**** BEGIN LOGGING AT Tue Feb 05 03:00:01 2019 Feb 05 03:53:54 Is anyone having trouble updating and upgrading their BBBs? Feb 05 04:02:15 It seems connecting to the repos is taking a bit of time. Feb 05 09:41:11 zmatt: I hope you received my thank-you message yesterday. The LCD cape is now working fine! Thanks again! Feb 05 09:47:40 glad to hear Feb 05 11:20:27 is there a windows usb driver for beagleboard x15? Feb 05 11:22:50 ahnkle: what do you mean? driver for what? Feb 05 11:23:20 no driver should be needed Feb 05 11:24:22 i see a usb com port but it doesnt work: no data received Feb 05 11:24:30 (assuming the standard bbx15 image sets up the usb device port for similar purposes as the standard beaglebone image... I'm not actually sure of that, it's also quite possible the usb device port has no default functions at all) Feb 05 11:24:39 which port are you referring to? Feb 05 11:25:07 never mind, it only has one Feb 05 11:25:20 (I was mixing it up with another dev board) Feb 05 11:26:05 i power up x15, connect it to windows 10 with usb cable; then i get a serial port and a (non-working) CDC ECM device Feb 05 11:26:20 no rndis? Feb 05 11:27:58 iirc a beaglebone with a current image shows up as four devices: mass storage, serial (CDC-ACM), usb-standard networking (CDC-ECM), and microsoft-proprietary networking (RNDIS) Feb 05 11:28:33 (the latter being required because microsoft didn't bother implementing the usb standard in windows) Feb 05 11:29:41 I don't know if it's the same on the bbx15, but it seems plausible the same setup code is used Feb 05 11:30:50 i've been thinking i needed a serial connection: i will take a look at network connection Feb 05 11:30:57 thanks Feb 05 11:30:59 CDC-ECM and RNDIS provide basically the same functionality, so it suffices if only one of them works (RNDIS on Windows, CDC-ECM on Mac, both work on Linux) Feb 05 11:31:13 well typically you'd just connect the beagleboard to your network Feb 05 11:31:19 I've never used its usb port Feb 05 11:36:13 it works very well - one of the nicest features when it was a new idea, as you could plug it in and it Just Worked Feb 05 11:36:40 It actually seems to create 2 network interfaces on usb now. not sure why that is Feb 05 11:37:27 maybe one is cdc-ecmn and the other is rndis Feb 05 11:38:12 I literally just explained that Feb 05 11:38:56 ah, so you did. I read it as meaning it had those options, not that it exercised them all. sorry. Feb 05 11:40:06 is it possible to use the usb device port as a host ? Feb 05 11:40:22 on the bbb or bbx15 ? Feb 05 11:40:26 bbb Feb 05 11:40:39 or bbgw in fact, if that's important Feb 05 11:40:39 yesnomaybesorta Feb 05 11:40:54 bbgw... my condolences Feb 05 11:40:55 I love a definitive answer :) Feb 05 11:41:01 yeah i know Feb 05 11:41:18 but it's working for me at the moment. I'll know better next time Feb 05 11:42:21 you're missing the usb 5v power control normally required for a host port... iirc the driver nowadays can deal with vbus being fixed 5v though (I think it didn't used to in earlier days) Feb 05 11:42:40 power isn't too important. Feb 05 11:42:49 i can supply it elsewhere if need be Feb 05 11:43:00 well it very much is for the usb controller, it monitors vbus closely Feb 05 11:43:29 it is designed to support usb otg with its weird vbus signalling Feb 05 11:43:34 even when it's being a host port ? Feb 05 11:43:50 ok, i don't know how that works Feb 05 11:44:04 so I remember it used to get confused as heck if its vbus-enable signal didn't work Feb 05 11:44:14 other than that there's a 5th pin you can ground Feb 05 11:44:16 (which isn't pinned out) Feb 05 11:45:40 why would you care though? the bbgw already has multiple usb host ports Feb 05 11:45:53 I unfortunately want 5 :( Feb 05 11:46:18 so I can add a hub but if I could use the device port it would save that Feb 05 11:46:45 and yeah the ID pin should be grounded if you'd use usb0 as host port instead of device port Feb 05 11:47:07 would that be enough to switch it or do I need to load drivers too ? Feb 05 11:47:27 also, according to the schematic, the bbgw ties vdd_5v and usb_dc together, so it already backfeeds 5v to vbus Feb 05 11:47:39 (which is a gross violation of the usb spec) Feb 05 11:48:14 it would need a DT change Feb 05 11:52:13 on a barely-related subject, is there a cheap and convenient way to get good quality sound out of a bb ? I solved a problem with ground noise on a USB DAC with an isolating transformer, but then I found clicks on one type of DAC that didn't occur on another. not sure if it's an electrical problem or a driver problem Feb 05 11:52:52 the clicks occur when one of the 4 blue leds flashes Feb 05 11:53:00 audio quality is entirely dependent on your audio hardware, not on the bbb Feb 05 11:53:11 sounds like your analog circuit is picking up noise Feb 05 11:53:59 I don't have experience with usb dacs as such, we use i2s Feb 05 11:54:43 that was certainly the case when I had a common ground. but the clicks seem to be something that occur only with a PCM2704 dac and not a CM109 Feb 05 11:55:00 I was wondering if I could use I2s instead Feb 05 11:55:16 which would also save me a usb port .. Feb 05 11:55:37 I can't imagine why you'd use usb for this rather than i2s... for me, usb would be last resort Feb 05 11:55:37 is there a minimal adapter ? Feb 05 11:56:01 what do you mean by that? Feb 05 11:57:09 are there already pins on the headers dedicated to I2s and can I just fit a simple cape with an I2s dac, or does it need a more complex one ? Feb 05 11:58:11 there are a variety of pinouts possible for i2s. I don't know how many are made impossible by the bbgw's misdesign though Feb 05 11:59:29 what product do you use to pick it up ? or is your own hardware ? Feb 05 12:00:23 our own hardware Feb 05 12:01:21 what's your product (if that's not secret ) ? Feb 05 12:01:27 https://dutchdutch.com/ Feb 05 12:02:06 interesting Feb 05 12:03:25 what's the bb doing there ? presumably not the DSP ? Feb 05 12:03:53 the bbgw uses one of the pinouts for i2s for bluetooth audio, but that doesn't necessarily conflict Feb 05 12:04:14 no, it's used for control/management, programming the dsp, and network audio streaming Feb 05 12:05:25 (well, it does a little signal processing as part of network audio streaming, but just limited to mixing and asynchronous sample rate conversion) Feb 05 12:05:52 wired or wireless ? Feb 05 12:06:09 wired Feb 05 12:08:40 reliable low-latency streaming via wifi sounds like a headache to me, especially since you never know what sort of crappy access point people use Feb 05 12:09:04 yeah, i don't trust wifi much. it's useful but not reliable Feb 05 12:09:23 (plus, an RF transmitter right next to sensitive analog hardware... fun fun fun) Feb 05 12:10:33 that exactly. the ideal would be to use AVB, but I think it'll be a while before we properly support that Feb 05 12:12:11 to be fair, a decent commercial setup can be pretty good but not a random domestic router that's fighting next door's channel allocation Feb 05 14:49:58 once a USB bluetooth adapter is inserted how should i manage bluetooth devices pairing / serial communication ? Feb 05 14:52:23 (no connman used) Feb 05 15:12:31 hi, can someone please help me find documentation (schematic aso) for a cape ("Comms Cape A2" at top on "https://beagleboard.org/capes BeagleBoard.org")? Feb 05 15:13:06 fred__tv: connman would be irrelevant for this anyway Feb 05 15:13:38 just because I've found only examples using connman... Feb 05 15:14:36 huh, connman does anything with bluetooth? that's news to me. but then again I don't use connman nor track its development Feb 05 15:15:42 should I install additional software or is there already something "on board" (BB.ORG) Feb 05 15:15:43 ? Feb 05 15:16:23 I think bluetooth stuff is included but I've never done anything with bluetooth so no idea Feb 05 15:18:48 the "bluetooth" package (and its dependency "bluez") appear to be installed on the iot image Feb 05 15:19:11 if you're using a console image you may need to install it yourself Feb 05 15:19:59 (I just checked, bluetooth support in connman is just for network-tethering over bluetooth, and it depends on bluez) Feb 05 15:20:05 since the analyzer I was working to gives good results over bluetooth , I would have the power timing issues solved , as there would me no more physical connections between devices (and even better for portability) Feb 05 15:23:55 none with experiance of cape 'Comms Cape A2'? Feb 05 15:41:42 testit: yeah there are a bunch of poorly documented capes I think... is there no info on this one on github? Feb 05 15:43:01 jkridner[m]: https://github.com/beagleboard/capes/tree/master/beaglebone/Comms has no schematic pdf and the README has dead links and lists vendors as "TBD" Feb 05 15:43:16 I have paired bluetooth device , how can I see it as a serial to write/read into ? Feb 05 15:43:43 fred__tv: is there no info on how to use bluetooth on linux to be found on the entire internet? Feb 05 15:45:49 jkridner[m]: and the overview at https://github.com/beagleboard/capes lists the Comms cape and several other capes as "prototype" even though they seem to be in production Feb 05 15:46:47 I forgot the rfcomm step..... Feb 05 15:46:55 I find it very strange that one can buy a cape and then not find even a schematic for it Feb 05 15:47:56 testit: yep, I fully agree Feb 05 15:48:23 and I've poked jkridner[m] about it before, but I'm not sure if he's seen my messages Feb 05 15:49:07 who is jkrinder? Feb 05 15:50:44 Jason Kridner is beagleboard.org co-founder and community manager Feb 05 15:51:14 ok, thanks for the info Feb 05 15:55:23 fred__tv: btw, rcn has updated the bb-customizations package to include the improved udev rules :) Feb 05 15:56:20 (so, anyone who'd appreciate a significant reduction is boot time, be sure to update that package!) Feb 05 15:56:28 *in boot time Feb 05 16:00:29 fine Feb 05 16:29:53 using /dev/rfcomm0 doesn't work as expected into script , like /dev/ttyO1 does instead.... Feb 05 16:30:20 with minicom both devices work like a charm.... Feb 05 16:30:31 https://ghostbin.com/paste/ztxap Feb 05 16:30:53 replaced /dev/ttyO1 with /dev/rfcomm0 Feb 05 16:31:21 in the time you've spent struggling with abusing bash to do stuff it really isn't designed for, you probably could have learned enough python to get it working there Feb 05 16:31:47 :-(((((( Feb 05 16:33:29 Does anyone have experience with portable displays that connect via USB? Are there drivers available for the beagleboard x15? Feb 05 16:56:08 m Feb 05 16:58:07 raffo: there wouldn't be anything beagleboard-x15 specific about it. your question should be "are there drivers available for linux?" Feb 05 17:15:59 it does feel slightly ironic to connect a display to the x15 via usb, considering it has three digital video outputs (on the expansion connectors) in addition to the HDMI output Feb 05 17:34:04 I'm exploring options, we don't necessarily have those outputs available in certain configurations. I've never used those displays and that's why I'm asking. Feb 05 18:01:21 yeah I understand it can be a matter of convenience Feb 05 20:18:52 Hello, I just had a very weird problem with a beagle bone black: I flashed the newest debian iot on a sd card and booted from it. This worked. Then I wanted to flash the image to the emmc, but after ~5 Minutes of flashing it turned completly off. Now only the pwr led blinks once, when connecting it to usb/ power supply. Is there any hope to restore it? I did not connect anything to it. Feb 05 20:22:10 huh Feb 05 20:22:22 that sounds really bizarre Feb 05 20:22:40 zuim: are you sure it's not a problem with the power supply? Feb 05 20:23:22 how are you powering the bbb? Feb 05 20:23:43 zmatt: I tried using a usb cable connected to a pc and a 5v power supply. Feb 05 20:24:13 When I successfully booted from the sdcard I used the usb cable to the pc. Feb 05 20:26:04 there's no chance you accidently created a (partial) short-circuit, e.g. contact with some stray bit of metal (e.g. wire) Feb 05 20:26:08 ? Feb 05 20:26:41 also, is it a new bbb or one that has already seen a fair bit of use? Feb 05 20:27:00 I think not, because it was on a table with nothing else on it. Feb 05 20:28:29 It is a old bbb, which I used now for the first time. It is revision b6 Feb 05 20:29:24 hmm that's not a bbb revision but a pcb revision Feb 05 20:29:30 might be either rev B or rev C Feb 05 20:30:04 (or even A6A I think... not sure, I don't have a list of pcb revisions) Feb 05 20:31:48 I just found out that it probably is revision C, because it has 4gb emmc storage Feb 05 20:31:51 I have trouble believing reflashing caused this... I feel like it's more likely to be a coincidence, with the actual cause being some form of damage or maybe even just early failure (since it hasn't seen any prior use) Feb 05 20:31:58 yes, that implies rev C Feb 05 20:32:47 how long did you run it from sd card? Feb 05 20:33:12 like, a significant amount of time, or basically just enough to edit the /boot/uEnv.txt as needed for reflashing? Feb 05 20:33:28 only for 20 to 30 minutes, until I had figured out how to connect to it and write to the emmc Feb 05 20:35:20 Maybe it really was a short or some other issue with unfortunate timing. Everything I find on this issue says the cpu is probably dead... Feb 05 20:36:50 yeah... if you can exclude issues with power supply (including cable) and no external hardware is attached to the bbb, a brief power led flash = internal short-circuit somewhere, most commonly internal to the processor as a result of a damaged I/O cell Feb 05 20:37:22 but typically that happens because someone does something stupid with aforementioned I/O cell Feb 05 20:37:55 (e.g. exposure to overvoltage or overcurrent) Feb 05 20:40:57 it's really weird, because nothing was connected and I did not find anything that could have shorted it. But I just found a post saying that the voltage across C34 is lower than 5V, when the AM335X failed and I measured only 0.8V. So I think it's definetly a dead chip. Feb 05 20:41:20 Thanks for your help with this! Feb 05 20:43:14 zuim: well, measuring <1V there just means the beaglebone is powered off, which is not news Feb 05 20:43:45 bbl, shopping Feb 05 21:06:42 auto x = 21; is valid C Feb 05 21:20:57 K&R maybe, I highly doubt it is valid C99 :P Feb 05 21:22:14 yeah it isn't Feb 05 21:22:44 why not ? Feb 05 21:22:46 :( Feb 05 21:22:48 so, it's not valid C (since nobody nowadays still means K&R C when they say "C") Feb 05 21:22:57 auto is the automatic storage class Feb 05 21:23:25 auto is valid C, implicit int isn't Feb 05 21:23:35 ah Feb 05 21:23:39 lame Feb 05 21:23:53 it would have to be written auto int x = 21; and you'd suffer strange looks from people Feb 05 21:24:08 #define auto __auto_type Feb 05 21:24:10 here it works Feb 05 21:25:06 that's a gcc extension Feb 05 21:31:35 you can of course just use c++ if you want an actual auto :P Feb 05 21:32:29 -fconcepts you can even write auto add( auto x, auto y ) { return x + y; } ;P Feb 05 21:32:47 *with -fconcepts Feb 05 21:38:05 (not sure why auto-parameters for functions isn't part of the c++17 standard when the equivalent closure []( auto x, auto y ) { return x + y; } *is* valid c++17) Feb 05 21:40:09 that's nice Feb 05 21:42:58 it's a bit weird that g++ 7 accepted it as gnu extension even without -fconcepts but g++ 8 gives a warning on it Feb 05 21:46:44 I want to be able to use c++2x already damnit :P Feb 05 21:47:44 yeah when there's a new standard it's like christmas Feb 05 21:47:48 I read it and try the new things Feb 05 21:50:46 template decltype(std::declval() + std::declval()) add(A&& a, B&& b) { return a+b; } Feb 05 21:51:06 beautiful Feb 05 21:51:20 just use decltype(auto) as return type Feb 05 21:51:54 yeah I saw that it worked too, but since auto isn't allowed in the parameter type I thought that I might as well not use it at all so that's it's even prettier Feb 05 21:52:23 also please don't write "A&& a" instead of "A &&a" .. just because you *wish* it were parsed the way suggested by the whitespace doesn't make it a reality Feb 05 21:52:37 *a reality Feb 05 21:52:42 in C I write int *a Feb 05 21:52:45 oh I spelled it right the first time Feb 05 21:52:50 jeez I need sleep Feb 05 21:52:53 lol Feb 05 21:53:06 so why would you write int& a in C++ ? Feb 05 21:53:14 it should be int &a for exactly the same reason Feb 05 21:53:19 I wrote it int &a at first but my teachers kept picking on me Feb 05 21:53:26 yeah Feb 05 21:53:31 wow that's bad Feb 05 21:53:33 but the heritage that made C write int *a is far, now Feb 05 21:53:42 they picked on me for int *a too but I kept doing it Feb 05 21:53:45 because I knew the origin Feb 05 21:54:38 the thing is, writing int& a will make people think that int& a = (expr), b = (expr); declares both a and b to be int & Feb 05 21:54:53 while in reality a is int & and b is int Feb 05 21:55:04 since the & binds to the identifier, not to the type Feb 05 21:55:05 yeah Feb 05 21:55:31 so especially teachers should abhor this practice, as it is awful from a didactic point of view Feb 05 21:55:38 yes Feb 05 21:56:27 another weird thing is const for instance; I know it applies to the type on its left Feb 05 21:56:34 but the parser allows it on the left of the type Feb 05 21:56:41 const int/int const Feb 05 21:56:43 yes that's also bizarre Feb 05 21:56:47 I write int const Feb 05 21:56:57 I knew it Feb 05 22:06:05 I also wonder whether people who write int& x still understand how to write more complex types, like a reference (or pointer) to an array Feb 05 22:06:08 :P Feb 05 22:06:41 (fun exercise: change unflatten() in https://pastebin.com/7eH4rmP4 to use an explicit return type instead of auto) Feb 05 22:08:58 int (*(&x)[42])(double); Feb 05 22:10:00 lol Feb 05 22:10:12 it's not for your unflatten lol Feb 05 22:10:16 I'm looking at it Feb 05 22:10:48 I believe I wrote a lvalue reference to an array of 42 pointers to `int function(double);' Feb 05 22:11:02 yup Feb 05 22:11:10 so you need extern or an initializer ;P Feb 05 22:11:30 right Feb 05 22:15:05 I can cheat Feb 05 22:15:28 template using unflatten_type = T (*)[n]; ... Feb 05 22:15:40 no cheating Feb 05 22:15:41 no auto there ! Feb 05 22:15:44 :( Feb 05 22:15:54 no trailing return type either Feb 05 22:16:15 template< size_t n, typename T > T (*unflatten(T x[]))[n]; Feb 05 22:17:02 no wonder C is despised next to hipster languages Feb 05 22:17:03 isn't C declaration syntax beautiful :D Feb 05 22:17:08 lol Feb 05 22:17:15 yeah, that's the word Feb 05 22:17:52 when I was writing it I was like... "wait how do you... is it? .. no.. really? ... it has to be..." Feb 05 22:18:21 lol Feb 05 22:18:26 yeah you have to be in autopilot Feb 05 22:18:30 it can't possibly be anything else, but it feels so wrong to put those brackets after the parameter list Feb 05 22:18:40 just work backwards the syntax Feb 05 22:19:10 I also added the comment Feb 05 22:19:28 // (yes, I could have used auto as return type here, but I wanted to give Feb 05 22:19:30 // you the opportunity to reflect on this syntax for a bit.) Feb 05 22:19:30 calling unflatten(T x[]) is a pointer, so *unflatten(T x[]) is an array, so (*unflatten(T x[]))[n] is a T, so T (*unflatten(T x[]))[n]; Feb 05 22:19:39 lol Feb 06 00:40:01 I see someone was discussing Capes earlier! Feb 06 00:42:02 Schematics should be a mfg. ordeal, right? I am sure the people at the mfg. co. should be able to show the schematics. Feb 06 00:42:24 GHI! Feb 06 00:43:01 But, since the ideas are still prototypes, the info. is just not out yet. This may be the reason. Feb 06 00:44:53 I am not sure about every Cape but some of them have markings for the components on the Capes. And, like w/ the L298 MotorCapes, looking at any schematic of the L298 will give some insight. **** ENDING LOGGING AT Wed Feb 06 02:59:57 2019