**** BEGIN LOGGING AT Tue May 17 02:59:58 2016 May 17 03:53:31 userland running May 17 03:53:39 PoS i2c crap May 17 04:15:06 hi May 17 04:15:18 I need Help May 17 04:17:15 hi May 17 04:40:46 we can't help you here May 17 05:25:53 ds2: i2c itself is pretty okay (although there are two things that should have been done differently with 20/20 hindsight) ... most i2c controllers however suck :P May 17 05:30:09 zmatt: not that... the silly usage of I2C for system setup May 17 05:30:54 zmatt: all the kernel and uboot code is entangled with I2C crap May 17 05:33:09 ah... well the eeprom stuff in uboot can be easily ripped out of course, the pmic stuff is a bit trickier depending on your power management needs May 17 05:38:58 zmatt: the kernel side also does a bunch of eeprom stuff May 17 05:39:06 pmic stuff is fine May 17 05:39:31 I don't see the point of paying $$ to buy the I2C, paying $$ for a more expensive PCB, paying $$ to get it programmed when it can be fixed in SW May 17 05:40:31 I2C chip May 17 05:44:21 I'm quite sure the kernel does nothing with the eeprom on my system May 17 05:45:00 its existence is declared in DT but that's it, the kernel knows nothing about it other than "it's an eeprom" May 17 05:47:02 there's also nothing useful for it in there May 17 09:20:53 i have an element14 beaglebone black industrial (the red PB one) with a non working ethernet interface. the yellow led blinks weirdly at boot, and 2 boots out of 3 the ethernet communication does not work (the interface is UP but not LOWER_UP). can't seem to reproduce it. any hint? May 17 09:21:27 (when not working, the orange led stays mute) May 17 10:33:46 define "blinks weirdly" May 17 10:34:27 it is unfortunately typical for link to go up, down, up, down, up May 17 10:35:21 because the phy automatically negotiates a link after powering on, yet both u-boot and the kernel obnoxiously decide to reset the phy again May 17 10:36:24 and you say you have "2 boots out of 3" yet "can't seem to reproduce it" ? eh? May 17 10:37:31 in the non-working situation, check what happens when you pull the ethernet cable May 17 11:19:38 hi May 17 13:02:14 hi. anyone familiar with writing device trees and handling dma on bbb? May 17 13:03:34 i need to use mmc2 (the third mmc interface). but the device tree only defines dma for mmc0 and mmc1. i don't use mmc1 (it's disabled). could i just copy the dma config from mmc1 to mmc2? May 17 13:04:18 or does it require any extra config? or is dma channel 2 and 3, that mmc1 use in the device tree, specifically bound to mmc1 and not usable by mmc2? May 17 13:50:42 jkridner, if you dont mind me asking when will the beaglebone blue/strawson robotics cape be available? I have been waiting on this thing for more than a year now May 17 14:00:45 hello everyone! Is there any GNU/Linux distribution that comes by default with 4.6.x.x Linux kernel? May 17 14:01:00 Without needing to compile a custom one May 17 14:07:44 danbilzerian: wasnt 4.6 just released ? May 17 14:08:38 rob_w: I think so May 17 14:08:53 rob_w: Or any 4.x should be good May 17 14:09:21 any 4.x is quite possible May 17 14:09:55 rob_w: Do you know any distros that are using it? May 17 14:10:40 Usually ArchLinux uses the last kernel versions May 17 14:11:02 But I can't find anywhere the version specified May 17 14:11:23 i bet angstrom is already on a 4.x and so is poky May 17 14:12:12 as the official TI kernel is around 4.1.18 May 17 14:13:31 rob_w: thanks May 17 14:14:09 4.1.18 was released version ( together with their sdk) , i also see a 4.4.y branch so i guess that one is next May 17 14:18:05 I will first try the 4.1.18 version, I expect to obtain more throughput on UART peripherals with this newer kernel May 17 14:18:40 On my actual 3.8.13 is pretty slow May 17 15:28:07 gentoo May 17 16:39:24 hi guys May 17 16:39:34 I'm the person who questions related to the ADC with DMA. May 17 16:39:44 The problem has been solved, thank you answer to my question. May 17 19:40:45 Hello all. I am now almost convinced I have bricked my first beagle bone. I have a FTDI attached and nothing comes out when plugging power. Only the power LED goes on. The power button works a bit as keeping it pressed 5s will make the power LED go off. May 17 19:41:15 But nothing else apart from that. I am confident on my FTDI setup as I am able to use it on another BBB to login through serial. May 17 19:41:35 Anything else I should check before putting it in the bin ? May 17 19:47:46 sounds ok to me. If the power LED is steady its alive. May 17 19:48:52 you should see a string of 'CCCCCC' if the BBB cannot find any image to boot .. otherwise, try a known good uSD debian image, hold the BOOT button whilst you *power up* (not reset) the BBB and see if it comes to life then May 17 19:49:17 if the power LED were to blink instantaneously and extinguish .. then you need to RMA it :D May 17 19:56:22 The problem is that I do not have single character out of the debug port. Even when trying to boot on a uSD (with BOOT button) May 17 19:58:27 So it seems the BBB does not even start to look for an image to boot. it 'stuck' even before that May 17 20:05:37 Bobox: definitely turned flow control off in your serial terminal .. checked there is signal levels on the pins? not a connection issue? May 17 20:06:33 Unfortunately not. I just unplugged the serial cable and plug it on another BBB, and it work straight on. May 17 20:06:46 I need to look if there is signals levels on the pins May 17 20:06:48 you might have zapped the buffer chip on the serial lines if you're unlucky .. but it sounds as though the power management 'block' is active if you can turn it on/off with the button .. that "uses" the processor .. May 17 20:07:32 Yes, the power managment part seems to work, behaviour of turning on/off seems to be the normal one May 17 20:09:23 I need to look if i have properly turned flow control off May 17 20:11:09 arf, nothing more with screen -fn May 17 20:12:02 I Tend to use minicom .. so you can go into Ctrl-A/O to change the serial settings :) May 17 20:12:24 save as default/dfl and then restart to be sure May 17 20:12:35 but there are optionss :) May 17 20:12:49 I will try minicom May 17 20:24:00 * zmatt uses screen May 17 20:24:44 zmatt: thoughts on a very quite BBB?! May 17 20:24:51 quiet** May 17 20:25:39 sysboot having a weird value for some reason May 17 20:25:57 i.e. due to esd damage to one of the lcd_data pins May 17 20:25:59 yea that's about the only other thing ran through my mind .. unplug all capes/etc .. May 17 20:26:05 s/i\.e\./e.g./ May 17 20:26:27 of course a bbb should never be declared dead before removing anything plugged into its expansion headers May 17 20:28:41 also, check the voltages on the sysboot pins... on the P8 tab of my pins spreadsheet ( https://goo.gl/Jkcg0w ) the critical sysboot pins are marked in yellow, and right next to them is a column showing L in light green or H in light red May 17 20:28:54 those pins should have the logic levels indicated May 17 20:29:27 the most common way for a damaged I/O to manifest itself is an internal short to gnd or vdd May 17 20:31:37 if that is the cause you may be out of luck.. you can try checking for a usable bootmode compatible with the new stuck value and avoid driving it at runtime May 17 20:32:45 ^ sound advice May 17 20:34:59 there's also at least one person who managed to "repair" an am335x damaged in such a way by taking a serious power supply (2-3A) to that short circuit to internally obliterate it and managed to get it apparently functional again, minus some esd protection diode(s) May 17 20:35:20 Thanks both for the advices May 17 20:35:30 while an epic achievement, I would not recommend trying that unless the garbage bin is the next stop ;P May 17 20:35:31 I'll look at the sysboot pins May 17 20:35:51 I won't try that. I won't know what I am doing, and I don't like that :) May 17 20:36:55 what were you doing prior to "bricking" it ? May 17 20:37:43 I was trying to understand the output of a hall effect based wheel encoder May 17 20:38:58 So I think I probably damaged one of the GPIO pins, with perhaps too high voltage or something like that May 17 20:39:11 one concern I have is... long-pressing the power button causes it to power *off* you say, not power cycle? May 17 20:40:03 oh right with a serial cable attached that might actually make sense May 17 20:40:31 yes long-pressing the power button causes to power off (power led goes off). May 17 20:40:38 normally a bbb which has no external connections and which has not enabled the hdmi framer will power-cycle when you long-press the power button May 17 20:40:41 Even without a serial cable attached May 17 20:41:53 Need to press for 10 seconds May 17 20:42:02 power-off suggests there's more than usual power consumption while the system is held in reset, causing the pmic to go into temporary FAULT state and then OFF state one second later May 17 20:42:33 the hdmi framer is the usual cause since it completely unaware of reset (it has no reset pin) May 17 20:43:47 for comparison, this is a poweroff without hdmi framer enabled: http://elinux.org/File:Bbb-c-3v3u-shutdown-dc5v.png May 17 20:43:54 versus with it enabled: http://elinux.org/File:Bbb-c-3v3u-shutdown-dc5v-hdmi.png May 17 20:44:58 so that would suggest there's higher than normal current being drawn... maybe a "weak short", just not enough to trip the pmic right away May 17 20:45:20 (the power led blip of doom is when the pmic goes into fault state immediately at power on) May 17 20:46:05 then again maybe it's just random variation, consider how close call it is to avoid faulting at poweroff, so maybe this symptom means nothing May 17 20:47:00 still, the power button is managed entirely by the pmic... for it to work only the RTC logic of the am335x needs to work (which is in a special isolated power domain) May 17 20:47:34 e.g. dropping the bbb and busting the main oscillator crystal could also cause the state you're describing May 17 20:48:07 anything that keeps the SoC from working normally May 17 20:48:59 Thanks zmatt. Don't brainstorm too much for me, I am learning debugging and will come back when I have more information, particularly voltage informations May 17 20:49:14 sorry :) May 17 20:49:50 I won't thow the BBB to the bin soon, I'll try to find out what happened. In the meantime I have another one to play with, I'll try not to brick it also :) May 17 20:49:53 also check the voltages of the power rails ... 3.3V and 1.8V (ADC) on expansion header P9 May 17 20:50:12 3.3v and 5v are ok, I just checked May 17 20:50:17 there's also a little test point next to the power led, it should also measure 3.3V May 17 20:50:57 the 1.8V and the testpoint-3.3V are actually more important since those are actual processor supplies (unlike the 5v and the external 3.3V) May 17 20:54:02 ADC 1.8v is ok May 17 20:54:45 Is the testpoint the 'hole' named TP4 on the board ? May 17 20:54:53 yep May 17 20:55:45 it's actually the EEPROM write-protect input, but it's pulled up to the processor 3.3V (VDD_3V3A) May 17 20:56:02 it's the most convenient place to measure that voltage May 17 20:56:46 jkridner: yo May 17 20:56:53 3.3v on the test point May 17 20:57:15 Bobox: ok, so you have a perfectly healthy bbb... apart from the minor detail of not workign May 17 20:57:55 wassup? May 17 20:58:10 Bobox: were any of the pins you were using within the range of the sysboot pins? May 17 20:58:20 mistawright: you were waiting on the Robotics Cape ahead of the Blue coming up? May 17 20:58:23 I need to find out where those are May 17 20:58:42 jkridner: @ pocketbone.... routing usb over slots in the ground plane, tsk tsk tsk May 17 20:59:07 zmatt: yeah.... just one spot. It was a 1.5hr hack. May 17 20:59:26 spent longer on the board outline than the layout. May 17 20:59:57 just wanted to prove it works. just got my proto yesterday. May 17 20:59:59 jkridner: 1 major discontinuity on the left usb, two smaller ones on the right usb May 17 21:00:10 also, there's a trace... missing... May 17 21:00:17 I need to reduce the size of the reset pull-up resistor... not sure when I'll get a chance to do that. May 17 21:00:31 also, I didn't even power the microSD card.... wire-modded that. May 17 21:00:53 zmatt: I didn't have the net there before... had it named wrong. May 17 21:01:28 ah, I just noticed a stub coming out of a via going nowhere :) May 17 21:01:40 zmatt: yeah, I noticed that yesterday... not sure how I did that. May 17 21:01:58 probably meant to to go the unconnected futher left? May 17 21:02:00 must have been a ripup that I didn't finish. May 17 21:02:25 anyway, I've recruited some collaborators now and we'll move it to Kicad. May 17 21:02:26 Bobox: I linked to my spreadsheet :) May 17 21:03:22 jkridner: still, two-layer pcb is not exactly a friend to anything high-speed... unless you manage a perfect breakout on the top layer you need to seriously hack up your ground plane, and you did May 17 21:04:17 as some commenter (probably on hackaday? don't remember) said... good luck ever getting something like that through FCC testing ;) May 17 21:04:56 zmatt: What value should I read on the sysboot pins ? May 17 21:04:58 zmatt: its a prototype ffs :P lol May 17 21:05:03 zmatt: oh, yeah. I should have a power plane too to try to give a bit more shielding. May 17 21:05:44 veremit: no, it's a hack :) a "can this be done on 2 layers?" challenge May 17 21:05:44 I really just wanted a super simple proof of concept to show just how crappy a board can be. :-) May 17 21:05:49 that May 17 21:06:09 zmatt: well see .. :P May 17 21:06:18 the sip itself is loaded with decoupling caps. May 17 21:06:21 veremit: a prototype is usually a prototype *for* something May 17 21:06:37 jkridner clearly has too much time on his hands .. hahahaha :D May 17 21:06:57 jkridner, not sure what im waiting for. I just remember seeing a youtube clip about the robotics cape which you may have been in. And then i heard about the beaglebone blue. In fact the video was some wowee deal or wte that mip toy is called. May 17 21:06:59 it is a hack prototype of a prototyping hack. May 17 21:07:08 jkridner: nice :) May 17 21:07:18 jkridner: your clearances also creep me out a bit... on your PNG export some adjacent things actually seem to touch May 17 21:07:32 even in cases where it should have had plenty of room May 17 21:08:14 presumably that also reduces the qualification for "crappy pcb" May 17 21:08:16 mistawright: we setup a manufacturer for the cape, but I have had difficulty getting the switch turned on because of concerns the Blue would just obsolete it... I think we've gotten past that.... I really need more of the capes to help finish the Blue development. May 17 21:08:28 Bobox: I actually mentioned that May 17 21:08:38 Oops sorry then May 17 21:08:44 need to reread May 17 21:08:51 mistawright: The SeeedStudio BeagleBone Green and other distractions (a new baby) have really slowed down the Blue... I've recruited some extra help to get it done. May 17 21:08:55 Bobox: 40 mins ago exactly May 17 21:09:28 jkridner: one thing that bothers me about the osd335x is that they copied the bbb power scheme directly May 17 21:09:40 complete with issues May 17 21:10:07 jkridner, if you want a beta tester I'm more than willing ;) May 17 21:10:32 in a new design I would seriously look to avoid the tps65217 pmic entirely and check out one of the alternatives May 17 21:11:06 * jkridner goes afk for food May 17 21:11:14 jkridner, i dont find the green all too inticing, their newer board with wifi and bluetooth is a nice and caught my attention. The sancloud board i am more interested May 17 21:11:36 the enhanced is neat May 17 21:13:41 Agreed, the enhanced is much more interseting May 17 21:13:51 anyone see how far sancloud got? May 17 21:14:01 I know the kickstarter/etc closed .. May 17 21:14:05 indiegogo wtfever May 17 21:15:02 bbl May 17 21:17:25 veremit, from what i understand they are still producing the board May 17 21:17:41 zmatt: Yellow sysboot bins are at the correct level ( if 2.7v for High is correct). Thanks for the help, I'll continue looking at my board with your spreadsheet tomorrow May 17 21:17:59 so hopefully i will be able to buy one. I love flying with the bbb-mini cape and my beaglebone black. its almost like flying with snoopy May 17 21:18:55 zmatt, mistawright: the only thing is overcoming that damn musb silicon/driver issue .. and hopefully getting zmatt's power domain fixes implemented :D (possibly the replacement of that troublesome PMIC!) May 17 21:19:33 bo May 17 21:19:57 Bobox: it may be yeah, the pullups are very weak May 17 21:22:02 veremit: is musb actually still buggy, or are we repeating old lore? I don't use usb often enough to tell but I've never had much problems May 17 21:22:22 zmatt: I don't -think- they conquered *all* the bugs lol May 17 21:22:32 I've not built a >4 kernel for the BBB yet May 17 21:23:27 wait you mean you're still on a 3.something kernel? May 17 21:26:12 lol @ xkcd May 17 21:28:01 busy doin stuff with *other* arm boards .. haha May 17 21:28:06 stm32 is my latest fun .. May 17 21:28:17 oh boy May 17 21:28:45 "sane pinmux? what's that?" May 17 21:29:58 pinmux .. you don't need one of those :p May 17 21:30:11 its only got 64 pins anyway! May 17 21:30:12 :D May 17 21:30:59 assuming you don't actually need to use any significant number of the peripherals it has May 17 21:33:07 let alone with dma... May 17 21:34:11 "spi 1 and usart 3 ... YES! they don't pin-conflict! ... oh they share the same dma channels" May 17 21:34:59 and seriously EVERY TIMER IS DIFFERENT May 17 21:36:30 mind you I only have experience with the f100 and some indirect experience with the f107 May 17 21:36:47 dunno if the grass is greener on higher models May 17 21:39:05 f407 is what work's targetting most May 17 21:39:08 !@^#%!@$#^%!@# what the fuck?! I just discovered my last train ... doesn't exist. not even any mention of maintenance or problems or whatever, it's just not there in the schedule May 17 21:39:10 tis quite a nice processor May 17 21:39:23 zmatt: congrats :) on yer bike, mate! May 17 21:40:25 what the f is this May 17 21:40:55 I also woke up too late to go shopping so plan was to grab food on route to work May 17 21:41:50 oh wait May 17 21:41:51 phew May 17 21:42:06 I looked up the reverse direction accidently May 17 21:42:42 hhahaha May 17 21:42:47 go get another cofffee :P May 17 21:42:48 lol May 17 21:44:35 ok, time to kick myself out the door and grab that train May 17 21:44:50 hehe May 17 23:10:02 Quiet... May 17 23:13:37 haruma: comes and goes .. like the tide. May 17 23:14:21 Haha :-D May 17 23:18:01 back May 17 23:18:09 or, I will be once I made tea May 17 23:18:29 tea is good .. bring me one pls :D May 17 23:26:21 lol what, "Node.js bridge for COBOL" May 17 23:28:08 * veremit met jython today :/ May 17 23:40:48 https://github.com/IonicaBizau/node.cobol/blob/master/lib/node-exec.cbl May 17 23:41:16 I feel the urge to file a bug report that it doesn't properly escape the NODEJS_CODE string, potentially leading to injection attacks May 17 23:43:02 heh **** ENDING LOGGING AT Wed May 18 02:59:58 2016