**** BEGIN LOGGING AT Sun Dec 20 02:59:57 2020 Dec 20 03:07:27 ayjay_t, https://gnucobol.sourceforge.io/ Dec 20 03:09:10 GenTooMan: ! Dec 20 03:17:51 set_ set set? (short names allow for added redundancy and repetition) Dec 20 03:18:19 Hello! Dec 20 03:18:49 greetings for as kremlin would say in not some many words "e" :D Dec 20 03:18:58 I just wanted to let you know I can see that you are typing AND! Dec 20 03:19:25 I cross-compiled some nice SDK onto the BBB for using a "hardened," integrated servo. Dec 20 03:19:28 Neat, heh? Dec 20 03:19:42 set_ nothing bad happened? Dec 20 03:19:50 Now, I need to try their...no. Dec 20 03:19:53 Nothing bad happened. Dec 20 03:20:14 Nice! Outside of nothing bad happening. Well. Dec 20 03:20:37 I suggest testing a few things first. Dec 20 03:21:01 I tapped some piece of machinery into "stupidville." One of the tapped holes is larger than the rest and I have no room for nuts now. Dec 20 03:21:02 Blah. Dec 20 03:21:15 But, on the software side, I need to compile the extras too. Dec 20 03:21:28 The hello_world and GPIO and so on. Dec 20 03:21:57 My g-stop is not in yet plus I need to find one that can handle 24v. Dec 20 03:22:18 I think. Dec 20 03:22:46 Like the Replicape, this add on to handle the shear power of the servo is 24v to 75vdc. Dec 20 03:23:19 I have been looking to find some linear power supplies and have come up empty so far. Dec 20 03:23:31 It seems switching is not what is needed. Dec 20 03:24:34 I was thinking in b/t 24v and 75vdc, i.e. about 50vdc would be preferred. Dec 20 03:24:56 So, I do not go above the 75vdc and/or below the 24vdc. Dec 20 03:25:30 GenTooMan: What would you expect me to test first? Dec 20 03:25:49 Random motor tests? Dec 20 03:26:13 W/out the motor attached? Yes, I have thought about this idea and this is the first thing that will happen. Dec 20 03:30:32 I still need a couple of items for this motor to work as needed but clamping this motor down to test it w/ their set up software is going to be extraordinary. Dec 20 03:30:57 I cannot wait to hear it go cw and ccw. Dec 20 03:31:04 Zoom! Dec 20 03:33:25 Oh and GenTooMan, the global-stop might be able to be handled by 5v too. I just do not know how to handle their SDK yet. So, yea. I will need to keep reading their manuals and ideas on their hardware. Dec 20 03:34:34 GenTooMan: I wanted to tell you one thing too. Cross-compiling the sdk is done but the examples have not compiled correctly yet. Dec 20 03:34:48 I caught a few snags in the building of the examples. Dec 20 03:35:07 hmm you are building them on your PC under windows I assume? Dec 20 03:35:15 No. Debian. Dec 20 03:35:26 I am using Buster on the dev. desktop and on the BBB. Dec 20 03:36:10 I got handed a new one by asking why libstdc++.so.6 was different on Debian and Ubuntu. I had to switch to Debian to compile the sdk. Dec 20 03:37:37 I can put the dang examples on the BBB but first I will need to build them. I am getting stuck. GenTooMan: I will try to contact you later or if you have time now, I can boot the debian machine to get my errors. What say you? Dec 20 03:38:21 Heh? Dec 20 03:38:41 I see the time, too. Sorry for chiming in so later (as usual). Dec 20 03:39:08 20 minutes! Dec 20 03:40:28 GenTooMan: So, what do you think? Should I go to the debian machine and boot it to get the errors? I can then pastebin them? Dec 20 03:40:45 Actually, it is booted and eclipse is up. Dec 20 03:41:04 brb Dec 20 03:43:09 probably best to let everyone know, although if it's C++ ##c++basic might work better. I doubt it's a driver thing you are building. Dec 20 03:45:50 Hey. Dec 20 03:46:07 I got the SDK examples finally built. Dec 20 03:46:10 hello Dec 20 03:46:15 I needed to build them one at a time. Dec 20 03:47:06 I was building them under the .so file dir. and now I tried to build them one-by-one in another dir. and it worked! Dec 20 03:47:10 it worked! Dec 20 03:47:39 Computers? Sheesh. Just when you think they are good for nothing, they have surprises! Dec 20 03:47:47 Zoom! Dec 20 03:48:36 GenTooMan: The libstdc++.so.6 file needed that I built was libsFoundation20.so. Dec 20 03:49:19 renrelkha greetings and welcome Dec 20 03:49:53 I had to put the libstdc++.so.6 version from the build on the BBB. I am not sure why so far. But, they were the same .so.6, i.e. it had to be so I could use my BBB. Dec 20 03:49:55 Sheesh. Dec 20 03:49:59 This is why I am confused. Dec 20 03:50:06 sometimes the solution to the problem is to change the problem. Dec 20 03:50:11 Ha. Dec 20 03:50:21 The instructions fooled me. Dec 20 03:50:45 The instructions were for a '17 build of Ubuntu 16.04. Dec 20 03:50:54 I got lucky I could update them. Dec 20 03:51:15 Now, I just need to put the .elf files on the BBB. Dec 20 03:51:48 Then, then, then, plug it all in and hopefully pray for miracles! Dec 20 03:52:28 I never thought I would get so, pardon the pun, excited over Debian and motors. Dec 20 03:52:50 Get it, so, so, so? Dec 20 03:53:46 GenTooMan: I just do not understand why in the world I would need to put a completely new libstd++.so.6 lib. on my BBB. Dec 20 03:54:15 I mean, maybe back in '17, that was the way to do it. I would not know. Dec 20 03:59:23 A switch-a-roo is what I did and a build is what happened! Dec 20 03:59:36 Very satisfying. Dec 20 04:00:24 renrelkha: Hello! Dec 20 04:00:50 how goes it Dec 20 04:01:03 Well and not too shabby. Dec 20 04:01:15 good Dec 20 04:01:29 they probably update https://packages.debian.org/sid/libstdc++6 Dec 20 04:01:45 Yep. If you have not read all my funny jargon, I finally learned how to cross-compile things. Dec 20 04:01:45 then software depend on updated software Dec 20 04:02:02 It is five years in the making and just before the new year! Dec 20 04:02:15 exciting Dec 20 04:02:19 I know! Dec 20 04:02:37 I am fascinated w/ what can be accomplished. Dec 20 04:03:16 Some people just make their SDKs available w/ the Makefile format outside of building an entire cross-compilation. Dec 20 04:03:42 I finally needed to step up a bit and prove to myself I was not wasting my time. Dec 20 04:04:19 Are you still flying? Dec 20 04:04:51 i built new quad Dec 20 04:04:54 Nice. Dec 20 04:05:00 W/ what hardware? Dec 20 04:05:03 BBB? Dec 20 04:05:06 yeah Dec 20 04:05:09 Nice. Dec 20 04:05:47 I left the BBBlue alone for a bit b/c of all this rain we have been having. Dec 20 04:05:49 trying to make can peripheral Dec 20 04:05:57 Oh. Nice. Dec 20 04:06:03 with stm32 nucleo board Dec 20 04:06:15 The BeagleBone Black Primer has a short in it about CAN. Dec 20 04:06:27 Oh. Dec 20 04:06:34 i have the can working on bbb Dec 20 04:06:37 Aw. Dec 20 04:06:45 ardupilot handle the can side on that Dec 20 04:06:51 software Dec 20 04:06:58 Oh. Dec 20 04:07:00 Okay. Dec 20 04:07:19 just have to make sure interface is up Dec 20 04:07:24 I never got into CAN for some reason. Dec 20 04:07:28 and connected before starting ardupilot Dec 20 04:08:03 im not going to use the nucleo board long term but would be nice for prototyping or developing for devices Dec 20 04:08:10 I did find out how to get the compiled versions of ardupilot on the BBBlue before all this rain arrived. Dec 20 04:08:33 nice Dec 20 04:08:49 I have not been able to test out my new m5, left-handed lock nuts on the motors to hold the props on yet. Dec 20 04:09:14 So, I am a bit bummed out but it will get cold and I can then fly. For now, we are just receiving hot days and rain. Dec 20 04:09:58 all of my motors have same thread Dec 20 04:10:04 Really? Dec 20 04:10:05 use locktite Dec 20 04:10:08 Ha. Dec 20 04:10:20 * GenTooMan face palms. Dec 20 04:10:38 Finding the m5, left-handed lock nuts were so difficult. eBay? Yep! Dec 20 04:11:01 I got 10 of them w/ S & H for $2.00. I do not know how they do it. Dec 20 04:11:25 * set_ palms sand? Dec 20 04:12:08 I am just glad that Python is not ruining my life still. Dec 20 04:12:12 Ha. Dec 20 04:12:51 should they be threaded different? ha Dec 20 04:12:56 Yes. Dec 20 04:13:13 Two are cw and two are ccw. Dec 20 04:13:14 hmm python has certain issue. It's not bad it's just not great. Dec 20 04:13:31 right so the props don't come unscrewed? Dec 20 04:14:34 i never had a problem but i didnt shop motors much i need to get bigger ones anyway Dec 20 04:14:57 Does anyone understand how to create the button to stop the motor in source? I have tried everything. Python calls me right but then runs the source when the button is pressed which should stop the source from moving the motor any further. Dec 20 04:15:05 W/ Python? Dec 20 04:15:26 you need to do that one thing i think Dec 20 04:15:43 I thought a couple of while loops and some if statements would do it. Dec 20 04:15:48 Blah. Dec 20 04:16:36 I mean... Dec 20 04:16:47 I can get the button to turn the LED on and off. No issue. Dec 20 04:17:12 To stop the motor in a routine and act on the stop is another story, i.e. I think. Dec 20 04:18:00 B/c it would then need to go, stop, and resume in another direction which is attainable w/ the source I have currently. The set up is off. Dec 20 04:19:02 Motor(4500, 1000) <<< this is how to call action to the motor to step 4500 steps in 1000 microseconds. Dec 20 04:20:14 Right, so. I would then need to make this available before my while loop or call this in the while loop before the if statements that call action to the button press which is an input notified by a RISE. Dec 20 04:20:42 Clueless in Seattle? Blah. Dec 20 04:21:16 Let me show you the source if I can hash it up real quickly. Dec 20 04:22:49 I need to boot again. Brb. Dec 20 04:27:32 https://pastebin.com/0UyjsqAj does not do it just yet. I know these are small potatoes in source but I am failing to find what exactly needs to happen to stop and move back 50 steps in 1000 us when the button is hit. Dec 20 04:28:53 I'm not an expert at this but you want multiple states to keep track of what the motor is doing. States are just a reference ID for what the motor is "doing". For example stop starting run stopping would be states, Dec 20 04:28:55 That is just using the Adafruit_BBIO library for a GPIO and the MotorBridge.py library for the Cape in use. Dec 20 04:29:11 Okay. Dec 20 04:29:26 do I need to repeat it with flash cards? :D Dec 20 04:29:43 No. I got it. Multiple states. Dec 20 04:30:06 I only can put in two states b/c of the lib. I use. Dec 20 04:30:18 One is positive and the other is negative. Right? Dec 20 04:30:27 In words of one syllable no. Dec 20 04:30:34 Fine. Dec 20 04:30:35 Okay. Dec 20 04:30:50 So, how would I put in multiple states on one motor? Dec 20 04:30:59 I thought you didn't quite get it. So lets think about moving a motor. Dec 20 04:31:05 Okay. Dec 20 04:31:14 Let' us. Dec 20 04:31:16 when a motor is not moving it's what? stopped? Dec 20 04:31:22 Right. Dec 20 04:31:42 when a motor needs to accelerate to a speed it's what? starting Dec 20 04:31:47 Stopped, 0, off Dec 20 04:31:49 Right. Dec 20 04:31:50 Okay. Dec 20 04:31:56 Starting, 1, on. Dec 20 04:32:26 so what happens when the motor reaches speed? It's running Dec 20 04:32:32 Yes. Dec 20 04:33:03 what happens when you need to stop? It's de-accelerating so stopping. Dec 20 04:33:12 Okay. Dec 20 04:33:29 those are each unique states and are handled differently. Dec 20 04:33:37 Oh. Dec 20 04:33:38 Okay. Dec 20 04:33:41 starting accelerates Dec 20 04:33:49 stopping decelerates Dec 20 04:33:55 stop does nothing. Dec 20 04:34:08 running just stays at the same speed Dec 20 04:34:25 Okay. But, I do not have those available options on my lib. I have go, stop, reverse. Dec 20 04:34:40 no forward? Dec 20 04:35:00 well, depending on the what forward means, go represents that idea. Dec 20 04:35:24 So, on my slide, it can go backwards or forward depending on how the stepper is wired. Dec 20 04:35:38 Not the servo. Sorry. Dec 20 04:36:04 Let me show you the source for the .py lib. Dec 20 04:37:40 https://pastebin.com/vZTtybHA shows my options. Dec 20 04:40:18 Now...if the lib. needs to be altered, I might be able to accomplish this task. But, I need a starter. Where would I start? The firmware for the chip? It basically uses, like most things, pwm.h and gpio.h. Dec 20 04:42:01 Here: https://github.com/Seeed-Studio/MotorBridgeCapeFirmwareSourceCode/blob/master/Motor/motor_stepper.c Dec 20 04:42:26 That is the stepper source they plugged into the chip onboard the Motor Bridge Cape. Dec 20 04:42:56 So you have 2 things you can set direction and on off for stepping. Dec 20 04:44:11 Right. Dec 20 04:44:17 That is it. Dec 20 04:45:01 hmm how do you know the motor is moving? Dec 20 04:45:11 from the library? Dec 20 04:45:40 Well, I see the motor on the slide making the block on the slide move. Dec 20 04:46:05 I know it works. This is not the complication. It does work. Dec 20 04:46:36 I just cannot figure out how to install some type of python script to make my...oh. Dec 20 04:47:11 So, each input has to be able to handle each different, separate function from the lib? Dec 20 04:47:30 actually not you but the program how can it tell if it's moving? Dec 20 04:48:33 Hmm. I do not know. There are not any print statements or anything or any error messages. Is this what you mean? Dec 20 04:49:17 Feedback via source is what is needed? Is this what you are asksing me? Dec 20 04:50:22 It cannot. Dec 20 04:50:37 * GenTooMan hmms, "how do you know if the motor is stopped then? I mean you have to have something that indicates what the motor is doing for your code." Dec 20 04:50:47 Right. Dec 20 04:50:50 Okay. Dec 20 04:50:55 SOrry. Dec 20 04:51:10 I thought I gave you the source earlier. Dec 20 04:51:33 So, motor.ServoMotorBInit() starts the motor. Dec 20 04:52:14 motor.ServoMotorBMove(1000, 1000) makes the motor move 1000 steps in 1000 us. Dec 20 04:52:25 Oh. Dec 20 04:52:43 SO, motor.ServoMotorBMove(0, 1000) would be stop. Dec 20 04:52:51 Blah. Dec 20 04:52:57 I understand. Dec 20 04:53:48 and reverse would be motor.ServoMotorBMove(-1000, 1000) going in reverse, again (depending on the wiring), 1000 steps in 1000 us. Dec 20 04:54:00 Okay. So, I need to actually stop it. Dec 20 04:54:08 I need to stop it in the source. Dec 20 04:55:10 So, once the button gets hit, the init function does not get called since the motor is already in motion, the motor.ServoMotorBMove(0, 1000) gets called after the button gets hit. Dec 20 04:55:29 That seems like cheating. Dec 20 04:55:54 But, the damn machine only knows what I am telling it to do from what it already understands. So, okay. I will play along. Dec 20 04:56:30 I understand now. I was incorrect thinking that the GPIO could initialize the stop by itself. Yikes. Dec 20 04:57:21 I do not say it enough. Thank you. Dec 20 05:03:20 I just forced you to examine what was going on. Dec 20 05:03:53 * GenTooMan heads for bed, "not an early day but ..." Dec 20 05:06:27 No issue. Dec 20 05:06:29 Thank you. Dec 20 05:06:36 See you on another planet! Dec 20 06:31:47 I recently got the beaglebone blue. I have Linux working on it, but I've been having trouble powering it. Is it expected I can run it from a 2S 30C 1500mah lipo battery? It seems to power it for 30s before the board turns off. The board runs normally when plugged into 12V. **** BEGIN LOGGING AT Sun Dec 20 08:38:20 2020 Dec 20 08:40:02 bramp: I don't have a blue, but that battery sounds like it ought to be fine Dec 20 13:54:18 bramp: running for 30 seconds also seems odd... I wonder if unplugging the 12V unintentionally causes a power-off event Dec 20 13:54:33 bramp: can it be turned back on while on battery power? Dec 20 16:22:16 hi zmatt and bramp i posted on gitter but that battery should work unless it is dead, which can happen if power off but leave battery plugged in and have 5v devices connected Dec 20 16:22:44 zmatt shared schematic before to me regarding this issue Dec 20 16:23:06 maybe a napkin Dec 20 16:23:26 https://photos.app.goo.gl/6SyBtF3oUzcwn5G98 that one Dec 20 16:24:17 yep Dec 20 16:27:35 i use following method, use board on battery, plug in 12v until battery charged, disconnect battery, power down board with button, disconnect barrel jack Dec 20 16:28:14 2s lipo extension is handy Dec 20 16:30:31 unplug 12v when powered and on charged battery does not trigger power down Dec 20 16:31:08 ok, good. it shouldn't, but I know some kernels had some misbehaviour in that regard Dec 20 16:31:20 well i have not experienced it anyway Dec 20 16:31:25 but that might be ancient history Dec 20 16:32:09 actually, never mind, I'm a derp Dec 20 16:32:23 on the blue the battery and the 12V are indistinguishable to the PMIC Dec 20 16:37:11 I was thinking of the case where the PMIC's battery terminals are used (or I think it also used to happen when powering via both usb and the barrel and then disconnecting one) Dec 20 16:38:12 i dont power from usb much only to initiall connect to set up the wifi Dec 20 16:43:01 would be possible to write program that allow blue to take in analog video signal and overlay text on the video ? Dec 20 16:43:45 and output it as analog video on another gpio ? Dec 20 16:45:12 or would additional chips be necessary to handle the video Dec 20 16:45:46 if you only want to overlay text, I wouldn't capture the video at all but perform the overlaying with external hardware Dec 20 16:46:21 especially if it suffices to have black or white text, that just means holding the signal low or high respectively Dec 20 16:46:54 (signal or signals, depending on video format) Dec 20 16:48:35 and extract the sync from the incoming video and just feed that into the beaglebone Dec 20 16:54:23 I think PRU should be able to generate composite video, with (crappy) color even... especially since PRUSS can be clocked from the Display PLL allowing its clock frequency to be tuned to something appropriate (e.g. 196.875 MHz in case of NTSC, which is exactly 55 times the NTSC color burst carrier frequency) Dec 20 16:55:27 use a second PRU output to control a high-speed analog mux to switch between its signal and the incoming video you want to overlay onto Dec 20 16:57:12 some sort of sync extraction to a PRU input so it can position its output correctly onto each frame, and to be able to tune the display PLL to the correct frequency Dec 20 16:57:25 and a video driver on the mux output Dec 20 16:57:30 i have external device similar to this that can do it i just wondered if possible to do without additional hardware on blue https://hobbyking.com/en_us/minim-osd-for-apm-or-pixhawk-flight-controllers.html Dec 20 16:59:57 no Dec 20 17:00:50 i will save you comments there i would like to learn to make that kind of device one day ;) Dec 20 17:01:07 if you look at the diagram on page 1 of the chip they use: https://datasheets.maximintegrated.com/en/ds/MAX7456.pdf Dec 20 17:01:30 you can actually see exactly the structure I outlined: Dec 20 17:02:20 sync extractor, something that extracts the video timing from that, something that generates the text , a mux that combines it with the incoming video, and a video driver Dec 20 17:03:16 the beaglebone (more specifically PRU) could do the digital parts of this chip, but not the analog parts Dec 20 17:06:49 but unless you need the extra flexibility it would offer (e.g. being able to draw almost anything, as well as having crude colors), or just think it would be a fun project, I'd just stick with this chip that was purpose-made for it Dec 20 17:07:49 though you could connect it to the AM335x directly via SPI instead of going through the ATmega32P on this board Dec 20 17:08:40 hmm, the sparkfun breakout for this chip is also retired... https://www.sparkfun.com/products/retired/9168 Dec 20 17:09:30 i have HF?4067 multiplexer on breakout board Dec 20 17:09:50 HP4067 maybe hard to read Dec 20 17:10:30 i guess would be this CD74HC4067SM96 Dec 20 17:12:14 oh its CD74HC4067SM96G4 i guess Dec 20 17:16:11 I mean, the mux is the least interesting part :P using a 16-to-1 mux seems a bit excessive :P Dec 20 17:17:47 its what i already have on hand and is just for learning anyway Dec 20 17:18:34 like, if you want to do this it would make more sense to start with just attempting to use PRU to generate video, before bothering with the additional complication of overlaying it onto another video signal Dec 20 17:18:55 it's definitely not a simple project either way Dec 20 17:19:09 would more components be necessary Dec 20 17:19:21 video driver/buffer ic Dec 20 17:19:41 and possibly some analog filtering Dec 20 17:19:50 not sure, I think ds2 has some experience with this Dec 20 17:19:56 iirc Dec 20 17:20:10 its a bit over my head at the crrent time i think Dec 20 17:20:35 but is something i would like to explore in the future Dec 20 17:42:49 Thanks @zmatt and @renrelkha. I don't know what's up then. A couple things to note. This is a new lipo battery, and i tried to charge it with the beagleboard. Using a multimeter the battery has 7.8V, using the rc_battery_monitor it shows similar voltage. When i unplug the 12V connect, the 25% charge LED lights up. So I'm wondering if it thinks the battery is under charged and dies... Dec 20 17:43:35 Also, I have no additional sensors/devices connected yet. Just battery, and 12V. Dec 20 17:44:21 weird, 7.8V indicates plenty of charge Dec 20 17:47:59 none of my boards have had working or accurate battery level lights Dec 20 17:49:05 maybe it thinks battery is too dead Dec 20 17:49:36 I mean, the voltage measurement seems correct so why would it think that? Dec 20 17:51:00 i guess i am only familiar with low battery voltage threshold in software, not sure mthod used here Dec 20 17:51:08 but yeah, evidently something on the system is doing this... if you keep the system log open (journalctl -f) while unplugging 12V you may catch a log message of whatever is responsible for shutting down the system Dec 20 17:52:52 also find out if shut down or lop Dec 20 17:53:38 i assume protect circuit just shut off output to protect battery Dec 20 17:54:32 well, the fact that it runs for 30s suggests software shutdown, and the fact that the battery indicator seems to falsely imply the battery is low further hints at software being at fault Dec 20 17:55:08 good point Dec 20 17:55:16 And sorry, just to double check. This battery is only connected to the jst connector. While I understand that's for charging, I assumed it can also be used for powering. Dec 20 17:55:39 I'll try and capture journalctl as it shutsdown Dec 20 17:55:44 mine powers just fine from that without using barrel jack at same time Dec 20 17:55:57 or any other power source Dec 20 17:56:13 the battery header is both for charging and for powering the beaglebone obviously, it would be kinda useless otherwise Dec 20 17:56:48 Yes :-) I find so little information online, so I'm just double checking my assumptions :-) Dec 20 17:58:42 they are very sensitive to shorts Dec 20 17:59:53 renrelkha: hardware overdischarge protection is 2.5 V per cell Dec 20 18:00:18 so that's definitely not going to have kicked in here Dec 20 18:02:43 yeah if were that imbalanced i would expect battery failure Dec 20 18:03:26 the charger balances the cells Dec 20 18:03:37 is there out of balance protection tho? Dec 20 18:04:20 maybe the battery have bad balance from factory Dec 20 18:05:02 i guess easy way would be to check the voltage on balance connector Dec 20 18:05:02 if one cell were at 2.5V you could absolutely never have 7.8V total Dec 20 18:05:46 yes but some charge circuitry will not charge a battery that has 3.2v on one cell and 4.1v on another cell Dec 20 18:05:59 or whatever Dec 20 18:06:15 if large variance Dec 20 18:08:09 the balancer will disable charging if either of the two cells has an overvoltage condition Dec 20 18:09:27 at 4.3V .. which seems very uncomfortably high to me Dec 20 18:09:31 anyway, I don't think any of this is relevant Dec 20 18:09:48 ok Dec 20 18:09:51 7.8V indicates plenty of charge Dec 20 18:12:43 i am not ee but to me it mean one cell could be 4.2 and another 3.6 do i have fundamental misunderstanding? Dec 20 18:13:45 it could be, but 1. it is implausible and there's no reason to think this is the case 2. even if it were the case, it would be irrelevant for the problem at hand Dec 20 18:26:50 yeah i guess the mp2615 doesnt distinguish between cells Dec 20 18:29:16 sorry Dec 20 18:55:34 g Dec 20 19:00:39 is there a way for me to get a specific kernel and not just the latest one Dec 20 19:01:05 i am on 4.19.94-ti-r57 Dec 20 19:01:14 and I need r45 Dec 20 19:01:23 why exactly? Dec 20 19:01:31 but yeah, you can just install a specific kernel with apt Dec 20 19:01:52 sudo apt install linux-image-4.19.94-ti-r45 Dec 20 20:35:28 zmatt: Was messing with the panel thing for the ai. Dec 20 20:35:34 Lorforlinux sent me a video Dec 20 20:35:43 https://photos.google.com/share/AF1QipPgQUUt9Yz-VXaPSznXfgKxw7yeCjj-lamwGVAS43_VrqForHVaJ2BlhvfVLj2Qqg/photo/AF1QipMO5QEKVNMtM3gXvvHScj1nlwHc5Cli55Ww18Dp?key=c3VxRTBPejE5MDF4ZGtsSlhERFU2YnExQ2JfMFRn Dec 20 20:35:53 he is using a down level kernel Dec 20 20:36:03 so I was hoping switching back would fix Dec 20 20:36:12 it didnt =( Dec 20 21:23:47 GenTooMan: Sorry for misunderstanding yesterday about the on/off mechanism and input/output. Dec 20 21:24:22 On a nice note, the source from the SDK works but I have to rewrite some of it to work w/ Linux SBCs, e.g. THE BBB! Dec 20 21:27:20 set_ why is it that you have to rewrite stuff all the time Dec 20 21:30:07 Well. Dec 20 21:30:16 There are many reasons. Dec 20 21:30:25 ttyS2 is used a lot on my BBB. Dec 20 21:30:34 zmatt / renrelkha. I ran both rc_battery_monitor and journalctl. https://pastebin.com/QfDCx82x. The hightlights are, once switched to battery, the board lasted ~20second, and it was obvious the voltage dropped significantly (as reported by the rc_battery_monitor). I also measured each cell, they were 3.950v and 3.953V (with a total of 7.91) measured from the JST charger connector. Dec 20 21:31:14 So, does it just draw too much current from the battery, and the battery gives up? I'm not a EE person :) Dec 20 21:31:27 mattb00ne: Sometimes things change in python and sometimes things change in Linux. Dec 20 21:31:31 It really depends. Dec 20 21:31:43 Not everything is "cutting edge" current. Dec 20 21:32:17 bramp: you measured the battery voltage while the beaglebone was running off battery power? because this looks like your battery voltage drops quite a bit once the bbb is actually running off of it Dec 20 21:33:48 it also looks like you're not supposed to run rc_battery_monitor manually since evidently there's a service running for it :P though I don't know where you'd obtain the output of that service Dec 20 21:34:34 I measured with my multimeter when the battery is disconnected. While it's connected the only measurement I have is the rc_battery_monitor Dec 20 21:35:11 And yes, it looks like a service, but seems very happy to be run manually. I read over the source last night while trying to figure this out :-) Dec 20 21:35:41 So anyways, the drop in voltage is why the board is dying? Dec 20 21:35:43 do you have another load you can put on battery or smart external battery charger Dec 20 21:36:18 This stuff is all new to me, so I don't have anything else to really test the battery Dec 20 21:36:31 like, it makes sense that the system shuts down if it's measuring 6.46 V Dec 20 21:36:55 since that would normally imply your pack has virtually no charge left Dec 20 21:38:15 recharge and balance? Dec 20 21:39:05 The only charger I have is the beaglebone blue. Leaving it plugged in for a while, the lights indicate it is charging Dec 20 21:39:10 But maybe not Dec 20 21:39:20 it should charge when plugged in yes Dec 20 21:40:06 By 12v to barrel jack, the 2S charges. Dec 20 21:40:31 Or...is that by 5v? Dec 20 21:40:48 the blue's input is 12v Dec 20 21:41:05 Right...just checking. Dec 20 21:41:15 charge light comes on even with no battery attached Dec 20 21:41:21 Yes, I have a 12V power supply plugged in (coming from the wall). And the lipo plugged into the 2S jst port. Dec 20 21:41:31 Okay. Dec 20 21:41:39 Just checking. Trying to catch up here. Dec 20 21:41:57 No worries :-) I appreciate all help Dec 20 21:42:06 Does the BBBlue do balancing too? Dec 20 21:42:17 it does not appear to Dec 20 21:42:22 I think I should just buy a separate charger, maybe a separate battery. To rule some things out Dec 20 21:42:35 set_: yes Dec 20 21:42:37 battery is cheaper Dec 20 21:42:44 Did anyone ever get a photo of the set up? Dec 20 21:42:46 the blue has a balancer Dec 20 21:42:50 Right. Dec 20 21:42:57 That is what I thought. Dec 20 21:43:06 Well the voltage, as measured by my multimeter, shows each cell has the same voltage. So that would mean they are balanced? Dec 20 21:43:12 yeah Dec 20 21:43:27 Okay...is there an older image on your board? Dec 20 21:43:33 so, it appears the problem is voltage drop under load Dec 20 21:44:12 This is a photo of the set up: https://usercontent.irccloud-cdn.com/file/V5z1vv8E/PXL_20201220_214341176.jpg Dec 20 21:44:14 meaning either a problem with the battery, or a weirdly high resistance in the path from your battery to the point where the bblue measures it Dec 20 21:45:40 So, did we already see the output of the LEDs on the BBBlue when the 12v barrel jack connector is plugged in w/ the 2S? Dec 20 21:46:02 random unrelated remark, the antennas should be rotated outward like this: https://photos.app.goo.gl/tgXV8iQo3SQCrmdG8 (they're just rotated in over the pcb as a way to protect them during shipping) Dec 20 21:46:24 Oh! Thanks Dec 20 21:46:35 Yea. That will give a longer reach of WiFi too! Dec 20 21:47:11 yeah, having the antennas hovering over the pcb is not going to be great for reception Dec 20 21:47:34 ha i did not get to the balance charge part of diagram Dec 20 21:48:32 https://inst.eecs.berkeley.edu/~ee192/sp20/files/BeagleBone_Blue_sch.pdf Dec 20 21:49:03 Are we using the current image and does are we using librobotcontrol? Dec 20 21:49:26 we are talking about a battery Dec 20 21:49:53 or they are and im giving bad info i guess Dec 20 21:50:31 it doesn't seem likely a software issue would cause it to think the voltage is slightly lower than the actual voltage, but only while 12v is disconnected Dec 20 21:50:31 Right...I get that idea but the source for librobotcontrol has to be run as a .service, i.e. sudo systemctl start xxxx.service. Dec 20 21:50:48 Oh. Okay. Dec 20 21:51:08 a problem with the battery seems more likely Dec 20 21:51:11 https://photos.app.goo.gl/eUNVth6menM7tBwV6 is a video of the board being turned on. It lasts a minute this time before turning off Dec 20 21:51:18 Yeah, I agree it's a battery issues Dec 20 21:51:56 bramp: my guess would be the battery is being monitored by a service and that's what's causing the shutdown Dec 20 21:53:06 Hmm. That is odd. You something... Dec 20 21:53:22 It may be the way that specific battery is wired up. Dec 20 21:53:26 Let me check mine. Dec 20 21:53:28 so, if this battery just dips rather low under load even if it still has significant charge (e.g. high internal resistance for some reason) it might be fine to lower the shutdown threshold Dec 20 21:54:02 my theory is the battery bad drop voltage and the protection turn off power abruptly Dec 20 21:54:11 Oh and one thing to try too, just push the power button once it "dies." Dec 20 21:54:48 renrelkha: well, the voltage seemed to stabilize around 6.46 V ... that's way too high for the hardware protection to kick in Dec 20 21:54:53 I have Black, White, Red on my battery connector. Dec 20 21:55:02 In that order. Dec 20 21:55:28 mine is black black red Dec 20 21:55:56 your photo shows black white red Dec 20 21:56:02 Hmm. Dec 20 21:56:17 Okay, so, outside of theory and knowledge, some poking around may help. Dec 20 21:56:19 (or red white black, dunno which side is pin 1) Dec 20 21:56:38 i just looking at a battery pack Dec 20 21:56:41 Oh. Let me check to make sure, @zmatt. So, we can rule that out. Dec 20 21:56:42 regardless, I'm inclined to just put the blame on the battery pack Dec 20 21:56:59 set_: rule what out? I have no idea why you're talking about wire colors Dec 20 21:57:27 if the leads were in the wrong order, the board will not power on and quite possibly something will start emitting smoke Dec 20 21:57:39 I have a red wire closest to the outside of the board. Dec 20 21:57:41 Okay. Dec 20 21:57:50 So, that is ruled out. Dec 20 21:58:33 updating the bootloader Dec 20 21:58:45 that should like be kernel independent right ? Dec 20 21:59:11 mattb000ne: mostly yes Dec 20 21:59:41 does it complaining about something being mounted already on / make sense Dec 20 21:59:51 mattb000ne: what? Dec 20 21:59:51 i had not problem installing on r57 Dec 20 21:59:58 mount: /tmp/uboot: /dev/mmcblk1p1 already mounted on /. Dec 20 22:00:13 mattb000ne: wtf are you doing? Dec 20 22:00:20 So, are we sure we tested the DMM to LiPo correctly? BLK to BLK and RED to RED and RED to White? Dec 20 22:00:31 set_: please just stop talking Dec 20 22:00:35 Fine. Dec 20 22:00:38 Sheesh. Dec 20 22:00:40 Control! Dec 20 22:00:44 LOL Dec 20 22:01:02 taking another run at this Dec 20 22:01:03 https://deepaklorkhatri.me/GSoC2020_BeagleBoard.org/ Dec 20 22:01:22 mattb000ne: that message looks like you're trying to run something while booted from emmc that isn't meant to be run while booted from emmc Dec 20 22:01:24 so rather than installing the latest kernel i am using the one which he had panel working Dec 20 22:01:32 i c Dec 20 22:01:34 he has this specific panel working? Dec 20 22:01:39 this cape I mean? Dec 20 22:01:41 yes Dec 20 22:01:48 https://photos.google.com/share/AF1QipPgQUUt9Yz-VXaPSznXfgKxw7yeCjj-lamwGVAS43_VrqForHVaJ2BlhvfVLj2Qqg/photo/AF1QipMO5QEKVNMtM3gXvvHScj1nlwHc5Cli55Ww18Dp?key=c3VxRTBPejE5MDF4ZGtsSlhERFU2YnExQ2JfMFRn Dec 20 22:01:55 that is his video Dec 20 22:02:29 I can switch to SD card Dec 20 22:02:38 but it did not complain when i was on the latest kernel Dec 20 22:02:42 which is weird Dec 20 22:04:03 on second thought I don't think I have the energy right now, and I don't really want to watch a video Dec 20 22:04:13 lol Dec 20 22:04:17 fair enough Dec 20 22:04:24 sounds like the answer is to boot from the SD Dec 20 22:06:10 I will watch it! Dec 20 22:10:51 So, are you using the AI still for this Cape? Dec 20 22:11:16 trying to Dec 20 22:11:24 just to have it and mess around Dec 20 22:11:30 It seems the fellow has it working w/ a particular overlay w/ a specific release. Dec 20 22:11:33 my AI empire starts with the bbai Dec 20 22:11:38 Aw. Dec 20 22:11:47 that is what we are thinking Dec 20 22:11:50 so I am testing Dec 20 22:12:06 So, he shows which release he used in the video. This is a start. Dec 20 22:12:09 Oh. Dec 20 22:12:10 Okay. Dec 20 22:12:27 Is there a way to dictate which release you use? Dec 20 22:13:10 B/c...usually I just update the kernel to the latest release w/out thinking of all the goodies in b/t those releases. Dec 20 22:15:50 yes you can load a specific kernel Dec 20 22:15:54 I asked zmatt earlier Dec 20 22:16:04 now for whatever reason it does not like boot loader Dec 20 22:16:23 Aw. Dec 20 22:16:45 So, if you update the bootloader and revert back to older image releases, it may be an issue. Dec 20 22:16:56 I remember now. Dec 20 22:17:19 I gave @zmatt a couple "handfulls of glory" on this issue years ago. Dec 20 22:17:47 I remember one fellow new how to fix it. Dec 20 22:18:07 Me, I do not. Dec 20 22:18:25 reverting to earlier releases of the bootloader is not easy. Dec 20 22:33:46 I know the script for updating the bootloader cannot be put in reverse but it may accept args. Did you look in the file yet? Dec 20 22:34:05 I am booting now to check. Dec 20 22:40:12 500 lines of beauty and I am not sure I exactly understand how to "revert" to earlier revisions of the bootloader. It shows a lib. on github. This is a bonus in case you need to learn how to revert. Off to look at github-a-roo. Dec 20 22:43:59 Nope. I got a bunch of backslosh instead of good evidence on how to revert the bootloader. Dec 20 23:02:20 Whelp, that is all from me for now. I have gone googlie eyed again. Dec 21 01:28:22 SUCCESS!!!!! Dec 21 01:32:04 Kablooie? Dec 21 01:34:31 Images! Dec 21 01:34:46 I knew you could do it. Faith and trial by fire! Dec 21 01:35:04 AI, AI, AI! **** ENDING LOGGING AT Mon Dec 21 02:59:56 2020