**** BEGIN LOGGING AT Thu Feb 18 02:59:59 2016 Feb 18 04:40:57 rfkill Feb 18 05:19:46 Good morning Feb 18 05:20:14 any one here ? Feb 18 05:21:26 how to reconfigure the eqep Feb 18 05:21:39 so that i can read the frequency Feb 18 05:23:16 the freq of one signal Feb 18 05:23:26 i mean with 1 input Feb 18 05:41:00 Hi! Who worked with pulse-code on PRU? Feb 18 05:43:47 Some values, 226 or 118 coding fine, but 100 - incorrectly. Checked on oscilloscope. (I'm talking about the output signal from PRU). Feb 18 06:15:25 up!) Feb 18 07:12:31 Hello I have some questions about a Beaglebone Black ' ADC. is it possible to attain a 16 bit through oversampling? my signal should be sampled at 1KHz. Feb 18 07:13:26 how noisey is the beaglebone? the arduino Uno is quite noisy over its internal bus, if there are many processes running (RTC, SPI, I2c and 1 Wire) Feb 18 07:17:33 if I connect the ILI9341 TFT to the beaglebone how much do I deminish my processing time for other processes? (I would like to display in 10ms) delay Feb 18 07:17:51 Bastiaan: in my experience the built in ADC sucks Feb 18 07:18:16 do you have a good solution? Feb 18 07:18:29 if you need 16 usable bits, get a high quality ADC and hook it up Feb 18 07:21:50 if 16 bits is achievable, why not 24 bits or are the requirements (board layout, traces etc pose a problem?) Feb 18 07:23:16 is the SPI of the beaglebone faster then the ADC with its 200KHz sample rate? Feb 18 07:27:16 yes, SPI can go much faster Feb 18 07:28:14 probally can over sample Feb 18 07:28:31 will it be useful is a better question Feb 18 07:29:07 I personally get about 9-9.5 useful bits from it with either 4x or 16x oversample Feb 18 07:30:11 is there a good manual for building a 24bit ADC and to hook it up to the BBB? what other variables do I have to look at? Feb 18 07:30:49 paging KotH for some reality readjustment Feb 18 07:43:02 Bastiaan: given that this looks like a research project. I'd recommend you start out by understanding how ADCs work and what 'useful bits' are. Then you could put that into relation of what you are actually trying to measure. Feb 18 08:00:20 Thanks TBR so you are trying to say that by trying to get a Resolution of 12 Bit or more, Certain Factors contribute to the ADC Useful Bits (Noise from resistors, Aliasing, Sampling Rate) as we all would like to achieve the Maximum from a Device your information was useful. what is a good approach in doing so? sincerely yours B Feb 18 08:36:45 Bastiaan: if you need more than 10bit resolution, you need to know how to do analog designn Feb 18 09:22:23 good point, any good leads where to start? or which I should read through? Feb 18 09:24:43 read the appnotes written by bob pease and jim williams on ADCs Feb 18 09:24:53 or generally what those two wrote Feb 18 09:25:01 alternatively you can hire me as conslutant :) Feb 18 09:36:10 Hi there! Feb 18 09:36:25 does BB black accepts any uSD card? Feb 18 09:48:09 pretty much, yes Feb 18 09:51:18 ok, thank you. Feb 18 10:12:29 so, should i be able to boot the latest images from the SD card and "chat" over usb to the BBB? Feb 18 10:14:01 at the moment, having tried 2 images, the BBB seems to boot, and my local kernel sees a new ethernet interface, but getting a dhcp ip for that interface from the BBB fails Feb 18 10:14:53 KotH: yeah that's one thing about analog audio that really sucks... the required SNR -.- Feb 18 10:15:07 although it's hilarious that some manufacturers try to peddle 32-bit ADCs Feb 18 10:15:42 (with the same performance specs as their 24-bit ADCs, but more expensive of course) Feb 18 10:17:01 zmatt: there are 32bit ADCs Feb 18 10:17:11 zmatt: TI and ADI have one each Feb 18 10:17:26 zmatt: and they actually get an ENOB of ~28 bits Feb 18 10:17:41 (but only at <10sps) Feb 18 10:18:10 folks, does cloud9 terminal in terms of functionality the same as a normal debian console? Feb 18 10:18:16 hehe, yeah I was talking audio ADCs... Feb 18 10:18:31 i.e. 48 kHz and *then* claim 32-bit Feb 18 10:18:46 or well, it does transfer 32 bits of course Feb 18 10:35:07 how long will be executed command LSL in compare with ADD? Feb 18 10:37:28 on which core? Feb 18 10:37:38 PRU0 Feb 18 10:39:47 all PRU instructions take exactly 1 cycle except for memory access (LBBO, SBBO, LBCO, SBCO) and things like HALT and SLP of course Feb 18 10:41:04 XIN/XOUT can also take multiple cycles in same cases (e.g. contention when accessing scratchpad, or when doing a direct core-to-core transfer) Feb 18 10:42:01 hm...thanks.. Feb 18 10:42:04 well, not "e.g.", those are the only cases I'm aware of Feb 18 10:42:17 there's also a stall counter in the PRU core control registers Feb 18 10:42:38 that counts the number of cycles PRU was unable to execute an instruction Feb 18 10:44:05 (as long as the cycle counter is running) Feb 18 10:47:58 i have a problem and i know, it exactly connected with delay: http://beagleboard.org/Community/Forums?place=msg%2Fbeagleboard%2FZISc8SFdWqA%2FiATOjPELBAAJ Feb 18 10:48:46 I need to generate a pulse signal with of one delay Feb 18 11:04:32 hello. How should i power beaglboard rev C5 Feb 18 11:04:40 5V 2.5 A? Feb 18 11:04:56 yes Feb 18 11:05:12 1A is fine for a start Feb 18 11:10:05 I have problem while flashing Feb 18 11:10:24 USR0 D is flashing and it is everything what happeds Feb 18 11:10:33 after 1,5 h nothing has changed Feb 18 11:11:09 Which diodes should be blinking if everything is OK? Feb 18 11:11:23 it's not flashing, you've just booted a system Feb 18 11:13:41 make sure you actually put a flasher (and not a standalone image) on the card, and hold down the S2 button (the button nearest to the card slot) while powering up Feb 18 11:15:32 (flasher images have a filename starting with "BBB-eMMC-flasher-") Feb 18 11:16:49 Hmh. I am a litte bit confused. I used that wiki http://elinux.org/Beagleboard:BeagleBoard#BeagleBoard_Rev_C4.2FC5_Reflashing_Procedure Feb 18 11:17:00 And image from http://beagleboard.org/latest-images Feb 18 11:17:39 And wiki says that it is flashing to internal memory Feb 18 11:18:02 and the button which should be pushed is USER BUTTON S1 Feb 18 11:18:52 not pushed Feb 18 11:18:55 given that this info still refers to angstrom it must be really really old Feb 18 11:18:55 held at power up Feb 18 11:19:10 remove all power. push button. hold button. put power Feb 18 11:19:19 zmatt: note that he doesn't have a BBB but a BBclassic Feb 18 11:19:25 oh Feb 18 11:19:31 my bad Feb 18 11:20:30 my sugestion was to forget about nand and use SD Feb 18 11:20:41 alexx600: sorry, I don't really know much about the classic BB Feb 18 11:21:01 tbr: I use SDHC card Feb 18 11:21:18 IIRC we pointed out some working images that don't need flashing yesterday Feb 18 11:21:37 I'd suggest to log in via RS232 console Feb 18 11:22:46 hmh, sorry i am a litte bit confused. I got from my friend old beagleboard. I don't know if in eMMC is good working "image". So i tried to reflash it. Feb 18 11:22:59 its does not have emmc Feb 18 11:23:03 it has raw NAND Feb 18 11:23:09 there are flasher images for it too Feb 18 11:23:22 but as suggested, the best is to "nand erase" Feb 18 11:23:26 and just boot from SD Feb 18 11:23:30 and ignore the NAND Feb 18 11:24:40 what about a BBB? should i ignore the eMMC and only use the SD? Feb 18 11:24:40 OK. SO when you say NAND you mean internal memory not NAND SD CARD TECHNOLOGY? Feb 18 11:24:44 yes Feb 18 11:25:07 ScrumpyJack: if it works for you, use it Feb 18 11:25:16 ScrumpyJack: K Feb 18 11:25:19 ScrumpyJack: no Feb 18 11:25:20 ScrumpyJack: that's up to you. the BBB has a bit saner on board storage as it's eMMC, but you can also ignore it and use SD Feb 18 11:25:48 hmh, So i have angstrom on my SD card image from http://beagleboard.org/latest-images . I didn't know taht there is another image to boot and another to flash Feb 18 11:26:11 if i could get the dhcp stuff working over usb that would help :) at the moment, booting off the SD card blocks at that point Feb 18 11:26:24 what image did you put on it? Feb 18 11:26:46 lemme check Feb 18 11:26:46 (also, consider ethernet instead if possible) Feb 18 11:26:49 Angstrom Distribution (BeagleBoard and BeagleBoard-xM - 4GB SD) 2012-01-11 - more info Feb 18 11:27:05 only this is possible on beaglebone Feb 18 11:27:28 i used raspberry, and when i put sd card in slot, and power board it starts works Feb 18 11:27:28 debian 8.3 from here -> http://beagleboard.org/latest-images Feb 18 11:27:53 alexx600: sorry for any confusion, my question was intended for ScrumpyJack Feb 18 11:28:17 debian is only to BeagleBone, BeagleBone Black and Seeed BeagleBone Green via microSD card Feb 18 11:28:24 not to BeagleBoard Feb 18 11:28:30 ScrumpyJack: ok, that *should* bring up usb networking by default I think... I'm not sure Feb 18 11:28:36 alexx600: this isn't about your issue Feb 18 11:28:45 ok, thanks for help. Feb 18 11:28:45 two conversations going on here Feb 18 11:28:52 zmatt: thanks Feb 18 11:29:46 ScrumpyJack: I never use USB networking so I don't really keep track of whether there are any problems with it in any particular image. afaik it should work in any image by default Feb 18 11:31:10 ok, i got it working Feb 18 11:32:00 tbr: It would be great if you could give me advice how to boot system from SD Card? I power board 1,5 A DC. Have SDHC 4GB Card which Anstrem linux, and when i power it. Nothing happends. What RS232,UART,USB cable would give me if nothing boots? Feb 18 11:34:26 so i'm a bit further, but javascript doesn't seem to run Feb 18 11:34:28 http://st.ilet.to/bbb.png Feb 18 11:39:17 tbr: will RS232 console works? It doesn't boot Feb 18 11:39:17 ? Feb 18 11:51:39 alexx600: I think for the BB classic you need a breakout cable and a USB-to-RS232 converter Feb 18 12:01:47 ds2: what low-pass filter did you use for the 16x oversampling? (given that the ADC's only built-in capability is averaging, which is a rather crappy filter: http://goo.gl/vOCyNO ) Feb 18 13:43:02 Hello! Feb 18 13:43:15 I need information! Feb 18 13:43:36 Do you have a Beaglebone Black Revisão C in stock? Feb 18 13:45:26 Beagles don't have a programmable microcontroller do they? Feb 18 13:50:59 ScrumpyJack: they have the PRUs Feb 18 13:51:10 which are essentially micro controllers Feb 18 13:51:42 they are on the AM335x SoCs, so BBW, BBB, BBG Feb 18 13:54:07 tbr: this? http://elinux.org/Ti_AM33XX_PRUSSv2 Feb 18 14:26:27 ScrumpyJack: yes Feb 18 14:32:53 is there a nice IDE to program them? :) Feb 18 14:37:24 i've found this https://github.com/biocode3D/prufh Feb 18 15:05:47 So.. I've got squashfs on a partition... Identical deployed.. Feb 18 15:05:58 And now one of them is returning a different checksum... WTF. Feb 18 15:27:57 [ 9.080834] SQUASHFS error: lzo decompression failed, data probably corrupt Feb 18 15:27:57 [ 9.087907] SQUASHFS error: squashfs_read_data failed to read block 0x594c11 Feb 18 15:27:59 Right Feb 18 15:45:53 So. is there a way to properly erase the on-board flash on the BBB? Feb 18 15:57:21 Spidler: just overwrite it Feb 18 15:58:05 av500: Considering that I'm seeing corruption on squashfs on top of it, I don't really want to just overwrite it and hope that it won't re-corrupt, but to actually perform a nand-erase on it. Feb 18 15:58:30 there is no NAND erase Feb 18 15:58:39 Spidler: it's not a NAND, but an eMMC Feb 18 15:58:42 its an emmc, so it behaves like any other SD card Feb 18 15:58:43 Ah, right Feb 18 16:57:03 hello to everybody Feb 18 16:57:47 I need help trying to make my beaglebone boot after fat partition mess up.... Feb 18 16:58:25 I formatted that partition with windows utility by mistake..... Feb 18 16:59:10 now I writo again the files in there by coping them from a new sd card with same image... Feb 18 17:00:09 but when i connect the BBB with the usb cable I've got only 3 led that always stay on....and nothing else Feb 18 17:00:36 Thanks in advance for the help.. Feb 18 17:16:44 Fix it......!!! Feb 18 17:17:13 I hadn't formatted well the partition. Feb 18 17:17:33 Now it works..... Feb 18 17:17:49 sorry for boring with my questions.!! Feb 18 17:44:26 ScrumpyJack: I'm woring on PRU code right now. So far it's been a real tragic experience, but I was trying to use remoteproce. Now with uio_pruss at least I have an idea my PRU isn't enable. Don't know why not; the .dtbo I put in should have fixed that. Feb 18 17:45:39 where did you get the uio_pruss module? Feb 18 17:45:49 At least that's what "prussdrive_open failed" turns up when researched; the PRU isn't enabled. So far I haven't found out how to *tell* if it's enabled. Feb 18 17:46:12 Yesterday zmatt said I should be using a bone kernel so I put in 4.1.18-bone19, which has uio_pruss. Feb 18 17:46:29 what's in /sys/class/uio/uio0 Feb 18 17:47:05 As usual /sys/class/uio is empty. Feb 18 17:48:07 uio & uio_pruss claim they're loaded. Simple .dbto with "status = okay" for <&pruss> claims it's loaded. Feb 18 17:50:10 kernel log agrees the "cape" for the .dtbo is loaded. Don't see anything obvious for uio_pruss on a quick scrolling around. Feb 18 17:51:03 does it need a driver in /lib/firmware ? Feb 18 17:52:16 From what I've researched, uio_pruss does not need anything in /lib/firmware except the .dbto to load the "cape" that is supposed to enable the PRU. One provides a file to uio_pruss to tell it what code to execute. Feb 18 17:53:03 syslog nicely agrees with kern.log; "cape" is loaded, nada regarding uio_pruss. I would expect a LKM to say *something* when it goes in. Feb 18 17:53:42 ah, what happens if you load is and provide it some empty file? any log noise? Feb 18 17:54:22 Well since prussdrv_open fails it can't get to step where I could try to load anything. Feb 18 17:55:08 i need to get the bone kernel too and see what i get. gtg. Feb 18 17:57:15 Anyone know if there's a reason modding the main .dtb would work to enable a PRU where the simple overlay would not? Feb 18 17:58:11 hey! Feb 18 17:59:20 i am attempting to flash my beaglebone black via uart as an exercise Feb 18 17:59:41 but i am having difficulty navigating the emmc from my bootloader (barebox) Feb 18 18:00:44 i have managed to boot to the final stage bootloader but i don't know how/where to write my MLO and bootloader.bin to on the emmc Feb 18 18:01:10 if anyone has some pointers on how to do this (even from das uboot) that would be awesome :) thanks Feb 18 18:27:08 As expected, no change: uio_pruss is loaded, overlay is loaded, /sys/class/uio is empty, /dev has no uio device. I've read some 4.1.x kernels don't play nicely with uio_pruss. Are we sure 4.1.18-bone19 does? Feb 18 18:40:39 Ragnorok: well I got the pointer of switching to the -bone kernel because I saw rcn-ee mentioning this, so generally I'd trust he is right Feb 18 18:41:51 Roit toe. Feb 18 18:42:20 Just checking. Nothing so far has caused /sys/class/uio to populate. Feb 18 18:42:22 can you pastebin your overlay source? Feb 18 18:42:46 Sure. Let me figure out how to get it off the BBB and I'll do that. Feb 18 18:42:48 and there are no kernel messages when you loaded the overlay? Feb 18 18:43:21 There are when I load the overlay. There are none when I load uio_pruss. Feb 18 18:43:40 loading the overlay should automatically load uio_pruss afaik Feb 18 18:44:15 Ah. That doesn't happen. Then again if the overlay is not actually enabling the PRU, it makes sense. Feb 18 18:45:25 you can also inspect the live state of the device tree with... ehm... hold on... Feb 18 18:46:57 http://pastebin.com/uMkkPTrY Feb 18 18:47:19 This is from a SO post regarding enabling PRUs. Feb 18 18:47:39 dtc -I fs -O dts -o tmp.dts /proc/device-tree converts your live device-tree into a dts Feb 18 18:48:11 But it's pretty much the same all over. I also changed my dtb to include status="okay" in each of the two pru@ sections. Feb 18 18:50:01 I don't see any pru@ nodes Feb 18 18:50:56 Not in that overlay. They're in the main dtb that uEnv points to. Feb 18 18:51:39 The device capture shows them. I'll add that bit to the pastebin. Feb 18 18:52:11 I don't see them in neither 4.1-bone nor 4.4-bone Feb 18 18:53:24 http://pastebin.com/sq2YsAC8 Feb 18 18:53:26 wait, are you using a custom main dtb also? if so, what kernel did it derive from? Feb 18 18:53:30 That's what in the dbt I'm loading. Feb 18 18:54:06 Sadly I don't know, as I didn't write it, but it's some 4.1.x version. Feb 18 18:54:08 those DT declarations are for remoteproc Feb 18 18:54:17 Ok. Feb 18 18:54:19 yes, some 4.1.x-ti version you mean Feb 18 18:54:51 uio_pruss uses compatible = "ti,pruss-v2"; Feb 18 18:54:54 Possibly. I've been working on code while others have been working on dtb, so I am sadly not very conversant with this area. Feb 18 18:55:24 So I should scrap this dtb and craft one based on a bone version, seems like. Feb 18 18:56:36 as an ugly hack you could probably patch up this pruss node using your overlay, but there are quite a few more DT differences so yes basing one on the original dts of the appropriate kernel version would be better Feb 18 18:57:11 Roit toe. I knew it would be time to learn this dt stuff some day. Looks like today. Thanks! Feb 18 19:02:13 http://pastebin.com/EMhw63x3 Feb 18 19:02:20 this is untested, just a guess Feb 18 19:02:27 no idea if overlays can actually *do* this Feb 18 19:03:03 lol Well what the heck. Worst it'll do is fail, right? Feb 18 19:03:25 well, maybe save any open documents ;) Feb 18 19:03:35 wut? Feb 18 19:03:44 before running it Feb 18 19:03:47 in case of kernel panic Feb 18 19:03:57 lol, wink Feb 18 19:06:57 DTs are... not really complicated, just annoyingly verbose and repetitive, and not always very well documented Feb 18 19:07:15 they really ought to be generated from a more high-level format Feb 18 19:07:46 Yeah. It's that annoying repetition that has kept me from learning them, while I had someone else to do it. (grin) Feb 18 19:09:33 there's also the little problem of the fact the device hierarchy isn't actually a tree but a directed graph, which means some aspects are not only poorly represented but also with no obvious way on how to fix it Feb 18 19:10:22 lol Still and all, after JCL it's probably not that hard to figure out. I should just bite that bullet. Feb 18 19:13:35 JCL... Feb 18 19:13:40 * zmatt carefully steps backwards Feb 18 19:17:26 Serious. Fortunately it's been so long I've forgotten nearly all the arcana for that. it does however provide a yardstick for obtuse complexity that's rarely exceeded. Feb 18 19:32:18 Hello would be possible to power up the beagle bone black from the 5V jack using a lipo battery? lipo battery is 11.1 V so i just need to step down to 5V and connect it to the 5V jack ? Feb 18 19:39:55 linkages: as long as it's exactly 5V and with sufficient current, sure it doesn't matter what's behind that Feb 18 19:53:23 tbr, The TPS65217C power management device is used along with a separate LDO to provide power to the system. i copyed from the manual... does it provide any sort of voltage regulation or current protection? Feb 18 19:53:56 thre is no mention in the manual Feb 18 19:54:16 linkages: no, the board expects exactly 5V, as stated in the SRM Feb 18 19:54:43 ok thanks Feb 18 19:55:12 in theory the TPM has capability to be run from LiIon batteries, but that's not properly available due to how the hardware was designed (as far as I understand) Feb 18 19:56:03 there are those 4 pins next to the power connector Feb 18 19:56:54 for more details you'd need to check the schematics, data sheets and kernel support Feb 18 19:57:01 googling ...ive seen some hacking to connect a battery .. but power is taken from 3.7 connection Feb 18 19:57:31 yes, it's 3.7 (so 4.3 max) Feb 18 19:57:34 so single cell Feb 18 19:57:38 not 3 cell Feb 18 19:57:59 yes... my battery is 3 cell Feb 18 19:58:12 so i prefer to use the 5v plug Feb 18 19:58:27 then you need an external regulator Feb 18 19:58:43 yes ... i step down the voltage from 11.1 to 5V Feb 18 19:59:14 my concern is... what happen if the battery is dischagning and is going below 5.v Feb 18 19:59:47 probably nothing... BBB just power off Feb 18 20:00:18 I suspect the TPM will throw a fit about undervoltage and it will turn off Feb 18 20:01:18 yep... voltage regulator give a max 2 amp ... seems good if i want to plug some usb perifericals Feb 18 20:02:03 I simply use those USB battery packs for those use cases Feb 18 20:02:22 they have a decent charge regulator and will turn off cleanly when discharged Feb 18 20:02:26 only 500 mA current Feb 18 20:02:36 no, there are some that will do 2A Feb 18 20:02:40 1A is common Feb 18 20:03:04 ok Feb 18 20:03:09 ill check Feb 18 20:03:28 im trying to build a quadcopter Feb 18 20:03:42 need to use PRU also Feb 18 20:03:52 will be probably complex Feb 18 20:05:30 you'll probably want to have a discharge gauge or at least wire the pack voltage (through a voltage divider!) to the BBB ADC Feb 18 20:05:40 that way you have a chance to know when you're running low Feb 18 20:23:24 Hello all am planning to learn some embedded linux Feb 18 20:23:50 could you please point me to some good links Feb 18 20:23:51 ? Feb 18 20:34:34 zmatt: I see in modules.alias that pruss-v2* appears to be mapped to uio_pruss. I presume that's how it knows to load that module? It's still not working; I'm researching debugging options for dto load. Feb 18 20:57:44 bou6: do you have a beaglebone? Feb 18 21:46:02 I'm comparing am335x-boneblack.dts (from the dtb) with my custom dtb. The former is from 4.1.18-bone19, the latter is for a 4.1.x TI kernel. They are largely the same, but I see the sixth line under ocp{...} that linux,phandle = <0x45> in the form and <0x55> in the latter. Since this ia defining the same hardware I would expect the same value. Any reason why it wouldn't match up? Feb 18 21:54:17 Ragnorok: phandles are just arbitrary identifiers assigned by dtc. it's normal for those to vary Feb 18 21:54:30 Oh. Cool. Thanks. Feb 18 21:58:31 wkup_m3 seems to be related to remoteproc, but if it's hardware, wouldn't it still exist and its declaration be okay? Feb 18 21:59:20 Or is its removal good in the name clutter reduction for things that may not be implemented in this kernel? Feb 18 22:01:03 This under l4_wkup@44c00000. It is declared in other places in boneblack, just not under that label. Feb 18 23:13:02 Ragnorok: remoteproc is also used for the cortrex-m3 used for some power management tasks (e.g. suspend) Feb 18 23:13:07 it is not related to pruss Feb 18 23:13:32 it is not implemented in the -bone kernel I think, hence the DT entries related to it will have no effect Feb 18 23:13:53 (at least, not implemented in 4.1-bone, it might be in later ones, not sure) Feb 18 23:14:14 I figured. I took that bit out. I left in a emic entry. Feb 18 23:14:50 I'm thinking extra hardware defined but not implemented isn't really a show stopper for the dbt. Feb 18 23:14:52 dtb. Feb 18 23:15:03 correct Feb 18 23:15:34 \o/ I guessed something right. Baby steps. Feb 18 23:16:11 btw I hope you're not editing a decompiled dtb... that's really making life hard on yourself Feb 18 23:17:13 Well sort of. It's a dtb built by someone else from a decompiled dtb. I'm comparing to a decompiled dtb to have a guess what should at least be defined. Feb 18 23:17:33 you can't just get the actual source from them? Feb 18 23:17:55 Well I should say a dts built by someone else, and I'm comparing to a decompiled dtb. Feb 18 23:18:05 is there an explanation what the different kernel images are for like linux-image-4.4.1-bone5 or linux-image-4.4.1-ti-rt-r7 Feb 18 23:18:12 what is the difference ? Feb 18 23:18:30 tat: I'm not sure what the deal is with the 4.4-ti series since the current official kernels from TI are 4.1 Feb 18 23:18:54 but what is bone4 bone5 bone17 stand for ? Feb 18 23:19:08 I've always assumed version Feb 18 23:19:21 version of the patchset applied on top of the baseline kernel Feb 18 23:19:45 eliasbakken: it is me morris, i am still on the hunt of that boot problem after loading BB-BONE-REPLICAPE-00B3 Feb 18 23:19:54 the main kernel series right now are 4.1-ti, 4.1-bone, and 4.4-bone Feb 18 23:20:00 4.1 is the long-term support series basically Feb 18 23:20:21 i am using 4.1.15-bone17 Feb 18 23:20:55 i exidently installed systemd and now i get net eth0: phy 4a101000.mdio:01 not found on slave 1 on boot Feb 18 23:20:59 -ti diverges more strongly from mainline and has e.g. fancy power management and some other stuff that -bone doesn't have Feb 18 23:21:05 it stalls the init for about 3 minutes Feb 18 23:21:21 every image uses systemd Feb 18 23:21:28 (every image from rcn-ee that is) Feb 18 23:21:35 that message has nothing to do with the stall Feb 18 23:21:35 it is stated that this problem was fixed in a mail that is more than one year old ??? Feb 18 23:21:39 tat, yes the network lsb raise thing. Feb 18 23:21:46 yes Feb 18 23:21:55 the message is harmless (though fixable) Feb 18 23:22:16 so where comes my stall from Feb 18 23:22:22 trying to bring up ethernet probably Feb 18 23:22:35 systemd-analyze blame tels me Feb 18 23:22:40 when using ifupdown Feb 18 23:22:41 3min 1.911s networking.service Feb 18 23:22:43 yeah Feb 18 23:22:46 ifupdown Feb 18 23:22:55 so what is the solution, Feb 18 23:23:00 trash ifupdown ? Feb 18 23:23:33 yeah, use an actual network manager. the big images use connman, but I have no experience with it. I'm using systemd-networkd but that requires upgrading to debian stretch since jessie's systemd is too old Feb 18 23:23:56 ok i give connman a try Feb 18 23:23:58 systemd-networkd does work like a charm for me Feb 18 23:24:10 also has the dhcp client integrated Feb 18 23:24:26 and coordinates well with other services like systemd-resolved and systemd-timesyncd Feb 18 23:25:08 the -rt kernels additionally have real-time patches. if you don't specifically need it, use a normal kernel instead Feb 18 23:25:47 ok, gotto switch trains and won't have internet, will be back when I arrive at work Feb 18 23:25:53 bbl Feb 18 23:26:11 so would you recommend updating a bbb to stretch or would i spend my days hunting way more issues than working on my own stuff ? Feb 18 23:47:54 yes that fixed it, switching to connman made an end with that Feb 19 00:23:39 I've never had much trouble with upgrading, but I'm fairly experienced with debian so my view may be biased Feb 19 00:24:01 I prefer it to having to cope with old crap :P Feb 19 00:30:53 *sigh* I pointed out a power domain violation recently in one of our prototypes, their argument was "the signal is fairly high impedance when high so it shouldn't be a problem" ... Feb 19 00:31:22 aaand now I find a bbb on my desk with a post-it "uart1_rxd defect" on it Feb 19 00:32:58 * zmatt hates to say "I told you so." Feb 19 00:33:29 nm, scrap that, I'm actually pretty okay with saying "I told you so" Feb 19 00:34:17 lol Feb 19 00:34:53 ""don't be a duffer, always buffer..."" Feb 19 00:36:52 they could really use a nanny-cape Feb 19 00:38:50 yeah a bus-switch or something would help a lot... especially since correctly reasoning about power domains and power sequencing doesn't seem to be on the horizon for them unfortunately -.- Feb 19 00:54:44 does "unzip" in Debian unzip files, esp. .zip files? Feb 19 00:54:51 yes Feb 19 00:55:05 I am getting command not found Feb 19 00:55:14 then it's probably not installed Feb 19 00:55:19 Oh. Feb 19 00:55:20 Okay. Feb 19 00:55:24 sudo apt-get install unzip Feb 19 00:55:39 (first do "sudo apt-get update" unless you've already done that less than a day ago) Feb 19 00:55:45 Oh...I probably do not have unzip. I gotcha now. Feb 19 00:55:54 Okay. Feb 19 00:57:25 You were right. Thank you. Feb 19 00:58:47 I am currently using my BBG to produce temp. readings and have it signaled by a colored LED. Feb 19 00:59:15 When does that BBBlue come out? Feb 19 00:59:30 Robots! Feb 19 01:01:25 "Additional details to be announced in February. Production planned for May. Feb 19 01:01:27 " Feb 19 01:02:07 looks like it's just a BBB with a cape? Feb 19 01:02:17 Cool! Feb 19 01:02:23 I will be patient. Feb 19 01:02:35 Capes! Feb 19 01:35:29 damn, the mating connector for the bbx15 expansion connectors is $8 Feb 19 01:37:23 zmatt: yes, and you need 8 of them if you build a stacking board Feb 19 01:38:28 I think the connector is nice, but i don't think that aspect of it was thought through well enough, $64 just on connectors for expansion boards is too much. Feb 19 01:38:35 what's the hardware details, besides the cpu is the same ... Feb 19 01:38:54 [yes, of the blue...] Feb 19 01:39:29 "The insertion loss-to-crosstalk ratio (ICR) with five-aggressor differential FEXT meets the extrapolated IEEE802.3ap specification for 10 Gbps with plenty of margins." <-- is that not a little bit excessive even for the highspeed signals that an am572x has? :P Feb 19 01:39:50 veremit: no details yet but it looks to me like it's just a plain BBB with a cape Feb 19 01:40:00 ugh gues I'll have to wait til .. oh its February Feb 19 01:40:01 but it's hard to say Feb 19 01:40:17 they can't make two boards for less than the price of one, but yeah .. pics are deceiving Feb 19 01:40:55 I didn't see anything about pricing? Feb 19 01:41:05 they're billing it as cheaper than the bbb Feb 19 01:41:13 ie. to compete with RPi /etc Feb 19 01:41:22 to a limited extent Feb 19 01:41:36 ok, so it's not a bbb Feb 19 01:41:57 well it probably lacks hdmi then, since I see no mention of it (nor use for it) Feb 19 01:42:16 but all other features of the bbb seem to be listed Feb 19 01:42:17 no .. who needs hdmi on an arm board? Feb 19 01:42:25 ethernet will probably go Feb 19 01:42:26 with wifi Feb 19 01:42:46 what's that big connector at the top then? :P Feb 19 01:42:55 its not a blue .. thats an old pic Feb 19 01:42:58 ah Feb 19 01:43:01 bbb with cape :P see?! Feb 19 01:43:08 its a robocape afaik Feb 19 01:43:27 http://hackaday.com/2016/01/11/introducing-the-beaglebone-blue/ Feb 19 01:43:57 http://beagleboard.org/blue looks like that pcb pic might be the actual blue Feb 19 01:44:03 (if you scroll down a bit) Feb 19 01:44:16 yeah the pcb outline/etc Feb 19 01:44:36 looks noticeably smaller Feb 19 01:45:37 unless thats one massive bga.... Feb 19 01:45:44 weird indeed Feb 19 01:46:13 based on the dimensions of e.g. the usb host connector it looks bbb-sized to me Feb 19 01:46:25 yeah could be altoids tin factor again .. Feb 19 01:46:50 but if so .. what's that *** massive bga?! Feb 19 01:46:51 also the same difference in curvature of the corners Feb 19 01:47:06 fpga?! Feb 19 01:47:23 400 balls Feb 19 01:47:30 no .. cost prohibitive surely. Feb 19 01:49:21 * zmatt scratches back of head Feb 19 01:50:26 not *that* many balls actually... the am335x is 18x18 (versus that thing 20x20), which means its pitch is relatively big too Feb 19 01:50:53 yea.. Feb 19 01:51:14 seems like way too many pins for the wireless chipset though Feb 19 01:51:28 waay too mny Feb 19 01:52:15 hope they used an sdio channel or something .. not tacked it to the usb .. hahaha Feb 19 01:52:34 also, if the am335x isn't on this side, then where's the fuckton of decoupling caps that's obligatory on the reverse side of the am335x? Feb 19 01:52:46 Can a phandle be just any old value? Long as it's one byte? Feb 19 01:53:11 Ragnorok: phandle values are never picked manually, they're selected by the device tree compiler Feb 19 01:53:30 and the values inside <> are 32-bit, not bytes Feb 19 01:53:34 So I should delete all the phandle lines? Feb 19 01:53:45 you shouldn't edit a decompiled dtb Feb 19 01:53:52 *shudder* Feb 19 01:54:03 because you've lost all the original label-references Feb 19 01:54:33 device trees are basically untyped, so when decompiling dtc can't tell the difference between a number that's just a number versus one that references a phandle Feb 19 01:56:10 I actually have no idea how it even knows whether to present a property as string(s) or as array... it probably just guesses based on some heuristic Feb 19 02:03:06 zmatt: All the high speed signals are on one connector (the Sata Expansion and the 2 x PCIe) so that could have been an expensive one if required, and the oher just normal cheapies. Feb 19 02:03:45 zmatt: Also, i detected a cross labeling on UART1, TX is named RX and vice versa. Hope they fix that because it could be confusing to expansion designers. Feb 19 02:04:25 Strontium: well it's nice to use a somewhat okayish connectors for HD video signals, although I'm less worried about the connector and more about the serious deficiency of ground pins Feb 19 02:06:39 zmatt: yes, there isnt a lot of GND, although there are a few around the SATA and PCIe signals. Feb 19 02:07:38 zmatt: I am worried that the $64 budget for connectors on expansion boards will be a huge disincentive for making expansion boards. Feb 19 02:07:43 yeah that's how I remember it, that connector is the only one with sufficient grounds Feb 19 02:08:48 One up side with surface mount connecotrs though is the expansion boards can choose to route what they like through and cut off what they want to keep exclusively. Feb 19 02:11:07 zmatt: I also don't understand all the N/C pins on the expansion connectors. They could be GND's or could route out some more of the IO (there is a bunch of IO that isnt exposed). Feb 19 02:11:27 I dont see the point in having N/C pins on an expansion connector at the very least they should be GND Feb 19 02:13:14 A1 and A4 are nc on all four of 'em though, there might be something special about them Feb 19 02:13:36 could be for debugging/jtag programming/test purposes Feb 19 02:13:47 no Feb 19 02:14:08 sometimes a reason for NC .. dunno wha case you're talking about though :p Feb 19 02:14:26 same two pins on each of the four expansion connectors Feb 19 02:14:51 and they're special pins of some sort Feb 19 02:14:59 veremit: They are truly N/C they have no connection at all, and they even have GND jumping around them. For example pin 2 and 3 on P19 when 1,4,5 is GND. Feb 19 02:15:14 there are also a few more NCs, those may have a history of some sort perhaps Feb 19 02:15:18 -shrug- Feb 19 02:15:47 Im hoping a chinese clone source of the connector pops up Feb 19 02:16:49 pins 2,3,33 of P19 are indeed weird, those look like they were used for something in some design iteration and this was later removed Feb 19 02:18:40 Strontium: there's another highspeed pair done right: pins 29+59 of P16, surrounded by grounds Feb 19 02:19:25 you can use supplies too if you're careful. Feb 19 02:19:59 one hd camera block we're using actually has a co-axial multicore cable on it. Horrible thing, but i guess it works. Carries LVDS. Feb 19 02:20:29 have to be careful you don't break the cores, but the grounding is incorporated into the connector. Feb 19 02:20:37 KEL is the manufacturer iirc. Feb 19 02:24:19 zmatt: Actually checking the data sheet would make me think that stacking more than one board isn't feasible on an X15 because they have through hole mounts which would prevent a connector under it. Which kinda sucks. Feb 19 02:25:26 But on the upside its got big fat power pins. Feb 19 02:30:47 zmatt: I am looking at it because I want to make a "bare" expansion template schematic and board for the X15 for KiCad. Feb 19 02:32:15 veremit: you can reference signals to supply planes, but apart from P19 it's similar to the BBB: just a lot of IOs crammed together: http://gerbil.xs4all.nl/bbx15-expansion.png Feb 19 02:34:55 oh joy Feb 19 02:43:45 veremit: at least one of the video output ports has 22 ohm series resistors near the cpu Feb 19 02:47:39 zmatt: FX18-60S-0.8SV10 on mouser is $4.42 so thats better than $8.18 but its lower profile (i guess, cant find data on it). **** ENDING LOGGING AT Fri Feb 19 02:59:58 2016