**** BEGIN LOGGING AT Wed Jan 20 02:59:57 2021 Jan 20 03:09:02 mattb000ne: Get the LFS211. Jan 20 03:09:18 It will teach you how to network w/ superiority. Jan 20 03:09:45 Well, until the next level. Jan 20 03:09:58 I think they call it super-superiority. Jan 20 03:10:22 superduperiority Jan 20 03:10:51 Yes sir! Jan 20 03:11:09 You can run through their highlights and notes quickly, if careful! Jan 20 03:11:34 Let me see. I have it from years back. Some stuff has changed but it is still in working order. Jan 20 03:11:34 heh? Jan 20 03:12:13 It really depends on how people set up their distros. Jan 20 03:12:39 If you just plug and play, yep, things change. Jan 20 03:21:15 Hey GenTooMan! It is me, it is really me. Jan 20 03:21:16 i will give it a shot Jan 20 03:21:36 mattb000ne: Yea. They have some good print outs. Jan 20 03:22:16 Their end-of-'school' sessions have a list of commands one was supposed to learn in depth during the courses. Jan 20 03:22:41 That list of commands is super nice! Jan 20 03:22:46 ... Jan 20 03:22:49 for review...that is. Jan 20 03:23:50 What are they making you use instead of bind9? Jan 20 03:35:07 What exactly does unprivileged mean in the RISC-V world? Does it mean that one would not be allowed to use specific instances of its instruction? Jan 20 03:35:32 I am doing some catch up work. Jan 20 03:36:40 It seems that traps are okay (privileged) and their unprivileged worksheet is almost 300 *'s in depth. Heh? Jan 20 03:37:33 So, it is like they are saying less is allowed from the RISC-V instruction set. Jan 20 03:38:19 Is that good or bad? I am not sure exactly how this could affect the use of...oh. I keep forgetting coprocessors. Jan 20 03:38:20 Blah. Jan 20 03:52:54 It has 8GB of memory on it. What is a TF Card? Is the TF just what the micro Sd card fits into? Jan 20 03:55:50 This is like an entire new world to me. The ARM instruction set has been looked over by me. I cannot explain why or for how long. Jan 20 03:55:59 it's you for a moment anyhow Jan 20 03:57:03 Anyway, starting w/ the RISC-V instruction set might prove useful. I was unaware on the new board that there is a specific coprocessor for a neural network. How could you let me overlook it, GenTooMan? Jan 20 03:57:17 set_ TF - is TransFlash Jan 20 03:57:21 Oh. Jan 20 03:57:30 Trans. Jan 20 03:57:36 Okay. Jan 20 03:57:55 at least they didn Jan 20 03:58:11 err did not call it FlashUniversal or something. Jan 20 03:58:29 One thing says SD Card and one thing says TF Card. Jan 20 03:58:36 How do you boot this thing? Jan 20 03:59:08 Is SD Card equal to TF Card? Jan 20 04:00:18 oh. Jan 20 04:00:28 TransFlash is a Micro SD Card. Jan 20 04:00:32 Blah. Jan 20 04:00:42 Why, GenTooMan, why? Jan 20 04:00:42 Well anyhow :D Jan 20 04:00:46 Ha. Jan 20 04:00:52 because they knew you would be confused. Jan 20 04:01:00 Yep! Jan 20 04:01:09 Blue is red. Jan 20 04:01:11 they ARE out to get you no need to be paranoid. Jan 20 04:01:20 What is that noise? Jan 20 04:02:00 Did I forget to mention that the food at the Maker Faire '19 was ______. Jan 20 04:02:26 and you are still remembering it? Jan 20 04:02:29 Literally. Someone wanted everyone to die. I should show disgust but it will not help. Jan 20 04:02:32 Hahha. Yes! Jan 20 04:02:45 I should have known. Jan 20 04:02:51 so it was so bad it was memorable. Jan 20 04:02:56 Right. Jan 20 04:03:06 And... Jan 20 04:03:16 why do we always remember the bad things. Jan 20 04:03:34 Jambalaya does not make you hear verbal sounds of chimes. Jan 20 04:03:38 I promise I am right. Jan 20 04:03:56 I met the lady that was in charge of getting the food. Jan 20 04:04:01 After the faire. Jan 20 04:04:11 At the local library. I could have kicked her. Jan 20 04:05:10 I should have kicked her for being so naive. Jan 20 04:05:16 umm she went for Tom's used shoes for supply food? Jan 20 04:05:23 Probably. Jan 20 04:05:35 I think they gave her the garbage. '= Jan 20 04:06:16 well sometimes these things happen. You go with a vendor and they do a bad job. It's like a dice roll. Some places have a better reputation than others. Jan 20 04:06:32 GenTooMan: Did you stay up to date on the RISC-V stuff for the kernel? Jan 20 04:06:35 Either... Jan 20 04:06:41 Oh and yes, you are right. Jan 20 04:07:02 Sometimes people flawlessly fail at food production. Jan 20 04:07:13 I burnt a pizza once. Jan 20 04:07:19 right I remember I never ate at Chi'Chi's after getting food poisoning from them. Jan 20 04:07:35 they eventually went out of business. Jan 20 04:07:40 You too. Oh wait, ours in the south is called something else. Jan 20 04:07:48 Right. The one I went to ran out of biz too. Jan 20 04:08:10 Damn people! Jan 20 04:08:12 Boston Market has become BM Jan 20 04:08:21 The...ha. Jan 20 04:08:23 use to be good now it's just expensive. Jan 20 04:08:38 I rememember Boston Market in VA. Jan 20 04:08:46 It was good. Food. Jan 20 04:09:13 then they change CEO's and prices and began to shrink ever since. Jan 20 04:09:23 I thought I was going to call the Ambulance at the Faire. Jan 20 04:09:31 Yep. I get what you are sayin'. Jan 20 04:09:44 It was a close call. Jan 20 04:09:50 So you got a T-Shirt "I survived marker Faire!" Jan 20 04:09:54 No! Jan 20 04:09:58 NOt even a damn shirt. Jan 20 04:10:17 I got a laser cut Maker Faire emblem. Jan 20 04:10:42 I was like, really? Come on city folk! Where is my T? Jan 20 04:11:15 I am just glad no one got hurt. Jan 20 04:11:35 There were children playing and oddities in our brains. Jan 20 04:11:43 WE got very lucky. Jan 20 04:12:10 I was ready for 2020 w/ my own meal. Jan 20 04:12:17 Then...you know. Jan 20 04:13:33 Pandemic stuff happened. 1/3 of the people at work got it. Family got it. I got lucky. It was probably that damn jambalaya that saved me and my life. "Stay away and approach me w/ caution!" Jan 20 04:15:58 Aw...the good, ole days. Jan 20 04:16:08 * set_ sips ma' tea. Jan 20 04:16:45 well I'm not sure you would want a cursed shirt. So CV-19 visited. I find it odd though that their are an lot of variants already. It normally takes a bit longer for that. Also the spread is a bit strange too. Even with airline tavel way down. Jan 20 04:20:08 You are right. Jan 20 04:20:23 Why are things mutating? Jan 20 04:20:38 influenze has been here for years. Jan 20 04:20:48 Of course there are many alternate strains. Jan 20 04:21:28 Bird flu, swine flu, and etc. Jan 20 04:22:54 COVID is not well known around these parts...USA. It usually happens in countries where drainage is poor and water is scarce. Jan 20 04:24:52 COVID? that virus spreads significantly faster has the humidity goes down below 50%\ Jan 20 04:25:01 Plus...if indeed these mutations are correct/evident, how? How? I am not a microbiological engineer but cancer is similar. If you got it, mutations happen. Jan 20 04:25:26 I have read that too. Heat is a good way to kill it off. AUS should be okay and S. America too. Jan 20 04:26:30 Actually, in the local news and national news, 'Microsoft News', I have not seen any international stuff in a long time. Jan 20 04:27:04 that's because the "rest of the world news" has been suppressed. Jan 20 04:27:12 It was a popular thing for England to be in the news for a bit. Now, it is like nothing is being of an international concern when...right. Jan 20 04:27:17 It's called sedating the masses. Jan 20 04:27:21 Why is that happening right now? Jan 20 04:27:52 the US "news" isn't news at all didn't you realize that? Jan 20 04:28:01 I would think that it would be a good idea to see how others are handling things w/ COVID. Jan 20 04:28:06 Ha. You are right. Jan 20 04:28:25 Even the AP is bellowing low on this one. Jan 20 04:28:34 I might have to switch to Rueters. Jan 20 04:28:37 It's just a business they are trying to sell advertising and also entertainment. No news. Jan 20 04:28:53 You should at least you will know what is going on then. Jan 20 04:28:59 In the US Jan 20 04:29:22 Weird right. I get a lot of who is the next fashion model, who is popular on TV, and who is going to buy what and when. Jan 20 04:29:25 * GenTooMan started watching BBC for news on the US. Found a lot of things that weren't reported here going on. Jan 20 04:29:47 Yes all the unimportant things. Jan 20 04:29:50 Right. Jan 20 04:30:26 Things that hold value and are thought of as knowledgeable are gone in Microsoft, Facebook, and blah-blah.com for now. Jan 20 04:30:37 I get so upset. Jan 20 04:30:46 go to bed by the way. Jan 20 04:30:56 Where are my _____. OH yea. I forgot. Jan 20 04:31:00 Okay. Jan 20 04:31:04 snore! Jan 20 04:31:21 well if you don't discipline your body it will get back at you. :D Jan 20 04:31:29 * set_ has to deal w/ tomorrow again! Jan 20 04:31:51 GenTooMan: Although we might never meet, thank you for making me understand. Jan 20 04:32:22 Close and closure. Like in the military, they say, "Cluster ______________/beep!" Jan 20 04:32:32 and I am out! Jan 20 04:34:31 GenTooMan: Jan 20 04:34:43 where did you got for news Jan 20 04:36:22 virus mutated all the time Jan 20 04:36:29 any place but a US news source. Canada, BBC, Reuters, just any place but a US source. Jan 20 04:36:40 ok just checking Jan 20 04:37:11 that is probably a good way to go Jan 20 04:37:32 most of the US news is from a single source. That's why it's so bad. Jan 20 04:37:41 the reason we are in this mess now is rather than building stuff attention is worth money now Jan 20 04:38:07 If I could get 1 million people watch me taking a dump on youtube I would make more than any engineer Jan 20 04:38:13 that is what is wrong with america Jan 20 04:39:32 I remember talking to someone in Comp Sci. "You can make a lot of money in it", then I asked, "do you think you can enjoy it?" This child didn't understand work is too be enjoyed not endured. If you can't stand doing something find something else to do. Jan 20 04:40:09 their is a lot wrong with people in the US mostly it's in their heads. Jan 20 04:44:44 pardon my spelling I still have trouble with there and their. Jan 20 04:45:46 i am with you Jan 20 04:46:00 that is why i have changed careers like 5 times Jan 20 04:46:32 got to believe in what you do since you are going to spend a lot of time doing it Jan 20 05:14:48 how might I send out a signal on the spi_sclk pin? Jan 20 07:35:18 has anyone ever had their spi send the wrong data Jan 20 19:45:35 has anyone ever had issue with spi data not being received correctly? i.e. I load 167 into the Tx buffer and receive 135 on the beaglebone Jan 20 19:54:51 that's one bit flipped Jan 20 19:55:10 what speed are you running at? Jan 20 20:00:54 8 mhz Jan 20 20:00:59 on both devices, I believe Jan 20 20:02:01 are setup and hold times in spec? Jan 20 20:02:29 and levels, obviously Jan 20 20:18:07 what do you mean? Jan 20 20:37:44 https://www.ti.com/document-viewer/AM3358/datasheet/mcspi-master-mode-lit54013320 Jan 20 20:44:37 I'm using the AdaFruit SPI python library, with the frequency set to 8 mhz Jan 20 20:45:02 and the other device? Jan 20 20:46:21 other device is an STM32 F401RE with default clock configuration (16 MHz), but in the SPI register for that device, the quickest buad rate is the clock frequency divided by 2, so it should be 8MHz on that device as well Jan 20 20:50:58 and are the signals in spec? Jan 20 20:51:23 check with a scope or logic analyser Jan 20 21:19:36 https://gyazo.com/a07baf8e7343bc5d402f45ed2e4c19a2 Jan 20 21:19:53 what's that? Jan 20 21:19:54 This is all I get on the MISO pin on the slave device Jan 20 21:20:08 when the transfer is initiated Jan 20 21:20:28 this is data coming from the slave device Jan 20 21:20:55 Are the lines supposed to be like that Jan 20 21:22:16 what speed did you configure? Jan 20 21:22:27 8 MHz Jan 20 21:22:45 uh, check the scale there Jan 20 21:22:52 I set spi.msh = 8000000 on the beaglebone Jan 20 21:22:56 you need to blow that up about 1000x Jan 20 21:24:22 8 MHz means one byte is 1 μs, i.e. 0.001 ms Jan 20 21:24:38 holy crap this software is annoying. I can't zoom in without losing the marks Jan 20 21:25:27 also, what you said doesn't make sense... the number of bytes transmitted and received is inherently always the same with spi Jan 20 21:25:40 every clock cycle transfers 1 bit in each direction Jan 20 21:25:53 oh never mind Jan 20 21:25:55 I misread what you said Jan 20 21:26:11 somehow I thought you said you loaded 167 bytes, not the number 167 Jan 20 21:26:37 https://gyazo.com/f54c03161d51579210639c0b11dfef28 Jan 20 21:26:38 make sure both sides use the same spi mode Jan 20 21:26:54 yeah, I have it set to mode 0 on the beaglebone and 0 on the slave Jan 20 21:27:27 if I have spi.cshigh = true, should it be in mode 3 on both devices? Jan 20 21:27:30 and 8 MHz is definitely fast enough to get signal integrity issues depending on how you wire it up Jan 20 21:27:41 cshigh has nothing to do with spi mode, and should be false Jan 20 21:28:08 I had cshigh set to false, but the slave wasn't receiving any signals when it was Jan 20 21:28:14 (since the pin used for cs is high by default at power-on, so using it for an active-high chip select is ill-advised) Jan 20 21:28:19 what would be the ideal frequency? Jan 20 21:28:30 can the slave not be configured for an active-low chip select? Jan 20 21:28:38 I think it can, gonna check Jan 20 21:28:48 what sample rate is that logic analyser set to? Jan 20 21:29:00 8 mhz Jan 20 21:29:07 uh, that won't work Jan 20 21:29:08 that's useless Jan 20 21:29:14 it needs to be 80 MHz or more Jan 20 21:30:04 more generally, it should be a lot higher than whatever clock frequency you're using, especially if you were hoping to use it to investigate signal integrity (for which a scope is actually better suited than a logic analyzer) Jan 20 21:30:45 and you need to look at both clock and data lines Jan 20 21:32:19 and there's no "ideal frequency". typically your intended use will impose a minimum clock speed necessary, and the maximum possible will depend on the two devices and especially how you connect them Jan 20 21:33:02 it's usually best not to push too close to either boundary Jan 20 21:33:13 from the slave datasheet: "For devices set as slave, the NSS pin acts as a classical NSS input: the Jan 20 21:33:13 slave is selected when NSS is low and deselected when NSS high.". Also, I set the spi.cshigh to true because I had tested it, and for some reason it was high, even when I didn't modify it Jan 20 21:33:40 I have the NSS pin on the slave attached to the spi_cs pin on the beaglebone Jan 20 21:34:06 that sounds like cshigh should be false... assuming cshigh means "active-high chip-select" Jan 20 21:34:09 by testing it, I mean I plugged the wire coming from the spi_cs pin and it lit an led Jan 20 21:34:28 before I ahd cshigh = true Jan 20 21:34:33 yes that's correct Jan 20 21:34:49 active-low chip-select is the most common in my experience Jan 20 21:34:50 if cshigh is false, cs will be low during a transfer and high otherwise Jan 20 21:35:02 which is almost always how chip select works Jan 20 21:35:04 but when cshigh = false, the microcontroller doesn't respond, even though the microcontroller is at active low Jan 20 21:35:33 that sounds very strange Jan 20 21:45:49 yeah, having spi.cshigh = true sets the pin low when calling spi.xfer Jan 20 21:45:50 weird Jan 20 21:46:20 unfortunately it seems py-spidev didn't bother documenting the setting Jan 20 21:47:20 ok it directly corresponds to SPI_CS_HIGH Jan 20 21:47:25 which does mean active-high Jan 20 21:47:54 did you check with your logic analyzer? Jan 20 21:48:22 the only pin I've checked on the logic analyzer is the miso pin Jan 20 21:48:36 so how do you know what the cs pin does during the transfer? Jan 20 21:49:38 when it's connected to the led, the led turns off during the transfer Jan 20 21:49:51 you can see a led turning off for 1 microsecond? Jan 20 21:50:12 well no Jan 20 21:50:20 I'm gonna hook it up to the analyzer real quick Jan 20 21:50:32 connect all the signals Jan 20 21:50:44 and sample at 100 MHz or more Jan 20 21:50:50 or lower your speed Jan 20 21:51:16 also, are you using py-spidev or Adafruit_GPIO.SPI ? Jan 20 21:51:53 I lowered the speed to 1 mhz on both devices, and I still get 135 when sending 167 Jan 20 21:52:02 I'm using Adafruit_GPIO.SPI Jan 20 21:52:16 there's no speed to configure on the slave, the slave has no influence over speed Jan 20 21:52:57 depending on what frequency (and what value I send from the slave), I either get back 135, 195, 0 or 255 Jan 20 21:53:03 but I'm never sending those numbers Jan 20 21:54:09 again, you're sure you configured the right spi mode? since last time I asked you seemed configured about what spi mode is Jan 20 21:54:13 *confused Jan 20 21:54:57 yeah, I have the spi.mode = 0 on the beaglebone and the slave by default has cpol and cpha set to 0 Jan 20 21:56:05 are the device grounds connected together properly? Jan 20 21:56:14 do those settings have the same meaning on both devices? Jan 20 21:56:27 that's also a good thing to check Jan 20 21:56:55 and fire up the scope if you have one Jan 20 21:57:06 or, capture cs, sclk, and miso together on your logic analyzer, at a sample rate at least 10 times your spi clock speed Jan 20 21:58:10 maximum sample rate on the logic analyzer is 24mhz Jan 20 21:58:20 then you can't use it for this Jan 20 21:58:25 sure you can Jan 20 21:58:30 just configure your spi clock speed low enough Jan 20 21:58:37 then the problem might go away anyway Jan 20 21:58:51 22:51 < Zalymo> I lowered the speed to 1 mhz on both devices, and I still get 135 when sending 167 Jan 20 21:59:06 oh, I missed that Jan 20 21:59:47 okay, looks like Ima have to build a messy circuit Jan 20 21:59:57 so set the spi to 1 MHz and the logic analyser to max Jan 20 22:20:16 will I need diodes for the logic analyzer if I connect all the wires to a breadboard? Jan 20 22:21:15 ? Jan 20 22:21:31 first, does your logic analyzer actually have 3 or more channels? Jan 20 22:21:37 yes, it has 8 channels Jan 20 22:21:48 okay, then I have no idea what your question means Jan 20 22:21:53 I need to have the beaglebone and slave connected but then I need to connect it to the analyzer Jan 20 22:22:12 right? Jan 20 22:22:34 logic analyzers and scopes are designed to avoid influencing the signals being measured as much as possible... "look but don't touch" Jan 20 22:23:50 so no diodes then Jan 20 22:23:57 does the logic analyzer not have clamps or something on its leads? you don't really want to change your wiring setup significantly, especially not insert a breadboard (which will no doubt deteriorate the signal integrity) Jan 20 22:24:54 all the analyzer has are male extension headers (it's a cheap usb logic analyzer) Jan 20 22:30:05 Hi everyone. I have a beaglebone with 2GB eMMC, I believe. I'm trying to flash a 2gb image to a 4gb sd card and it seems to continuously fail Jan 20 22:30:07 any ideas? Jan 20 22:30:24 do you use balanea etcher Jan 20 22:30:27 yes Jan 20 22:30:34 what 2gb image? Jan 20 22:30:47 first i'd like to say that I'm a newbie to beaglebone Jan 20 22:30:49 and what's the failure reported by Etcher? Jan 20 22:31:01 same Jan 20 22:31:12 from https://beagleboard.org/latest-images, the newest 2GB I could find Jan 20 22:31:13 Debian 7.5 (BeagleBone, BeagleBone Black - 2GB SD) 2014-05-14 - more info - md5: 35877ce21e8ed0eb1bdc6819ad71c317 Jan 20 22:31:19 that's extremely obsolete Jan 20 22:31:26 whoa, 6 years ago Jan 20 22:31:30 is there a newer 2gb flasher image i could use? Jan 20 22:31:53 My problem was that, as a newbie, I tried to flash a 4 GB image to the 2 GB eMMC ;-( Jan 20 22:32:02 so now I'm just trying to get something that'll work Jan 20 22:32:31 which type of beaglebone are you using Jan 20 22:32:32 the console images from https://elinux.org/Beagleboard:Latest-images-testing are tiny, though also minimalistic so they're intended for users who are comfortable with just installing whatever programs they actually want Jan 20 22:32:43 alternatively you can just wipe eMMC and boot from sd card instead Jan 20 22:33:27 I would not recommend using an ancient image just because it happens to (barely) fit in eMMC Jan 20 22:33:37 zmatt: if I reduce the clock speeds below 1 mhz, will connecting to a breadboard be acceptable Jan 20 22:35:19 Cool thanks! Let me try this real quick:) Jan 20 22:35:30 Zalymo: I mean, depends on the purpose. to verify that the slave has the same understanding of what cpha/cpol mean as the beaglebone, yes Jan 20 22:36:00 to investigate potential signal integrity issues with however you currently have your devices connected? no, since you'll be changing how you have the devices connected Jan 20 22:36:24 but confirming the first thing is probably a useful first step Jan 20 22:37:06 and since you had the exact same result on 1 MHz as on 8 MHz it doesn't really sound like a signal integrity issue Jan 20 22:37:27 also you never answered 22:56 <@zmatt> are the device grounds connected together properly? Jan 20 22:37:43 (as in, the ground of the bbb and the ground of the microcontroller) Jan 20 22:37:58 are the ground of the slave device and the beaglebone supposed to be connected to eachother? Jan 20 22:38:11 could be mismatched clock polarity Jan 20 22:38:12 I don't have any ground connection between them Jan 20 22:38:17 *facepalm* Jan 20 22:38:24 wait Jan 20 22:38:35 I even listed that explicitly in the "quick tips for not frying your hardware" yesterday Jan 20 22:38:43 well yes but no hardware has been fried Jan 20 22:38:55 was it yesterday? or whenever Jan 20 22:39:08 so I need to have a ground connection between them? like take a wire and plug it into both ground pins on each board? Jan 20 22:39:21 how the hell else would anything work? Jan 20 22:39:28 ground should be the very first connection you make between two devices you intend to interconnect Jan 20 22:39:36 ohh Jan 20 22:40:03 (with some rare, very specific exceptions you don't need to worry about) Jan 20 22:40:18 GND! Jan 20 22:41:00 My ufw firewall was creating issues w/ my camera. Oops. @zmatt is right again. My fault, i.e. as usual. Jan 20 22:41:04 there's a reason a voltmeter has two leads.. voltages are relative, there's no such thing as a line being "3.3V", it's 3.3V relative to _something_ ... and if not mentioned, that something is ground Jan 20 22:41:26 so if two devices don't have a shared ground, they don't have a shared meaning of voltages Jan 20 22:41:43 which makes communication pretty arbitrary Jan 20 22:42:06 thanks zmatt, trying to flash my BBB with that 1 GB image now Jan 20 22:42:44 I'd like to install a custom linux image, from Mathworks Simulink, to use Embedded Coder for BBB Jan 20 22:43:11 still happening >:| I really thought that would fix this Jan 20 22:45:49 is there a proper way to shutdown beaglebone black if I can't ssh to it to shutdown -h now Jan 20 22:45:49 ? Jan 20 22:46:05 press power button Jan 20 22:46:44 thanks! Jan 20 23:01:28 how does 167 come out to 63!? how does 200 end up as 192?!? Jan 20 23:02:24 200 vs 192 is one bit flipped Jan 20 23:02:42 63 or 163 ? Jan 20 23:03:01 163 would be one bit off from 167 Jan 20 23:03:20 putting 167 into tx outputs 63 on the beaglebone Jan 20 23:03:30 anyway, captured it with a logic analyzer yet? Jan 20 23:04:30 working on that now Jan 20 23:19:39 https://gyazo.com/423992ebd66cda1d8171848a898c35ff Jan 20 23:19:44 this is what each segment looks like Jan 20 23:20:23 I want to see the signals, not some interpretation thereof, I'll interpret them myself thank you Jan 20 23:20:29 wtf? according to the miso bits, the data is correct Jan 20 23:21:12 https://gyazo.com/20f77b086a74f04bec8bc14b6dcb1cda Jan 20 23:21:16 raw signals Jan 20 23:21:41 zoom in on the action please Jan 20 23:22:03 also wtf is that chip select doing at +1100 Jan 20 23:22:57 https://gyazo.com/3ee987ebce6ab2d0fa0e1013f6b68bd2 Jan 20 23:23:13 oh 1100 us, I feared it might be volts Jan 20 23:23:24 lol Jan 20 23:23:24 speaking of frying hardware Jan 20 23:24:27 is the receiver sampling on rising clock edge? Jan 20 23:24:42 should be, if it's configured to mode 0 Jan 20 23:25:57 so then something is wrong with the beaglebone configuration Jan 20 23:27:33 you're transmitting 0x01 ? Jan 20 23:27:36 on the beaglebone Jan 20 23:27:42 yes Jan 20 23:27:45 when using xfer Jan 20 23:29:05 this does look like both sides are configured as mode 0 Jan 20 23:30:02 so I'm not sure why you're still receiving garbage... perhaps there are still signal integrity issues (may be hard to detect with a logic analyzer) Jan 20 23:30:30 though it's weird if you keep getting the same wrong values Jan 20 23:30:53 but the logic analyzer is receiving the proper signals, and it's still getting the same value of 63 even with the messy wiring I had to do to hook it up to the logic analyzer Jan 20 23:31:05 yeah, that part is really strange Jan 20 23:31:16 so I'm sort of convinced that the beaglebone is not interpreting the signals properly Jan 20 23:31:39 yes but *why* Jan 20 23:32:20 because it thinks it can get away with it Jan 20 23:32:23 we use both spi buses of the beaglebone in our applications, at high speeds, so I know there's no problem with that on the beaglebone side itself Jan 20 23:34:50 https://gyazo.com/6a7d37bd87158c0618d41d9e7c7379c0 Jan 20 23:34:53 meant to show this as well Jan 20 23:35:09 maybe the empty transfer is the reason? Jan 20 23:35:12 yeah that's just weird as hell Jan 20 23:35:13 no it's not Jan 20 23:36:23 py-spidev is doing that intentionally, rudely Jan 20 23:36:35 presumably as workaorund for a driver bug on some platform Jan 20 23:36:46 but also, why does it think this is CS_HIGH mode? Jan 20 23:36:52 even though the cs is clearly active-low Jan 20 23:37:19 never trust a snake Jan 20 23:38:10 I definitely do not set the CS_HIGH flag in my own code Jan 20 23:38:34 and our chip selects are (as always really) active-low too Jan 20 23:40:12 here's the bit I presume is responsible for that brief chip select assertion: https://github.com/doceme/py-spidev/blob/master/spidev_module.c#L605-L610 Jan 20 23:40:31 yeah, I have the cshigh set to false, turned out it was not working right because I had a certain spi mode set on the slave device Jan 20 23:40:41 ah ok Jan 20 23:40:42 hmm Jan 20 23:40:52 then this bit shouldn't be it Jan 20 23:41:36 then I don't really get where that's coming from Jan 20 23:42:16 magic I guess Jan 20 23:43:22 or maybe spi.readybytes(1) isn't working right Jan 20 23:43:33 readbytes() Jan 20 23:44:34 uhh, I thought you said you were using xfer, not readbytes ? Jan 20 23:44:45 I'm using xfer and readbytes Jan 20 23:44:55 first xfer, then readbytes Jan 20 23:45:10 why? regardless, the transfer we're looking at on the scope is the first one (i.e. the xfer) Jan 20 23:45:32 since I can see it's transmitting 0x01 (while readbytes will transmit 0x00) Jan 20 23:56:37 oh yeah, I forgot to mention that when reading the rx buffer in the slave device, the value isn't 1 as it should be Jan 20 23:56:45 maybe that has something to do with it Jan 21 00:00:22 nvm, it is correct **** ENDING LOGGING AT Thu Jan 21 02:59:57 2021