**** BEGIN LOGGING AT Wed May 06 03:00:05 2020 May 06 03:00:11 That would be odd for the GHI people and their Capes. May 06 03:01:46 sys_5 is different then vdd_5 ? May 06 03:02:20 I think so. I am not completely sure why or for what reasons. I can look it up again if necessary. May 06 03:02:31 KenUnix: vdd_5v is input (to power the BBB), sys_5v is output May 06 03:03:24 Oh. May 06 03:03:26 Yikes. May 06 03:04:06 vdd_5v can sometimes also be abused as output if ihe BBB is powered via the 5V barrel jack (it ties directly to it), but you shouldn't do that May 06 03:04:21 Not familar with servo cape. zmatt help set_ May 06 03:04:22 Now, I know. May 06 03:04:47 KenUnix: He helped me more than one can fathom. Computers! May 06 03:05:51 based on a quick glance at the schematic the servocape uses sys_5v like it should, so no problem there May 06 03:06:10 WOuld I need to apply another voltage source to the Cape? May 06 03:06:21 Outside of what the BBB provides it? May 06 03:07:11 I don't see any reason for that May 06 03:07:21 Okay. Dang GenTooman! May 06 03:08:08 So... May 06 03:08:19 What are the 5v in and GND on the Cape doing on it? May 06 03:09:13 does it have a "5v in" ? I do see a VIN but that just seems to be connected directly to pin 2 of each output header May 06 03:09:34 so you can use it for whatever voltage you need for your servos I guess, or just not use it at all May 06 03:10:09 Okay. I was wondering b/c GenTooMan told me to attach that part of the 5v In and GND to the BBB. May 06 03:10:20 He is bad news, bad! May 06 03:10:30 he probably misread the schematic May 06 03:10:33 it's a weird schematic May 06 03:10:35 Oh. Okay. May 06 03:11:15 It really does not matter. The screw terminal on the Cape came damaged when I purchased it. May 06 03:11:30 "Loose as a goose." May 06 03:11:30 lol May 06 03:11:40 I know. Boo! May 06 03:13:44 On the source, how would I know which servo header I am using? May 06 03:15:15 So yea. It does not work unless I am probing it w/ the DMM. May 06 03:18:44 Loose pin to cape? May 06 03:19:00 Nope. Maybe. May 06 03:21:20 I think it is one of the poorly placed servos I was using. May 06 03:21:25 I will try w/ another one. May 06 03:21:44 it slipped from 0 to 180 and back by hand. I think it was the servo. May 06 03:23:34 Do I need a special servo? May 06 03:23:53 Nice name : Maya Kaczorowski May 06 03:26:40 either my battery just died on my DMM or the Cape killed it! May 06 03:30:04 Conference tomorrow noon eastern github githubsatellite.com May 06 03:31:01 Whelp, that is all she wrote for the ServoCape for tonight. I need to figure out about the DMM and a better way to test the ServoCape headers. May 06 03:32:33 Good night. Good luck set_ May 06 03:36:43 Thanks. May 06 03:47:54 Yep. DMM shot! May 06 07:10:04 Hi All, I was wondering if someone could give me some pointers on C++, I am trying to get a sample application running on a beaglebone AI but keep getting errors, I have even tried to compile this on a Linux VM but am getting the same issue. I have researched the problem for days without avail. May 06 07:15:09 without actually giving the problem, the answer is a clear "nobody can" for sure. May 06 07:25:06 LetoThe2nd, /usr/bin/cmake --build /home/picozense/Downloads/PicoZenseSDK_Ubuntu14.04_20190605_v2.4.0.4_DCAM710/Samples/DepthFrameReader/cmake-build-debug --target all -- -j 1 May 06 07:25:06 DepthFrameReader May 06 07:25:07 /home/picozense/Downloads/PicoZenseSDK_Ubuntu14.04_20190605_v2.4.0.4_DCAM710/Samples/DepthFrameReader/DepthFrameReader.cpp:19: undefined reference to `PsGetDeviceCount' May 06 07:25:08 /home/picozense/Downloads/PicoZenseSDK_Ubuntu14.04_20190605_v2.4.0.4_DCAM710/Samples/DepthFrameReader/DepthFrameReader.cpp:37: undefined reference to `PsSetDataMode' May 06 07:25:09 /home/picozense/Downloads/PicoZenseSDK_Ubuntu14.04_20190605_v2.4.0.4_DCAM710/Samples/DepthFrameReader/DepthFrameReader.cpp:46: undefined reference to `PsGetFrame' May 06 07:25:11 CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/DepthFrameReader.dir/all' failed May 06 07:25:26 The files exist, and are referenced May 06 07:26:23 sounds massively like you just forgot to link against some library. namely the one that provides those three functins. May 06 07:29:17 This is where my issue lies, I am a C# programmer, my C++ is not great. I have picked up quite a bit but am struggling with this at the moment. Just hasnt clicked yet. The files I have referenced are header files, could you give a brief description of libraries. I am running a sample project that is provided by the hardware supplier so trying to May 06 07:29:18 get my head around why it would not exist as is May 06 07:30:41 Ryan52: just including a header tells the compiler that you want to use something that "looks like this". then you can compile. but in the end, you have to link, and the linker needs to know where "this" actually is, then. May 06 07:35:09 LetoThe2nd makes sense, I am assumig library files are he standard .lib. The project does not contain any of these. This can probably be rectified with the hardware provider. One last question if you dont mind, how do you link the libraries? May 06 07:36:52 Ryan52: .lib is not a suffix used for libraries on linux May 06 07:37:20 static libraries are .a, shared libraries are .so May 06 07:37:40 Ryan52: for the first, PsGetFrame is certainly not in the standard library but something in your "PicoZenzeSDK". for the second, googling "cmake link library" is certainly wroth trying. as well as doing some introductory c++ stuff :) May 06 07:38:35 its probably possible to call c functions from mono.. May 06 07:38:43 Ryan52: but to be honest, if you're having trouble building an example included with a commercial sdk, it feels more appropriate to seek support from the company that made the sdk May 06 07:39:02 humpelstilzchen[: it certiainly is, its only software after all. May 06 07:39:05 there's nothing beaglebone-specific about your questions May 06 07:39:31 zmatt: yup May 06 07:40:32 so much for "I have researched the problem for days without avail" May 06 08:19:28 * rob_w dma breaks my head May 06 08:41:50 rob_w: how so? May 06 08:42:19 zmatt, too many options, examples and too less expierence May 06 08:42:35 but i am crawling towards my goal ;-) May 06 08:43:14 i ve i am correc then my pru might use dma now for real soon May 06 08:44:26 what do you mean by that? any access by pru to main memory could be considered a form of dma, or do you mean using pru in combination with edma? (which would be mostly pointless but might have its rare uses) May 06 08:45:16 actually i dont know all the insights .. let me give my "state of the nation" May 06 08:45:53 i successfully setup pru firmware and a iio wrapped pru rpmsg driver to talk backn n forth via rpmsgs May 06 08:46:48 yet i see a bottleck with that as rpmsg buffers are single and the pru waits till the arm releases the virtio ring May 06 08:47:02 I've never quite understood the point of rpmsg... they murder performance May 06 08:47:42 hmm, I don't think rpmsg intrinsically has any limitation on the number of buffers available May 06 08:49:04 well i tend to adiffernt setup now .. where i dma_alloc_coherent some dma , send the address to that via rpmsg to the pru ( use 2 of those buffers) and then the pru should fill up the dma with sweet data and notify the arm via rpmsg again May 06 08:49:44 oh you're talking about memory allocated via linux' API to allocate dma-memory, not about actual DMA May 06 08:50:10 that sounds like you're just making rpmsg even more inefficient though May 06 08:50:14 i do , it seems ;) as stated ive virtually no idea about all this May 06 08:50:36 rpmsg already works by sending buffers (which can be allocated independently) via queues May 06 08:51:40 well ur quite right i guess about the inefficient things May 06 08:52:15 the main inefficiency is that pru reading from ddr3 memory is just very slow and ruins deterministic timing May 06 08:52:34 and rpmsg has unnecessary complexity and overhead for many applications May 06 08:53:16 when u say reading ? do yo also mean writing ? May 06 08:53:24 no, just reading May 06 08:53:32 pru writing to ddr3 memory is fire-and-forget May 06 08:53:37 it doesn't block on it May 06 08:53:46 ok , so my application only writes from pru to arm May 06 08:54:06 yep, that could be very fast if you use a simple ringbuffer May 06 08:54:08 only the small control bits ( and the new dma address) might now be written from arm to pru May 06 08:54:38 just use some chunk of pruss memory for that instead of ddr3 memory May 06 08:55:14 i want 2 buffers , so the pru can fill up 1st, irq the arm about finishiing the first and continue the 2nd while arm copies the 1st and clears the flag May 06 08:56:04 what would arm copy it to? May 06 08:56:21 copying from ddr3 memory to some other place in ddr3 memory is typically fairly pointless May 06 08:56:51 well ... true , what i meant is from there on it gets pushed inside the iio_buffers May 06 08:57:05 dont know if that is copy but i guess May 06 08:57:14 I think BeagleLogic just uses a single ringbuffer to which pru writes and from which data is consumed by the reader May 06 08:58:25 (so that would get copied once to userspace when it does read() ) May 06 08:58:25 ok I don't know anything about IIO May 06 09:00:09 yeah ..but i will worry later .. or switch over to a alsa wrapping and use gstreamer or such but iio gives me some nice adc settings and environments which is what i ve behind the pru , 6 x 16bit 400khz adc May 06 09:00:19 I'd assume you'd allocate buffers via the kernel iio interface and then have pru write its data into it May 06 09:00:34 or something? May 06 09:00:37 yes basicly May 06 09:01:32 I'm looking at the iio API right now, it's not super clear yet May 06 09:01:33 and integretiy of that adc data is crucial , so i do like to have some control which feels rpmsg gives me May 06 09:02:15 the key is iio_push_to_buffers_with_timestamp() May 06 09:02:36 yeah I saw it though I don't really understand it May 06 09:02:56 also with timestamp? that sounds awkward May 06 09:03:47 you configure it via sysfs to eitehr include a stamp or not or how big your buffers are or which channels you want to see or not May 06 09:04:03 I wonder if alsa would be more suitable May 06 09:04:04 then your /dev/myiiodev is ready to be read May 06 09:04:38 this looks more like it's more meant for stuff that's occasionally samples May 06 09:04:50 alsa is definitly a good alternative i my mind, as gstreamer then coudl start mangling data with filters or pipe it to the network etc May 06 09:05:43 anyway, I'm probably not the best person to give advice since I'd never bother writing a custom kernel driver for this in the first place, since doing it entirely from userspace seems infinitely easier May 06 09:06:24 hmm May 06 09:06:27 (and less likely to cause kernel panics :P ) May 06 09:06:50 ur talking about the uio pruss stuff right ? May 06 09:06:56 but I'm also a uio-pruss user, not remoteproc-pru May 06 09:07:09 yep May 06 09:08:44 although that doesn't hugely matter here I think... I guess it depends on how much memory you need, since uio-pruss also allows allocating a configurable chunk of shared ddr3 memory, and I'm not sure if remoteproc-pru supports that yet May 06 09:09:08 so far I've not needed buffers that big anyway May 06 09:09:32 so what where ur usecases for uio pru ? May 06 09:09:51 you mean what I use pru for? May 06 09:09:57 yes May 06 09:11:09 we currently use it to capture high-resolution timestamps for an audio stream, because the timestamps from alsa are way too jittery to be of any use whatsoever May 06 09:11:31 though I want to switch to using it for audio data as well, and ditch alsa entirely May 06 09:11:48 nice May 06 09:12:43 yeah iirc alsa timestamps had like tens of microseconds of jitter, with outliers that were even bigger, and we're now more in the tens of nanoseconds ranges iirc May 06 09:13:25 and if pru is entirely responsible for audio then 5-10 ns accuracy should be no problem May 06 09:16:23 thats all so beyond me ( it feels) i am soo happy to finished the pru code to match the specs needed , now i need to shuffle it to arm .. then i will lay down and die in peace May 06 09:18:20 I'm curious how you're using pru to receive a continuous stream of data from an adc yet somehow have timing needs that are so loose that you can afford to use rpmsg May 06 09:20:00 I guess it's not that bad at 400 kHz... 500 pru cycles per sample is still a lot of time to waste May 06 09:21:07 well so the setup is as follows May 06 09:22:39 6 highend adc are connected via spi to pru pins ..so the spi clocks out the 6 channels data ... well we then might avrg the samples x4 May 06 09:22:56 so basicly i get 100khz worth data May 06 09:23:49 this is then packed into the payload of the rpmsg, the transport is still too slow May 06 09:24:08 having pru do averaging doesn't really relax any timing constraints unless the ADC has a fifo to buffer 4 samples May 06 09:24:11 as i ve only 1 rpmsg buffer pru->arm May 06 09:24:37 this definitely sounds like a normal ringbuffer would be more appropriate May 06 09:24:39 yeah i need to still trace the costs etc May 06 09:25:55 when i get back to my office i will recalc my timings via my scope May 06 09:26:09 so put things into perspective: BeagleLogic can sample up to 12 (or 14 if eMMC is disabled) channels, packed into 16 bits, at 100 Msps (i.e. one sample every 2 pru clock cycles), and it offloads that data to ddr3 memory May 06 09:26:16 but the already know that rpmsg with a single buffer is not working May 06 09:26:47 rpmsg isn't constrained to having a single buffer available though, at least the protocol doesn't May 06 09:26:51 which is uio pruss right ? May 06 09:27:13 beaglelogic? I don't think it uses uio-pruss no, I think it might use a custom kernel driver but I'm not sure May 06 09:27:59 Hi everyone, I have a question. Is it possible to flash the beaglebone black eMMC directly from mmc1 pins on expansions headers ? May 06 09:28:43 rob_w: it definitely doesn't use rpmsg, I can assure you that ;) May 06 09:28:49 hehe May 06 09:29:07 rob_w: but that's more because it predates it and because rpmsg would be completely pointless. I think rpmsg, as a protocol, could deal with that data flow May 06 09:29:30 u think ? May 06 09:29:37 sebc87: that should be possible yes, just make sure to keep the processor in reset by pulling P9.10 down to ground May 06 09:30:48 rob_w: well, almost... I'm not sure if rpmsg could be convinced to put its control structures in pruss memory May 06 09:31:18 rob_w: the very high data rate of BeagleLogic almost certainly precludes doing *any* read from ddr3 memory May 06 09:32:05 i gues it doent need to read maybe only during setups ,, then its a oneway street pru->arm May 06 09:32:19 as logicanalyzers do ? May 06 09:32:42 zmatt thank you, Have you an idea of hardware that I could connect on these pins to emulate a mass storage on my laptop ? May 06 09:33:31 rob_w: yeah but even just dumping data into a ringbuffer normally still requires *some* read, at the very least to prevent overrun May 06 09:33:43 ok right May 06 09:35:42 rob_w: rpmsg requires more reads than that due to its indirection (it uses an array of packet descriptors and two ringbuffers of indices into that packet descriptor array, one for each direction of communication... more or less) May 06 09:37:04 well i could imagine my setup to stripped to only setup via rpmsg on start , show the pru its buffers, then only use the irqs to kick and read the dma May 06 09:38:09 I personally just place the buffer pointers into pru registers as part of the pruss initialization, but that's something I can do with uio and you can't with remoteproc :) May 06 09:38:49 yeah , i am happy with that , doesnt hurt me if that works May 06 09:41:28 rob_w: my whole pru program is... rather simple as a result: https://pastebin.com/raw/wfn849ip May 06 09:42:41 sweet ... brb phone May 06 09:48:14 it could actually be simplified further.. it still samples both the pruss eCAP timer and the system timer, but I've since then written a linux driver that allows using an eCAP as system timer, and have switched to using the (200 MHz) pruss eCAP timer as system timer for linux May 06 09:49:19 that was needed in part because otherwise the raw monotonic time would have a small frequency error compared to the main oscillator (i.e. 1 second of raw monotonic time would not equal 24 million cycles of the 24 MHz main oscillator) May 06 09:49:44 resulting in slow drift of timestamps that made a mess of things May 06 09:51:31 sebc87: another beaglebone could do it... maybe whatever logic is in an sd card reader could also do it (at reduced performance since only 4-bit data instead of 8-bit) but the minor protocol differences between SD and MMC might end up being an issue May 06 09:57:57 Can I connect a sd card reader directly and work in 4-bit mode ? May 06 09:58:45 there is a non-zero probability that that might work. unfortunately also a substantial probability that it won't May 06 10:02:46 Ok thank you May 06 10:20:05 Hello any one suggest me enable all UART steps. May 06 10:20:17 in Beagle bone AI May 06 10:55:42 please any one suggest me May 06 11:04:38 Any one suggest me Enable all uart in Beagle bone AI through updating device tree or other best way? May 06 12:05:08 no ^ haha :) May 06 13:12:21 shouldn't be that hard May 06 13:12:49 except the bbai probably still has that annoying giant DT node that occupies all pins May 06 20:01:04 zmatt finally got around to testing encoder and getting it all wired up today. This is what I've got in Python, and when I print cur_position, it's always printing 0's. You mind taking a peek to see if I've got it setup right? https://pastebin.com/J0jBFH3v May 06 20:02:52 I'm getting a voltage on the output of the differential to single chip, but I've only got a meter here to test it. Guess I need to get a scope May 06 20:41:33 biggi_: did you configure the pins correctly? **** BEGIN LOGGING AT Wed May 06 22:40:13 2020 May 06 22:42:03 zmatt sorry I just saw this. I'll post my whole code May 06 22:42:51 zmatt I assume not. I just looked over the documentation https://pastebin.com/2JFC4qxs May 06 22:50:52 donno to scream at TI or MS :( ARRRRGGGGGGGGGGGGGGGGGGGGGGGGGGGGG May 06 22:51:36 so much waste time... 4 windows boxes, only one works May 06 22:51:57 and TI doesn't know what's wrong either May 07 00:35:17 (y) Hi all ;D May 07 00:37:36 zmatt, With the release of Debian 10.3 4-6-2020 and forward setting up USB has been changed. Have you seen?:O May 07 00:37:47 no? May 07 00:37:59 Explain? May 07 00:38:08 yes, explain what you mean May 07 00:38:53 How & where the USB is defined is moved to a boot file May 07 00:39:06 ? May 07 00:39:24 Wait I'll push up a copy..... May 07 00:40:04 I have a 2020-04-06 system running here May 07 00:40:10 just explain what you're talking about May 07 00:40:40 I don't see anything different to previous images May 07 00:41:29 Look here https://pastebin.com/pabav14G May 07 00:42:21 oh you're talking about network configuration May 07 00:43:16 that whole script still seems really gross to me, I don't understand why they don't use a decent network manager (e.g. systemd-networkd) so this sort of hack isn't needed May 07 00:43:31 Yes I saw a few questions posted on BB forum on Gllgle. May 07 00:45:13 So IF you use /etc/network/interfaces it won't work anymore. I wonder where you setup Ethernet now? May 07 00:46:25 ethernet has always been managed by connman May 07 00:46:34 If you use USB Ethernet it don't see anywhere to set default route except manually May 07 00:46:55 You mean connmanctl ? May 07 00:47:00 Ethernet uses dhcp which will set the default route automatically May 07 00:47:22 connmanctl is a commandline utility that controls connman May 07 00:47:56 I should cook up a guide on how to replace all this mess by systemd-networkd May 07 00:48:06 maybe he'll consider it for future releases May 07 00:49:01 So, you can have both USB & Ethernet active? USB <> PC & Eth0 >Internet via router. May 07 00:49:14 those normally don't have anything to do with each other May 07 00:49:22 and work entirely independently May 07 00:50:01 The device tree he points dosen't seem relative May 07 00:50:10 ? May 07 00:50:27 the post you quoted doesn't reference anything in device tree May 07 00:50:47 At the bottom of his message is a link. Go check out. May 07 00:51:09 yeah I saw, it links to setup scripts May 07 00:52:06 Well maybe I'll try 4-26-2020 on my spair BBB. May 07 00:53:01 So, would you edit image before burning to SD-card? May 07 00:53:15 no, why? May 07 00:53:53 To preset USB I don't use address specified I se 192.1668.137.x May 07 00:54:02 ?? May 07 00:54:09 why? May 07 00:54:57 To pre-configured setting for USB I don't use the address specified I use 192.168.137.x May 07 00:57:03 why? May 07 00:57:25 does he default conflict with your LAN ? May 07 00:57:29 *the default May 07 00:58:38 yes because the do internet sharing on Win-10 and it won't accept 192.168.7.xx is you want to share internet. May 07 00:59:20 I thought you were using ethernet for internet access? May 07 00:59:29 not yet May 07 01:00:55 192.168.7.xx is ok for pc <> BBB but not for internet sharing May 07 01:01:12 pretty sure internet sharing can be configured to use a specific range May 07 01:02:00 but regardless, when using internet sharing your PC is the dhcp server and the bbb ought to be the dhcp client and not another dhcp server. with a proper network manager that would be easy to change, not so much with the current scripts probably May 07 01:02:37 That's what I thought. But windows kept complaining in network config when for that device any time i added internet access. May 07 01:02:51 complaining about what? May 07 01:04:42 It would pop a dialog box saving I could not use that address. In fact it changed it to the 137 address and said this change may interrupt connectivity. May 07 01:05:37 Windoze May 07 01:05:38 odd, windows I guess May 07 01:06:35 Well I see in the future Windoze moning to a good kernel like linux. May 07 01:06:55 this has nothing to do with the kernel, this is a policy decision May 07 01:09:05 Look M$ came out with their anti-virus for linux on WSL. I tried to install it using their really long procedures and it blue up. M$ was no help so I removed it. M$ solution was to re-install linux. Nut huh? May 07 01:11:30 M$ likes to REBOOT, POWER DOWN, after update REBOOT, an on on May 07 01:11:35 anti-virus for Linux? So they are going out of business? May 07 01:13:13 Yoou had to install a special service MTAP or some thing and after installing it my McAfee firewall warned me about unknown tafiic attempts. That was that. May 07 01:14:52 Then those Ubuntu folks told be the 'sleep bug was M$ fault. I said it worked fine on18.04 anyway they closed myticket. May 07 01:16:44 Close your eyes. Close the ticket. Make it go away by pointing fingers. They could have tried it. May 07 01:18:21 Simple shell command sleep like sleep 5 should pause for 5 seconds but I was getting hardware clock problem. Gee mabe I should write my own sleep 'c' program May 07 01:18:54 I am all about pointing some fingers. "He did it!" May 07 01:19:41 I broke down and got a new DMM. Off to test the servo headers! May 07 01:20:15 set_ I was telling zmatt that usb to PC setup has changed and moved to a file now as of 4-2-2020 May 07 01:20:39 dunno where it was previously, I think hardcoded in the script? May 07 01:20:45 Heh? May 07 01:20:46 Odd. May 07 01:20:56 Well, that sort of makes sense. May 07 01:21:04 File system and all... May 07 01:21:28 I think things should be files. Right? May 07 01:21:46 Set see this Look here https://pastebin.com/pabav14G May 07 01:22:33 zmatt If there is truly a but in Ubuntu regarding time then a 'c' program won't work? May 07 01:22:49 Yea. That is for the server config. for the BBB. May 07 01:22:51 KenUnix: ? May 07 01:23:07 KenUnix: I can't parse that question May 07 01:23:53 zmatt Ubuntu 20.04 If there is truly a bug in Ubuntu regarding reading time then a 'c' program won't work to pause for 5 seconds? May 07 01:24:17 KenUnix: what on earth are you talking about? May 07 01:25:04 The 'sleep' bug in the shell on Ubuntu 20.04. I use sleep to pause scripts for a timed period May 07 01:25:36 I very much doubt there's a "sleep bug" in ubuntu 20.04, please clarify May 07 01:26:52 OK. At the CLI if you type in sleep 5 it comes back with a hardware clock problem. That's 20.04 on Windoze WSL May 07 01:27:07 oh so you mean a WSL bug, not an ubuntu bug May 07 01:27:34 no, because version 18.04 and Debain 10 is OK May 07 01:28:32 20.04 is supposed to be the LTS for Windoze WSL May 07 01:28:36 so it's a wsl bug that only affects certain releases May 07 01:28:42 sounds like it has to do with the glibc version May 07 01:29:31 Do you have something I could use in 'c' to check it? May 07 01:30:08 this appears to be the core issue: https://github.com/microsoft/WSL/issues/4898 May 07 01:32:34 "clock_nanosleep(CLOCK_REALTIME, ...) has always been broken on WSL" May 07 01:33:57 workaround: https://github.com/microsoft/WSL/issues/4898#issuecomment-612622828 May 07 01:34:09 (for ubuntu specifically) May 07 01:35:14 KenUnix: adding https://launchpad.net/~rafaeldtinoco/+archive/ubuntu/lp1871129 and updating your system should hopefully fix it May 07 01:36:04 actually it looks like upstream has a newer version of glibc so you may need to forcibly install this particular version May 07 01:36:12 (unless upstream has implemented a workaround) May 07 01:39:23 Has the ServoCape been known to kill off DMMs? May 07 01:39:48 set_: you're the only one using it May 07 01:39:49 One bit the dust yesterday after attaching it to the ServoCape. May 07 01:39:54 Ha. May 07 01:39:55 what's a DMM ? May 07 01:40:09 Digital Multi Meter. May 07 01:40:30 there's no plausible way the servo cape can damage a multimeter, everything is max 5V May 07 01:40:59 I attach the leads of the DMM to the ServoCape, it renders the DMM sensorless. May 07 01:41:14 Yesterday, it did the same thing but to an older model of what I got today. May 07 01:42:35 Now, this DMM can handle 12v easy. I have seen it handle the 12v supply from a dc. May 07 01:42:46 But... May 07 01:43:52 For some reason, when I attach the DMM positive to positive on the Cape servo headers and GND of my DMM to negative of the Cape servo header (Same Pin), I get corruption on my DMM. May 07 01:44:10 I mean. The DMM refuses to show any detail of what i am connecting it to. May 07 01:44:23 set_: ?? May 07 01:44:30 I have no idea what you mean by that May 07 01:44:34 Luckily, this one did not hash out totally compared to the last one. May 07 01:44:43 Okay. May 07 01:44:57 I am going to ask around at the mfg. site. May 07 01:44:59 and what pin of what cape servo header did you mean? May 07 01:45:04 Maybe, someone knows something. May 07 01:45:09 S16 May 07 01:45:20 That is the easiest one to access. May 07 01:46:25 It is the three pin/Servo Pin config. closest to the screw terminal on the Cape. May 07 01:47:16 for those, pin 1 is ground, pin 2 only connects to other connectors hence floats by default, pin 3 is the output (presumably 0V by default) May 07 01:51:04 Okay. May 07 01:51:13 I see the parts on the silkscreen May 07 01:52:06 This is what I am trying to explain. It basically renders my voltmeter useless. I use the DMM as a voltmeter, i.e. just to check if 5v is coming through. May 07 01:53:10 I do not apply power to the board until my system has been set up for testing. May 07 01:54:05 The DMM will not turn on but after a couple of seconds, 15 to 30 seconds, the DMM will come back to life once the connection has been released. May 07 01:54:15 Odd! May 07 01:54:21 I have no idea what you do with your equipment May 07 01:54:27 I know. May 07 01:54:34 I am just trying to test this equipment. May 07 01:54:39 B/c... May 07 01:54:50 I have been unable to make it work w/ source. May 07 01:54:52 but presumably it's just operator error May 07 01:55:15 I will let it go. I get the point. May 07 01:57:09 zmatt, no good the ppa didn't fix sleep or htop May 07 01:57:30 KenUnix: did you install the libc provided by that package? May 07 01:57:42 2.31-0ubuntu8+lp1871129~1 May 07 01:58:04 like I said, it looks like merely updating won't install this package anymore May 07 01:58:08 i followed the steps in he ppa May 07 01:59:05 try apt install libc6=2.31-0ubuntu8+lp1871129~1 May 07 01:59:28 sudo add-apt-repository ppa:rafaeldtinoco/lp1871129 and sudo apt-get update May 07 01:59:46 yeah those steps just make the repo available, that doesn't install anything May 07 02:00:18 Well it said it was updating 17 things May 07 02:02:29 you mean downloading 17 package lists probably May 07 02:06:40 Wouldn't the sudo apt update activate them? May 07 02:06:53 no that just updates the package lists May 07 02:07:13 also, like I already said, even upgrading your packages probably won't install this one May 07 02:07:28 since a newer libc has been released and this ppa hasn't been updated since May 07 02:07:42 so what was the exercise for? May 07 02:08:01 03:59 <@zmatt> try apt install libc6=2.31-0ubuntu8+lp1871129~1 May 07 02:08:17 to forcibly install that specific version May 07 02:10:38 OK. With all this cow-dung all I want was to sleep for a few seconds May 07 02:11:00 dat works May 07 02:11:07 date works May 07 02:11:35 KenUnix: blame microsoft, it's a long-known issue May 07 02:11:45 it just got more visible May 07 02:12:34 Looking deeper into that ppa it refers to htop. OK got something in 'c' to test for 5 second delay May 07 02:12:47 Is there a way to easily go back and see comments here since I'm on a different computer on a different browser? Seems the log isn't up to date yet May 07 02:13:39 biggi_: I don't think adafruit_bbio does pinmux setup for you? just in case double-check it with my show-pins utility https://github.com/mvduin/bbb-pin-utils/#show-pins May 07 02:14:21 biggi_: and no, the only way to be assured to know what was said is by being here May 07 02:15:38 I will bookmark that and try when I get back out to the work bench. https://adafruit-beaglebone-io-python.readthedocs.io/en/latest/Encoder.html is what I was going off zmatt May 07 02:17:23 actually; i'm trying showpins now. Forgot I've got it on the home network so I can ssh in May 07 02:19:46 zmatt so I've got it on eQEP1 which is P9.33 and P9.35 and I'm not seeing those on showpins May 07 02:20:29 biggi_: thehm May 07 02:20:32 ehm May 07 02:20:57 P9.33 and P9.35 are (NON-3.3V-TOLERANT) analog inputs May 07 02:21:35 Sorry; 8.33 and 8.35 May 07 02:21:58 those are definitely in the output of show-pins May 07 02:22:39 Yea; I had my mind crossed. Sec, I'll put it in an imgur. May 07 02:22:46 What would happen if you put this on P9.33? May 07 02:22:48 Would it blow the input? May 07 02:23:13 there'd be significant risk of frying the board May 07 02:23:23 https://imgur.com/gxnsIpl May 07 02:23:36 Well board isn't dead, I think I might have hooked it up that way earlier XD May 07 02:24:08 biggi_: that's saying the pins are configured for use by video output, connecting to that would also be a bad idea May 07 02:24:19 disable video by uncommenting the appropriate line in /boot/uEnv.txt May 07 02:24:54 you may want to pick different pins anyway if possible May 07 02:25:14 I can use eQEP2 May 07 02:25:16 externally driving pins that have sysboot function is risky May 07 02:25:47 specifically, both pins must be low during power-on-reset May 07 02:26:09 Here's 8.11 and 8.12 for eqep2 https://imgur.com/GD7AWGq May 07 02:26:13 and I don't think you can ensure that in your case May 07 02:26:39 eqep0 shows p9.27 and p9,92 on the documentation https://adafruit-beaglebone-io-python.readthedocs.io/en/latest/Encoder.html May 07 02:26:43 I don't know where 9.92 is? May 07 02:27:02 a silly naming convention of cape-universal May 07 02:27:41 a few expansion header pins (including P9.42) connect to two cpu pins May 07 02:27:50 and cape-universal allows those cpu pins to be configured separately May 07 02:28:03 so to distinguish them it adds 50 to one of them May 07 02:28:38 so I can use p9.42 for that? May 07 02:29:04 I just set it up for eqep2, going to get in and try that out May 07 02:29:35 aha! May 07 02:29:41 it does set it up aumatically May 07 02:30:07 I'm guessing it was just me using the eqep1 without disabling the video (I used the iot image, so thought it was diabled by default) May 07 02:32:53 board isn't fried either, so here's to hoping that analog input still works too May 07 02:36:48 I'm new to a lot of this pin mux stuff. A lot of what I do is setting pins on micros on a custom pcb for a single task, not flipping them around like this haha May 07 02:37:02 So thank you again for your knowledge and putting up with my...slowness **** ENDING LOGGING AT Thu May 07 02:59:57 2020