever seen before... Oct 30 01:18:00 tht-shrouded, 2.54mm/0.1" pitch, 4-ways Oct 30 01:18:40 but with internal corners chamfered Oct 30 01:19:16 I suspect this is a situation where an image would work a lot better than text ;) Oct 30 01:19:19 which makes it keyed, together with the pins offset to the top Oct 30 01:19:27 zmatt: yeah probably right .. hold on :p Oct 30 01:24:43 lol, one of our speakers somehow got lost/misplaced in transit a while back... and yesterday someone (newly registered user) showed up in the Gearslutz forum thread about our product advertising one speaker for sale Oct 30 01:24:53 d'oh! Oct 30 01:25:01 that guy must not be the brightest tool in the shed Oct 30 01:25:41 https://imgur.com/a/sLVFAja - clear as mud .. its the white one at the bottom Oct 30 01:27:02 hmm, I'm pretty sure I've seen those... Oct 30 01:27:07 \o/ Oct 30 01:27:27 its got 'HY' and '12' stamped inside the connector .. Oct 30 01:27:35 if that means anything at all Oct 30 01:27:43 which mean I can say with a reasonable degree of certainty that they exist Oct 30 01:27:51 LMAO :D Oct 30 01:30:26 https://imgur.com/a/inYH7lJ - slightly less shadowy Oct 30 01:31:25 oh look! the pinout is printed on the board. That's how good my naked eyesight is LOL Oct 30 01:31:41 hah Oct 30 01:31:51 isn't* :D Oct 30 01:32:00 if you wanted to know the pinout then identifying the connector would have been of no use Oct 30 01:32:17 this is clearly just a random 4-pin header used as console port with a random pinout Oct 30 01:32:26 :P Oct 30 01:32:39 I think GND is pin1 at least... Oct 30 01:33:00 yup it is .. that's the square pad on the base. Oct 30 01:33:24 I'm kinda hoping to find a mating connector .. or else I'm gonna end up butchering the board :/ Oct 30 01:33:37 some colleagues of mine might have a better chance of being able to identify the connector... but they're not around atm Oct 30 01:33:41 not wild about that idea, but could be he only way Oct 30 01:33:43 probably because it's 02:33 Oct 30 01:34:10 no problem, if you wouldn't mind passing it on, I've lost a couple of weeks on this board already .. another 48h ain't gonna hurt ;) Oct 30 01:35:20 I was hoping to start some arm64 dev .. but got stymied with the stupid boot process from friendlyarm Oct 30 01:35:38 I've reflashed once, but I'd started setting up the OS image Oct 30 01:35:52 so I know its serial debug time, alas! Oct 30 01:37:04 veremitz: some of the friendly arm stuff has labelled schematics on the site, just it's quite well hidden... Oct 30 01:37:10 I can't even think of any thin crimps that I could slip on the pins even .. the standard 0.1" ones are too big Oct 30 01:37:15 at least there LCD did Oct 30 01:37:35 suprothunderbolt: yeah I had a dig around, the schematic has them unlabelled, and I can't see anything on PCB and there's no BoM :( Oct 30 01:37:48 wtb BOM :( Oct 30 01:37:48 oh :( Oct 30 01:38:15 I'm sure its not unique to the T3 board either, but nobody seems to have posted what it IS anywhere Oct 30 01:38:23 I'm using one of their LCDs. It was not a fun experience. Documentation was creative. I.e. totally random numbers. Oct 30 01:38:35 next step will be to contact FA themselves, but that's usually chancy Oct 30 01:38:44 yyup, I can imagine :( Oct 30 01:39:14 then again not sure if there are decent LCD suppliers... Oct 30 01:39:36 I think you just have to use a supplier who has vendor backup Oct 30 01:39:49 which often costs more in the short term, but saves years of headaches in the long term .. Oct 30 01:40:11 there's certainly one or two I've come across I'd try out at least . Oct 30 01:40:20 yeah, I tore down one of the LCDs to see if I could work out where the actual panel was coming from Oct 30 01:40:58 mmm I gonna do that with an old monitor which seems to have a power-up supply glitch which causes the panel to strobe for about 30s Oct 30 01:41:01 veremitz: oh? who? Oct 30 01:41:15 Review Display Systems (UK co.) Oct 30 01:41:29 Time for me to get this clock stuff working with this codec... i'm wondering mclk-fs Oct 30 01:42:01 There was another panel display supplier at the show I went to earlier this month .. don't recall the co. name though Oct 30 01:42:21 I've bought a knock-off display from Taiwan(?) where chalk-elec.com hang out Oct 30 01:42:30 they aren't *too* bad Oct 30 01:42:53 beware of customs/import duty though Oct 30 01:43:03 Taiwan knockoff? Taiwan is usually high end kit. Oct 30 01:43:05 I had a nice bill when the courier came wiht mine Oct 30 01:43:24 I'm asking which supplier in Shenzhen :) Oct 30 01:43:28 they're ex-tablet displays, but don't ask me 'what' Oct 30 01:43:39 oh interesting Oct 30 01:43:46 https://www.chalk-elec.com/ Oct 30 01:43:49 refurb. Oct 30 01:44:08 Malaysia, my bad Oct 30 01:44:19 that's 10X what I'd be looking to pay. Oct 30 01:44:19 I knew it wasn't china Oct 30 01:44:32 yeah you don't get much for nothing sadly. Oct 30 01:44:54 I dismissed the Rpi-7" display cos the resolution is shit Oct 30 01:45:19 The $10 touch display i'm using is okay, I'm looking for 4.3" Oct 30 01:45:20 but you don't get even near to 1xHD for less than about $80 Oct 30 01:45:30 but the $15 is a big step up. Oct 30 01:45:38 mhmm Oct 30 01:45:57 didn't .. err .. 4D do a decentish 4.3" BBB display? Oct 30 01:46:04 cape even? Oct 30 01:46:13 its probably nearer $30-50 Oct 30 01:47:04 but I do like my 15" touch panel .. I just need to frame it and mount it with its B-Pi-Zero on the wall :D Oct 30 01:47:16 (Smart mirror project) :D Oct 30 01:47:19 mmm... I'm wondering if the mclk-fs will help... Oct 30 01:47:21 oh neat :) Oct 30 01:47:53 not a cheap project that one .. its heading steadily towards $200 :( Oct 30 01:49:49 I want to do one of these: https://www.indiegogo.com/projects/nybble-world-s-cutest-open-source-robotic-kitten but with a BBB or BBBlue Oct 30 01:50:21 ditch -both- the Pi and the arduino .. cos the PRU will run the calcs fine, and the processor any 'AI' Oct 30 01:55:08 mmm if i add dai-tdm-slot-num the driver doesn't load... Oct 30 01:57:02 suprothunderbolt: mclk-fs is definitely not what you want Oct 30 01:57:31 if you add dai-tdm-slot-num you may also need to add dai-tdm-slot-width Oct 30 01:57:33 I've just been attempting to read how all the clock stuff in mcasp interacts. I am confused. Oct 30 01:57:37 it shouldn't be necessary though Oct 30 01:57:42 yeah, I added both . Oct 30 01:57:49 I didn't add a a rx-mask Oct 30 01:58:16 have you tried configuring the cpu as master instead of the codec? Oct 30 01:58:48 no Oct 30 01:59:13 might be worth a try Oct 30 01:59:31 of course the "proper" solution would be to fix this clearly broken driver :P Oct 30 01:59:38 I'm not sure how. Oct 30 01:59:50 ahh that codec driver? Oct 30 01:59:54 yes Oct 30 02:00:22 I'm mostly assuming it's my fault somewher Oct 30 02:00:23 easier said than done :p Oct 30 02:00:35 no it's not your fault Oct 30 02:02:16