**** BEGIN LOGGING AT Fri Apr 11 02:59:58 2014 Apr 11 03:03:22 Hi, I am planning a BBB application that will require dealing with pin IO and binary data: bit shifting and operating on byte arrays. I find that the javascript support for the latter is weak. Is there documentation on accessing IO pins of the BBB at a lower level than bonescript? Apr 11 03:07:37 Derek Malloy's videos look promising Apr 11 03:07:43 jaw22, yes try googing pasm Apr 11 03:09:14 jaw22, I am in contact with an MIT student who is writing assembly level code for the PRU cores in the BBB, his mastery of the platform is nothing less then magical. Apr 11 03:12:51 that's a little lower than I was aiming for... I'm thinking more like C/C++ or Go (with CGo calls). Hopefully Go isn't a taboo word in this channel Apr 11 03:18:53 from what I have seen, playing in the mud can build up some awesome foundations for future castles Apr 11 03:20:27 :) Apr 11 03:21:08 I agree, I'm not afraid to really dig in. But If I write code for the PRU cores, will I be able to get it into userspace? Apr 11 03:22:10 do you really need that level of speed or fine control over timing? Apr 11 03:23:03 I can post github references if you like, in the case of the HAM writing the code, yes he wrote a software R2R DAC that runs in the PRU hardware Apr 11 03:23:39 he managed to write several sets of code to date, including AM and DSBSC Apr 11 03:26:01 pjustice: no, I need basically no control over timing or speed. I just need access to IO pins to read analog values, and a type system that allows me to operate on raw binary, whereas Javascript appears to use IEEE floating point for all numbers, except when using bitwise operations, and I need deeper control than that Apr 11 03:38:46 seems relevant, though all the crazy about "use bonescript" at the top seems a bit nuts to me - http://www.element14.com/community/community/knode/single-board_computers/next-gen_beaglebone/blog/2013/10/10/bbb--beaglebone-black-io-library-for-c Apr 11 03:40:39 but the library seems quite simple, and shows you how to do pin i/o Apr 11 03:57:13 pjustice: thanks, it looks like that has evolved into https://github.com/VegetableAvenger/BBBIOlib Apr 11 04:02:16 yeah, that version is not so simple :) Apr 11 04:07:59 but the simpler one you linked doesn't support analog, apparently. the header file only lists is_high() and is_low() Apr 11 04:30:26 sure Apr 11 04:30:47 the point is mapping /dev/mem gets you access to the control registers; the simple version is an easy to understand version of that Apr 11 04:31:08 the trm would give you the details for analog Apr 11 04:35:11 trm? Apr 11 04:36:09 technical reference manual Apr 11 04:36:34 http://www.ti.com/lit/ug/spruh73j/spruh73j.pdf Apr 11 04:37:01 jaw22: sorry if this is a dumb question, but are you talking about being able to get analogue off all the gpio pins or only the 7 current analogue pins? Apr 11 04:53:44 just the analog pins :) Apr 11 04:59:47 jaw22: ahh yea ok, i thought you may have found some mythical way to convert the other gpio's as well. had my hopes up for a little bit ;) Apr 11 05:04:53 Dan1: there are always external ADCs. Usually they have multiple inputs and higher resolution than those built into general purpose microprocessors Apr 11 05:06:08 yea once i get finished with what im currently playing with im going to look into that sort of thing for other uses... at the moment my brains just about in overdrive just trying to get working what i want and how i want it Apr 11 05:10:17 Quick question: If I want to install debian onto a SD for a BeagleBoard Rev.C2, does the netinstall script on github cover all revisions of the board or do I have to do something specific? Apr 11 05:44:32 jaw Apr 11 05:44:42 im about to gpl a full featured ruby lib for the bb Apr 11 05:47:39 supports all the features and a nice interface Apr 11 05:47:43 prob going to post it up this weekend Apr 11 05:48:22 you got anything online about what your actually doing nyt-? blog or something yet Apr 11 05:51:15 nope i just wrote it this last week Apr 11 05:51:36 would have been done quicker had I not had to troubleshoot stupid adc kernel patches that broke adc functionality Apr 11 05:52:15 i'll pastebin my test file so far so you can see what using it looks like Apr 11 05:52:22 but im about to make it more oop friendly Apr 11 05:52:38 will function both ways in the end i reckon since some people are more happy with procedural Apr 11 05:53:48 http://pastebin.com/mH3zATDP Apr 11 05:57:29 nyt-, I do declare that ruby is even worse than javascript for what I need to do Apr 11 05:57:38 you can glean most of the functionality from there Apr 11 05:57:54 jaw22: how is ruby worse than js? works fine for handling binary data and bitshifting Apr 11 05:58:04 are you just not familiar? Apr 11 05:59:51 no, I guess not Apr 11 06:00:16 see what i pasted Apr 11 06:00:19 theres some binary operations in there Apr 11 06:00:30 for dealing with an mcp3008 over spi and some i2c stuff further down Apr 11 06:00:45 anyway was just mentioning it Apr 11 06:01:03 you can just access everything raw in c its not that difficult Apr 11 06:01:06 just ... a lot of code Apr 11 06:01:11 check out the adafruit python library Apr 11 06:01:17 they basically do everything in C Apr 11 06:01:28 and just push the data back up to python Apr 11 06:03:48 I highly appreciate that you have added BBB functionality in another language Apr 11 06:05:00 but if I were to use my favorite tool for this job (python), I would be using the wrong tool. I think ruby is not the right tool either Apr 11 06:05:39 what is the job? Apr 11 06:05:51 but yeah like i said, check the adafruit library if you want examples on how ot do things in c Apr 11 06:07:26 if you need really fast adc you'll have to write it in ASM and access the PRU but i think someone is working on a limited distribution c compiler to do that as well Apr 11 06:07:29 was mentioned earlier Apr 11 06:08:44 I think I'm going to rewrite BBBiolib. I like their approach of using the memory mappings. I do not like the file opening and reading in the Adafruit python library. That seems like unnecessary overhead Apr 11 06:08:52 jkridner: ^^^ Apr 11 06:08:57 [07:51] http://pastebin.com/mH3zATDP Apr 11 06:08:59 But the BBBiolib api is silly Apr 11 06:09:35 the ADC access doesn't need to be PRU fast Apr 11 06:09:36 nyt-: please ping jkridner or me when you have released it Apr 11 06:10:34 av500: sure, you have interest in it? Apr 11 06:12:07 of course, the platform has limited options right now Apr 11 06:12:47 I basically need to write the docs, gemify it, and add an oop interface Apr 11 06:12:52 but functionality at this time is complete Apr 11 06:13:23 drove up to my friends house to pickup an mcp3008 the other night he was using for his pi since they dont have an adc Apr 11 06:13:40 so was able to finish the spi codes Apr 11 06:13:49 which is quite interesting how the bbb does it Apr 11 06:13:57 not sure i like having only one dedicated CS pin per controller Apr 11 06:14:13 wonder if you can just leave it discon'd and use a gpio pin for CS instead Apr 11 06:14:16 dont see why not Apr 11 06:23:34 nyt-: personally no, but we have an interest to share it to a larger audience Apr 11 06:24:05 ah ok Apr 11 06:24:11 i still have no idea who anyone is or what their involvement is Apr 11 06:24:44 i see youre opped but that doesn't tell me much ;) Apr 11 06:25:01 I've been here for years, and I don't know. Don't worry. Apr 11 06:48:42 nyt-: av500 is the second biggest oss developer known to men Apr 11 06:48:52 nyt-: and jkridner is beagle mom Apr 11 07:06:48 lol whos the first Apr 11 07:54:11 oh, PS nyt- the application is a true random number generator, which means that I need to access the ADC and do linear algebra in the same application xD Apr 11 07:56:27 and make http calls Apr 11 07:57:50 I am beginning to think that python actually *is* the right tool after all, due to numpy and urllib2 Apr 11 07:59:56 nyt-: gkh :) Apr 11 08:00:32 jaw22: python is becomming like xml Apr 11 08:01:28 KotH: what? what do you mean? Apr 11 08:04:31 jaw22: python is getting applied to all kind of problems, even those it has never been designed for Apr 11 08:05:17 jaw22: doing numbercrunching with a scripting language was considered among the most stupid things you could do as a programmer, just 5-6 years ago Apr 11 08:07:15 python was designed to be general-purpose, for one thing, and 2, modern python implementations are very fast... Apr 11 08:07:58 linear algebra is a difficult thing. LAPACK is insane. A high-level solution makes sense, especially for an application where performance isn't critical Apr 11 08:09:51 although if I was going to use the Adafruit BBIO library, I would have to use the thrashing of opening and closing files to read the ADC values instead of memory mapping, so that really turns me off Apr 11 08:11:13 linear algebra is easy as long as you stick just to vectors and matrices Apr 11 08:12:02 and python is a general purpose _scripting_language_ Apr 11 08:12:08 the distinction is important Apr 11 08:12:25 and just because people use it for all kind of shit does not mean it's intelligent to use it for all kind of shit Apr 11 08:15:55 so, what tool do you suggest I use? Apr 11 08:16:12 for reading the ADC, linear algebra, and http Apr 11 08:36:05 KotH? Apr 11 08:38:35 hai Apr 11 08:39:51 erm, hi.. Apr 11 08:51:37 jaw22: what does that whole thing do? Apr 11 08:53:09 jaw22: you would use a separate http server for a start, so that is irrelevant; and the "one true language" for the rest if you want performance Apr 11 08:53:51 ah sorry, you mean make http calls, like POSTS and GETS? Apr 11 08:54:14 in which case, still the one true language Apr 11 08:59:14 it's to gather entropy, whiten it, and POST it to a remote machine Apr 11 09:02:13 datarate? Apr 11 09:02:34 and you know that sending random data over the network is kind of stupid? Apr 11 09:04:59 if you want true entropy, just use some recorded speech of $POLITICIAN Apr 11 09:05:24 KotH, why is it stupid? Apr 11 09:17:14 jaw22: the whole point of random data is to have something that nobody else knows about Apr 11 09:17:21 sending it over the network defeats that purpose Apr 11 09:20:29 bored~! Apr 11 09:21:01 ..... Apr 11 09:21:19 I agree with your first statement Apr 11 09:21:25 however Apr 11 09:21:35 I am not a moron Apr 11 09:21:52 and I disagree with your second statement Apr 11 09:22:21 you can always encrypt the random data :) Apr 11 09:22:25 I plan to use ecsda Apr 11 09:22:29 ^^ exactly Apr 11 09:23:21 use a random key Apr 11 09:23:24 oops, I mean ecdsa Apr 11 09:25:02 only problem is, I can't figure out the TSC_ADC FIFOs Apr 11 10:06:20 any horny girls is there? Apr 11 10:11:24 av500: danke Apr 11 10:33:20 after reading the TRM and https://github.com/VegetableAvenger/BBBIOlib/blob/master/BBBio_lib/BBBiolib_ADCTSC.c#327 I have come to the conclusion that when the ADC FIFO registers are read, they automatically populate with the next ADC readings... that seems like voodoo... is that possible? Apr 11 10:41:37 Hi Apr 11 10:45:56 jaw22: a lot of adcs do something like that, so its certainly possible Apr 11 10:46:11 jaw22: probably you just should look at the IIO driver for them. Apr 11 10:58:12 I'm looking for information about drive servomotors with the beaglebone black in c++, basically I need to know how to use the PWM, Anyone knows any web page of this topic? Apr 11 10:58:50 https://www.google.com/search?q=beaglebone+pwm Apr 11 10:59:50 isn't there some finished cape for you as a template? Apr 11 11:58:43 LetoThe2nd: I don't think you can determine from C when a memory address is read, I think that would have to be built into the chip Apr 11 11:59:37 jaw22: read again. i specifically said that the adc does it. Apr 11 12:00:00 jaw22: the reference to iio was more meant as an example on how to use it correctly. Apr 11 12:01:11 ah, ok Apr 11 12:12:41 Don't buy anything from Farnell today, they still haven't patch for heartbleed Apr 11 13:39:50 Hi, I have a strange problem with my BBB, when I try to use GPIO pins, my BBB become laggy/crash. The heartbeat led become irregular and the command line become very slow. Any idea ? :/ Apr 11 13:40:20 use how? Apr 11 13:42:19 I make a scheme, 2mn Apr 11 13:45:04 I do that : http://image.noelshack.com/fichiers/2014/15/1397223885-img.png Apr 11 13:47:34 I didn't send any command to my BBB it becomes unstable before Apr 11 13:48:25 Tork: the power strip lines are all connected no? Apr 11 13:48:30 on your breaboard Apr 11 13:48:32 bread Apr 11 13:48:56 you shorted pin 1 to pin 45 Apr 11 13:49:27 Oops, It's a mistake on the scheme, I did'nt do that with my breadboard :p Apr 11 13:50:01 maybe a photo of the real thing... Apr 11 13:50:28 Tork: also, i dont know if the BBB can drive enough current to get your LED to emit anything Apr 11 13:50:51 Tork: if you need more than 1mA for the led, it's not going to happen. or worst case, you might fry the processor Apr 11 13:51:21 Koth -> Ok, Thx :) Apr 11 13:51:36 Tork: the TRM or datasheet says how much current a GPIO pin can deliver Apr 11 13:51:39 Tork: read it Apr 11 13:52:17 Tork: oh: and just in case: absolute maximum ratings are exactly that: excede those by even a tiny bit, and your BBB is a brick Apr 11 13:52:30 Tork: never exceed the recomended operating conditions (which are much lower) Apr 11 13:52:31 limit is 4ma or so Apr 11 13:52:32 isnt pin 1 and pin 45 both grounds on a BBB? Apr 11 13:52:38 low current leds should work Apr 11 13:52:42 standard 20mA led, no Apr 11 13:53:40 on the p9 header Apr 11 13:53:52 or is that pic just wrong? Apr 11 13:54:16 (11:15:11 PM) Tork: I do that : http://image.noelshack.com/fichiers/2014/15/1397223885-img.png <------- Apr 11 13:55:53 Hi! Apr 11 13:56:05 I don't understand why in this case it works http://learn.adafruit.com/downloads/pdf/blinking-an-led-with-beaglebone-black.pdf Apr 11 13:56:36 How can I specify GPIO base number for i2c based IO expander via device tree? Apr 11 13:56:48 panto: ^^^ Apr 11 13:56:56 By default the expander gets 496 on am335x Apr 11 13:57:39 you don't Apr 11 13:57:49 you use a DT gpio chip reference Apr 11 13:58:31 off the top of my head something like: pin = <&i2c-expander 5 0>l Apr 11 13:58:32 *; Apr 11 13:58:33 Tork: low current leds Apr 11 13:58:44 Tork: if your led is designed to be drivven by <4mA then it works Apr 11 13:58:44 where 5 is the index of the gpio on the expander Apr 11 13:59:26 Tork: but be advises: it's not that the io pins can only delvier 4mA. it's rather that you are not allowed to draw more than 4mA. ie if you draw more than 4mA per pin (and a total maximum which is mentioned somewhere) then you will fry the CPU Apr 11 13:59:49 Koth oki thx Apr 11 14:00:45 panto: Documentation/devicetree/bindings/gpio/gpio.txt talks about gpio-ranges, but it looks wired so far Apr 11 14:02:58 what do you mean wired? Apr 11 14:05:26 Hi all, I am looking for a raspberry pi replacement. I have several usb power issues with the pi. I was wondering if anyone could steer me to the techincal specs on usb for the beagleboard. How much current/etc it can handle. design specs. etc. Any help would be appreciate. Apr 11 14:05:39 ioudas: beagleboard.org? Apr 11 14:05:54 having a hard time finding it on that Apr 11 14:06:08 or http://beagleboard.org/black and http://eLinux.org/Beagleboard:BeagleBoneBlack Apr 11 14:06:52 you're better off powering multiple devices with a powered USB hub Apr 11 14:09:44 panto: I've read it once more and I think I understand it now. They mean it for SoC's internal GPIO controller, where it makes sense to specify, what GPIO is routed to what pin specified by pinctrl Apr 11 14:10:09 panto: trying to get info from Documentation/devicetree/bindings/gpio/gpio-pcf857x.txt Apr 11 14:10:09 right Apr 11 14:10:22 it's not biggie Apr 11 14:10:31 panto: which file describes pin = <&i2c-expander 5 0>l? Apr 11 14:10:43 err, >l is an error Apr 11 14:10:48 it should be >; Apr 11 14:10:58 panto: I also would like to know, where 496 as base comes from? Apr 11 14:11:04 the same one Apr 11 14:11:41 from the implementation of the board Apr 11 14:11:44 is that beaglebone? Apr 11 14:12:14 IIRC the SoC GPIOs start numbering from 0, and the others from a high integer Apr 11 14:12:37 * ioudas hmmms Apr 11 14:12:40 cant seem to find it Apr 11 14:13:00 panto: it is a custom board based on am335x Apr 11 14:13:09 panto: with tca6416 Apr 11 14:13:35 the gpio expander when it loads, registers it's own range of gpios Apr 11 14:13:52 where those gpios are located is not defined Apr 11 14:14:07 there's some kind of allocation logic in gpiolib Apr 11 14:14:41 my point is that if you're using DT, you should not care about the global number, only about the index of the gpio in the gpio controller Apr 11 14:14:49 panto: that's bad. In 3.2 there was something like MAX_GPIO for am335x in the board file and the gpio expnader was assigned MAX_GPIO as base number Apr 11 14:15:02 why is it bad? Apr 11 14:15:31 if you make assumption about the gpio # you're getting from gpiolib you're asking for trouble Apr 11 14:15:35 *assumptions Apr 11 14:16:43 do you know https://github.com/jackmitch/libsoc? It has its MAX_GPIO at 256, so with my 496 I hit this at once :-) Apr 11 14:18:15 yegorich, I'm not familiar with it, but I know the developer is a reasonable guy Apr 11 14:18:21 he's going to accept patches :) Apr 11 14:18:36 panto: I just would like to have a chance to define an gpio expander in DT and say, that it's base number should follow direct after the referenced GPIO controller, like a linked list Apr 11 14:18:53 you don't get to decide Apr 11 14:19:04 panto: I've already contributed about 3 patches for libsoc Apr 11 14:20:45 numerical namespace api is a deadend, you should change thinking to using the modern gpiod_* apis Apr 11 14:20:59 and forget about the broken numbering system Apr 11 14:21:15 mdp: do you have a link? Apr 11 14:21:47 include/linux/gpio/consumer.h Apr 11 14:24:17 mdp: is it suitable for userspace? Apr 11 14:24:32 not that I'm aware of Apr 11 14:25:03 user-space should not deal with number anyway Apr 11 14:25:20 the board should set up symbolic names that correspond to the schematics and then map Apr 11 14:25:54 panto: symbolic links in /sys/? Apr 11 14:26:02 yegorich, https://www.kernel.org/doc/Documentation/gpio/sysfs.txt Apr 11 14:26:15 yegorich, the gpiod_* api can exports to userspace as well Apr 11 14:28:15 the sysfs api needs your patches to update it to sanity..unique label so you can identify your gpio controller and get rid of the integer nonsense in the node name Apr 11 14:30:48 mdp: OK. I understand Apr 11 14:31:43 coupled with the descriptor based approach, a sane userspace api would have a unique label per instance of control and integer identifier only matters within that one controller instance namespace Apr 11 14:32:14 and that unique label is already sitting there in the kernel anyway Apr 11 14:34:05 mdp: something like this http://www.spinics.net/lists/devicetree/msg07894.html ? Apr 11 14:35:49 mdp: seems like 3.15 still doesn't have this functionality :-( Apr 11 14:36:36 yegorich, yeah, I was hoping to inspire you ;) Apr 11 14:36:59 yegorich, it will be a huge improvement and remove so many product-specific hacks Apr 11 14:38:14 mdp: oh, yes! Apr 11 14:45:07 mdp: This was the last patch version (3) https://www.mail-archive.com/linux-gpio@vger.kernel.org/msg00864.html Apr 11 14:46:32 yegorich, yeah, this was timely cause I was wondering what happened to it last week and forgot to check Apr 11 14:46:49 yegorich, whilst hacking around something here in a similar manner Apr 11 14:47:03 maybe one of us will come up with some time to revive it ;) Apr 11 14:47:17 mdp: it would be great Apr 11 14:47:36 mdp: I'll put it on my TODO list Apr 11 14:48:02 how long is that list? I wonder if it's like mine where it's sorta like putting it in the trash ;) Apr 11 14:48:15 mdp: have to finish some real life projects first :-) As usual :-) Apr 11 14:48:31 same story here Apr 11 14:48:38 mdp: something like that :-) Apr 11 15:15:18 I see a lot of material about sampling the ADC in continuous mode, eg http://beagleboard-gsoc13.blogspot.de/2013/07/sampling-analogue-signals-using-adc-on.html Apr 11 15:16:02 what's wrong with mmap'ing /sys/bus/iio/devices/iio\:device0/in_voltage0_raw and sampling whenver I like? Apr 11 15:16:16 whenever* Apr 11 15:17:51 Hi, someone has try to connect BBB with a capacitive touchscreen based on ft5306 controller Apr 11 15:17:57 ? Apr 11 15:18:50 I have some problem to find a driver for i2c capacitive control Apr 11 15:19:54 jaw22: because its one shot sampling mode by default using that interface and limited to about 20k/s on a good day Apr 11 15:22:07 twenty thousand samples per second? that's plenty for my application Apr 11 15:23:21 Does the BBB require a high speed USB connection? I can't seem to get to work through full speed link. The F: drive pops up but is inaccessible. Apr 11 15:23:40 I know it works when directly connecting it to the PC (aka high speed). Apr 11 16:00:10 Has anyone run into a problem of running C++ threads on the beaglebone black running angstrom? From reading online I believe it's something to do with linaro toolchain using the wrong pthread library, but I don't know what to do about it. Apr 11 16:23:07 cinix: have you tried building natively? Apr 11 16:26:30 That's the only way I have tried Apr 11 16:26:49 I just got my beagle bone black and I'm not able to get the GUI to show in chrome/firefox. Apr 11 16:27:03 any suggestions? Apr 11 16:27:04 I actually just found (sort of) the answer, more of a work around. Adding these compile flags fixes it: -D__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 -D__GCC_HAVE_SYNC_COMPARE_AND_SHAVE_SYNC_COMPARE_AND_SWAP_4 -D__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 Apr 11 16:27:56 This is where I read about it: https://stackoverflow.com/questions/16171401/why-does-this-simple-c11-threading-example-fail-when-compiled-with-clang-3-2 Apr 11 16:28:03 Nevermind, got it working Apr 11 16:31:34 jkridner: hmm you think they will make it? https://www.indiegogo.com/projects/arduino-shield-to-beaglebone-black-cape--2 Apr 11 16:31:53 doesn't look like they will get the funding to me. Apr 11 16:32:09 I threw a banner up on beagleboard.org today to try to get them a bit more attention... Apr 11 16:32:18 I'm not seeing people click on it... Apr 11 16:32:34 they mostly scroll past directly to the "buy" links. Apr 11 16:32:41 jkridner: that's a good thing to do... but that's too bad Apr 11 16:34:46 I think they need to find a distributor to buy enough to pay for the campaign. Apr 11 16:35:03 pdp7: know of any distributors who'd want a bunch of these? Apr 11 16:40:18 * pdp7 looks at the periodic table Apr 11 16:44:39 BBB post on hackaday: http://hackaday.com/2014/04/11/beaglebone-black-and-fpga-driven-led-wall/ Apr 11 16:53:38 Hi all, I am looking for a raspberry pi replacement. I have several usb power issues with the pi. I was wondering if anyone could steer me to the techincal specs on usb for the beagleboard. How much current/etc it can handle. design specs. etc. Any help would be appreciated. I cant seem to find them on the beagle board site. Apr 11 17:03:22 ioudas: good queetsion, i'm trying to think of where i've seen that inof Apr 11 17:03:25 *info Apr 11 17:04:26 page 78 of the Tech Ref Manual has a schematic of the USB Host Apr 11 17:04:41 https://github.com/CircuitCo/BeagleBone-Black/blob/master/BBB_SRM.pdf?raw=true Apr 11 17:05:08 from http://elinux.org/Beagleboard:BeagleBoneBlack#Hardware_Files Apr 11 17:06:02 there is a power mgmt chip unlike Pi so it handles power more manua Apr 11 17:06:43 interesting Apr 11 17:06:50 * ioudas looks Apr 11 17:07:09 yeah, i have a lot of pi issues with usb. we are running a symbol scanner and wifi and the power issues are insane Apr 11 17:07:22 so USB port VBUS comes from TPS2051 http://www.ti.com/product/tps2051b Apr 11 17:07:28 yeah, me too Apr 11 17:07:43 cus there is no power mgmt on pi, its really hit or miss Apr 11 17:07:59 yeah Apr 11 17:08:31 so i'm thinking current for the USB host port is limited by the TPS2051 Apr 11 17:08:33 you post in their forums... you get nailed to the wall for suggesting the usb power issues are real. and banned if you arent bending over.. i want something i know will work Apr 11 17:08:39 to how much Apr 11 17:08:41 * ioudas is reading Apr 11 17:08:56 how are you powering the pi? Apr 11 17:09:41 i power the pi directly from the wall, tried a 2a and 1a charger. and some phone chargers... same story. devices are powered via a 2a hub Apr 11 17:09:59 keyboard, wifi and symbol scanner are on hub.. symbol scanner has its own powered wedge Apr 11 17:10:51 i think you'll find the results are better with open source hardware engineered for a constraint beyond making the cheapest possible computer :) Apr 11 17:11:14 i would agree. just curious on maximum output Apr 11 17:11:37 yeah, datasheet for that chip shows 500mA continous Apr 11 17:11:37 its interesting though, i put a volt meter across the wedge... the thing sucks up 1.78amp. during scanning the delta is 200ma Apr 11 17:11:44 looks like it can handle spikes Apr 11 17:11:46 interesting Apr 11 17:11:55 i'm not 100% sure though Apr 11 17:12:05 jkridner: do you know max current for the host port? Apr 11 17:12:32 if not, the beagleboard list is the goto for tech questions usually if IRC doesn't work out Apr 11 17:13:02 rgr Apr 11 17:13:06 ioudas: https://groups.google.com/forum/#!categories/beagleboard if you've not seen Apr 11 17:13:08 There isn't much of a drop from the input Apr 11 17:13:39 within reason, you should be able to pass on current from the input fairly easily Apr 11 17:14:58 jkridner: so a device on the host port should be able to pull more than 500mA? Don't know much about the TPS2051, but taking a look at the datasheet Apr 11 17:15:02 i guess doesnt hurt to order one to test. Apr 11 17:15:07 :) Apr 11 17:15:26 i have all the other devices. see if it works. Apr 11 17:15:37 but the scanner wedge portion blows my mind in terms of amperage. Apr 11 17:15:53 im waiting on another meter to confirm Apr 11 17:15:53 ya, that's a lot to source from the USB port Apr 11 17:16:11 the meter i tested with... makes me think it may be wrong Apr 11 17:16:21 there are those Y adapter cables to combine current from multiple ports Apr 11 17:16:51 like 500mA + 500mA to get 1A Apr 11 17:16:57 pdp7: yes, that's my understanding from Gerald, when powered through the barrel. Apr 11 17:17:04 jkridner: cool, thanks Apr 11 17:18:52 the devices are a normal 100ma keyboard (dell 5 dollar special) an atheros card (netgear n150) and this symbol scanner adapter. http://barcodescannersdiscount.com/noname4.html and a ls 4004i scanner. (power wedge for scanner is dc 5v 1amp) which i have in place. Apr 11 17:19:03 so i think 500 vs a 140 on the pi Apr 11 17:19:07 will be interesting to test Apr 11 17:19:27 im actually building a wyse 60 terminal.... the code is done, the only stuff we have problems with is the pi. Apr 11 17:20:48 wyse 60 terminal... like this is for point of sale? Apr 11 17:21:09 yea, kind of. for a company Apr 11 17:21:25 the code works beatifully. everything works.. except the pi just locks up during use Apr 11 17:21:33 you can remove a wifi... bam locks up Apr 11 17:21:36 etc. Apr 11 17:21:41 dvi resets. Apr 11 17:22:34 that's an interesting application, i like to see SBCs like the BBB being used in commercial/industrial applications to replace old computer equipment Apr 11 17:22:46 hopefully you'll have better luck with BBB Apr 11 17:23:13 i can say that i did many projects with Pi and BBB, and I've never had power issues with BBB unlike Pi Apr 11 17:23:26 or USB issues Apr 11 17:23:48 this is my favorite powered hub for any system Apr 11 17:23:49 http://www.adafruit.com/products/961 Apr 11 17:25:30 http://upload.wikimedia.org/wikipedia/en/thumb/8/8b/NoPC_Wyse.png/150px-NoPC_Wyse.png Apr 11 17:25:38 The "no-PC" logo used by Wyse to promote thin clients[17] Apr 11 17:25:44 i actually use that hub Apr 11 17:25:53 hey, single board computers like BBB are making that dream a reality now :) Apr 11 17:26:02 ioudas: ah, cool Apr 11 17:26:02 the code is written in c Apr 11 17:26:07 it can compile on anything really Apr 11 17:26:14 that's good Apr 11 17:26:43 Hi, I was wondering if anyone could tell me if Beagleboard Rev D works with simulink? Apr 11 17:26:44 what i did was stick a pi in a http://www.hubbell-wiegmann.com/ nema rated junction box Apr 11 17:27:04 ah ok Apr 11 17:27:05 put the stuff in there to hold down wires.... and hooked up a monitor Apr 11 17:27:24 its really slick... literally you cant buy the wyse terminals... the company i work for has a aix cobol back end Apr 11 17:27:32 they use the wyse emulation for screens to load product Apr 11 17:27:44 and do inventory type things... we are going to roll out 60 units Apr 11 17:27:50 brb, going to buy a real multimeter Apr 11 17:28:11 hah, my company also has an old IBM backend behind all the modern web stuff Apr 11 17:28:44 ioudas: very interesting use case for bbb... kinda like replacing old PCs running CNC machines Apr 11 17:29:32 PaulRainey: sorry, i've not used simulink... hopefully other here have Apr 11 17:41:36 How do I enable the built-in remote desktop in Ubuntu from ssh? http://www.makeuseof.com/tag/ubuntu-remote-desktop-builtin-vnc-compatible-dead-easy/ - Everything I am finding is people wanting to install 3rd party solutions. Apr 11 17:44:48 stop using those non compliant USB devices Apr 11 19:07:05 pdp7, got some new news on that pi. Apr 11 19:07:10 you around Apr 11 19:12:52 ioudas: howdy, whats the news Apr 11 19:14:26 the pi's with a wna n 150 netgear will freeze upon ap disconnection or removal entirely. similar cards will not produce a freeze of the os. Did manage to measure voltage across its main. was lower than i wanted, but problem occours with a 2amp 5v charger. The n150 netgear just will instantly freeze all io. Apr 11 19:14:55 ah Apr 11 19:15:14 no kern log, not able to dmesg or see anything on console. just bam froze Apr 11 19:15:27 interesting, with the pi, there is tp1 and tp2 test points that show the 5v rail... i would often see that being below 4.75V when having trouble Apr 11 19:15:32 ah dang Apr 11 19:15:38 that's a tough one Apr 11 19:15:41 yeah Apr 11 19:16:05 with the 2.1 amp charger the tp1 and tp2 run at 4.95. disconnection of wifi will cause a .001 delta in voltage. Apr 11 19:16:12 i do know that the IP block for the USB OTG on the Pi's processor (BCM2835) is not that great Apr 11 19:16:27 i think it is source of lot of finicky behavior Apr 11 19:16:56 of course, kinda of what happens when using a media accelerator chip as a general purpose GNU/Linux computer Apr 11 19:17:30 i think if i did a bbb a lot of these power/odd issues may just vanish. Apr 11 19:17:36 i will most likely have the company order one Apr 11 19:17:48 great Apr 11 19:17:59 so we will see Apr 11 19:18:20 problem is, the other cards, do a deauth reason code 3 with the ap's compared to my atheros wna card. Apr 11 19:18:26 then wont rejoin Apr 11 19:18:26 super, it can be a bit tough to find in stock... the beagleboard mailing list has a good pulse on the situation Apr 11 19:18:29 until reboot Apr 11 19:19:10 ah, so the ones that don't freeze the pi end up being disconnected, eh? Apr 11 19:19:21 that's no fun Apr 11 19:19:33 teah Apr 11 19:19:34 yeah* Apr 11 19:19:53 re: order, for today, specialcomp has in stock: http://specialcomp.com/beaglebone/ Apr 11 19:20:37 Adafruit gets weekly shipment, 1 per customer, so sign up to be notified: https://www.adafruit.com/products/1278 Apr 11 19:20:56 rgr Apr 11 19:20:59 thank you Apr 11 19:21:02 Newark element14: http://www.newark.com/element14/bbone-black/dev-board-am3358-mpu/dp/41X3868?ost=bbone+black+embest Apr 11 19:21:09 (an authorized clone) Apr 11 19:21:14 sure, np Apr 11 19:21:33 is there a limitation for these in commercial settings such as business uses Apr 11 19:21:35 point of sale Apr 11 19:21:40 your first like said non commercial use Apr 11 19:21:52 no, it just won't be supported Apr 11 19:22:29 you can engage with the manufacturers, CircuitCo or element14/embest, directly for commerical arrangments Apr 11 19:22:35 ah Apr 11 19:22:41 gotcha Apr 11 19:23:13 its basically gerald who does the HW for BBB doesn't want to get back 100 devices RMA'd cus they were in some companies product and failed Apr 11 19:23:50 also, one reason for stock shortage is that companies embedding in products have been sucking up stock from distributors Apr 11 19:24:05 small scale: buy where ever Apr 11 19:24:19 large scale: engage with CircutiCo, or now also, element14 Apr 11 19:24:57 bbiab, lunchtime! \o/ Apr 11 20:49:41 Is the default desktop sharing Vino server in the BBB Ubuntu image? (http://youtu.be/f9F3EE2jCPo?t=34s) - Is there a way to enable it from a ssh terminal? I don't have a mini hdmi to hdmi adapter at the moment so I can't check the GUI like in the video Apr 11 21:21:20 Greetings. Anyone know anything about these Element14 built beaglebone blacks? Apr 11 21:21:46 Should I wait for a "real" one instead? Apr 11 21:22:14 It seems that they are only guranteed for the first party stuff. Angstrom. Apr 11 22:19:13 does anybody have completed beagle projects they'd like to demo? Apr 11 22:34:54 jaw22: demo online? Apr 11 22:35:43 dsmith-work: i have an element14/embest bbb (bought from newark element14 in USA) Apr 11 22:36:09 dsmith-work: it is same functionality as CircuitCo BBB rev B (the most recent Rev) Apr 11 22:36:29 dsmith-work: all software works the same on it Apr 11 22:38:16 dsmith-work: my gallery if you're interested https://plus.google.com/photos/117542001281850354871/albums/6000002045485150529?banner=pwa Apr 11 22:39:44 pdp7: Cool, thanks. Apr 11 22:40:43 any created a project page on beagleboard.org? Apr 11 22:40:49 I added this beagleboard.org/project/bbb-bicolor-matrix Apr 11 22:40:51 errr Apr 11 22:40:55 http://beagleboard.org/project/bbb-bicolor-matrix Apr 11 22:41:08 but would like to upload images, not sure if it supports hosted media Apr 11 22:41:23 like wordpress or wikimedia would Apr 11 22:41:59 it does let you upload 1 image as the thumbnail Apr 11 22:45:48 yeah demo online Apr 11 22:49:19 jaw22: like g+ hangout? Apr 11 23:01:56 is there a reason one would use debian armel over armhf? Apr 11 23:06:11 dogrocket: i thought there was some java issue Apr 11 23:07:08 Hello, I had a quick question about the latest images (debian for the BB/B). I upgraded and I noticed I can't connect to the bone via usb/ethernet. Has the address changed? Apr 11 23:07:21 ok... well i need armhf :/ Apr 11 23:08:48 dogrocket: ya, i can't think of a reason actually... there was some historical issue with images avail for pi but that got nothing to do with beagle Apr 11 23:09:02 i think armhf was always the leader for Java in general on ARM Apr 11 23:09:57 jg_: is it this one? http://debian.beagleboard.org/images/bone-debian-7.4-2014-03-27-2gb.img.xz Apr 11 23:10:07 Yes Apr 11 23:10:42 I've not seen anything about address changing... i believe i have the 3-27 at home on bbb but would need to double check Apr 11 23:10:50 usb should be 192.168.7.2 Apr 11 23:11:06 ethernet should be set to dhcp, so whatever your local network does Apr 11 23:11:13 I have both a white and black, the white is running the new debian image but black is on angstrom still Apr 11 23:11:48 I'm using 192.168.7.2, works with angstrom but not on the white (debian) Apr 11 23:11:49 ah, gotcha, interesting... been awhile since i've used white. i think it has a usb to serial adapter so you can get the serial console easily Apr 11 23:12:34 have you used the serial port over USB before? Apr 11 23:12:42 I have Apr 11 23:12:44 it would be easy way to see output of ifconfig Apr 11 23:13:11 iirc, white had that built in versus needing an ftdi cable on black Apr 11 23:13:19 I was just seeing how the usb/ethernet works, normally I'm powered with a ethernet Apr 11 23:13:24 ah Apr 11 23:13:31 oh good point! Apr 11 23:13:54 I'll try that, if I can't get it working I'll flash my BBB with debian to compare Apr 11 23:14:00 i think i'll have to try my white this weekend... its kinda been sad lately Apr 11 23:14:20 but i was thinking about how the new debian beagleboard.org image would run ok for both Apr 11 23:14:30 well the non-emmc flasher, that is Apr 11 23:14:32 same here, that's I forced myself to use it with debian... it's been sitting around collecting too much dust :D Apr 11 23:14:53 other than the usb/ethernet it's running great! Apr 11 23:15:00 I'm glad the switch was made Apr 11 23:15:30 cool... i could see that maybe being something not many folks have tested... bbw with debian 3-27 beagleboard.org Apr 11 23:15:38 ya! Apr 11 23:15:42 well, I'm off for now. I'll report back later when I've played around a bit Apr 11 23:15:58 me too... it was this weird bizzaro land for me without apt-get or yum :) Apr 11 23:16:09 cool, cya Apr 11 23:42:53 i need to go and buy a door Apr 12 00:51:57 when he buys a door does he sell a window? Apr 12 00:54:55 in deed. **** ENDING LOGGING AT Sat Apr 12 02:59:59 2014