**** BEGIN LOGGING AT Thu Feb 09 03:00:03 2017 Feb 09 04:26:14 LuaStoned: same as always: the spi controller needs to be configured (incl pinmux) and enabled in DT, and you define one or more child nodes for individual spi devices (and their settings) Feb 09 04:27:22 as with most things this can be done either in the main DT, or at runtime using overlays Feb 09 04:27:55 my overlay-utils include an example -> https://github.com/mvduin/overlay-utils/blob/master/spi0.dtsi Feb 09 07:08:05 hello! Feb 09 09:30:52 hello, i'm currently reversing a PCB and i'd like to archive my findings on a PCB designer for Linux, what are you guys using? Feb 09 09:31:53 I tested oregano but the component bank is pretty weak Feb 09 09:38:14 gquere: kicad Feb 09 10:02:32 KotH: thanks, trying that and eagle but both seem to have incomplete libraries Feb 09 10:17:00 Hey everyone, I have a doubt regarding the GPIO1_28(12th pin on P9). I'm working on a project to implement a push button interrupt and I noticed that the GPIO1_28 pin is always set to rising(/sys/class/gpio/gpio60/value file shows value is 1), whether i push the button or not. Can someone tell me if this is normal? And if so, how can I change it to falling? Feb 09 10:36:23 ITYM: "I have a question" Feb 09 10:37:21 tbr, not "a doubt" ? Feb 09 10:37:45 ogra_: I doubt they have a doubt. So to say. Feb 09 10:38:00 :) Feb 09 10:40:56 kkk0306, did you check your pinmux configuration? Feb 09 10:41:01 kkk0306: regarding your _question_ - you should look at the device tree and overlays that you load. Those define the pinmux Feb 09 11:12:43 tbr: Sorry. I have "a question". :) And I've just started working on BeagleBone. Not very familiar with it yet. I will take a look thank you. Feb 09 11:15:00 see you folks Feb 09 11:16:15 lore4: The pinmux configurations shows this for the pin 60 "pin 60 (44e108f0.0): (MUX UNCLAIMED) (GPIO UNCLAIMED)" Feb 09 12:07:11 gquere: you will never find a tool with complete libraries Feb 09 12:07:21 gquere: there will be always that one part that isnt in there Feb 09 13:04:45 KotH: it was easier for me to add libraries in eagle so I used this Feb 09 13:05:18 (element14 has a whole lot for each vendor) Feb 09 13:52:14 zmatt: remember when I burnt a BBW like 2 weeks ago? I know now why Feb 09 13:54:14 on the PCB I connected it to, the reference for low voltage components was the PSU phase Feb 09 14:10:18 hello :) Feb 09 15:00:14 gquere: ouch Feb 09 15:18:03 zmatt: my understanding is that I have to place an isolation transformer before the PCB I want to measure, and bind the BBB's GND to the PSU phase Feb 09 15:18:11 think that's correct? Feb 09 15:18:39 what? Feb 09 15:18:58 bind the BBB'd GND to earth ground / safety ground you mean Feb 09 15:19:45 well, since the ground reference for the PCB's components is the input's PSU phase Feb 09 15:20:18 shouldn't the BBB be made aware of that by also connecting its ground to the same phase? Feb 09 15:20:21 *that* is the problem you should fix, a phase should never be used as low-voltage ground reference Feb 09 15:20:34 that's an electrocution hazard waiting to happen Feb 09 15:20:52 probably yeah, but I didn't design the product Feb 09 15:21:41 if there's no way to fix that in the product, use an isolation transformer and tie whatever pin the product uses as ground to actual ground Feb 09 15:23:38 not sure what's gonna happen if I invert P and N... Feb 09 15:23:58 neutral is rarely a substitute for ground Feb 09 15:24:33 when I said ground I meant GND on the BBB right? Feb 09 15:24:39 btw is it really using that phase as reference, as in a low-ohm connection? Feb 09 15:25:06 I suppose it must be given the damage.. Feb 09 15:25:15 no idea Feb 09 15:26:53 also, I'm guessing you don't have a residual-current circuit breaker? ( https://en.wikipedia.org/wiki/Residual-current_device ) .. since that should have tripped immediately before components got fried Feb 09 15:26:53 [WIKIPEDIA] Residual-current device | "A residual-current device (RCD), or residual-current circuit breaker (RCCB), is a device that instantly breaks an electric circuit to prevent serious harm from an ongoing electric shock. Injury may still occur in some cases, for example if a human falls after receiving a shock. In the United States and..." Feb 09 15:27:23 this one tripped in the building yeah Feb 09 15:27:41 the BBW and USB ports of my computer also tripped :p Feb 09 15:28:10 (btw we call these differential circuit breakers) Feb 09 15:28:21 yeah wikipedia lists a whole bunch of synoyms Feb 09 15:29:33 lots of different words/meanings for the same object Feb 09 15:30:02 yeah I guess it can't necessarily protect hardware... Feb 09 15:30:19 can confirm, my BBW is half black now Feb 09 15:30:57 right... that doesn't sound like it tripped soon enough to protect a human either Feb 09 15:31:25 do you even have earth / safety ground connections on either of the two devices? (your computer and that.. thing) Feb 09 15:31:34 computer yeah Feb 09 15:31:51 thing doesn't Feb 09 15:32:36 (thing I'm hacking is a smartmeter so it seems logical that it doesn't have ground) Feb 09 15:34:52 if your neutral is close enough to earth then *maybe* you can get away without an isolation transformer, but I think you'll probably still end up frying hardware and/or tripping the RCD Feb 09 15:35:23 (if you swap phase/neutral to make neutral the reference plane I mean) Feb 09 15:35:40 you think it's wrong to connect the BBB's GND to the input phase of the thing (which comes from an isolation transformer) ? Feb 09 15:35:44 but that's easy enough to test: short neutral to earth in advance :P Feb 09 15:35:56 wait there's an isolation transformer? Feb 09 15:36:06 well I need to add one :p Feb 09 15:36:16 otherwise it'll blow up again Feb 09 15:36:55 did some research in the meantime, learnt stuff but 'a little knowledge is a dangerous thing' so I'd rather check Feb 09 15:38:11 so, after isolation there's no phase vs neutral anymore, and you can simply make that input pin neutral by tying it to earth ground Feb 09 15:39:59 in general, make sure there's a good solid connection between the ground planes of all devices Feb 09 15:42:51 still unsure about how that'd work, i'll try to read up more on that before connecting anything Feb 09 15:42:55 that's also how you easily find out if you can get away without an isolation transformer and simply use your neutral as reference (i.e. swap phase/neutral compared to your previous setup)... tie that to earth ground through a resistor. if your RCD trips then you need an isolation transformer, if the resistor starts emitting smoke your RCD isn't working Feb 09 15:43:15 if it works you can tie a good solid connection between the two Feb 09 15:44:34 in all cases you want whatever the device considers "ground" to be at the same potential Feb 09 15:47:25 it doens't make sense for me yet, i'll study it more Feb 09 15:47:53 hmm, I think i'd never assume N and Earth be related Feb 09 15:49:44 it might work today and then the load situation on the electric net changes and suddenly N is away several 10 volts Feb 09 15:53:21 thinkfat: yeah it sounds like a bad idea in general, unless they're already combined (TN-C / TN-C-S) Feb 09 15:56:14 it's mostly TT over here Feb 09 15:58:49 isolation transformer. Feb 09 16:03:02 @zmatt thanks, i just stumbled upon some issues on github where rcn mentioned loading spi at runtime is too late.. so I was curious Feb 09 16:04:18 LuaStoned: iirc the problem was that it enabled spi without configuring it properly, and it couldn't be configured properly later since the DT would only be scanned by the driver at probe time Feb 09 16:04:58 hmm, well im using the console image and loading via uEnv, seems to work Feb 09 16:05:11 loading what via uEnv ? Feb 09 16:05:33 note that since the bugfix spi isn't enabled by default in the main dts Feb 09 16:05:43 so then you can later configure and enable it by loading an overlay Feb 09 16:08:49 also, loading the overlay at runtime (rather than via uEnv) might not work due to conflict with cape-universal Feb 09 16:09:06 cape-universal is however automatically disabled if you load any overlay via uEnv.txt Feb 09 16:33:57 Hello Feb 09 17:29:21 Hi, I want to run a home server with my BBB, but I wonder what's the maximum time that BBB can work (2 months ?), did anyone know ? Feb 09 17:35:12 .o0(people are so used that stuff breaks, that they even expect everything to fail within months?) Feb 09 17:44:45 KotH: the facebook generation :) Feb 09 17:45:40 [ 5:45pm] michael@beagleblack ~ $ uptime Feb 09 17:45:40 17:45:28 up 295 days, 17:25, 1 user, load average: 0.08, 0.06, 0.10 Feb 09 17:47:46 ah.. you are doing the yearly reboot thing aswell ? :) Feb 09 17:48:13 I have no idea why that got rebooted then .. but its my dhcp/bind9 cache .. so .. its kinda network-central :D Feb 09 17:49:17 miraculously, it hosts a USB2 external HDD .. and it just works™ Feb 09 17:50:07 Linux beagleblack 3.13.6-bone7.2 #1 SMP Fri Mar 21 03:59:54 GMT 2014 armv7l <= fine vintage XD Feb 09 17:53:27 :) Feb 09 20:20:50 hi# Feb 09 21:07:38 join Feb 09 21:08:20 Could anyone help me configuring CAN bus for beaglebone black? Feb 09 22:34:18 can someone tell me what the http://octavosystems.com/octavo_products/osd335x/ on the BBB wireless is about? is it just some BBB components packaged as an integrated chip? Feb 09 22:37:01 yes Feb 09 22:37:15 it is 90% of a BBB integrated into a BGA package Feb 09 22:37:24 makes it easy to do AM335x based designs Feb 09 22:37:38 what's making up the bulk? i guess it's just a ton of resistors and capacitors? Feb 09 22:38:13 it is the processor plus PMIC plus memory Feb 09 22:38:25 so it's three ICs? Feb 09 22:38:33 and some glue Feb 09 22:38:40 and it's 90% of the cost? Feb 09 22:38:51 isn't soldering all those small parts a lot more work? Feb 09 22:38:55 cost depends on how many you buy :D Feb 09 22:39:05 not in volume Feb 09 22:39:11 =) i'm just wondering where the big save is compared to these 3 discrete components Feb 09 22:39:56 have you designed with DDR memory? Feb 09 22:40:05 no, i have not. that's why i'm asking =) Feb 09 22:40:29 the original BBB is a 6L PCB.... the OSD3358 makes is reasonably easy to do it in 4L Feb 09 22:40:46 ah, i see. so the 90% refers to the routing effort here? Feb 09 22:40:55 for example, I did the VKS with the OSD3358 - http://www.hy-research.com/VKS.html Feb 09 22:41:02 not just routing Feb 09 22:41:26 DDR memory requires the traces to be matched so it is more then making it easy to connect point A to B Feb 09 22:41:50 sorry, i'm new. matched == length matched? Feb 09 22:42:01 you can in theory do a design with the OSD3358 on a 2L board Feb 09 22:42:14 match as in length and impedence... so what you run underneath it matters Feb 09 22:42:29 that sounds... complicated Feb 09 22:42:29 every member of a signal group needs to arrive at the other end at the same time Feb 09 22:43:29 it is more of a nuisance then complicated Feb 09 22:43:41 but it does impose requirements Feb 09 22:47:12 well, it sounds like a nightmare to debug at least Feb 09 22:54:49 matter of getting used to it Feb 09 23:34:43 ds2: thanks for the insights =) Feb 10 02:58:38 I'm having an issue where building PRU code with clpru says that EINVAL is not defined. What includes folder should I be using to get the right version of errno.h **** ENDING LOGGING AT Fri Feb 10 03:00:01 2017