#qi-hardware IRC log for Saturday, 2015-02-07

DocScrutinizer05wpwrak: ((battery is a bit tricky at the moment)) thought about asking? ;-)11:48
wpwrakhmm ? all i do is available for review and i document my progress in quite some detail :) for example, here's the current battery issue: http://lists.en.qi-hardware.com/pipermail/discussion/2015-February/010811.html12:24
wpwrakbasically OLED's Vih(min) being quite high, higher than i tought, and i somehow missed that until now. so i need to rearrange the voltage domains.12:26
wpwrakbut first i'm checking whether my batteries are really still okay. wouldn't do to spend a lot of effort to solve problems that are just caused by dead batteries :)12:32
xiangfuwpwrak: Hi. I just uploaded some pics [brd-back-close.jpg  brd-back.jpg  brd-front-1.jpg  brd-front.jpg  brd-two.jpg  brd-with-UBB-back.jpg  brd-with-UBB-front.jpg] to here: http://downloads.qi-hardware.com/people/xiangfu/anelok/tmp/13:34
xiangfuthis time is small files. 13:34
xiangfuI double make sure. the wire connect is correct. there is no short.13:35
xiangfuI think there are something wrong with my nanonote. trying the ubbctl now..13:35
wpwrak(photos) whee, nice ! thanks !13:36
wpwrak(nanonote) oh, did you disable the mmc driver ? echo jz4740-mmc.0 >/sys/bus/platform/drivers/jz4740-mmc/unbind13:37
wpwrakand if your kernel has an atben driver, then also  cd /sys/bus/spi/drivers/at86rf230; echo spi*.0 >unbind13:37
wpwrakokay, kernel and user space both trying to use the port at the same time may explain some issues :)13:38
wpwrak(pics) you have front and back reversed in the file names :)13:39
xiangfudownload the nanonote toolchain for compile ubbctl13:45
wpwrakhmm, and you may have placed too many of the resistors around the boost converter. e.g., if you're using the AAT1217-3.3, R4-R6 should be DNP. it looks as if they're populated13:45
wpwrakor is this just an optical illusion ?13:48
xiangfunothing on F4-R613:54
xiangfuwpwrak: http://dpaste.com/0VN8S2514:18
xiangfuubbctl + swd.14:18
wpwrakDAT0 looks unhappy14:25
wpwrakseems to be shorted to ground in some way14:25
xiangfuhmm... I am sure the DAT0 means SWD_CLK is not shorted at UBB or Anelok.14:28
wpwrakthe DAT0=1!0  says that the ben is trying to output a 1 but it reads back a 014:29
wpwrakbut you can check this by setting the outputs manually (with ubbctl) and measuring14:31
wpwrakand you may also want to check the supply voltage. if there's a problem somewhere, it may be too low, but still not low enough to just crash the ben as well14:32
wpwrak(though shorts usually crash the ben. so one has instant feedback ;-)14:32
xiangfuone of the board : the SWD_CLK between GND have 60ohm14:32
xiangfuHere is another board output: http://dpaste.com/0VMV98X14:35
wpwrak60 ohm is quite low. if you short the probes via the board, how many ohm do you get ?14:35
wpwrakin any case, 60 ohm is too low. i measured for both SWD_CLK and SWD_DIO abour 7.8 MOhm if + is on SWD_* and 0 on GND, and about 27 kOhm if - is on SWD_* and + on GND14:38
xiangfuhmm... wait. 14:39
xiangfuthis device is not easy to understand. I borrow a Agilent U1252B.14:39
wpwrakresults may be different with different meters (that was with a picotest m3500a, alas not the most exact device but my fluke is busy watching the batter discharge), but anything lower than at least a few kOhm is suspicious14:40
wpwraksimilar to joerg's critter :)14:40
xiangfu./ubbctl dat0=0: give me 0.0091v14:42
xiangfu./ubbctl dat0=1: give m1 0.5xxxv14:42
xiangfu./ubbctl dat0=1: give m1 0.65xxv14:43
xiangfu./ubbctl dat0=1: give m1 0.5xxx ~ 0x65xxv14:43
xiangfuthis one ./ubbctl dat1=1: give me 3.3V. I like 3.3 :)14:44
xiangfuU1251B is expensive...14:45
xiangfuwpwrak: have you saw my last message:14:46
infobotARGH!!! STOP IT xiangfu!!!14:47
xiangfu./ubbctl dat0=1: give m1 0.5xxx ~ 0x65xxv14:47
xiangfubut ./ubbctl dat1=1: give me a very stable 3.3V.14:47
wpwrakyes, 3.3 V is good. 0.65 V suggests that there's a short. apparently not directly with ground but with some other pin15:01
wpwrak(u1252b) yeah, these agilent/keysight meters are a bit overpriced in my opinion. i'd get a good lab meter. not portable, but much nicer to use in the lab. and you don't have to worry about batteries ;)15:02
wpwrak#s/lab meter/bench meter/  that's the correct term15:03
wpwrakthe good ones are even more expensive, but if you're going to spend many hundreds of dollars on a meter, you may as well go the extra mile and get a really good one15:04
xiangfuYEAH: http://dpaste.com/24MHQ0715:04
wpwrakthe 0.65 V may be a short between DAT0 and some pin that outputs a "0". maybt try CLK=1 and see what happens15:05
xiangfuhmm... he sugest me this one: F179C .15:05
xiangfuFLUKE. expensive too. :)15:05
wpwrakif it's not that, things will get a bit more complicated. you could then test all the other pins of the KL26 and see if there's another one that has 0.65 V15:05
xiangfuwpwrak: it is the my taobao UBB soldering problem.15:06
xiangfufirst try: http://dpaste.com/2TQCYQE15:06
Action: xiangfu again15:06
xiangfusecond try: http://dpaste.com/2F2K93R15:07
wpwrak(fluke) maybe this ? http://www.testequipmentdepot.com/fluke/dmm/8808a.htm15:07
xiangfuflashed: http://dpaste.com/0JKC8ZH15:07
xiangfuwhat should I do next?15:08
wpwrak(fluke) i have this one (the 8808a didn't exist back then): http://www.testequipmentdepot.com/fluke/dmm/8845a.htm15:08
wpwraklemme check ...15:08
wpwrak(flashing) excellent ! it's normal that it takes a few tries15:08
wpwrakyou could now connect USB and see if it enumerates15:09
wpwrakwhen the led goes on after you connect usb, it's probably happy15:09
wpwrakah, and i would disconnect the ben first. just in case :)15:09
wpwrakand if it enumerates, you can then "make dfu" to flash the firmware15:11
wpwrakafter flashing the firmware, the led will go off. you'll then have to wait for about 30 seconds and then you should be able to operate the slider15:12
xiangfuconnect usb: nothing happen on kernel msg15:12
wpwrakwill be a bit tricky without a case but you should at least get some response15:12
wpwrakokay, not so easy then :)15:12
wpwrakwith usb connected, please check that the voltages are right: especially that VBUS > GND :)15:13
wpwrak(there is a big VBUS pad right next to the usb connector)15:13
wpwrakvbus should be roughly 5 V. if it's very different, somethign is wrong.15:14
xiangfuVBUS: -0.15v15:17
wpwrakhmm. are you sure you got the right type of conenctor ? :)15:19
wpwrakyours don't look reversed15:19
wpwrakin a reversed connector, the two little slots should be at the bottom15:20
wpwraknot sure if there are reversed connector that have them on the other side15:20
wpwrakare your connectors AB or B ?15:21
xiangfuhmm... -4.5V15:23
wpwrakhow to identify a reversed connector: the little plastic inside the connector should be close to the PCB, with the contacts facing away from the PCB 15:23
wpwrakthat looks even more like the wrong connector15:24
wpwrak(reversed = what anelok needs)15:25
xiangfuYes. wrong connecter...15:25
wpwraklet's hope you have a good electronic shop nearby :)15:26
wpwrakwell, at midnight, it may not be open, though. or do you have all night electronics shopping in china ? :)15:26
xiangfunot in Beijing. maybe in Shengzhen.15:28
wpwrak(regarding the usb connector, i hope to eventually replace this with a regular - non-reversed - connector. but so far i haven't been able to find a nice way to do this. layout gets very tricky.)15:28
wpwrakokay, what we can do now is ... a) wait until you can get right kind of connector, b) you can try to make a reversed cable, c) you could try to flash the application via SWD. but that's a bit messy because SWD is so unreliable and the application is larger than the boot loader15:31
xiangfulet me try the SWD.15:32
xiangfuwhat is the command for reflash teh application?15:33
wpwrakfirst you have to construct a combined binary15:33
wpwrakthis should do the trick: dd bs=16k conv=sync if=boot-2014.bin of=combined.bin; cat anelok-2014.bin >>combined.bin15:35
wpwrakthen flash with    swd combined.bin15:36
wpwrakhmm. my statistical estimates suggest that it may take something like 80 attempts before swd will work for somethign this large, given that it took you three attempts with boot-2014.bin. well, you can use a while loop :)15:39
wpwrakwhile ! ./swd combined.bin; do : ; done   # then take a tea or a nap ;-)15:40
xiangfua lot of access error15:43
wpwrakyes, there's something wrong but i don't know what. i already spent a lot of time trying to figure out what may be the problem :-(15:43
xiangfufinished one.15:44
wpwrakmaybe a fresh pair of eyes will find it in minutes ... ;)15:44
xiangfusuccessed one of them15:44
wpwrakexcellent ! :)  now .. swd -r to power and reset anelok15:48
wpwrakand then we'll see if it works :) (i've never tried this mode of operation myself)15:48
wpwrakyup, should work. just booted one of mine with  ./swd -r15:51
wpwrakah, but now i have to leave. people trickling in.15:52
xiangfuyes. now I am doing 'while'... :)15:52
xiangfuhave to sleep.15:54
--- Sun Feb 8 201500:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!