**** BEGIN LOGGING AT Sat Oct 31 02:59:59 2015 Oct 31 11:26:51 Hey when can I buy a X15? Oct 31 11:29:15 hey Oct 31 11:29:23 Anyone else here? Oct 31 11:32:57 at some point in the near future Oct 31 11:35:41 you could join the beagleboard-x15 google group. I'd expect that it's going to be announced there Oct 31 11:43:28 k ty Oct 31 12:44:14 my guess is they would shoot for December Oct 31 13:29:44 yes, X15 is delayed until December, see http://elinux.org/Beagleboard:BeagleBoard-X15#BeagleBoard-X15_Description Oct 31 14:42:50 hi all Oct 31 14:42:58 i move from arduino to begle Oct 31 14:43:07 and i have a little problem Oct 31 14:49:35 well, what is your little problem? or do you not want help with it? Oct 31 14:49:55 sory Oct 31 14:50:11 i`m trying to do an arduino map in to beaglebone python Oct 31 14:50:15 i mean this: Oct 31 14:50:31 "int workTemp = map(lightLevel, minLight, maxLight, minTemp, maxTemp)" Oct 31 14:51:36 sounds like a python problem Oct 31 14:53:51 it is Oct 31 16:15:35 Anyone familiar with cape and eeprom i2c address? Oct 31 16:16:02 what's your actual problem? Oct 31 16:17:28 If the eeprom is going to be the only i2c device, can I set its address to 0x00? Oct 31 16:19:57 http://elinux.org/BeagleBone_and_the_3.8_Kernel#Cape_Manager_requirements Oct 31 16:20:15 Those addresses are 0x54-0x57 on I2C1 bus, hence the 4 capes limitation. Oct 31 16:21:37 but if you are doing something custom that nobody else will ever use, you can obviously do what you want (and what works) Oct 31 16:25:13 Ok thanks for the information Oct 31 17:18:37 lol, the new X15 photo looks like someone tried a little too hard to insert the jtag connector the wrong way around Oct 31 17:20:53 oh wow lots of new photos Oct 31 17:24:57 where? now you have me curious Oct 31 17:25:53 still the same wiki page Oct 31 17:26:01 http://elinux.org/Beagleboard:BeagleBoard-X15 Oct 31 17:26:25 I don't know how new they are, but I hadn't noticed them before Oct 31 17:27:26 seems they were added a few days ago Oct 31 17:32:57 The SoC that drives the thing has 3 LCD interfaces. It would be nice if the BB X15 had at least 2 display. Oct 31 17:33:24 expansion headers Oct 31 17:33:51 Hmmm they have a IO header detail anywhere? Oct 31 17:34:09 I made a spreadsheet, https://docs.google.com/spreadsheets/d/1mSqEpV_BAUHfeNApytxHcGhgTZwypy564GyOr66Nphs/view Oct 31 17:35:21 expansion pins are labeled "exp" in the EVM/BB-X15 column (I was too lazy to include the actual connector / pin number) Oct 31 17:36:18 feel free to figure out which features could be used simultaneously :P I've started documenting IOSets in the Signals tab but that work is still incomplete Oct 31 17:36:42 video in/out iosets are already marked though Oct 31 17:36:57 (near the bottom of that tab) Oct 31 17:37:09 Heh better than what I have :D is the audio codec as crappy as usual? Or did they decide that it might be nice to have good audio? Oct 31 17:37:16 no idea Oct 31 17:37:25 check the schematic? Oct 31 17:37:45 What you mean actually LOOK OMG, good idea I'll do that :D Oct 31 17:38:43 lookie, a new version of the schematic Oct 31 17:38:48 too bad you can't easily diff those Oct 31 17:39:08 the file size more than doubled though o.O Oct 31 17:40:06 yeah... it's possible to put data into PDF files you can export I found out. That is supported in v 1.4 of the PDF format. I suppose the shifted it to being more a container format by then. Oct 31 17:40:28 doesn't really help to get a useful diff though Oct 31 17:41:22 Hmmm Libre office can read pdf's and allow you to modify them. It might be possible to use the writer to do a PDF diff like the spread sheet can do a diff between 2 spreadsheets. Oct 31 17:41:55 I was doing that with xlsx documents. Oct 31 17:42:07 a pure visual diff wouldn't be useful either... you'd need something that understands the semantic value of those little squiggles on a schematic to be able to highlight genuine differences (versus just a line having moved slightly) Oct 31 17:42:38 "this squiggle means they deleted something 2 times! Oct 31 17:43:16 Well since the PDF's are generated it's more likely to yeild something than a standard PDF when edited. I found that it's like word it keeps everything you do (bloody ...) Oct 31 17:43:36 27 pages .. hmm Oct 31 17:45:24 nice versioning too... there's rev A2 (2015-08-28), then rev A2b (2015-09-10), then rev A2 with revision history deleted (2015-10-22) Oct 31 17:46:07 I noticed that. A isn't released the versioning doesn't make sense. Oct 31 17:46:27 also, there are many previous revisions in circulation Oct 31 17:47:16 not only is the new PDF more than twice as big, it also renders noticbly slower Oct 31 17:48:15 I noticed that too. Oh well it's up to them how they control documents I am just glad they don't have to meet governmental regulation requirements or they would be shut down. Oct 31 17:48:16 never mind, that was just due to different zoom levels Oct 31 17:49:56 Nice thing about their PDF's is that they use REAL text in them unlike PADS which makes images for text output so you can't search the bloody schematics. Oct 31 17:50:09 oh don't get me started Oct 31 17:50:53 I can't believe there are still schematics programs which dare to shit out crap like that Oct 31 17:51:07 They left out the mechanical ... it says 28 pages but their are 27 hate it when that happens. Oct 31 17:51:37 but apparently quite a few schematics programs are still incapable of producing proper PDFs Oct 31 17:52:09 zmatt... LOL yeah is all I can say KiCAD as 'crazy' as it is works better than pads. I think they are just shitting out an image and using LZH compression on it. Oct 31 17:52:40 I think KiCAD finally puts out real PDFs. Oct 31 17:52:57 Actually I can check that hmmm. Oct 31 17:53:07 we recently acquired an Altium license, so hopefully I will never have to see a piece of shit PDF like that anymore again Oct 31 17:54:54 good luck with that. Oct 31 17:55:26 they previously used a mix of different programs, all of which produced equally shit PDFs Oct 31 17:56:23 no that's not true, some were considerably more shit than others Oct 31 17:57:03 Just be glad you aren't using PADs I sware EDA tools have hit to crapper big time these days. Oct 31 17:58:13 I can't fathom how they can make them worse then come out with a "new" version that fixes the bugs. Taking a page out of Microsofts book of bad business practices. Oct 31 17:58:19 at least some produced an unsearchable PDF of a mere 331 KB (no embedded graphics) rather than a 16 MB one (again nothing but a bunch of lines) Oct 31 18:00:25 so, have you figured out yet which combinations of video interfaces could be used practically? Oct 31 18:00:53 I'm perusing the data sheet and user guide. Oct 31 18:01:37 the TRM might still be somewhat useful to check which of an interface are actually needed for a particular mode of operation Oct 31 18:02:13 I think I've already merged the relevant info (for this particular question) from the datasheet into my spreadsheet Oct 31 18:02:28 *check which pins of an interface Oct 31 18:03:55 I find you need a map to find stuff on the SoC I can see why it takes an entire team just to make a product out of one these parts. I think the Prya from open pandora is using an IC from this family. Oct 31 18:04:25 they claim to be using the omap5, in which case they're absolutely insane Oct 31 18:05:17 I guess they are going for a tablet size? (IE 10" diagonal) with big ass battery. They are already at a 4Ah battery on the pandora. Oct 31 18:06:05 yes but the omap division has been killed off while the omap5 was still in development Oct 31 18:07:14 TI offers no support for it Oct 31 18:07:35 IT looks as if the X15 is using an OMAP repurposed design too me. They may have reused the core IP and added the other stuff from the industrial line. Oct 31 18:07:47 Vayu is definitely omap5-derivative Oct 31 18:08:07 but under the umbrella of the automotive department Oct 31 18:08:15 and actually maintained/supported Oct 31 18:09:44 Interesting they choose to call the video pins VOUT. Why are they mixing analog nomenclature with digital? That's ignorant and confusing. Perhaps they should send the people who decide this stuff to "communication fundamentals in engineering". Oct 31 18:10:30 ehm, dunno I don't think that there's really any risk of confusion Oct 31 18:10:59 I wondered why their were so many VOUT pins ... now I know it's not a built in LDO. Oct 31 18:12:14 power supply pins use very different naming schemes anyway Oct 31 18:12:29 Looks like you could add another 2G to the memory. Oct 31 18:14:25 it could support 4 GB with just a BOM change, and I suspect (contrary to the TRM's statement) 8 GB if the pcb is changed to use 8 x8 RAM chips instead of 4 x16 Oct 31 18:15:00 the required RAM chips are pretty expensive though Oct 31 18:15:38 I browsed at some of the manufacturers... usually a RAM chip of the required size turns out to be dual-rank Oct 31 18:16:29 but single-rank ones do exist Oct 31 18:17:56 GenTooMan: and honestly, "vout1_hsync" wasn't a hint to you it might be video? ;P Oct 31 18:19:59 hey I said peruse. :D looks like V1 goes to the IO pins the HDMI port is built in so the question is does V1 go to both or just one of this points. TRM time I guess. Oct 31 18:20:35 there are four video pipelines Oct 31 18:20:57 one for each vout port and one for hdmi Oct 31 18:21:32 though vout0 (vout1 in TRM, see also the Note tab in my spreadsheet) has a mux allowing it to display any of the four pipelines Oct 31 18:21:55 Only one seems to be brought out on the IO pins and then their is the HDMI Oct 31 18:22:27 ehh, why do you say only one if brought out? Oct 31 18:22:58 when right at the top of my spreadsheet is visible that vout2 is also on the expansion headers Oct 31 18:23:52 well I was just looking at the labels in the schematic myself :D Oct 31 18:24:20 every pin has up to 14 functions (in addition to gpio and safe mode) Oct 31 18:24:30 ignore the labels, look at the spreadsheet :P Oct 31 18:28:12 looks like all three video outputs are available, in addition to hdmi Oct 31 18:28:32 all at full 24-bit Oct 31 18:31:47 however, only vout0 has series resistors so it may be challenging to maintain signal integrity on the other two if used at high speeds... I guess we'll have to wait to see what's achievable in practice Oct 31 18:34:47 I also worry slightly about how few ground pins the expansion connectors have Oct 31 18:35:45 but then again, parallel video isn't *that* highspeed anyway Oct 31 18:41:49 brb shopping Oct 31 19:21:04 hmm spread sheet useful (hides club) Oct 31 19:23:43 that's why I make 'em ;) though figuring out which combinations of functionality are possible is still a challenge Oct 31 19:25:17 most pins have multiple functions, most functions map to multiple pins, for many interfaces there are iosets to consider (though I suspect you may be able to violate them if you figure out the correct manual iodelay config) Oct 31 19:25:40 the same functionality may be available in more than one way Oct 31 19:26:47 e.g. if you need only 18-bit color, you can either use d0-d17 or d0-d5,d8-d13,d16-d21 Oct 31 19:27:07 which may be useful if some of those pins have alternate functions you want to use Oct 31 19:32:21 but, if we ignore other functionality, it seems the full video-out capabilities of the processor are accessible, except for vout0_fld (which is only used for interlaced video, and even there is optional) Oct 31 19:36:12 and possibly clock rate limitations on vout1 and vout2, depending on factors Oct 31 19:40:26 well adding more heads for a display is less a problem that way. You are right to be concerned about the "io" not being designed for video output. Screw in the IO lines is a concern for DVI/HDMI output. Oct 31 19:40:58 depends a bit on the trace length from the processor to the dvi/hdmi framer Oct 31 19:43:15 plus the processor's output driver characteristics, and how sensitive the framer is to suboptimal signal integrity Oct 31 19:43:28 really needs proper simulation or a lot of trial and error Oct 31 19:44:42 T&E tends to be expensive. What EDA tool are they using for the board design? It might be hard to calculate exacting length capacitance and inductance for a gerber plot. Oct 31 19:46:22 "they" are no doubt using something expensive that is perfectly capable of doing signal integrity analysis... but that's of no use unless "they" are also the ones designing this hypothetical expansion board for the video outputs Oct 31 19:47:15 zmatt "they" are hypothetically just bringing some IO out at this point :D Oct 31 19:47:45 yes, the only expansion board that exists at this point is the one used by the EVM Oct 31 19:48:08 at least it isn't LVDS signals we are talking about here. Oct 31 19:48:57 well, if it were then they would certainly have made sure it arrives at the connector safely Oct 31 19:49:08 see e.g. the pcie lanes Oct 31 19:49:46 note also how P19 does have plenty of ground pins :P Oct 31 19:54:05 I'll avoid making a bad pun on that :D Oct 31 19:55:43 this is the expansion board I was referring do btw -> http://www.ti.com/tool/tmdxevm5728 Oct 31 19:56:07 or rather, x15 + expansion board... dunno if you can buy them separately Oct 31 19:57:10 hi Oct 31 19:58:23 howdy Oct 31 19:58:27 anyone here familiar with rtklib ;D? Oct 31 19:58:32 hi zmatt <(^_^<) Oct 31 19:58:41 never heard of it Oct 31 19:58:47 oh Oct 31 19:59:04 well i want to develop a programm on beaglebone and i have never done something like this before Oct 31 19:59:21 i think i need a bootloader Oct 31 19:59:50 ;x Oct 31 19:59:56 i need to come back later Oct 31 20:00:00 i cant ask good questions yet Oct 31 20:00:06 ;( Oct 31 20:02:43 o.O Oct 31 20:04:17 GenTooMan: so it seems Pyra really has gone forward with using the omap5432... brave... the TI support forum has been closed and removed from their site index, support for the Tesla DSP on the omap4 and omap5 has been dropped in the current compiler version... Oct 31 20:23:08 zmatt indeed kind of "OH WELL" unfortunately OMAP was targeted at a market TI didn't actually try to promote or didn't win enough design wins. Consumer markets are finicky that way. Oct 31 20:23:39 well they had their time Oct 31 20:24:33 but I can understand them for wanting out.... it's just not a funny market Oct 31 20:29:19 I think the big issue in consumer market is the fickleness. My thinking is "less consumer" and "more practical" too much "look it can play video and show HD content" and not enough "its secure and hard to screw with works reliably". A lot of companies practically FAKE they are doing something useful with some of the systems deployed in homes these days. Oct 31 20:30:26 uhuh, more features, more speed, less power, MORE FEATURES, no development time since two months from now our new design will be obsolete again Oct 31 20:31:14 that's just bad management and bad planning. Following fads did not make apple do well. They take some serious engineering time in their stuff. Oct 31 20:38:22 interesting, apparently the *full* omap5 TRM was briefly on TI's website Oct 31 20:38:35 so some people have that thing... Oct 31 20:39:08 zmatt well better than nothing I guess. Still the issue is they are using a dead part. That makes no sense. Oct 31 20:39:58 well it's not really dead either, in fact an updated datasheet was released this month, and ISEE also built a board with it Oct 31 20:40:25 it's just that TI doesn't want to support it, at all, possibly because they laid off the people who actually know this chip Oct 31 20:40:49 it's in a really weird limbo Oct 31 20:41:02 and I would really like to get my hands on that TRM Oct 31 20:43:33 Well seeing that TI is trying to make an SoC out of everything it can make everything they offer ... confusing they have so much parts that have so many peripherals that it's an overload. Keep it simple doesn't seem to be their method these days. Oct 31 20:44:21 well stuffing many peripherals onto a die makes a lot of sense actually, and keeps the number of dies low Oct 31 20:45:02 (the number of "parts" that are actually the same die is mostly irrelevant except for the marketing dept) Oct 31 20:45:55 I'm actually a fan of integration... especially when large amounts of data need to be moved around Oct 31 20:46:45 Integration is fine if you actually have things well documented. Most things are marginally documented of late. Oct 31 20:47:26 not sure I agree... documentation quality varies, although TI's still seems quite good Oct 31 20:47:43 Vayu's documentation is definitely a lot better on all the integration details than subarctic Oct 31 20:48:16 but I don't see how documentation quality relates to chip integration Oct 31 20:52:57 It's hard to use something that's poorly documented that's all. Anything that is from TI's burr brown operation is well documented. MSP line well mileage varies. BenchMarq God help you if you use one of those parts (really might be the only person aside from the designer who knows exactly how to use a part from those people). So yeah it varies. The industrial side of things isn't too bad. The aquisition of Luminary micro Oct 31 20:52:57 introduced a lot of bad documentation into TI I noticed. MSP is a TI part from the beginning still squirrelly documented. Oct 31 20:55:05 ok I'm not really familiar with TI docs apart from SoCs and some audio parts... and actually of those the one I really had issues with was a burr brown part iirc Oct 31 20:55:25 hehehe. Oct 31 20:56:49 I've used a whole variety of parts from INA's to Audio D2A to small industrial SoCs. mostly small stuff. Power supply stuff from National seems mostly top notch. Oct 31 20:58:17 I also got a good impression from the docs of the xilinx coolrunner-ii (which I was exploring since it's present on the xds100v2) Oct 31 20:58:28 their IDE otoh... /o\ Oct 31 20:59:07 cool runner is actually from philips semi which is no nxp. They developed those parts I was very impressed on how well they documented stuff on it. Oct 31 20:59:20 s/no/now/ Oct 31 21:00:07 ahh, didn't know that Oct 31 21:00:28 nxp in general also gives me a good impression Oct 31 21:01:04 been a while since I worked with one of their μCs but I recall it was a pretty sane thing Oct 31 21:01:20 (otoh I've grown to dislike STM32 quite a bit) Oct 31 21:02:14 the problem with the IDE is of course that it makes no sense to have the same development tools for a 32-macrocell CLPD all the way to FPGA of millions of cells Oct 31 21:03:39 and at least I got that IDE to run, can't say the same about the one from microsemi (was previously different company, forgot the name) Oct 31 21:03:51 actel Oct 31 21:14:19 * zmatt considers buying an IGEPv5 as a toy to play with Oct 31 21:14:54 ick, that price is of course ex VAT ... hmz Oct 31 21:19:12 :x Oct 31 21:20:13 Oi All those logic design parts are kind of pricey Oct 31 21:20:31 tools not parts brain dead it's only permanent Oct 31 21:20:34 I have a question about bbb Oct 31 21:20:50 well the IGEPv5 is quite reasonably priced Oct 31 21:22:03 I just got one and i only want to use it as i did with stm32discovery board Oct 31 21:22:29 Put a program on it and check if it doesnt crash Oct 31 21:22:34 I used mine as a door stop so ... Oct 31 21:22:37 you mean baremetal programming Oct 31 21:23:27 on a processor like the am335x that is a bit more challenging than on a typical microcontroller Oct 31 21:24:42 I was thinking i could make the whole program and os bootable with u-boot? Oct 31 21:25:15 Or how would u put a c program on it :x? Oct 31 21:25:37 ROM can load a program from μSD card or even via the network Oct 31 21:26:19 I personally wouldn't bother with getting u-boot involved Oct 31 21:26:53 Hmm i am not familar with u boot at all so i am glad u say that Oct 31 21:27:21 I usually use netbooting when working on baremetal code... it's convenient Oct 31 21:28:01 since you just recompile, reset the BBB and ROM grabs the freshly compiled code from my laptop via the network Oct 31 21:28:41 U can connect that to eclipse ? Oct 31 21:28:53 I am very new to this Oct 31 21:29:17 Havnt used a raspberry pi either Oct 31 21:29:24 I don't use eclipse Oct 31 21:29:59 baremetal programming on an rpi is probably a lot ickier than on an am335x Oct 31 21:30:14 Zmatt thank u for ur help Oct 31 21:30:46 I once made an example in pure assembly (someone asked for it, no idea why he wanted pure assembly) -> https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/791/4452.trivial_2D00_a8.tar.gz Oct 31 21:31:23 ;( Oct 31 21:31:39 there's also a C++ version here https://e2e.ti.com/support/arm/sitara_arm/f/791/p/389752/1377125#1377125 Oct 31 21:32:45 Dont want to take any more of ur time but i want to know if i got the main idea down Oct 31 21:32:48 having jtag is highly recommended when doing baremetal programming though Oct 31 21:33:20 So for baremetal i dont need linux on it ;×? Oct 31 21:33:38 eh, baremetal means "without OS" Oct 31 21:33:59 Sorry Oct 31 21:34:06 it's not exactly the easiest road to take though Oct 31 21:34:19 given the complexity of a big processor like that on the beaglebone Oct 31 21:35:17 So the easier way is just get the program on the system and tell the system to run the program on start up? Oct 31 21:36:20 just write it as a linux application? and if you need it to run at boot, make it a service Oct 31 21:37:43 Ok thank u sir have a nice day <(^_^<) Oct 31 21:42:35 essentially the beagle bone is the complexity level of a PC Oct 31 21:42:41 oh gone I see Oct 31 21:43:13 also that's not quite true Oct 31 21:44:05 you can still, with time and effort, grasp what's going on in a cortex-a8 and on an am335x Oct 31 21:45:07 while the complexity of a modern x86 processor is insane, and SMM means you basically have no idea what's going on on your PC Oct 31 22:06:44 Such is weirdness. You have a PS3? did you know one of the cell processors isn't used on it for games or anything as such? Sony brilliantly decided to make a software network interface using one of the cores of the Cell processor to handle all the ethernet level processing the only on the PS3 that's real is the PHY. Oct 31 22:14:23 I don't have any gaming console, and I don't know enough about the cell processor or the PS3 design to judge whether or not this was a reasonable design decision or not Oct 31 22:38:17 I suppose the beagle board is just TI's way of making cheap Eval boards. Oct 31 22:50:40 also true, if you don't know about something it's hard to comment on it. At least I've found that so. For example I don't know all the IO on the BBB so some questions are well heh "donno" :D Oct 31 22:52:57 then just open up the TRM every now and then, pick a random module you're not familiar with, and read about what it does :) Oct 31 22:54:06 heh... right now I want a micro HDMI cable :D but as you say read the TRM might help a bit. My keenest interest is the PPU and SPI bus at the moment Oct 31 22:54:19 PRU you mean Oct 31 22:54:58 Well I guess that's better than calling them a "PU" :D Oct 31 22:55:02 yeah I also still really really need to find time to play with PRU, since it's definitely one of the most awesome subsystems Oct 31 22:58:10 SPI is not really particularly interesting... McSPI is a sort-of-okayish implementation, though the Mc aspect seems pretty useless to me (and isn't supported by the linux driver), but I guess it very much depends on what sort of devices you attach to it Oct 31 22:58:59 the linux driver is kinda sucky and e.g. doesn't support using arbitrary GPIOs as chip-selects Oct 31 23:00:38 the PWMSS peripherals are also worth knowing about Oct 31 23:02:11 like, we're going to have a product with fans installed, and the eQEP peripherals are perfect for tracking fan speed and it has integrated watchdog for alerting if the motor stalls Oct 31 23:03:56 and I'm still curious what precision the PWM outputs are truly capable of... the "micro edge positioning" step size isn't specified in the datasheet (hmm, should still complain about that on E2E) Oct 31 23:04:11 but on the omap-L13x it's around 200 ps Oct 31 23:05:07 and those were fabricated on 65 nm process while the am335x is 45 nm process, which probably means even finer precision Oct 31 23:12:52 is the PWM sourced from the core clock or from a seperate "peripheral" clock? Oct 31 23:21:02 GenTooMan: do I really have to point you to the "how to ask smart questions" doc as if you're a noob? :P the clocks used for each subsystem/peripheral are listed right at the front of the relevant TRM chapter Oct 31 23:21:39 and an overview is also given in the PRCM chapter, with pretty pictures even Oct 31 23:23:33 ohhh pictures I like pictures (LOL) Oct 31 23:24:39 (the pwmss modules run on 100 MHz, but really learn to fish that info from the TRM :P ) Oct 31 23:25:50 ( specifically l4ls = l3s = core-pll-m4 / 2 ) Oct 31 23:27:05 but the eHRPWM modules are capable of finer control than a single clock cycle, they have a programmable delay line thingy attached to them (that's the HR in the name) Oct 31 23:29:08 hence my statement that it seems likely to me their resolution will be finer than the 200 ps listed in the omap-L13x datasheet, since afaik smaller transistors usually mean lower gate delays Oct 31 23:31:50 It's doubious you would need ultra fine resolution for driving a fan. However if you were setting a window for capturing a radar return pulse that would be a different matter Oct 31 23:32:15 some examples are mentioned in the chapter I think Oct 31 23:32:33 the PWMSS modules originally come from TI's C2000 series real-time microcontrollers Oct 31 23:33:34 I have a 28XXX kit for motor control looking like I should do something with it. I have to finish my driver but you can guess how time is sometimes. Nov 01 01:48:10 GenTooMan: I don't really claim to grasp time Nov 01 02:07:07 Does anyone knows how to read analog input (AIN*) on the BBB? Looking on the internet, it should be in /sys/devices/ocp.3/helper.12#, but I don't have this directory on my system Nov 01 02:07:38 helper? I've never seen such a thing Nov 01 02:08:14 zmatt.. probably a hack :p Nov 01 02:08:31 veremit: my first thought is "might be some machinekit thing..?" Nov 01 02:08:57 ocp.3 .. hrm .. Nov 01 02:09:00 found it, it's in helper.15 :) Nov 01 02:09:10 o,O Nov 01 02:09:21 thanks Nov 01 02:09:28 what kernel is this? Nov 01 02:09:32 and dtb Nov 01 02:09:33 debian Nov 01 02:09:40 that's a distribution, not a kernel Nov 01 02:09:48 3.8 Nov 01 02:09:53 uname -r Nov 01 02:10:04 I got a ocp.3 on my ancient build .. Nov 01 02:10:09 but no helpers Nov 01 02:10:12 cat /etc/dogtag Nov 01 02:10:14 3.8.13-bone79 Nov 01 02:10:32 BeagleBoard.org Debian Image 2015-03-01 Nov 01 02:10:36 grep ^dtb /boot/uEnv.txt Nov 01 02:10:40 curious Nov 01 02:11:25 [ 2:10am] michael@beagleblack ~ $ ls /sys/devices/ocp.3/44e0d000.tscadc/TI-am335x-adc/ Nov 01 02:11:25 driver iio:device0 modalias power subsystem uevent Nov 01 02:11:44 nearest I got Nov 01 02:12:02 on a 3.13 RN kernel Nov 01 02:13:03 Liir: the collective knowledge here about such old kernels isn't very high anymore I think, and that will probably not improve over time ;) Nov 01 02:13:38 zmatt... afaik 3.8 is still 'official' ... Nov 01 02:13:52 I know Nov 01 02:13:59 how should I update my kernel, I updated my BBB with the most recent image Nov 01 02:13:59 although a fine vintage also Nov 01 02:14:04 in fact new releases are still being done periodically Nov 01 02:14:40 I think RN has a build server no? Nov 01 02:14:48 or similar Nov 01 02:14:49 Liir: 2015-03-01 is most definitely not the most recent image Nov 01 02:14:53 or even a recent image Nov 01 02:15:06 you can "apt-get upgrade" Nov 01 02:15:15 that won't update the kernel though I think Nov 01 02:15:21 hrm Nov 01 02:15:35 what's the apt-cache for linux-image? Nov 01 02:15:48 I'm too used to emerge/portage commands .. Nov 01 02:16:18 Liir: the most recent official release is 2015-07-28, though personally I've always used the most recent testing image whenever I needed an image Nov 01 02:17:15 veremit: you mean apt-cache search -n linux-image Nov 01 02:17:15 ? Nov 01 02:17:23 interesting, I did that last week following links on the BBB website, and that's what I got. Let me see if I can find this one Nov 01 02:17:31 md5sum: 716fbfe023ad2b64807346503f9951cf Nov 01 02:17:33 whoops Nov 01 02:17:36 http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Debian_Releases Nov 01 02:17:42 zmatt .. yea that'll do Nov 01 02:19:07 veremit: or, assuming your pkgcache is intact, just do "apt-get install linux-image-" and then press tab twice Nov 01 02:19:14 although I don't recall the portage command to list slot options :/ Nov 01 02:19:27 ah thats nifty Nov 01 02:20:08 veremit: if it doesn't work, check /etc/apt/apt.conf.d/02apt-get-clean Nov 01 02:21:53 removing *.deb files is okayish to save space (though I'm pretty sure there must be a better way than this), however it should _not_ remove the *.bin files nor set the cache files to "" Nov 01 02:22:32 that just causes apt to regenerate the files every time, and breaks tab completion Nov 01 02:22:47 (I hope rcn fixed that in later images) Nov 01 02:54:30 zmatt thanks for the spreadsheet really useful Nov 01 02:55:15 GenTooMan: thanks, although there's still plenty of work to do Nov 01 02:55:20 it could really use an interactive tool though Nov 01 02:55:34 (I mean, one that doesn't suck, unlike TI's) Nov 01 02:59:21 * zmatt gasps Nov 01 02:59:42 I just discovered the TRM is... wrong.. about something :o Nov 01 02:59:49 * zmatt can't believe it **** ENDING LOGGING AT Sun Nov 01 02:59:58 2015